Voluntary Employee Withdrawal and Inattendance - A

Voluntary Employee
Withdrawal and
Inattendance
A Current Perspective
INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY:

THEORY, RESEARCH, AND PRACTICE
Robert D. Smithers, Series Editor, Rollins College, Winter Park, Florida
VOLUNTARY EMPLOYEE WITHDRAWAL AND INATIENDANCE
A Current Perpsective
Edited by Meni Koslowsky and Moshe Krausz
A Continuation Order Plan is available for this series. A continuation order will bring
delivery of each new volume immediately upon publication. Volumes are billed only upon
actual shipment. For further information please contact the publisher.
Voluntary Elllployee
Withdrawal and
Inattendance
Edited by
Meni Koslowsky
Bar-Ilan University
Ramat Can, Israel
and
Moshe Krausz
Bar-Ilan University
Ramat Can, Israel
Springer Science+Business Media, LLC
ISBN 978-1-4613-5151-1
ISBN 978-1-4615-0599-0 (eBook)
DOI 10.1007/978-1-4615-0599-0
2002 Springer Science+Business Media New York

Originally published by Kluwer Academic / Plenum Publishers, New York in 2002
Softcover reprinl of Ihe hardcover 1sI edilion 2002
http://www.wkap.nl/
ro 9 8 7 6 5 4 3 2 1
A c.I.P. record for this book is available from the Library of Congress
AII rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in
any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher, with the
exception of any material supplied specificalIy for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work.
Contributors
Helena Mensah Addae
University of the West Indies, St. Augustine Campus, Trinidad & Tobago
Vishwanath V. Baba
Michael G. DeGroote School of Business, McMaster University, Hamilton,
Ontario L8S 4M4, Canada
Julian Barling
School of Business, Queen's University, Kingston, Ontario K7L 3N6,
Canada
Gary Blau
Human Resource Administration Department, Temple University, Fox
School of Business Administration, Philadelphia, Pennsylvania 19122
Yair Amichai Hamburger
Department of Psychology, Bar-Han University, Ramat-Gan, Israel 52900
Kathy A. Hanisch
Iowa State University, Ames, Iowa 50011
David A. Harrison
Department of Management and Organization, Pennsylvania State
University, University Park, Pennsylvania 16802
Muhammad Jamal
Department of Management, Concordia University, Montreal, Quebec
H3G 1MB, Canada
Gary Johns
Department of Management, Concordia University, Montreal, Quebec
H3G 1M8, Canada
v
vi
Contributors
Rabindra N. Kanungo
Faculty of Management, McGill University, Montreal, Quebec. H3A IG5,
Canada
E. Kevin Kelloway
Department of Management, Saint Mary's University, Halifax, Nova
Scotia, B3H 3C3, Canada
Meni Koslowsky
Moshe Krausz
Manuel Mendonca
Faculty of Management, McGill University, Montreal, Quebec H3A IG5,
Canada
Abraham Sagie
School of Business Administration, Bar-Han University, Ramat-Gan, Israel
52100
Caroline Weber
School of Industrial Relations, Queen's University, Kingston, Ontario,
K7L 3N6, Canada
Preface
When we started editing this volume, the world seemed to be a worker's

heaven: the high-tech industry was flourishing then and the "New
Economy" seemed to have the leading role in the business world.
Moreover, the burgeoning need for new recruits created a labor market
where the employees had the upper hand. Job candidates and new
recruits often dictated the salary and benefit packages that they wanted
whereas employers-start-ups as well as more established companieshad to yield to their demands. R&R, i.e., recruitment and retention, were
the major concern of many companies due to the disparity between need
for, and supply of, qualified and often even unqualified applicants. In
addition to their constant demands throughout the recruitment process,
employees were also involved in voluntary turnover, leaving one company for the next one, sometimes foregoing a thorough search of the available alternatives and, frequently, having more than one offer at hand.
This state of affairs has taken a marked downturn starting shortly
after the beginning of the new millennium and exacerbated by the terrorist attacks in New York City and Washington. The world has undergone a
deep economic crisis affecting, in particular, high-tech where burgeoning
numbers of employees are losing their jobs across the Western world. The
power balance has shifted toward organizations that now have wider latitude in dictating salary and benefits to job candidates as well as restricting
their current employees. As a result, employees are more likely to stick to
their current jobs or, at least, to be much more timid and cautious when
contemplating mild as well as more severe forms of voluntary withdrawal
such as late arrival to work and full-day absences, let alone leaving the job
and the organization, topics that form the core of the current volume.
In addition, other questions and doubts are raised about the "New
Economy" and the outcomes of globalization. Verbal, as well as physical,
attacks on the concept of globalization have just culminated in the murderous terrorist attacks in the United States, which the heinous perpetrators as well as various more objective observers have attributed to the
negative outcomes that globalization has had on poorer parts of the
vii
viii
Preface
world. During the current phase of the business cycle, companies can dictate the terms of employment; the pressure is on to cut many jobs and
severely curtail the choice now available to job searchers.
To borrow a term that has been one of the key concepts explaining
employee turnover-perceived ease of mobility (March & Simon, 1958)labor markets have moved, within one year, from a high and often insatiable quest for new workers to large scale downsizing, with some
companies near bankruptcy. This situation has created fewer labor market
alternatives for those beginning their careers as well as for "old-timers"
wishing to make a job change or midlife transition. The reader must be
cognizant of the fact that these recent developments will undoubtedly
have an effect on the issues dealt with in the current volume. The question may, therefore, be raised as to whether voluntary withdrawal behavior is at all an important issue under these circumstances. Moreover,
whereas economic and labor market status reflect situational trends in the
demand and supply of employees, affected by cyclical fluctuations, the
world has for quite some years been undergoing changes that appear to
be more enduring and stable.
A related question raised by Harrison in his chapter concerns the
present status of the milder or temporary forms of withdrawal, namely,
lateness and absence. Are they still viable and important organizational
measures in an era where work communication may be considered as
"continuous," i.e., not bound by time or distance? Thus, working at home,
on the road, or over the Internet does not require an office and is not limited to the hours between 9 and 5.
While current knowledge does not provide firm answers to such
questions, it is our belief that the various forms and modes of withdrawal
behavior will continue to prevail, though possibly on a somewhat smaller
scale. First, good workers and performers are still a premium and are still
in demand (Donovan, 2001) even during these times. Second, for the average worker, even in labor markets that are indeed more risky and uncertain for employees, individuals will continue to be absent voluntarily and
ponder the possibility of finding a "better job," even if chances of doing
so are slim. Such behaviors and attitudes are just natural reactions and
are, to a large extent, independent of the economy. For example, the oftenquoted turnover model by Mobley, Griffeth, Hand, and Meglino (1979)
does include job alternatives as an antecedent but also incorporates
within it many individual difference variables and company-related
measures. Thus, psychological reasons such as job dissatisfaction or the
need to alleviate the stress associated with working on dull and unsatisfying jobs would be expected to explain a large amount of turnover
variance. In addition, people will still arrive late or take days off, perhaps,
Preface
ix
less frequently and less conspicuously than before so as not to jeopardize

their present employment or to prevent other less-drastic employerinitiated sanctions. Although it may be expected that throughout difficult
labor market periods, such as currently prevail, the traditional measures
of withdrawal will decline, other forms of withdrawal may very well
develop (e.g., high frustration with a job that one cannot easily leave may
be conducive to psychological problems, lack of concentration and motivation, and, in the extreme, more violence and abuse in the workplace).
As for the communication technologies that allow employees to be
"absent" but also at the same time to work individually or collaboratively,
it should be noted that the benefits as well as the original forecasts about
the future of teleworking or telecommuting appear to have been exaggerated and the number of employees working from home is far less than
predicted. In fact, even companies that have favored such arrangements
are more sensitive to some of their shortcomings (Kurland & Bailey, 1999)
and seem to be shying away from them, "retreating" to the more traditional form of face-to-face work communication.
In conclusion, we contend that employee withdrawal will continue to
be an important research and practical issue despite transitory and longrange environmental changes.
A brief review of the various chapters will illustrate quite clearly the
wide gamut of issues represented here. The order of chapters follows, in
general, a progression from theoretical presentations at the beginning to
more empirical and practical applications toward the end.
In the first chapter, Sagie, Koslowsky, and Hamburger present a multilevel model of lateness. Unlike Blau, there is no distinction between
chronic and incidental lateness, nor is length considered. Instead, the
authors emphasize the usefulness of analyzing lateness in wider contexts-individual, organizational, and cultural-and draw the effects on
each of those circles. Mensah Addae and Johns (Chapter 2) claim that in
cross-cultural absence research both individual- and cultural-level
absence legitimacy are distinctive criterion measures since absence is
influenced by factors beyond employees' control. The model posits several variables that influence absence legitimacy perceptions such as work
centrality and time perceptions. Also, the meaning of time varies by culture. Additional cultural variables relevant to absence legitimacy are locus
of control, gender role differentiation, and social support systems. The
authors assert that by examining cross-cultural factors, the role of social
influence on absenteeism is clarified and permits generalizations beyond
single site studies which are constrained by specific and unique factors.
In Chapter 3, Krausz outlines some newer forms and unexplored
antecedents of employee turnover behavior, in particular considering
Preface
turnover as part of an individual's preplanned career program, regardless

of work satisfaction and not contingent on unsolicited job opportunities.
It is also suggested that turnover may be an outcome of other employees'
behavior or of nonwork changes such as relocation of a spouse. The
author argues that more attention should be given to intraorganizational
job mobility where employees move to another unit or branch. Such
moves are similar to extraorganizational turnover since they may involve
meaningful losses to the employee him- or herself, to colleagues, to supervisors, and to entire organizational units or branches.
In Chapter 4, Kanungo and Mendonca discuss three types of withdrawal: avoiding or reducing membership behavior such as tardiness,
absenteeism, and turnover, avoiding or reducing performance behaviors
such as work restriction and social loafing, and avoiding extra role behaviors which are not part of the role itself. All of these withdrawal forms
hamper innovative behavior and reduce the organization's effectiveness.
The authors propose that withdrawal is actually a manifestation of
employee powerlessness and weak commitment to the job and the organization. Each of these three types represents a relief from tension-ridden
environments. Kanungo and Mendonca outline the performance management process-a strategy to reduce employee withdrawal, enhance
mutual trust, and increase commitment.
In Chapter 5, Harrison contends that although lateness and absence
differ quantitatively from each other, both of them differ qualitatively
from turnover, thus justifying a distinction between two types of withdrawal-temporary and permanent. In addition, the author proposes
studying turnover as a multiple- as opposed to a single-occurrence variable with individual differences describing a continuum of propensity to
turnover. In addition to the organizational definitions, Harrison offers a
workflow or psychological definition of both lateness and absence, which
do not necessarily agree with the organizational definitions. A co-worker
may consider a fellow employee as arriving "late" because of work overload rather than as a result of a violation of organizational rules. In the
same manner, lateness or absence may be considered a 360-degree phenomenon as seen by suppliers, customers, and superiors regardless of the
organizational definition ("1 don't care what your company thinks, you
were too late for my needs and plans").
Blau contributed two chapters (6 and 7) to this volume. The first of
these is a reprint of the now-classical paper on employee lateness which
David Harrison described (Chapter 5) as the "data set [that] probably
contains the most meticulous and time-sensitive data measurement in the
literature. Few lateness researchers can hope to have the detailed records
he was able to exploit, if they can get archival data at all." Blau developed
Preface
xi
and tested a lateness taxonomy consisting of three types-increasing

chronic, stable periodic, and unavoidable/random lateness. By breaking lateness into three categories based on the nature, frequency, and duration of
lateness, attitudinal measures such as involvement and commitment were
found to be stronger predictors of one category-increasing chronic,
especially when compared to the findings from previous studies where
lateness was examined as a unitary construct.
In Blau's second chapter, the author contends that although there has
been only limited research on lateness since the 1994 paper and despite
the new employment options, empirical evidence suggests that lateness
continues to be an important "time abuse behavior, albeit for specific samples. The author highlights the importance of multilevel studies on lateness and of incorporating more than one withdrawal form in the context
of a single study. Such approaches may raise organizations' interest in
lateness studies, particularly if the definition of lateness is expanded,
turning it into a more costly occurrence than under current definitions.
Kelloway, Barling, and Weber report in Chapter 8 the results of a
meta-analysis comparing absence rates of smokers and nonsmokers.
Overall, the data support the hypothesis that smokers are more frequently
absent than nonsmokers. Those differences held across various country
groupings. The authors, while acknowledging the mediating role of
employee health, which guided their hypothesis, contend that other considerations such as the relatively high smoking rates among younger and
older employees may also be a factor in explaining absence variance. The
authors believe that smoking may account for an even higher proportion
of nonworking time by its impact on partial absence such as time taken
off for smoking breaks. While partial absence appears to be less harmful
than full-day absences, the accumulation of many short-duration absence
events can eventually be quite costly to an organization.
In Chapter 9, Baba and Jamal present an empirical study of job involvement and absence. The authors contend that although the reasons for
absence are multilevel-individual, group, or organizational-nevertheless,
the attitudes of individual employees' are still central. Their findings show
that job involvement is a direct predictor of two absence measures: time lost,
an indicator of involuntary absence, and frequency, a measure of voluntary
absence. Their main finding is that psychological and behavioral constraints
moderate the association between job involvement and absence. Work (e.g.,
job scope) and nonwork (family roles) constraints seem to weaken the control that individuals have over their work behavior. Empirical support is
given to the conceptual distinction between controllable/voluntary and
uncontrollablelinvoluntary absence. A low-constraint environment will
reduce absence among job-involved employees.
xii
Preface
Finally, in Chapter 10, Hanisch presents an empirical study that

focuses on some of the implications of the turnover process duration.
Specifically, she deals with the time lag between initial turnover thoughts
and actual departure for preturnover behavior. Unlike most turnover
studies, the current research deals with various forms of withdrawal,
rather than the more common focus on a single form in the context of one
study. Attitudes and behaviors prior to quitting differed between individuals who left in a rather quick process compared to those for whom the
process was much longer. The research suggests that the longer one thinks
about quitting before making the break from the organization, the more
detrimental are the effects on the organization. The data also allowed the
author to test some models of withdrawal.
ACKNOWLEDGMENTS We would like to thank each of the authors for their
contribution to this volume as well as for their patience. We also acknowledge helpful financial support from the Research Authority of Bar-Han
University, Ramat Gan, Israel. Assistance in preparing the manuscript for
publication was provided by Elie Lebowitz.
REFERENCES
Donovan, A. (2001, April 22). StiU hearing ') quit' at work. The New York Times,
p.4(L).
Kurland, N. B., & Bailey, D. E. (1999). Telework: The advantages and disadvantages
of working here, there, anywhere, and anytime. Organizational Dynamics,
5,53-68.
March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley.
Mobley, W. H., Griffeth, R. w., Hand, H. H., & Meglino, B. M. (1979). Intermediate
linkages in the relationship between job satisfaction and employee turnover.
Psychological Bulletin, 86, 493-522.
Contents
Chapter 1 ................................................
Antecedents of Employee Lateness: A Multiple-Level Model
Abraham Sagie, Meni Koslowsky, and Yair Amichai Hamburger
Chapter 2 .......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
21
National Culture and Perceptions of Absence Legitimacy

Helena Mensah Addae and Gary Johns
Chapter 3 .......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
The Many Faces of Voluntary Employee Turnover:

A Multifacet and Multilevel Perspective
Moshe Krausz
Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
Employee Withdrawal Behavior: Role of the Performance

Management Process
Rabindra N. Kanungo and Manuel Mendonca
Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
95
Meaning and Measurement of Work Role Withdrawal:

Current Controversies and Future Fallout from Changing
Information Technology
David A. Harrison
Chapter 6 ....................... . . . . . . . . . . . . . . . . . . . . . . . . .
133
Developing and Testing a Taxonomy of Lateness Behavior

Gary Blau
Chapter 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
161
New Conceptualizations of Lateness since Blau, 1994

Gary Blau
xiii
xiv
Contents
Chapter 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
167
Smoking and Absence from Work: A Quantitative Review

E. Kevin Kelloway, Julian Barling, and Caroline Weber
Chapter 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Job Involvement and Absence: The Role of Constraints as
Moderators
Vishwanath V. Baba and Muhammad Jamal
179
Chapter 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
193
The Timing of Thinking about Quitting: The Effect on

Job Attitudes and Behaviors
Kathy A. Hanisch
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
213
Voluntary Employee
Withdrawal and
Inattendance
1
Antecedents of Employee Lateness
A Multiple-Level Model
Abraham Sagie, Meni Koslowsky, and
Yair Amichai Hamburger
Lateness behavior can be described as arriving at work after the scheduled

time (Shafritz, 1980). Typically, lateness is addressed in the literature as a
withdrawal behavior, a category that includes absence and turnover as well
(Herzberg, Maunser, Peterson, & Capwell, 1957; Mobley, 1987). Compared
to the latter behaviors, lateness is a milder form of withdrawal; not every
case of late arrival can be considered a violation of company rules. Another
distinctive characteristic of lateness is that typically it is a volitional behavior, or, at least, within the employee's control. Whereas worker dismissal is
an involuntary type of turnover, and certified sickness or funeral attendance
can be classified as involuntary absence (March & Simon, 1958; Sagie, 1998),
most of the lateness occurrences can be avoided. Indeed, although some
causes of lateness (e.g., health problems and difficulties in transportation;
Koslowsky, Kluger, & Reich, 1995) are not made intentionally, nevertheless,
as Blau (1994) has indicated, proper time management may reduce tardiness
even when the immediate cause is not within the employee's direct control.
For the organization, lateness has many financial and nonfinancial
costs; it may negatively affect the performance of the late employees as well
as the entire unit's productivity (Blau, 1994; Sagie, Birati, & Tziner, 2002;
ABRAHAM SAGlE School of Business Administration, Bar-Han University, Ramat-Gan,
Israel 52100
MENI KOSLOWSKY AND YAIR AMICHAI HAMBURGER Department
of Psychology, Bar-Han University, Ramat Gan, Israel 52900
1
Abraham Sagie et al.
Steers & Rhodes, 1978). In addition, late employees are likely to reflect low
motivation, dissatisfaction at work, and low commitment to the organization (Hanisch & Hulin, 1990), which may diffuse among their co-workers
(Jamal, 1984). From the late employee's point of view, lateness behavior
may indicate an initial withdrawal from work that may deteriorate toward
more severe forms of withdrawal such as absenteeism and turnover
(Herzberg et al., 1957; Koslowsky, Sagie, Krausz, & Dolman-Singer, 1997).
Additionally, management responses to late arrivals may include the
recording of the late occurrences in the individual's file, disciplinary
reviews, and outright punishment. In view of such responses, lateness
could indeed be considered a costly behavior for the involved employee(s).
Being a volitional, generally unexcused, dysfunctional behavior for
both the self and the organization, one may ask why the employee does
not avoid tardiness behavior, and why frequently he or she consciously
chooses to come late to work. As these questions relate to the individual
level of analysis, the answers provided in the literature generally adhere
to this level. Most of the explanations concentrate on the individual's personality (Blau, 1994; Knatz, Inwald, Brockwell, & Tran, 1992) or on his or
her work relationships, experiences, and expectations (Hanisch & Hulin,
1991; Koslowsky et al., 1997; Steers & Rhodes, 1978). Detailed explanations focusing on the individual are addressed below. The overall aim
of the present chapter is to suggest that individual-level variables are part
of a multiple-level model that also identifies and describes various
antecedents at the group (e.g., work-team), organizational, and extraorganizational (e.g., national culture) levels. The primary notion is that the
process of lateness behavior and its antecedents would better be understood if one considers all of these levels of analysis.
Figure 1 shows the proposed multiple-level model; causal linkages
within the same level of analysis are depicted by horizontal arrows connecting antecedents to outcomes (i.e., lateness behavior). Vertical arrows
represent causal linkages leading from a higher level (e.g., organizational)
to a lower one (e.g., group). Table 1 presents the antecedents at each level.
The discussion will be presented first for each level separately and then
for the links across levels. In addition to describing the model and its multiple levels, we will provide here the results of a survey of the causes of
lateness that were gathered from a sample of managers. The fit between
the empirical data and the underlying model will be assessed.
INDIVIDUAL LEVEL
In the literature, there are five main categories of factors used to explain
the individual-level lateness behavior: personality traits, organizational
Employee Lateness
Extraorganizational
Factors
....
Ex traorganizationalLevel
Lateness Behavior
,It
Organizational
Factors
Organizational-Level
Lateness Behavior
,It
Group
Factors
-"
Group-Level
Lateness Behavior
,It
Individual
Factors
Indi vidual-Level
Lateness Behavior
Figure 1. A multiple-level model of lateness behavior.

Note: Causal linkages within the same level of analysis are depicted by horizontal arrows
connecting antecedents to outcomes. Vertical arrows represent causal linkages leading from
a higher level to a lower one.
attributes, work-related situational factors, nonwork situational factors,

and commuting, a combined work and nonwork factor (see the first column of Table 1).
Several researchers (Blau, 1994; Richard & Slane, 1990) considered
lateness or tardiness to be a personality disposition, the opposite of punctuality or promptness. Blau (1994) argued that punctuality is an underlying personality construct affecting behavior in both work and nonwork
situations. Whereas some workers are always punctual and arrive at work
on time or, sometimes, even before the scheduled time, certain others
Table 1. Multiple-Level Antecedents of Employee Lateness

Variables at the
Individual level
Personality dispositions
Role, status, power
Work-related situational
factors (e.g., voluntary
absence)
Nonwork situational
factors (e.g., work-family
conflict)
Commuting
Group or organizational
level
Extraorganizational
level
Rules and norms

Reward system
Leader's self-example
Work arrangements (e.g.,
flextime)
Organizational change
(e.g., downsizing) or
crisis
Industry type, ownership

National culture
Cross-cultural dimensions:
sequential! synchronic,
individualism/ collectivism,
and power distance
Labor market conditions
Political situation
exhibit either stable levels or, for various reasons, increasing chronic lateness behavior. The latecomers may "cut things too close," i.e., feel that
they have all the time in the world to get to work, and will be quite surprised when the train already left at the station or the traffic jam turns out
to be worse than expected. Blau's analysis clarifies that the delay time is
not the only component that should be taken into account when one calculates the impact of mismatched personalities on attendance behavior.
Consider a meeting of prompt and tardy participants. Whereas the former
may enter the meeting room a short time (e.g., 15 minutes) before the
scheduled time, the latter will be there some time (e.g., half an hour) after
the appointed time. Hence, the punctual person would wind up waiting
during the period between both arrivals (in our example, 45 minutes).
A personality dimension that is especially relevant to attendance
behavior is the Type A behavioral pattern. Baker, Dearborn, and
Hamberger (1984) found that Type A individuals arrive earlier for
appointments than their counterparts, Type B individuals. Also, Burman,
Pennebaker, and Glass (1975) have reported that Type A persons judged
the lapse of 1 minute sooner than did Type B persons. In a study by
Levine and Bartlett (1984), the authors found a moderate, though significant, relationship between Type A behavior and self-reported measures of
punctuality. One specific component of Type A pattern, time urgency,
holds considerable promise as a potential predictor of lateness behavior.
Time urgency can be defined as working at an "accelerated pace"
(Burman et al., 1975; Landy, Rastegary, Thayer, & Colvin, 1991). It is the
tendency to consider time as a scarce resource and to plan its use very
carefully. An individual with a high level of time urgency always feels
that he or she does not have enough time. Indeed, Bercovits (1996) found
that time urgency significantly discriminates between never-late and late
Employee Lateness
workers. Workers who showed higher levels of time urgency had fewer
lateness occurrences at work.
An entirely different approach to promptness and lateness in the
workplace considers these behaviors as a result of one's differential social
or organizational resources (e.g., role, status, prestige, and power) rather
than as an outcome of personality. From an organizational point of view,
the employee's working time is not considered to be under his or her
exclusive control, but is considered an organizational resource. Further,
the cost of employee time is not constant across hierarchy levels. Senior
workers' time is believed to be more expensive than that of junior workers. Hence, it is typically accepted that higher-ranked members (e.g.,
superiors) are allowed to control the time of lower-ranked workers.
Consequently, in many organizations, different informal rules define the
entitlements and obligations of superiors and their subordinates with
regard to punctuality and lateness behaviors, in general, and, more specifically, when both parties meet with each other. The inequality among the
various rules reflects the diversity of roles, power, and status in the workplace. In Table 2, we propose an informal protocol of conduct for two
roles: boss (or superior) and subordinate.
Table 2. An Informal Protocol of Attendance Behaviors for Superiors and
Subordinates in the Organization
Subordinate
Superior
The subordinate's time is money; hence,

his/her attendance at work is controlled
by strict regulations.
When they have an appointment, the
subordinate is expected to precede the
superior at the meeting room. He/she
should not come late to the meeting. If
he/she does, a detailed excuse is
required.
Normally, the meeting takes place in the
superior's room. If, however, the
subordinate is the host, he/she
must stop performing current
activities (e.g., interview, phone call,
document writing) before the
meeting's onset.
The subordinate's right to initiate
unscheduled appointments is limited.
The superior's time costs more money;

nevertheless, his/her attendance at work
is controlled by loose regulations.
When they have an appointment, the
superior is allowed to arrive after the
subordinate at the meeting room. In fact,
he/she is expected to come late; in such a
case, no or only a general excuse is
required.
If the meeting takes place in the superior's
room, the subordinate has to wait (in or
outside the room) until the superior
completes current activities (e.g.,
interview, phone call, document writing).
In the case of a pending appointment,

the subordinate has to wait until
he/she is called in.
The superior's right to initiate

unscheduled appointments is generally
honored.
The superior is typically allowed to
initiate a pending appointment that will be
opened on his/her request.
Based on Table 2, one may see that different attendance rules control
the superior and subordinate during their meeting. The rules related to
the superior are more generous and those related to the subordinate are
stricter. The subordinate is expected to precede the superior at the meeting room. He or she should not come late to the meeting; if he or she does,
a detailed excuse is required. On the other hand, the tardy superior does
not need to provide a detailed explanation; often a few general words are
adequate. If the meeting takes place in the superior's room (as normally
happens), the subordinate has to wait in or out of the room until the superior finishes doing the current activities, such as interview, phone call, or
document writing. Conversely, if the meeting takes place in the subordinate's room, he or she must stop doing the current activities before the
meeting's onset.
Indeed, these differential rules are not universal. As described in
detail below, the organizational culture and the norms of the surrounding
society may affect the discrepancy in attendance rules between the
superior and subordinate. Additionally, the quality of the superiorsubordinate relationship and, particularly, their mutual trust (Bauer &
Green, 1996; Dienesch & Liden, 1986) are inversely related to the divergence between both sets of rules. Similarly, the higher the subordinate's
commitment to the organizational goals, the higher his or her performance, and the more opportunities for delegation, empowerment, and
participative decision making offered by the superior to the subordinate
(Sagie & Koslowsky, 2000), the smaller is the discrepancy in their respective attendance rules. The rationale is that achieving better relations
between both parties or some resemblance in their authorities (e.g., both
are involved in making work decisions) is incongruent with a large gap in
the protocols applying to their attendance and lateness. It appears, therefore, that various measures at the dyadic, organizational, or extraorganizational levels may interact with role, status, and power, and then
influence employees' lateness behavior. In the following sections, this
issue is described in further detail.
According to a personality-based framework, as well as an organizational resource one, it is difficult to attribute either an aim or an external
cause to one's lateness behavior. Alternatively, however, Blau himself
(Blau, 1985, 1986, and the reprinted chapter in this book) explained that
lateness might be a result of work-related situational variables. Rather
than internal (personality) causes, the focus here is on the causes or aims
of one's behavior, primarily concerned with his or her relationship with
the job or the company. Because of opportunities (e.g., important meeting,
proximity of raise discussions) the worker may come earlier to work; conversely, because of difficulties in the work setting (e.g., poor interpersonal
Employee Lateness
relations, unfair supervision, low salary, boredom, chronic stress,

burnout; Jamal, 1984; Koslowsky et al., 1995), the worker may come late.
In an attempt to integrate the diverse lines of thought, we propose the
following scheme. Whereas personality and differential organizational
resources account for interpersonal differences in lateness behavior, the situational factors primarily account for intrapersonal variations in lateness
behavior. Thus, a tardy worker typically arrives later than his or her punctual colleague; however, even for the tardy employee, lateness behavior is
not uniform. Frequently, this person may come on time to the workplace,
especially if he or she expects disciplinary sanctions from the supervisor,
if a significant meeting has been scheduled, or a very important client is
expected to arrive.
Out of the different situational antecedents, the literature emphasizes
situational variables linking lateness with withdrawal from the
work. Although there are conflicting views (see Hill & Trist, 1955, and
Nicholson & Goodge, 1976, for an alternative, compensatory approach),
the bulk of the lateness research indicates that lateness is typically a signal of withdrawal and that the late employee inclines to exhibit other
types of withdrawal behaviors including volitional absenteeism and/or
turnover (Koslowsky et al., 1997; Sagie et al., 2002). Also, he or she is likely
to reflect psychological withdrawal, including dissatisfaction at work
and low organizational commitment (Beehr & Gupta, 1978; Sagie, 1998).
Based on a meta-analysis, Koslowsky et al. (1997) found that lateness
correlates positively with several other withdrawal behaviors (e.g.,
voluntary absence and turnover intentions) and negatively with performance, job satisfaction, and organizational commitment.
Several situational variables that are rooted, at least partially, in the
employee's nonwork life (e.g., home, health) may influence lateness and
other withdrawal behaviors. Family-work conflict (Greenhaus & Beutell,
1985) may be particularly relevant for lateness. Demands stemming from
the individual's role as an employee (e.g., arriving at work on time) may
conflict with demands in the family domain (e.g., taking care of young
children in the morning). When individuals are involved in juggling work
and family demands, it could be expected that from time to time they
would not manage to arrive for work on time (Blau, 1994). Bercovits
(1996), for example, found that employees with younger children arrived
later for work. The author argued that during the morning hours the
demands of a very young child are greatest and a parent must decide
between providing for the child and leaving for work at a reasonable
time. Hence, the interplay between time resources allocated to work
and nonwork activities may account for one's attendance behavior at
work.
Abraham Sagie et at.
While discussing the antecedents of absence, Johns (1997) reviewed

several individual-level nonwork variables that could be relevant as well
to lateness. These include medical reasons (e.g., absence and lateness rates
would rise in cases of sickness), lifestyle variables (e.g., lateness increases
with smoking or drinking), misconduct or deviant behavior (e.g., unexcused time lost would positively correlate with other deviant behaviors
and negatively with altruism and organizational citizenship), and economic considerations relating either to the employee (e.g., withdrawal
increases if one's nonlabor income rises) or to his or her family (e.g., the
secondary income earner in a family could "afford" to be late more often;
Blau, 1985, 1995).
Finally, one of the most important individual-level factors associated
with one's work as well as nonwork life is commuting. Koslowsky et al.
(1995) have posited that commuting related stressors such as traffic jams,
difficulty in parking, or car breakdown may affect work behaviors including performance, absenteeism, and lateness. However, research on the
commuting-lateness relationship is not conclusive. Leigh and Lust (1988)
found that lateness increased as the commuting distance became greater.
In contrast, Nicholson and Goodge (1976) found that lateness correlated
negatively with commuting distance. This could be explained by the fact
that one allocates spare time for long but not for short commutes.
Consequently, waste of time due to traffic jams and irregular public
transportation could not be compensated for in the case of short
commutes.
Drawing on these arguments, we framed the following propositions:
Proposition 1: Personality traits, differential organizational resources, and three types of
situational factors-work-related, non work, and commuting-account for the variance in
individual-level lateness.
Proposition 2: Personality traits and the organizational differential resources primarily
account for interpersonal differences in lateness behavior; the three types of situational
factors primarily account for intrapersonal variations in lateness behavior.
GROUP AND ORGANIZATIONAL LEVELS
As displayed in Fig. 1, the immediate level beyond that of the individual

is the group level (e.g., team, organizational unit, subordinates of a common leader). In accordance with Klein, Dansereau, and Hall (1994), we
must make two underlying assumptions before applying the term
"group-level lateness behavior." First, although interpersonal differences
in every behavior are inevitable, lateness is supposed to be relatively
homogeneous within the group. This means that group members are
Employee Lateness
more similar than dissimilar with regard to their lateness behaviors.

Second, for different groups coexisting within a common organization,
lateness differs from one group to another. The two assumptions indicate
that the group, and not a lower (Le., the individual) or higher unit (e.g.,
the organization), is the appropriate level of analysis. If the second
assumption does not hold and quite similar patterns of lateness behavior
characterize most of the organization members, then the group level is not
appropriate here. If, in addition, lateness differs from one organization to
another, then the organization is the correct level of analysis.
How is it that group (or organization) members exhibit similar attendance behaviors? As shown in the second column of Table I, common factors, at the group or organization level, may affect the behaviors of all (or
most) members and shape their patterns of work attendance. For instance,
formal company rules indicate what is a tolerable or justifiable level of
lateness and what is intolerable and unjustifiable. Similarly, reward systems and discipline may account for the uniformity of employee attendance behavior (Sagie, Elizur, & Koslowsky, 1996). As demonstrated by
Caplice and Mahmassani (1992), workers may arrive earlier to the workplace in an environment with low management tolerance to lateness. In the
case of cohesive groups, informal attendance norms are not less important
than formal rules (Kanekar & Vaz, 1993; Spink & Carron, 1992). Indeed,
highly interdependent tasks and strong group cohesiveness should lead to
low-lateness-tolerance norms among co-workers (Blau, 1995).
The variables-rules, norms, rewards, and sanctions-should be considered in combination. Kerr (1995) demonstrated that norms are determined not by what management says but what it actually does, through
the company's reward system. It follows that when the firm preaches the
significance of prompt arrival at work but, in practice, does not measure
arrival time, the operative significance of this norm may be very low. If,
as another example, the manager demonstrates the importance of arrival
on time through a personal example, this would be considered a highly
significant norm.
Organizational work schedules and arrangements also impact
employee punctuality. As implied by its name, flexible work time sets
more liberal boundaries on what is considered on-time arrival. Observing
workers who come to work at an hour that was once considered as late
may arouse an impression of increased lateness. Actually, however, fewer
occurrences of real lateness can be expected. Hence, one is able to arrive
at work at an acceptable time despite the influence of various individuallevel factors including commuting problems, family obligations, or even
getting up late. Further, in addition to the practical help in the confrontation with these constraints, the change from fixed into flexible work time
10
generally provides greater motivation for effective functioning at work and

for better adherence to the attendance rules. Several other work arrangements, such as part-time employment (Krausz, Sagie, & Bidermann, 2(00)
and working at home or in a geographically distributed workplace (Le.,
switching locations in the course of the work day), may seem to an outside
observer that people are missing, but in actuality they reduce the difficulties associated with lateness. Conversely, shift work (Barton, 1994) and rigid
service time (e.g., at shop or bank) make these difficulties more acute, but
help in reducing their magnitude.
All of these variables are relatively constant characteristics of the relevant group or organization and have a relatively uniform influence on
the attendance behaviors of the involved members. The effects of some
other variables on lateness vary, however, as a result of different group or
organizational events and occurrences (e.g., change or crisis). For example, when there is a change in the company (e.g., restructuring; downsizing) and reduction in manpower is considered, punctuality tends to
improve. In this case, the violation of the company rules may be too
expensive. Conversely, in times of stability, more employees allow themselves to arrive at work after the scheduled time. As in the case of individual-level lateness, we propose that while the group (or organizational)
constant characteristics account for intergroup differences in lateness
behavior, the situational factors primarily account for intragroup variations in lateness behavior.
We conclude this section by framing two group (or organizational)level propositions:
Proposition 3: Group or organizational factors, like rules, norms, reward system, leader's
self-examples, work arrangements, and organizational change, cumulatively account for
the variance in group- or organizational-level lateness.
Proposition 4: Group or organizational constant characteristics (e.g., rules, norms, reward
system, leader's self-example, work arrangements) primarily account for intergroup differences in lateness behavior; the situational factors (e.g., organizational change) primarily
account for intragroup variations over time in lateness behavior.
EXTRAORGANIZATIONAL LEVEL
Organizations do not exist within a vacuum; the impact of the environment within which the company operates on the staff attendance behavior is often crucial and can be more significant than the influence of the
specific company for which the employee works. The rightmost column
in Table 1 presents several extraorganizational antecedents of punctuality
and tardiness. Starting with industry type, it appears that as compared to
Employee Lateness
11
high-tech industrial firms, tardiness is less tolerated in the low-tech

industry. The reason may be that as the low-tech worker is expected to
operate a machine throughout the entire day without any slack time, his
or her lateness is more visible. Conversely, the high-tech employee's lateness is less felt, as a significant portion of his or her workday is devoted
to outside activities such as meetings, coordination, planning, and training. In such an environment, performance is measured mostly by number
of projects completed regardless of time spent (or not spent) on each project. Also, the extended workday for many high-tech employees actually
enables them to return missing time. Thus, to the degree that one remains
longer time after the workday officially ends, his or her late arrival is not
considered to be a serious discipline problem (Ralston, 1989).
Low- and high-tech companies differ also with regard to the variability of attendance rules across organizational levels (see Table 2). In the
case of low-tech companies, the gap between the rules applying to senior
and junior workers is much larger than the respective gap in high-tech
companies. Indeed, in the former, attendance rules for the rank and file
are much stricter and more demanding than the rules that are relevant to
supervisors and managers. Conversely, in the case of high-tech companies, rules for both junior and senior employees are quite similar and
allow each to start the workday late, but to devote as much time as
needed for task completion. Ownership of the organization makes a difference as well. Generally, the public sector resembles the low-tech privately owned companies; in both cases, overtime is usually limited, and
the difference between attendance codes of senior and junior clerks is
pretty high. Nevertheless, management policy in companies and institutions within the public sector is very often more liberal than that of the
privately owned low-tech companies.
National culture is also an important determinant of work time and
lateness behavior (Manrai & Manrai, 1995). Adler (1993) showed that the
value of punctuality is more important for workers in Japanese than
American companies. This does not mean, however, that workers in the
United Sates are typically latecomers; in comparison with some other
countries they appear to be quite punctual. For example, in their detailed
study of cultural differences and lateness perceptions, Levine, West, and
Reis (1980) found that punctuality in the United Sates is a more salient
issue than it is in Brazil. Compared with the Americans, the longer intervals of time must pass before the Brazilians consider someone late. In
addition, Brazilian are more likely to attribute lateness to external causes
that are beyond their own control. Other countries can also be classified
according to their normative extent of punctuality and time urgency. A
case described by Trompenaars and Hampden-Turner (1998:133) may
12
Abraham Sagie et at.
illustrate this point. A meeting of managers working for a multinational

company was scheduled for precisely 2:00 PM. At 1:50 PM most participants were present in the meeting room. The meeting, however, could not
open on time, as the Italian and Spanish representatives remained in a
nearby hall until 2:20 PM, making telephone calls.
What accounts for the differences in punctuality among nations?
Trompenaars and Hampden-Turner (1998) distinguished between two
basic types of societies: sequential and synchronic. According to the
sequential point of view, events are organized in an ordered sequence or
a straight line; any change in the sequence appears as a disturbance or a
source of uncertainty. "In the Netherlands you could be the Queen, but if
you are in the butcher's shop with number 46 and you step up for service
when number 12 is called, you are still in deep trouble. Nor does it matter if you have an emergency; order is order" (p. 126). The synchronic
method, in contrast, allows people to track multiple tasks in parallel.
Returning to the former anecdote, it appears that the Italian and Spanish
executives, representatives of synchronic societies, saw no problem with
simultaneous events: their telephone calls and the group meeting.
Additionally, two of Hofstede's (1980) cultural dimensions, individualism/ collectivism and power distance, could be relevant as well to
human diversity in lateness. The individualism/collectivism dimension
describes the way in which the individual defines him- or herself as either
an independent agent or a part of the collective. It may be proposed that
lateness would be less forgivable in individualistic than collectivistic cultures. The rationale for this is that in the former culture, one's time is considered as one's private resource, and others are not allowed to waste it
(e.g., by coming late to a meeting or the workplace). "Time is money" is a
slogan particular to certain individualistic environments. Conversely, for
collectivists, one's time could be perceived as a communal resource, so
that others may feel they are allowed to use it freely.
Power distance is the extent to which the society and its institutions
accept power hierarchy and inequality as legitimate. The larger the distance, the higher is the probability that time at different hierarchical levels would not be considered as equal and diverse rules would be applied
to using the superiors' and subordinates' time. The protocol in Table 2
would be more justifiable in societies with a higher rather than lower
power distance.
Finally, the frequency of lateness occurrences is contingent on various
situational aspects of the extraorganizational environment. For example,
Leigh (1985) found that absenteeism decreases when unemployment
rises. Leigh suggested two explanations for this phenomenon. First,
employees tend to exhibit good attendance when layoffs are imminent
Employee Lateness
13
and prospects of alternative employment are low. Second, some employers use hard times to divest themselves of problematic employees. Both
explanations are equally relevant to lateness; hence, it could be proposed
that higher unemployment would be associated with a decrease in lateness. Another situational variable operating in the extraorganizational
environment is the political climate; it could be predicted that in times of
higher political stability lateness will decrease. Conversely, in times of
higher political tension or conflict, there will be more interference with
the employees' orderly attendance behaviors. Similar to our previous
analyses, we suggest that constant extraorganizational characteristics
account for differences between units (e.g., national cultures) in lateness
behavior and situational factors primarily account for within-unit variations over time in lateness behavior.
In summary, the following extra organizational-level propositions are
suggested:
Proposition 5: Extraorganizational factors, like industry type, ownership type,

national culture, cross-cultural dimensions, labor market conditions, and the
political situation, cumulatively account for the variance in lateness at the extraorganizational unit.
Proposition 6: Extraorganizational characteristics that are constant (e.g., industry type, ownership type, national culture, and cross-cultural dimensions) primarily account for interunit differences in lateness behavior; the situational
factors (e.g., labor market conditions and the political situation) primarily
account for intraunit variations over time in lateness behavior.
AN EMPIRICAL SURVEY
The current analysis indicates that employee lateness is often an outcome
of diverse antecedents operating in multiple levels of analysis. A comprehensive explanation of lateness requires, therefore, the consideration of all
relevant levels. An implication of this is that when the manager considers
the phenomenon in the company, he or she should remember that
although lateness is typically a volitional behavior, it is affected by a wide
range of variables, both within and out of the employee'S direct control.
Are managers aware of the different levels of analysis and the various
antecedents that could be expected to occur in each level? Discussions
with several CEOs and human resource executives clarified that they typically agree with the current trend in the literature and limit their attention
to only one or, at most, two levels of analysis. More specifically, asking
the managers about causes of lateness yielded most often personality
14

Table 3. A Multiple-Level Questionnaire of the Causes Attributed by
Managers to Employee Lateness
Item
Cause of lateness
1. Think of two employees in a team. What are

the causes of their late arrivals
to work? List all relevant causes.
Personality dimensions
Situational variables
Nonwork life
Commuting
20
17
8
7
2. Think of employees in two different teams.

What are the causes of their late arrivals
to work? List all relevant causes.
Group norms
Leadership
Rewards and sanctions
Flextime
13
7
7
3
3. Think of employees in two different

companies or even different sectors. What are
the causes of their late arrivals to work?
List all relevant causes.
Reward systems
Organizational norms
Organizational leadership
Discipline
10
7
2
2
4. Think of employees in two different

countries. What are the causes of their
late arrivals to work? List all relevant
causes.
Cultural differences
Industry type
Public transport
Political tension
22
7
Frequency
6
5
dimensions (e.g., punctuality) or situational variables (e.g., commuting) at

the individual level. Therefore, we designed a questionnaire that explicitly addresses all of the aforementioned levels, and administered it to 40
managers from several private and public organizations in Israel. All of
the respondents were participants in a management course for executives:
32 men and 8 women ranging in age from 23 to 50. The number of
employees in an organization ranged from 5 to 100. The four questionnaire items are given in Table 3. Each item required the respondent to concentrate on a specific level of analysis and to draw from his or her work
experience relevant causes of lateness. The results of the survey are presented in the Table 3.
The findings clarify that for our managerial sample, the employees'
personalities were the most common individual-level category of causes
attributed to their late arrivals at work; 20 out of the 52 responses to question #1 dealt with this category (the respondents were allowed to list more
than a single cause). In line with Blau (1994), the current managers felt that
one's punctuality or unpunctuality is the most important antecedent of
one's late occurrences at work. Other factors at the individual level contributing to late behavior were situational work variables (e.g., ending late
the work of the previous day, interpersonal relations, stressor factors),
Employee Lateness
15
nonwork causes (e.g., home obligations, taking care of children), and difficulties in commuting.
Ouring the preparatory interviews with managers (they were not
included in the present sample), the interviewees rarely mentioned variables at the group and organizational levels as antecedents of lateness. The
questionnaire included, therefore, a direct question requiring the respondents to focus on the interteam factors. At these levels, the most important
causes of lateness were group norms, leadership, rewards and sanctions,
and flexible time. At the interorganizational level the respondents raised
such causes as reward systems, company norms, leadership, and discipline. Finally, when the respondents focused on national origin, cultural
differences were the most frequent explanation of tardiness (22 out of 40).
Other responses to this question were concerned with typical industry
characteristics (especially the existence of flextime arrangements), public
transport system, and political tension. It appears that our respondents
believed that an inadequate public transportation system increases the frequency of lateness occurrences. Also, they assumed that the higher the
political tension in a given society the more numerous are the interferences
with one's orderly attendance behavior. In summary, although not
designed to test the aforementioned propositions, these results supported
the notion that individual, group, organizational, and extraorganizational
factors may explain the phenomenon of employee lateness.
CONCLUSIONS
Up to this point the chapter has discussed parallel tracks of influence

starting from antecedents of lateness behavior at different levels linked
independently to appropriate outcomes. Klein et a1. (1994) used the term
"mixed determinants multiple-level effect" to describe a situation where
several independent variables at various levels of analysis influence a
dependent variable at a lower level. Indeed, one's attendance or lateness
behavior may be a product of antecedents at all of the above-mentioned
levels. Thus, the employee's nationality is significant, contributing to the
prediction of his or her extent of lateness; Latin Americans, for instance,
will come late more often than will Germans. Further, the organizational
setting is important for enhancing the accuracy of lateness prediction.
Hence, even for a Latin American environment, working in a multinational company (especially, German owned) rather than in local ones
ensures a stronger adherence to the company's attendance rules. Beyond
this, knowing one's personality and current work and nonwork situational characteristics may help in predicting lateness.
16
The fact that multiple determinants at different levels of analysis may

affect employee tardiness raises an interesting question: can we identify
here a "pecking order" of importance, so that the highest level (e.g.,
national) is most important, followed by the next level (organizational),
and so on, until the lowest level (individual) which would be considered
as the least important?} The rationale for this scheme is that features of the
environment such as education, tradition, literature, folklore, and media
shape the punctuality habits of individuals so that two different persons
from the same cultural environment (again, two Germans or two Latin
Americans) are more similar than dissimilar to one another. In a like
manner, units within each additional level use their mechanisms of socialization and control (e.g., values, formal rules, leadership) in order to affect
attendance or lateness behavior. Each successive unit inserts more conformity in behavior; hence, two persons from the same national background who are working for the same company are more similar to each
other than either of them is to a person from a different company.
Alternatively, one may offer other hierarchies; for example, regardless
of the specific level of analysis, inherited factors (e.g., cultural tradition or
personality) are more significant than acquired factors (e.g., formal attendance rules). Multiple-level studies are required to examine the relative
weights of units in the diverse levels. Few such studies have been reported.
In the only known study that involved group- and individual-level determinants of employee lateness, Blau (1995) found that the contributions of
individual-level determinants (previous lateness patterns and personal attitudes) exceeded those of the work group norms. Individual-level factors
accounted for 34 and 33% of the variance, for samples of bank tellers and
hospital workers, respectively, whereas group norms accounted for only 4
and 3%. Nevertheless, using hierarchical regression analysis, Blau entered
the individual factors first, followed by the group factor. Thus, the order of
variables entry into the regression analysis can explain Blau's results. This
issue, therefore, still awaits a systematic and strictly controlled investigation.
A situation involving two or more clashing antecedents of lateness,
from the same or different levels of analysis, is instructive in understanding the lateness phenomenon in the current globalized society. In the above
example of the German parent company and its overseas subsidiaries,
applying the German rules would be easier and more successful in a similar cultural environment (e.g., The Netherlands) than a dissimilar one
(e.g., Latin America). In the second case, the clash between the German
and local time orientations would be higher. Consequently, though lateness occurrences are expected to be less common in the Latin American
1
This idea was suggested by an anonymous reviewer.
Employee Lateness
17
subsidiary than in neighboring local organizations, they would be more

acceptable in this subsidiary than in a Dutch company. The situation in our
sample subsidiary can be more complicated due to the possible influence
of other variables at additional levels of analysis, such as strict or loose
organizational rules, rigid or permissive leader, and an employee's punctuality or tardiness. Nowadays, as we face a significant increase in the
amount of international transactions and cross-cultural interactions, an
accurate assessment of the degree of compatibility among diverse
antecedents of lateness is vital for senior managers in multinational firms.
From an organizational perspective, controlling lateness behavior
should be a goal worth striving for. A model that allows the researcher or
the practitioner to predict, explain, or intervene at the appropriate point
would indeed be a welcome addition to the organizational behavior field.
The current model may help in this direction; however, additional propositions, especially those that deal with clashing determinants of lateness
at diverse levels of analysis, need to be developed, and empirical work to
test these propositions is essential. Furthermore, proceeding beyond the
lateness model, it may be quite informative to determine which of the
antecedents discussed above are common to the other withdrawal behaviors, such as absence. As the organizational rules are more permissive
with lateness (it can often be made up on the same day; hence, a late
arrival may not be noted in personnel records) than absence (which is
generally recorded in an individual's file), we expect that the variance in
lateness behaviors would be higher. Nonetheless, it is interesting whether
all of the aforementioned multiple-level determinants are similarly relevant to other withdrawal behaviors such as absence.
Finally, an interesting theoretical question concerns the adaptation of
the current model that is based on contemporary data to future companies with flexible organization-home boundaries. Consider, for example,
the idea of virtual teams that their physical meeting at a specific geographical location is the exception rather than the rule; most of a team's
task activities can be remote controlled. In such a team, working time is
flexible as well. In view of such developments, the traditional terms of the
withdrawal literature, including late arrival at the workplace, would
seem obsolete. Nevertheless, the basic human motives, attitudes, and
behaviors, including the withdrawal phenomenon, would still be relevant. Therefore, we should deal in the future with less clear-cut dysfunctional conduct like withholding efforts at work or social loafing (Sagie
et aI., 2002). Despite the difficulties in assessment, Sagie et al. showed that
the latter withdrawal forms are equally costly to both the employee and
the employer. Furthermore, like lateness, these behaviors could not be
understood fully unless one considers all relevant levels of analysis,
18
including variables at the individual, group, organizational, and extra organizationallevels.

REFERENCES
Adler, P. S. (1993). The "learning bureaucracy": New United Motor Manufacturing,
Inc. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior
(Vol. 15, pp. 111-194). Greenwich, CT: JAI Press.
Baker, L. J., Dearborn, M., & Hamberger, K. (1984). Type A behavior in women: A
review. Health Psychology,S, 477-497.
Barton, J. (1994). Choosing to work at night: A moderating influence on individual tolerance to shift work. Journal of Applied Psychology, 79, 449-454.
Bauer, T. N., & Green, S. G. (1996). Development of leader-member exchange: A
longitudinal test. Academy of Management Journal, 39, 1538-1567.
Beehr, T. A., & Gupta, N. (1978). A note on the structure of employee withdrawal.
Organizational Behavior and Human Performance, 21, 73-79.
Bercovits, M. D. (1996). Time urgency: A personality attribute for predicting lateness.
Unpublished master's thesis, Bar-Han University, Ramat Gan, Israel.
Blau, G. (1985). Relationship of extrinsic, intrinsic, and demographic predictors to
various types of withdrawal behaviors. Journal of Applied Psychology, 70, 442-450.
Blau, G. (1986). Job involvement and organizational commitment as interactive
predictors of tardiness and absenteeism. Journal of Management, 12, 577-584.
Blau, G. (1994). Developing and testing a taxonomy of lateness behavior. Journal of
Applied Psychology, 79, 959-970.
Blau, G. (1995). Influence of group lateness on individual lateness: A cross-level
examination. Academy of Management Journal, 38, 1483-1496.
Burman, M. A., Pennebaker, J. W., & Glass, D. C. (1975). Time consciousness,
achievement striving and the Type A coronary-prone behavior pattern.
Journal of Abnormal Psychology, 84, 76-79.
Caplice, c., & Mahmassani, H. S. (1992). Aspects of commuting behavior: Preferred
arrival time, use of information, and switching propensity. Transportation
Research, 26, 409-418.
Dienesch, R M., & Liden, R C. (1986). Leader-member exchange model of leadership: A critique and further development. Academy of Management Review,
11,618-634.
Greenhaus, J., & Beutell, N. (1985). Sources of conflict between work and family
roles. Academy of Management Review, 10, 76-88.
Hanisch, K. A., & Hulin, C. L. (1990). Job attitudes and organizational withdrawal:
An examination of retirement and other voluntary withdrawal behaviors.
Journal of Vocational Behavior, 37, 60-78.
Hanisch, K. A., & Hulin, C. L. (1991). General attitudes and organizational withdrawal: An evaluation of causal model. Journal of Vocational Behavior, 39,110-128.
Herzberg, E, Maunser, B., Peterson, R, & Capwell, D. (1957). Job attitudes: Review
of research and opinion. Pittsburgh: Psychological Services.
Employee Lateness
19
Hill, J. M., & Trist, E. L. (1955). Changes in accidents and other absences with
length of service: A further study of their incidence and relation to each other
in an iron and steel works. Human Relations, 8, 121-152.
Hofstede, G. (1980). Culture's consequences: International differences in work-related
values. Beverly Hills: Sage.
Jamal, M. (1984). Job stress and job performance controversy: An empirical assessment. Organizational Behavior and Human Performance, 33, 1-21.
Johns, G. (1997). Contemporary research on absence from work: Correlates,
causes, and consequences. In C. I. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (Vol. 12, pp. 113-173).
New York: Wiley.
Kanekar, S., & Vaz, L. (1993). Effects of gender and status upon punctuality norms.
Journal of Social Psychology, 133, 377-384.
Kerr, S. (1995). On the folly of rewarding A, while hoping for B. Academy of
Management Executive, 9 (1), 7-16.
Klein, K. J., Dansereau, E, & Hall, R. J. (1994). Levels issues in theory development,
data collection, and analysis. Academy of Management Review, 19, 195-229.
Knatz, H. E, Inwald, R. E., Brockwell, A L., & Tran, L. N. (1992). IPI and MMPI
predictions of counterproductive job behaviors by racial groups. Journal of
Business and Psychology, 7, 189-200.
Koslowsky, M., Kluger, AN., & Reich, M. (1995). Commuting stress: Causes, effects,
and methods of coping. New York: Plenum Press.
Koslowsky, M., Sagie, A, Krausz, M., & Dolman-Singer, A (1997). Correlates of
employee lateness: Some theoretical considerations. Journal of Applied
Psychology, 82, 79-88.
Krausz, M., Sagie, A., & Bidermann, Y. (2000). Actual and preferred work schedules and scheduling control as determinants of job-related attitudes. Journal of
Vocational Behavior, 56,1-11.
Landy, E J., Rastegary, H., Thayer, J., & Colvin, C. (1991). Time urgency: The construct and its measurement. Journal of Applied Psychology, 76, 644-657.
Leigh, J. P. (1985). The effects of unemployment and the business cycle on absenteeism. Journal of Economics and Business, 37,159-170.
Leigh, J. P., & Lust, J. (1988). Determinants of employee tardiness. Work and
Occupations, 15, 78-95.
Levine, R. V., & Bartlett, K. (1984). Pace of life, punctuality, and coronary heart disease in six countries. Journal of Cross-Cultural Psychology, 15, 233-255.
Levine, R. V., West, L. J., & Reis, H. T. (1980). Perceptions of time and punctuality
in the United Sates and Brazil. Journal of Personality and Social Psychology, 38,
541-555.
Manrai, L. A, & Manrai, A K. (1995). Effects of cultural-context, gender, and
acculturation on perceptions of work versus social/leisure time usage. Journal
of Business Research, 32,115-128.
March, J. G., & Simon, H. A (1958). Organizations. New York: Wiley.
Mobley, W. H. (1987). Intermediate linkages in the relationship between job satisfaction and turnover. Journal of Applied Psychology, 62, 237-240.
20
Nicholson, N., & Goodge, P. (1976). The influence of social, organizational, and
biographical factors on female absence. Journal of Management Studies, 13,
234-254.
Ralston, D. (1989). The benefits of flextime: Real or imagined? Journal of
Organizational Behavior, 10, 369-373.
Richard, D., & Slane, S. (1990). Punctuality as a personality characteristic: Issues of
measurement. Journal of Psychology, 124, 397-402.
Sagie, A. (1998). Employee absenteeism, organizational commitment, and job satisfaction: Another look. Journal of Vocational Behavior, 52, 156-171.
Sagie, A., Birati, A., & Tziner, A. (2002). Assessing the costs of behavioral and psychological withdrawal: A new model and an empirical illustration. Applied
Psychology: An International Review, 51, 67-89.
Sagie, A., Elizur, D., & Koslowsky, M. (1996). Work values: A theoretical overview
and a model of their effects. Journal of Organizational Behavior, 17, 503-514.
Sagie, A., & Koslowsky, M. (2000). Participation and empowerment in organizations:
Modeling, effectiveness, and applications. Beverly Hills: Sage.
Shafritz, J. M. (1980). Dictionary of personnel management and labor relations. Oak
Park, IL: Moore.
Spink, K. S., & Carron, A. V. (1992). Group cohesion and adherence in exercise
classes. Journal of Sport and Exercise Psychology, 14, 78-86.
Steers, R. M., & Rhodes, S. R. (1978). Major influences on employee attendance: A
process model. Journal of Applied Psychology, 63, 391-407.
Trompenaars, E, & Hampden-Turner, C. (1998). Riding the waves of culture:
Understanding cultural diversity in global business (2nd ed.). New York:
McGraw-Hill.
2
National Culture and Perceptions
of Absence Legitimacy
In recent years there has been increasing interest in the cross-cultural aspects
of organizational behavior. This interest is a joint effect of the globalization
of business, increasing cosmopolitanism among researchers, and recognition that cultural contrasts can provide a new optic on traditional domains
of organizational research. Despite the growth in cross-cultural organizational research, one area has remained almost immune to its influence-the
study of so-called work withdrawal behaviors such as lateness, absenteeism, and turnover. This is especially curious when it is recognized that
these behavioral manifestations of withdrawal are themselves culture-free,
unlike many of the hypothetical constructs favored by organizational
researchers. Indeed, in the domain of absenteeism research, Martocchio and
Harrison (1993:295) allude to a "gaping cross-cultural hole." This chapter is
an attempt to repair this gap by proposing a model of some factors that
influence the perception of absence legitimacy across national cultures.
EFFECTS OF NATIONAL CULTURE ON ABSENTEEISM:
THE BASIC ARGUMENT
There are at least three reasons to suspect that cross-national cultural
differences will have an impact on attitudes toward absenteeism and
HELENA MENSAH ADDAE University of the West Indies, St. Augustine Campus,
Trinidad & Tobago
GARY JOHNS Department of Management, Concordia University,
Montreal, Quebec H3G 1MS, Canada
21
22
attendance, and ultimately on attendance behavior. One is growing evidence for the impact of social dynamics on absenteeism within cultures,
or at least within Western cultures where the extant research has been
conducted. Given the marked differences in values that have been
observed between cultures (Earley & Gibson, 1998; Hofstede, 1980, 1991;
Triandis, 1995) and the impact that values have on norms and other social
influence mechanisms, there is every reason to expect that views concerning absenteeism might differ across cultures.
Absenteeism is a behavior that has at least two qualities that would
seem to make it particularly susceptible to social influence, and by extension, cultural influence. For one thing, in the abstract, it is far from clear
just what constitutes a reasonable, legitimate level of absence. This follows from the rather large differences in absence rates that have been
observed between social units such as work groups, departments, plants,
industries, and nations (reviewed by Johns, 1997). This inherent ambiguity provides entree for social influence. At the same time, absenteeism is
perceived as mildly deviant behavior (Robinson & Bennett, 1995). At least
in the West, people make negative attributions about the absence of others (Nicholson, 1975), feel guilty about being absent (Hackett, Bycio, &
Guion, 1989), and get into disputes with employers about absenteeism
(e.g., Clay & Stephens, 1994). In both the East and the West, people underreport their own absence behavior and see their attendance as superior to
that of their work colleagues (Johns & Xie, 1998). These negative connotations do not deny individual differences in the perceived legitimacy of
absenteeism, but they do again suggest that people may be socially sensitive to how others view the behavior, thus setting the stage for social
influence (see Gellatly & Luchak, 1998).
Johns (1997, 2001) reviews the growing evidence concerning the
impact of social influence on absenteeism within national cultures. For
instance, a wide variety of operationalizations of absence norms tend to
be correlated with actual absence behavior (e.g., Baba & Harris, 1989;
Gale, 1993; Gellatly, 1995; Gellatly & Luchak, 1998; Harrison, 1995; Xie &
Johns, 2000). Similarly, work group structure and process variables, in
particular group cohesiveness, have been shown to predict absence.
Although cohesive work groups tend toward less absence (Johns, 1997),
moderators such as job dissatisfaction and collusive tendencies can stimulate absence in cohesive groups (Drago & Wooden, 1992; Xie & Johns,
2000). Finally, cross- and multi-level research has detected the existence of
distinctive absence cultures (Chadwick-Jones, Nicholson, & Brown, 1982;
Johns & Nicholson, 1982; Nicholson & Johns, 1985) at the unit (usually
work group) level of analysis (e.g., Iverson, Buttigieg, & Maguire, 1999;
Johns, 1994; Markham & McKee, 1995; Martocchio, 1994; Mathieu &
National Culture and Absence Legitimacy
23
Kohler, 1990; Xie & Johns, 2000). Such research frequently reveals the
impact of the group on individual absenteeism or related perceptions.
A second reason why absenteeism might be expected to vary across
cultures rests in cross-national differences in factors such as economic
development, infrastructure support for families, social welfare provisions
concerning absenteeism from work, and so on. For example, Griindemann,
de Winter, and Smulders (1994) documented a range of legislative differences and a corresponding range of absenteeism rates between nations of
the European Union. Kaiser (1998) interpreted such differences in terms
of the economic implications of Hofstede's (1980) cross-cultural typology of
values. If the social influence described above constitutes discretionary
stimuli on the part of the culture that might shape absence attitudes and
behavior fairly directly, the influence being described here is ambient stimulation (Hackman, 1992), background factors that condition views about
the legitimacy of absenteeism and stem indirectly from cultural values.
A third reason why perceptions and attitudes about absenteeism
might be expected to vary across cultures rests in a small body of research
that has actually examined culture and job withdrawal. In the domain of
turnover, Abrams, Ando, and Hinkle (1998) found that low identification
with one's organization predicted intentions to quit in both Japan and
Britain. However, subjective norms regarding turnover had a much more
potent influence in collectivistic Japan than individualistic Britain. In a
case study of a General Dynamics plant in Arizona staffed by Navajo
Native Americans, Winfield (1995) explored how Navajo culture and tradition influenced work attendance. In particular, she described how participation in a nine-day healing ceremony led to absence and how plant
management accommodated this ritual. Kuzmits (1995) studied absenteeism among employees in a midwest U.S. food processing plant, some
of whom were Vietnamese refugees. Based on Confucian values, he predicted and confirmed that the Vietnamese exhibited less voluntary
absence and less absence-related discipline than non-Vietnamese, while
not differing on involuntary absence. Johns and Xie (1998) predicted and
found that Chinese and Canadians were equally self-serving in underestimating their own absence from work and seeing their attendance record
as better than that of their peers. However, in line with collective values,
the Chinese were also more inclined to group-serve, seeing the attendance
of their work peers as much superior to the occupational norm. The
authors also found cultural differences in the perceived legitimacy of various reasons for absenteeism. Finally, at the same Chinese research site,
Xie and Johns (2000) found that work group cohesiveness and absence
culture salience interacted to predict absence at the individual, work
group, and cross-levels of analyses. Although this was a uniculture study,
24
the collective orientation of the Chinese was particularly conducive to the

observed effects.
In summary, we believe that the above lines of argumentation
and empirical evidence suggest that there is considerable merit to examining potential cross-cultural differences in the perceived legitimacy of
absenteeism.
ABSENCE LEGITIMACY AT CULTURAL AND INDIVIDUAL
LEVELS
A key dependent variable in the model of absence to be presented is

perceptions of absence legitimacy. Legitimacy can be defined as the extent
to which employees perceive absenteeism as an acceptable work behavior, and it is embedded in the social context within which the behavior is
enacted. The study of absence legitimacy has at least three distinct advantages for examining absence cross-culturally. First, one of the most useful
ways to understand cultural differences and similarities is to consider
what behaviors are considered more or less legitimate and thus sanctioned by the society in each culture. This more direct, more behavioral
approach is thus complementary to the study of more general, more
abstract cultural values. Next, the concept of absence legitimacy would
seem to have the potential for construct validity at both the cultural and
individual levels of analysis. That is, cultures might differ in the extent to
which absenteeism itself or certain causes of absence are seen as legitimate Oohns & Xie, 1998), but so might individuals within these cultures.
As will be detailed below, there is considerable merit to developing
a model that applies at both the cultural and individual levels of analysis.
Finally, absence legitimacy may be a more proximal and useful criterion
for initial cross-cultural studies of absence than absenteeism itself. This is
because cross-cultural studies almost necessarily involve a host of uncontrolled variables, some of which might influence work attendance in spite
of the perceived legitimacy of absence. Thus, just because a behavior
is viewed as legitimate does not guarantee that it will be enacted.
Nevertheless, we do conceive of absence legitimacy as a mediating variable between more distal cultural causes and absenteeism, with more perceived legitimacy resulting in a higher absence rate. However, moderators
of the legitimacy-absence connection are quite feasible.
Harvey and Nicholson (1999) studied the perceived legitimacy of
various illnesses as causes of absence. They found considerable variation
by illness type and by respondents' job grade and age. They also found
that those who attributed greater legitimacy to absence due to colds,
25
upset stomach, or mild backache were absent from work more frequently.
As they note, these are high-discretion illnesses in terms of attendance.
Directly relevant to the cross-cultural thesis, Johns and Xie (1998) found
similarities and differences between Chinese and Canadians in the perceived legitimacy of a wide variety of causes of absenteeism. For example,
the Chinese were less likely to endorse illness, stress, and depression as
reasons for absence and more likely to endorse house maintenance and
personal business. No difference was observed regarding problems with
bosses or co-workers. This limited evidence suggests promise for a theoretically sound model of cross-cultural differences in absence legitimacy.
In the model to be presented below, it will be argued that the following variables influence perceptions of absence legitimacy at both the cultural and individual levels of analysis: work centrality; time orientation;
locus of control; perceptions of gender role differentiation; perceptions of
the efficacy of social support systems. These variables were chosen for
consideration because of their documented or suspected variation both
between individuals and between cultures. This choice was intended to
accommodate cross-national variation while recognizing individual differences in values and perceptions within cultures. As such, the proposed
model is a multilevel model (Klein, Dansereau, & Hall, 1994; Klein,
Tosi, & Cannella, 1999; Rousseau, 1985) in that the proposed effects for
absence legitimacy are expected to occur at both the cultural and
individual levels. As Klein et al. (1994:223) explain, such models are
"uniquely powerful and parsimonious."
WORK CENTRALITY
There is a general recognition that work occupies a great deal of an

individual's time in industrial societies (England & Harpaz, 1990). Is work
therefore the activity of most value and importance to individuals? In an
attempt to provide answers to this question, Dubin (1956) posited that the
social world of individuals is subdivided into various areas of fairly independent social activity and interest. Moreover, all social settings do not
have equal salience to people, and they will spend more time at a given
location and on an activity that reflects their central life interests.
Dubin (1956) found that although most people spend a considerable
amount of their time at work, only 24% of the respondents surveyed indicated that work was a central life interest. A number of researchers have
criticized the validity of his measures and the conclusions drawn from his
research (Maurer, Vredenburgh, & Smith, 1981). In spite of the criticisms,
subsequent studies have investigated the relationships between central
26
life interests and attitudinal variables. Dubin, Champoux, and Porter

(1975) found that blue-collar workers in a bank and a telephone company
with high work centrality had higher levels of organizational commitment than those with nonwork central life interest. Baba (1989) found a
significant positive correlation between work as a central life interest and
job involvement among teachers in Nigeria and Trinidad.
Drawing on studies conducted on central life interest, the Meaning of
Work (MOW) International Research Team (1987) conceptualized work
centrality as (1) an individual's identification with working or the significance that he or she attaches to the work role compared to other life roles
and (2) the value placed on working as a life role in absolute terms.
Similarly, Kanungo (1982) defined work centrality as the extent to which
individuals identify psychologically with work and view it as important
in their lives and central to their conceptions of self.
This conceptualization of work centrality can be linked to Kluckhohn
and Strodtbeck's (1961) cultural dimension of activity orientation. They
advocate that people's activity orientation affects their approach to work
and leisure and how preoccupied they are with work and work-related
concerns. People in doing-oriented cultures have a tendency toward
viewing work and work-related activities as central to their lives. On the
other hand, people from being-oriented cultures will be more spontaneous and leisure-oriented. Thus, work becomes a means to achieve
leisure pursuits.
Work centrality should be contrasted with job involvement. Whereas
work centrality denotes identification with the broader concept of work,
job involvement represents a more specific construct that measures the
extent to which individuals identify psychologically with the kind of
work they do. Thus, from a normative perspective, work may be central
to an individual's life but he or she may not necessarily psychologically
identify with the type of work performed. However, we do recognize that
an individual's job involvement may be related to his or her work
centrality.
Differences in level of work centrality exist within and across nations.
In their investigation of the perception of work in eight nations, the MOW
International Research Team (1987) found evidence of cross-national differences in work values. For example, the Japanese had the highest mean
work centrality score while the British had the lowest mean score.
Generally, it was found that individuals with high work centrality tended
to work longer hours. As such, it is expected in nations where work is central to the lives of most individuals absenteeism will less likely be viewed
as a legitimate work behavior than in those nations where individuals
may be more leisure oriented.
27
Kanungo (1990) found that in India there exists a family-centered

work ethic. Work is therefore central to individuals to the extent to which
they believe that work helps to maintain the needs of one's family (which
includes parents and relatives), friends, and even strangers. Theyemphasize leisure pursuits that foster security and social needs.
Thus, it is anticipated that work attendance would be instrumental in
attaining economic and social needs. As such, if a certain level of attendance is sufficient to maintain one's needs, absenteeism may be seen as a
legitimate behavior.
Similarly, Ali and AI-Shakhis (1989) found that in Saudi Arabia, family and then work were the most important activities in the lives of the
respondents, followed by leisure and community service. It is expected
that family needs would supersede work needs. Family-related activities
would be more central and thus to the extent that individuals have to be
absent from work in order to attend to family matters, absenteeism would
be viewed as a legitimate work behavior.
Work centrality is transmitted through socialization of the individual.
Without exception, all of the researchers who conducted studies on the
meaning of work and its centrality to individuals in a particular society
cited economic, religious, and cultural reasons for their findings. For
example, Ali and AI-Shakhis (1989) indicated that from a cultural standpoint, Saudi Arabians' obligation to their family takes precedence over
their individual needs. Work is instrumental in pursuing family interest
and enhancing an individual's reputation in society. England and Misumi
(1987) posited that national differences in work centrality between Japan
and the United States could be explained by cultural and economic factors. They indicated that collectivism in Japan facilitates uniformity in the
socialization practices of organizations and that leads to higher levels of
work centrality.
Although Kanungo (1982) postulated a theoretical link between work
centrality and absenteeism, a review of the literature identified one study
that attempted to link work centrality to absenteeism (Shapira & Griffith,
1990). They found no relationship between work centrality and the number of days absent for employees over a 6-month period. They suggested
that the reason for the low correlation between the variables could have
been due to the low absence rate of the respondents. Brooke and Price
(1989) found that work involvement influenced self-reported absenteeism
through employee job satisfaction.
Research examining the effects of academic involvement on student
absenteeism suggests that there is a negative relationship between the
two. Farrell and Mudrack (1992:707) operationalized academic involvement as "the extent to which individuals identify psychologically with
28
their school activities and regard these as important in their lives and central to their conceptions of self." This definition is consistent with that of
work centrality and Kanungo's (1982) conceptualization of work involvement. However, the domain investigated here is different from the work
environment, which may suggest that the dynamics of academic involvement may be different from those of work involvement. These findings
provide some evidence that work centrality may not strongly influence
absenteeism. Perceptions of absence legitimacy may therefore be a viable
mediator in the relationship between work centrality and employees'
absenteeism.
In spite of the paucity of research linking work centrality with absenteeism, most studies of absenteeism have implied that work is central to
the lives of individuals and have ignored other life spheres that may be of
more importance and influential in explaining absence. This is seen in the
large amount of research that links job variables to absence (Harrison &
Martocchio, 1998; Johns, 1997) and in the relative lack of research on nonwork factors.
It is postulated that, in general, if work is a central life interest for individuals, their work involvement will be high and consequently they will
engage more in work and work-related activities than those whose involvement in work is low. As such, people with high work centrality would not
view absenteeism as a legitimate behavior and would exhibit lower levels
of absenteeism. Based on the preceding we offer the following proposition:
Proposition 1: There will be a negative relationship between employees' work centrality and their perceptions of absence legitimacy at the individual and national
levels of analysis.
TIME ORIENTATION
Time is one of the most important dimensions of human and organizational behavior. Despite its inevitability and pervasiveness, the dimension of time and its effect on employee behavior have not been accorded
appropriate attention in research in organizational behavior. There are
cultural differences in people's concepts of time, and societies can be
differentiated as either monochronic or poly chronic (Hall, 1966).
Monochronic societies view time as money, take time commitments such
as deadlines and schedules seriously, and emphasize promptness. On the
other hand, in polychronic societies people perceive time as a "free" good,
view time commitments as a goal to strive for if possible, and are likely to
change their plans easily and often (Hall, 1966, 1983).
29
In Western industrialized societies, time is viewed as homogeneous,

unidirectional, irreversible, divisible, linear, measurable, singular, objective, and abstract (Byrne, 1982; Denhardt, 1986; McGrath & Rotchford,
1983; Schriber & Gutek, 1987). On the other hand, in developing countries, time is viewed as an intangible that is ever-present, and therefore
need not be managed. For example, in India, people "live physically in the
present, but psychologically in the past, and are unconcerned about the
future ... time is often considered in an abstract philosophical way as
being eternal i.e., ever present but never passing. Thus, delays in actions
or slowness at work are easily tolerated as normal. Deadlines, time
targets, punctuality, etc. are meaningless" (Kanungo, 1990:805).
Similarly, in Trinidad and Tobago, people live in apparent indifference to time, and there is no apparent notion of exact times for appointments. The following excerpt from the Trinidad Express (July 5, 1996)
illustrates this assertion.
For the 4th of July celebration of American Independence, the bell is
usually rung exactly at midday. But with some of the guests arriving
after the bell-tolling hour at the annual party ... the traditional bell
ringing took place just after 1 p.m. With a complete understanding of
the culture, however, Brian Donnelly [the U.S. Ambassador] simply
dubbed it "Trinidad time."
Time is therefore a subjective social phenomenon based on the temporal experiences of individuals across nations. In a 12-nation study,
Robinson, Converse, and Szalai (1972) determined cross-national disparity in how individuals allocate their time among such activities as work,
personal needs, childcare, household chores, education, recreation, and
other leisure activities.
In most nations, the temporal cycle of activities tends to be geared
primarily to the cycles of work organizations. Thus, family and other
social organizations become residual claimants to the allocation of time.
The patterning and allocation of time based on organizational work
schedules pose constraints on individuals' ability to meet competing
demands on their time from their various life spheres. Thus, organizational work schedules may have a direct effect on how individuals apportion their work and nonwork time.
Work schedules may be uniform for most institutions in some
nations, whereas in others, the schedules may differ considerably across
institutions. In some nations most organizations have fixed opening
hours such as 8 AM to 4 PM, or 9 AM to 5 PM. Thus, offices and shops will
generally open and close at about the same time. On the other hand, in the
30
United States and Canada for example, the general tendency is for shops
to close later than offices. Furthermore, a number of services such as
banking and airline reservations can be done via 1-800 numbers or the
Internet, and some grocery stores are open 24 hours a day.
In nations where there is uniformity of work schedules across institutions, this will lead to greater constraints on individuals' ability to
accomplish both work and nonwork activities. Absence from work to
accomplish nonwork activities that coincide with work hours may therefore be seen as a legitimate work behavior. In contrast, where there are
differences in work schedules, individuals will have relatively more flexibility in allocating their work and nonwork time, making it more difficult
to legitimize absenteeism as a means of performing nonwork activities.
For example, Johns and Xie (1998) found that while the Chinese perceived
house maintenance and personal business as legitimate reasons for absenteeism, their Canadian counterparts did not.
In spite of the fact that differing or uniform work schedules within a
nation will lead to disparate constraining effects on the allocation of time
cross-nationally, individuals' general values will also influence the
amount of time they allocate to various activities and subsequently influence their absences and their perception of absence legitimacy.
Johns and Nicholson (1982) posited that although there is an undeniable negative relationship between work and nonwork time, there is a
paucity of research in the area. Absence is a constrained behavior (Johns,
1991); thus, it temporarily redraws the boundaries between work and nonwork. If inflexible work schedules impose constraints on individuals'
behavior, non-conventional work schedules, such as flextime and the compressed workweek, should in theory improve employee attendance behavior. This is because individuals are expected to have more liberty in the
allocation of their time to meet competing demands from their various
social spheres. For example, both flextime and the compressed workweek
permit employees to engage in personal business or family activities during standard business hours. Therefore, to be absent in order to meet competing demands from nonwork spheres may not be perceived as legitimate.
Research findings on the relationship between the various nonconventional work schedules and absenteeism are mixed. The evidence for
flextime suggests that it stimulates attendance. For example, in a longitudinal field experiment, Dalton and Mesch (1990) found the introduction
of a flextime program in a large utility company led to a considerable
decrease in absenteeism, but absence levels returned to baseline rates with
the termination of the program. Kim and Campagna (1981) found that flextime reduced absenteeism of employees in a public sector organization.
31
Interviews with employees revealed that they felt they had more control
over the use of their time, and this allowed them to balance work and nonwork demands.
On the other hand, McGuire and Liro (1987) reported that traditional
flextime did not reduce absenteeism. Instead, they found that a staggered
fixed schedule reduced absenteeism among employees. Pierce,
Newstrom, Dunham, and Barber (1989) reported that researchers have
generally found a reduction in absenteeism with the introduction of flextime in organizations, a result confirmed in a recent meta-analysis by
Baltes, Briggs, Huff, Wright, and Neuman (1999).
These findings suggest that since flextime attempts to minimize or
eliminate constraints to attendance, employees on such work schedules
may not be able to legitimize their absences. A lack of justification for
absenteeism under a flextime schedule may be even more pronounced in
nations where there are differential work schedules across institutions.
This is because flextime allows them considerably more flexibility than
employees working under standard hours. For example, recent interviews with employees in a large multinational organization that introduced a flextime work schedule in Trinidad and Tobago suggest that
whereas employees may have viewed absenteeism as a legitimate behavior before the introduction of the system, they can no longer justify their
absences as a means of meeting nonwork demands.
Results for the effect of the compressed workweek on absence have
been less favorable. Ivancevich and Lyon (1977) found no change in
absence among operating employees in a manufacturing company. Nord
and Costigan (1973), however, found decreased absence in a pharmaceutical company after the introduction of a compressed workweek. Based on
five such studies, Baltes et al. (1999) concluded that the compressed workweek does not reduce absenteeism. It is likely that the compressed
workweek may have more effect on employees' nonwork activities than
work activities.
Another stream of research that is relevant to an individual's ability
to distribute his or her time accordingly is research on shift work. It is postulated that shift work allows employees discretion in the use of their time
to meet extra work responsibilities.
Taylor, Pocock, and Sergean (1972a) reported that absenteeism
among shift workers was lower than that of day workers. However, they
speculated that although shift work influences absence, certain factors
other than shift work may playa role in the effect of shift work on absenteeism. In another study (Taylor, Pocock, & Sergean, 1972b), in which they
examined factors such as sick pay schemes, weekly hours of work, and
32
earnings, they found that biographical factors, occupational factors, high

identification with one's job, and shift premiums moderated the relationship between shift work and absence. Importantly, employees who opted
to work under shift systems were less absent than their counterparts who
did not choose to work under such systems. Thus, self-selection moderated the relationship between shift status and absenteeism.
Nicholson and Jackson (1978) indicated that while shift work influences absence, temporal boundaries of shift work would have independent effects on absenteeism. They reported that the length of the shift cycle
was related to the frequency of absence at different positions within the
cycle. Furthermore, they found days of the week had strong independent
effects on absence, thus confirming Pocock's (1973) assertion of the existence of daily variations in absence levels.
Results from studies on alternative work schedules have been equivocal, but those from studies that have attempted to link work and nonwork factors to absenteeism have not been controversial. Rousseau (1978)
found that employees' perceptions of work and nonwork activities correlated with their frequency of absence. As such, employees who were more
satisfied with their non work experiences were more absent than those
who were more satisfied with their work. Morgan and Herman (1976)
found a correlation between perceived achievement of off-the-job outcomes and leisure time. Moreover, perceived off-the-job social outcomes
were a better predictor of absence than perceived on-the-job penalties. In
a direct evaluation of the relationship between work and nonwork and
absenteeism, Youngblood (1984) found that employees who valued their
nonwork time more were absent more. These studies demonstrate that
absence is an attempt to "balance the quantity and quality of time spent
in various endeavors" (Johns & Nicholson, 1982:151). Therefore, employees who value and are more satisfied with their nonwork activities may
not perceive absenteeism as illegitimate.
Kacapyr (1996) reported that the American Demographics Index of
Well-Being indicated that, in the United States, the number of nonwork
hours has been on the decline since 1990. In less than 6 years, recreation
spending increased by 35%. In contrast, nonwork hours decreased by 2%
over the same period. However, the decline in leisure hours seems to be
tapering off. In the first month of 1996, the number of nonwork hours
grew more than 1% but not at the same rate as the decrease in nonwork
hours. It is concluded that while Americans may have less nonwork time
available to them, they are spending much more on entertainment
(Kacapyr, 1996). If nonwork hours are decreasing it would be interesting
to find how people allocate their time between work and non work.
Would such changes lead to increases in absence and/ or the perception of
33
its legitimacy? In light of the preceding:
Proposition 2: The flexibility of the boundaries between work and nonwork time
will influence perceptions of absence legitimacy at the individual and national
levels of analysis.
If competing demands on individuals' time have an effect on how
individuals apportion their time between their work and nonwork life
spheres, their time perspective is also expected to influence their behavior.
Landy, Rastegary, Thayer, and Colvin (1991) indicated that people differ in
their concern about time and its passage. Thus, people with a time urgency
perspective will function at a more accelerated pace than those who do not.
Central to this line of research are Type A and Type B behavior patterns.
Type A individuals have a heightened sense of time urgency, while Type Bs
have a more relaxed view of time. Although Landy et al. (1991) conceptualized time urgency as an individual-difference construct, it can be
extended to the national level of analysis. For example, Jones (1988) linked
Type A and B behaviors to temporal perspectives associated with different
cultural groupings. He advocated that in cultures where future time orientation is predominant, Type A behavior outcomes are expected to prevail.
Levine and Wolffe (1985) demonstrated that the pace of life differs
across nations. In a six-nation study, Japan was the most time-conscious
nation, while Indonesia was quite leisurely on the indicators of accuracy
of bank clocks, walking speed, and post office service speed. In a followup study of 31 countries, Levine (1997) found differences in the pace of
life among the countries surveyed. It can be argued that in nations where
the pace of life is fast, and time management is emphasized, individuals
would more likely have a time-urgent orientation. For example, in Latin
America, one might expect to wait hours even in the face of appointments, since appointment does not have the same meaning as it does in
the United States (Hall, 1973). It is therefore safe to assert that in general
Latin Americans would be more Type B than North Americans. Moreover,
it is estimated that 50% of North Americans are Type As (Friedman &
Rosenman, 1974).
Conte, Mathieu, and Landy (1998) posited that time urgency is
related to organizational outcomes such as performance (Friend, 1982;
Glass, Snyder, & Hollis, 1974). Explicitly relevant to current purposes,
Conte and Jacobs (1999) found that time awareness was negatively related
to absenteeism. They posited that employees who schedule their work
and are more aware of the passage of time are expected to be more punctual and would therefore be less absent. Furthermore, consistent with Hall
and Hall's (1990) assertion that polychronics would be less concerned
about the passage of time, they found a positive relationship between
34
both lateness and absence and polychronicity. In general, individuals with

a time-urgent perspective will perceive absence as a less legitimate behavior. This view is consistent with Schriber and Gutek's (1987) suggestion,
in their explication of the time dimensions of organiZations, that individuals whose perceptions of the importance of punctuality or schedules and
deadlines are at variance with those of their co-workers might exhibit elevated absenteeism. Thus, we offer the following proposition:
Proposition 3: Employees' time orientation will influence their perception of

absence legitimacy at the individual and national levels of analysis.
LOCUS OF CONTROL
Rotter's (1966) internal-externallocus of control construct assesses
the extent to which individuals attribute the control or causes of their
actions and behavior to themselves or to factors in their environment.
Individuals with an internal locus of control believe that they are personally responsible for their behaviors. On the other hand, externals believe
that luck and fate determine their behavior and what happens to them.
Lefcourt (1992) pointed out there has been a great deal of interest in
locus of control since these beliefs have implications for employees' attitudes, perceptions, and behaviors in organizations (Blau, 1987). Locus of
control has been found to be related to a number of organizational factors
such as performance, motivation, and turnover (Renn & Vandenberg,
1991; Spector, 1982) as well as organizational commitment (Coleman,
Irving, & Cooper, 1999). Generally, employees with an internal locus of
control were rated higher on performance and satisfaction and were more
committed and motivated than those with an external locus of control.
Although the construct was conceptualized to assess individual differences, studies have found cross-national differences in locus of control. In
fact, Aycan, Kanungo, Mendonca, Yu, Deller, Stahl, & Kurshid (2000) postulated that cultures that engender external locus of control are more fatalistic
than those that foster internal locus of control. Similarly, Kluckhohn and
Strodtbeck (1961) indicated that people's relationships to their environment differ according to culture. For example, in the North American
culture, people believe that they can dominate their environment whereas
in Middle Eastern cultures, people believe that life is preordained.
These beliefs are consistent with the concept of locus of control. We should
therefore expect a greater proportion of employees in North America to be
internals than employees in the Middle East.
Although most cross-cultural studies of locus of control have
involved students, their findings provide evidence of cross-cultural
35
differences and similarities in perceptions of locus of control. In general,

studies (Hamid, 1994; Harari, Jones, & Sek, 1988; McGinnies, Nordholm,
Ward, & Bhanthumnavin, 1974; Parsons & Schneider, 1974; Reitz & Groff,
1974) have concluded that American, Swedish, New Zealand, and Canadian
students were relatively internal, while Japanese, Indian, Mexican, Chinese,
and Thai students were more external. Men have been found to be more
internal than women across a wide variety of cultural comparisons (Harari
et al., 1988; Parsons & Schneider, 1974; Reitz & Groff, 1974).
Judge, Martocchio, and Thoresen (1997) indicated that personality
and dispositional traits have not been given sufficient attention in absenteeism research. They found that extraversion and conscientiousness predicted absence. It is therefore viable to investigate the influence of locus of
control, a personality trait, on perceptions of absence legitimacy. Judge and
Martocchio (1996) found that people with an external locus of control gave
more excuses for their absences and attributed external causes to their
absence behavior. It can therefore be inferred that because externals attribute external causes to their absences they are more likely to view absenteeism as a legitimate work behavior. Rahman and Kumar (1984) found no
relationship between locus of control and employee absenteeism.
However, Keller (1984) and Johns (1994) found that employees with
perceived internal control over their health were less absent than those
with external health locus of control. Also, research has found a negative
relationship between students' academic locus of control and their absence
behavior (Trice, 1987; Trice & Hackburt, 1989). Frayne and Latham (1987)
found that self-management training in overcoming obstacles to attendance increased self-efficacy for attendance and reduced absence. Such
self-efficacy is analogous to internal control for a specific behavior.
Given the negative connotations of absenteeism alluded to earlier, it
follows that people who perceive low control over their attendance
behavior will view absenteeism as more legitimate. Thus, the following
proposition is advocated:
Proposition 4: Employees' locus of control will influence their perception of

absence legitimacy at the individual and national levels of analysis.
GENDER ROLE DIFFERENTIATION
Gender role differentiation is the extent to which the roles played by
females and males are clearly distinct. Thus, the extent of role differentiation can be represented on a continuum. In a high-role-differentiated
nation, males will tend to engage in roles that are ascribed to them and
will be inclined to disassociate themselves from those regarded as female
36
roles. On the other hand, in places of low-role-differentiation, the separation of roles according to gender will be at a minimum.
Beach (1989) asserted that ideological perspectives on the societal
roles of males and females led to a gender-based model of work.
Accordingly, these ideological viewpoints have been sustained by a theory of structural functionalism that has upheld the need for such a division of labor. Traditionally, females have been charged with taking care of
the family and attending to household responsibilities. In comparison,
men have worked outside the home and have been expected to be the
dominant income earners in the family (Lewis, 1992; Major, 1993).
This is consistent with Hofstede's (1980) stipulation that in masculine
societies, gender roles are clearly differentiated and males are expected to
dominate society and females are expected to be nurturing. In contrast,
feminine societies believe that gender roles are more fluid and therefore
they espouse gender equality. As such, like females, males can also take
on nurturing roles. Hofstede's views are congruent with those of
Trompenaars (1993), who also asserted that societies differ on how roles
are ascribed to males and females. It is thus possible to locate nations
along a gender role continuum.
Scott, Alwin, and Braun (1996) found evidence of cross-national differences in gender role differentiation. In a comparative study of Britain,
the United States, West Germany, the Netherlands, Italy, and Ireland they
found that the Netherlands was the least differentiated according to gender. Attitudes toward gender role differentiation in the United States and
Britain were on par and ranked second. This was followed by Ireland and
Italy, respectively, and it was found that West Germany was the most
gender-role-differentiated nation.
The demographic composition of the work force is changing internationally, and this should have some impact on gender role differentiation.
For example, in the United States, females constitute one of the fastestgrowing sectors of the work force. The proportion of working females
increased from 36% to 54% between 1960 and 1991 (U.S. Department of
Commerce, 1992). While the rate of growth of participation in the work
force is different in other nations, the trend is similar. In Trinidad and
Tobago, there was a steady increase of females in the work force between
1990 and 1993 (Central Statistical Office, 1993). The Statistical Institute of
Jamaica (1994) reported a total population of about 2.5 million in 1993,
comprising 50% females. In the same year, employment of female workers increased by 7400; however, there was a decline of 6800 in male
employment (Economic and Social Survey, 1995).
With the influx of females into the work force, changes in the
compartmentalization of gender roles have occurred. For example, the
37
division of roles between females and males in the United States has
changed over time due to economic and historical reasons (Beach, 1989),
and men have continued to increase their family roles (Pleck, 1985).
Notwithstanding the changes in role differentiation, females continue to be
more involved in traditional roles than their male counterparts. In addition,
females in high-role-differentiated nations will probably be employed in
lower-paying jobs than their counterparts in low-role-differentiated
nations. This is because low-role-differentiated nations will tend to be more
egalitarian. As such, it is likely that females will be employed in jobs that
are traditionally considered male jobs, whereas in high-role-differentiated
nations, certain jobs remain exclusively reserved for males.
In Japan for example, there is high gender role differentiation. In fact,
there is a ban on females engaging in overtime, midnight, and holiday
work. Females are also rarely promoted to higher-level positions in organizations. Despite the fact that Japan has the largest number of employed
women in the developed world (Friedland, 1994), only 2% of them are in
managerial and professional positions (Nihon Horei Kyokai, 1989).
It is anticipated that females will generally be more absent than males
in most nations. Females in high-role-differentiated nations, however, are
expected to be relatively more absent than those in low-role-differentiated
nations. However, absence levels for dominant income earners may not
be affected by gender role differentiation, so both males and females who
are dominant income earners may generally be less absent than those who
are not. For example, females are reported to be less absent than males in
Jamaica (Employers' Consultative Association of Trinidad and Tobago,
1993). This is because a majority of females in the survey headed their
households.
A meta-analysis by Cote and Haccoun (1991) determined that women
exhibit higher absence levels than men. For example, in the United States,
the Bureau of Labor Statistics (1982) reported that the incidence of absenteeism for women was 58% higher than that for men. For instance,
Meisenheimer (1990) reported that during an average week in 1989, the
U.S. absence rate for women was 6.6% compared to 4% for their male
counterparts. Furthermore, females had higher rates of absenteeism than
males in each age group. Due to the widespread acceptance of women's
additional nonwork roles, there seems to be a psychological contract that
has allowed for the institutionalization of higher absenteeism among
females. This institutionalization of female absenteeism is also prevalent
in other nations. For example, Hein (1984) posited that employers in
Mauritius perceived that women who were absent from work had legitimate reasons such as family responsibilities or illness. However, the managers were not equally tolerant of male absenteeism and perceived that
38
males who were absent did so for "absolutely no good reason." Thus, we
offer the following proposition:
Proposition 5: Employees' perception of gender role differentiation will be related

to their perception of absence legitimacy at the individual and national levels of
analysis.
SOCIAL SUPPORT SYSTEMS

In theory, social support systems are services available to employees
that are expected to enhance their ability to attend work and thus minimize their absences. Employers, the government, care providers, or family members may provide these services. Social support that is
company-based may include benefits that are either mandated by law or
provided for employees at the discretion of the employer. Obviously,
there will be cross-national differences in the benefits available to employees and the effects of employee benefits on absence would differ according to the type of benefit.
A literature search revealed a plethora of information on employee
benefits in practitioner journals but less research in academic journals.
The most frequently cited benefits in the United States and Canada are
drug and alcohol abuse rehabilitation, counseling for personal and family
problems, health facilities, and child care and eldercare benefits. In
England, ("Work/Life programs," 1996) employers offer special leave
comprising bereavement, medical, family sickness, domestic crisis, studying and examinations, caring, moving house, religious observance, and
marriage, but the outcomes of these special leave provisions were not
evaluated. In fact, Clause 3 of the 1995 European Community framework
agreement on parental leave stipulates that "member States and/ or social
partners shall take the necessary measures to entitle workers to take time
off from work ... on the grounds of force majeure for urgent family reasons in cases of sickness or accident making the presence of the worker
indispensable" ("Special-leave," 1997).
In the United States and Canada, both the practitioner and academic
literature suggest that absence is reduced with the provision of employee
assistance programs. In 1996, in a survey of 800 American companies with
a combined work force of 7 million, 50% of the organizations indicated
that employer-sponsored work-life programs led to decreases in absenteeism ("Work/Life programs," 1996). In a meta-analysis of 21 studies,
Unckless, Mathieu, and Kelley (1998) found that employee assistance
programs were particularly effective in reducing absenteeism. This effect
39
probably stems from direct reduction in constraints on ability to attend as

well as indirect improvements in attendance motivation.
Kossek and Nichol (1992) asserted that employees with available
benefits tend to be more loyal to employing organizations. Drawing on
social exchange theory, Eisenberger, Fasolo, and Davis-LaMastro (1990)
demonstrated that when employees feel that the organization is committed to them, they are likely to be less absent from work. Employee assistance programs may signal such commitment. The bottom line here is that
absence will be perceived as less legitimate when constraints against
attendance are removed and when organizational support helped to
remove them.
Friedman (1990) asserted that 12% of firms in the United States with
at least 100 employees grant child care benefits. One-fourth of them offer
on- or near-site facilities. Investigations of the effect of benefits on users
and nonusers of these benefits have reported uneven results. Milkovich
and Gomez (1976) found a significant relationship between center use and
absenteeism, with users exhibiting lesser absence than nonusers. Unlike
Milkovich and Gomez (1976), Goff, Mount, and Jamison (1990) reported
that the provision of on-site childcare did not reduce employee absenteeism. They indicated, however, that the quality of childcare irrespective
of the location tended to reduce work! family conflict and consequently
lowered absenteeism.
In Sweden and England, researchers have investigated the effects of
sickness insurance on absenteeism (Dean & Taylor-Gooby, 1986;
Johansson & Palme, 1993, 1996). Johansson and Palme (1996) stated that
in Sweden, sickness insurance is a compulsory component of their
National Insurance Scheme. The insurance allows employees to be absent
from work for up to 8 days without a physician's certificate. They also
indicated that 97% of blue-collar workers' absenteeism was covered by
sickness insurance. Obviously, this is not due to significant decrements in
the health of blue-collar workers and other employees in Sweden.
Interestingly, Johansson and Palme (1996) indicated that between 1990
and 1991, the government reduced the levels of sickness insurance contributions. This reduction, coupled with a rise in unemployment, led to a
considerable 8.1 % decline in absenteeism.
Watkins (1994:1) noted that according to the European Union
Statistics, the "British worker is second only to the Dutch in the European
'sick-note league.'" In addition, he pointed out that 60% of sickness
absence is for periods of less than 3 days, the period for which statutory
sick pay is covered. Consequently, in 1994 the government introduced
changes to the Statutory Sick Pay Bill. The 80% rebate paid to employers
for paying statutory sick pay on behalf of the government to employees
40
was abolished. Watkins (1994) noted that these changes were precipitated
by an increasing concern about the effects of statutory sick pay on
absenteeism.
In a comparative study of sickness insurance on absenteeism, Prins
(1990) as well as Prins and de Graaf (1986) asserted that unlike the United
States and Canada, employees in Belgium, Germany, and the Netherlands
are entitled to state-supported sickness insurance. They found that the
Netherlands had both a higher frequency of sickness absence and longerduration spells than either Belgium or Germany. These differences were
attributed to institutional disparity among the countries in the provision
of social insurance for sickness absence. Benefits in the Netherlands were
more generous than in the other countries. Moreover, eligibility criteria
were less stringent.
These findings concerning how national absence rates correspond to
the details of state-mediated sick leave plans parallel within-country findings that absenteeism is extremely sensitive to firm-level variations in sick
leave policies, control policies, and other incentives and disincentives
(Dalton & Mesch, 1991; Dalton & Perry, 1981; Ehrenberg, Ehrenberg, Rees,
& Ehrenberg, 1991; Kenyon & Dawkins, 1989; Ng, 1989). At both the
national and firm levels, it is likely that legitimization is a critical mediating variable. As economic and social provisions for absenteeism increase,
a signal is sent that absenteeism is a legitimate work behavior, not only
tolerated but also provided for (Kaiser, 1998).
Thus, while social support systems may serve to either increase or
decrease absenteeism, employees' perception of the efficacy of these
systems will influence their perceptions and behavior. In light of the
preceding, the following proposition is advocated:
Proposition 6: Employees' perception of the efficacy of social support will be

related to their perception of absence legitimacy at the individual and national
levels of analysis.
FUTURE THEORY AND RESEARCH

The cross-cultural model of absenteeism that we present in this
chapter is a multilevel, main effect model. As such, it is designed to introduce some of the factors that might explain differences in absenteeism
between cultures, using legitimacy as a mediating explanatory variable.
Although these variables are also expected to operate within cultures, this
is not the thrust of our argument. A reviewer of the chapter noted that
interesting questions may also be posed by developing cross-level models
41
(Klein et al., 1994; Klein & Kozlowski, 2000; Rousseau, 1985) of the impact
of culture, and we agree. Such models would posit main effects of culture
on individual absence or locate culture as a moderator of the relationship
between individual-level causes and absenteeism. In fact, we alluded to
one such moderator when we speculated how the tendency for women to
be absent more than men might be exacerbated in cultures with high
gender role differentiation.
Johns and Xie (1998) found several differences between Canadians
and Chinese in the People's Republic in terms of endorsing what constituted legitimate reasons for absenteeism. An extension of this finding is
that culture should moderate the relationship between potential causes
and the occurrence of absenteeism. For example, the needs of a distant
family member might be more likely to stimulate absenteeism in a more
collective culture than in an individualistic culture. Such more specific
hypotheses are not at odds with our more general propositions that
national cultures influence absence legitimacy through the mechanisms of
work centrality, self-control, time allocation, gender dynamics, and social
support.
Moving from theory to research, we have been conducting absence
research in Canada, the United States, China, Mexico, Trinidad, Nigeria,
Ghana, India, Pakistan, and Japan. It is well known that both cross-cultural
research and absenteeism research have special challenges associated
with them, and we will not review these here. One unique challenge when
the two are combined is accessing sites with relatively clean absenteeism
data that are otherwise reasonably comparable in terms of occupational
mix, gender mix, organization of work, and so on. We have not had great
luck in getting multinational firms to provide comparative cross-national
research sites; such access is almost always under local control. This
means that statistical control may be necessary to equate research sites
(e.g., Johns & Xie, 1998), a less palatable option than procedural control
via ideal sampling opportunities. In fact, Klein et al. (1999:244) posited
that "the analysis of multilevel data has been the topic of considerable
debate ... but perhaps even more daunting than multilevel data analysis is
the task of multilevel data collection." In a related vein, we have found
managers in less developed countries to be more receptive to absenteeism
research than North American managers. The reader can speculate on the
reasons for this.
A subtext to this chapter is that absence may have different meanings
in different cultures (see Johns & Nicholson, 1982). This having been
acknowledged, we would like to strongly emphasize one point: We have
not had particular difficulties asking respondents from various cultures to
reflect on absenteeism in questionnaires. We simply begin by defining
42
absence in a way that seems to be cross-culturally transportable (not showing up for scheduled work), adjusting the basic definition to reflect local
customs. For example, time off for jury duty is not absenteeism in Canada,
and time off to celebrate one's marriage is not absenteeism in China. Other
than these simple adjustments, we avoid attributions about cause in the
definition or measurement of absenteeism. Such attributions should be
restricted to predictor variables, not incorporated into the criterion.
CONCLUSION
The study of variations in the perceptions of the legitimacy of absenteeism across national cultures would seem to benefit both absenteeism
research and cross-cultural research. At the same time, it may have some
practical advantages for the management of attendance.
In the domain of absenteeism research, a cross-cultural perspective
virtually forces researchers to examine in detail the nonwork factors that
might affect attendance. As Johns (1997) notes, extant absence research is
highly deficient in this regard. Looking at absence cross-culturally, we
must confront off-the-job social influence and infrastructure differences
that surely affect absenteeism within our own cultures but have seldom
been studied. In addition, the cross-cultural approach highlights the more
general role that social influence may play in affecting attendance.
Although the value-added benefits of this theoretical position are now
well recognized (Johns, 1997,2001), the cross-cultural approach extends
and refines the role of social influence on absenteeism.
A cross-cultural approach has at least two other benefits for absenteeism research. The typical single-site research study is often dealing
with absence behavior that is highly constrained by social forces, industry norms, local absence culture, and organizational control systems (see
Johns, 1991). Studying absence cross-culturally provides requisite variety
in these ambient causal mechanisms such that one is able to examine less
constrained behavior. This provides both theoretical and methodological
leverage. In addition, and relatedly, cross-cultural research forces the
withdrawal researcher to attend more carefully to contextual forces on
withdrawal. It is our impression that such forces have been particularly
ignored in withdrawal research, including that on absenteeism.
In terms of cross-cultural research per se, the study of absenteeism
(and other so-called work withdrawal behaviors) may provide an interesting bridge between anthropological studies of mundane human behaviors (e.g., time usage) and psychological studies focused on hypothetical
constructs (e.g., time urgency). Because the act of absenting oneself from
work can be measured in a culture-free way, it is an attractive basis from
43
which to study organizational behavior cross-culturally. This can be contrasted with the study of concepts such as leadership, the meaning of
which is debated even within North American culture.
The cross-cultural study of absenteeism holds some relevance for the
practice of managing absence. One issue concerns the absence behavior of
migrants in an increasingly mobile global work force. How does an individual who has been socialized in a nation where absence is generally
viewed as a more legitimate behavior behave in a nation where it is
viewed as less so? The answer to this question is far from clear. Harrison,
Johns, and Martocchio (2000) speculated how ethnic diversity might
reveal normative differences about what constitutes a reasonable level of
absence, differences that might be especially salient in a team-oriented
work environment. For instance, employees with origins in collective cultures might be quite accepting of absence due to problems arising in their
extended families, a reason not acceptable to their individualistic work
colleagues or manager.
On the other hand, according to the cultural persistence and adaptation hypothesis (Joy, 1989), although individuals may adapt their cultural
values to suit different cultural settings, their original cultural values tend
to persist. It is therefore anticipated that in the short term, while individuals would adapt their absence behavior to suit the prevailing ethos in the
current culture in which they are employed, their perceptions of absence
legitimacy may not change. Thus, when such individuals return to their
country of origin, they would revert to their original absence behavior.
However, it is suggested that over time due to socialization in the new
culture, individuals would undergo a permanent change in their perceptions of absence legitimacy as well as their absence behavior.
The cross-cultural aspects of absence are also relevant to firms that
conduct global business. If our speculations are correct, a "global corporate attendance policy" is probably an ill-advised human resources initiative. Rather, expatriate managers need to be sensitized to indigenous
views about the legitimacy of absence. Knowing those views, attendance
management systems can be devised that correspond to and take advantage of the causal mechanisms behind them.
As we noted at the beginning of this chapter, work withdrawal
behaviors have been all but ignored in cross-cultural research on work.
We hope that we have demonstrated the value of examining the cultural
manifestations of absenteeism.
The preparation of this chapter was supported by
grant OO-ER-0506 from Quebec's Fonds pour la Formation de Chercheurs
et l' Aide a la Recherche and grant 410-99-1491 from the Social Sciences
and Humanities Research Council of Canada.
ACKNOWLEDGMENTS.
44
REFERENCES
Abrams, D., Ando, K., & Hinkle, S. (1998). Psychological attachment to the group:
Cross-cultural differences in organizational identification and subjective
norms as predictors of workers' turnover intentions. Personality and Social
Psychology Bulletin, 24, 1027-1039.

Ali, A., & AI-Shakhis, M. (1989). The meaning of work in Saudi Arabia.
International Journal of Manpower, 10, 26-32.
Aycan, Z., Kanungo, R N., Mendonca, M., Yu, K., Deller, J., Stahl, G., and Kurshid,
A. (2000). Impact of culture on human resource management practices: A tencountry comparison. Applied Psychology: An International Review, 49,192-220.
Baba, V. V. (1989). Central life interests and job involvement: An exploratory study
in the developing world. International Journal of Comparative Sociology, 3D,
181-194.
Baba, V. v., & Harris, M. J. (1989). Stress and absence: A cross-cultural perspective.
Research in Personnel and Human Resources Management, Suppl.1, 317-337.
Baltes, B. B., Briggs, T. E., Huff, J. w., Wright, J. A., & Neuman, G. A. (1999).
Flexible and compressed workweek schedules: A meta-analysis of their
effects on work-related criteria. Journal of Applied Psychology, 84, 496-513.
Beach, B. (1989). Integrating work and family life. The home-working family. Albany:
State University of New York Press.
Blau, G. J. (1987). Locus of control as a potential moderator of the turnover
process. Journal of Occupational Psychology, 60, 21-29.
Brooke, P. P., & Price, J. L. (1989). The determinants of employee absenteeism: An
empirical test of a causal model. Journal of Occupational Psychology, 62, 1-19.
Byrne, N. (1982). Time construction in everyday life. Paper presented at the meeting
of the American Sociological Association, San Francisco.
Central Statistical Office. (1993). Social indicators of the Republic of Trinidad and
Tobago. Port of Spain: Office of the Prime Minister.
Chadwick-Jones, J. K., Nicholson, N., & Brown, C. (1982). Social psychology of absenteeism. New York: Praeger.
Clay, J. M., & Stephens, E. C. (1994). An analysis of absenteeism arbitration cases:
Factors used by arbitrators in making decisions. International Journal of
Conflict Management,S, 130-142.
Coleman, D. E, Irving, G. P., & Cooper, C. L. (1999). Another look at the locus of
control-organizational commitment relationship: It depends on the form of
commitment. Journal of Organizational Behavior, 20, 995-1001.
Conte, J. M., & Jacobs, R R (1999). Temporal and personality predictors of absence and
lateness. Paper presented at the annual convention of the Society for Industrial
and Organizational Psychology, Atlanta.
Conte, J. M., Mathieu, J. E., & Landy, E J. (1998). The nomological and predictive
validity of time urgency. Journal of Organizational Behavior, 9,1-13.
Cote, D., & Haccoun, R R (1991). L'absenteism des femmes et des hommes: Une
meta-analyse. Canadian Journal of Administrative Sciences, 8,130-139.
Dalton, D. R, & Mesch, D. J. (1990). The impact of flexible scheduling on employee
attendance and turnover. Administrative Science Quarterly, 35, 370-387.
45
Dalton, D. R., & Mesch, D. J. (1991). On the extent and reduction of avoidable
absenteeism: An assessment of absence policy provisions. Journal of Applied

Dalton, D. R., & Perry, J. L. (1981). Absenteeism and the collective bargaining
agreement: An empirical test. Academy of Management Journal, 24, 425-431.
Dean, H., & Taylor-Gooby, P. (1990). Statutory sick pay and the control of sickness
absence. Journal of Social Policy, 19, 47-67.
Denhardt, R. B. (1986). Time consequences in organizational life. Paper presented at
the meeting of the Academy of Management, Chicago.
Drago, R., & Wooden, M. (1992). The determinants of labor absence: Economic factors and workgroup norms across countries. Industrial and Labor Relations
Review, 45, 764-778.

Dubin, R. (1956). Industrial workers' worlds: A study of the central life interests of
industrial workers. Social Problems, 3,131-142.
Dubin, R, Champoux, J. E., & Porter, L. W. (1975). Central life interests and organizational commitment of blue-collar and clerical workers. Administrative
Science Quarterly, 20, 411-421.
Earley, c., & Gibson, C. B. (1998). Taking stock in our progress on individualismcollectivism: 100 years of solidarity and community. Journal of Management,
24, 265-304.
Economic and Social Survey of Jamaica. (1995). Planning Institute of Jamaica. Kingston.
Ehrenberg, R G., Ehrenberg, R A., Rees, D. I., & Ehrenberg, E. L. (1991). School
district leave policies, teacher absenteeism, and student achievement. Journal
of Human Resources, 26, 72-105.
Eisenberger, R, Fasolo, P., & Davis-LaMastro, V. (1990). Perceived organizational
support and employee diligence, commitment, and innovation. Journal of
Employers' Consultative Association of Trinidad and Tobago. (1993). Absenteeism

in the workplace: A study of the prevalent costs and mediation of the
employee'S absence from duty. Port of Spain.
England, G. W., & Harpaz, I. (1990). How working is defined: National contexts
and demographic and organizational role influences. Journal of Organizational
Behavior, II, 253-266.
England, G. w., & Misumi, J. (1987). Work centrality in Japan and the United
States. Journal of Cross-Cultural Psychology, 17, 399-416.
Farrell, G. M., & Mudrack, P. E. (1992). Academic involvement and the nontraditional student. Psychological Reports, 71, 707-713.
Frayne, C. A., & Latham, G. P. (1987). Application of social learning theory to
employee self-management of attendance. Journal of Applied Psychology, 72,
387-392.
Friedland, J. (1994). Career lady's curse: Japanese recession is a setback for

women. Far Eastern Economic Review, 27, 44-45.
Friedman, D. E. (1990). Work and family: The new strategic plan. Human Resource
Planning Journal, 13, 79-89.

Friedman, M., & Rosenman, R H. (1974). Type A behavior and your heart. New York:
Knopf.
46
Friend, K (1982). Stress and performance: Effects of subjective workload and time
urgency. Personnel Psychology, 35, 623--633.
Gale, E. K (1993). Social influences on absenteeism. Unpublished doctoral dissertation, Purdue University.
Gellatly, I. R (1995). Individual and group determinants of employee absenteeism:
Test of a causal model. Journal of Organizational Behavior, 16, 469-485.
Gellatly, I. R, & Luchak, A A (1998). Personal and organizational determinants of
perceived absence norms. Human Relations, 51,1085-1102.
Glass, D. c., Snyder, M. L., & Hollis, J. F. (1974). Time urgency and the Type A
behavior pattern. Journal of Applied Social Psychology, 4, 125-140.
Goff, S. J., Mount, M. K, & Jamison, R L. (1990). Employer supported child care,
work/ family conflict, and absenteeism: A field study. Personnel Psychology, 43,
793-809.
Griindemann, R W. M., de Winter, C. R, & Smulders, P. G. W. (1994). Absenteeism in
the European Union (Working paperWjPj94j29jEN). Shankill, Co. Dublin, Ireland:
European Foundation for the Improvement of Living and Working Conditions.
Hackett, RD., Bycio, R M., & Guion, R M. (1989). Absenteeism among hospital
nurses: An idiographic longitudinal analysis. Academy of Management Journal,
32, 424-453.
Hackman, J. R (1992). Group influences on individuals in organizations. In
M.D. Dunnette & L.M. Hough (Eds.), Handbook of industrial and organizational
psychology (2nd ed., Vol. 3, pp.199-267). Palo Alto, CA: Consulting
Psychologists Press.
Hall, E. T. (1966). The hidden dimension. Garden City, NY: Doubleday.
Hall, E. T. (1973). The silent language. Garden City, NY: Anchor Press.
Hall, E. T. (1983). The dance of life: The other dimension of time. Garden City, NY:
Anchor Press/Doubleday.
Hall, E. T., & Hall, M. R. (1990). Understanding cultural differences. Yarmouth, ME:
Intercultural Press.
Hamid, P. (1994). Self-monitoring, locus of control, and social encounters of Chinese
and New Zealand students. Journal of Cross-Cultural Psychology, 25, 353-368.
Harari, H., Jones, c. A, & Sek, H. (1988). Stress syndromes and stress predictors
in American and Polish college students. Journal of Cross-Cultural Psychology,
19,243-255.
Harrison, D. A. (1995). Volunteer motivation and attendance decisions:
Competitive theory testing in multiple samples from a homeless shelter.
Journal of Applied Psychology, 80, 371-385.
Harrison, D. A, Johns, G., & Martocchio, J. J. (2000). Changes in technology, teamwork, and diversity: New directions for a new century of absenteeism
research. Research in Personnel and Human Resources Management, 18, 43-91.
Harrison, D. A., & Martocchio, J. J. (1998). Time for absenteeism: A20-year review
of origins, offshoots, and outcomes. Journal of Management, 24, 305-350.
Harvey,J., & Nicholson, N. (1999). Minor illness as a legitimate reason for absence.
Journal of Organizational Behavior, 20, 979-993.
Hein, C. (1984). Jobs for the girls: Export manufacturing. International Labour
Review, 123, 251-265.
47
Hofstede, G. (1980). Culture's consequences: International differences in work-related

values. Beverly Hills: Sage.
Hofstede, G. (1991). Culture and organizations: Software of the mind. New York:
McGraw-Hill.
Ivancevich, J., & Lyon, H. (1977). The shortened workweek: A field experiment.
Iverson, R. D., Buttigieg, D. M., & Maguire, C. (1999). Absence culture: The effects of
union membership within work groups and industrial relations climate. Paper presented at the meeting of the Academy of Management, Chicago.
Johansson, P., & Palme, M. (1993). The effect of economic incentives on worker absenteeism: An empirical study using Swedish microdata. Umea Economic Studies,
314, University of Umea.
Johansson, P., & Palme, M. (1996). Do economic incentives affect work absence?
Empirical evidence using Swedish micro data. Journal of Public Economics, 59,
195-218.
Johns, G. (1991). Substantive and methodological constraints on behavior and attitudes in organizational research. Organizational Behavior and Human Decision
Processes, 49, 80-104.
Johns, G. (1994). Medical, ethical, and cultural constraints on work absence and attendance. Presentation made at the International Congress of Applied
Psychology, Madrid.
Johns, G. (1997). Contemporary research on absence from work: Correlates, causes
and consequences. International Review of Industrial and Organizational
Johns, G. (2001). The psychology of lateness, absenteeism, and turnover. In
N. Anderson, D. S. Ones, H. K. Sinangil, & C. Viswesvaran (Eds.), Handbook of
industrial, work and organizational psychology (Vol. 2, pp. 232-252). London: Sage.
Johns, G., & Nicholson, N. (1982). The meanings of absence: New strategies for
theory and research. Research in Organizational Behavior, 4, 127-172.
Johns, G., & Xie, J. L. (1998). Perceptions of absence from work: People's Republic
of China versus Canada. Journal of Applied Psychology, 83, 515-530.
Jones, J. M. (1988). Cultural differences in temporal perspectives. Instrumental
and expressive behaviors in time. In J. E. McGrath (Ed.), The social psychology
of time (pp. 21-38). Beverly Hills: Sage.
Joy, A. (1989). Ethnicity in Canada: Social accommodation and cultural persistence
among the Sikhs and the Portuguese. New York: AMS Press.
Judge, T. A., & Martocchio, J. J. (19%). Dispositional influences on attributions
concerning absenteeism. Journal of Management, 22, 837-861.
Judge, T. A., Martocchio, J. J., & Thoresen, C. J. (1997). Five-factor model
of personality and employee absence. Journal of Applied Psychology, 82, 745-755.
Kacapyr, E. (1996, June). Jumping for joy. American Demographics, 10-11.
Kaiser, C. P. (1998). Dimensions of culture, distributive principles, and decommodification: Implications for employee absence behavior. Journal of SocioEconomics, 27, 551-564.
Kanungo, R. N. (1982). Work alienation: An integrative approach. New York: Praeger.
48
Kanungo, R N. (1990). Work alienation in developing countries: Western models

and Eastern realities. In A. Jaeger & R N. Kanungo (Eds.), Management in
developing countries (pp. 93-208). London: Routledge.
Keller, R T. (1984). The role of performance and absenteeism in the prediction of
turnover. Academy of Management Journal, 27,176-183.
Kenyon, P., & Dawkins, P. (1989). A time series analysis of labour absence in
Australia. Review of Economics and Statistics, 71, 232-239.
Kim, J. S., & Campagna, A. E (1981). Effects of flexitime on employee attendance
and performance: A field experiment. Academy of Management Journal, 24,
729-741.
Klein, K. J., Dansereau, E, & Hall, R J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19,
195-229.
Klein, K. J., & Kozlowski, S. W. J. (Eds.). (2000). Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions. San Francisco:
Jossey-Bass.
Klein, K. J., Tosi, H., & Cannella, A. A. (1999). Multilevel theory building: Benefits,
barriers, and new developments. Academy of Management Review, 24, 248-254.
Kluckhohn, E R., & Strodtbeck, E L. (1961). Variations in value orientations. New
York: Row, Peterson.
Kossek, E. E., & Nichol, V. (1992). The effects of on-site child care on employee attitudes and performance. Personnel Psychology, 45, 485-509.
Kuzmits, E E. (1995). Differences in incidence of absenteeism and discipline
between Vietnamese and non-Vietnamese employees. The International Journal
of Organizational Analysis, 3, 303-313.
Landy, J. E, Rastegary, H., Thayer, J., & Colvin, C. (1991). Time urgency: The construct and its measurement. Journal of Applied Psychology, 76,644-657.
Lefcourt, M. M. (1992). Durability and impact of the locus of control construct.

Levine, R (1997). A geography of time. New York: Basic Books.
Levine, R, & Wolffe, E. (1985). Social time: The heartbeat of culture. Psychology
Today, pp. 28-35.
Lewis, S. (1992). Work and families in the United Kingdom. In S. Zedeck (Ed.),
Work, families, and organizations (pp. 395-431). San Francisco: Jossey-Bass.
McGinnies, E., Nordholm, L. A., Ward, C. D. & Bhanthumnavin, D. L. (1974). Sex
and cultural differences in perceived locus of control among students in five
countries. Journal of Consulting and Clinical Psychology, 42, 451-455.
McGrath, J. E., & Rotchford, N. L. (1983). Time and behavior in organizations.
Research in Organizational Behavior, 5, 57-101.
McGuire, J. B., & Liro, J. R (1987). Absenteeism and flexible work schedules. Public
Personnel Management, 16, 47-59.
Major, B. (1993). Gender, entitlement, and the distribution of family labour. Journal
of Social Issues, 49,141-159.
Markham, S. E., & McKee, G. H. (1995). Group absence behavior and standards: A
multilevel analysis. Academy of Management Journal, 38,1174-1190.
Martocchio,
J. J.
49
(1994). The effects of absence culture on individual absence.
Human Relations, 47, 243-262.
Martocchio, J. J., & Harrison, D. A (1993). To be there or not to be there?

Questions, theories, and methods in absenteeism research. Research in
Personnel and Human Resources Management, 11,259-329.
Mathieu, J. E., & Kohler, S. S. (1990). A cross-level examination of group absence

influences on individual absence. Journal of Applied Psychology, 75, 217-220.
Maurer, J. G., Vredenburgh, D. J., & Smith, R. L. (1981). An examination of the centrallife interest scale. Academy of Management Journal, 24,174-182.
Meisenheimer, J. R., Jr. (1990, August). Employee absences in 1989: A new look at
data from the CPS. Monthly Labor Review, pp. 28-33.
Milkovich, G., & Gomez, L. R. (1976). Childcare and selected work behaviors.
Academy of Management Journal, 19, 111-115.

Morgan, L. G., & Herman, J. B. (1976). Perceived consequences of absenteeism.
MOW International Research Team. (1987). The meaning of work. New York:
Academic Press.
Ng, I. (1989). The effect of vacation and sickleave policies on absenteeism.
Canadian Journal of Administrative Sciences, 6(4), 18-27.
Nicholson, N. (1975). Industrial absence as an indicant of employee motivation and job

satisfaction. Unpublished doctoral dissertation, University of Wales, Cardiff.
Nicholson, N., & Jackson, P. (1978). Shiftwork and absence: An analysis of temporal trends. Journal of Occupational Psychology, 51, 127-137.
Nicholson, N., & Johns, G. (1985). The absence culture and the psychological contract-Who's in control of absence? Academy of Management Review, 10, 397-407.
Nihon Horei Kyokai. (1989). Chingin Sensasu (Wage census). Tokyo: Author.
Nord, W. R., & Costigan, R. (1973). Worker adjustment to the four-day week: A
longitudinal study. Journal of Applied Psychology, 58, 60-66.
Parsons, 0., & Schneider, J. (1974). Locus of control in university students from
Eastern and Western societies. Journal of Consulting and Clinical Psychology, 42,
456-461.
Pierce, J. L., Newstrom, J. W., Dunham, R. B., & Barber, A E. (1989). Alternative
work schedules. Boston: Allyn & Bacon.
Pleck, J. H. (1985). Working wives/working husbands. Beverley Hills: Sage.
Pocock, S. J. (1973). Daily variations in sickness absence. Applied Statistics, 22,
375-392.
Prins, R. (1990). Sickness absence in Belgium, Germany and the Netherlands: A comparative study. Amsterdam: Nederlands Instituut voor Arbeidsomstandigheden.
Prins, R., & deGraaf, A (1986). Comparison of sickness absence in Belgian,
German and Dutch firms. British Journal of Industrial Medicine, 43, 529-536.
Rahman, M. A, & Kumar, N. J. (1984). Absenteeism as related to locus of control.
Psychological Studies, 29,73-75.

Reitz, J., & Groff, G. K. (1974). Economic development and belief in locus of
control among factory workers in four countries. Journal of Cross-Cultural
Psychology,S, 344-355.
50
Renn, R. W., & Vandenberg, R. J. (1991). Differences in employee attitudes and

behaviors based on Rotter's (1966) internal-externallocus of control: Are they
all valid? Human Relations, 44,1161-1178.
Robinson, J. P., Converse, P. E., & Szalai, A. (1972). Everyday life in twelve
countries. In A. Szalai (Ed.), The use of time: Daily activities of urban and suburban populations in twelve countries (pp. 113-144). The Hague: Mouton.
Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors:
A multidimensional scaling study. Academy of Management Journal, 38, 555-572.
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of
reinforcement. Psychological Monographs, 80 (Whole No. 609).
Rousseau, D. M. (1978). Relationship of work to nonwork. Journal of Applied
Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and
cross-level perspectives. Research in Organizational Behavior, 7, 1-37.
Schriber, J., & Gutek, B. (1987). Some time dimensions of work: Measurement of
an underlying aspect of organizational culture. Journal of Applied Psychology,
72, 642-650.
Scott, J., Alwin, D. E, & Braun, M. (1996). Generational changes in gender-role attitudes: Britain in a cross-national perspective. Sociology, 30, 471-492.
Shapira, Z., & Griffith, T. L. (1990). Comparing the work values of engineers with
managers, production, and clerical workers: A multivariate analysis. Journal
of Organizational Behavior, 11, 281-292.
Schriber, J. B., & Gutek, B. A. (1987). Some time dimensions of work: Measurement
of an underlying aspect of organizational culture. Journal of Applied
Special-leave for personal reasons-a survey. (1997). IRS Employment Trends,
606,5-10.
Spector, P. E. (1982). Behavior Bulletin or in organizations as a function of employees' locus of control. Psychological Bulletin, 91, 482-497.
Taylor, P. J., Pocock, S. J., & Sergean, R. (1972a). Absenteeism of shift and day
workers. British Journal of Industrial Medicine, 29, 203-208.
Taylor, P. J., Pocock, S. J., & Sergean, R. (1972b). Shift work and day workers'
absence: Relationship with some terms and conditions of service. British
Journal of Industrial Medicine, 29, 338-340.
The Statistical Institute of Jamaica. (1994). Demographic statistics. Kingston.
Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview.
Trice, A. (1987). Academic locus of control and required and optional class attendance. Perceptual and Motor Skills, 64, 1158.
Trice, A., & Hackburt, L. (1989). Academic locus of control, Type A behavior, and
college absenteeism. Psychological Reports, 65,337-338.
Trompenaars, E (1993). Riding the waves of culture. London: The Economist Book
Publishers.
Unckless, A. L., Mathieu, J. E., & Kelley, P. L. (1998, August). The relative
effectiveness of absence interventions: A meta-analysis. Paper presented at the
annual meeting of the Academy of Management, San Diego.
U.S. Department of Commerce. (1992).
51
u.s. Bureau of Labor Statistics. (1982).

Watkins, I. (1994, April). Sick excuses for swanning around. Personnel Management,
26,42-45.
Winfield, F. E. (1995). General dynamics in the Navajo nation. In C. P. Harvey &
M. J. Allard (Eds.), Understanding diversity: Readings, cases, and exercises
(pp. 186-197). New York: HarperCollins.
Wolfgang, A., & Weiss, D. (1980). A locus of control and social distance comparison of Canadian and West Indian-born students. International Journal of
Intercultural Relations, 4, 295-305.
Work/Life programs on the rise, survey finds. (1996, August). National Underwriters,
pp.9-1O.
Xie, J. L., & Johns, G. (2000). Interactive effects of absence culture salience and
group cohesiveness: A multi-level and cross-level analysis of work absenteeism in the Chinese context. Journal of Occupational and Organizational
Youngblood, S. A. (1984). Work, nonwork, and withdrawal. Journal of Applied
Psychology, 69,106-117.
3
The Many Faces of Voluntary
Employee Turnover
A Multifacet and Multilevel Perspective
Moshe Krausz
Recent statistics show that throughout their careers, American workers

hold an average of eight jobs (Wegmann, 1991). This implies that employees leave their jobs seven times, on average, throughout their active career
life. It also means that seven companies will have to deal with the turnover
of that employee. The impact of turnover also spreads to other entities
such as the employees' families, their communities, and the wider society,
affecting economic, social, and cultural systems. For theoreticians and
researchers, the obvious implication is that employee turnover has many
faces and levels and has, therefore, to be analyzed from a multilevel perspective. The wider theoretical and practical implication is that multilevel
constituencies affect turnover and are potentially being affected by it.
Turnover may be caused by personal and personality factors, family
issues, work team relations, career plans, the overall level of human
resource demand and supply in the community or the occupational field,
as well as organizational structure and strategic human resource practices
(Huselid, 1995). Previous experiences of organizational departure and the
socialization into the next organization (Adkins, 1995) can also be expected
to affect intentions, as well as actual turnover from the current job.
For many years, the psychological study of the employee turnover
process has centered around the two constructs of desirability of moving
MOSHE KRAUSZ Deparment of Psychology, Bar-Han University, Ramat-Gan, Israel 52900
53
54
Moshe Krausz
and ease of mobility, or their derivatives, that have been proposed by

March and Simon (1958). Years later, Lee and Mitchell (1994:85) asserted
that "researchers and managers can't predict much better than chance
whether a particular employee is going to quit, even after 30 years of
rigorous research effort. Turnover research badly needs some new
theory." One of the efforts in that direction was indeed made by those
authors (Lee & Mitchell, 1994; Lee, Mitchell, Wise, & Fireman, 1996).
The aim of the current chapter is neither to outline a new theory nor
to present a comprehensive and in-depth overview of existing theoretical
approaches to the topic of employee turnover. Instead, while some previous models are discussed, the main goal is to suggest some new ways of
looking at the employee turnover phenomenon in light of emergent realities in the world of work, organizations, and careers, and to suggest some
less researched approaches in this field.
In addition to a brief overview of existing approaches to employee
turnover, this chapter presents several new research issues that must be
dealt with in light of the emerging changes. In addition, the chapter highlights some management problems that are prevalent in organizations
today. Each of the following sections focuses on a certain modality of
turnover with respect to its antecedents, manifestations, and outcomes.
The many variations that are described highlight the fact that turnover is
far from being a simple phenomenon with one straightforward definition.
The following illuminates the many faces, facets, and levels affecting and
being affected by turnover.
THE CHANGING MEANING OF EMPLOYEE TURNOVER:

MATERIAL AND PSYCHOLOGICAL ASPECTS
Although the literature has rarely bothered to define the concept of

voluntary turnover, it is commonly viewed as the termination of formal
relations between an employee and an organization. This definition is
rather narrow and not up-to-date in light of current societal and organizational realities. Global competition and fast-changing technologies have
raised the need for more flexibility in the operation of organizations in
order to survive and thrive. One result of that need is the increasing rate
of new and diverse employment forms. The changes relate to the psychological contract between individuals and organizations (McLean Parks,
Kidder, & Gallagher, 1998; Rousseau, 1995) as well as to the tangible
and concrete employment contracts. There has been a steady and steep
increase in the number of employees working under personal and mostly
Voluntary Employee Turnover
55
temporary contracts; they are labeled contingent employees (Belous, 1989;

Cohany, 1998). Under the new circumstances and based on changes in the
organization's environment, it has become quite common for a person
and an organization to terminate relationships under one contract and
resume contact under another. It is not rare to encounter people who
either voluntarily or involuntarily sever their ties with an organization on
internal permanent contracts to resume the links with the same organization on temporary or consultancy contracts. The meaning of turnover
itself seems to take a new form under contingent arrangements such as
working through external agencies, outsourcing, leased work, or on-call.
This may be portrayed by examining the nature of the employment contract of temporary help agency workers, the fastest-growing sector among
contingent forms (Belous, 1989; Cohany, 1998) which is, in fact, the fastestgrowing sector of employment in general in the United States. Under that
arrangement, individuals are hired and paid by the external agency but
their work is performed for another organization-the client company,
mostly on the client's premises (Krausz, Brandwein, & Fox, 1995). In reality, many of the agency employees are not strictly temporary in terms of
the duration of the contract. Many agency employees are hired for
lengthy projects or long-term replacement so that their association with
the agency as well as with the client may extend for many months or even
several years (Krausz, Brandwein et aI., 1995). If turnover occurs among
such employees, who is the forsaken organization-is it the agency or the
client company? Indeed, it may be asked what turnover actually means in
that case. The answer may depend on the individual employee's decision.
If the temporary employee decides to leave a specific client, then she or he
"returns" to the employee pool of the external agency which implies that
there has been neither objective nor psychological turnover from the
agency. The employee is now available for placement with another client
who may have been searching for an employee with just his or her mix of
skills and abilities. For the agency this may imply higher income since the
new client may be willing to pay a higher price for the services rendered
by that employee. In a different situation, the employee may voluntarily
terminate her or his ties with the agency itself, a situation where both the
agency and the client had suffered turnover. Both of those constituencies
have to seek replacement, a process that is time consuming and has clear
financial ramifications such as the cost of recruitment itself as well as costs
due to hampered organizational efficiency throughout that process.
Another alternative may exist: Despite various limiting and binding
provisions in this triangular contract involving the individual, the agency,
and the client, employees sometimes defect from their agency and defy
the contract by continuing to work with the same client on a two-party
56
Moshe Krausz
contract between the employee and the organization (formerly, the client
company of the external agency).
For individuals working under triangular arrangements, the
turnover process may be more complex than for employees on more traditional arrangements. It is likely that the cognitive processes that have
been proposed in Mobley's (1977) withdrawal model are duplicated:
Dissatisfaction, thoughts of quitting, evaluation of alternatives (jobs?
client organizations? agencies?), and withdrawal intentions need to be
considered for both entities-the client company and the external agency.
The above analysis focused on formal and tangible losses and gains
for the involved parties. Turnover, however, also incorporates psychological losses and gains. The decision whether termination relates to the
client company alone or to the agency as well, has implication for the individuals, for their supervisors from the client organization and from the
external agency, and for two social networks-that which prevails in the
client company and the other involving colleagues employed by the same
agency who may be working for different client companies.
The implication of the above for researchers is that both a conceptual
and operational definitions of turnover need to be explicitly formulated
for each study of employee turnover. Only in that way can valid comparisons be made between findings of various studies. One implication for
management is that a more differentiated outlook is needed when considering recruitment and retention of employees. Whereas the financial'
cost of both recruitment and turnover of contingent employees is lower
than for core employees, the psychological costs and benefits that are
involved are not yet sufficiently understood due to the scarcity of research
on behavioral aspects involved in the employment of workers on atypical
contracts. Are the psychological losses associated with the departure of
agency workers, who often work side by side with core employees, similar to those involved in turnover of core colleagues?
THE ROLE OF SATISFACTION AND JOB ALTERNATIVES IN

TURNOVER
Satisfaction and Job Alternatives in the Context of Progression
Models
Mobley (1977) presented a model of intermediate linkages connecting satisfaction with ti.lrnover. In that model, turnover is an intrapsychic
developmental proCtSS that is initiated by the individual's dissatisfaction
with a current job or job environment. The model posits that the common
57
turnover process starts with evaluation of one's job and the ensuing level
of satisfaction. Perceptions of work alternatives, search of such alternatives, and their evaluation are further links in the chain that eventually
leads to turnover. Mobley acknowledges the possibility of certain variations in the temporal order of the various steps but in general, the process
is assumed to be developmental and linear.
Herzberg, Mausner, Peterson, and Capwell (1957) proposed another
outlook on progression. In contrast to Mobley's (1977) intrapsychic model,
their idea suggests that withdrawal from an unsatisfying work environment progresses from mild to more severe modes of behavior. The progression starts with lateness, continues to absenteeism, and eventually
culminates in turnover. Several reasons for such a progression may be proposed. First, actual turnover in the sense of dissociation of formal ties with
the organization is deemed by the individual as risky, difficult, or unrealizable. In that case, the employee may resort to easier solutions in the form of
lower and less complicated forms of withdrawal that are used as substitutes
for actual departure. A second rationale for such progression is the level of
discontent with the job. Mild dissatisfaction does not warrant turnover
from the organization. Only when satisfaction has further deteriorated and,
in addition, the mild withdrawal forms no longer provide effective coping
mechanisms, a higher (more severe) form of withdrawal may be sought.
Alternative job opportunities are not necessarily a factor that employees
consider in the initial and milder levels of dissatisfaction. This factor may
be probed only when the lower withdrawal forms have failed. Empirical
literature adopting the progression notion has been quite scant. Some
studies, however, found evidence of withdrawal progression (Krausz,
Koslowsky, & Eiser, 1998; Wolpin, Burke, Krausz, & Freibach, 1988).
Differential Importance of Satisfaction and Job Alternatives
An important conceptual treatise presented by Hulin, Roznowski,
and Hachiya (1985) proposed that the two constructs-satisfaction
and employment opportunities-may operate differently among different
populations. Thus, for example, turnover of marginal and temporary
employees may be more influenced by perceived opportunities than by
job satisfaction whereas the two variables will play an important role
among permanent and core workers. Peters, Jackofsky, and Slater (1981)
also found that elements of Mobley'S (1977) model significantly predicted
turnover among full-time but not part-time workers.
Hulin (1991) later suggested that turnover should be seen as one
out of several behavioral manifestations of an underlying adaptation/
withdrawal construct.
58
Moshe Krausz
DISSATISFACTION WITHOUT TURNOVER
Literature reviews show that most of the negative correlations

between satisfaction and turnover have been modest (Carsten & Spector,
1987; Crampton & Wagner, 1994) insinuating that even high levels of dissatisfaction do not necessarily result in employee turnover. Those findings
point out that the turnover process is more complex than a simple reaction
to an unsatisfying work environment and that many antecedents have a
potential to affect this process. Even when dissatisfaction is extremely
high, leaving a job in an organization is a radical reaction which may be
problematic or even traumatic from the individual's perspective (Landy,
1989). A wide array of demographic and personality factors deter employees from leaving a job and the organization. For example, older and more
tenured employees were found to be higher in organizational commitment
and lower in turnover compared to younger and less tenured workers
(Mathieu & Zajac, 1990). Similarly, certain personality traits and factors,
most notably openness for change, neuroticism, and conscientiousness,
three of the Big 5 personality factors, have been shown to be differentially
associated with levels of turnover (Barrick & Mount, 1991).
The major environmental factors that may block or deter dissatisfied
employees from turnover are the objective conditions in the labor market
and the subjective perception of availability of employment alternatives.
Indeed, the literature shows a much higher negative correlation between
satisfaction and turnover in low compared to high unemployment situations (Carsten & Spector, 1987). It should be noted, though, that perceptions of labor realities are not a perfect reflection of the objective
conditions, because (1) employees may not be aware of the wider picture
and (2) due to interactions between their personality characteristics and
the environmental realities; employees low in general openness to change
may be partly blind to their surroundings or may actively distort them.
TURNOVER WITHOUT DISSATISFACTION
Whereas dissatisfaction without turnover is a very common phenomenon, it seems that turnover without prior dissatisfaction is a less
prevalent situation. Three types of such turnover are proposed.
Impulsive/N on-Satisfaction-Dependent Turnover
Mobley (1977) acknowledged the possibility that the process
leading to turnover may be less rational and calculated than implied by
59
the intermediate linkages model. At times, the process is more impulsive,

initiated by factors such as unsolicited job offers. Such an offer, rather
than job satisfaction, may act as a catalyst for the chain of cognitive
processes assumed by the model.
A more developed model-the unfolding model-has been presented by Lee and Mitchell (1994) and empirically tested by Lee et al.
(1996). The model postulates that turnover occurrences may be an outcome of what the authors designate "shock to the system," the system
being the person's beliefs or values. In some respects, "shock to the system" resembles the concept of "contract violation" (Robinson, 1996;
Rousseau, 1995) as well as the concept of "stimulus event" (Rosse &
Miller, 1984). Lee and Mitchell (1994) argue that their concept differs from
that of Rosse and Miller (1984). Whereas the "stimulus event" pertains
even to mild but noticeable events, a shock to the system is a "jarring
event." Another distinction between the two models is that Rosse and
Miller (1984) contend that all external events are filtered through the individual's perceptions whereas Lee and Mitchell assert that in some cases,
the effect of the objective environment is more direct. The difference, then,
is mainly with respect to the magnitude of the environmental stimulus.
It may also be proposed that regardless of a belief or value system
(as in Lee & Mitchell, 1994), a shock may be fundamentally emotional
such as when as part of work, an employee (police officer, first-aid paramedic) is faced with an exceptionally brutal crime or a shocking accident
that prompts an immediate quitting decision. In such situations, the issue
is neither the perception that "I could have been better off" (Rosse &
Miller, 1984) nor a conflict with a prevailing belief and value systems
(Lee & Mitchell, 1984; Lee et aI., 1996), but a more direct effect on the
person's physical and emotional level of tolerance.
The main conceptual contribution of the unfolding model is that
turnover may be entirely unrelated to continuous work dissatisfaction
and is not the culmination of a progressive process. Furthermore, some
turnover events may seem irrational as they occur on the spur of the
moment, a situation that is less well explained by traditional turnover
models.
Turnover Incited by Others' Behavior
Turnover of co-workers may itself incite turnover intentions and
behavior by employees in the same work environment. Ostensibly, this
kind of environmental occurrence is merely an example of a "jarring
event" that causes a "shock to the system" (Lee & Mitchell, 1994) or
a "stimulus event" (Rosse & Miller, 1984). There are grounds to suggest,
60
Moshe Krausz
however, that co-workers' departure may be a unique cause of turnover

that is not necessarily encompassed by the previous concepts, despite the
essential resemblance.
There has been growing interest in the phenomenon of burnout contagion (Schaufeli & Enzmann, 1998). Golembiewski, Munzenrider, and
Stevenson (1986:184) reviewed data on that issue and noted that "very
high and very low scores on burnout tend to concentrate to a substantial
degree" and further concluded that the findings are an indication of
"contagion" or "resonance." Bakker, Le Blanc, and Schaufeli (1998), in a
comprehensive empirical study of nurses in intensive care units, demonstrated that independently of the nurses' workload, burnout levels were
higher in some units compared to others, thus supporting the contagion
hypothesis. One explanation for this phenomenon is that individuals
automatically mimic and synchronize facial and vocal expressions as well
as movements of other persons and, as a consequence, they converge
emotions (Hatfield, Caciopo, & Rapson, 1984). In addition to this unconscious imitation process, Schaufeli and Enzmann (1998) suggest that the
contagion may also be conscious. The contagion of emotional reactions is
probably one of the causes for some companies to have more difficulties
than others in attracting job candidates and retaining them. It is also noteworthy that studies associating the burnout components of the Maslach
Burnout Inventory with tunover intention report correlations ranging
between .24 and .45 with the three components and show that among the
four studies relating burnout to actual turnover, at least one burnout
dimension in each study correlated significantly with turnover (Schaufeli &
Enzmann, 1998).
Studies have shown that turnover itself may also be contagious;
employees who had witnessed co-workers' departure tend to react in
a similar manner, a phenomenon that has been likened to a snowball
effect. It was shown that such contagion effects operate mainly within
subgroups in organizations, such as units, departments, or teams, and the
authors suggested that behavior modeling is one possible explanatory
mechanism of such behavior which is not necessarily based on satisfaction considerations (Krackhardt & Porter, 1986).
A recent study by Krausz, Bizman, Yaakobovitz, and Caspi (1999)
dealt with the effects of co-workers' departure on turnover intentions of
the remaining employees. The study revealed a pattern of results that contradicted the expectation. Attributions of negative outcomes such as
increased workload, to a colleague's departure, did not affect the stayers'
intention to leave. Conversely, positive evaluations such as enhanced
opportunities for learning and personal growth and improved promotion
prospects, were associated with higher intentions to leave, rather than
61
stay. One interpretation given by the authors was that colleagues' departure may have acted like a torch that indicated the "light"-the availability of job alternatives. Such alternatives were either unnoticed prior to the
co-worker's departure or were deemed too difficult, risky, or unrealizable. The same study also found that the stayers' own work satisfaction
was not associated with the intention to leave, illuminating another variety of turnover that is unrelated to satisfaction.
If a contagion effect operates on turnover, then new labor market
realities such as the heterogeneity of employment contracts within organizations may have a substantial effect on turnover. A reality where core
and contingent employees work side by side and often collaborate within
work teams may imply contagion from the contingent to the core employees concerning loyalty and continuity of affiliation. As witnesses to the
mobility of contingent workers, core workers may react in one of two
ways. Some may bless their good luck for having a stable job whereas others may envy the opportunities for flexibility and variety that are afforded
to the contingent employees who are employed through external agencies
and who as part of that arrangement experience frequent changes of work
environment.
A research implication is the need to add co-workers' intention or
actual turnover as an antecedent of an employee's turnover. One way to
do that is by including items inquiring into employees' perceptions of coworkers' turnover intention. Management may use a similar approach by
including such questions in exit interviews.
Imposed Voluntary Turnover: The Difficulty of Disentangling
Voluntary from Involuntary
In certain situations or under some conditions, the distinction
between voluntary and involuntary turnover is not clear and in some
cases it is a mixture of both.
Moving employees from one site to another is an increasing phenomenon (Munton, Forster, Altman, & Greenbury, 1993). Most of the literature on relocation has dealt with individual relocation, namely,
situations when employees, most often in managerial positions, are
requested to relocate to another site of the same organization such as
another industrial plant or bank branch. In most cases, this is an individual move. In other instances, an entire plant, or a whole unit of an organization, relocates its premises to a different geographical location, often
requiring that employees move their home if they undertake the move.
Individual and plant relocation differ in several respects of which only
those that are relevant to voluntary turnover are mentioned here. In the
62
Moshe Krausz
individual situation, undertaking relocation means that while retaining

membership in the organization at large, one has to dissociate the ties
with his or her work unit. Avoiding relocation, on the other hand, may
involve one of two consequences, depending on the company's human
resource policy. In some cases, refusal to move is acceptable and the
employee stays in the same location. In other cases, however, failure to
move implies severance of ties with the company. When an entire plant
relocates its premises, moving implies continuity of the affiliation with
the team and department, while remaining in the former geographical
area is, by definition, an act of turnover from that organization. For the
employees as individuals, this means that they are faced with a dichotomous choice: they may either move with the entire company or, in case
they are unable or unwilling to relocate, they will sever all ties with that
organization.
Adopting the Lee and Mitchell terminology (Lee & Mitchell, 1994;
Lee et al., 1996), individual or group relocation can be regarded as a shock
to the system not only for the employee but also for her or his family
members. Similarly, it can be regarded as a violation of the psychological
contract between the individual and the organization (Rousseau, 1995).
From the organization's viewpoint, a major concern in plant relocation
is the continuous and smooth operation with as little disruption and time
and money loss. The larger the proportion of employees who choose to
move with the organization, the higher are the prospects that production
will be minimally interrupted and relocation will be mostly successful.
Conversely, employees' turnover is very costly to the company both financially and psychologically (e.g., see Fox & Krausz, 1987; Kilgore & Shorrock,
1991; Noe, Steffy, & Barber, 1988; Sagie, Krausz, & Weinstein, 2001). The
main point is that the turnover decision process and the final act of employees facing relocation, is a complex mix of voluntary choice (capturing an
opportunity to get away from a dissatisfying environment) and involuntary
factors (an unfitting climate for a family member or having a sick parent in
the old area). Even job alternatives will act differently in that situation; even
if the labor market in the old community is tight, employees may nevertheless refrain from relocation if this implies abandoning a sick and old parent or if the location has some undesirable qualities.
TURNOVER WITHOUT PRIOR SEARCH FOR ALTERNATIVES
Rational models of employee turnover since March and Simon (1958)

have asserted that prior to actual turnover, employees search for alternative jobs elsewhere. Lee and Mitchell's unfolding model of turnover
63
departs from those models also by proposing that in some instances,

turnover is not preceded by search. The empirical test of that model (Lee
et aI., 1996) supported that assertion. Lee et a1. (1996) noted, however, that
this finding should be treated with caution since the sample consisted of
nurses for whom ample job alternatives are available. In contrast with this
admonishing remark, findings from a longitudinal study of nurses' intention to leave the ward, the hospital, and the nursing profession (Krausz,
Koslowsky, Shalom, & Elyakim, 1995) suggest that despite the abundance
of nursing jobs, many nurses will first seek internal turnover prior to
developing intentions to leave the hospital, thus corroborating results
of a meta-analysis conducted by Hom, Caranikas-Walker, Prussia, and
Griffeth (1992).
INITIAL INTENTION TO LEAVE
A study by Mael and Ashforth (1995:326), entitled "Loyal from day
one ... ," showed that "strong identification does not require extended
tenure, career successes, or satisfaction as an organizational member.
Rather, simple designation as part of an organization can engender almost
immediate in-group identification." The authors also found, however,
large individual differences in the propensity for identification. Those
empirical findings also corroborate Rousseau (1998) who maintains that
despite force and situations that undermine identification, individuals
still need identification as part of their extended self or for more instrumental reasons.
From two seemingly opposing perspectives, I would like to suggest
that growing numbers of employees are coming to a new job either with
preliminary low levels of identification and commitment or with a premeditated short-term continuance commitment (Meyer & Allen, 1991) and
a prespecified plan to turn over at a certain point. The length of the preplanned affiliation may be based on sheer time ("I intend to work here no
more than 3 years"), on life changes and transitions ("till I give birth to my
first child," "till my spouse gets the degree"), or on certain predetermined
achievements ("till I have mastered certain skills or acquired a specific
body of knowledge"). Such considerations are typical mainly of hi-tech
occupations or of other individuals in high-demand occupations such as
nursing. This type of behavior is probably more prevalent among women
who are secondary breadwinners in the household. In hi-tech areas, quick
moves are required if one desires to be professionally viable and updated.
This type of premeditated turnover has not been accounted for either
by the traditional models (e.g., Hulin, 1991; Mobley, 1977) or by more
64
Moshe Krausz
recent ones (Lee & Mitchell, 1994). Additionally, it is not generated by

work dissatisfaction, by alternative job offers, or by "stimulus events"
(Rosse & Miller, 1984) or "shocks to the system" (Lee & Mitchel, 1994;
Lee et al., 1996).
What is the reasoning for suggesting that some individuals start a
new job with time-bound commitment and a preplanned program for
turnover? First, employees, even career starters, are not blind to the
changing world of work and employment and are well informed about
the increasing variety as exemplified by the number of workers who are
employed on contingent rather than core contracts (Belous, 1989; Cohany,
1998). Individuals have learned that they need to become their own career
managers as opposed to reliance on organization-managed careers (Hall,
1996; Hall & Mirvis, 1995). For those who have internalized the new
norms, accepting a job in an organization is viewed as one layer in a longterm career program. In addition, the high degree of specialization in
many hi-tech companies imposes frequent job and organization changes
in order to expand and update knowledge and professional skills.
It appears that this notion has not yet received conceptual attention
or empirical inquiry. For management in organizations this may imply
that recruitment and retention are Sisyphean efforts; they never really end
and should better be taken as such. It suggests that every employee is, in
fact, a "passive job seeker" regardless of organizations' efforts to improve
various aspects of work itself and of the work environment.
WITHIN-ORGANIZATION MOBILITY: FLEXIBILITY OR
TURNOVER?
Large and versatile organizations may be an arena for withinorganization lateral mobility. Employees may move between teams, units,
departments, functions, branches, and geographical location and still be
members of the same overall organization. Such moves may indicate personal as well as organizational flexibility. Taken from the organization's
standpoint, it is a way of coping with fluctuating work loads and with
structural and functional changes by utilization of the current human
resources thereby reducing the need to embark on expensive processes of
external recruitment and selection. For the employees, self-initiated internal mobility may be a safeguard against foreseeable decline in work volume in one's current unit thus strengthening the individual's job security.
Furthermore, internal mobility provides a route for voluntary change of a
current job or unit without the need to cross organizational boundaries.
Note that change of unit within the same organization may involve a
65
radical change of the culture of one's work environment. In addition, selfsought internal change is a way of enriching one's own job or just a way
to accomplish work and social variety.
Seen from another angle, however, within-organization moves may
be regarded as turnover. While the person making the change remains in
the same organization, the unit that she or he left still needs to find
a replacement, just as in the case of external turnover. Similarly, there are
at least two parties who may regard to move as a loss: colleagues of the
person who moves and her or his supervisor.
For co-workers, in addition to the loss of a role partner, it also has the
socioemotional implications of losing a friend (Riordan & Griffeth, 1995;
Weiss, 1990). In some respect, employees who leave their unit for another
one in the same organization may be perceived as rejecting their teammates even more than one who leaves the entire organization. Attributions
of the causes for departure may be easier and more comforting for the latter than for the former. Similarly, internal moves are a setback for the head
of the unit. The effect on the supervisor may be even more harmful: Loss
of employees to another organization can be rationalized by attributing its
reasons to factors beyond the control of the supervisor or the team, such as
higher payor improved promotion chances. Conversely, loss of employees
to departments in the same organization may be seen as rejection of the
supervisor and the work team as leaders, co-workers, and friends.
Organizations are, therefore, caught in a conflict between the desire to
retain employees by offering and allowing internal mobility to revitalize
work, and the negative emotions that such moves may evoke among peers
and superiors. This may be the reason that many organizations do not
encourage internal mobility or even actively block employees' desires
to change.
Authors have differed in outlook on this phenomenon. Some authors
offered various approaches for understanding external and withinorganization turnover. Whereas Jackofsky and Peters (1983) asserted that
within- and between-organization turnover should be combined and
treated as a single criterion, Krausz, Koslowsky, et al. (1995), based on
longitudinal data on nurses' turnover intention, supported the initial
assumption that within- and between-organization turnover merit
separate treatment. The latter findings favor the recommendation made
by Dalton and Todor (1993) that one way to reduce turnover from an
organization is by offering job alternatives within its boundaries. It has
to be acknowledged that organizational policies to encourage internal
mobility are not necessarily appreciated by all employees; willingness and
capability for change differ between individuals. Personality
constructs and behavioral styles such as openness for change, neuroticism,
66
Moshe Krausz
and conscientiousness-three of the Big 5 dimensions (Barrick & Mount,

1991), sensation seeking (Zuckerman, 1994), and attachment style
(Bowlby, 1988; Hazan & Shaver, 1970; Mikulincer, 1995)-are potential
moderators of the association between organizations' mobility policies
and actual changes undertaken by employees.
Some recent writings have drawn attention to conceptual similarities
and connections between employee transfers within the same organizations, such as those associated with relocation, and employee turnover
(Dalton, 1997; Dalton & Wimbush, 1998; Eby & Dematteo, 2000). It has
been proposed that the two phenomena deserve more connected handling in theory and research, compared with the separate and disconnected treatment traditionally given to those phenomena (Dalton, 1997;
Dalton & Wimbush, 1998). Even from a cost analysis perspective, Dalton
(1997) suggested that the iterative nature of within-organization transfer
creates a costly domino effect where the vacated position needs to be
filled by another employee (and the same applies to his or her pOSition
too). Those iterative placements require training which is often associated
with financial costs to the company.
An implication for researchers is that constructs such as commitment
which playa significant role in many turnover models (Hom & Griffeth,
1995; Mobley, 1977), and the concept of turnover itself, require more
refined and differentiated treatment. It may be needed to measure commitment to various foci, as has been proposed by Reichers (1985). From
the management's perspective, encouragement and tangible mechanisms
for job changes inside the organization have to be reciprocated by the person's needs and desires as well as by her or his personality makeup. It has
to be acknowledged, however, that although the employee remains a
member of the same organization, internal mobility arouses several reactions that are typical of real turnover.
SUMMARY AND CONCLUSIONS: THE NEED FOR MULTILEVEL

OUTLOOK ON TURNOVER
This chapter points out that the March and Simon model or its later
developments are no longer the sole explanation or antecedents of
turnover. Although for many employees, dissatisfaction is a major
antecedent and many of them will not leave an organization without prior
searching (and finding) an alternative job, in many cases, job dissatisfaction
and search for an alternative are not primary antecedents of turnover. There
seem to be growing numbers of career starters as well as more experienced
67
employees who prefer to be their own career masters who may preplan the
points and stages for career transition and change.
This chapter highlights the fact that turnover is not a unitClry concept.
Its traditional definition as termination of formal relations "w',ith an organization is limited in scope; within-organization job changes are a clear
departure from that definition. Moreover, within-organization changes
highlight the need to adopt a multilevel outlook on turnover; whereas
from the organization's perspective this is definitely not turnover,
although it is perceived as such by employees (supervisors or colleagues)
or entire units within that organization. In another situation that has been
illustrated, namely, employee relocation, different work or non work
parties differ in their views about the move; some may regard it as
turnover, based on its impact on them, whereas others, mainly the individual making the move and the entire organization, will conceive it
entirely differently. It has also been shown that in light of the heterogeneity of employment contracts that prevails today, the entire essence of
turnover may need to be turned over. The growing use of contingent
workers can itself be considered as an indication of a whole change in the
meaning of turnover.
The above also highlights the need for a multilevel outlook on
turnover. In some way, today's organizations are caught in an insolvable
conflict. Whereas Hall and Mirvis (1995:335) adamantly proclaim that lIan
organization should not be in the business of career planning," it is clear
that employees do expect training and personal development and would
turn down job offers from companies that are known to provide little of
that. Thus, while companies, mainly in hi-tech, are highly concerned with
issues of recruitment and preservation of high-quality employees, provision of challenging jobs accompanied by training and personal development entails a risk that some employees will seek further levels of
professional upgrading and career promotion elsewhere.
Social and economic welfare of communities and nations may be
improved by high rates of employee voluntary turnover since this may be
an indication of adaptation to the dynamic changes required by global
competition. Conversely, high numbers of employees sticking to a job
for very long periods should be a sign of rigidity and difficulties in timely
coping with required changes. The different facets and faces of turnover
that have been presented demonstrate the need for further expansion
of multilevel models and empirical research designs. This need has already
received attention: The 1999 volume of Research in Organizational
Behavior consists entirely of papers that present theory and propositions
that make linkages between different levels of analysis (Sutton & Staw,
1995). The topic of voluntary turnover clearly deserves a similar treatment.
Moshe Krausz
68
REFERENCES
Adkins, C. L. (1995). Previous work experiences and organizational socialization.
Bakker, A. B., Le Blanc, P. M., & Schaufeli, E. W. B. (1998). Burnout contagion: Are
leU nurses at risk for infection with the burnout virus? Unpublished manuscript,
Psychological Department, Utrecht University.
Barrick, M. R, & Mount, M. K. (1991). The big five personality dimensions and job
performance: A meta-analysis. Personnel Psychology, 44,1-26.
Belous, R S. (1989). The contingent economy: The growth of the temporary,
part-time and subcontracted workforce. Washington, DC: National Planning
Associates.
Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New York: Basic Books.
Carsten, J. M., & Spector, P. E. (1987). Unemployment, job satisfaction, and
employee turnover: A meta-analytic test of the Muchinsky model. Journal of
Applied Psychology, 72, 374-38l.

Cohany, S. R (1998). Workers in alternative employment arrangements: A second
look. Monthly Labor Review, 12(11), 3-21.
Crampton, S. M., & Wagner, J. A. (1994). Percept-percept inflation in micro organizational research: An investigation of prevalence and effect. Journal of

Dalton, D. R (1997). Employee transfer and employee turnover: A theoretical and
practical disconnect? Journal of Organizational Behavior, 18, 411-413.
Dalton, D. R, & Wimbush, J. C. (1998). Employee transfer: Review and research
agenda. Research in Personnel and Human Resource Management, 16,331-358.
Dalton, D. R, & Todor, W. D. (1993). Turnover, transfer, and absenteeism: An independent perspective. Journal of Management, 19, 193-219.
Eby, L. T., & Dematteo, J. S. (2000). When the type of move matters: Employee outcomes under various relocation situations. Journal of Organizational Behavior,
21, 677-687.
Fox, S., & Krausz, M. (1987). Correlates of relocation intention and emotional
responses to an Israeli plant relocation. Journal of Occupational Behavior, 8, 325-338.
Golembiewski, R. T., Munzenrider, R E, & Stevenson, J. G. (1986). Stress in organizations: Toward a phase model of burnout. New York: Praeger.
Hall, D. T. (1996). The career is dead-Long live the career. San Francisco: Jossey-Bass.
Hall, D. T., & Mirvis, P. H. (1995). Careers as lifelong learning. In A. Howard (Ed.),
The changing nature of work. San Francisco: Jossey-Bass.
Hatfield, E., Caciopo, J. T., & Rapson, R L. (1994). Emotional contagion. London:
Cambridge University Press.
Hazan, c., & Shaver, P. S. (1990). Love and work: An alternative theoretical
prospective. Journal of Personality and Social Psychology, 59, 270-280.
Herzberg, E, Mausner, B., Peterson, RD., & Capwell, D. E (1957). Job attitudes:
Review of research and opinion. Pittsburgh: Psychological Services of Pittsburgh.
Hom, P. w., Caranikas-Walker, E, Prussia, G. E., & Griffeth, R W. (1992). A metaanalytic structural equations analysis of a model of employee turnover.
69
Hom, P. W, & Griffeth, R W (1995). Employee turnover. Cincinnati, OH: SouthWestern Publishing.
Hulin, C. L. (1991). Adaptation, persistence, and commitment in organizations. In
M. D. Dunnette & L. Houghs (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 445-506). Palo Alto, CA: Consulting Psychologists
Press.
Hulin, C. L., Roznowski, M., & Hachiya, D. (1985). Alternative opportunities and
turnover decisions: Empirical and theoretical discrepancies and an integration. Psychological Bulletin, 97, 233-250.
Huselid, M. A. (1995). The impact of human resource practices on turnover,
productivity, and corporate financial performance. Academy of Management
Journal, 38, 635-672.
Jackofsky, E. E, & Peters, L. H. (1983). Job turnover versus company turnover:
Reassessment of the March and Simon participation hypothesis. Journal of
Kilgore, J. E., & Shorrock, M. B. (1991). International relocation: The future of
spouse relocation assistance. Journal of Career Development, 17, 271-284.
Krackhardt, D., & Porter, L. W. (1986). The snowball effect: Turnover embedded in
communication networks. Journal of Applied Psychology, 71, 50-55.
Krausz, M., Bizman, A., Yaakobovitz, N., & Caspi, T. (1999). Stayers' intention to
withdraw following turnover of a coworker. Journal of Business and Psychology,
14,96-107.
Krausz, M., Brandwein, T., & Fox, S. (1995). Work attitudes and emotional responses
of permanent, voluntary, and involuntary temporary-help employees: An
exploratory study. Applied Psychology: An International Review, 44, 217-232.
Krausz, M., Koslowsky, M., & Eiser, A. (1998). Distal and proximal influences on
turnover intentions and satisfaction: Support for a turnover progression theory. Journal of Vocational Behavior, 52, 59-71.
Krausz, M., Koslowsky, M., Shalom, N., & Elyakim, N. (1995). Predictors of the
intention to leave the ward, the hospital, and the nursing profession: A longitudinal study. Journal of Organizational Behavior, 16, 277-288.
Landy, E J. (1989). Psychology of work behavior. Pacific Grove, CA: Brooks/Cole.
Lee, T. W, & Mitchell, T. R (1994). An alternative approach: The unfolding model
of voluntary employee turnover. Academy of Management Review, 19, 51-89.
Lee, T. W., Mitchell, T. R, Wise, L., & Fireman, S. (1996). An unfolding model of
voluntary employee turnover. Academy of Management Journal, 39, 5-36.
McLean Parks, J., Kidder, D. L., & Gallagher, D. G. (1998). Fitting square pegs into
round holes: Mapping the domain of contingent work arrangements onto the
psychological contract. Journal of Organizational Behavior, 19, 697-730.
Mael, E A., & Ashforth, B. E. (1995). Loyal from day one: Biodata, organizational
identification, and turnover among newcomers. Personnel Psychology, 48,
309-333.
Mathieu, J. E., & Zajac, D. M. (1990). A review and meta-analysis of the
antecedents, correlates, and consequences of organizational commitment.
70
Moshe Krausz
Meyer, J. P., & Allen, N. J. (1991). A three component conceptualization of organizational commitment. Human Resource Management Review, I, 61-89.
Mikulincer, M. (1995). Attachment style and the mental representation of the self.
Journal of Personality and Social Psychology, 69, 1203-1215.
Mikulincer, M. (1997). Adult attachment style and information processing:
Individual differences in curiosity and cognitive closure. Journal of Personality
and Social Psychology, 72, 1217-1230.
Mobley, W. H. (1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology, 62, 237-240.
Munton, A. G., Forster, N., Altman, Y., & Greenbury, L. (1993). Job relocation, managing people on the move. New York: Wiley.
Noe, R A., Steffy, B. D., & Barber, A. E. (1988). An investigation of the factors
influencing employees' willingness to accept mobility opportunities.
Personnel Psychology, 41, 559-580.
Peters, L. H., Jackofsky, E. E, & Salter, J. R (1981). Predicting turnover: A comparison
of part-time and full-time employees. Journal of Occupational Behavior, 2, 89-98.
Reichers, A. E. (1985). A review and conceptualization of organizational commitment. Academy of Management Review, 10, 465--476.
Riordan, C. M., & Griffeth, R W. (1995). The opportunity for friendship in the
workplace: An underexplored construct. Journal of Business and Psychology, 10,
141-155.
Robinson, S. L. (1996). Trust and breach of the psychological contract.
Administrative Science Quarterly, 41, 574-599.
Rosse, J. G., & Miller, H. E. (1984). Relationships between absenteeism and
other employee behaviors. In P. Goodman & R Atkin (Eds.), Absenteeism
(pp. 194-228). San Francisco: Jossey-Bass.
Rousseau, D. M. (1995). Psychological contracts in organizations. Beverly Hills: Sage.
Rousseau, D. M. (1998). Why workers still identify with organizations. Journal of
Organizational Behavior, 19, 217-233.
Sagie, A., Krausz, M., & Weinstein, Y. (2001). To move or not to move: Factors
affecting employees' actual relocation when an entire plant moves. Journal of
Occupational and Organizational Psychology, 74,343-358.
Schaufeli, w., & Enzmann, D. (1998). The burnout companion to study and practice.
London: Taylor & Francis.
Sutton, R J., & Staw, B. M. (1995). What theory is not. Administrative Science
Quarterly, 40, 371-384.
Wegmann, RG. (1991). From job to job. Journal of Employment Counseling,2S, 8--12.
Weiss, R S. (1990). Losses associated with mobility. In S. Fisher & c. L. Cooper
(Eds.), On the move: The psychology of change and transition. New York: Wiley.
Wolpin, J., Burke, R. J., Krausz, M., & Freibach, N. (1988). Lateness and absenteeism: An examination of the progression hypothesis. Canadian Journal of
Administrative Sciences, 5, 49-54.
Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation seeking.
London: Cambridge University Press.
4
Employee Withdrawal Behavior
Role of the Performance Management Process

INTRODUCTION
In today's turbulent environment brought about by globalization and

continuing technological change, innovation is a critical ingredient for a
successful organization. Innovation means more than new technology
and products. It means that employees use their knowledge, abilities, and
experience to respond creatively to the constantly changing demands of
the environment. It includes "ideas for reorganizing, cutting costs, putting in new budgeting systems, improving communication, or assembling
products in teams" (Kanter, 1983:20). This view of innovation recognizes
that all employees have the potential to be innovative.
However, the phenomenon of employee withdrawal behavior is a
grave impediment to innovation because it deprives the organization of
the benefits when employees experience a strong commitment to the job
and the organization. A primary and necessary condition for the realization of the employees' potential for innovation is the existence of an
almost indissoluble bond between employees and their organization.
Such a bond predisposes employees to strive for successful job performance, and the set of extra-job-role behaviors that are indispensable for
innovative endeavors that give a competitive edge to the organization.
The gradual erosion of the competitiveness of an organization and
the consequent loss of its markets suggests that the necessary level of
RABINDRA N. KANUNGO AND MANUEL MENDONCA Faculty of Management,
McGill University, Montreal, Quebec H3A IG5, Canada
71
72
innovation is not taking place in that organization. The generally higher

education level of today's employees would likely rule out competence as
the probable cause of organizational ineffectiveness. Therefore, the more
likely inference is that the prevailing organizational conditions not only
do not foster in employees the desire and willingness to innovate but, in
fact, contribute to their alienation in relation to the job and the organization. In other words, the bond between the employees and the organization is seriously ruptured or weakened. As a result, the organization,
despite its core competencies, loses its vitality because the employees,
who exercise these competencies, have withdrawn their support and
commitment, or are callously indifferent to the organization's interests
and objectives. The withdrawal behaviors create a physical and psychological distance between employees and the work environment that is
most detrimental to innovation and organizational effectiveness.
Conditions with potential to
increase
employee withdrawal behaviors
Organizational Factors
Supervisory Style
Reward System
Job Design
Employee Powerlessness
Reduced Employee's
-Job involvement
-Organizationa
commitmentl
Strategy to
reduce
employee witbdrawal behaviors
Perfonnance Management
Process
Employee Empowennent
Increased Employee's
-Job involvement
-Organizational
commitment
,
Increased Employee
Withdrawal Behaviors
-Membership
-Performance
-Extra-role
Reduced Employee
Withdrawal Behaviors
-Membership
-Performance
-Extra-role
Figure 1. An overview of the chapter: conditions with the potential to increase, and strategy
to reduce employee withdrawal behaviors.
73
What are the major types of employee withdrawal behaviors? What

are the factors and the underlying processes that explain employee withdrawal behavior? What can organizations do to create conditions that prevent employee withdrawal behavior? This chapter addresses these
questions-more specifically, in the following terms, as outlined in Fig. 1:
1. The major categories of employee withdrawal behavior
2. Employee withdrawal behaviors as a manifestation of a lack of job
involvement and organizational commitment that is the result of
work alienation or a sense of powerlessness
3. The notion of empowerment in management practice
4. Performance Management Process as a powerful strategy to create
organizational and work conditions that empower employees,
increase their job involvement and organizational commitment,
and, as a result, reduce employee withdrawal behavior
THE MAJOR CATEGORIES OF EMPLOYEE WITHDRAWAL

BEHAVIORS
Employee withdrawal behaviors can be described as the ways in

which an employee who is dissatisfied with a work situation, responds to
it (Hulin, 1991). Clearly, these responses do not positively contribute to
the performance of the job tasks and the set of voluntary, extra-job-role
behaviors that are critical to organizational effectiveness.
Following an exhaustive review of the literature of organizational
adaptation/withdrawal models, Hulin (1991:482) proposed a Modified
Organizational Adaptation/Withdrawal model that suggests a set of
behavioral intentions with corresponding behaviors that are the employees' responses to job or work role dissatisfaction, as indicated in Table 1.
Table 1. Modified Organizational AdaptationlWithdrawal Model
Behavioral intentions
Adaptation/withdrawal behaviors
To increase job outcomes
Stealing; using work time for personal tasks;

moonlighting on the job
Missing meetings; long coffee breaks; drinking/using
drugs before work; wandering around looking busy;
talking with co-workers about trivia
Tardiness; absenteeism; quitting; retirement
Unionization activity; transfer attempts; shift changes;
demotion attempts
To reduce job inputs

To reduce work role inclusion
To change work role
Note. From Hulin (1991:482).
74
As existing theories of motivation would suggest, the withdrawal behaviors seek to either increase job outcomes or decrease work output in order
to equitably compensate for the perceived work inputs (Adams, 1965;
Kanungo & Mendonca, 1997).
If we focus exclusively on employee withdrawal behaviors, including those suggested by Hulin (1991), then we can group them into three
broad categories:
1. Avoiding or reducing membership behaviors such as: tardiness

(Rosse, 1988), absenteeism (Johns, 1987; Nicholson & Payne, 1987;
Reichers, 1985; Somers, 1995; Steers & Rhodes, 1978, 1984), and
turnover (March & Simon, 1958; Mobley, 1977; Mowday, Porter, &
Steers, 1982; Somers, 1995).
2. Avoiding or reducing performance behaviors, as specified by the job
role, such as: work-to-rule, or work restriction (Roethlisberger &
Dickson, 1939; Whyte, 1955), social loafing (Latane, Williams, &
Harkins, 1979), grievances (Stagner, 1956), strikes (James, 1951).
3. Avoiding extra role behaviors. These behaviors are not part of the
employee's job content (Katz & Kahn, 1978), but are performed voluntarily, and found to contribute to organizational effectiveness.
Examples of these behaviors are: helping co-workers, protecting
the organization, making constructive suggestions, developing
oneself, spreading goodwill (Bishop & Scott, 1997; George & Brief,
1992).
WITHDRAWAL BEHAVIOR-A MANIFESTATION OF EMPLOYEE

POWERLESSNESS, JOB ALIENATION, AND WEAK
ORGANIZATIONAL COMMITMENT
In order to examine the evidence that employee withdrawal
behavior is indeed a manifestation of employee powerlessness, job alienation, and weak organizational commitment, we explore the concepts of
job involvement and alienation, and their effects in the workplace-in
relation to the withdrawal behaviors of membership, performance, and
extrarole behaviors.
Job Involvement, Job Alienation, and Withdrawal Behavior
Employees are involved in the job when they perceive that it has the
potential to satisfy their salient psychological needs (Kanungo, 1979,
1982). The work environment satisfies the employees' psychological
75
needs when employees: perceive work to be meaningful; can control the

work methods and schedule; find their superiors and co-workers are supportive and consistent with regard to the work norms and behaviors;
receive feedback on task accomplishments and equitable rewards; see
work as an opportunity for personal growth and development (Brown,
1996). It enables employees to satisfy their self-realization and ego-esteem
needs, to develop their abilities to the fullest potential, and to experience
dignity and self-respect. Such a work environment enhances the employees' self-efficacy and ability to cope with the job demands-as related to
the job tasks, unexpected events and situations, and the people they
would confront. As a result, employees experience feelings of being
empowered that heighten the motivation for task accomplishment and to
invest the necessary psychological and physical resources in job performance, attendance, and retention behaviors (Brown, 1996).
On the other hand, when work is highly prescribed and contractual,
with no autonomy, except for rudimentary and insignificant decisions,
workers are not involved, but alienated. According to Karl Marx (1932),
workers experience a state of alienation or a sense of powerlessness and
self-estrangement because they do not own and control the product and
the means of production; such control being the essential means by which
workers develop their "universal nature". Since Marx, many sociologists
have suggested that worker alienation results from workplace conditions
that neither allow workers to exercise freedom and control nor provide
them with the opportunity to utilize their potential (Blauner, 1964;
Shepard,1971).
A parallel conceptualization of the powerlessness variant of alienation also exists among psychologists. As Kanungo (1982:25) points out,
"if one analyzes the sociological concept of powerlessness in motivational
terms, it becomes obvious that if a situation constantly frustrates an individual's need for autonomy and control, it will create a state of alienation." Sashkin (1984) also considers the lack of autonomy and control in
the workplace to be the main reason for the feelings of powerlessness
among workers. For example, research studies suggest that the withdrawal behavior of absenteeism is largely due to job dissatisfaction arising from the job content (Hackett & Guion, 1985; Scott & Taylor, 1985), and
also from supervisory arbitrariness and lack of clear guidelines and
norms (Nicholson & Johns, 1985).
Thus, when the job does not permit employees to experience work as
meaningful, to be responsible for the work outcomes, and to know the
actual results of the work activities, when the work environment is characterized by ambiguous and conflicting work norms and behaviors, they
experience feelings of powerlessness. Such powerlessness creates a state
76
of job alienation and reduces the employees' self-efficacy belief, decreasing their motivation to invest in the psychological and physical resources
needed to accomplish the job tasks. As a consequence, they have been
found to engage in withdrawal behaviors such as tardiness, absenteeism,
turnover, and decrease in job effort and performance (Brown, 1996).
Organizational Commitment and Withdrawal Behavior
Employee empowerment also fosters strong organizational commitment because the empowered employees now perceive greater competence and control over the organizational resources and environment
(Menon, 1999). What is organizational commitment? Organizational commitment has been defined in different ways. In essence, it is a psychological state that characterizes the employee's bond or relationship with the
organization. Meyer and Allen (1997:11) describe organizational commitment in terms of three components: affective commitment, continuance
commitment, and normative commitment. "Affective commitment refers
to the employee's emotional attachment to, identification with, and
involvement in the organization .... Continuance commitment refers to
an awareness of the costs associated with leaving the organization ....
Normative commitment reflects a feeling of obligation to continue
employment."
Employees high on all three types of commitment have low turnover.
Employees high on affective commitment have better attendance records,
work harder, and perform better. Employees high on affective commitment and normative commitment are more inclined to engage in extrarole
behaviors that contribute to organizational effectiveness-such as helping
co-workers and protecting the organization (Bishop & Scott, 1997;
Brown, 1996; Mathieu & Zajac, 1990; Meyer & Allen, 1997; Somers, 1995).
Compared to employees with a low organizational commitment, employees with a high organizational commitment have internalized the organization's vision that creates a stronger bond with the organization and,
therefore, are more likely to engage in behaviors that contribute to its
effectiveness (Dessler, 1999).
On the other hand, the feeling of powerlessness weakens the employees' organizational commitment because they do not experience the psychological state of bonding with the organization and, therefore, engage in
withdrawal behavior (Dessler, 1999). Employees with weak affective, continuance, and normative commitment have high turnover. Employees
with weak affective commitment demonstrate a higher rate of voluntary
absenteeism, low effort, and poor performance. Employees with weak
affective and normative commitments avoid extrarole behaviors
77
(Bishop & Scott, 1997; Brown, 1996; Mathieu & Zajac, 1990; Meyer & Allen,
1997; Somers, 1995).
From the discussion in this section, it is clear that employee feeling of
powerlessness is the root of the malaise of job alienation and weak
organizational commitment, with the resulting withdrawal behaviors.
The next section examines the notion and practice of empowerment
in order to lay a foundation for the development of an appropriate
strategy to empower employees, and implement a commitment-oriented
management practice that seeks to effectively address the withdrawal
behavior.
THE NOTION OF EMPOWERMENT IN MANAGEMENT
PRACTICE
In order to overcome the adverse effects of powerlessness in workers,
organizational researchers have suggested "empowerment" as the key
dealienation strategy (Block, 1987; Kanter, 1979). "The buzzword of the
1990s is 'empowerment' and it's one of the hottest topics in management"
(Stern, 1991:B4). What is empowerment? What is the specific process by
which it overcomes the debilitating effect of worker alienation?
Conger and Kanungo (1988) proposed that empowerment can be
viewed in two different ways: as a relational and as a motivational construct (see Fig. 2). Most theorists in industrial sociology and organizational behavior have tended to view empowerment as a relational
construct. Derived from the literature of social exchange theory, the relational dynamics revolves around the relative power status of the parties
in the exchange situation that might be one of dependence and/ or interdependence (Blau, 1964; Emerson, 1962; Homans, 1974; Thibaut & Kelley,
1959). In the organizational context, power is interpreted to mean the possession of formal authority or control over organizational resources.
Social exchange theories suggest that subordinates are more likely to
experience alienation because of their dependence on their managers who
have the power to determine whether they will achieve their desired outcomes. By sharing power, managers seek to overcome the alienation
brought about by subordination and control (Schacht, 1970). Viewed in
terms of the relational dynamics, empowerment has come to mean the
process (flow # 1, Fig. 2) of delegation or decentralization of decisionmaking power in which management gives away or shares power with
workers (Burke, 1986; Kanter, 1983). This connotation of empowerment is
also reflected in the Merriam Webster's Dictionary meaning of empower as
"to authorize or delegate or give legal power to someone."
78
;--
Organizational Conditions:
Delegation/Power Sharing
(Stems out of Relational Dynamics in
Social Exchange Theories)
De-alienation Through Empowerment
A Belief State of Self-Efficacy
*2
Individual's Motivational State:

EnablinglEnhancing Self-Efficacy
(Stems Out of Motivational Dynamics
in Expectancy Theories)
I-
I) Relational approach to the determinants of empowerment

2) Motivational approach to the determinants of empowerment
3) Organizational conditions that enhance self-efficacy belief
Figure 2. Determinants of empowerment.
By treating empowerment as a relational construct, the practical

managerial consequence has been to empower employees through a
plethora of participative management techniques. Some of these are subordinate involvement in goal-setting as is the case in the widely used
management by objectives (MBO), job enrichment, and quality control
circles (Herzberg, 1966; Likert, 1961, 1967; McGregor, 1960). The techniques have had mixed results (Fein, 1974; Luthans & Reif, 1972, 1973),
and the relational-based approach to empowerment is inadequate in
explaining the success or failure of the participative management techniques in empowering employees. The wisdom of this approach can also
be questioned for other reasons. Its exclusive focus on participative management techniques does not encourage managers to develop a greater
repertoire of equally, if not more, effective empowerment strategies.
The more fundamental questions, however, relate to its failure to
explore the nature and the psychological mechanisms underlying the
empowering experience. Does the sharing of power and resources with
subordinates automatically empower them? Are the effects of an empowering experience the same as the effects of delegation, participation in
79
decision-making, and resource sharing? As discussed later, these techniques might well be a means of empowerment only to the extent that
they lead to enhancing the individual's self-efficacy belief (flow # 3,
Fig. 2). But, the mere act of sharing or delegating power, in the relationalbased approach, is inadequate to fully understand the nature and experience of empowerment because it ignores the basic motivational processes
involved in it. As will be seen in the ensuing discussion, these processes
are better and more fully captured when empowerment is interpreted
as a motivational construct.
The social psychological literature provides the basis for interpreting
empowerment as a motivational construct. Here empowerment is used as
a motivational and/or expectancy belief state internal to individuals. For
instance, individuals are assumed to have a need for power (McClelland,
1975) where power connotes an internal urge to influence and control
other people. A related but more inclusive disposition to control and
cope with life events has also been proposed by several psychologists
while dealing with the issues of primary/secondary control (Rothbaum,
Weisz, & Snyder, 1982), internal/external focus of control (Rotter, 1966),
and learned helplessness (Abramson, Garber, & Seligman, 1980).
Individuals feel empowered when they perceive that they can adequately
cope with events, situations, and/or the people they confront. On the
other hand, individuals feel powerless when they believe that they are
unable to cope with the physical and social demands of the environment.
Empowerment in this motivational sense also refers to a belief in
self-determination (Deci, 1975) or a belief in personal self-efficacy
(Bandura, 1986, 1997) as the individual copes with environmental
demands. Any organizational strategy or technique that strengthens this
self-determination or self-efficacy belief of workers will tend to make
them feel empowered at work (flow # 2, Fig. 2) and, consequently,
dealienated. Conversely, any strategy that weakens their self-determination or efficacy belief will increase their sense of personal powerlessness
or alienation in relation to the work context. The Oxford English dictionary meaning of empower as "to enable" also reflects the motivational sense
of empowerment. In contrast to the earlier definition of empowerment as
"delegation" (of authority and resource sharing), the connotation of
"enabling" implies motivation through the enhancement of one's personal efficacy and ability to cope with environmental demands.
The motivation-based approach to empowerment requires two sets
of actions:
1. To identify organizational conditions or contextual factors that
cause employees to feel powerless and self-estranged-resulting
in withdrawal behavior
80
2. To modify such conditions and provide employees with information that enhances personal efficacy, and increase job involvement
and organizational commitment in order to reduce the withdrawal
behaviors
We now deal with the first set of actions. The next section discusses
the Performance Management Process-a potent strategy to modify the
contextual conditions in order to empower employees and thereby reduce
withdrawal behavior.
Contextual Factors Lead to Powerlessness
Conger and Kanungo (1988:477) have identified four categories
of contextual factors that contribute to the lowering of employees' selfefficacy beliefs (See Table 2).
Table 2. Context Factors Leading to a Feeling of Powerlessness that
Contributes to Employee Withdrawal Behavior
Organizational factors
Significant organizational change/transitions
Start-up ventures
Competitive pressures
Impersonal bureaucratic climate
Poor communications/network-forming systems
Highly centralized organizational resources
Supervisory style
Authoritarian (high control)
Negativism (emphasis on failures)
Lack of reason for actions/ consequences
Reward system
Noncontingency (arbitrary reward allocations)
Low incentive value of rewards
Lack of competence-based rewards
Lack of innovation-based rewards
Job design
Lack of role clarity
Lack of training and technical support
Unrealistic goals
Lack of appropriate authority/discretion
Low task variety
Limited participation in programs, meetings, decisions that have a direct impact on job
performance
Lack of appropriate/necessary resources
Lack of network-forming opportunities
Highly established work routines
81
Table 2. (Continued)
Job design
High rule structure
Low advancement opportunities
Lack of meaningful goals/tasks
Limited contact with senior management
Note. From Conger & Kanungo (1988).
We draw on a few factors from each category to illustrate how each has
the potential to lower the employees' self-efficacy beliefs, increase their
feeling of powerlessness, and reduce their job involvement and organizational commitment that contribute to the withdrawal behaviors.
Organizational Factors. Events such as an acquisition, a merger,

significant technological changes, rapid growth, or serious financial emergencies induce major organizational change that alter organizational
structures, norms, communication and action patterns, and create considerable uncertainty and disorientation (Nadler, 1980; Tichy & Devanna,
1986). The upheavals, inherent in such change, can range from job responsibilities being added or diminished to downsizing. The added responsibilities can be perceived as a threat because these might require skill levels
beyond the employee's competencies. When responsibilities are diminished there is the fear that the jobholder might now be subordinated to
others in the structure. Inherent in such changes is the possibility of role
ambiguity and conflict that creates demands that the employees do not
believe they can handle. Finally, downsizing is a direct threat to one's job
security with largely negative effects on motivation. Either of these
changes poses a serious challenge to the employees' sense of control and
competence, lowering self-efficacy and reducing job involvement and
organizational commitment.
In these situations, employee withdrawal behaviors-avoiding or
redUcing membership, performance, and extrarole-are seen as the only
possible options. Thus, absenteeism provides relief as they withdraw
from the tension-ridden work situation; it may also be an opportunity to
search for another job situation, which results in turnover. Reducing job
performance-in particular, avoiding extrarole behavior such as making
constructive suggestions-might be seen as an attempt to restore equity
in the context of the injustice that is perceived. These situations do not
encourage employees to spread goodwill about the organization. Instead,
openly talking about the perceived injustice might be seen as a way of
restoring equity, especially if the organization could have anticipated the
82
changes and their consequences, but failed to be proactive in assisting the

employees to cope with them. For example, the organization did not provide employees with opportunities to acquire the competencies needed
for the technological changes (Oessler, 1999). In the case where downsizing is the effect of the organizational change, employees see that there is
no "equality of sacrifice," where the management ranks are untouched by
the layoffs (Baron & Kreps, 1999).
Supervisory Style. Bureaucratic organizations and directing members by rigid rules and procedures tend to create inequities in the distribution of organizational power. Both of these characteristics deprive
employees of control and discretion and, thereby, lower the employees'
self-efficacy belief. This is largely because such a style does not favor,
much less foster, open communications.
In such a closed and oppressive climate, it is inevitable that employees
engage in withdrawal behavior such as tardiness, social loafing, and
increased grievances, and avoid the whole set of extrarole behaviors.
Through tardiness and social loafing-provided no adverse consequence,
such as a pay deduction, is attached to it-employees exercise their hidden
power to counterbalance that of the supervisor. The grievances, in particular the baseless ones, are an overt demonstration of employee power, especially if the grievance system is sanctioned by the collective agreement with
the union. One might also view the grievances as a vehicle through which
the employees are demanding a two-way communication in the organization. By avoiding the extrarole behaviors, employees communicate that
they too can play by those same rigid rules; in effect, they are making
a statement: "the extrarole behaviors are not part of my job description."
Reward System. Employees' sense of powerlessness increases
when they do not receive valued rewards; and when the rewards are not
based on employee competence and persistence in innovative behavior
(Sims, 1977; Szilagyi, 1980).
Rewards provide employees with a message on the value of their job,
and their performance. In reality, however, it is a message of what the
organization believes its employees are worth. Rewards are one aspect of
the workplace that directly affects the individual's self-worth. When
rewards do not reflect the job value and their performance, employees
experience serious inequity that leads to pay dissatisfaction. To restore
equity, employees engage in withdrawal behaviors of absenteeism,
reduced performance, and turnover (Brown, 1996; Kanungo & Mendonca,
1997). Since rewards perceived to be unfair are a serious reflection of one's
self-worth, employees' commitment to the organization will also be
reduced which will cause employees to avoid the extra role behaviors.
83
"You get what you reinforce"; therefore, if the performance and extrarole
behaviors are not reinforced, then employees will refrain from engaging
in these behaviors (Luthans & Stajkovic, 1999).
Job Design. When the job involves role ambiguity and conflict, the
employees' self-efficacy beliefs are greatly diminished and their job commitment is considerably weakened.
The direct impact of role ambiguity and conflict is withdrawal from
performance behavior. The task will not be performed not because the
employee does not wish to, but because the employee does not know
what it is that has to be done. The job has two components: what is to be
done and how it is to be done. It is the manager's responsibility to communicate clearly, in unambiguous terms, what is to be done because managers are expected to know how the performance of the job tasks fit or
relate to the successful attainment of the work unit's objectives.
Employees have been hired after ensuring that they have the job relevant
knowledge, skills, and abilities. Therefore, they do not expect to be told
how the job is to be done. Furthermore, employees will also withdraw
from task performance when the job involves role conflict. The reason for
the low self-efficacy and the resulting withdrawal is not because employees do not know how the job is to be done, but because the role conflict
creates uncertainties due to issues like ambiguous overlapping jurisdictions with regard to job tasks, responsibilities, and resources.
The discussion thus far has demonstrated that the phenomenon of
employee withdrawal behavior has its origin in the organizational conditions that lower employees' personal efficacy belief and create a feeling of
powerlessness which, in turn, cause employees to reduce job involvement
and commitment to the organization. The discussion has also examined
the conceptual framework of the empowering experience that provides a
foundation for commitment-oriented management practices. Such practices "commit to people-first values ... clarify the mission and ideology ...
guarantee organizational justice ... create a sense of community ... support
employee development" (Dessler, 1999:66). The Performance Management
Process, described in the next section, incorporates the commitmentoriented management practices needed to reduce employee withdrawal
behaviors.
THE PERFORMANCE MANAGEMENT PROCESS-AN ANTIDOTE
TO EMPLOYEE WITHDRAWAL BEHAVIOR
As can be seen from Fig. 3, the Performance Management Process
outlines a set of essential preconditions (Lawler, Mohrman, & Resnick,
84
1984) that constitute the foundation and rationale to ensure the effective
implementation of the procedural steps that follow. As we describe the
process, we will discuss how the essential conditions and the procedural
steps enhance the employees' self-efficacy belief in the workplace. The
empowerment experienced by employees increases their job involvement
and organizational commitment that prevent withdrawal behavior or
alleviate its adverse effects.
ESSENTIAL PRECONDITIONS
Organizational Work Culture
Employees are a vital resource
Management Commitment
~Communication
managers for employee

development
~Rewarding
PROCESS
Defining the Job
Job objectives
~
Organizational Objectives
-.
'--
Setting performance Standards and Rewards

Difficult and specific goals
Employee participation
Removal of environmental constraints
Monitoring Performance
Ongoing feedback
Coaching and mentoring
Formal Performance Review

Perfomlance assessment
Action plan for next period
~Goals
~Training and development

--Removal of environmental constraints
Figure 3. The performance management process.
85
Essential Preconditions
The Performance Management Process is founded on a fundamental

value that employees are a vital resource of the organization. To ensure that
it is not just wishful thinking, management commitment to it is ensured,
not only by its communication, but also by the policy of rewarding managers for employee development. The articulation and acceptance of this
value throughout the organization means that it will permeate the organization's policies and practices.
The practical consequence of a commitment to people-first values is
the creation of a work culture in which employees are encouraged and
supported to grow and develop. In such a culture, employees are no
longer mere disposable resources, but are persons who can and are
expected to fully utilize their talents and abilities to the attainment of job
objectives and contribute to organizational effectiveness. Such a culture
fosters in employees the willingness to take risks and to innovate (Dessler,
1999; Meyer & Allen, 1997; Sommer, Bae, & Luthans, 1996).
The empowering effects of this work culture are easy to visualize. For
example, in times of major organizational changes such as an acquisition
or a merger, employees will receive timely information on the reasons and
consequences of such changes. They will be reassured that every effort is
being made to assist them in coping with the uncertainty and disorientation existing during the transition stage of such major changes. As a
result, major organizational changes will not pose a serious threat to the
employees' sense of control and competence.
Procedural Steps
Step 1. The Performance Management Process is not a one-shot,

once-a-year performance appraisal ritual which managers plough
through most reluctantly. It is essentially the management of employee
performance as an ongoing, cyclical process. It begins with a clear definition of the subordinate's job. In this first step, the manager must identify
all of the important aspects of the job and clarify how it is related to the
objectives of the organization.
One way of identifying job objectives that are concrete and appropriate to organizational objectives is to view the recipient of the job's output
as the "customer" of that job. In this approach, the managers assist their
employees in identifying the "cutomers" of their output. For example,
the sales analyst who prepares the monthly sales statistics which are
routed to the Sales Manager, the Marketing Manager, and Senior
Managers, sees each recipient as a "customer," and becomes conscious of
86
the job's responsibility to satisfy the specific needs of the customers. The
customer approach enables the employees to experience the significance
of their tasks and promotes, in varying degrees, identification with the
organization's objectives.
This step addresses the factors that lower the employee's self-efficacy
belief-factors such as role ambiguity, role conflict, and lack of meaningful job goals. It gives managers the opportunity to communicate the organization's vision-including its values, mission, and goals-and to clearly
explain how the employee's job tasks fit and contribute to the attainment
of the vision. When employees understand the job's relationship to the
organization's vision and its underlying higher purpose, they internalize
the vision and identify with it because they share the beliefs and values
represented by the vision, and it also enables them to feel a sense of community (Kanter, 1972). Such internalization transforms the employees'
values and beliefs and heightens their sense that their job is a particularly
significant endeavor (Collins & Porras, 1997; Conger & Kanungo, 1998).
The resulting internalization and identification with the organization
greatly enhance the employees' affective and normative commitment,
which increase their willingness to undertake the extrarole behaviors
(Meyer & Allen, 1997). Also facilitated are the employees' ability to understand and their willingness to cope with organizational changes even at
some personal inconvenience or cost (Kanungo & Mendonca, 1996).
Fundamental to this and the other steps as well, is the manager's participative approach that involves the employee in the process of job definition that leads to an agreed understanding of and commitment to the
appropriate job behaviors.
Step 2. The second step of the cycle is the setting of performance standards and the corresponding rewards. The first step spelled out what needs to
be done. The second step specifies how well the tasks must be done. It sets
the standards by which the job performance will be assessed. Goal-setting
theory (Locke & Bryan, 1968), abundantly supported by empirical evidence (Locke, Shaw, Saari, & Latham, 1981), stipulates that employee
performance is greatly enhanced when the assigned goals are difficult,
but attainable; and specific, but appropriate to the objectives of the organization. This step also establishes the performance period, and ensures
that the rewards are contingent on performance and are valued by the
employee (Kanungo & Mendonca, 1988).
The process of performance assessment, despite the most sophisticated attempts to introduce objective procedures, will always remain a
subjective process and, as such, highly vulnerable to fallible judgments
even with all the goodwill in the world. Therefore, the participation of
87
employees in the establishment of standards and measures of performance is also crucial. The knowledge and expertise of both the supervisor
and the employees contribute to performance standards and measures
that are reasonable, realistic, and appropriate. This is not to suggest that
subordinates would set unrealistically low standards. In fact, there is considerable evidence that subordinates often tend to set unrealistically high
goals (Lawler, 1977). The joint mutual-influence process would, in addition to injecting realism, ensure manager-subordinate agreement and
give the subordinate "ownership and control" which go a long way to
generating the trust, acceptance, and commitment so necessary to this
highly subjective system (Locke & Latham, 1984).
In addition to establishing the job objectives and measures of performance, the manager should also identify the needed resources for task
attainment and ensure that these are available to the employee. Managers
should also explore the possible environmental constraints that might
interfere with the employee's job performance, and work to removing
them. If all of the required resources are not available, or the constraints
cannot be removed, then the performance standards would need to be
appropriately adjusted to reflect the fact that job performance at such levels is beyond the control of the employee.
The activities in this step ensure role clarity; realistic and meaningful
goals; performance-contingent and valued rewards; participation in decision making; communications; allocation of the needed resources; and
removal of environmental constraints. As discussed previously, these
activities are most conducive to enhancing the employees' self-efficacy
belief.
Step 3. The third step of the performance management cycle is monitoring the performance. During this phase, the manager provides informal
ongoing feedback, which is to be viewed not in terms of faultfinding but
rather as on-the-job coaching. When managers function as coach, they
seek to help employees grow and reach desired performance levels.
Coaching involves knowing how well the employees are performing relative to performance standards in terms of specific, measurable behaviors, and discussing areas for improvement. The coach gives praise for
work well done and offers constructive criticism when appropriate. In the
latter case, the manager cites specific behaviors with specific suggestions
on how to correct the performance problems.
Performance standards can never be etched in stone; their validity
always assumes a relatively stable environment. The manager, who functions like a coach, will be sensitive to changing environmental factors, and
make suitable adjustments to the performance standards. Perhaps the
88
shortfall in performance may be due to the employee being deficient in

certain skills. The coaching approach will cause the manager to be immediately aware of such deficiencies and provide remedial measures, such
as training. In their coaching stance, managers are careful to create an
open, relaxed atmosphere, which encourage employees to seek guidance
in sorting out priorities or in resolving problems.
As mentioned in the previous section, the employees' self-efficacy
belief is lowered by the supervisory style characterized by high control
(authoritarian), emphasis on failures (negativism), and lack of reason for
actions and consequences. The supervisory style adopted in the activities
in this step provides employees with an empowering experience, and
greatly enhances their self-efficacy belief.
Step 4. The fourth step is the formal appraisal review at the end of the
predetermined performance period, during which managers record their
assessment of the individual's performance. This phase usually poses the
greatest problem for most managers as it demands that they play two
apparently conflicting roles-as coach and judge. The most frequently
recommended approach to this phase is termed the "problem-solving"
approach because its focus is on the removal of obstacles to good performance-obstacles such as inappropriate or obsolete work procedures,
lack of resources and certain skills, lack of a clear understanding of the job
role and requirements. This approach also encourages a joint discussion
between the manager and the subordinate.
The mutual exchange of information provides a much clearer picture of
the individual's job performance and the context in which it was carried out.
As a result, the subordinates' trust in the fairness of the process increases
because they are now certain that the manager does indeed have all of the
information needed for a reasonable assessment. Furthermore, it provides
an opportunity to discuss the short- and long-term career objectives of the
subordinates, as well as their training and development needs. The problem-solving approach creates a climate of mutual trust, is nonthreatening,
and, therefore, makes the appraisal review the ideal event to discuss and set
goals for the next performance period. In this approach, managers function
as both coach and judge, but the obvious emphasis is on their mentoring role
as they seek to nurture the subordinates' strengths and minimize the negative effects of their weaknesses, if these cannot be completely eliminated.
If the performance review is to aid in performance management, the
assessment recorded should provide information that facilitates equitable
compensation decisions and identifies training and development needs.
When managers ensure equity in compensation decisions, they unleash
the tremendous motivational power of the compensation program with
89
two positive consequences, among many, that inevitably follow. First, the
correct and desirable performance behaviors are reinforced which greatly
increases the probability of such behaviors being repeated in the future
(Luthans & Stajkovic, 1999). The employees get a clear, direct, and unambiguous message of the type and level of performance behaviors expected
from them. Second, the satisfaction that follows equitable compensation
increases the value employees place on the rewards they receive from the
organization. The empirical evidence is overwhelmingly conclusive that
employee motivation is high when rewards are contingent on performance behavior, and are valued by the employees (Kanungo & Mendonca,
1988). One of the fundamental principles underlying this process is procedural justice. Therefore, the proper use of this process throughout the
organization, requires that the individual's performance measures, and
corresponding rewards, cover all facets of the job in relation to the organization's objectives, and duly recognize the support and collaboration of
colleagues who contribute to the individual's successful performance
(Bloom, 1999).
Identifying training and development needs is the necessary, but
not sufficient, first step to improving performance. It must be followed
through with appropriate programs, which remedies specific performance deficiencies or provides opportunities for the acquisition or enhancement of certain skills and abilities. This activity is especially critical to
organizations that compete in a highly dynamic and rapidly growing
industry. Therefore, continuing commitment to employee development
and growth cannot be overemphasized. As an employee observed of the
company's development program, "Federal Express made me a man. It
gave me the confidence and self-esteem to become the person I had the
potential to become" (Dessler, 1999:63).
The Performance Management Process and
the New Work Environment
The discussion of the Performance Management Process as a strategy

to reduce withdrawal behavior, did not specifically touch on the significant and even radical change in the work environment, and to work itself,
brought about by continuing developments in information technology.
The use of information technology is increasingly widespread, with beneficial results, in the "ordinary workplace"-in the business and manufacturing facilities to which employees go when they leave home each
workday. However, with the rapid growth of telecommuting, the individual's workplace can be any site, even a mobile one, that is supported by
the appropriate technological connections. Recent estimates suggest that
90
the number of telecommuters is expected to surpass 108 million by 2002

(Pinsonneault & Boisvert, 2001). Information technology is known to be
"user-friendly," and has been received with great enthusiasm by employees as a means of expediting and facilitating task performance.
Would employee withdrawal behavior be a concern in such an
exciting work environment-either in the ordinary location or offsite? In
the present state of turbulent change, it would be purely speculative to
suggest in what way withdrawal behavior might be manifested in this
environment. Information technology can, undoubtedly, improve productivity and facilitate speedy and accurate accomplishment of job tasks.
However, it can also induce withdrawal behavior in the area of performance-as evidenced by news reports of employees using company time
and resources to attend to their investment portfolios, and surf the Net to
engage in activities that hinder job performance (MacGregor, 2000). The
potential for withdrawal behavior is greatly increased in an offsite location because of the isolation from the regular work community. Indeed,
there is an increasing recognition that, in addition to adequate physical
and technological support, the effectiveness of telecommuting depends
on a relationship of confidence between the employees and managers,
establishing precise job objectives and goals, regular evaluation and feedback, and equitable rewards (Pinsonneault & Boisvert, 2001). The essential preconditions and procedural steps of the Performance Management
Process are most conducive to fostering a relationship of confidence
between employees and managers. As discussed in earlier sections, it
allows managers to create a work culture and organizational conditions
that enable employees to accept ownership and control over the job. Such
ownership and control enhance and reinforce the employee's job involvement and organizational commitment and, at least for this reason, should
be the critical first step in addressing any manifestation of employee
withdrawal behavior.
CONCLUSION
The Performance Management Process is a comprehensive strategy

to reduce employee withdrawal behavior. It incorporates two powerful
approaches that have proven to be effective in managing employee
behavior. One is the employee empowerment expereince founded on
social cognitive theory (Bandura, 1986, 1997). The other is the process
of behavior modification through reinforcement based on the operant
conditioning approach (Luthans & Stajkovic, 1999). The Performance
Management Process is at the core of managing human resources at the
91
organizational and individual levels. It ensures that the contextual factors-organizational conditions, supervisory style, reward system, and
job design-provide employees with an empowering and a reinforcing
experience to increase their job involvement and organizational commitment that is essential to reducing the withdrawal behavior. Developed on
a sound conceptual foundation, the Performance Management Process
provides a systematic approach to remedy the malaise of employee withdrawal behavior-be it in the "ordinary workplace" location or away
from it in telecommuting programs. However, it is recognized that the
evidence, cited in the chapter, relates to specific aspects of withdrawal
behavior-such as the relationship of absenteeism to its antecedent conditions <e.g., Nicholson & Payne, 1987). It does not validate the relationship of withdrawal behaviors to the various stages of the Performance
Management Process suggested in this chapter. This remains the task of
future research.
REFERENCES
Abramson, L. Y., Garber, J., & Seligman, M. E. P. (1980). Learned helplessness in
humans: An attributional analysis. In J. Garber & M. E. P. Seligman (Eds.),
Human helplessness: Theory and applications (pp. 3-34). New York: Academic
Press.
Adams, J. S. (1965). Injustice in social exchange. In L. Berkowitz (Ed.), Advances in
experimental psychology (Vol. 2, pp. 267-299). New York: Academic Press.
Bandura, A. (1986). Social foundations of thought and action: A social-cognitive view.
Englewood, NJ: Prentice-Hall.
Bandura, A. (1997). Self-efficacy: The exercise of control. San Francisco: Freeman.
Baron, J. N., & Kreps, D. M. (1999). Strategic human resources: Frameworks for general
managers. New York: Wiley.
Bishop, J. W., & Scott, K. D. (1997, February). How commitment affects team performance. HRMagazine, pp. 107-111.
Blau, P. M. (1964). Exchange and power in social life. New York: Wiley.
Blauner, R. (1964). Alienation and freedom. Chicago: University of Chicago Press.
Block, P. (1987). The empowered manager. San Francisco: Jossey-Bass.
Bloom, M. (1999). The performance effects of pay dispersion on individuals and
organizations. Academy of Management Journal, 42, 25-40.
Brown, S. P. (1996). A meta-analysis and review of organizational research on job
involvement. Psychological Bulletin, 120,235-255.
Burke, W. (1986). Leadership as empowering others. In S. Srivastra (Ed.), Executive
power (pp. 51-77). San Francisco: Jossey-Bass.
Collins, J., & Porras, J. (1997). Built to last. New York: Harper Business.
Conger, J. A., & Kanungo, R. N. (1988a). The empowerment process: Integrating
theory and practice. The Academy of Management Review, 13(3), 471-482.
92
Conger, J. A., & Kanungo, R. N. (1998b). Charismatic leadership in organizations.

Beverly Hills: Sage.
Deci, E. L. (1975). Intrinsic motivation. New York: Plenum Press.
Dessler, G. (1999). How to earn your employees' commitment. Academy of
Management Executive, 13, 58--67.
Emerson, R. M. (1962). Power-dependence relations. American Sociological Review,
27,31-41.
Fein, M. (1974). Job enrichment: A re-evaluation. Sloan Management Review, 15,
69-88.
George, J. M., & Brief, A. P. (1992). Feeling good-doing good: A Conceptual analysis of the mood at work-organizational spontaneity relationship. Psychological
Bulletin, 112, 310-329.
Hackett, R. D., & Guion, R. M. (1985). A reevaluation of the absenteeism-job satisfaction relationship. Organizational Behavior and Human Decision Processes,
35, 340-381.
Herzberg, F. (1966). Work and the nature of man. Cleveland: World.
Homans, A. (1974). Social behavior: Its elementary forms. New York: Harcourt Brace
Jovanovich.
Hulin, C. (1991). Adaptation, persistence, and commitment in organizations. In
M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational
psychology (2nd ed., Vol. 2, pp. 445-505). Palo Alto, CA: Consulting
Psychologists Press.
James, J. (1951). An experimental study of tensions in work behavior. University of
California Publications in Culture and Society, 2(4), 203-242.
Johns, G. (1987) Understanding and managing absence from work. In S. L. Dolan
& R. S. Schuler (Eds.), Readings in Canadian personnel and human resource management (pp. 324-335). St. Paul, MN: West.
Kanter, R. M. (1972). Commitment and community: Communes and utopias in sociological perspective. Cambridge, MA: Harvard University Press.
Kanter, R. M. (1979). Power failure in management circuits. Harvard Business
Review, July-August, 65-75.
Kanter, R. N. (1983). The change masters. New York: Simon & Schuster.
Kanungo, R. N. (1982). Work alienation. New York: Praeger.
Kanungo, R. N., & Mendonca, M. (1988). Evaluating employee compensation.
California Management Review, 31, 23-39.
Kanungo, R. N., & Mendonca, M. (1996). Ethical dimensions of leadership. Beverly
Hills: Sage.
Kanungo, R. N., & Mendonca, M. (1997). Compensation: Effective reward management (2nd ed.). New York: Wiley.
Katz, D., & Kahn, R. L. (1978). The social psychology of organizations. New York: Wiley.
Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work:
The causes and consequences of social loafing. Journal of Personality and Social
Psychology,S, 189-202.
Lawler, E. E. (1977). Reward systems. In J. R. Hackman & J. L. Shuttle (Eds.),
Improving life at work (pp. 163-226). Santa Monica, CA: Goodyear.
93
Lawler, E. E., Mohrman, A. M., & Resnick, S. M. (1984). Performance appraisal

revisited. Organizational Dynamics, Summer, 20-35.
Likert, R. (1961). New patterns of management. New York: McGraw-Hill.
Likert, R. (1967). The human organization. New York: McGraw-Hill.
Locke, E. A., & Bryan, J. F. (1968). Goal-setting as a determinant of the effect of
knowledge of score on performance. American Journal of Psychology, 81,
398--406.
Locke, E. A., & Latham, G. P. (1984). Goal-setting: A motivational technique that
works! Englewood Cliffs, NJ: Prentice-Hall.
Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P. (1981). Goal-setting and
task performance: 1969-1980. Psychological Bulletin, 90, 125-152.
Luthans, E, & Reif, W. E. (1972). Does job enrichment really payoff? California
Management Review, 15, 30-36.
Luthans, E, & Reif, W. E. (1973). Job enrichment: Long on theory, short on practice.
Organizational Dynamics, 3, 30--43.
Luthans, E, & Stajkovic, A. D. (1999). Reinforce for performance: The need to go
beyond pay and even rewards. Academy of Management Executive, 13,49-57.
McClelland, D. C. (1975). Power: The inner experience. New York: Irvington Press.
MacGregor, A. (2000, February 4). Surfed 'n' turfed: on-job Net abuse:
Productivity, legal issues key concerns. The Montreal Gazette, pp. A1-A2.
McGregor, D. (1960). The human side of enterprise. New York: McGraw-Hill.
Marx, K. (1932). Economic and philosophical manuscripts. In Marx-Engels
Gesamtansgabe (Vol. 3). Berlin: Marx-Engels Institute.
Mathieu, J. E., & Zajac, D. M. (1990). A review and meta-analysis of the
antecedents, correlates, and consequences of organizational commitment.
Menon, S. T. (1999). Psychological empowerment: Definition, measurement, and
validation. Canadian Journal of Behavioral Science, 31,161-164.
Meyer, J. P., & Allen, N. J. (1997). Commitment in the workplace: Theory, research, and
application. Beverly Hills: Sage.
Mowday, R. T., Porter, L. W., & Steers, R. M. (1982). Employee-organizationlinkages:
The psychology of commitment, absenteeism, and turnover. New York: Academic
Press.
Nadler, D. (1980). Concepts for the management of organizational change. New York:
Organizational Research & Consulting.
Nicholson, N., & Johns, G. (1985) The absence culture and the psychological
contract-Who's in control of absence? Academy of Management Review, 10,
397--407.
Nicholson, N., & Payne, R. (1987). Absence from work: Explanations and attributions. Applied Psychology: An International Review, 36,121-132.
Pinsonneault, A., & Boisvert, M. (2001). The impacts of telework and telecommuting on organizations and workers: What do we know. In N. Johnson (Ed.),
94
Telecommuting and virtual office: Issues and opportunities (pp. 163-185). Hershey,
PA: Idea Group Publishing.
Reichers, A. E. (1985). A review and reconceptualization of organizational commitment. Academy of Management Review, 10, 465-476.
Roethlisberger, F. J., & Dickson, W. J. (1939). Management and the worker.
Cambridge, MA: Harvard University Press.
Rothbaum, F. M., Weisz, J. R., & Snyder, S. S. (1982). Changing the world and
changing self: A two process model of perceived control. Journal of Personality
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of
reinforcement. Psychological Monographs, 1 (Whole No. 609).
Sashkin, M. (1984). Participative management is an ethical imperative.
Organizational Dynamics, Spring, 5-22.
Schacht, R. (1970). Alienation. New York: Doubleday.
Scott, K. D., & Taylor, G. S. (1985). An examination of conflicting findings on
the relationship between job satisfaction and absenteeism: A meta-analysis.
Shepard, J. M. (1971). Automation and alienation. Cambridge, MA: MIT Press.
Sims, H. P. (1977). The leader as a manager of reinforcement contingencies.
In J. G. Hunt & L. L. Larson (Eds.), Leadership: The cutting edge (pp. 121-137).
Carbondale: Southern Illinois University Press.
Somers, M. J. (1995). Organizational commitment, turnover and absenteeism: An
examination of direct and interaction effects. Journal of Organizational Behavior,
16,49-58.
Sommer, S., Bae, S. H., & Luthans, F. (1996). Organizational commitment across
cultures: The impact of antecedents on Korean employees. Human Relations,
49, 977-993.
Stagner, R. (1956). Psychology of industrial conflict. New York: Wiley.

Steers, R. M., & Rhodes, S. R. (1978). Major influences on employee attendance: A
Steers, R. M., & Rhodes, S. R. (1984). Knowledge and speculation about absenteeism. In P. S. Goodman & R. S. Atkin (Eds.), Absenteeism: New approaches
to understanding, managing, and measuring employee absence (pp. 229-275).
San Francisco: Jossey-Bass.
Stern, M. (1991, December 9). Empowerment frees employees. The Globe &
Mail, p.B4.
Szilagyi, A. D. (1980). Casual inferences between leader reward behavior and
subordinate goal attainment, absenteeism, and work satisfaction. Journal of
Occupational Psychology, 53,195-204.

Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York:
Wiley.
Tichy, N. M., & Devanna, M. A. (1986). The transformational leader. New York:
Wiley.
Whyte, W. F. (1955). Money and motivation: An analysis of incentives in industry.
New York: McGraw-Hill.
5
Meaning and Measurement of
Work Role Withdrawal
Current Controversies and Future Fallout from
Changing Information Technology
David A. Harrison
Despite a massive empirical effort, there has been no consensus on appropriate solutions to lateness, absence, or withdrawal problems (Johns,
1997). A vigorous debate about the meaning and measurement of these
behaviors has continued for decades (Blau, 1998; Hanisch & Hulin, 1990,
1991; Hanisch, Hulin, & Roznowski, 1998; Harrison & Martocchio, 1998;
Hulin, 1984; Johns, 1984, 1998; Johns & Nicholson, 1982; Martocchio &
Harrison, 1993; Mobley, 1982; Muchinsky, 1977; Rosse & Miller, 1984). The
purpose of this chapter is to evaluate some of the arguments in that
debate, and to explore the implications of those arguments under conditions of rapidly changing information technology.l begin with a review of
the constitutive definitions, or the stipulated meanings ascribed to singleand multiple-behavior types of withdrawal. I then examine the implications of those meanings for improved measurement. Finally, I predict how
those meanings are likely to evolve as the information revolution severs
the performance of many work tasks from time and place. This change in
work arrangements might necessitate a new look at the ways withdrawal
behaviors are defined and studied.
DAVID A. HARRISON Department of Management and Organization, Pennsylvania
State University, University Park, Pennsylvania 16802
95
96
David A. Harrison
CONSTITUTIVE DEFINITIONS
Withdrawal research, as with most topics of study in the behavioral
sciences, is currently dominated by the hypothetico-deductive approach
(Task Force on Statistical Inference, 1996). Initially, constructs are defined
and linkages between them are proposed before operationalizations are
formed and empirical relationships are estimated (Schwab, 1998). In the
postpositivistic paradigm under which most of this research is conducted,
constructs are not "real." Instead, they are abstractions meant to provide
a more parsimonious representation of the psychological and behavioral
world. Much of the current controversy surrounding the meaning and
measurement of withdrawal stems from disputes over the most appropriate level of such abstraction.
Constructs are initially defined constitutively. That is, the first step that
researchers are urged to take in the hypothetico-deductive approach is to
clearly state what a construct is-to stipulate its meaning (Runkel & McGrath,
1972). Constitutive definitions state the "what" of constructs; theories in
which they are embedded state the "why." Clear constitutive definitions or
stipulated meanings are critical for ensuring comparability of findings
across studies. As Hulin (1984) has noted, however, research on most
forms of withdrawal did not begin with stipulated meanings. Lateness,
absenteeism, and turnover began and stayed for many years as organizational facts or variables in search of underlying constructs and theories
that would help explain them (see Organ's 1988 work on organizational
citizenship behavior for a criterion that started in a different way).
In other words, scientific meanings were generally not a priori considerations in the earlier decades of exploratory studies that preceded
full-blown models of each form of withdrawal (Johns & Nicholson, 1982;
Martocchio & Harrison, 1993). The seriousness of that problem continues
to haunt current investigations, and should not be downplayed (e.g., the
prevalent but rather loose usage of the word "withdrawal" to refer to
anyone of the behaviors discussed, including its use in this chapter).
Constitutive definitions are the foundation for every stage of research that
follows. Lack of attention to constitutive definitions can lead to poor
measurement, and therefore uninterpretable empirical findings (StoneRomero, 1994).
However, withdrawal research is not alone in having a long history
of study before constitutive definitions are given serious consideration.
Most major streams of behavioral science research in general, and of
organizational science research in particular, started with descriptive
work on social phenomena using labels that had been part of the common
vernacular before they were the subject of rigorous investigations (see
Meaning and Measurement of Withdrawal
97
Katzell & Austin, 1992). Induction from these descriptive findings led to
insights about possible underlying constructs and relationships. One of
the best-known examples of this process is the observation of positive correlations between any two tests of intellectual performance, and the ensuing century of controversy about the constitutive and operational
definitions of intelligence (Carroll, 1993). The construct of job performance has had a similar, although somewhat shorter, scientific history
(Campbell, McHenry, & Wise, 1990). More recently, legislative actions and
judicial rulings regarding sexual harassment have clearly preceded clear
scientific definitions and comprehensive investigations of it (Fitzgerald,
Drasgow, Hulin, Gelfand, & Magley, 1997), although the behavior has
arguably been going on for as long as women have worked with men.
As with any realm of science, constitutive definitions for single- and
multiple-behavior withdrawal constructs are, to a large degree, arbitrary.
One cannot argue about their construct validity, or use data to support or
refute them. On the other hand, one can argue about how useful such definitions are, by referring to the amount of agreement scientists have regarding their stipulated meaning, and by evaluating the degree to which those
definitions support the logic, processes, and goals of scientific inquiry (e.g.,
analysis and synthesis; understanding, prediction, and control). Armed
with solid constitutive definitions of withdrawal behaviors, one can then
argue about the validity of particular operational definitions.
OPERATIONAL DEFINITIONS
Both stipulated and nomological meanings are necessary for knowledge to accrue about any phenomenon. In the second step of the idealized
research sequence mentioned above, investigators develop operational
definitions or measurements for their constitutively defined constructs.
Once developed, researchers can use a variety of data-based tools (e.g.,
multitrait multimethod matrices; Campbell & Fiske, 1959) to refine those
measurements so that they (1) fully span the construct space and (2) provide high-fidelity, low-error numerical representations of it (Hinkin,
1995). Measurements shown to be less susceptible to these dual problems
of deficiency and contamination should have higher construct validity
(Schwab, 1980). Only after operational definitions have been developed
can researchers empirically examine relationships of withdrawal behaviors to a proposed network of other constructs, and thereby come to
conclusions about their nomological meaning(s).
Deficiency and contamination issues are a strong source of contention
in withdrawal research. In current debates, proponents of a multibehavior
98
David A. Harrison
meaning for withdrawal believe that lateness, absenteeism, and turnover

data have nominal construct validity, because no construct has been clearly
defined for any of them, or each of their underlying constructs has been
backed into rather than stipulated beforehand (e.g., Hulin, 1991). And, if
each behavior is indeed part of a more general withdrawal orientation,
studying them separately creates intractable deficiency and contamination
problems. If a single behavior such as lateness is measured, it does not
cover the full spectrum of withdrawal, and behavior-specific (itemspecific) error variance cannot be distinguished from construct-relevant
variance (e.g., Roznowski & Hanisch, 1990). On the other hand, proponents of single-behavior approaches to lateness, absenteeism, and turnover
argue that current measures of general forms of withdrawal are ill-formed,
contaminated by attributions, attitudes, and intentions, and by self-report
response biases (e.g., Blau, 1998; Johns, 1998).
SINGLE WITHDRAWAL BEHAVIORS

Despite having a controversial scientific past, constitutive definitions
(stipulated meanings) do exist for single forms of withdrawal. For this
chapter, I will limit the discussion to the three that are studied most often:
absence, lateness, and turnover. All are forms of behavior. They are not
beliefs, attitudes, desires, intentions, or expectations. These latter constructs are constitutively and empirically distinct from behaviors
(Fishbein & Ajzen, 1975; Greenwald, 1989). Behaviors are sequences of
motor acts observable by a party other than the actor, but some behaviors
(e.g., absence from work) are more easily observed than others (e.g.,
drinking on the job). Single behaviors are defined in terms of an action,
target, context, and time (Ajzen & Fishbein, 1980). Multiple-behavior constructs involve several actions, perhaps aggregated over many targets,
contexts, and/ or times as well (Ajzen & Fishbein, 1977).
Absence
Constitutive Definition. An absence from work is typically defined

as lack of physical presence at a given setting and time defined by one's work
schedule (Harrison, Johns, & Martocchio, 2000). Price (1997) offered a nearly
identical definition, as did Atkin and Goodman (1984) and many others
(e.g., U.S. Bureau of Labor Statistics: Miner, 1977). Schedules are codifications of social expectations about when and where a person will be
engaged in some activity. If one's work schedule is not clearly specified, an
absence from work is not clearly defined. A similar argument could be
made about social expectations and almost all other job-related criteria.
99
The constitutive definition given above is an admittedly narrow one.

However, by moving to a slightly higher level of abstraction, the definition could be broadened to be the tendency to miss scheduled work over a
given interval. The construct label would then become absence propensity or
absence proneness. Except in rare contexts (e.g., Harrison, 1995; Smith,
1977), researchers are not interested in absence behavior on a single occasion, but instead over an extended time. How "extended" that time
should be is also a subject of debate, as different aggregation periods tend
to build up different components of variance, and therefore strengthen or
weaken connections with some theoretical antecedents at the expense of
others. Harrison and Martocchio (1998) argued for three time intervals,
but noted some arbitrariness in their demarcation. Short-term, mid-term,
and long-term absence propensity were defined as <3 months, 3 months
to 1 year, and> 1 year, respectively. Furthermore, by dropping the word
"work" and moving to a wider array of settings, one could conceptualize
an absence propensity that involved the tendency to miss many different
forms of scheduled attendance. These might include school classes, training exercises, religious services, sports practices, social club meetings,
medical appointments, and family gatherings (Harrison & Price, 2003,
in press; Johns, 1984).
Note that modifiers such as "voluntary" or "involuntary," and
"avoidable" or "unavoidable," are not part of the construct labels or constitutive definitions. These adjectives change the nature of what is studied
in the definition by going beyond the patterned expression of behavior
(Martocchio & Harrison, 1993). Instead, they stipulate a broad set of
motives, reasons, or processes that generate the behavior. It may be possible to explain a greater proportion of criterion variance by partitioning
absences into such categories, but some of that added explanatory
power is spurious, as it was stipulated beforehand. Similarly, stipulating
the meaning of an absence as "a response to X" or "because of Z" undermines its status as a distinct behavior or separate constuct, and
instead renders it as an indicator or operational definition of X or Z
(Johns, 1998). More pointedly, defining a behavior in terms of its cause(s)
makes gathering data about those causes meaningless. The cause is
already a given. One cannot support constitutive definitions and stipulated meanings with data, only operational definitions and nomological
meanings.
Note also that some absence (lateness, and turnover) researchers
might not be willing to claim their dependent variables are approximations of any particular construct, refusing to impute any additional meaning to these behaviors beyond their occurrence in organizational records.
For them, the imputation of excess meaning to the behaviors is one of the
problems they see in the multiple-behavior paradigm. Without some
100
David A. Harrison
underlying abstraction, however, my contention is that what they are

studying is not portable from situation to situation; findings regarding the
antecedents or consequences of absenteeism will always be at the level of
raw data rather than generalizations that could be made from underlying
theory.
Operational Definition. Organizational Records. Although a great

deal has been written about how to analyze absenteeism data (see
Martocchio & Harrison, 1993, for one summary), somewhat less attention
has been given to the construct validity of absence propensity measuresexcept for the aforementioned concerns about how well they capture a
voluntary versus involuntary distinction (e.g., Hackett & Guion, 1985).
There is a strong preference or orthodoxy for using organizational records
as absence measures (Johns, 1994b). This is not so much because they represent a gold standard of measurement fidelity, but because they are ordinarily taken from a different data source than measures of independent
constructs, a process that mitigates problems of common method variance
(however, see Crampton & Wagner, 1994, for little systematic evidence of
this problem in the absenteeism domain).
There has been a considerable effort to understand the reliability
of records-based measures of absence, which puts an upper bound on
construct validity. Through the extensive work of Chadwick-Jones,
Nicholson, and Brown (1982), the field of 40+ possible records-based
operational definitions has been winnowed to 2 frontrunners. Absence
frequency refers to the number of "episodes," "incidents," or "spells" of
absenteeism within a specified time period, usually without regard to the
length of time spent in any particular spell. Time lost is the total number
of days missed within any given period. Hackett and Guion (1985)
reported a slightly higher reliability for time lost than frequency measures, averaged over all test-retest, split-half, and coefficient alpha estimates in their meta-analysis. Another look at their data shows that if
fluctuations over time are taken to be the chief source of error, then frequency measures have greater reliability (stability) than time lost. The
mean and median test-retest reliability estimates for absence frequency
were .55 and .58, respectively. For time lost, they were .40 and .23.
Neither set of numbers is comforting. Both sets suggest a rather high
percentage of error variance, and therefore meager construct and criterionrelated validity for the most common absence measures. Another explanation for these substandard reliabilities is offered by Leonard, Dolan, and
Arsenault (1990) and Steel (1990). Both note observed mean or variance
differences in absenteeism over months, seasons, and years. These data
could mean there are either changing true scores on absence propensity, or
101
changing proportions of error in absenteeism measures over time. Or, they

could mean that conventional measurement theory is not appropriate for
the absenteeism domain (Atkin & Goodman, 1984). Additive true score
theory is not axiomatic. It makes assumptions that are subject to empirical
verification. If the data violate those assumptions, then some other measurement approach is warranted.
The latter conclusion was a premise in the work of Fichman (1988,
1989) and Harrison and Hulin (1989). They applied event history models
to organization records of absenteeism. Such models assume nonlinear
relationships between a time-varying absence propensity or hazard rate, a
baseline hazard function, and a vector of predictors (and possible unmeasured heterogeneity) that uniquely define each person. All three papers
demonstrated the utility of hazard rates as operational definitions of
absence propensity. Since then, however, use of these techniques has languished, perhaps because of their requirement of detailed absence records
over time rather than marginal totals. Usage has also been stymied by
their early omission in major statistical programs and their incorporation
of parameter estimates that are not readily (or profitably) convertible to
correlations. Conventional linear models and statistics are better understood by most researchers in the withdrawal area, or in any realm of organizational science.
Other Sources of Data. Organizational records of absenteeism are
archived through some process. Many times that process involves the perception and judgment of supervisors or administrators rather than use of
an automated time-keeping system. In other words, at the point of measurement, many absenteeism records are detailed social reports rather than
truly nonreactive observations. Even if records were fully automated or
nonreactive, the purposes of record-keeping systems are often at odds
with the constitutive definition of absence propensity (Webb, Campbell,
Schwartz, Sechrest, & Grove, 1981). For example, it is widely recognized
that absenteeism data for managerial or salaried employees are not faithful documentations of when they missed work (Martocchio & Harrison,
1993). For these reasons, and for concerns about confidentiality and rising
restrictions on access to organizational records, absence researchers sometimes turn to other sources of data (Johns, 1994b).
Other sources of data have the advantages of allowing investigators
to design their own measures, to fully span the stipulated meaning of
absence propensity without having to rely on some of the construct validity compromises associated with archives. Those measures are typically
self-reports but can be social (peer, supervisor) reports as well. Although
they have reasonable convergent validity (Johns, 1994b, reports a metaanalytic correlation of r = .64 of self-reports with records-based data on
102
David A. Harrison
a sample of six studies), they come with several well-known sources of

contamination. These errors arise from cognitive processing stages
involved in generating answers to self- or social-report questions
(Harrison & McLaughlin, 1996).
Perceptions of one's own absences tend to be biased downward relative to other measures, because of implicit attention and anchoring problems (Harrison & Shaffer, 1994) and the motivation to present oneself in a
positive way, even when data are collected anonymously Oohns, 1994b).
Reports on the absenteeism of others tend to be biased upward for the
same reasons, unless the target persons' absences cast the perceiver in a
bad light (e.g., a supervisor's own work group; Harrison & Shaffer, 1994;
Johns, 1994a). In both of these cases, as well as in the use of organizational
records, it is imperative that resear{:h~rs carefully examine the implicit and
explicit incentives for providing particulc,r patterns of data. If incentives exist
to provide the most accurate possible assessment (e.g., as in some types
of auditing), then these forms of contamination will be reduced.
Operationally, recognition of this problem might involve asking respondents to take the perspective of their peers, which is a common technique
in personality measurement: "How many days of work would your closest co-workers say that you missed in the last three months?" Implying
that self-report data could be checked for accuracy against records or
against other reports would also help to align the incentive system.
Noting a(n ambiguous) norm of "everyone misses work once in a while"
(Johns, 1994b) could be helpful as well.
Encoding of absenteeism questions will be best when those questions
are worded as directly and clearly as possible. Given the definition of
absence propensity, this means specifying the time interval, target, and
substance of the action, and even using context-specific jargon if it helps
to clarify meaning to the respondents. Constitutive definitions should be
universal; operational definitions can be local. In my own university, the
common phrase for professors canceling class times on their syllabus is to
give students "a walk." End-of-the-semester teacher evaluations contain
social reports by students of the number of times a professor "missed
class this semester," but that phrase might be taken by the students to
mean only unplanned absences. The difference in meaning creates a deficiency error in measurement.
Recall of absence behaviors is perhaps the toughest processing stage
for withdrawal researchers to deal with. Asking about absences in natural
time intervals such as pay periods might help to improve recall, but the
slope of the decay function for accurately remembering absenteeism is
relatively uncharted. Hackett, Bycio, and Guion (1989) found perfect correspondence between self-reported absenteeism on daily postcards and
103
organization records (a recall period of 1 or 0 days). From this starting

point, it seems reasonable to suggest that the decay of recall for absenteeism is monotonic and decreasing over the past. Johns (1994b) suggests
an outer bound of 6 months in self- or social-report questions. It may be
possible to improve recall over that period with cues or probes about the
time of year in question (including dates; see Nicholson & Payne, 1987, for
an excellent example). More valid "distributed ratings" used in self-evaluations of performance (Mabe & West, 1982) might be put to good use
here as well. For example, a researcher might be able to ask a series of
questions such as: "How many months since last December have you had
perfect attendance at work (not missed a single day)?" "In the same time
period, how many months did you miss only 1 day? .. only 2 days?" and
so on, until some open-ended upper bound such as "8 or more."
It is often difficult for respondents to go through the mapping of
recalled information to the judgment or response format required by survey questions. Does missing 4 days of work in a month mean being absent
"seldom," "sometimes," "frequently," or "a lot"? It might be better to provide no response format at all (Schwarz, 1999). That is, rather than use a set
of adverbs or numbers to generate what will be seen as a normatively
appropriate response range (Harrison & McLaughlin, 1996), open-ended or
free response answers might be preferable. Altogether, following these suggestions would lead to different-looking but perhaps more construct-valid
operational definitions of absence. Their effectiveness can be easily tested if
researchers gather both archival and questionnaire (self- and social-report)
data in the same study. And if their effectiveness is supported, they can be
applied to measures of other withdrawal behaviors for which organizational records typically do not exist (e.g., Hanisch et al., 1998).
The final point to made about these alternative ways to operationalize absence has to do with the direction of time. Absence from work is
stipulated to be a behavior. Measuring the occurrence of behaviors means
aggregating over the near or far past. The tendency to use absenteeism as
a dependent variable, when it is collected over a time period earlier than
the independent variables, has slowed but not stopped (Harrison &
Martocchio, 1998). And it is not clear whether absences are best thought
of as outcomes or inputs or both for variables such as job attitudes.
Predictive and postdictive correlations involving absenteeism are of
roughly similar size (Clegg, 1983; Rosse & Hulin, 1985), although some
postdictive correlations have been found to be larger (Tharenou, 1993).
Therefore, if self-report or social-report questions are used to operationally define absenteeism, they need to be asked at a time that is well
beyond the measurement of cognitive or attitudinal constructs thought
to be predictors. Not only will this weaken most of the mechanisms
104
David A. Harrison
proposed to create common method variance (e.g., Podsakoff & Organ,

1986), it will also bring the constructs into the proper causal order for testing most single- and multiple-behavior theories of withdrawal.
Lateness
Constitutive Definition. Lateness is constitutively defined as arriving at work after the start, or leaving before the end, of a scheduled workday
(Koslowsky, Sagie, Krausz, & Singer, 1997). The latter part of that definition (leaving early) is probably the only portion that differs from what
might be found in an English dictionary, and what might generate disagreement among withdrawal scholars (e.g., Blau, 1985). Further distinctions are sometimes made about how much time must elapse before an
instance of lateness turns into an absence. For example, Price (1997) followed Chadwick-Jones, Nicholson, and Brown (1982) in suggesting a time
threshold of 4! hours. A different threshold of 2 hours might distinguish
lateness from a "partial absence." Exact time limits may be more a function of organizational rules about sick leave or wages than behavioral
substance, and for that reason might be better handled as an operational
rather than constitutive issue.
The considerable overlap between absence and lateness definitions is
obvious. Both are temporary separations from one's work role. Both lend
themselves readily to observation in work organizations. Both involve not
performing an expected action (working), on the same target (job-related
tasks), in the same context (workplace). As with absence, the lateness definition can be abstracted to lateness propensity or lateness proneness and can
be taken to include behaviors enacted over extended time intervals or
broad sets of social settings (Harrison & Price, 2003, in press). Indeed,
Blau (1994) defines different forms of lateness in terms of patterning over
time, although it is not clear whether the operational or constitutive definitions came first for each form.
Operational Definition. Blau's (1994) dataset probably contains the

most meticulous and time-sensitive lateness measurement in the literature. Few lateness researchers can hope to have the detailed records he
was able to exploit, if they can get archival data at all. It is unlikely that
firms have as careful or as automated a system for documenting lateness
as absence. If lateness records are available, they may be social reports by
supervisors or administrative personnel. As such, they might reflect more
than individual lateness propensity. Contaminating factors in social perception, personal relationships between "data-makers" and employees,
variation in enforcement of formal rules, and differences in informal work
105
unit norms for what constitutes arriving late or leaving early, would likely
contribute to variance in that kind of operational definition.
The fact that there are fewer available lateness (than absence) data
has weakened the unspoken mandate to use only records-based measures. In a comprehensive, meta-analytic summary of lateness correlates,
Koslowsky et al. (1997) did not distinguish self-reported versus recordsbased measures, although 20% of the original studies used self-reports.
Moreover, lateness researchers do not have the benefit of published studies devoted to the reliability, stability, or validity of lateness operationalizations (lateness data were included in Chadwick-Jones et aI., 1971, but
they were not the focus of that study). Aside from Blau's (1994) contention
that different forms of lateness represent different behavioral processes,
there are also no compelling arguments about what constitutes the most
valid records-based index of lateness. Most studies appear to use a frequency measure, although time lost measures are possible. In sum, if lateness is to remain a viable, single-behavior criterion in organizational
research, a good deal of psychometric work needs to be done.
Some of that work should compare the convergent validity of selfreports, social reports, and archival records of lateness. All of the suggestions given above for improving self- and social-report instruments for
absence propensity would apply to lateness propensity as well. As a
minor, tolerated, or perhaps even ignored event in some organizations, it
is improbable that respondents clearly perceive and store instances of
lateness behavior in their memories. Hence, recall and judgment errors
will be more likely. There may be a need for stronger memory cues or a
priori questions that begin to activate more effortful cognitive processing
about coming to work late or leaving early (Harrison & McLaughlin,
1996). Although it might generate contamination due to attributional content, it could be helpful to start a series of questions with: "Occasionally,
things happen to almost everyone that make them late for work. Can you
recall events that made you late for work with in the past 3 months? What
were they? For each one, how many times did it happen over the same
3-month period? Because of their relative mildness or even banality, it
may be necessary to use a shorter recall interval for lateness probes than
for questions about other forms of withdrawal. If lateness was a serious
or highly visible problem in a research context (e.g., for flight attendants
or city bus drivers), these suggestions would be less critical. However, it
would also be more likely in such situations that the organization had put
together its own system for monitoring lateness behavior. And, in all cases
for which lateness propensity is specified as a dependent construct, lateness measurement must have a clearly defined time interval that occurs
after the measurement of independent constructs.
106
David A. Harrison
Turnover
Constitutive Definition. The stipulated meaning of turnover does

not have the level of conceptual overlap with absence and lateness as the
latter two have with each other. Price (1997) defined turnover as individual movement across the membership boundary of an organization. Similarly,
Tett and Meyer (1993) referred to it as termination of an individual's
employment with a given organization. In both cases, the target of the
behavior is implied to be the organization (Hom & Griffeth, 1995). A further distinction must be made between turnover, which is a behavioral
construct, and intentions to quit or withdrawal cognitions, which are
cognitive constructs.
Withdrawal researchers are usually interested in employee-initiated
turnover behavior, or quitting, but that form is often impossible to distinguish from firm-initiated turnover, firing or layoff (this might also be
thought of as an issue of operational rather than constitutive definitions).
Likewise, "avoidable" versus "unavoidable" and "functional" versus
"dysfunctional" labels for turnover mix the behavior with the motivational engine that may be pushing it along. In the former case, the attributions are to unforeseen or uncontrollable events (e.g., spouse
relocation). In the latter case, the attributions are to prior performance
levels that make the separation more or less desirable for the organization
(e.g., Campion, 1991). Both sets of reasons predict different components of
overall variation in turnover behavior (Harrison, Virick, & William, 1996).
In another contrast with absence and lateness propensity, there is
seldom a clear conceptual or empirical aggregation of turnover over
repeated occasions. That is, in most research on turnover, it can happen
only once per person. The tendency to leave one's organization is usually
thought of and studied as short-term turnover propensity. A likely reason is
that firms serve as the conduit for data collection rather than employees.
When a sampled set of individuals can be followed in a panel or a longitudinal design, such as in the National Longitudinal Surveys of Youth
(NLSY), researchers have been willing to investigate a long-term turnover
propensity that spans many years and multiple jobs (e.g., Judge &
Watanabe, 1995).
There are other, less frequently studied work behaviors that do share
conceptual foundations with turnover. All cases involve the permanent
separation of an employee from his or her current work role. Transfer is an
overtly similar action, but is defined as movement across work unit
boundaries, or leaving one's job but staying within the same organization
(Dalton & Todor, 1993). The opportunity to seek a transfer may only exist
in large firms with multiple units or many locations. Although early return
107
from an expatriate assignment is typically referred to as turnover and is

thought of as a form of withdrawal (Shaffer & Harrison, 1998), it fits better with the constitutive definition of transfer, as the employee usually is
repatriated within the original firm. Again, this type of behavior is only
viable for organizations with international operations. Retirement is a special, final form of turnover, defined as leaving a career or occupation in
addition to leaving one's role and organization (Hanisch & Hulin, 1990).
Operational Definition. Except in the studies mentioned above that

rely on national surveys, almost all measures of turnover propensity are
taken from organizational archives. It pays for firms to keep high-fidelity
data on turnover; separation and severance costs can be quite high. That
is, in terms of simple indicators of who is and who is not a member of an
organization at any particular time, there is a strong incentive for measurement accuracy and construct validity in organizational (especially
payroll) records. However, if researchers are interested in separating
employee-initiated from firm-initiated turnover, Campion (1991) provides convincing arguments and evidence about a lack of construct validity in those same archives for organizationally ascribed turnover reasons.
He developed a short set of questions for supervisors and former employees that could be used as a psychometrically sound index of the voluntariness of turnover.
Although turnover data are readily available from most firms, any
measure of turnover propensity involves explicit or implicit choices from
withdrawal researchers. Perhaps the most important of these involves
time. All employees leave their organizations. The only real question is,
when? In yet another contrast to lateness and absence behaviors, there is
rarely an issue of causal order with records-based turnover measures.
There is, however, a nontrivial issue involving the length of an observation window, which has ranged from a few weeks to several years (Price,
1997). Unfortunately, a rationale for that time interval is rarely offered in
turnover measurement, and what is usually distilled from organizational
records during that window is a simple dichotomy of 0 = stayer and
1=leaver (Williams & Livingstone, 1994). Much has been written about
whether investigators should correct for suboptimal base rates of this
dichotomy or >50%; e.g., Williams, 1990). Both substantive constraints
and aggregation periods have an effect on this base rate, which in turn
sets an upper bound on correlations between turnover propensity measures and normally distributed predictors.
Event history techniques circumvent both the time and base rate
problems. Such techniques (described earlier in the Absence section) estimate a time-dependent turnover propensity for each employee (Dickter,
108
David A. Harrison
Roznowski, & Harrison, 1996). On the other hand, they require moving
out of the realm of linear thinking and correlation coefficients, even
though their general specification is similar to that of a regression model.
Since their introduction to withdrawal researchers in the late 1980s (e.g.,
Morita, Lee, & Mowday, 1989), nearly a dozen studies in the organizational literature have applied event history techniques to turnover, and
many have provided insights into its determinants that would not have
been possible with conventional methods (e.g., Sheridan, 1985). However,
it is disappointing to see that the use of these demonstrably superior
methods remains far outpaced by conventional and more fallible ways to
operationalize turnover propensity in withdrawal research (Somers &
Birnbaum, 1999).
MULTIPLE WITHDRAWAL BEHAVIORS
Broad Is Preferable to Narrow

Higher levels of abstraction make constructs and theories more
portable (Sutton & Staw, 1995). A theory of real estate sales performance
cannot be generalized as well as a theory of generic work performance
(Hunt, 1996). Some researchers see greater abstraction and a movement to
multiple behaviors (as outcroppings of a broader construct) as part of the
recipe for doing better science in the withdrawal area. For example,
Hanisch et al. (1998:464) state: "Scientifically these single behaviors are
less interesting and useful with regard to building a framework of behavior generalizability in the employee withdrawal area. Good theories,
rather than case-by-case or behavior-by-behavior evaluations, are critical;
they provide the basis for generalizations across manifestations of constructs, situations, populations, and time." Under this perspective, the
study of lateness, absenteeism, or turnover as single behavioral propensities does not pay high enough scientific dividends, a claim that has
enjoined proponents of single- versus multiple-behavior approaches in a
lively debate (e.g., Hanisch et al., 1998; Johns, 1998).
However, the argument given above can be made at any level of specificity or generality of scientific constructs. Multiple sclerosis (MS) is a lowbase-rate manifestation of underlying autoimmune dysfunction. Reading
comprehension is a distinguishable, but narrow, slice of general mental
ability. Sexual harassment is a form (perhaps no more frequent than
lateness or absence) of workplace aggression or misconduct. Job satisfaction is one component of life satisfaction or subjective well-being.
Withdrawal is a single part of a broader syndrome of counterproductive or
109
deviant work behavior (e.g., Robinson & Bennett, 1995), which is, in turn,
a part of the more general construct of job performance (Van Scotter &
Motowidlo, 1996).
Yet, in each of these substantive domains, progress can be and has
been made on the molar construct without calls for abandoning the
molecular one, even though progress in the molecular approach has been
stunted at various times. The key appears to be keeping compatible levels
of abstraction within each research approach-to keep the same level of
specificity or generality between independent and dependent constructs
(Ajzen & Fishbein, 1977). For example, the demyelinization that occurs in
MS provided immunology researchers with clues about different classes
of T cells and the mechanisms by which they detect foreign tissue. New
knowledge about immune functioning led MS researchers to develop
tailored immune-suppression drugs to help slow the progress of that
disease. In the withdrawal domain, studies have shown that group
absence norms or cultures, which are shared beliefs about legitimate types
and amounts of absences, are consistently related to absence propensity
(Johns, 1997). From these investigations of more specific forms of withdrawal, researchers can learn about the transmission and reception of
normative information about work, or the development and maintenance
of within-group standards for behavior, and use them to build comprehensive models of response thresholds and temporal patterning in generalized forms of withdrawal.
Narrow Is Preferable to Broad
Looking at the controversy from the other perspective, lateness (e.g.,
Blau, 1998), absence (e.g., Martocchio & Harrison, 1993), and turnover
(e.g., Hom & Griffeth, 1995) investigators have been reticent to acknowledge the utility of a generalized withdrawal approach. This has occurred
despite the fact that most of those taking the specific approach have
adopted the "withdrawal" label for their research domain. Using that
descriptor implicitly endorses the oldest proposition-in-use (Johns &
Nicholson, 1982) for each of the single withdrawal behaviors: that they all
stem, partly, from a tendency to distance oneself from dissatisfying work.
Single-behavior advocates have also been slow to recognize the many
theoretical frameworks and empirical studies that posit or support some
form of generalized withdrawal, albeit using other labels. Table 1 shows
"neglect," "avoidance," "disengagement," "noncompliance," and "production deviance" are all labels for similar dimensions of individual work
behavior used in over a dozen different theories, taxonomies, or typologies. A thesaurus-driven literature search might find a dozen more. In
David A. Harrison
110
Table 1. Constructs Overlapping with Withdrawal in Various Theories,

Typologies, or Taxonomies of Work Behavior or Job Performance
Authors/theorists
Overall framework/ theory
Construct label
Becker & Martin (1995)
Impression management
Withdrawal (- )
Not working to potential ( - )
Campbell, McHenry, &

Wise (1990)
Dimensions of job
performance
Personal discipline (+)
Farrell (1983); Rusbult,

Farrell, Rogers, &
Mainous (1988)
Responses to job
dissatisfaction
Neglect (-)
Exit (-)
Fisher & Locke (1992);

Henne & Locke (1985)
Behavioral response to
(dis)satisfaction
Physical avoidance of job ( - )

Avoidance of work itself ( - )
Hanisch (1995); Hanisch &

Hulin (1990, 1991)
Organizational
withdrawal
Work withdrawal (-)

Job withdrawal (-)
Hogan & Hogan (1989)
Organizational
delinquency
Absenteeism ( - )
Hunt (1996)
Generic work behavior
Attendance ( + )
Off-task behavior (-)
Kahn (1990)
Calibration of self-in-role
Personal disengagement ( - )
Mangione & Quinn (1976);

Boye & Wasserman (1996)
Organ (1988); Smith,
Organ, & Near (1983)
Puffer (1987)
Prediction of job
performance
Organizational citizenship
behavior
Counterproductive
behavior (- )
Compliance ( + )
Robinson & Bennett (1996)

Rosse & Miller (1984);
Rosse & Hulin (1985);
Hulin (1991)
Deviant workplace behaviors

Adaptation to work
Production deviance ( - )
Avoidance ( - )
Van Scotter &

Motowidlo (1996)
Contextual performance
Job dedication ( + )
Prediction of job performance Noncompliance (-)
groups studies, labels such as "shirking," "free-riding," and "socialloafing" cover similar conceptual territory (Kidwell & Bennett, 1993).
Withdrawal, by any other name, would seem to have the same conceptual
sweetness.
Different Goals
Another wedge driving the single- and multiple-behavior withdrawal camps apart seems to be differential emphasis on one of two different goals of science. Those who prefer the multiple-behavior approach
111
couch most of their arguments in terms of greater abstraction and better

prediction (Roznowski & Hanisch, 1990). Aggregating over multiple, positively covarying withdrawal behaviors should yield higher proportions
of variance in the resulting composite that can be predicted by broad
independent constructs (Epstein & O'Brien, 1985; Rushton, Brainerd, &
Pressley, 1983). In contrast, those who prefer the single-behavior approach
couch most of their argument in terms of greater precision and deeper
understanding (Johns, 1998). These goals of science are not necessarily
transitive. One can have nearly perfect prediction but a weak understanding of the phenomenon under study, knowing little more than that
its current state deviates only slightly from its past state. On the other
hand, one can have nearly perfect understanding of the mechanisms
underlying the phenomenon, but make only weak and only short-term
predictions (as in chaotic systems; Cheney & Harrison, 1999).
Another difference deals with the main emphasis in either research
stream, on dependent or dependent constructs in withdrawal. Those who
study single-behavior forms of withdrawal tend to treat that behavior as
their primary interest, as a criterion to be explained. This is evident from
the variety of theories and wide scope of independent constructs that
investigators such as Blau (e.g., 1985, 1994) and Johns (e.g., Johns, 1997;
Johns & Nicholson, 1982) have brought to bear on lateness and absence.
In contrast, those interested in a multiple-behavior form of withdrawal
appear to be primarily concerned with job attitudes, and want to build
theories that bring together the many reactions that employees might
have to feeling negatively about their work. For instance, one justification
given for studying a broad withdrawal construct is the desire to take a
"multivariate perspective of employee responses to satisfaction" (Hanisch
et al., 1998:467). This emphasis on what might follow from job satisfaction, rather than what might help to explain withdrawal, is also evident
in the constitutive definitions that I turn to next.
Constitutive Definition
A final source of opposition to general, withdrawal-like constructs

stems from murkiness in their stipulated meanings. Perhaps the strongest
proponents of an overall withdrawal construct are Hanisch, Hulin, and
their colleagues (e.g., Hanisch, 1995; Hanisch & Hulin, 1990, 1991; Hulin,
1984, 1991; Roznowski & Hanisch, 1990). They define it to be "composed
of a variety of acts, or surrogate intentions, that reflect both the negativity
of the precipitating job attitudes and the target of these negative job attitudes" (Hanisch & Hulin, 1991:111). They distinguish two forms. Work
withdrawal is defined as "behaviors dissatisfied individuals use to avoid
112
David A. Harrison
aspects of their specific work role or minimize time spent on their specific
work tasks while maintaining their current organizational and work role
membership." Job withdrawal is defined as "employees' efforts to remove
themselves from a specific organization and their work role" (Hanisch &
Hulin, 1991:111).
Even after being taken to task for problems with these definitions
(e.g., Johns, 1998; Martocchio & Harrison, 1993), no revision has been
offered (Hanisch et al., 1998). The definitions are not confined to a set of
behaviors. Instead, job attitudes and intentions are stipulated to be part of the
meaning of the withdrawal construct. Under these terms, work withdrawal
is meaningless unless it is accompanied by dissatisfaction. As the cause is
constitutively defined (rather than hypothesized) in the effect, this stipulated meaning does not support the logic of research that would investigate job satisfaction as an antecedent of work withdrawal. The definition
of job withdrawal does not carry a stipulation of its cause, but both definitions give rise to operational definitions that mix attitude, intention,
and behavior variance.
Operational Definition
Current Measures. All of the behaviors presumed to span these

broadly defined withdrawal constructs are unlikely to be recorded in organizational archives. Therefore, current operational definitions are self- or
social reports. The withdrawal measures that have received the most empirical attention were developed by Hanisch, Hulin, and their colleagues.
The operational definition of work role withdrawal contains items
that deal with "unfavorable work behaviors," lateness, and absence.
Job withdrawal items deal with turnover and sometimes retirement
(Hanisch & Hulin, 1990, 1991; Roznowski & Hanisch, 1990). Questions
about unfavorable behaviors contain self-reported rates of minor forms
of poor performance such as "did poor-quality work," personal use of
company resources, and "drinking or using illicit drugs at work because
of things that occur at work" (italics added). At least three questions
are devoted to both absence and lateness in the work withdrawal measure: "How often are you _ _?" "How desirable is being _ _?" "How
likely is it that you will be _ _?" In one sample, seeking an internal
transfer was specified as a form of work withdrawal, and the first question is referred to as "thoughts of ... " being late, absent, or transferred,
rather estimates of the frequency with which each behavior occurred
(Roznowski & Hanisch, 1990, p. 368). Responses formats for all of these
questions are 4-, 5-, or 7-point frequency or semantic differential-type
scales. No time span is specified for the work withdrawal questions.
113
The operational definition of job withdrawal contains a set of three

items for turnover (thoughts of quitting, desirability of quitting, and
expectations/likelihood of quitting) "within the next few months." Also,
feeling toward, desirability of, and intentions to retire are included in the
job withdrawal measure for some samples, along with intended retirement age. Response formats for these question are similar to those given
above, except for the last question, which involves writing in a number.
Subsets of the work withdrawal and job withdrawal items are each standardized and summed. Total scores are then correlated with predictors
such as job attitude measures. Resulting correlations are fit to structural
models (e.g., Hanisch & Hulin, 1991).
Contamination. Relative to (other) constitutive definitions of withdrawal as behavior, these two measures appear to contain a large amount
of contamination. The desirability of an act is most closely related to attitude toward that act. Attitudes, intentions, and expectations are conceptually distinct from one another and from enacted behavior (Fishbein &
Ajzen, 1975; Fishbein & Stasson, 1990). The authors themselves point out
that many of the questions follow directly from operationalizations of
withdrawal cognitions (see Hom & Griffeth, 1995, for a recent summary),
which they note are hypothesized to be precursors of withdrawal
(Roznowski & Hanisch, 1990) rather than the behavior itself. At least
some portion of the fit of the authors' withdrawal models is likely to be
preordained by the overlap of attributional, and especially attitudinal,
content in these operational definitions (Schwab, 1980; Stone-Romero,
1994).
Common Method Variance. Proponents of the multiple-behavior
approach to withdrawal point to the positive relationships among the
work withdrawal or job withdrawal indicators and note in some cases
that these covariances produce "impressive" internal consistencies (e.g.,
a = .81 for job withdrawal in one sample; Hanisch et al., 1998:473).
Without positive covariation in these operational definitions, there would
be no empirical basis for a general withdrawal construct. Opponents (or
at least critics) of the multiple-behavior approach suggest that this consistency is biased upward by the use of overtly similar questions and
response formats, all asked of the same person at the same time (Blau,
1998; Johns, 1998). Counterarguments (Hanisch et al., 1998) cast the claim
of common method bias as a nebulous one (it is in most cases, but is
better specified in this one), and suggest that it would disallow any
content-based covariation, factor discrimination, or differential relationships between constructs. It does not.
114
David A. Harrison
Given the prevalence of self- and social-report measurements of

withdrawal, it is important to explore these criticisms more fully. It also
should be noted that organizational records are not free from common
method variance, especially if different persons are responsible for making observations of the lateness and absence for persons j, k, 1... versus
persons q, r, s..... There is a great deal of empirical evidence about biasing effects of common method variance, which is a catch-all term for a
category of systematic errors that can arise from transient or permanent
psychological factors contaminating responses to questions asked of the
same source. That is, all forms of common method variance arise from
systematic response tendencies that are irrelevant to the trait in question
(Podsakoff & Organ, 1986). Viswesvaran, Schmidt, and Ones (1996) used
meta-analysis to estimate that same-source variance components inflated
correlations between performance ratings by an average of r == .10-.20.
Meta-analyses of the antecedents of organizational citizenship behavior
(Organ & Ryan, 1995) and the correlates of organizational commitment
(Mathieu & Zajac, 1990) found effect sizes that were sometimes twice as
large for same-source versus different-source correlations, with absolute
amounts of inflation that were similar to those reported by Viswesvaran
et al. Doty and Glick (1998) meta-analyzed several dozen MTMM matrices and estimated that common method variance biased correlations or
path coefficients between constructs upward by 30-40%.
The amount of common method variance was not a constant in these
meta-analyses. In the most comprehensive study of it, Crampton and
Wagner (1994) suggested that common method (percept-percept) variance
depends on the types of constructs being measured and the ways that operational definitions are constructed. Therefore, if data from theoretically
related constructs corne from the same source, researchers can limit the
biasing effects of common method variance by changing features of the
measurement context. Harrison, McLaughlin, and Coalter (1996) provide
several suggestions on how to do this. For example, questions about very
narrowly defined behaviors are more likely to be susceptible to consistency
or carryover errors. The same kind of false consistency is also more likely
with items that follow one right after the other, with items having overtly
similar substantive content, and with items having overtly similar question
formats. Finally, common method errors are exacerbated by the consistent
use of response formats that do not have a natural (linguistic or numerical)
equivalent to the type of information being retrieved by a respondent.
The work withdrawal and job withdrawal instruments described
above appear susceptible to most of these potential problems. Most reliability estimates for these instruments gauge the internal consistency of
someone's agreement with statements about what they have done, how
115
often they think about it, how much they might like to do it, and how
strongly they expect to do it. Although the authors of these instruments
are unwilling to consider perception of covariation as evidence about the
structure of withdrawal (i.e., from multidimensional scaling; Hanisch
et al., 1998: 469), they are willing to consider covariation of perceptions as
unambiguous evidence of consistency within families of withdrawal
behaviors.
A more critical perspective on the data gathered with these measures
might regard them as a mix of content and common method variance.
Table 2 shows the expected reliability estimates one would obtain under
differing levels of (item-item) true score and error correlations, for scales
that have from 3 to 12 items. The range of true-score correlations, .15 ::::::
P'TT :::::: .35, was taken from the values of disattenuated meta-analytic correlations involving lateness, absence, and turnover (e.g., Koslowsky et al.,
1997; Mitra, Jenkins, & Gupta, 1992). The range of error correlations, .05::::::
PEE:::::: .20, was taken from the studies of common method variance listed
Table 2. Expected Reliability Estimates for Self- or Social-Report Measures of

Withdrawal under Varying Instrument Sizes and Amounts of True Score and
Common Method Variance
Item-item error correlation
Item-item true
correlation
Instrument size
PEE=O
3-item scale
6-item scale
9-item scale
12-item scale
PIT = .15
PIT = .20
PIT = .25
PIT = .30
PIT = .35
PIT = .15
PTT = .20
PTT = .25
PTT= .30
PTT= .35
PIT =.15
PIT = .20
PTT= .25
PTT = .30
PIT = .35
PTT=15
PTT= .20
PTT = .25
PTT = .30
PTT = .35
.35
.43
.50
.56
.62
.51
.60
.67
.72
.76
.61
.69
.75
.79
.83
.68
.75
.80
.84
.87
PEE = .05
PEE = .10
PEE =.15
PEE = .20
.43
.50
.56
.62
.67
.60
.67
.50
.56
.62
.67
.71
.67
.56
.62
.67
.71
.75
.72
.72
.76
.80
.83
.75
.79
.83
.86
.88
.80
.84
.87
.89
.91
.76
.80
.83
.86
.79
.83
.86
.88
.90
.84
.87
.89
.91
.92
.62
.67
.71
.75
.79
.76
.80
.83
.86
.88
.83
.86
.88
.90
.91
.87
.89
.91
.92
.94
.76
.80
.69
.75
.79
.83
.86
.75
.80
.84
.87
.89
.72
116
David A. Harrison
above. By specifying possible amounts of the two types of covariance,

researchers can make an educated guess at what the internal consistencies
of their own estimates would have been without the biasing effects of correlated error. For example, the ex = .81 estimate of reliability obtained in
one sample for job withdrawal might have been in the ex = .60-.70 range if
there was a typical amount of common method error. Less impressive reliability estimates found in other samples and for the work withdrawal
scale might have been much lower if correlated error was removed.
Evidence of strong or even moderate behavioral consistency using these
operational definitions remains suspect.
Recommendations and a (Kind of) Reconciliation
To be fair to the authors of the multiple-behavior withdrawal instruments, they have always been candid in stating that their research and
their instruments are works in progress (Hanisch et al., 1998). And, their
constitutive definition of withdrawal as a psychological soup of attitudes,
intentions, and behaviors matches their operational definitions. However,
if the stipulated meaning of withdrawal was that of a process (a sequence
of relationships between constructs) instead of a construct per se, it is
likely there would be much less resistance to the multiple-behavior line of
reasoning and research. A more direct and distinct set of constitutive definitions for withdrawal behaviors could then be stipulated.
Unified Meaning. There may be more agreement between the singleand multiple-behavior approaches to withdrawal than is reflected in the
tenor of recent discussions. If the fairly thin distinction between absence
and lateness is removed, it becomes apparent that both approaches to this
research domain specify two sets of withdrawal behaviors, one temporary
and one permanent. Moreover, these two forms of withdrawal share a conceptual basis that is both clear and at a fairly high level of abstraction. Each
one can be defined as part of a propensity to reduce or withhold inputs from
one's work role. This definition is simple, straightforward, and does not stipulate causes or effects. Use of the word "inputs" helps to rope in some of
the well-articulated concepts and predictions of equity theory into the withdrawal area (e.g., Johns, 1997; Schwarzwald, Koslowsky, & Shalit, 1992). If
the word "contributions" were substituted for "inputs," this underlying
theme could be traced back to original propositions by March and Simon
(1958). A similar kind of conceptual integration has been achieved in the
groups literature, where Kidwell and Bennett (1993) tied "shirking," "freeriding," and "social loafing" together as slightly different manifestations of
the propensity to withhold effort during team performance.
117
Measurement Implications. What if this unifying constitutive definition were applied to the multiple-behavior perspective on withdrawal?
There would still be an opportunity to differentiate temporary and permanent withholding of inputs. A variety of actions in addition to lateness
and absence could be gainfully applied in the operational definition of
temporary withdrawal, including some behaviors that appear in current
theories and in withdrawal-like instruments (e.g., Roznowski, Rosse, &
Miller, 1992; Rusbult et aI., 1988): taking long breaks, missing scheduled
meetings, leaving tasks for someone else to do, or performing any number of "neglectful" off-task behaviors during work hours (Hunt, 1996).
Theories of common method variance would suggest dispersing
questions about those behaviors throughout a larger surveyor interview
instrument, using items that tap unrelated constructs as psychological
"filler" tasks (Harrison et aI., 1996). To the same end of reducing potentially false consistency, the withdrawal questions should have different
response formats, perhaps asking for open-ended percentages and frequencies, as well as conventional forced choice responses that involve
numerical ranges when appropriate (e.g., "once a month"), or adverbs
("never" ... "all the time"). When opportunities for the stimulus behavior
might be limited, or might vary widely across respondents, such as for
"missing scheduled meetings," it would be beneficial to first ask how
many times the behavior was scheduled before finding out how many
times it was missed. The final number to be standardized and added to
the work withdrawal composite would be a ratio of the two (Harrison &
Price, 2003, in press).
Measurement of the behavioral tendency toward permanent withdrawal would be more circumscribed, but might benefit from the same
guidelines. Questions could probe forms of job search behavior (Blau,
1998): "how many times have you contacted other companies about job
openings," " ... checked newspaper or Internet listings of other job opportunities?" Overt behaviors such as " ... asked for a transfer" or " ... told
a co-worker that you planned to leave" would also fit within the constitutive definition.
Time. Current work withdrawal and job withdrawal instruments
have self-report questions about actions, thoughts, attitudes, and intentions spread throughout the past, present, and future. If the measures
suggested above were developed, questions or items should be about the
occurrence of (near) past behavior. Just as all behaviors are defined by
time, all items should have clear time boundaries specified (Atkin &
Goodman, 1984). Finally, as with the casual order concerns raised about
measurements of lateness and absence, these behavioral reports would
118
David A. Harrison
need to be made well after cognitive or affective predictors were measured

(Martocchio & Harrison, 1993). Until future studies of withdrawal adopt
some of these recommendations regarding measurement and take more
fine-grained approaches to timing, evidence for both single- and multiplebehavior theories of withdrawal will be ambiguous.
WITHER WITHDRAWAL? EVOLUTION OF MEANING AND
MEASUREMENT UNDER CHANGES IN INFORMATION
TECHNOLOGY
What will those future studies of withdrawal behaviors look like?
Will such fine-grained data be available to organizational researchers?
Will lateness, absence, and withdrawal have the same stipulated meaning
in the next decades as they had in the previous ones?
To answer those questions about the future, it is necessary to reexamine the past. The scientific study of withdrawal (especially absence and
lateness) was an outgrowth of the high levels of employee synchronization required by the circumscribed jobs (mainly manufacturing) of the
Industrial Revolution (Harrison et al., 2000). Near the end of the nineteenth century, being "on the clock" meant the same thing as being "on
the job" for employees in large firms, because performance of work
tasks was tightly bound to precisely scheduled times and places. Lack of
presence at those places, for those times, meant that work tasks were not
being done.
Near the end of the twentieth century, information technology
started to unbind tasks from times and places. The upshot for withdrawal
research is that the simultaneity of being "on the clock" and "on the job"
today is not as complete as it once was. For many people, both notions
have ceded to being "on-line" or being "connected," as high-speed digital communications have become a mainstay of many jobs. As this
chapter goes to press, Ford Motor Company, the prototype of a large manufacturing firm that rode (or pushed) the wave of the Industrial
Revolution, recently announced it is buying a desktop computer and
nearly free Internet access for all of its 350,000 worldwide employees. The
primary reason is to have everyone from line-level assembly workers to
top-level executives connected to e-mail and a company intranet, which
will permit low-cost and fairly immediate messages to be sent from anyone to anyone in the firm. However, another important reason is that it
removes barriers for (mostly white-collar) employees to work from home.
Delta Airlines announced a similar plan for all of its employees soon
afterward. Chrysler, GM, and other airlines are scrambling to follow suit.
119
Distributed Work
More than simply making communication easier or cheaper, information technology (IT) is allowing work tasks to he done without historically strong constraints on schedule and location (Cascio, 1998). That is,
IT is facilitating the coordination of employee inputs in ways that could
only be accomplished in the past through spatial proximity and face-toface interaction (Agres, Edberg, & Igbaria, 1998). Thus, IT enables a wide
array of flexible work arrangements. Various forms of these arrangements, such as satellite work centers, virtual offices, or telecommuting,
can be characterized as distributed work (Belanger & Collin, 1998).
Distributed work is a term describing a variety of practices that allow
employees and their tasks to be dispersed over settings outside a central
work location or times outside a synchronous work schedule (Lindstrom,
Moberg, & Rapp, 1997). Employees take part in distributed work to the
extent they work remotely from the physical and temporal locations of
their managers, co-workers, or subordinates, but on an interdependent
task. The most well-known and often-researched type of distributed work
is telecommuting, which is almost always done from an employee's home
(Feldman & Gainey, 1997).
Estimates of the number of full-time employees in the United States
working as telecommuters have ranged from 4 million people in 1990 to
11 million in 1997 (Shellenbarger, 1997). Annual growth in telecommuting
has averaged 15%. Conservative projections place the number of telecommuters at 15 million by the end of the year 2000, approaching 15% of the
U.S. work force. By 2010 that number may exceed 25% of full-time
employees (U.S. Department of Commerce, 1998), given expanding bandwidths and faster speeds for digital communication.
Change in the Constitutive Meaning of Withdrawal Behaviors
Turnover or job withdrawal means essentially the same thing in the

types of jobs described above as the types of jobs that have been studied
in the last 50 years. However, the growth in distributed work has fundamental implications for future stipulated meanings, and therefore future
operational definitions, of absence and lateness. Both of these temporary
forms of withdrawal are defined in terms of their spatial and temporal
context, as either lack of physical presence at a setting and time prescribed by
one's work schedule (absence), or as arriving at work after the start or leaving
before the end, of a scheduled workday (lateness). However, the nature of distributed work strips away parts of this defining context. Under flexplace,
physical presence or "arriving at" one's specified work setting is not
120
David A. Harrison
meaningful. Under flextime, conforming to one's specified work schedule

also is not meaningful. Under multiflex, neither the work schedule nor the
work setting is meaningful. In most of these scenarios, absence and lateness cannot be clearly defined. Therefore, single-behavior withdrawal
researchers will have two options. They can (1) ignore, at least in part,
individuals with distributed work arrangements, or (2) move to a more
abstract constitutive definition of these withdrawal behaviors.
The first option may already be the default choice. Withdrawal
researchers tend to follow the data (Hulin, 1984). Because of strong preferences within the discipline for using organizational records to measure
absence and lateness, a large and nonrandom proportion of working
adults is rarely studied. This includes many types of salaried or professional employees such as those in middle- to upper-level management, for
whom attendance records are rare or not consistently kept (Martocchio &
Harrison, 1993). Lateness records for such employees are rarer still.
Another segment of the population noticeably absent from absence
research includes the many millions of people who work for small businesses ("Sickness absence monitoring," 1994). Although attendance and
lateness records might exist in some of these local or family-owned firms,
there is a clear preference among organizational researchers to seek out
data from companies that would have a large enough work force for
reasonably powerful statistical tests (e.g., n > 50). Employees who will
be working away from work may simply join these other unstudied or
understudied populations.
On the other hand, those who take part in distributed work arrangements typically do not work from home all the time. The typical telecommuter evenly splits the time spent at work and the time spent working at
home (an average of 19.3 hours per week away from a centralized workplace; Telecommute America, 1997). And, even if the amount of necessary
"face time" at the traditional workplace declines, it is unlikely to disappear altogether (McCloskey & Igbaria, 1998). For these conventional
workdays, absenteeism retains its conventional definition.
The second option in single-behavior withdrawal research would be to
redefine absence and lateness in terms of the integrative definition given
above: as instances of temporarily withholding inputs from one's work role. For
those whose work schedules are not clearly defined by both time and place,
"absenteeism" and "lateness" might instead be labeled "task avoidance" or
"short-term withdrawal." These more inclusive terms and concepts share
conceptual foundations with some of the theories presented in Table I, and
they are more closely aligned with the multiple-behavior approach to withdrawal. That is, under distributed work, short-term withdrawal would
manifest itself as lack of (complete) participation in a specific assignment,
121
project, or set of tasks. Moving up to that level of abstraction would let

investigators cast a wider data collection net, and would allow the findings
of withdrawal research to be more robust to changes in the nature of jobs
and in the types of records kept by organizations.
Also at that level of abstraction, the historical divisions between
work role withdrawal and task motivation research would start to dissipate. When work moves out of the workplace, the process of allocating
time and energy to it might become the basic mechanism under study in
both the withdrawal and motivation research streams (e.g., Naylor,
Pritchard, & Ilgen, 1980). In addition, behaviors (or the lack of them) that
reflect a withholding of inputs would have to be mapped with greater
precision into the criterion space, alongside dimensions of organizational
citizenship behaviors (Organ, 1988), contextual performance (Van
Scotter & Motowidlo, 1996), and task performance (Campbell et al., 1990).
More immediately, the challenge of measuring this withholding of inputs
to one's work role will need to be met. Fortunately for researchers (but
perhaps less fortunately for employees), distributed work arrangements
enabled by IT have built-in means through which this measurement
might be accomplished.
Changes in the Operation Definitions of Withdrawal Behaviors
Two broad classes of jobs are easiest to accommodate under distributed work: those that involve a high division of labor and external control, and those that involve a low division of labor and internal control
(Ford & Butts, 1995). Although withholding of inputs is possible in any
work role, it is the former type of job (e.g., clerical tasks; customer service;
programming) for which distributed work is growing the fastest
(Dickisson, 1997). In these arrangements, the physical separation of
employees from supervisors and co-workers makes them less susceptible
to absence and lateness measures such as time sheets and punch cards
(Handy & Mokhtarian, 1995).
Process-Based Measures. Most forms of distributed work are

enabled through bidirectional information technologies. Just as if an
employee were working at a desk in a centralized office, computerized
performance monitoring (CPM) systems can continuously track and
report any action performed by an employee working remotely and mediated through some IT. Anything that can be digitized can be recorded by
employers in real time through high-speed Ethernet, cable, or digital subscriber lines (or within 10 years, by wireless media)-if there a main
processor or server at a centralized workplace or if the employee has a
122
David A. Harrison
networked computer. These records can provide supervisors with information about "electronic attendance," work speed, work completed (keystrokes, pages, lines of code), completed forms, correspondence, call logs,
connect time, and errors-as soon as an employee works on a connected
telecommunications system (Ambrose & Alder, 2000). Other forms of IT
can even allow a firm to monitor physical presence in front of a computer
(e.g., a small video camera or "webcam").
Despite legislative and scientific disputes surrounding the intrusiveness of CPM (George, 1996), it is already commonplace in U.S. businesses.
In a survey conducted by the American Management Association, over
two-thirds of its members reported doing some form of electronic
employee surveillance. Over half reported using CPM (Lavelle, 1999).
Some of the critical issues in the disputes are the recording of taskrelevant versus irrelevant information, and disclosure to employees that
monitoring is occurring (Ambrose & Alder, 2000). If these procedural justice hurdles can be overcome, CPM-based records will become ubiquity
for employees both in and out of distributed work arrangements.
With employee permission, observations could be taken from these
records and modified by researchers to use as measures of generic inputs
to one's work role. The same IT that allows bidirectional communication
might also be used to administer short, time-sensitive measures of job attitudes and perceptions in a more-or-Iess continuous fashion (e.g., at the
beginning and end of each work session; Hackett et al., 1989). Such data
would permit unambiguous testing of some of the major within-person
hypotheses regarding withdrawal (Harrison & Martocchio, 1998). A final
benefit of this technology is its ability to detect out-of-bounds responses
and other random errors at the point of measurement, and to immediately
record and structure data files.
Outcome-Based Measures. The other approach to withdrawal
measurement under distributed work might take advantage of the elaborate processes by which a supervisor and employee in such an arrangement are urged to agree on what constitutes good performance, and how
such performance will be evaluated. For example, employees who work
remotely might be asked to sign a "telecommuter agreement." This collaborative contract states exactly what is expected of both employee and
supervisor, covering such items as the regularity of communication, what
constitutes completion of particular objectives, the deliverable quality
and quantity of task products, deadline parameters, the format in which
results will be reported and assessed, and specification of the frequency
and timing of conventional attendance at a central workplace (Cooper,
1996). Although these outcome-based measures would not contain the
123
same level of detail as process-based measures, development of the agreement described above usually requires more careful thought from the
organization about what should be measured for the telecommuter than
for employees in conventional work arrangements (e.g., Dickisson, 1997;
McNerney, 1995).
That is, under telecommuting and other forms of distributed work,
the attendance measurement system (Gardiner, 1992) is likely to be
expanded, or perhaps be aligned completely with the firm's performance
measurement system (Naylor et aI., 1980). Such a change would also be
consistent with a change in the constitutive definition of absenteeism.
Obviously, researchers would need to work closely with firms and
employees to make these weekly or monthly types of records into useful
assessments of withdrawal behavior. Current withdrawal models could
still be tested with powerful within-subject designs, if attitudinal and
perceptual data could be collected at regularly spaced intervals (e.g.,
Westman & Eden, 1997).
Other Sources of Data. It is possible that organizational records created from IT-enabled work will be unavailable or unconnected to
researchers' constitutive definitions of withdrawal. In that case, tacit prescriptions against the use of other sources of data will need to be relaxed
(Johns, 1998). And, given the nature of such work arrangements, social
reports of withdrawal behaviors from co-workers or supervisors would
have little construct validity. These circumstances underscore the importance of developing sound self-report measures of everyday forms of
withdrawal (behaviors, not feelings) using the guidelines suggested
above for self-reports of absence, lateness, or work-role separation.
If withdrawal is redefined as the reduction of inputs to one's work
role, it may be necessary to develop sets of job-specific (tailored to the
context) and generic (portable across contexts) items. Reverse-scored
measures of discretionary (OCB-like) behaviors could fall into either category depending on the jobs being studied, and would be especially sensitive to changes in levels of purported independent variables (they
would also be a source of deficiency in measures based solely on organizational records). Under distributed work, reductions of these kinds of
prosocial inputs to one's work role are the easiest to enact, involve the
lowest cost, and entail the least risk. The potency of interpersonal consequences is reduced by the limited bandwidth of communication offered
by most information technologies. Limited face-to-face interaction, and a
more permeable membrane between work and family demands is also
likely to change the nomological meaning of withdrawal behavior
(Harrison et a1., 2000).
124
David A. Harrison
CONCLUSION
For nearly a century, absence, lateness, and turnover have been

viewed as expensive organizational problems and intriguing social phenomena. For nearly a half-century they have been studied as different
forms of withdrawal from one's work role. Yet, controversies about how
these behaviors should be constitutively and operationally defined are
still smoldering. Even within the last 5 years, debates have been rekindled
regarding how different research approaches provide a weaker or
stronger scientific foundation from which to investigate withdrawal (e.g.,
Blau, 1998; Hanisch et al., 1998; Johns, 1998).
The strongest and most consistent argument I have made in this
paper is that withdrawal researchers need to be more thoughtful about
the stipulated meanings and subsequent measures of the constructs they
are investigating. To that end, I offered simple constitutive definitions for
single withdrawal behaviors, and I made a series of suggestions about
how they might be operationally defined through organizational records,
social reports, and self-reports. I also offered criticisms and suggestions
for both types of definitions currently being used in multiple-behavior
withdrawal research. A careful comparison of both of these broadly construed approaches that have presumably been at odds with each other
shows that they share more than might be recognized at first glance. Each
"camp" of investigators distinguishes temporary and permanent forms of
withdrawal, and each acknowledges that negative job attitudes play some
role in the withdrawal process.
Because changes in technology will redefine the temporal and spatial
context of jobs, perhaps removing those contextual elements altogether, it
seems that it would be helpful if research in the withdrawal domain
moved toward an integrative definition of withdrawal-as a propensity
to withhold inputs from one's work role. Adopting this definition would
not only provide a unifying set of meanings to the withdrawal area, but it
would lend itself to straightforward development of measures. Clearer
meanings and sounder measures will pay both theoretical and empirical
dividends for many years to come.
REFERENCES
Agres, c., Edberg, D., & Igbaria, M. (1998). Transformation to virtual societies:
Forces and issues. The Information Society, 14, 71-82.
Ajzen, 1., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis
and review of empirical research. Psychological Bulletin, 84, 888-918.
125
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.
Ambrose, M. L., & Alder, G. S. (2000). Designing, implementing, and utilizing
computerized performance monitoring: Enhancing organizational justice. In
G. Ferris (Ed.), Research in personnel and human resources management (Vol. 18,
pp. 187-219). Greenwich, CT: JAI Press.
Atkin, R. 5., & Goodman, P. S. (1984). Methods of defining and measuring absenteeism. In P. S. Goodman & R. S. Atkin (Eds.), Absenteeism: New approaches to
understanding, measuring, and managing employee absence (pp. 47-109). San
Francisco: Jossey-Bass.
Becker, T., & Martin, S. (1995). Trying to look bad at work: Methods and motives
for managing poor impressions in organizations. Academy of Management
Journal, 38,174-199.
Belanger, E, & Collins, R. W. (1998). Distributed work arrangements: A research
framework. The Information Society, 14, 137-152.
Blau, G. J. (1985). Relationship of extrinsic, intrinsic, and demographic predictors
to various types of withdrawal behaviors. Journal of Applied Psychology, 70,
442-450.
Blau, G. J. (1994). Developing and testing a taxonomy of lateness behavior. Journal
of Applied Psychology, 79, 959-970.
Blau, G. J. (1998). On the aggregation of individual withdrawal behaviors into
larger multi-item constructs. Journal of Organizational Behavior, 19, 437-451.
Boye, M. W., & Wasserman, A R. (1996). Predicting counterproductive behavior
among drug store applicants. Journal of Business and Psychology, 10, 337-349.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by
the multitrait-multimethod matrix. Psychological Bulletin, 56,81-105.
Campbell, J. P., McHenry, J. J., & Wise, L. L. (1990). The modeling of performance
in a population of jobs. Personnel Psychology, 43,313-333.
Campion, M. A (1991). Meaning and measurement of turnover: Comparison of
alternative measures and recommendations for research. Journal of Applied
Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies.
London: Cambridge University Press.
Cascio, W. E (1998). The virtual workplace: A reality now. The IndustrialOrganizational Psychologist, 37, 32-36.
Chadwick-Jones, J. K., Brown, C. A, Nicholson, N., & Sheppard, C. (1971).
Absence measures: Their reliability and stability in an industrial setting.
Chadwick-Jones, J. K., Nicholson, N., & Brown, C. A (1982). The social psychology
of absenteeism. New York: Praeger.
Cheney, M.S., & Harrison, D. A (1999). Uncertainty revisited: Nonlinear, chaotic
changes on the journey toward decision resolution. Paper presented at the national
meeting of the Academy of Management, Chicago.
Clegg, C. W. (1983). Psychology of employee lateness, absence, and turnover:
A methodological critique and an empirical study. Journal of Applied
126
David A. Harrison
Cooper, R. C. (1996). Telecommuting: The good, the bad and the particulars.
Supervision, 57, 10-12.
Crampton, S. M., & Wagner, J. A. (1994). Percept-percept inflation in microorganizational research: An investigation of prevalence and effect. Journal of
Dalton, D. R., & Todor, W. D. (1993). Turnover, transfer, absenteeism: An interdependent perspective. Journal of Management, 19, 193-219.
Dickisson, K. H. (1997). Telecommuting: Got your homework done? CMA
Magazine, 3-14.
Dickter, D., Roznowski, M., & Harrison, D. A. (1996). Temporal tempering: An
event history analysis of the process of voluntary turnover. Journal of Applied
Doty, D. H., & Glick, W. H. (1998). Common methods bias: Does common methods
variance really bias results? Organizational Research Methods, 1(4), 374-406.
Epstein, S., & O'Brien, E. J. (1985). The person-situation debate in historical and
current perspective. Psychological Bulletin, 98, 513-537.
Farrell, D. (1983). Exit, voice, loyalty, and neglect as responses to job dissatisfaction: A multidimensional scaling study. Academy of Management Journal, 26,
596--607.
Feldman, D. c., & Gainey, T. W. (1997). Patterns of telecommuting and their consequences: Framing the research agenda. Human Resource Management Review,
7,369-388.
Fichman, M. (1988). Motivational consequences of absence and attendance:
Proportional hazard estimation of a dynamic motivation model. Journal of
Fichman, M. (1989). Attendance makes the heart grow fonder: A hazard rate
approach to modeling attendance. Journal of Applied Psychology, 74,325-335.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Fishbein, M., & Stasson, M. E (1990). The role of desires, self-predictions, and perceived control in the prediction of training session attendance. Journal of
Applied Social Psychology, 20,173-198.
Fisher, C. D., & Locke, E. A. (1992). The new look in job satisfaction research and
theory. In C. J. Cranny, P. C. Smith, & E. E Stone (Eds.), Job satisfaction: How people feel about their jobs and how it affects their performance. New York: Lexington,
165-194.
Fitzgerald, L. E, Drasgow, E, Hulin, C. L., Gelfand, M. J., & Magley, V. J. (1997).
Antecedents and consequences of sexual harassment in organizations: A test
of an integrated model. Journal of Applied Psychology, 82, 578-589.
Ford, R. c., & Butts, M. A. (1995). Questions and answers about telecommuting
programs. Business Horizons, 38, 66-72.
Gardiner, R. C. (1992). Tracking and controlling absenteeism. Public Productivity &
Management Review, 15, 289-307.
George, J. E (1996). Computer-based monitoring: Common perceptions and
empirical results. Management Information Systems Quarterly, 20, 459-480.
127
Greenwald, A G. (1989). Why are attitudes important? In A R. Pratkanis,

S. J. Brecker, & A G. Greenwald (Eds.), Attitude structure and junction
(pp. 1-10). Hillsdale, NJ: Erlbaum.
Hackett, R. D., Bycio, P., & Guion, R. M. (1989). Absenteeism among hospital
nurses: An idiographic-longitudinal analysis. Academy of Management Journal,
32, 424-453.
Hackett, R. D., & Guion, R. M. (1985). A re-evaluation of the absenteeism-job satisfaction relationship. Organizational Behavior and Human Decision Processes,
35,340-38l.
Handy, S. L., & Mokhtarian, P. L. (1995). Planning for telecommuting. Journal of the
American Planning Association, 61 (1), 99-11l.
Hanisch, K. A (1995). Behavioral families and multiple causes: Matching the complexity of responses to the complexity of antecedents. Current Directions in
Psychological Science, 4, 156-162.
Hanisch, K. A, & Hulin, C. L. (1990). Job attitudes and organizational withdrawal:
Hanisch, K. A, & Hulin, C. L. (1991). General attitudes and organizational withdrawal: An evaluation of a causal model. Journal of Vocational Behavior, 39,
110-128.
Hanisch, K. A, Hulin, C. L., & Roznowski, M. (1998). The importance of individuals' repertoires of behaviors: The scientific appropriateness of studying multiple
behaviors and general attitudes. Journal of Organizational Behavior, 19,463-480.
Harrison, D. A, & Hulin, C. L. (1989). Investigations of absenteeism: Using event history models to study the absence-taking process. Journal of Applied Psychology,
74,300--316.
Harrison, D. A, Johns, G., & Martocchio, J. J. (2000). Changes in technology, teamwork, and diversity: New directions for a new century of absenteeism
research. In G. Ferris (Ed.), Research in personnel and human resources management (Vol. 18, pp. 43-91). Greenwich, CT: JAI Press.
Harrison, D. A, & Martocchio, J. J. (1998). Time for absenteeism: A 20-year review
of origins, offshoots, and outcomes. Journal of Management, 24, 305-350.
Harrison, D. A, & McLaughlin, M. E. (1996). Structural properties and psychometric qualities of self-reports: Field tests of connections derived from cognitive theory. Journal of Management, 22, 313-338.
Harrison, D. A, McLaughlin, M. E., & Coalter, T. M. (1996). Context, cognition,
and common method variance: Psychometric and verbal protocol evidence.
Organizational Behavior and Human Decision Processes, 68, 246-26l.
Harrison, D. A, & Price, K. H. (2003, in press). Context and consistency in absenteeism: Studying dispositional and social influences across multiple settings.
Harrison, D. A, & Shaffer, M. A (1994). Comparative examinations of self reports
and perceived absenteeism norms: Wading through Lake Wobegon. Journal of
Applied Psychology, 79,240-251.
128
David A. Harrison
Henne, D., & Locke, E. A. (1985). Job dissatisfaction: What are the consequences?
Intemational Journal of Psychology, 20, 221-240.
Hinkin, T. (1995). A review of scale development practices in the study of organizations. Journal of Management, 21, 967-988.
Hogan, J., & Hogan, R. (1989). How to measure employee reliability. Journal of
Hom, P. w., & Griffeth, R. W. (1995). Employee tumover. Cincinnati, OH: SouthWestern.
Hulin, C. L. (1984). Suggested directions for defining, measuring, and controlling
absenteeism. In P. S. Goodman & R. S. Atkin (Eds.), Absenteeism: New
approaches to understanding, measuring, and managing employee absence (pp. 391420). San Francisco: Jossey-Bass.
Hulin, C. L. (1991). Adaptation, persistence, and commitment in organizations. In
M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (2nd
ed., pp. 445-505). New York: Wiley.
Hunt, S. T. (1996). Generic work behavior: An investigation into the dimensions of
entry-level, hourly job performance. Personnel Psychology, 49, 51-83.
Johns, G. (1984). Suggested directions for defining, measuring, and controlling
absenteeism. In P. S. Goodman & R. S. Atkin (Eds.), Absenteeism: New approaches
to understanding, measuring, and managing employee absence (pp. 360-389).
Johns, G. (1994a). Absenteeism estimates by employees and managers: Divergent
perspectives and self-serving perceptions. Journal of Applied Psychology, 79,
229-239.
Johns, G. (1994b). How often were you absent? A review of the use of self-reported
absence data. Journal of Applied Psychology, 79, 574-59l.
Johns, G. (1997). Contemporary research on absence from work: Correlates,
causes, and consequences. International Review of Industrial and Organizational
Johns, G. (1998). Aggregation or aggravation? The relative merits of a broad withdrawal construct. Journal of Organizational Behavior, 19,453--462.
theory and research. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 4, pp. 127-172). Greenwich, CT: JAI Press.
Judge, T. A., & Watanabe, S. (1995). Is the past prologue?: A test of Ghiselli's Hobo
Syndrome. Journal of Management, 21, 211-229.
Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at work. Academy of Management Journal, 33, 692-724.
Katzell, R. A., & Austin, J. T. (1992). From then to now: The development of industrial-organizational psychology in the United States. Journal of Applied
Kidwell, R. E., Jr., & Bennett, N. (1993). Employee propensity to withhold effort: A
conceptual model to intersect three avenues of research. Academy of
Management Review, 18,429-456.
Koslowsky, M., Sagie, A., Krausz, M., & Singer, A. D. (1997). Correlates of
employee lateness: Some theoretical considerations. Journal of Applied
129
Lavelle, L. (1999, April 16). Survey shows increase in firms that monitor employees' communications. The Record. Hackensack, NJ: Knight-Ridder Tribune
Business News.
Leonard, c., Dolan, S. L., & Arsenault, A. (1990). Longitudinal examination of the
stability and variability of two common measures of absence. Journal of
Occupational Psychology, 63, 309-316.
Lindstrom, J., Moberg, A., & Rapp, B. (1997). On the classification of telework.
European Journal of Information Systems, 6, 243-255.
Mabe, P. A., III, & West, S. G. (1982). Validity of self-evaluation of ability: A review
and meta-analysis. Journal of Applied Psychology, 67, 280-296.
Mathieu J. E. & Zajac D. M. (1990). A review and meta-analysis of the antecedents,
correlates, and consequences of organizational commitment. Psychological
Bulletin, 108, 171-194.
McCloskey, D. W., & Igbaria, M. (1998). A review of the empirical research on
telecommuting and directions for future research. In M. Igbaria & M. Tan
(Eds.), The virtual workplace (pp. 338-358). Hershey, PA: Idea Group Publishing.
McNerney, D. J. (1995). Telecommuting: An idea whose time has come. HR Focus,
72(11), 1-4.
Mangione, T., & Quinn, R. (1975). Job satisfaction, counterproductive behavior,
and drug use at work. Journal of Applied Psychology, 60,114-116.
Martocchio, J. J., & Harrison, D. A. (1993). To be there or not to be there?
Questions, theories and methods in absenteeism research. In K. Rowland &
G. Ferris (Eds.), Research in personnel and human resources management (Vol. 11,
pp. 259-329. Greenwich, CT: JAI Press.
Miner, M. G. (1977). Job absence and turnover: A new source of data. Monthly
Labor Review, 100, 24-31.
Mitra, A., Jenkins, G. D., & Gupta, N. (1992). A meta-analytic review of the relationship between absence and turnover. Journal of Applied Psychology, 77, 879-889.
Mobley, W. H. (1982). Some unanswered questions in turnover and withdrawal
research. Academy of Management Review, 7, 111-116.
Morita, J. G., Lee, T. W., & Mowday, R. T. (1989). Introducing survival analysis to
organizational researchers: A selected application to turnover research.
Muchinsky, P. M. (1977). Employee absenteeism: A review of the literature. Journal
of Vocational Behavior, 10, 316-340.
Naylor, J. c., Pritchard, R. T., & Ilgen, D. R. (1980). A theory of behavior in organizations. New York: Academic Press.
Nicholson, N., & Payne, R. (1987). Absence from work: Explanations and attributions. Applied Psychology: An International Review, 36(2), 121-132.
Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome.
Lexington, MA: Heath.
Ostroff, c., & Harrison, D. A. (1999). Meta-analysis, level of analysis, and best estimates of population correlations: Cautions for interpreting meta-analytic
results in organizational behavior. Journal of Applied Psychology, 84, 260-270.
Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research:
Problems and prospects. Journal of Management, 12, 531-544.
130
David A. Harrison
Price, J. L. (1997). Handbook of organizational measurement. International Journal
of Manpower, 18,303-558.
Puffer, S. M. (1987). Prosocial behavior, noncompliant behavior, and work performance among commission salespeople. Journal of Applied Psychology, 72,
615-62l.
Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors: A multidimensional scaling study. Academy of Management Journal, 38,
555-572.
Rosse, J. G., & Hulin, C. L. (1985). Adaptation to work: An analysis of employee

health, withdrawal, and change. Organizational Behavior and Human Decision
Processes, 36, 324-347.
Rosse, J. G., & Miller, H. E. (1984). Relationship of absenteeism and other

employee behaviors. In P. S. Goodman & R. S. Atkin (Eds.), Absenteeism:
New approaches to understanding, measuring, and managing employee absence

(pp. 194-228). San Francisco: Jossey-Bass.
Roznowski, M., & Hanisch, K. A. (1990). Building systematic heterogeneity
into job attitudes and behavior measures. Journal of Vocational Behavior, 36,
361-375.
Roznowski, M., & Hulin, C. L. (1992). The scientific merit of valid measures of
general constructs with special reference to job satisfaction and job withdrawal. In C. J. Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction:
Advances in theory and research. New York: Free Press.
Roznowski, M., Rosse, J. G., & Miller, H. E. (1992). On the utility of broad-band measures of employee behavior: The case for employee adaptation and citizenship. Paper
presented at the annual meeting of the Academy of Management, Las Vegas.
Runkel, P. J., & McGrath, J. E. (1972). Research on human behavior: A systematic guide
to method. New York: Holt, Rinehart & Winston.
Rusbult, C. E., Farrell, D., Rogers, G., & Mainous, A. G. (1988). Impact of exchange
variables on exit voice, loyalty, and neglect: An integrative model of responses
to declining job dissatisfaction. Academy of Management Journal, 31, 599-627.
Rushton, J. P., Brainerd, C. J., & Pressley, M. (1983). Behavioral development and
construct validity: The principle of aggregation. Psychological Bulletin, 94, 18-38.
Schwab, D. P. (1980). Construct validity in organizational behavior. In
L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 2,
pp. 3-43). Greenwich, CT: JAI Press.
Schwab, D. P. (1998). Research methods for the organizational sciences. Hillsdale, Nl:
Erlbaum.
Schwarz, N. (1999). Self-reports: How the questions shape the answers. American
Psychologist, 54, 93-105.
Schwarzwald, J., Koslowsky, M., & Shalit, B. (1992). A field study of employees'
attitudes and behaviors after promotion decisions. Journal of Applied
Shaffer, M. A., & Harrison, D. A. (1998). Expatriates' psychological withdrawal

from international assignments: Work, non-work, and family influences.
131
Shellenbarger, S. (1997, August 20). Madison Avenue may need to alter image of
'90s telecommuter. The Wall Street Journal, p. B1.
Sheridan, J. E. (1985). A catastrophe model of employee withdrawal leading to low
job performance, high absenteeism, and job turnover during the first year of
employment. Academy of Management Journal, 28, 88-109.
Sickness absence monitoring and control: A survey of practice. (1994, September).
Industrial Relations Review and Report, 4-16.

Smith, F. J. (1977). Work attitudes as predictors of attendance on a specific day.
Somers, M. J., & Birnbaum, D. (1999). Survival versus traditional methodologies
for studying employee turnover: Differences, divergences and directions for
future research. Journal of Organizational Behavior, 20, 273-284.
Steel, R P. (1990). Psychometric theory and the problem of relating prior and subsequent absences. Journal of Organizational Behavior, 11, 407-411.
Stone-Romero, E. F. (1994). Construct validity issues in organizational behavior
research. In J. Greenberg (Ed.), Organizational behavior: The state of the science
(pp. 155-180). Hillsdale, NJ: Erlbaum.
Sutton, R I., & Staw, B. M. (1995). What theory is not. Administrative Science
Quarterly, 40, 371-384.

Task Force on Statistical Inference. (1996). Initial Report. Washington, DC: Board
of Scientific Affairs, American Psychological Association.
Telecommute America. (1997). "News Release July 2, 1997." Available from
http://www.att.com/press/0797/970702.bsa.html#facts.
Tett, R P., & Meyer, J. P. (1993). Job satisfaction, organizational commitment,
turnover intention, and turnover: Path analysis based on meta-analytic
findings. Personnel Psychology, 43, 259-293.
Tharenou, P. (1993). A test of reciprocal causality for absenteeism. Journal of
Organizational Behavior, 14(3),269-287.
U. S. Department of Commerce. (1998, April). The emerging digital economy.
Washington, DC: U. S. Government Printing Office. NTIS order number:
PB98-137011INQ.
Van Scotter, J. R, & Motowidlo, S. J. (1996). Interpersonal facilitation and job dedication as separate facets of contextual performance. Journal of Applied
Viswesvaran, c., Schmidt, F. L., & Ones, D. S. (1996). Comparative analysis of the
reliability of job performance ratings. Journal of Applied Psychology, 81, 557-574.
Webb, E. J., Campbell, D. T., Schwartz, R D., Sechrest, L., & Grove, J. B. (1981).
Nonreactive measures in the social sciences. Boston: Houghton Mifflin.
Westman, M., & Eden, D. (1997). Effects of a respite from work on burnout:
Vacation relief and fade-out. Journal of Applied Psychology, 82, 516-527.
Williams, C. R (1990). DeCiding when, how, and if to correct turnover correlations.
Williams, C. R., & Livingstone, L. P. (1994). Another look at the relationship
between performance and voluntary turnover. Academy of Management
Journal, 37, 269-298.
6
Developing and Testing a
Taxonomy of Lateness Behavior
Gary Blau
Employee lateness has long been recognized by organizations as

a behavior necessitating monitoring and control (e.g., Motley, 1926).
Company rules and regulations for employees almost always contain
a section that communicates the organization's policies for handling
employee lateness. The costs associated with employee lateness include
the loss of late-employee productivity, the administration time of supervisors (e.g., counseling-disciplining late employees), and the negative
impact on other workers who have to pick up the slack (Cascio, 1987). For
many years, application-oriented articles have been written discussing
how to prevent or reduce employee lateness (e.g., Kempen, 1982;
Kennedy, 1984; Kite, 1984; Ruchti, 1967).
Despite the necessity of organizations to at least monitor employee
lateness, there has been very little attempt to develop a systematic
approach for studying employee lateness. This lack of conceptual work
on employee lateness stands in stark contrast to the large body of literature devoted to conceptual models on the related physical withdrawal
behaviors (Blau & Boal, 1987; Rosse & Hulin, 1985) of absenteeism (e.g.,
Brooke, 1986; Gibson, 1966; Steers & Rhodes, 1978) and turnover (e.g.,
March & Simon, 1958; Mobley, 1977; Mobley, Griffeth, Hand, & Meglino,
1979; Steers & Mowday, 1981). The purpose of this chapter is to develop
GARY BLAU Human Resource Administration Department, Temple University, Fox
School of Business Administration, Philadelphia, Pennsylvania 19122.
133
134
Gary Blau
and test a taxonomy of lateness behavior. It will be argued that different

categories of lateness behavior exist and can be differentiated on the basis
of (1) the pattern, frequency, and duration of lateness incidents; (2) the relationships of lateness-behavior categories to other physical withdrawal
behaviors; and (3) different antecedents for the lateness-behavior categories.
Unmasking different categories of lateness behavior will improve the
ability to describe and understand the construct of lateness. Increased
explanation of lateness behavior may help organizations to more effectively control such behavior. Empirical work has already been done to
develop taxonomies of other physical withdrawal behaviors, including
turnover behavior (Abelson, 1987; Campion, 1991) and absenteeism
(Dalton & Mesch, 1991). It is my position that further research is needed
to understand lateness, before combining lateness with other withdrawal
behaviors, such as absence, into a more general work withdrawal construct (Hanisch & Hulin, 1990, 1991). Such a general work withdrawal
construct will undoubtedly be useful. However, our current lack of
knowledge regarding categories or types within each withdrawal behavior would seem to require first fully understanding the construct validity
of each behavior (Schwab, 1980) before aggregating those behaviors into
a more general construct.
Table 1 describes the proposed taxonomy of lateness behavior to be
tested. It is assumed that a more accurate meaning of lateness can be objectively determined by examining the pattern, frequency, and duration of
employee-lateness incidents, as opposed to employee- or supervisor-cited
reasons. Borrowed from attribution theory (Jones & Nisbett, 1972), the
Table 1. A Taxonomy of Lateness Behavior

Pattern, frequency,
and duration
Relationship to
other behaviors
Antecedents of such
lateness
Increasing
chronic
Nonrandom,
increasing,
and increasing
Stable
periodic
Nonrandom, stable,
and stable
Low job satisfaction

Low job involvement
Low organizational
commitment
Leisure-income
Trade-off
Work-family conflict
Unavoidable
Random
Positive to voluntary
absence, voluntary
turnover, leaving
work early
Independent of voluntary
absence, voluntary
turnover, leaving work
early
Positive to nonwork
lateness
Category
Transportation
Bad weather
Personal illness or
accident
A Taxonomy of Lateness Behavior
135
causes for employee lateness assigned by the supervisor would tend to be

employee controlled (Le., avoidable), whereas attributions of the same
behavior made by the employee would tend to be circumstance controlled
(Le., unavoidable). Empirical support for employees attributing absence
behavior frequency more to unavoidable factors (e.g., illness) than controllable factors (e.g., difficulties getting up or disagreements with boss) is
provided by Payne and Nicholson (1987).
Within this lateness taxonomy, a nonrandom pattern means that the
individual exhibits a fairly predictable display of lateness behavior. This
display will generally be increasing or stable in frequency and duration.
A random pattern is unpredictable, so frequency and duration cannot be
classified. This taxonomy of lateness, including relationships to other
behaviors and antecedents, is largely derived from the cumulative findings of prior research. It is important to note that, in general, most previous studies of lateness (e.g., Adler & Golan, 1981; Angle & Perry, 1981;
Clegg, 1983; Farrell & Robb, 1980; Gupta & Jenkins, 1983; Macy & Mirvis,
1976; Nicholson & Goodge, 1976; Rosse & Hulin, 1985) focused on bluecollar types of jobs (e.g., telephone operators, bus transit personnel, manufacturing employees, printing-company employees, food processors,
and hospital employees). In lateness studies with more professional samples such as nurses (e.g., Blau, 1985b; Jamal, 1981), shift work was
required. Thus, the proposed lateness taxonomy in Table 1 is most applicable to either blue-collar or shift-related jobs.
This lateness taxonomy is not intended to predict that employees in
one type of lateness category will show more or less overall lateness (frequency and duration of incidents) than employees in another category. It
is possible for employees falling in any of the categories to have more
overall lateness than employees in the other two categories. Instead, categorizing individuals into types of lateness on the basis of pattern, frequency, and duration of incidents should improve understanding of the
antecedents related to each category, as well as the relationships between
lateness and other work behaviors. An explanation of each lateness
behavior category is appropriate.
INCREASING CHRONIC LATENESS
As shown above, increasing chronic lateness is characterized by a

nonrandom pattern of increasing frequency and duration. It is expected
that this type of lateness will show a positive relationship to the physical
withdrawal behaviors of voluntary absence, voluntary turnover, and
leaving work early. This can be part of a progressive withdrawal model
136
Gary Blau
(Herzberg, Mausner, Peterson, & Capwell, 1957) or spillover model

(Beehr & Gupta, 1978), for which there is some empirical support (e.g.,
Adler & Golan, 1981; Clegg, 1983; Rosse, 1988). Relevant antecedents to
such lateness are the distinguishable work attitudes (Brooke, Russell, &
Price, 1988; Mathieu & Farr, 1991) of job satisfaction, job involvement, and
organization commitment. Although the empirical evidence is mixed,
there is at least some support for these work attitudes being significantly
negatively related to lateness (e.g., Angle & Perry, 1981; Beehr & Gupta,
1978; Blau, 1986; Clegg, 1983). However, even where significant work
attitude-lateness relationships were found, the category of lateness being
measured was not clearly defined. Mixing different categories of lateness
behavior together will weaken relationships with the specific antecedents
of each lateness category.
STABLE PERIODIC LATENESS
Stable periodic lateness is characterized by a nonrandom pattern of

stable frequency and duration. This type of lateness is not expected to
show a significant relationship to voluntary absence, turnover, or leaving
early. It is not that employees exhibiting this type of lateness dislike their
work situations (as they do for increasing chronic lateness), but they have
"better or more important things to do." Leisure-income trade-off and
work-family conflict should be positive antecedents to this type of lateness. Leisure-income trade-off suggests that despite being docked for the
time they are late, some employees may consciously choose to be a relatively fixed amount of time late on a regular basis because of their preference for some "leisure-related" activity (e.g., reading the morning paper
or sleeping). Leigh and Lust (1988) found a positive relationship between
employee wages and lateness and suggested that the higher wage motivated the late worker to make a leisure-income trade-off.
Work-family conflict is a form of interrole conflict in which the role
pressures from work (e.g., on-time arrival) and family domains (e.g.,
dropping a child off at day care or school) are mutually incompatible
(Greenhaus & Beutell, 1985). Despite the lack of previous research,
work-family conflict is expected to be positively related to this type of
lateness behavior (Schultz & Henderson, 1985). For example, parentsespecially working mothers (Williams, Suls, Alliger, Learner, & Wan,
1991)-may place a higher priority on their children's needs. For such
employees, following Lobel's (1991) utilitarian argument for career versus
family role investment, the higher perceived level of family rewards over
career rewards results in greater family versus career investment. This
137
may be manifested by an employee systematically attending to a child's

needs and sacrificing on-time arrival to work.
UNAVOIDABLE LATENESS
Unavoidable lateness is characterized by a more random pattern of

behavior. As noted earlier, frequency and duration cannot be classified. To
help distinguish this category of lateness from the other categories, it is
expected that the randomness of this kind of lateness should show a positive relationship to the randomness of an employee's nonwork lateness
behavior (e.g., being late to a social event). This is based on the assumption that such lateness transcends work and also includes nonwork situations. Employees who exhibit unavoidable lateness do so not in response
to a bad work situation (increasing chronic) or because they have better or
more important things to do (stable periodic) but because of less controllable factors such as transportation concerns, bad weather, personal illness or accident. Such factors would be expected to lead to work (and
nonwork) lateness.
Gupta and Jenkins (1983) and Leigh and Lust (1988) found a significant positive relationship between employee-related transportation problems and lateness, whereas Farrell and Robb (1980) found a negative
relationship between employee-reported ease of travel to work and lateness. Motley (1926) found seasonal fluctuations in employee lateness, with
the winter months showing the highest lateness rates. Jamal (1981) found
a negative relationship between employee mental health and lateness.
On the basis of this literature review, the following hypotheses will
be tested:
H t : Increasing chronic lateness, which is nonrandom, of increasing
frequency and increasing duration, will show a stronger positive
relationship to voluntary absenteeism, voluntary turnover, and
leaving work early than stable periodic lateness, which is nonrandom, of stable frequency and duration.
H 2 : Unavoidable lateness, which is random, will be significantly positively related to nonwork lateness.
H3: Job satisfaction, organizational commitment, and job involvement will be significantly negatively related to increasing chronic
lateness.
H 4 : Leisure-income trade-off and work-family conflict will be significantly positively related to stable periodic lateness.
Hs: Transportation concerns, bad weather, and personal illness or accidents will be significantly positively related to unavoidable lateness.
138
Gary Blau
METHOD
Hospital Sample and Procedure
A questionnaire was constructed and administered in October 1990
on a voluntary basis to all nonexempt hospital employees for whom daily
attendance records were kept. This hospital, located in a large Eastern
city, offers patient care in many medical specialties. Sample respondents'
jobs included secretary, clerk, receptionist, lab technician, maintenance
worker, housekeeper, bill processor, and food service worker. Hospital
management was concerned about employee lateness and supported the
study. Union approval was also obtained. Subjects were given time at
work to fill out the survey. The hospital's human resource department
assisted with the survey administration across all departments and shifts.
Complete confidentiality of subject responses was maintained and participants mailed their survey directly to the investigator in a preaddressed
business reply envelope.
Of the 641 surveys distributed, 509 (79%) were voluntarily completed
and mailed to the investigator. Two follow-up letters helped to increase
subject participation. Of the 509 subjects, 483 (95%) were willing to give
the last four digits of their social security number so that subsequent lateness and other behavioral data could be collected. Eighteen months after
the survey (April 1992), record-based data (i.e., lateness, absence,
turnover, and leaving work early) were collected. Of these 483 subjects,
102 had no recorded lateness incidents over the 18-month period.
A demographic breakdown of the 381 subjects with lateness incidents
indicated that: the average age was 34 years; 71 % were women; 55% were
married; 47% were black, 39% were white, and 14% were either Hispanic
or Asian; 70% had at least a high school degree; 64% worked on the day
shift; average organizational tenure was about 11 years; and 42% had one
or more children living at horne. A comparison of these demographics to
the 102 subjects with no lateness incidents revealed that significantly less
"never late" participants (28%) had children living at home.
Bank Sample and Procedure
A questionnaire was constructed and administered in September
1990 on a voluntary basis to 844 nonexempt bank employees for whom
daily attendance records were kept. This nonunionized bank is headquartered in a large Eastern city. Sample respondents' jobs included teller,
data entry clerk, secretary, receptionist, mail clerk, check processor, and
maintenance worker. Bank management was concerned about employee
139
lateness and supported the study. Subjects were given time at work to fill
out the survey. Supervisory personnel assisted with the survey administration at each work site (e.g., branch or department). All 66 branches participated in this study. Complete confidentiality of subject responses was
maintained, and participants mailed their survey directly to the investigator in a preaddressed business reply envelope.
Of the 844 surveys distributed, 720 (85%) were voluntarily completed
and mailed to the investigator. Two follow-up letters helped to increase
subject participation. Of the 720 subjects, 619 (86%) were willing to give
the last four digits of their social security number so that subsequent lateness and other behavioral data could be collected. Eighteen months after
the survey (March 1992), record-based data (i.e., lateness, absence,
turnover, and leaving work early) were collected. Of these 619 subjects,
171 had no recorded lateness incidents over the 18-month period.
A demographic breakdown of the 448 subjects with lateness incidents
indicated that: the average age was 37 years; 63% were women; 61 % were
married; 58% were white, 25% were black, and 17% were either Hispanic or
Asian; 81 % had at least a high school degree; average organizational tenure
was about 9 years; and 40% had one or more children living at home.
A comparison of these demographics to the 171 subjects with no lateness
incidents indicated that the "never late" sample had significantly higher
organizational tenure (12 years) and less children (25%) living at home.
Following the same survey and record-based data collection procedures
across the hospital and bank samples will allow for stronger comparison.
Measures
Lateness. As noted in the Introduction, it is necessary to measure

the pattern, frequency, and duration of an employee's lateness behavior.
Each of these three components must first be operationalized. The same
decision rules in categorizing lateness incidents were used across both
samples to test the genera liz ability of this measurement approach.
Human resource personnel in the hospital and bank helped to develop
these decision rules. Each organization had an escalating disciplinary
policy for dealing with employee lateness. This disciplinary policy was
based on a I-year calendar-based time frame.
Assuming that an employee was not suspended after 1 year, his or
her lateness slate was supposedly "wiped clean" and he or she started
over. If an employee was suspended, he or she was placed on a probation
schedule for the following year. Both the hospital and bank human
resource personnel were also concerned about the lack of supervisor
diligence in recording the frequency and duration of employee lateness
140
Gary Blau
incidents. Part of the reason for both organizations approving this study
was to increase supervisors' awareness about employee lateness, including the application of organizational policy for dealing with lateness.
Rosse and Hulin (1985) found a similar negligent supervisory attitude in
recording employee lateness incidents and applying company policy.
To help overcome this lack of diligence in recording lateness and to
allow a stronger application of the pattern, frequency, and duration decision rules for determining the type of lateness behavior an employee
exhibited, bank and hospital personnel agreed that a longer (i.e., 18
month) time frame should be used. For example, after 12 months when
employee records were wiped clean, both organizations wanted to know
if (and by how much) employee lateness behavior increased. Analysis of
12-month versus 18-month lateness behavior for 100 randomly selected
participants across both organizations indicated a significant increase in
overall lateness behavior. This suggests that some employees may "work
the system."
For the hospital, lateness was defined as when an employee reported
to work at least 8 min past the beginning of his or her shift. Over a I-year
period, a certain number of lateness incidents led to the following penalties: 3 incidents-general counseling; 5 incidents-verbal warning
with documentation; 9 incidents-written warning; 13 incidents-3
working days' pay suspension; 15 incidents-5 working days' pay suspension; and 16 incidents-job termination. The hospital's policy also
stated that employees would be docked wages in IS-min increments
beginning 8 min after start time as follows: 8 to 15 min late-IS min
docked; 16 to 30min late-30min docked; 31 to 45min late-45min
docked; and so on.
For the bank, lateness was defined as when an individual reported
to work at least 5 min past the beginning of his or her shift. The bank
used the following incident to penalty schedule over a 12-month period:
4 incidents-general counseling; 6 incidents-verbal warning with
documentation; 8 incidents-written warning; 11 incidents-5 working
days' pay suspension; and 14 incidents-job termination. The bank's
policy also stated that employees would be docked wages in 30-min
increments as follows: 5 to 30 min late-3D min docked; 31 to 60 min late60min docked; and so on. Thus, the bank's lateness policy was more
punitive.
Employees were initially classified as having either a nonrandom or a
random lateness pattern. Nonrandom was defined as when at least twothirds (>66%) of an individual's lateness incidents occurred at the beginning of a workweek (usually Monday) or after a legal holiday. Random
was defined as when less than one-third 33%) of an individual's
141
lateness incidents occurred at the beginning of a workweek (usually

Monday) or after a legal holiday. Supervisors generally indicated that
lateness was most often a problem at the beginning of a workweek or just
after a legal holiday. Of the 30 supervisors interviewed across both organizations, no one mentioned employees having a midweek or end-of-theweek lateness problem. An analysis of 100 randomly selected subjects
across both organizations did not support a midweek or other nonrandom lateness pattern. Once an individual was placed in the nonrandom
category, frequency and duration were operationalized. Frequency was
either increasing or stable, and duration was either increasing or stable.
Increasing frequency was defined as when the number of lateness
incidents per month increased across at least 7 of the 18 months. Stable
frequency was defined as when the number of lateness incidents per
month increased 3 or less of the 18 months. These frequency and duration
cutoffs were derived from conversations with the 30 interviewed supervisors mentioned above and human resource personnel. Hospital and
bank human resource personnel indicated that typically after an
employee was disciplined for excessive lateness, his or her behavior
would at least temporarily improve. However, with some employees this
improvement began only with stronger sanctions and even then such
improvement was transitory (Le., 1 or 2 months) before the employee
would regress back until the next penalty was implemented. Paralleling
the frequency cutoffs, increasing duration was defined as when the number of minutes late per month increased across at least 7 of the 18 months.
Stable duration was defined as when the number of minutes late per
month increased 3 or less of the 18 months.
Increasing and stable were defined on a month-by-month basis by
comparing frequency and duration to the previous month. For example,
if a participant exhibited the following 18-month frequency of lateness
incidents per month: (1) 1, (2) 2, (3) 0, (4) 1, (5) 2, (6) 1, (7) 0, (8) 1, (9) 2, (10)
0, (11) 2, (12) 2, (13) 0, (14) 0, (15) 1, (16) 2, (17) 0, (18) 0, he or she would
fall into the increasing frequency category because lateness incidents
increased eight times (i.e., 1 to 2, 3 to 4, 4 to 5,7 to 8, 8 to 9, 10 to 11, 14 to
15, and 15 to 16).
As noted in the above coding scheme, the middle one-third was
deliberately left out in measuring random 33%) versus nonrandom
(>66%) pattern, increasing (at least 7 out of 18 months) versus stable (3 or
less out of 18 months) frequency, and increasing (at least 7 out of 18
months) versus stable (3 or less out of 18 months) duration. This was
designed to have more confidence that individuals falling into the top
versus bottom third of each measure truly exhibited the measured
characteristic. Personality research has used a similar approach (Le., using
142
Gary Blau
top-third versus bottom-third subjects on a personality measure) to represent "truer" personality distinctions (Blass, 1977).
This approach also acknowledges the dynamic nature of lateness for
individuals (i.e., that employees can display more than one category of
lateness in a given period). The middle-third group of subjects on pattern,
frequency, and duration should best represent employees with varying
types of lateness. This dynamic lateness category group is defined by onethird to two-thirds of the lateness incidents occurring at the beginning of
a workweek or after a legal holiday and where frequency and duration
increased 4 to 6 of the 18 months. This dynamic lateness group will be
compared to the three previously defined lateness category groups:
increasing chronic, stable periodic, and unavoidable. Because the
dynamic lateness category represents a mix of the other three lateness
behaviors (increasing chronic, stable periodic, and unavoidable), it is not
possible to a priori hypothesize relationships of dynamic lateness to other
work behaviors or antecedents. Therefore, dynamic lateness was not
included in Table 1.
Using these decision rules, five raters coded the lateness behaviors of
the study participants in each sample. For the hospital sample, 353 out of
381 (93%) subjects clearly fell into one of the four categories as follows:
increasing chronic lateness (n = 59); stable periodic lateness (n = 70);
unavoidable lateness (n = 104); and dynamic lateness (n = 120). The
remaining 28 subjects met only one of the frequency or duration cutoffs
and were excluded from further analysis. For the bank sample, 402 out of
448 (90%) subjects clearly fell into one of the four categories as follows:
increasing chronic lateness (n = 66); stable periodic work lateness (n = 79);
unavoidable lateness (n = 125); and dynamic lateness (n = 132). The
remaining 46 subjects met only one of the frequency or duration cutoffs
and were excluded from further analysis.
To check rater reliability, all five raters coded a common group of 50
randomly selected participants into the four latensss categories. Interrater
reliability was assessed using coefficient kappa, which estimates the
agreement between pairs of raters while correcting for chance agreement
(Cohen, 1960). With five raters there are 10 between-rater comparisons.
The average kappa coefficient for these comparisons was .87, and the
range was .97 to .81. Given the four types of lateness, three dummy coded
variables were created for later canonical analysis.
Voluntary Absenteeism. From hospital and bank records, frequency

counts for the number of voluntary (unexcused) absences for subjects
over the same 18-month period (after the survey) as the lateness data
were recorded. Unexcused absence was commonly defined across the
143
hospital and bank as absence without permission. All other behavioral

data listed below (i.e., turnover and leaving work early) were collected for
the same 18-month period (after the survey) as the lateness incidents.
Collecting low base-rate behaviors such as absence and leaving early over
a longer period of time can reduce data-distribution problems (e.g., skewness and kurtosis), but at a cost of attenuating the relationships between
antecedents and such behaviors (Harrison & Hulin, 1989). However,
because the main purpose of this study was to explore the lateness behavior construct, with other withdrawal behaviors being secondary, such
a trade-off was felt to be acceptable.
Turnover. From hospital and bank records, frequency counts for the
number of subjects who voluntarily left their organization after 18 months
were recorded (1 = stay and 2 = leave). For the hospital, there were 247
(70%) stayers and 106 (30%) leavers. Of these 106leavers, exit interviews
indicated that 23 said that they left for "unavoidable" (Abelson, 1987)
non-job-related reasons (e.g., family related or medical). For the bank,
there were 288 (72%) stayers and 114 (28%) leavers. Of these 114 leavers,
exit interviews revealed 29 "unavoidable" leavers. On the basis of
Abelson's (1987) research, these unavoidable leavers were deleted from
each sample when analyzing the turnover data.
Leaving Work Early. Because daily attendance records were kept
for nonexempt employees at the bank and hospital, the number of times
a subject left work early without permission was tabulated.
Nonwork Lateness. In the survey, subjects were asked to retrospectively indicate the number of times they were late over the last 18 months
in each of the following nonwork categories: (a) religious (e.g., mass or
wedding); (b) volunteer related (e.g., PTA or charity work); (c)
family-social (e.g., picnic or party); and (d) community related (e.g., town
meeting, athletic event, garden club, or political event). The following
response scale was used: 1 = never (0 times); 2 = rarely (1 or 2 times);
3 = occasionally (3 to 5 times); 4 = frequently (6 to 9 times); and 5 = very frequently (at least 10 times). A frequency range is given within each verbal
anchor under the assumption that subjects are more likely to correctly
recall an approximate (versus exact) number of times they were late to an
event category. Across both samples 40 spouses were willing to answer
(after getting subject permission) this three-item nonwork lateness measure about their participating spouse. The resulting correlation of .69
between subjects and spouses suggests that the participants' retrospective
nonwork lateness self-reports are reliable.
144
Gary Blau
Work Attitudes. Job satisfaction was measured using the 20-item

short-form version of the Minnesota Satisfaction Questionnaire (Weiss,
Dawis, England, & Lofquist, 1967). Job involvement was measured with 9
items from Kanungo's lO-item scale. Blau (1985a) recommended that 1
item from Kanungo's (1982) measure be dropped because of its poor factor loading. Organizational commitment was measured using Meyer and
Allen's (1984) 8-item affective scale. Recent empirical work (e.g., Dunham
& Grube, 1990) suggests that Meyer and Allen's (1984) affective scale is
superior to the more popular Porter, Crampon, and Smith (1976) scale. A
6-point response scale (1 = strongly disagree, 6 = strongly agree) was used
for these attitude measures.
Leisure-Income Trade-off. On the basic of the work of Leigh and Lust
(1988) and Youngblood (1984), a three-item measure of leisure-income
trade-off was used. The items were: "my personal time is more important
than the money I lose by being late for work," "I am willing to trade-off
my pay for coming to work when I want to," and "given my salary, I can
afford to lose money by coming to work when I wish." A 6-point response
scale was used (1 = strongly disagree, 6 = strongly agree).
Work-Family Conflict. Work-family conflict was measured with
Kopelman, Greenhaus, and Connolly's (1983) 8-item measure. A sample
item is: "my work schedule often conflicts with my family life." A 6-point
response scale was used (1 = strongly disagree, 6 = strongly agree).
Transportation Concerns. On the basis of the prior work of Farrell
and Robb (1980) and Gupta and Jenkins (1983), a three-item measure was
developed. The three items are: (1) "How many miles do you travel to get
to work?" (2) "How many minutes does it generally take you to get from
home to work?" (3) "Considering such things as traffic, weather, and
availability of public transportation, rate the difficulty of getting to
work." The response scales for each item were as follows: (1) 1 = 0-6 miles,
2 = 7-12 miles, 3 = 13-18 miles, 4 = 19-24 miles, 5 = 25-29 miles, and
6 = more than 30 miles; (2) 1 = 0-15 minutes, 2 = 16-25 minutes, 3 = 26-35
minutes; 4 = 36-45 minutes, 5 = 46-60 minutes, and 6 = more than 60 minutes;
(3) 1 = very easy, 2 = easy, 3 = slightly easy, 4 = slightly difficult, 5 = difficult,
and 6 = very difficult.
Bad Weather. Weather was measured by dummy coding the winter or

"bad weather" months (November through mid-March) versus the other
months on the basis of Motley's (1926) study. Two cycles of bad-weather
months (1990-1991 and 1991-1992) were collected over the 18-month time
145
frame. Mean levels of unavoidable lateness were compared for the badweather versus other months.
Personal Illness-Accident. A three-item retrospective scale was

developed for the survey. The items were: "Over the last 18 months, how
many times have you been involved in some type of traffic accident on the
way to work?" "Over the last 18 months, how many times did feeling ill
(e.g., cold, headache, stomach upset, or allergy) delay your arrival to
work?" "Over the last 18 months, how many times did some kind of
unusual event (e.g., flat tire or slow alarm clock) delay your arrival to
work?" The following response scale was used: 1 = never (0 times);
2 = rarely (1 or 2 times); 3 = occasionally (3 to 5 times); 4 = frequently (6 to 9
times); and 5 = very frequently (at least 10 times). The answers to this scale
by the same 40 spouses mentioned earlier had a correlation of .38 to their
participating spouses' answers. Thus, the retrospective nature of this scale
is more problematic.
RESULTS
Canonical analysis was initially done to examine the overall relationship of the dummy-coded lateness behavior categories to the antecedents.
Table 2 presents the correlations of study variables with canonical variates
broken down by sample. With the Bartlett residual test (Barcikowski &
Stevens, 1975), three significant canonical correlations (R)-representing
uncorrelated antecedent variable-lateness behavior composites-were
found for each sample. The pattern of canonical variate-variable correlations indicates that lower job satisfaction, job involvement, and organizational commitment lead to increasing chronic lateness; higher
leisure-income trade-off and work-family conflict are related to more stable periodic lateness; and transportation concerns, and to a lesser extent
personal illness or accidents, lead to unavoidable lateness. Overall, these
results support the proposed lateness taxonomy. Given the N to K (subjects to variables) ratio for both samples, these canonical correlations
should be stable (Barcikowski & Stevens, 1975).
Means, standard deviations, reliabilities, and correlations among
study variables are shown for the hospital sample and the bank sample
(see Table 3). As noted earlier, 353 of the 381 hospital employees and 402
of the 448 bank employees could be coded into one of the four lateness
categories. The lower variable means on the work attitudes for the bank
sample suggest that these employees are somewhat less committed to
their work than the hospital employees. The self-report scales, in general,
146
Gary Blau
Table 2. Correlations of Study Variables with Canonical Variates Broken

Down by Sample
Correlation-canonical
variate pairs
Antecedent variables
Job satisfaction
Job involvement
Organizational
commitment
Leisure-income trade-off
Work-family conflict
Transportation concerns
Bad weather
Personal illness / accident
Lateness behaviors
Increasing chronic
Stable periodic
Unavoidable
R
Hospital sample
First
Bank sample
Second
Third
First
Second
Third
-.51
-.46
-.43
-.21
-.13
-.18
-.14
-.11
-.15
-.53
-.44
-.45
-.17
-.20
-.22
-.05
.18
.20
.34
.37
.15
.08
.08
.19
.39
.22
.27
.10
.12
.17
.06
.05
.40
.38
.13
-.01
.04
.16
.21
.32
.23
.35
.15
.21
.47
.49
.27
.23
.52
.18
.19
.17
.50
.11
-.04
.09
.45
.24
.08
R! = .43**
.10
.26
.40
.13
R2 = .35*
R3 = .30*
.11
-.11
-.09
R! = .48** R2 = .38* R3= .32*
Note . Canonical variate-variable correlations >.30 are in boldface type.
p< .05. p<.01.
exhibit at least adequate (>.70) coefficient alphas (Nunnally, 1978), with

the exception of the personal illness or accident scale. The items making
up this scale were heterogeneous. The correlations between study variables are similar across the two samples.
Significant correlations were found between antecedents to different
categories of lateness. For example, job attitudes had significant negative
relationships to leisure-income trade-off and work-family conflict.
However, the maximum overlap between these variables was 6% or r =
(- .25)2. Such results are consistent with prior research (e.g., Thompson &
Blau, 1993; Youngblood, 1984).
As noted in the Method section, subjects were classified into one of
the four lateness categories. Because the proposed taxonomy of lateness
uses both frequency and duration information, overall lateness scores
were calculated within each category by adding the sum of incident frequency by duration for each participant. This created overall "minutes
late" measures in Table 3. Separate correlations between lateness frequency and duration measures were strong enough (average correlation
was .80 across lateness categories and samples) to support combining the
frequency and duration data. Adler and Golan (1981) also found very
147
strong relationships between lateness frequency and duration measures.

Overall lateness measures will enhance the comparability of this study's
results to previous lateness research (e.g., Adler & Golan, 1981; Gupta &
Jenkins, 1983).
Tests of Hypotheses
Hypothesis 1 (increasing chronic lateness will have a stronger positive relationship to voluntary absence, voluntary turnover, and leaving
work early than stable periodic lateness) was supported across both samples for voluntary absence and leaving early but not for turnover. As
shown in Table 3, significant differences in correlations (McNemar, 1969)
were found (p < .05, one-tailed) for these relationships.
Hypothesis 2 (a significant positive relationship will be found
between unavoidable lateness and nonwork lateness across both samples)
was supported. Hypothesis 3 (significant negative relationships will be
found between job satisfaction, job involvement, and organizational commitment to increasing chronic lateness across both samples) was supported. Hypothesis 4 (significant positive relationships will be found
between leisure-income trade-off and work-family conflict to stable periodic lateness for both samples) was supported. Hypothesis 5 (a significant
positive relationship between transportation concerns and unavoidable
lateness will be found) was supported for both samples. Personal accident
or illness was positively related to unavoidable lateness for the bank sample but not the hospital sample. Correlation results with only the first 12
months of lateness data were consistent with the results cited above.
To test the relationship between bad weather and unavoidable lateness, a t test was done between winter months (i.e., November through
mid-March) versus the other months to see if unavoidable lateness was
higher during the winter months. The results did not support this
hypothesis. The mean minutes level of unavoidable lateness during winter months was 219.4 versus 217.7 for the other months (t = 1.18, P > .05)
for the hospital sample and 219.1 versus 215.5 (t = 1.47, P > .05) for the
bank sample.
Tables 4 and 5 show the mean levels of study variables for both samples, broken down by lateness category. Consistent with prior lateness
research, the never late group is also shown. An analysis of variance was
used to test for significant overall mean differences on each variable; this
was followed by the Duncan multiple range test to examine the differences between pairs of means at the .01 level of significance. The Duncan
multiple range test is favored over the least significant difference test
(Nie, Hull, Jenkins, Steinbrenner, & Bent, 1975) because it uses a different
148
Gary Blau
Table 3. Means, Standard Deviations, Reliabilities, and Correlations
Hospital sample
Bank sample
M
SD
.86
.85
73.8
28.7
11.9
6.8
.87
.88
7.5
3.1
.84
.73
33.6
10.8
7.7
3.3
.77
.75
.38-.20"
.41"
-.24"
-.16-
24.3
4.6
.81
25.7
4.9
.80
-.25"
-.17"
-.19"
.14"
9.6
2.4
.70
8.3
2.1
.69
-.07
-.03
0.2
.11"
9.1
2.2
.61
8.2
2.0
.65
.04
.01
-.04
.05
255.3'
68.5
NA
231.8'
64.2
NA
-.39"
-.41-
-.38"
.15
217.4'
71.7
NA
212.1'
63.4
NA
-.15
-.17
-.19
.35-
218.6'
73.3
NA
217.3'
67.1
NA
-.11
-.12
-.10
.16
230.4'
85.8
NA
222.5'
88.3
NA
-.19-
-.22-
-.18-
.17
8.3
3.2
NA
7.1
2.5
NA
-.24-
-.20-
-.17"
.19"
1.3'
13.5
12.1
0.6
4.6
4.3
NA
NA
.75
1.3'
12.1
11.6
0.6
3.8
4.2
NA
NA
.72
-.22-.25-.06
-.23-.21-.05
-.14"
-.16"
.02
.10
.15"
.12-
Variable
SD
1. Job satisfaction
2. Organizational
commitment
3. Job involvement
4. Leisure-income
trade-off
5. Work-family
conflict
6. Transportation
concerns
7. Personal illness
or accident
8. Increase chronic
lateness'
9. Stable periodic
lateness'
10. Unavoidable
lateness
11. Dynamic
lateness'
12. Voluntary
absence
13. Turnover
14. Leaving early
15. Nonwork
lateness
75.6
30.9
12.4
7.2
35.1
12.2
'"
'"
.41"
.39-
3
.36"
.39"
4
-.18"
-.16-.13"
Note . Correlations for the hospital sample are below the diagonal, and correlations for the bank sample
Reported in minutes based on ~(Frequency " Duration).
'N = 59 for increase chronic lateness, N = 70 for stable periodic lateness, N = 104 for unavoidable
'N = 66 for increase chronic lateness, N = 79 for stable periodic lateness, N = 125 for unavoidable
dNA = not applicable because there were separate groups of subjects within each lateness category.
e Significant difference in correlations, p < .05, one-tailed test.
" p<.05.
range value for different size subgroups. The results in Tables 4 and 5 provide further support for the study hypotheses. Increasing chronic late
subjects exhibited significantly higher voluntary absence and leaving
early behavior and had significantly lower job satisfaction, organizational
commitment, and job involvement than subjects in the other lateness
categories. Stable periodic late subjects had significantly higher
leisure-income trade-off and work-family conflict, whereas unavoidably
late subjects had a significantly higher level of nonwork late behavior and
transportation concerns (hospital sample only).
Perhaps most interesting is the comparison of mean level variables
between the never late versus lateness categories. The results across both
149
among Study Variables for Bank and Hospital Samples
10
11
12
-.17
-.18
-.08
-.13
-.19*
-.18*
-.14
.32*
-.10
.16
.10
.31*
.07
.11
-.14*
-.15*
-.05
-.07
-.06
-.02
-.45*
-.38*
-.11*
.10*
-.01
.08
-.05
.01
-.35*
-.14
.12
.09
.14
.09
13
14
-.23*
-.17*
-.26*
-.20*
-.24*
-.16*
.03
.04
-.20*
.15
-.15*
.13*
-.17*
.07
-.19*
.18*
.02
.09
.12
.19*
.11*
.14*
.15*
-.03
.13
.27*
.08
.04
-.03
.11*
.05
.07
.29*
.11
.09
.05
.10*
.12*
NA d
NA d
NAd
.42*'
.19
.38*'
.01
NA d
NAd
.16'
.13
.14'
.04
NAd
.12
.08
.06
.34*
.20*
.17
.16
.13
.31 *
.29*
.10*
.07
.15*
.13
.08
.09
.28*
.12
.07
NAd
.14
.25*
.18
NAd
NAd
.20*
.12
.09
NA d
NA d
NA d
.18*
.08
.11*
.39*'
.12e
.10
.20*
.11 *
.24*
.09
.03
.13*
.06
-.02
.10*
.07
.20
.43*'
.03
.14
.13e
.08
.04
-.01
.32*
.10
.18*
.15*
.30*
.26*
.14*
.15*
.10*
.07
15
.02
-.11*
-.09
are above the diagonal.

lateness, N
lateness, N
= 120 for dynamic lateness, N = 330 for turnover, other N = 353.

= 132 for dynamic lateness, N = 373 for turnover, other N = 402.
samples show that the never late participants had significantly higher job
satisfaction, organizational commitment, and job involvement and significantly lower leisure-income trade-off, work-family conflict, personal
illness or accident, voluntary absence, and leaving work early. These
patterns of results are supportive of the lateness taxonomy and reinforce
the importance of antecedent and work behavior variables for distinguishing late versus never late behavior.
The demographic difference of never late subjects being less likely to
have dependent children at horne would help to explain some of these
results, including the lower levels of work-family conflict, leisure-income
trade-off, and leaving work early. Both samples in this study consisted
Gary Blau
150
Table 4. Means and Standard Deviations of Study Variables by Lateness

Category for Hospital Sample
Variable
Job satisfaction
M
SO
Organizational
commitment
M
SO
Job involvement
M
SO
Leisure-income
trade-off
M
SO
Work-family
conflict
M
SO
Transportation
concerns
M
SO
Personal illness or
accident
M
SD
Voluntary
absence
M
SO
Turnover
M
SO
Leaving early
M
SO
Nonwork late
M
SO
Chronic
Periodic
(N=59)
(N= 70)
Unavoidable
(N= 104)
Dynamic
Never late
(N= 120)
(N=102)
67.3_
11.9
78.4 e
12.9
76.1 e
12.4
81.2d
13.1
5.23**
12.3
25.4_
6.6
30.2b
7.2
32.3b
7.1
30.7b
7.3
35.1e
7.4
3.86*
29.8_
7.1
34.~
35.7b
7.6
36.2b
7.8
41.6e
7.5
4.11*
11.5_
3.1
14.~
12.3_
3.0
11.9_
2.8
7.4e
3.1
3.22*
3.3
22.7_
4.8
29.2b
4.9
24.1_
4.7
23.8_
4.4
20.1e
4.3
4.37**
9.1_
2.2
8.9_
2.3
11.5b
2.6
9.4.
2.5
8.3.
2.4
2.15*
8.8.
2.3
8.7.
2.2
9.3.
2.1
9.1.
2.2
3.~
4.68**
2.3
11.5.
3.6
8.1b
3.0
7.9b
3.2
5.1 e
2.0
6.02**
3.1
1.7
0.6
1.4
0.5
1.2
0.5
1.1
0.4
1.37
0.4
18.5.
4.9
15.9b
4.8
12.5e
4.4
12.3e
4.5
7.3d
3.1
9.22**
10.3.
4.5
11.6.
4.2
13.9b
4.4
12.0_
4.3
9.5_
2.7
3.53*
73.~
7.2
1.1
7.~
Note. Across each row, cell means that do not share the same subscript are significantly
different at the .01 level by Duncan's multiple range test.
p< .05. p<.01.
151
Table 5. Means and Standard Deviations of Study Variables by Lateness

Category for Bank Sample
Variable
Chronic
(N= 66)
Periodic
(N= 79)
Unavoidable
(N = 125)
Dynamic
(N= 132)
Never late
(N = 171)
67.1 a
12.0
72.9 b
12.1
74.1b
11.8
74.3 b
11.9
80.3 c
4.77**
25.9a
6.5
29.2b
6.7
30.6b
7.0
28.4b
6.9
34.0 c
30.3a
7.9
33.1b
7.8
34.5 b
7.7
33.9 b
7.5
37.7c
9.8a
3.6
12.6b
3.5
9.9a
3.2
10.6a
3.3
6.7c
25.7a
5.0
28.3 b
5.1
24.2 c
24.5 c
20.2c
4.6
4.8
4.4
8.0a
2.2
7.8a
2.1
9.2a
2.2
8.1 a
2.0
6.8b
1.9
2.29*
8.1 a
2.0
7.3a
1.9
8.2a
2.1
8.4 a
1.9
3.8b
1.4
4.46*
9.5a
2.7
7.0b
2.6
6.6 b
2.4
6.7b
2.5
4.9 c
6.35**
1.6
0.6
1.3
0.5
1.1
0.4
1.2
0.5
1.1
0.4
1.09
15.5a
4.0
12.3b
3.9
11.2b
3.7
11.5b
3.8
6.9c
8.97**
1O.5a
4.1
11.4a
4.2
13. 7b
4.3
10.9a
4.2
9.8a
3.7
Job satisfaction
M
SD
10.8
Organizational
commitment
M
SD
4.06**
7.1
Job involvement
M
SD
4.50**
8.0
Leisure-income
trade-off
M
SD
5.51**
2.6
Work-family
conflict
M
SD
5.13**
Transportation
concern
M
SD
Personal illness or
accident
M
SD
Voluntary absence
M
SD
1.8
Turnover
M
SD
Leaving early
M
SD
2.6
Nonwork late
M
SD
3.39*
Note. Across each row, cell means that do not share the same subscript are significantly
different at the .01 level by Duncan's multiple range test.
p< .05. p<.OJ.
152
Gary Blau
primarily of women. Using a sample of primarily female employees,

Thompson and Blau (1993) found a significant positive relationship
between number of dependent children at home and perceived conflict
between family and work. Over 20 years ago, Hall (1972) noted that
work-family conflict is higher for employed women versus men because
women are more likely to deal with work and family roles simultaneously
(i.e., juggle work and family demands) rather than sequentially. Duxbury
and Higgins (1991) argued that gender differences in antecedents and
consequences of work-family conflict were largely attributable to societal
expectations and behavioral norms. These authors suggested that a more
equal distribution of family roles to match increased responsibilities
outside the home was needed.
DISCUSSION
The results of this study empirically demonstrate, across two samples, that three distinct categories of lateness behavior can be operationalized by pattern, frequency, and duration. By isolating such lateness
categories, it was argued earlier that stronger antecedent-lateness relationships could be found than in previous research that did not break
down general lateness behavior into specific categories. For example, job
involvement accounted for 14% (hospital) and 12% (bank) of increasing
chronic lateness behavior versus other studies (e.g., Beehr & Gupta, 1978;
Cummings & Manring, 1977) where 1% to 8% of lateness has been
accounted for Organizational commitment accounted for 17% (hospital)
and 14% (bank) of increasing chronic lateness behavior versus other studies (e.g., Blau, 1986; Clegg, 1983) where 4% to 7% of lateness has been
accounted for Transportation concerns accounted for approximately 7%
of unavoidable lateness behavior in both samples versus 2% to 5% of lateness in other studies (e.g., Farrell & Robb, 1980; Gupta & Jenkins, 1983;
Leigh & Lust, 1988).
Furthermore, although the antecedents used were generally related
to specific lateness behavior categories as hypothesized, such antecedents
were generally not significantly related to other categories of lateness.
This lack of statistically significant relationships when crossing
antecedents to other lateness categories provides additional support for
the suggested lateness behavior taxonomy. Also, a dynamic lateness
behavior category was created that represented subjects with a mixture of
increasing chronic, stable periodic, and unavoidable lateness. The correlational results found between antecedent variables and work behaviors to
dynamic lateness are quite similar in magnitude to the results found by
153
previous research (e.g., Blau, 1986; Clegg, 1983; Cummings & Manring,
1977; Gupta & Jenkins, 1983). This suggests that previous lateness studies
may have measured lateness with some combination of different types of
lateness behaviors.
Various limitations of this study need to be noted. The use of retrospective nonwork late and personal illness or accident measures, especially over a long time period, is problematic. For example, there is a
longer time frame (i.e., 3 years) between the retrospective non work lateness and personal illness measures to the behavioral data versus the 18month time frame between the attitudinal and perceptual measures to the
behavioral data. Having a greater time frame between variables would be
expected to weaken their relationships (e.g., Mobley et al., 1979). Also,
asking subjects to remember if their arrival to work was delayed by an
unusual event over a retrospective 18-month time frame could pick up
antecedents (e.g., work-family conflict), which may spill over into different lateness categories. Impression management (Paulhus, 1986), or trying to look good to others, could further distort subject responses (e.g.,
more likely to report feeling ill than having a slow alarm clock).
Future researchers could consider collecting such self-report data
over shorter time periods or having subjects use a diary or log to help
them more accurately remember incidents. Similarly, finding the selfreport measure of transportation concerns, but not the objective bad
weather measure, to be significantly related to unavoidable lateness reinforces the need to take precautions against retrospective justification bias
affecting results. Certainly other methodological concerns (e.g., range
restriction or lower measure reliability) may also affect the amount of
lateness category variance explained.
It is also important to acknowledge that, despite the inductive support within both organizations studied for developing random versus
nonrandom patterns and stable versus increasing frequency and duration
classifications, the decision rules and cutoff values applied were organization specific. In other organizations, the decision rules and cutoff values
to create such lateness behavior categories may well be different.
However, the key issue seems to be that some type of decision rule-cutoff
value scheme consistently applied to frequency and duration lateness
data can result in identifying meaningful and useful categories of
employee lateness behavior.
Future research might well expand the number of antecedents used
in the proposed lateness taxonomy. For example, time management skills
or procrastination (Burka & Yuen, 1983) may be a useful individual difference variable to measure, particularly for subjects in the unavoidable
lateness category. By "cutting things too close" in planning to be on-time
154
Gary Blau
for work, an individual is more susceptible to any type of less controllable

antecedent (e.g., traffic or weather) causing lateness. Richard and Slane
(1990) found that punctuality style was a persistent personality characteristic across work and nonwork situations. Self-efficacy may be another
individual difference worthy of future investigation. Frayne and Latham
(1987) found that increasing employee-perceived self-efficacy through
training led to better subsequent job attendance. Employees with low selfefficacy are more likely to perceive that they cannot cope with environmental demands. Such environmental demands could include not only
bad weather and transportation problems but also work-family conflict
situations.
What are the implications of this study's results for organizations?
The positive links of increasing chronic lateness to voluntary absence and
leaving work early with low job involvement, job satisfaction, and organizational commitment as antecedents support a progressive withdrawal
pattern from work (Herzberg et al., 1957; Rosse, 1988). For example,
with conditional probability analysis, Rosse (1988) found support for a
lateness-to-absence progression. It is important for supervisors to monitor
and recognize changes in the pattern, frequency, and duration of their
employees' lateness and to prevent this progression. By talking to the
employee and understanding the cause of his or her work-related unhappiness, a supervisor may be able to take remedial steps (e.g., job rotation,
lateral transfer, or active counseling; Kennedy, 1984). The stable periodic
lateness pattern may be more difficult to break. A flextime schedule
(Nollen & Martin, 1978), if feasible, may help an employee deal with
a work-family conflict or leisure-income trade-off concern.
Even for employees falling into the unavoidable lateness category,
managers should investigate if the pattern, frequency, and duration of
incidents are truly random. For example, is the employee always late for
the same reason (e.g., heavy traffic or bad weather)? Are there other
employees with approximately the same commute (e.g., distance or same
route) who show similar lateness behavior? Discussing the employee's
commute and encouraging him or her to do more advance planning,
including steps to allow more time to get into work, may be helpful.
Stricter enforcement of the company's lateness disciplinary policy by
supervisors (Rosse & Hulin, 1985), as well as further toughening of this
policy, may also be necessary. This is particularly true for employees who
"work the system." In a study of absences among 60 blue-collar employees, Morgan and Herman (1976) found that for some subjects absenteeism
provided an opportunity to experience consequences encouraging
absence that were not offset by organizational consequences for deterring
absence. Besides a stronger "stick," having a "carrot" to encourage
155
employees to be prompt everyday may be useful. Hermann, deMontes,

Dominquez, Montes, and Hopkins (1973) found that daily small cash
bonuses for promptness over a 49-week treatment period led to a strong
reduction in tardiness of six frequently late hourly employees.
What are some research implications of this study? As noted in the
Introduction, recent research by Hanisch and Hulin (1990, 1991) has suggested that specific withdrawal behaviors such as lateness and absenteeism can be aggregated into a more general work withdrawal construct.
This study'S position is that the creation of such a construct may be premature. Hanisch and Hulin (1990, 1991) found in both studies, using academic and nonacademic subjects from a large Midwestern university, that
lateness, absence, and unfavorable job behaviors (i.e., making excuses to
get out of work) loaded strongly onto a general factor that they labeled
work withdrawal. Hanisch and Hulin (1991: 111) defined work withdrawal as "behaviors dissatisfied individuals use to avoid aspects of their
specific work role or minimize time spent on their specific work tasks
while maintaining their current organization and work-role memberships." However, in neither study did Hanisch and Hulin (1990, 1991)
discuss what the university's disciplinary policy was for dealing with
lateness, absence, or undesirable behaviors. Most organizations, including the bank and hospital used here, have escalating disciplinary policies
that culminate in terminating the employee. The impact of such disciplinary policies on a general work withdrawal construct needs to be
addressed.
In addition, the self-report methodology used by Hanisch and Hulin
(1990,1991) to collect the lateness, absence, and undesirable job behaviors
suggests that method variance may be contributing to the positive relationships between these behaviors (Schwab, 1980). Relying on supervisor
or interviewer ratings to examine relationships between lateness, absence,
and other withdrawal behaviors (e.g., Beehr & Gupta, 1978; Gupta &
Jenkins, 1983; Rosse, 1988; Rosse & Hulin, 1985) can lead to problems,
such as measurement error because of uneven recording of behaviors
(Rosse & Hulin, 1985) or attribution errors (Payne & Nicholson, 1987). The
rating-based absence and leaving work early measures collected in this
study suffer from these limitations.
Furthermore, in studies testing the links between withdrawal behaviors, general or overall measures of these behaviors were most often used
(e.g., Gupta & Jenkins, 1983; Hanisch & Hulin, 1990, 1991; Rosse, 1988;
Rosse & Hulin, 1985). Such general measures may hide or mask relationships between more conceptually similar specific types of withdrawal
behaviors. For example, the relationships between unavoidable lateness
and unavoidable absence or increasing chronic lateness and avoidable
156
Gary Blau
absence may be stronger than the link between general lateness and general absence. Unfortunately, testing more specific relationships between
withdrawal behaviors was beyond the scope of this study, because only
limited measures of other withdrawal behaviors were gathered.
However, in future research a broader taxonomy of withdrawal
behaviors-including lateness, absence, and leaving early-could be
developed and tested with inductively defined thresholds to determine
pattern, frequency, and duration categories for all of the measured withdrawal behaviors. Ideally, the same threshold values would be applicable
across all of the withdrawal behaviors, but the investigator should be prepared to develop different cutoff values for each withdrawal behavior.
Such different behavioral thresholds would reflect the organization's
sensitivity to each withdrawal behavior. For example, with a sample
of nurses Blau (1985b) found that a hospital's control policy for lateness
was not as strict as it was for unexcused absence.
To conclude, despite the inductive approach used to test the proposed taxonomy of lateness categories, the results seem promising.
Certainly, additional work-operationalizing lateness-behavior categories
with other samples is needed. The dominance of blue-collar and shiftbased jobs in lateness research may be at least partially due to lateness
incidents being more carefully monitored for such jobs. It would be interesting to see the extent to which this taxonomy was supported with managerial or other jobs traditionally given more flexibility regarding
lateness. Finally, this study suggests that investigators must be more sensitive to how lateness is operationalized. More careful operationalization
can result in a greater understanding of the lateness construct.
ACKNOWLEDGMENTS. I gratefully acknowledge the input of Celia
Kamath, John Deckop, Frank Linnehan, and Diana Deadrick on earlier
versions of this article. This research project was funded in part by a
grant-in-aid-of-research from Temple University. Portions of this article
were presented at the 1993 National Academy of Management meeting in
Atlanta.
REFERENCES
Abelson, M. (1987). Examination of avoidable and unavoidable turnover. Journal
Adler, S., & Golan, J. (1981). Lateness as a withdrawal behavior. Journal of Applied
Angle, H., & Perry, J. (1981). An empirical assessment of organizational commitment
and organizational effectiveness. Administrative Science Quarterly, 26,1-14.
157
Barcikowski, R., & Stevens, J. (1975). A Monte Carlo study of the stability of
canonical correlations, canonical weights, and canonical variate-variable
correlations. Multivariate Behavioral Research, 10, 354-364.
Beehr, T., & Gupta, N. (1978). A note on the structure of employee withdrawal.
Blass, T. (1977). Personality variables in social behavior. Hillsdale, NJ: Erlbaum.
Blau, G. (1985a). A multiple study investigation of the dimensionality of job
involvement. Journal of Vocational Behavior, 27, 19-36.
Blau, G. (1985b). Relationship of extrinsic, intrinsic, and demographic predictors
to various types of withdrawal behaviors. Journal of Applied Psychology, 70,
442-450.
Blau, G. (1986). Job involvement and organizational commitment as interactive
Blau, G., & Boal, K. (1987). Conceptualizing how job involvement and organizational commitment affect turnover and absenteeism. Academy of Management
Review, 12, 288-300.
Brooke, P. (1986). Beyond the Steers and Rhodes model of employee attendance.
Academy of Management Review, 11, 345-361.
Brooke, P., Russell, D., & Price, J. (1988). Discriminant validation of measures
of job satisfaction, job involvement and organizational commitment. Journal
Burka, J., & Yuen, L. (1983). Procrastination. Reading, MA: Addison-Wesley.
Campion, M. (1991). Meaning and measurement of turnover: Comparison of
alternative measures and recommendations for research. Journal of Applied
Cascio, W. (1987). Costing human resources: The financial impact of behavior in organizations (2nd ed.). Boston: Kent.
Clegg, C. (1983). The psychology of employee lateness, absence and turnover:
A methodological critique and empirical study. Journal of Applied Psychology,
68,88-101.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and
Psychological Measurement, 20, 37-46.

Cummings, R., & Manring, S. (1977). The relationship between worker alienation
and work-related behavior. Journal of Vocational Behavior, 10, 167-179.
Dalton, D., & Mesch, D. (1991). On the extent and redaction of avoidable absenteeism: An assessment of absence policy provisions. Journal of Applied Psychology,
76,810-817.
Dunham, R., & Grube, J. (1990, August). Organizational commitment: The construct
and its measurement. Paper presented at the meeting of the National Academy
of Management, San Francisco.
Duxbury, L., & Higgins, C. (1991). Gender differences in work-family conflict.
Farrell, D., & Robb, D. (1980, August). Lateness to work: A study of withdrawal
from work. Paper presented at the meeting of the National Academy
of Management, Detroit, MI.
Frayne, c., & Latham, G. (1987). Application of social learning theory to employee
self-management of attendance. Journal of Applied Psychology, 72, 387-392.
158
Gary Diau
Gibson, R. (1966). Toward a conceptualization of absence behavior of personnel in

organizations. Administrative Science Quarterly, 11, 107-133.
Greenhaus, J., & Beutell, N. (1985). Sources of conflict between work and family
roles. Academy of Management Review, 10, 76--88.
Gupta, N., & Jenkins, G. (1983). Tardiness as a manifestation of employee withdrawal. Journal of Business Research, 11, 61-75.
Hall, D. (1972). A model of coping with role conflict: The role behavior of college
educated women. Administrative Science Quarterly, 17, 471-486.
Hanisch, K., & Hulin, C. (1990). Job attitudes and organizational withdrawal: An
examination of retirement and other voluntary withdrawal behaviors. Journal
Hanisch, K., & Hulin, C. (1991). General attitudes and organizational withdrawal:
An evaluation of a causal model. Journal of Vocational Behavior, 39,110-128.
Harrison, D., & Hulin, C. (1989). Investigations of absenteeism: Using event history models to study the absence-taking process. Journal of Applied Psychology,
74,300-316.
Hermann, J., deMontes, A., Dominquez, B., Montes, F., & Hopkins, B. (1973).
Effects of bonuses for productivity on the tardiness of industrial workers.
Journal of Applied Behavioral Analysis, 6, 563-570.
Herzberg, F., Mausner, B., Peterson, R., & Capwell, D. (1957). Job attitudes: Review
of research and opinion. Pittsburgh: Psychological Services.
Jamal, M. (1981). Shift work related to job attitudes, social participation and
withdrawal behavior: A study of nurses and industrial workers. Personnel
Jones, E., & Nisbett, R. (1972). The actor and the observer: Divergent perceptions
of the cause of behavior. In E. Jones, D. Kanouse, H. Kelley, R. Nisbett,
S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the cause of behavior
(pp. 63-86). Morristown, NJ: General Learning Press.
Kanungo, R. (1982). Measurement of job and work involvement. Journal of Applied
Kempen, R. (1982). Absenteeism and tardiness. In L. Frederiksen (Ed.), Halldbook

of organizational behavior and management. New York: Wiley.
Kennedy, D. (1984). A counseling approach to chronic tardiness. Supervisory
Management,29,25-29.
Kite, P. (1984). Preventing employee tardiness. Supervision, 46,14-15.

Kopelman, R., Greenhaus, J., & Connolly, T. (1983). A model of work, family and
interrole conflict: A construct validation study. Organizational Behavior and
Human Performance, 32, 198-215.
Leigh, J., & Lust, J. (1988). Determinants of employee tardiness. Work and Occupations,
15,78-95.
Lobel, S. (1991). Allocation of investment in work and family roles: Alternative theories and implications of research. Academy of Management Review, 16, 507-521.
McNemar, Q. (1969). Psychological statistics. New York: Wiley.
Macy, B., & Mirvis, P. (1976). A methodology for assessment of quality of work life
and organizational effectiveness in behavior-economic terms. Administrative
Science Quarterly, 21, 212-226.
159
March, J., & Simon, H. (1958). Organizations. New York: Wiley.

Mathieu, J., & Farr, J. (1991). Further evidence for the discriminant validity of
measures of organizational commitment, job involvement and job satisfaction. Journal of Applied Psychology, 76, 127-133.
Meyer, J., & Allen, N. (1984). Testing the "side bet theory" of organizational commitment: Some methodological considerations. Journal of Applied Psychology,
69,372-378.
Mobley, W. (1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology, 62, 237-240.
Mobley, W., Griffeth, R, Hand, H., & Meglino, B. (1979). Review and conceptual
analysis of the employee turnover process. Psychological Bulletin, 86, 493-522.
Morgan, L., & Herman, J. (1976). Perceived consequences of absenteeism. Journal
Motley, R (1926). Lateness of plant employees: A study of causes and cures.
Journal of Personnel Research,S, 1-3.
Nicholson, N., & Goodge, P. (1976). The influence of social, organizational, and
biographical factors on female absence. Journal of Management Studies, 13,
234-254.
Nie, N., Hull, c., Jenkins, J., Steinbrenner, K., & Bent, D. (1975). SPSS: Statistical
package for the social sciences. New York: McGraw-Hill.
Nollen, S., & Martin, V. (1978). Alternative work schedules: Flextime. New York:
AMACOM.
Nunnally, J. (1978). Psychometric theory. New York: McGraw-Hill.
Paulhus, D. (1986). Self deception and impression management in test responses.
In A. Angleitner, & J. Wiggins (Eds.), Personality assessment via questionnaire
(pp. 143-165). Berlin: Springer-Verlag.
Payne, R, & Nicholson, N. (1987). Absence from work: Explanations and attributions. Applied Psychology: An International Review, 36,121-132.
Porter, L., Crampon, W., & Smith, F. (1976). Organizational commitment and
managerial turnover: A longitudinal test. Organizational Behavior and Human
Performance, IS, 87-98.
Richard, D., & Slane, S. (1990). Punctuality as a personality characteristic: Issues of
measurement. Journal of Psychology, 124, 397-402.
Rosse, J. (1988). Relations among lateness, absence and turnover: Is there
a progression of withdrawal? Human Relations, 41, 517-531.
Rosse, J., & Hulin, C. (1985). Adaptation to work: An analysis of employee health,
withdrawal and change. Organizational Behavior and Human Decision Processes,
36, 324-347.
Ruchti, W. (1967). Is there an answer for lateness and absenteeism? Supervisory
Management, 12, 21-22.
Schultz, J., & Henderson, C. (1985). Family satisfaction and job performance:
Implications for career development. Journal of Career Development, 12, 33-47.
Schwab, D. (1980). Construct validity in organizational behavior. In B. Staw, &
L. Cummings (Eds.), Research in organizational behavior (Vol. 2, pp. 3-43).
Greenwich, CT: JAI Press.
160
Gary Blau
Steers, R., & Mowday, R. (1981). Employee turnover and post-decision accommodation processes. In B. Staw, & L. Cummings (Eds.), Research in organizational
behavior (Vol. 3, pp. 235-281). Greenwich, CT: JAI Press.
Steers, R., & Rhodes, S. (1978). Major influences on employee attendance:
A process model. Journal of Applied Psychology, 63, 391-407.
Thompson, c., & Blau, G. (1993). Moving beyond traditional predictors of job
involvement: Exploring the impact of work-family conflict and overload.
Journal of Social Behavior and Personality, 8, 635-646.
Weiss, D., Dawis, R., England, G., & Lofquist, L. (1967). Manual for the Minnesota
Satisfaction Questionnaire. Minneapolis: Industrial Relations Center,
University of Minnesota.
Williams, K., 5uls, J., Alliger, G., Learner,S., & Wan, C. (1991). Multiple role juggling and daily mood states in working mothers: An experience sampling
study. Journal of Applied Psychology, 76, 664-674.
Youngblood,S. (1984). Work, nonwork and withdrawal. Journal of Applied
7
New Conceptualizations of
Lateness since Blau, 1994
Gary Blau
Employee lateness is still conceptualized as a distinct and "neglected"

form of physical withdrawal behavior, compared to research investigating employee absenteeism (e.g., Brown, Fakhfakh, & Sessions, 1999;
Harrison & Martocchio, 1998) and employee turnover (e.g., Dalton,
Johnson, & Daily, 1999; Koys, 2001). As Blau (1994) noted, employee lateness has been traditionally defined and measured as arrival after the
beginning of scheduled work.
Is such lateness still an important variable to understand, given
the new work options being increasingly explored in organizations, such
as flextime, telecommuting, and telework (Apgar, 1998; Cascio, 2000;
Igbaria & Tan, 1998)? Blau's (1994) literature review suggests that lateness
research will remain somewhat sample-specific, i.e., focusing on shiftwork jobs (e.g., manufacturing, security), and customer-service jobs,
where start times are critical for the organization to monitor (e.g., transportation, hospital, banking). A recent survey-based study of 1281 business professionals on rudeness in business (Humphries, 2000) found that
"time abuse behaviors," including employee lateness to work, were identified by 34% of the respondents. Humphries (2000) further found that
such rudeness can be costly; 58% of customers said they would take their
business elsewhere, even if it was inconvenient or cost more. This suggests that employee lateness is still an important variable to study.
GARY OLAV Human Resource Administration Department, Temple University, Fox
School of Business Administration, Philadelphia, Pennsylvania 19122
161
162
Gary Blau
Hepburn and Barling (1996) have more recently expanded the definition of employee lateness to include any unexcused time missed during
the workday, such as taking too long of a lunch or work break, or leaving
work early, in addition to a tardy arrival. Using Hepburn and Barling's
(1996) broadened definition of lateness allows it to be understood as a
form of "production deviance" within Robinson and Bennett's (1995)
model of deviant work behaviors.
Blau (1994) did find significant correlations between increasing
chronic lateness and leaving work early for both bank (r = .38) and hospital (r = .43) employee samples.
EMPIRICAL RESEARCH ON LATENESS SINCE BLAU, 1994
There has been limited research on employee lateness since 1994. In a
longitudinal study, Blau (1995) found that a measure of presurvey group
(department)-levellateness explained a significant amount of postsurvey
individual lateness, beyond that due to presurvey individual lateness, and
various surveyed work measures, i.e., job involvement, organizational
commitment, job satisfaction, leisure-income trade-off, work-family conflict, transportation concerns, and illness or accidents. Describing the influence of a construct (lateness) at one level (i.e., group) on that variable
measured at a different level (i.e., individual) necessitates a cross-level
research design, in which the mean aggregate (group-level) value for a
variable (i.e., group lateness) is assigned to all members within a given
group. Then the aggregate variable can be correlated with individual-level
variables across all individuals to assess cross-level influence (Rousseau,
1985).
In Blau's (1995) study, 66 bank branches and 71 hospital departments
were used to measure group-level lateness. Group-level lateness scores
were obtained by averaging occurrences (frequency times duration) by
bank branch for tellers and by department for hospital employees, and
then assigning the respective mean values back to individuals (Rousseau,
1985).
Since the group measure would include the individual measure, to
control for measure contamination it is necessary to separate their measurements, i.e., pre- and postsurvey.
Why would group-level measures of lateness affect individual lateness? Applying Nicholson and Johns's conceptual work (1985) on the
social dynamics of absence in work settings suggests that social norms for
absence are based on three key referent groups: co-workers, supervisors,
and family and friends. Furthermore, the salience of norms and beliefs
New Conceptualizations of Lateness
163
about absence can change as these referents change (Harrison &

Martocchio, 1998; Miners, Moore, Champoux, & Martocchio, 1995). These
three referents may exert a similar influence on social norm development
for lateness. Research is needed to test such speculation.
Bercovits and Koslowsky (1996) found that time urgency, a multidimensional construct including time awareness, scheduling, and deadline
setting facets (Landy, Rastegary, Thayer, & Colvin, 1991), significantly discriminated between late and never-late workers. Workers with higher levels of time urgency had fewer lateness incidents. Applied to the Blau
(1994) taxonomy of lateness behavior categories, employees with greater
time urgency will be less unavoidably late because they will plan ahead
for traffic and bad weather factors.
In their meta-analysis of employee lateness (Le., tardy work arrival),
Koslowsky, Sagie, Krausz, and Singer (1997) found a stronger corrected
mean correlation of general lateness to voluntary absence (r = .41) versus:
involuntary absence (r = .16); turnover (r = .27); job satisfaction (r = -.11),
organizational commitment (r = - .27); and age (r = - .19). The strongest
correlation, however, was between intent to turn over and lateness (r = .46).
This suggests that employees who think they may be voluntarily leaving
their job may not be as concerned about their punctuality. In order to have
the critical number of samples needed to perform their meta-analysis,
Koslowsky et al. (1997) measured lateness in general, as opposed to specific
lateness behavior categories (Blau, 1994).
CONCLUSION
Although employee lateness remains important to study
(Humphries, 2000), there are clearly many practical and research challenges that may partially explain why lateness remains "neglected" compared to more prominent physical withdrawal behaviors, i.e., absence
and turnover. Finding organizations concerned enough about the "cost"
of employee lateness to allow for a stronger research design, i.e.,
employee identification to allow longitudinal and record-based data to be
collected, will remain difficult. Broadening the definition of lateness to
any unexcused time missed during the workday, including leaving early
and excessive breaks (Hepburn & Barling, 1996), can heighten the "costs"
associated with lateness to an organization, and further justify the organizational need to have it studied.
From a research standpoint, there are challenges to be faced. Even if
an organization allows for such a stronger research design, there may be
either poor organizational record-keeping or an insufficient subs ample
164
Gary Blau
size problem, which will prevent lateness behavior categories, i.e., pattern,
frequency, and duration of lateness behavior, from being operationalized
(Blau, 1994). Self-report lateness may be a necessary research strategy
option. Borrowing from Johns's (1994) discussion of using self-report
absence measures, one should consider: making the lateness accumulation
interval correspond to a salient and natural time unit (e.g., for teachers a
school year); lacking a salient time-frame, make the accumulation long
enough to enhance reliability but short enough to minimize memory loss
(e.g., 3-6 months); opt for a free-response format for obtaining time-lost or
frequency data (lower scale categories suggest to subjects that lower rates
are normal); assume that lateness questions may be threatening, so try to
ask longer questions which assume the behavior has been exhibited; and
to reduce self-generated validity and other artifacts, separate lateness
items within the survey.
Dalton and Mesch (1991) operationalized separate self-report measures of unavoidable (involuntary) and avoidable (voluntary) absence in a
creative manner, which could be extended to lateness behavior. By asking
employees "how many times were you late for work in the last 6 months
due to such factors as bad weather, getting stuck in traffic, having a car
problem or accident, unreliable public transportation, having a child or
other dependent care problem, or feeling sick?" this would measure
unavoidable (involuntary) lateness. Avoidable lateness would be measured as the difference between employees' response to this item and a
"total" lateness item, "how many times were you late to work for any reason in the last 6 months?" Using this approach, the avoidable lateness
measure would combine Blau's (1994) increasing chronic and stable periodic lateness categories, while the unavoidable lateness measure would
reflect Blau's (1994) same behavior category. Such a research design
should allow for a stronger understanding of lateness than using only a
general or overall lateness measure. It is hoped that this updated discussion of lateness will stimulate further research efforts.
REFERENCES
Apgar, M.1. (1998). The alternative workplace: Changing where and how people
work. Harvard Business Review, 76, May-June, 121-136.
Bercovits, M. D., & Koslowsky, M. (1996). Time urgency: A personality attribute for predicting lateness. Ramat Gan: Bar-Han University, Unpublished master's thesis.
Blau, G. (1994). Developing and testing a taxonomy of lateness behavior. Journal of
Applied Psychology, 79,959-970.
Blau, G. (1995). Influence of group lateness on individual lateness: A cross-level
examination. Academy of Management Journal, 38, 1483-1496.
New Conceptualizations of Lateness
165
Brown,S., Fakhfakh, E, & Sessions, J. (1999). Absenteeism and employee sharing:

An empirical analysis based on French panel data, 1981-1991. Industrial and
Labor Relations Review, 52, 234-251.

Cascio, W. E (2000). Managing a virtual workplace. Academy of Management
Executive, 14, 81-90.
Dalton, D. R., Johnson, J. L., & Daily, C. M. (1999). On the use of "intent to" variables in organizational research: An empirical and cautionary assessment.
Human Relations, 52,1337-1350.
Dalton, D. R., & Mesch, D. (1991). On the extent and reduction of avoidable absenteeism: An assessment of absence policy provisions. Journal of Applied

Harrison, D. A., & Martocchio, J. J. (1998). Time for absenteeism: 20 year review of
origins offshoots and outcomes. Journal of Management, 24, 305-350.
Hepburn, C. G., & Barling, J. (1996). Eldercare responsibilities, interrole conflict and
employee absence: A daily study. Journal of Occupational Health Psychology,
1,311-318.
Humphries, A. C. (2000). Rudeness: The hidden cost to business. Business and
Economic Review, 47, 22-25.
Igbaria, M., & Tan, M. (1998). The virtual workplace. Hershey, PA: Idea Group
Publishing.
Johns, G. (1994). How often were you absent? A review of the use of self-reported
absence data. Journal of Applied Psychology, 79, 574-591.
Koslowsky, M., Sagie, A., Krausz, M., & Singer, A. D. (1997). Correlates of employee
lateness: Some theoretical considerations. Journal of Applied Psychology,
82,79-88.
Koys, D. J. (2001). The effects of employee satisfaction, organizational citizenship

behavior, and turnover on organizational effectiveness: A unit-level, longitudinal study. Personnel Psychology, 54,101-114.
Landy, F. L., Rastegary, H., Thayer, J., & Colvin, C. (1991). Time urgency: The
construct and its measurement. Journal of Applied Psychology, 76, 64~57.
Miners, I. A., Moore, M. L., Champoux, J. E., & Martocchio, J. J. (1995). Time-serial
substitution effects of absence control on employee time-use. Human
Relations, 48, 307-326.
Nicholson, N., & Johns, G. (1985). The absence culture and the psychological
contract-who's in control of absence. Academy of Management Review, 10,
397-407.
Robinson, S., & Bennett, R. (1995). A typology of deviant workplace behaviors: A

multi-dimensional scaling study. Academy of Management Journal, 38, 555-572.
Rousseau, D. (1985). Issues of level in organizational research: Multi-level and
cross-level perspectives. In B. M. Staw & L. L. Cummings (Eds.), Research in
organizational behavior. (Vol. 7, pp. 1-37). Greenwich, CT: JAI Press.
8
Smoking and Absence from Work
A Quantitative Review
E. Kevin Kelloway, Julian Barling,
and Caroline Weber
Absenteeism continues to be one of the most costly problems facing

organizations today. For many years estimates of the cost of absenteeism in
North America have exceeded $25 billion annually (e.g., Steers & Rhodes,
1978) and, perhaps not surprisingly, researchers have focused considerable
attention on the identification of the causes of workplace absenteeism. As
noted by Hackett, Bycio, and Guion (1989), three perspectives have largely
driven organizational research on absenteeism. Absence has been seen as a
means of avoiding negative aspects of the job, a rational decision or choice
process, and as a result of socialization and job demands.
One consistent focus of attention in the research literature has been
the hypothesized link between work attitudes (e.g., job satisfaction) and
absenteeism. Hackett (1989) reviewed three meta-analyses of this relationship and concluded that there was consistent evidence for a modest
correlation between absence frequency and work satisfaction (r = - .21) as
well as between absence duration and overall job satisfaction (r = - .23).
Although these data are consistent, the observed relationship is small
in magnitude suggesting that either (1) work attitudes are only a minor
influence on absenteeism or (2) other factors plausibly affect absenteeism
E. KEVIN KELLOWAY Department of Management, Saint Mary's University, Halifax,
Nova Scotia B3H 3C3, Canada
JULIAN BARLING School of Business, Queen's
University, Kingston, Ontario K7L 3N6, Canada
CAROLINE WEBER School of
Industrial Relations, Queen's University, Kingston, Ontario K7L 3N6, Canada
167
168
E. Kevin Kelloway et al.
rates and result in lessening the magnitude of the observed attitudeabsence correlation (Johns, 1991). In this chapter we turn our attention to
the latter suggestion. Specifically, we present data suggesting a link
between employee lifestyle behaviors (Le., cigarette smoking) and workplace absence.
One of the major social questions in recent years has been the effect
of cigarette smoking on individual health and well-being. Sufficient data
have now been accumulated to conclude that smoking does exert adverse
health consequences (Bush & Wooden, 1995), especially increased risks
for coronary heart disease and cancer. Moreover, smokers are more likely
than nonsmokers to experience less severe ailments such as upper respiratory tract infections (Bush & Wooden, 1995; Henningfield et al., 1994).
The association between smoking and ill health begs the question of
whether there are other "hidden" costs associated with cigarette smoking;
"hidden" in that the impact of smoking on outcomes other than physical
ailments has to date received comparatively less attention.
The purpose of our study was to investigate one such potential cost,
namely, whether tobacco smokers are disproportionately more likely to be
absent from work than are nonsmokers. The basis for this question lies
in the observation that smokers are less healthy than nonsmokers
(Henningfield et al., 1994) leading to the suggestion that smokers' attendance at work will be negatively affected. Certainly there are numerous
references in the medical and human resource literatures suggesting an
association between smoking and absenteeism. However, close inspection
of this literature suggests that there is considerably more discussion about
this possible association then there are data supporting such a link. In
other words, the literature is currently driven more by rhetoric than by
empirical data. In addition, studies that do contain empirical data vary
markedly in quality. Therefore, the primary purpose of our study was to
statistically summarize the empirical data linking smoking and absenteeism. In doing so, we empirically integrate findings from numerous
studies to provide an estimate of the effect of smoking on absenteeism.
Second, a cursory inspection of the empirical literature on smoking
and absenteeism suggests that this question has drawn the attention of
researchers from many countries (e.g., Germany, Sweden, Poland, China,
Israel, Australia, the United States, the United Kingdom, and New
Zealand). Many of these countries differ from one another in terms of
both their "smoking culture" (e.g., public policy, attitudes, and legislation
regarding cigarette smoking) and workplace structures (including workplace policies regarding absenteeism). Accordingly, we also address
whether the magnitude of the association between smoking and absenteeism varies across studies.
Smoking and Absenteeism
169
Third, assuming that a reliable association exists between smoking and

absenteeism, the question of whether quitting smoking results in a decrease
in absenteeism assumes considerable importance (Bush & Wooden, 1995;
Wooden & Bush, 1995). Certainly, quitting smoking has beneficial effects on
medical outcomes including the reversal of some adverse smoking-related
conditions. Thus, because we believe that the intervening variable between
smoking and absence is ill health, it is reasonable to suggest that smoking
cessation results in decreased absenteeism.
To summarize, we present an empirical or quantitative study of the
relationship between smoking and absenteeism. In doing so we address
three specific research questions. First, is smoking associated with
increased absenteeism and, if so, what is the magnitude of this effect?
Second, does the magnitude of the effect vary according to the country in
which the data were collected? Finally, is smoking cessation associated
with a decrease in absence from work.
METHOD
Database
For a quantitative analysis to provide an accurate estimate of a relationship, it is critical that all available data be collected and analyzed. To
ensure that we collected as many studies on the relationship between
smoking and absenteeism as were obtainable, we conducted computerized searches of several databases, namely, Medline Medical Index,
American Business Index, Psychological Information Database, Social
Sciences Index, and Humanities Index. All studies purporting to provide
data on smoking and absenteeism were obtained. This included several
articles in languages other than English (e.g., German, Japanese,
Afrikaans) which were translated into English. Numerous articles that
purported to obtain data on the relationship of interest but did not contain any empirical data were also located. Because discussions on the link
between smoking and absenteeism frequently refer to the large database
available, a bibliography of these studies is available from the second
author on request. The articles that did produce usable data are presented
in Table 1.
Measures
In order to combine results across studies, it was necessary first to
develop common operationalizations of both cigarette smoking and
170
Table 1. Descriptive Information for the 25 Studies Providing Complete Data

Country
Data
year
smokers
Smoker
absence
Year
Anderson &
Malmgren
Athansou
Batenburg &
Reinkein
Bertera
Carmichael &
Cockcroft
Fawer et al.
Foerster et al.
Gabel & ColleyNiemeyer
Gallop
Green et al.
Hawker &
Holtby
Hayden et al.
Hendrix et al.
Hokomb&
Meiggs
Janzon et al.
Kark et al.
Kozak
Lowe
Manning et al.
Marti
McMillan
Qun&
Dobson
Ryan et al.
Smith et al.
Tsai et al.
1986
Sweden
1975
369
22.5
556
18.85
1979
1990
Australia
New Zealand
1978
1988
205
337
18.71
6.18
219
555
10.79
5.51
1990
1990
USA
UK
1988
1988
13,060
99
3.69
7
32,916
79
2.79
4
1982
1976
1990
UK
Germany
USA
1979
1975
1987
26
110
194
2.6
28.5
13.48
12
55
590
1.65
15.8
10.87
1991
1992
1988
New Zealand
Israel
UK
1988
1987
1987
20
2,067
40
10.3
10.99
4.63
33
3,759
121
7.9
10.3
3.39
1984
1991
1972
UK
USA
USA
1983
1986
1964
277
97
114
1981
1982
1987
1960
1989
1986
1981
1992
Sweden
USA
Czech
UK
USA
Germany
UK
China
1978
1981
1984
1957
1984
1984
1979
1987
517
186
373
2,284
96
299
1,962
1,479
13
1.36
25.9
6.57
5.74
3.57
5.58
0.39
230
163
250
1,057
206
3,660
929
377
4
0.91
28.1
5.49
3.95
2.1
4.11
0.27
1992
1981
1990
USA
Australia
Sweden
1989
1980
1986
825
226
2,388
12.96
2.94
7.79
1,712
479
4,609
9.84
2.11
5.12
28.5
6.4
6.9
nonsmokers
Nonsmoker
absence
Authors
239
366
29
15
5.8
4.3
absence from work. Studies varied considerably in their measurement of

smoking behavior. For example, some authors defined the smoking group
as individuals who reported smoking any amount of tobacco. Other
researchers limited the group of smokers to those who smoked specific
amounts (e.g., 15 cigarettes/day). Some researchers provided breakdowns of smoking intensity, dividing smokers into "heavy," "light," or
"occasional" smokers, often based on different criteria across studies.
Finally, some authors differentiated between pipe, cigar, and cigarette
smokers while others did not. In order to retain the maximum amount of
data in the analysis, we used the lowest common denominator in defining smokers as individuals who currently (i.e., at the time of the original
study) reported smoking any number of cigarettes. In cases where the
group of smokers was defined by the minimum daily amount of smoking
171
and the original researchers did not provide additional information, we

used the original author's definition of smoking.
Similarly, individual studies varied widely in their reporting of
absenteeism statistics. Some studies reported absence in terms of hours
per month, others in terms of days per year. In all articles used in the current analysis, sufficient information was presented to allow the calculation of absenteeism as days lost per year. In converting absenteeism
figures from hours to days we followed the frequently used definition of
a workday as 7! hours.
Method of Data Analysis
To be included in the current analysis, articles had to report absenteeism figures for, and the number of, both smokers and nonsmokers.
Our task was then to cumulate these findings across studies to provide
an estimate of the effects of smoking on absenteeism. In cumulating
findings across studies we followed the common meta-analytic practice
(Hunter, Schmidt, & Jackson, 1982) in choosing to use a weighted rather
than a simple average. 1 Thus, absenteeism data for smokers and nonsmokers were weighted by the sample size in each group. This procedure gives greater weight to findings derived from large studies and
less weight to findings based on smaller samples. This is important
because it well established that large-sample estimates are more precise
than are estimates based on small samples.
After deriving a weighted absenteeism figure for both smokers and
nonsmokers in each study, we then cumulated the measures across studies and divided by the total number for smokers (or nonsmokers as
appropriate). The end of this procedure was a single absenteeism figure
for all smokers across all studies and a comparable average for nonsmokers. These averages were then used in further analyses.
RESULTS
Is Smoking Associated with Absence from Work?
Twenty-five studies provided sufficient data to compare the
absenteeism rates of smokers and nonsmokers. Descriptive information
1
Ideally one would also derive an estimate of sampling error and be able to correct for study
artifacts (e.g., restriction in range, measurement error). Unfortunately, the data available
to us did not allow for such corrections-there were insufficient data reported to allow us
to estimate these effects nor were sufficient data available to allow the estimation of
artifact distributions.
172
from these studies is presented in Table 1. Absenteeism among smokers

(M = 6.37 days per year, N = 27,650) was significantly higher than was
absenteeism among nonsmokers (M = 4.29 days per year, N = 53,201),
t (242) = 3.31, P< .005. Thus, smokers missed 2.08 more days of work each
year than did nonsmokers. Stated somewhat differently, there was a
48.25% increase in absenteeism associated with smoking.
Does the Size of the Effect of Smoking on
Absenteeism Vary by Country?
To investigate this question, we initially combined data within five
geographically similar areas (Australia/New Zealand; Czechoslovakia,
Poland, and Germany; Finland and Sweden; the United Kingdom; and the
United States). A sixth residual category comprising one study from China
and one study from Israel was also formed. The difference in absenteeism
between smokers and nonsmokers in each region is presented in Table 2.
Overall there were no statistically significant differences in absence
from work as a function of smoking across geographical groups. That is,
although there are differences in the mean level of absenteeism in different
geographic regions, the difference between smokers' and nonsmokers'
absenteeism did not differ significantly across region. Alternative ways of
combining the data (e.g., USA versus all other countries; European countries versus all others) also resulted in no statistically significant differences between groups. Thus, the strength of the effect of tobacco smoking
on absence from work was deemed to be stable across countries.
Does Smoking Cessation Affect Absence from Work?
Of the original 25 studies that we located, 8 presented data on former
smokers and allowed for comparison of absenteeism rates among current
Table 2. Annual Average Absenteeism of Smokers and Nonsmokers by
Geographic Region
Region
Australia/New Zealand
Germany /Poland/Czechoslovakia
Sweden/Finland
United Kingdom
United States
China/Israel
Smokers
Nonsmokers
9.53
19.32
14.43
9.15
7.10
5.69
6.58
15.33
9.32
5.61
5.69
5.29
Note. No Significant differences between regions were observed.
173

Table 3. Annual Absenteeism of Current and Former Smokers
Authors
Country
Data
year
Smoker N
N
Nonsmoker
smokers absence nonsmokers absence
USA
1987
194
13.48
210
11.39
20
10.30
29
9.10
1972
1988
New
Zealand
USA
1964
114
6.90
44
6.18
1989
1981
1989
1981
1990
USA
Sweden
USA
UK
Sweden
1986
35
1978 517
1984
96
1979 1962
1986 2388
Mean
41
13
5.74
5.58
7.79
7.84
35
290
96
686
866
30
7
3.44
4.21
7.19
6.87
Year
Gabel & Colley- 1990

Niemeyer
1991
Gallop
Holcomb &
Meiggs
Jackson et al.
Janzon et al.
Manning et al.
McMillan
Tsai et al.
smokers, former smokers, and nonsmokers. Descriptions of these studies

are presented in Table 3.
Former smokers exhibited less absenteeism (M = 6.87, N = 2246) than
did current smokers (M = 7.84, N = 5326) although this difference was not
statistically significant (p> .05). Former smokers also exhibited more
absenteeism than did nonsmokers (M = 5.43, N = 6626) although again
this difference was not statistically significant.
DISCUSSION
The primary purpose of our investigation was to investigate the existence and size of an association between cigarette smoking and absence
from work by cumulating results gleaned from almost 40 years of published empirical research on this question. Our results strongly support
the existence of such an association with smokers being absent 2.07 days
per year more than nonsmokers; an increase of 48%. These results point to
the substantial organizational costs of smoking.
Two potential moderators of the smoking absenteeism relationship
were also investigated. First, we found that the size of the effects reported
in the literature did not vary according to where the data were collected.
Consequently, any effects of tobacco smoking on absence from work may
be considered to be stable across countries despite national! cultural
differences in smoking and absenteeism patterns. Second, a limited
number of studies reported data on former smokers in addition to data
from smokers and nonsmokers. This allowed us to assess the potential
174
effects of smoking cessation on absence. Although we did observe

that former smokers had a lower absenteeism rate than current smokers
(0.97 day /year), the difference was not statistically significant. However,
given the small number of studies providing complete information on
smoking cessation, it is possible that insufficient data, and the resultant
loss of power, accounted for the lack of a statistically significant finding.
Given that our strategy was to summarize and cumulate empirical
findings from a number of studies, our conclusions are potentially limited
by several observations.
First, the data that were used in our study were the direct comparison of absenteeism between groups of smokers and nonsmokers. While it
is convenient to summarize these data as suggesting the "effects" of
smoking on absenteeism, it should be noted that this phrasing implies a
causal inference that is not justified by the data. Most of the studies we
reviewed were cross-sectional, observational studies and therefore
allowed the possibility of reverse causality or "third-variable" effects.
For example, Ault, Ekelund, Jackson, Saba, and Saurman (1991) conducted a multivariate analysis and concluded that the association between
smoking and absenteeism was attributable to smokers being more likely to
be younger, moderate to heavy drinkers, blue-collar workers, and to have
shorter tenure in the workplace-factors also associated with higher levels
of absenteeism. Subsequent research that also implemented statistical controls (e.g., Bush & Wooden, 1995) provided conflicting findings in which
the association between smoking and absenteeism remained after controlling for a host of demographic, health, and lifestyle-related variables.
Clearly, quasi-experimental, experimental, and longitudinal data implementing controls over a wealth of potentially confounding variables (e.g.,
lifestyle factors, health status) are required in order to justify this implied
causal inference.
Ironically, when such multivariate analyses have been conducted,
researchers have chosen to report the multivariate estimates (e.g., logistic
regression weights; Bush & Wooden, 1995) rather than the bivariate estimates (e.g., mean differences) as in the earlier research. This reporting
strategy causes some difficulty for researchers wishing to cumulate
results across studies. The difficulty emerges from the fact that multivariate effects such as regression weights cannot be cumulated across studies
(Hunter et al., 1982). While such analyses are clearly required to estimate
the unique effect of smoking on absenteeism, there also remains a need
for reporting at the level of bivariate relationships.
Second, in deriving common measures of smoking and absence
across studies, considerable information is sacrificed. In particular, quantitative measures of smoking intensity (e.g., number of cigarettes smoked
175
per day) were frequently dichotomized to form groups of smokers and

nonsmokers. While this procedure allowed the comparison of results
across diverse studies, it also resulted in the loss of more precise measures. We suspect that this practice may have resulted in an underestimate
of the effect due to the loss of variance resulting from dichotomizing a
continuous variable.
Third, most of the source articles did not report complete descriptive
statistics for measures of absenteeism. In particular, measures of withingroup variance were not available. This is a particularly important omission because it prohibits the correction for sampling error that allows
precise population estimates to be derived from meta-analytic procedures
(Hunter et a1., 1982). In the current case we could cumulate the means but
have no information regarding the extent to which these "average" results
vary across studies.
Similarly, it should be noted that the studies contributing data to our
analyses focused on only full-day absenteeism (e.g., the use of "sick
days"). However, smoking is also associated with partial absence (i.e.,
taking cigarette breaks during the day) which can also result in significant
organizational costs. Greenberg, Finkelstein, and Berndt (1995) estimate
that such breaks constitute a loss of 3.125% of working time. Assuming
that 20% of the work force smokes, the value of this lost productive time
is estimated to approach $17 billion, U.S. each year (Greenberg et a1.,
1995). The lost time attributable to partial absence (i.e., approximately
8 days/year) may be substantially higher than the effect of smoking on
full-day absence that has been the focus of most empirical research.
The importance of considering partial absenteeism is highlighted by
the observation that Greenberg and co-workers' (1995) estimate is based
on smokers accruing 15 minutes each day in partial absenteeism. We suggest that this estimate is substantially lower than the actual amount of
time lost by smokers during the day and that doubling, or even tripling,
this estimate would not be unreasonable. Moreover, regional differences
in smoking patterns may make the estimate of smoking prevalence (i.e.,
20% of the work force) a very conservative assumption.
Likewise, the quality of data reporting on smoking was frequently
questionable. Researchers who have addressed the link between smoking
and absence have largely ignored variables such as the length of time
individuals had been smokers, why former smokers might have quit
smoking (e.g., ill health or the choice of a more healthy lifestyle), and a
host of other descriptors. Again this results in a very conservative assessment of the effects of both smoking and smoking cessation on absenteeism and research in these areas would be greatly enhanced if greater
attention would be paid to the measurement and reporting of smoking
176
behavior. Indeed, in evaluating the effects of smoking cessation on absenteeism, Wooden and Bush (1995) observed that absenteeism declines as a
function of the number of years as an ex-smoker, little effect being
observed in the first year.
In conclusion, the available data support the conclusion that cigarette
smoking is associated with increased absenteeism from work. Smokers
reported approximately 48% more absenteeism from work than did nonsmokers. Moreover, the magnitude of this effect was stable across countries. In addition, some empirical data suggest that smoking cessation
may be associated with decreased absenteeism. Taken together these findings suggest that there are significant organizational costs associated with
cigarette smoking and that organizations may be able to avoid these costs
by supporting smoking cessation and similar lifestyle programs.
ACKNOWLEDGMENTS. This research was supported by grants from
Marion Merrell Dow to all three authors and by Social Sciences and
Humanities Research Council of Canada grants to the first and second
authors. The authors express their appreciation to Doug Grant for his
assistance throughout the research.
REFERENCES
Andersson, G., & Malmgren, S. (1986). Risk factors and reported sick leave among
employees of Saab-Scania, Linkoping Sweden between the ages of 50 and 59.
Scandinavian Journal of Social Medicine, 14, 25-30.
Athansou, J. A. (1979). Smoking and absenteeism. Medical Journal of Australia, 1,
234-236.
Ault, R. w., Ekelund, R. B., Jackson, J. D., Saba, R. 5., & Saurman, D. S. (1991).
Smoking and absenteeism. Applied Econometrics, 23, 743.
Batenburg, M., & Reinkein,]. A (1990). The relationship between sickness absence
from work and patterns of cigarette smoking. New Zealand Medical Journal,
103,11-13.
Bertera, R. L. (1990). The effects of behavioral risks on absenteeism and health-care

costs in the workplace. Journal of Occupational Medicine, 33,1119-1124.
Bush, R., & Wooden, M. (1995). Smoking and absence from work: Australian evidence. Social Science and Medicine, 41, 437-446.
Carmichael, A, & Cockroft, A (1990). Survey of student nurses' smoking habits in
a London teaching hospital. Respiratory Medicine, 84, 277-282.
Fawer, R. F., Gardner, A. W., & Oakes, D. (1982). Absences attributed to respiratory
diseases in welders. British Journal of Industrial Medicine, 29,149-152.
Foerster, R. V., Blohmke, M., Koschorreck, B., & Allnoch, K. (1976). Disease,
physician-patient contracts and work disability in general practice. Sozial
und Praventivmedizin, 21,192-196.
177
Gabel, H. D., & Colley-Niemeyer, B. (1990). Smoking in a public health agency:

Its relationship to sick leave and other life-style behavior. Southern Medical
Journal, 83,1-17.
Gallop, B. (1991). Sickness absenteeism and smoking. New Zealand Medical Journal,
102,83.
Green, M. S., Luz, J., & Gofer, D. (1992). Absence from work among smokers and
nonsmokers in Israeli industries: The Cordis study. Israel Journal of Medical
Sciences, 28, 645-649.
Greenberg, P. E., Finkelstein, S. N., & Berndt, E. R. (1995). Economic consequences
of illness in the workplace. Sloan Management Review, 36, 26.
Hackett, R. D. (1989). Work attitudes and employee absenteeism: A synthesis of
the literature. Journal of Occupational Psychology, 62, 235-248.
Hackett, R. D., Bycio, P., & Guion, R. M. (1989). Absenteeism among hospital
nurses: An idiographic-longitudinal analysis. Academy of Management Journal,
32, 424-453.
Hawker, R., & Holtby, 1. (1988). Smoking and absence from work in a population
of student nurses. Public Health, 102(2), 161-167.
Hayden, S. P., Pincock, A. c., Hayden, J., Tyler, L. E., Cross, K. w., & Bishop, J. M.
(1984). Respiratory symptoms and pulmonary functions of welders in the
engineering industry. Thorax, 29, 442-447.
Hendrix, W. H., Steel, R. P., Leap, T. L., & Summers, T. P. (1991). Development of
a stress-related health promotion model. Antecedents and organizational
effectiveness outcomes. Journal of Social Behavior and Personality, 6(7), 141-162.
Henningfield, J. E., Ramstrom, L. M., Giorino, G., Barling, J., Weber, c.,
Kelloway, E. K., Strecher, V. J., & Jarvis, M. J. (1994). Smoking and the workplace: Realities and solutions. Journal of Smoking Related Diseases, 261-270.
Holcomb, H. S., & Meiggs, J. W. (1972). Medical absenteeism among cigarette, and
cigar, and pipe smokers. Archives of Environmental Health, 25, 295-300.
Hunter, J. E., Schmidt, F. L., & Jackson, G. B. (1982). Meta-analysis: Cumulating
research findings across studies. Beverly Hills: Sage.
Jackson, S. E., Cheneweth, D., & Glover, E. D. (1989). Study indicates smoking
cessation improves workplace absenteeism rates. Occupational Health and
Safety, 58(12), 13-15.
Janzon, L., Lindell, S., & TrelI, E. (1981). Smoking and disease: Attitudes and
knowledge in middle-aged men. Scandinavian Journal of Social Medicine, 9,
127-133.
Processes, 49,80-104.
Kark, J. D., Lebiush, M., & Rannon, L. (1982). Cigarette smoking as a risk factor for
epidemic A (h1n1) influenza in young men. New England Journal of Medicine,
307, 1042-1046.
Kozak, J. (1987). Absenteeism due to illness in smokers. Pneumonologia Polska, 55,
366-369.
Lowe, C. R. (1960). Smoking habits related to injury and absenteeism in industry.
British Journal of Preventative and Social Medicine, 14, 57-63.
178
E. Kevin Kelloway et at.
McMillan, G. H. (1981). The health of welders in Naval dockyards: Welding,

tobacco smoking and absence attributed to respiratory diseases. Journal of the
Society of Occupational Medicine, 31, 112-118.
Manning, M., Oland, J. S., & Osland, A (1989). Work-related consequences of
smoking cessation. Academy of Management Journal, 32, 606-62l.
Marti, B. (1986). Jogging is accompanied by low absenteeism. Sozial und
Preventivmedizin, 31, 313-318.
Qun, W. W., & Dobson, A J. (1992). Cigarette smoking and sick leave in an industrial population in Shanghai, China. International Journal of Epidemiology, 21,
293-297.
Ryan, J., Zwerling, c., & Orav, E. J. (1992). Occupational risks associated with
cigarette smoking: A prospective study. American Journal of Public Health, 82,
29-3l.
Smith, G. c., Athansou, J. A, Reid, C. c., Ng, T. K., & Ferguson, D. A. (1981).
Sickness absence, respiratory impairment and smoking in industry: An
Australian study. Medical Journal of Australia, 1,235-237.
Steers, R M., & Rhodes, S. R (1978). Major influences on employee attendance: A
Tsai, S. P., Cowles, S. R, & Ross, C. E. (1990). Smoking and morbidity frequency in
a working population. Journal of Occupational Medicine, 32, 245-249.
Wooden, M., & Bush, R (1995). Smoking cessation and absence from work.
Preventative Medicine, 24, 535-540.
9
Job Involvement and Absence
The Role of Constraints as Moderators
Constraint theory suggests that the relationship between work attitudes

and organizational behavior is influenced by a number of factors that are
present in the organizational environment. These factors can be subjective
or objective. They span both psychological and behavioral space Oohns,
1991). This study attempts to explore empirically the relationship between
a specific work attitude-job involvement-and a specific organizational
behavior-absence-in the presence of certain psychological and behavioral constraints.
Absence from work has occupied a central place in organizational
behavior research due to its theoretical and practical significance. Johns
and Nicholson (1982) suggest that the reasons for absence are multifaceted and located at the individual, group, and organizational levels.
Reviews on absence have corroborated this observation (Hackett &
Guion, 1985; Johns, 1997). Theoretical attempts at modeling absence
behavior have identified the role of attitudes to be crucial in understanding absence (Brooke, 1986; Hackett, 1989; Harrison, 1995; Johns, 1997;
Nicholson, 1977). Thus, most empirical studies on absence have concentrated on job satisfaction as the key explanatory factor (Hackett, 1989;
Hackett & Guion, 1985; Scott & Taylor, 1985). A much-cited meta-analysis
VISHWANATH V. BABA Michael G. DeGroote School of Business, McMaster University,
MUHAMMAD JAMAL Department of
Hamilton, Ontario LB5 4M4, Canada
Management, Concordia University, Montreal, Quebec H3G 1MB, Canada
179
180
of the correlates of absence by Farrell and Stamm (1988) concluded that

among all of the psychological or attitudinal correlates, only job involvement showed consistently significant correlations with absenteeism in
all of the analyses. Despite its theoretical importance (Brooke, 1986;
Nicholson, 1977; Steers & Rhodes, 1984) and the consistency of empirical
findings (Farrell & Stamm, 1988), job involvement has become a forgotten
variable in understanding absence (Brooke, 1986; Cheloha & Farr, 1980,
Mathieu & Kohler, 1990). Job involvement refers to an individual's psychological identification with work. Although the concept has suffered
from surplus meaning in the past, recent empirical attempts have clarified
the nature of the concept and clearly identify the above definition as the
most viable (Blau, 1985). Therefore, it stands to reason that job-involved
people tend not to absent themselves from work. Empirical studies reporting correlations between job involvement and absence seem to support
this view, although they are by no means consistent either in strength or
pattern. A majority of the studies reported a negative relationship between
job involvement and absence (Blumberg, 1980; Cheloha & Farr, 1980; Saal,
1978) while a significant minority reported no relationship between the
two (Blau, 1986; Hammer, Landau, & Stern, 1981). Breaugh (1981) reported
a negative correlation between job involvement and frequency of absence
but found no relationship when the time lost measure of absence was used.
This equivocality in the empirical findings could be attributed to a
number of factors. First, it is likely that the relationship between job
involvement and absence may be contingent on certain characteristics
specific to the samples used and may be revealed only in a multisample
study. The presence or absence of union policies or organizational policies
or norms may have certain contingency effects on the relationship which
need to be controlled. Second, there may be other variables within the
sample that moderate the relationship between involvement and absence
which need to be incorporated in the study. Third, the possible presence
of an intervening variable between involvement and absence such as
attendance motivation as suggested by Nicholson (1977) might mediate
the proposed relationship. Fourth, there may be methodological problems
associated with measures of involvement and absence which are inconsistent and therefore not strictly comparable across studies. In fact, the
studies cited here used four different instruments of job involvement
although the Lodahl and Kejner (1965) scale was widely used in other
studies of job involvement (Rabinowitz & Hall, 1977). Similarly, although
absence was conceptualized as time lost, frequency of absence, voluntary
and involuntary absence, excused or unexcused absence, sickness, and
so on, the generally accepted notion is to use both time lost and frequency
of absence (Johns, 1997). Finally, it is also possible that the studies used
Constraints as Moderators
181
inappropriate data analytic techniques in dealing with absence data. It

has been pointed out in the literature that absence distribution has a lower
limit of zero for a large portion of the sample and for the rest, takes on a
wide range of values above the limit. In other words, absence data typically manifest a truncated normal distribution for which the traditional
ordinary least squares (OLS) techniques of prediction may not be appropriate (Baba, 1990; Hammer & Landau, 1981). Yet, all of the predictive
studies cited here used OLS regression for predicting absence.
The present study attempts to overcome some of the problems mentioned above. By focusing on a sample of nonunion professional employees in an organization with no formally articulated absence control policy,
the study allows for free play on absence behavior (Johns, 1991) while
benefiting from the homogeneity of the sample in terms of both similarity
of occupation and organizational socialization.
Nicholson's (1977) notion of attendance motivation as an intervening
variable between job involvement and absence was tested using a pathanalytic model during preliminary investigation and was not found to be
empirically viable.
The pattern of results in the literature seems to suggest that a moderator approach may be appropriate for the current investigation.
Constraint theory suggests that while there may be an attitudinal inclination to engage in or refrain from certain behavior, constraint factors present in the environment might alter the pattern of relationship between the
attitudinal inclination and the intended behavior (Johns, 1991). This study
attempts to identify such constraints and explore their effect on the relationship between job involvement and absence.
Theoretical models of absence have identified a number of both psychological and behavioral constraints on attendance (Johns, 1991, 1997).
However, except for the work of Blau (1986) there has been no systematic
attempt to establish theoretically relevant constraint factors that moderate
the relationship between job involvement and absence. The empirical literature suggests that in the absence of constraints, a job-involved person
is less inclined to be absent from work. However, certain psychological
constraints like stress and lack of job challenge may interfere with that
relationship, especially among professional employees. Despite the attitudinal inclination to attend work, a job-involved person under stress or
who finds his job less challenging may decide more readily to absent
himself from work thus weakening the negative relationship or, in
extreme cases, even altering the direction of the relationship. Theoretical
work on absence seems to support this inference (Johns, 1997). Further,
behavioral constraints such as time pressure due to family responsibilities
as well as group and organizational norm pressures might also moderate
182

Job Involvement
--------+.
Absence
-Time lost
-Frequency of Absence
Constraints
-Psychological
-Behavioral
Figure 1. Model of job involvement and absence with constraints as moderators.
the relationship between job involvement and absence though in opposite

directions. In other words, presence of time pressure tends to weaken the
inclination to attend while the presence of strong attendance norms may
enhance the negative relationship between job involvement and absence.
The Steers and Rhodes (1978) model of attendance adds confidence to this
line of reasoning. Figure 1 shows a model suggesting a negative relationship between job involvement and absence which is moderated by
psychological constraints such as stress and job challenge as well as
behavioral constraints such as time pressure and norm pressure.
In order to overcome some of the problems associated with previous
studies of job involvement and absence, this study will use the established
Lodahl and Kejner (1965) scale of job involvement, and two measures of
absence, time lost and frequency of absence. Further, to accommodate
criticisms of inappropriate methods used for predicting absence due to
the nature of its distribution, this study will attempt to compare traditional OLS technique with the recommended Tobit analysis (Baba, 1990;
Hammer & Landau, 1981; Leigh, 1986). It is believed that unless we make
the explicit comparison of the two techniques with the same data set, it is
difficult to draw meaningful inferences regarding the relative merits of
OLS versus Tobit.
METHOD
Sample
The sample consisted of 193 nonunionized professional employees
from a largo aerospace facility located in a northeastern metropolis in
Canada. The average respondent was a university-educated male, about
183
40 years of age, married with one or two children and has been with the
organization between 2 and 5 years holding a permanent job.
Procedure
Data were collected in the form of a mail-back questionnaire

with prepaid envelopes which were mailed to 650 employees using the
organization's mailing system. With one follow-up letter, 227 usable questionnaires were returned. Of these, 193 were from male professional
employees which were retained for further analysis. Of the 227 respondents, 96% provided their employee identification number so that matching absence data could be obtained from company records. Though
nonrespondents were not contacted to verify systematically any nonresponse bias, organizational records indicated that they were similar to the
respondents both in terms of their demographic characteristics as well as
absence behavior pattern.
Instruments
Demographic information concerning age, gender, education, marital

status, number of children, and tenure on the job were obtained using
appropriate questions. Job involvement was measured using the six-item
short form of the Lodahl and Kejner (1965) scale. The Cronbach alpha reliability was .73. Job challenge was measured by a three-item scale using
appropriate items from a larger scale developed by Beehr, Walsh, and
Taber (1976) to measure higher-order need strength. The original anchors
referring to degrees of importance were changed to degrees of agreement
to reflect accurately the concerns of this study. The Cronbach alpha reliability was .79. The responses for both job involvement and job challenge
were obtained in a five-point Likert-type format ranging from strongly
disagree to strongly agree with high scores indicating higher levels of
involvement and challenge, respectively. Stress was measured using a 30item modified and shortened Cornell Medical Index using both somatic
and psychosocial symptoms with a four-point Likert-type response format ranging from never to often. High scores indicated higher levels of
stress. The Cronbach alpha reliability was .88. Norm constraint was measured by posing questions pertaining to the respondents' immediate work
group attendance record as well as whether the supervisor was positively
impressed by good attendance records. Similarly, time pressure in the
non work environment was measured by asking the respondent about
both time pressure cognition and time pressure behavior. Examples used
included taking the car for repairs, going to the bank, waiting at home for
184
repair personnel, or taking a family member for medical appointments.

Responses were obtained in a three-point Likert-type format. Absence
data were obtained in terms of both total time lost in days and frequency
of absence referring to the number of absence occurrences from company
records for the preceding 9-month period. The long time period used for
tracking absence provides for stability of the measures. Though in a strict
sense such data render the study postdictive rather than predictive, it has
been shown that prior absence is indeed a strong predictor of future
absence (Breaugh, 1981; Johns, 1997; Keller, 1983) and studies that are
postdictive in nature have made a useful contribution to our understanding of absence (e.g., Youngblood, 1984).
Data Analysis
Data analysis included descriptive statistics, correlations, and moderated multiple regression using both the OLS and Tobit procedures.
Since the absence data were found to be skewed in the case of both time
lost (0.65) and frequency (1.11), it was felt that the Tobit procedure would
be more appropriate (Baba, 1990). Methodological literature on absence
argues that heteroscedasticity can severely affect the standard error of the
regression coefficients overestimating it in some regions and underestimating it in others (Baba, 1990). This is perhaps why significant regression
weights from the OLS technique are unstable across studies reported in
the literature. However, in the Tobit model, regression weights are estimated through a maximum likelihood estimate (MLE) technique which is
considered to provide more consistent, less biased, and efficient estimates
compared to the OLS technique by avoiding the pitfalls of the latter (Baba,
1990; Hammer & Landau, 1981; Leigh, 1986).
Though it has been customary in the economics literature to compare
the regression coefficients obtained from the OLS technique with those
obtained using the MLE technique, it was pointed out by Leigh (1986) that
it is more appropriate to compare the OLS weights (B) to the partial derivative of the Tobit estimate B. This study reports both estimates Band B for
Tobit as well as the conventional B value for the OLS.
RESULTS
Table 1 shows the descriptive statistics and the correlations among
the dependent, independent, and moderator variables. The correlations
are generally moderate to low which suggest their appropriateness for
moderator analysis.
2.1
1.6
2.7
1.4
2.4
0.9
0.8
1-5
1-4
1-5
1-3
1-3
1-3
1-3
Note. Decimals are omitted from the correlation coefficients.

For r ~ 14, P oS .05.
For r ~ 22, P oS .001.
Job involvement (JI)

Stress (S)
Challenge (C)
Work group attendance (WA)
Supervisor's impression (SI)
Time pressure cognition (TC)
Time pressure behavior (TB)
Range
0.77
0.32
1.1
0.53
0.63
0.69
0.50
SD
42
10
23
-15
-23
-01
Job
involvement
-28
23
-32
-12
-08
16
21
Time lost
(M = 4.0, SD = 5.4)
Correlations (r)
-14
-04
22
26
-34
11
-34
Frequency
(M = 2.5, SD = 2.2)
Table 1. Descriptive Statistics and Correlations among Independent, Dependent, and Moderator Variables
(J1
.....
ex>
;jl
0"
o
c..
~
....
I),>
<Il
I),>
<Il
<Il
;;r
::s
n
o
186
Table 2 provides the results of the moderator analysis using both the
OLS and Tobit techniques. The results indicate substantial support for the
model linking job involvement and absence with both psychological and
behavioral constraints as moderators. Job involvement was a significant
predictor of both time lost and frequency of absence explaining 7.5% of
the variance (OLS) and 7.4% (Tobit) of time lost and 11.4% (OLS) and
12.2% (Tobit) of the variance in frequency of absence.
Time Lost
When the time lost measure of absence was used as a criterion, felt
stress significantly constrained the negative relationship between job
involvement and time lost providing positive regression coefficients for
both OLS and Tobit models. Similarly, job challenge acted as a significant
negative constraint on absence when interacting with job involvement
enhancing the negative relationship between job involvement and
absence. Work group norms of attendance likewise acted as a negative
constraint on absence behavior when interacting with job involvement.
Similarly, organizational norm measured through supervisor's positive
impression and emphasis of attendance acted as a significant negative
constraint on absence as indicated by the negative regression coefficients
and significant B values for both OLS and Tobit models. Further, both
time pressure cognition and time pressure behavior significantly constrained the observed negative relationship between job involvement and
time lost, encouraging absence behavior even among the job-involved.
However, this moderator effect was supported only in the OLS model.
In addition to the moderator effects mentioned above, stress, challenge, and time pressure behavior showed significant main effects on time
lost though they were all weaker compared to their moderator effects
except in the case of stress.
Frequency of Absence
When frequency of absence was used as the criterion measure, the
empirical support for the model was not as strong as in the case of time
lost. Significant constraining effects on the relationship between job
involvement and frequency of absence were observed for job challenge,
work group attendance, and time pressure cognition for both the OLS and
Tobit models in the same direction as in time lost. Time pressure behavior
showed significant moderator effect on the relationship between job
involvement and frequency of absence only for the OLS model which is
consistent with the previous observation for this variable. Both stress and
R2
B
R2
B
SE
R2
B
B
SE
R2
0.80
0.84 13.4
2.3
9.7
11.4
1.5
1.6
0.71
9.1
9.7
0.98 0.57
0.76
2.7
2.0
2.2
1.7
1.8
2.2
1.5
1.3
-2.0 -1.8
1.2
1.5
1.0
1.0
1.0
13.4
10.5
0.88 11.6
9.3
2.7
0.65 0.34 16.7
1.7
1.9
2.7
0.85 0.32 14.9
2.4
0.44
1.1
0.75
0.67
0.35 0.43 17.2
1.0
2.7
2.1
0.60 0.35 18.6

0.85 0.40 16.1
2.3
0.53 0.29 15.6
0.7 -0.28 -0.22 0.39 12.5 -0.7
0.77 0.28 17.9
0.55 0.23 14.3
11.8
2.2
1.7
1.8
0.96 10.2 -1.9 -0.20 0.31
0.88 0.71
-1.2
-2.1 -2.5 -1.7 1.03 10.1 -2.5 -0.97 0.34 18.4 -2.8 -1.3 -1.0 0.41 17.1 -3.1
-0.3 -0.12 -0.08 0.61 7.5 -0.2
0.14 0.26 11.6
0.5 -0.07
0.06 0.25 12.2 -0.3
-0.09 0.28 12.2 -0.4
Note. t-values 2: 1.9 are significant at .05 level. Brefers to the Tobit estimated coefficient. B refers to the regression coefficient in OLS and the
estimated partial derivative at the mean in Tobit.
Time pressure
behavior (TB)
JI X TB
Time pressure
cognition (TC)
JI X TC
-1.5
-1.8 0.87 10.7

-0.18 0.65 7.6
7.4 -0.4 -0.45 0.30 12.5 -1.5 -0.11
-1.3 -0.28 -0.19 0.69
8.4
-0.95 0.75
Workgroup
attendance (WA)
JI X WA
Supervisor's
impression (SI)
JI X SI
0.64 7.4 -4.1 -0.97 0.21 11.4 -4.7 -1.2 -0.95 0.26 12.2 -4.7
2.9
1.5 12.8
0.76 0.50 12.6
1.5
0.88
0.69 0.64 13.5
1.4
2.0 15.2
2.0
1.0 0.62 13.9
1.7
0.91 0.82 14.8
1.4
1.2
1.3
1.4
0.53
SE
Tobit
-1.8 -0.93 -0.63 0.49 9.6 -1.9 -0.24 0.16 12.5 -1.5 -0.30 -0.24 0.20 13.5 -1.5
-3.1 -1.4 -0.97 0.54 14.0 -2.6 -0.50 0.18 16.3 -2.8 -0.51 -0.41 0.22 16.9 -2.3
-1.8
3.1
2.7
OLS
-0.74 0.40 9.3

-1.4 0.44 14.1
0.50 0.75 -3.8 -2.6

1.2 12.5
3.1
4.4
1.5 15.3
2.4
3.9
SE
Tobit
Frequency
Challenge (C)
JI X C
Job involvement (JI) -1.9

Stress (S)
3.8
JIX HS
3.6
OLS
Time lost
Table 2. Moderator Analysis of Job Involvement-Absence Relationship
(Xl
'-l
....
0'
"'
'""'
Q.
'"
'"~
0
::I
-['""'
Ii
188
supervisor's impression failed to act as significant constraints of the relationship between job involvement and absence frequency. In addition,
both time pressure cognition and time pressure behavior showed significant positive main effects on absence frequency.
When the OL5 and Tobit models are compared in terms of their overall performance, their relative merits appear to be marginal at best. When
the regression coefficients are compared, Tobit seems to provide stronger
estim~tes at first glance though they vanish when the partial derivatives
from B are used in comparison. Out of 13 possible predictions of time lost,
9 were significant with the OL5 model while Tobit produced 8 significant
predictions. In terms of explained variance in time lost, OL5 estimates
were higher than Tobit in 6 cases while Tobit estimates were higher than
OL5 in 6 other cases. In 1 case, they were both identical. Maximum
increase in explained variance in time lost (.1R2) remained the same at
7.8% for both OL5 and Tobit model.
In the case of absence frequency, out of a possible 13 predictions, the
OL5 model produced 7 significant predictions while Tobit produced 6.
However, in terms of explained variance, 11 out of 13 times, the Tobit
model yielded higher R2 compared to OL5. Maximum increase in
explained variance in absence frequency (aR2) was 7% for the OL5 model
and 6.4% for Tobit.
DISCUSSION
The results provided substantial support to the notion that psychological and behavioral constraints moderate the relationship between job
involvement and absence (Johns, 1991). Absence literature suggests that
time lost measures reflect involuntary absence due, for example, to sickness or other calamities where an individual may have been absent for a
long duration though such occurrences may not be frequent at all
(Chadwick-Jones, Nicholson, & Brown, 1982). In other words, the individual may not have much control over his or her absence in this case. On
the other hand, the impact of such incidents are minimal when one deals
with frequency of absence occurrences. The degree of control an individual has over his or her absence is higher in this case compared to the former. Therefore, frequency of absence has been used as a measure of
voluntary absence as the term is referred to in the literature (ChadwickJones et aI., 1982). The differences in the pattern of results obtained for
time lost and absence frequency seem to underscore the theoretical
importance of differentiating between involuntary and voluntary absence
(Chadwick-Jones et aI., 1982; Johns, 1997). Time lost measure of absence
189
yielded better predictions in terms of significance compared to the frequency measure though variances explained were higher for the frequency measure. If the time lost measure is viewed as an indication of
involuntary absence, the results seem to suggest that constraints playa
more significant role in bringing about involuntary behavior than voluntary behavior. In other words, when a job-involved person encounters a
constraint to his behavior, it appears to bring about some form of involuntary behavior more than voluntary behavior. Simply put, constraints
seem to weaken the control an individual has over his behavior. Removal
of such constraints may indeed facilitate the free impact of attitude on
behavior, which in this case refers to the negative impact of job involvement on absence, especially if the latter is involuntary. It appears that the
general tenets of the path goal theory of leadership (House & Mitchell,
1974) apply to this situation in terms of designing an appropriate intervention strategy.
Specific results obtained appear to support these general observations. For example, while stress seems to interact significantly with job
involvement in influencing time lost, it had no impact, either directly or
as a moderator, on frequency of absence. Thus, the conceptual distinction
between involuntary and voluntary absence becomes empirically meaningful. Similarly, supervisor's positive impression or emphasis of good
attendance has no effect on the employee's voluntary absence behavior
simply because the employee is in control of his absence behavior and
perhaps makes compensatory accommodations for such impressions
either prior or subsequent to the occurrence of absence. The employee is
in control of his absence behavior in this case including its repercussions.
As reported earlier, stress and time pressure show a direct effect on
absence in addition to their moderator effects. These findings are supported in the literature on stress-induced absence (Johns, 1997) as well as
in the Steers and Rhodes (1978) model of attendance. In other words, they
influence the absence behavior of even those who are not involved in their
job which is understandable, but parenthetical to the concerns of this
study.
This study made an explicit attempt to compare the OLS model with
Tobit in exploring the relationship between job involvement and absence
with certain constraints as moderators. Contrary to the findings reported
by Price and Mueller (1986), the number of significant predictors actually
diminished in this study when the Tobit model was used. Since the Price
and Mueller (1986) study did not report either the regression coefficients
or the variances explained in the two models, other systematic comparisons could not be made. Based on the results of this study, one cannot
argue in favor of one model over the other. However, it seems clear that
190
the choice of Tobit over OLS depends on the degree of truncation exhibited by the distribution and its measure of skewness. As reported earlier,
the degree of skewness for absence frequency (1.12) was roughly twice as
high as for time lost (0.65). The Tobit model seems to perform better in the
case of absence frequency where in 11 out of 13 cases, the variance
explained was higher for Tobit. On the other hand, both OLS and Tobit
models provided comparable performance for time lost. This observation
is in line with the reasons provided by Hammer and Landau (1981) favoring the use of Tobit when the distribution is highly skewed. Therefore, it
is suggested that the researcher examine the nature of distribution of
absence data and choose either OLS or Tobit for data analysis depending
on the degree of truncation and the measures of skewness obtained for
that distribution.
In conclusion, it is clear that a job-involved professional male will be
less absent from work. However, he is also sensitive to stress, job challenge,
attendance norms, and nonwork time pressure in terms of his absence
behavior. Specifically, a job-involved person under stress will tend to absent
himself more than the one with less stress. Similarly, a job-involved person
with a challenging job tends to absent himself less than one with a less challenging job. Further, where attendance norms are strong, a job-involved
person tends to absent himself less compared to a situation where attendance norms are weak. Finally, a job-involved person with less nonwork
pressures tends to absent himself less than one with more time pressure.
Thus, a low-stress environment, challenging job, strong attendance norm,
and supportive nonwork environment will enhance attendance among jobinvolved people. Future research must concentrate on how such interventions can be effected as well as explore the actual processes by which
constraints influence both attitude and behavior.
ACKNOWLEDGMENTS. Financial support for this study from FCAR
(OO-ER-0506) and SSHRC (410-99-1491) is gratefully acknowledged. The
authors are thankful for the critical comments on an earlier draft by Gary
Johns and Willi Wiesner.
REFERENCES
Baba, V. V. (1990). Methodological issues in modeling absence: A comparison of
least squares and tobit analysis. Journal of Applied Psychology, 75, 428-432.
Beehr, T. A., Walsh, J. T., & Taber, T. D. (1976). Relationship of stress to individually and organizationally valued states: Higher order needs as a moderator.
191
Blau, G. J. (1985). A multiple study investigation of the dimensionality of job

involvement. Journal of Vocational Behavior, 27, 19-36.
Blau, G. J. (1986). Job involvement and organizational commitment as interactive
Blumberg, M. (1980). Job switching in autonomous work groups: An exploratory
study in a Pennsylvania coal mine. Academy of Management Journal, 23,
287-306.
Breaugh, J. A. (1981). Predicting absenteeism from prior absenteeism and work
attitudes. Journal of Applied Psychology, 66, 535-560.
Brooke, P. P., Jr. (1986). Beyond the Steers and Rhodes model of employee attendance. Academy of Management Review, 11,345-361.
Chadwick-Jones, J. K., Nicholson, N., & Brown, C. (1982). Social psychology of absenteeism. New York: Praeger.
Cheloha, R. S., & Farr, J. L. (1980). Absenteeism, job involvement and job satisfaction in an organizational setting. Journal of Applied Psychology, 65, 467-473.
Farrell, D., & Stamm, C. L. (1988). Meta-analysis of the correlates of employee
absence. Human Relations, 41, 211-227.
Hackett, R. D. (1989). Work attitudes and employee absenteeism: A synthesis of
the literature. Journal of Occupational Psychology, 62, (in press).
Hackett, R. D., & Guion, R. M. (1985). A reevaluation of the absenteeism-job satisfaction relationship. Organizational Behavior and Human Decision Processes,
25, 340--38l.
Hammer, T. H., & Landau, J. C. (1981). Methodological issues in the use of absence
data. Journal of Applied Psychology, 16,574-581.
Hammer, T. H., Landau, J. c., & Stern, R. N. (1981). Absenteeism when workers
have a voice: The case of employee ownership. Journal of Applied Psychology,
66,561-573.
House, R. J., & Mitchell, T. R. (1974). Path-goal theory of leadership. Journal of
Contemporary Business, Autumn, 81-97.
Processes, 49, 80-104.
Johns, G. (1997). Contemporary research on absence from work: Correlates, causes
and consequences. In C. I. Cooper & I. T. Robertson (Eds.), International review
of industrial and organizational psychology (Vol. 12, pp. 115-173). New York: Wiley.
theory and research. Research in Organizational Behavior, 4, 127-172.
Keller, R. T. (1983). Predicting absenteeism from prior absenteeism, attitudinal factors and nonattitudinal factors. Journal of Applied Psychology, 68, 536-540.
Leigh, J. P. (1986). Correlates of absence from work due to illness. Human Relations,
39,81-100.
Lodahl, T. M., & Kejner, M. (1965). The definition and measurement of job involvement. Journal of Applied Psychology, 49, 24-33.
192
Mathieu, J. E., & Kohler, S. S. (1990). A test of the interactive effects of organizational commitment and job involvement on various types of absence. Journal
Nicholson, N. (1977). Absence behavior and attendance motivation: A conceptual
synthesis. Journal of Management Studies, 14, 231-252.
Price, J. L., & Mueller, C. W. (1986). Absenteeism and turnover of hospital employees.
Greenwich, CT: JAI Press.
Rabinowitz,S., & Hall, D. T. (1977). Organizational research on job involvement.
Saal, F. E. (1978). Job involvement: A multivariate approach. Journal of Applied
Scott, K. D., & Taylor, G. S. (1985). An examination of conflicting findings on the
relationship between job satisfaction and absenteeism: A meta-analysis.
Steers, R. M., & Rhodes, S. R. (1978). Major influences on employee attendance:
A process model. Journal of Applied Psychology, 63, 391-407.
Steers, R. M., & Rhodes, S. R. (1984). Knowledge and speculation about absenteeism. In P .S. Goodman & R. S. Atkin (Eds.), Absenteeism (pp. 229-275).
Youngblood, S. A. (1984). Work, nonwork and withdrawal. Journal of Applied
10
The Timing of Thinking about
Quitting
The Effect on Job Attitudes and Behaviors
Kathy A. Hanisch
The decision to voluntarily quit or remain in a job is one that employees

make on a regular basis. The decision process can range from a job-related
event that causes an individual to resign immediately to one that is agonized about for several years; in limited cases it can be an impulsive act.
Within a turnover decision framework, individuals engage in the cognitive process of thinking about quitting which mayor may not translate
into quitting; it may also translate into alternative forms of withdrawal
(see Mobley, 1977, and Hanisch, 1995a). What occurs between the time
individuals think about quitting until their resignation date? Do individuals who think about quitting for long periods of time prior to quitting
have different attitudes and engage in different behaviors than those
employees who think about quitting and then shortly thereafter resign?
Are those individuals with a longer period of time between thinking
about quitting and quitting more disruptive to an organization than those
with a shorter time between their thoughts of quitting and resignation?
This research evaluated whether the decision lag was related to behaviors
employees engage in prior to exiting. It also examined how the timing of
the first thoughts of quitting relates to employees' attitudes in an organization prior to resignation.
KATHY A. HANISCH Iowa State University, Ames, Iowa 50011
193
194
Kathy A. Hanisch
Mobley (1977) described and charted the employee turnover process

arguing from Porter and Steers (1973) that we need to understand the
process of withdrawal and how the decision to terminate is made. The
intermediate steps as presented by Mobley (1977) include the following:
evaluation of existing job, experienced job satisfaction/dissatisfaction,
thinking of quitting, evaluation of expected utility of search and cost
of quitting, intention to search for alternatives, search for alternatives,
evaluation of alternatives, comparison of alternatives versus present job,
intention to quit/stay and quit/stay. There are myriad feedback loops
throughout the process and Mobley (1977) suggests that in addition to
quitting employees might select other forms of withdrawal in an effort to
deal with or manage their dissatisfaction. He suggests that research is
needed on how the expression of withdrawal behavior changes as a function of time and of changes in or reevaluation of the environment.
Research has supported a heuristic model based on Mobley's (1977)
turnover model (e.g., Mobley, Horner, & Hollingsworth, 1978). The
heuristic model posits that thinking of quitting is the most probable outcome of job dissatisfaction. Mobley et al. (1978) found support for this
portion of their model and acknowledge that this finding has received
support from others (e.g., Atkinson & Lefferts, 1972). They argue that to
obtain a fuller understanding of the withdrawal process, we must consider cognitive and behavioral phenomena in addition to affective variables (e.g., job satisfaction, organizational commitment). Somers (1996)
concurs that "it is not unreasonable to suggest that physical attachment is
accompanied by psychological attachment such that when the latter deteriorates, the former occurs."
Lee and Mitchell (1994) proposed the unfolding model of voluntary
turnover as a reaction to their tiring of studies not focusing on the psychological process of quitting. The major components in their model of
voluntary turnover include the "shocks to the system" and the amount of
psychological analysis that precedes a decision to quit and the act of quitting. They argue that the psychological processes can vary from a highly
rational comparison of alternatives to a quick judgment unfettered by
multiple considerations. The process of voluntarily leaving unfolds over
time, but the speed and reasons for leaving differ across employees. Some
employees have a precipitating event, some decide quickly, and some
search for alternatives. Lee, Mitchell, Wise, and Fireman (1996) encourage
future research on an individual's first thoughts about quitting and making the decision to quit. They suggest that this systematic difference in
time across the decision path may constitute a new and interesting
research direction.
Thinking about Quitting
195
LITERATURE REVIEW
Although no articles or empirical studies were located that have
addressed the issue dealt with in this chapter, a significant amount of
research is available that 0) has evaluated the relations among attitude
measures, turnover, and other withdrawal behaviors, (2) proposes the
relations among withdrawal behaviors, and (3) evaluates the different
withdrawal models oftentimes comparing stayers and leavers on relevant
attitude and behavior variables. Each will be briefly examined below.
JOB ATTITUDES AND TURNOVER
The relationship between turnover and attitudinal variables has been
examined repeatedly since Taft and Mullins (946) included dissatisfaction in their reasons for turnover. Most researchers find a significant negative association between quitting and attitudinal measures generally
focusing on job satisfaction and organizational commitment. For example,
George and Jones (1996) found a significant correlation between job satisfaction and turnover of - .36. Job satisfaction and organizational commitment were both significant variables in a regression analysis predicting
intent to leave in a study by Rosin and Korabik (1995). In a sample of
nurses, Eberhadt, Pooyan, and Moser (1995) reported a correlation of - .50
between intent to quit and job satisfaction. It should be noted that
some studies (e.g., Good, Page, & Young, 1996) did not find significant
associations between intent to leave and job satisfaction. Sometimes other
alternative attitudinal measures (e.g., organizational commitment) are
related to intent to leave or quitting.
Researchers have also found notable decrements in job satisfaction and
commitment shortly before turnover occurred (e.g., Youngblood, Mobley, &
Meglino, 1983). Job satisfaction was found to be a stronger predictor of
withdrawal cognitions/intentions than organizational commitment (Tett &
Meyer, 1993). Mobley (1977) found high negative correlations between satisfaction and frequency of thinking of quitting. Atkinson and Lefferts (1972)
found that the frequency of thinking about quitting was significantly
related to actual quitting.
ORGANIZATIONAL WITHDRAWAL
Organizational withdrawal was defined by Hanisch (1995b:604) as a
construct denoting "various behaviors employees engage in to remove
196
Kathy A. Hanisch
themselves from their job or avoid their work." In addition to turnover

and absenteeism-two frequently studied forms of withdrawal-being
late, leaving work early, playing computer games or surfing the Web on
work time, making excuses to go somewhere to get out of work, frequent
trips to the restroom, chatting with co-workers, missing meetings, and
drinking alcohol at work are examples of other organizational withdrawal behaviors (Fisher & Locke, 1992; Hanisch, 1995a; Hanisch &
Hulin, 1990, 1991). The antecedents of these behaviors have received considerable attention particularly with regard to the behaviors of absenteeism and quitting. Among the reasons are dissatisfactions with work,
supervision, and co-workers (see Hanisch, Hulin, & Roznowski, 1998,
for an overview of the relations among job attitudes and withdrawal
behaviors).
Models of Withdrawal Behaviors
There are several theories or models of what the relations among
withdrawal behaviors may be like. Each model hypothesizes different
functions of enacted withdrawal behaviors and implies different structures among the withdrawal behaviors. They include the Independent
Forms of Withdrawal Model, Compensatory Behaviors Withdrawal
Model, the Progression of Withdrawal Model, the Spillover Model of
Withdrawal, and the Alternate Forms of Withdrawal Model. In the original description of all of these models, two or three behaviors were the
focus: Absenteeism, quitting, and occasionally lateness. Since the descriptions of these theories or models, possible additional withdrawal behaviors have been identified and include, for example, leaving work early,
early retirement, and playing computer games at work (Hanisch, 1995a;
Hanisch & Hulin, 1990, 1991). Below I present a brief description of each
of the withdrawal models. For more specific details on them, see Hanisch,
Hulin, and Seitz (1996) and the references presented below to the original
authors of the models.
Independent Forms of Withdrawal Model. In the original statement
of the Independent Forms of Withdrawal Model by March and Simon
(1958), absenteeism and quitting were assumed to be unrelated because
they have different functions and consequences for employees who quit
or are absent. March and Simon hypothesized that being absent from
work and quitting one's job should be psychologically and statistically
independent. The Independent Forms Model describes the process of
organizational withdrawal as follows: Individuals choose to withdraw
from an organization by being absent or by quitting. The behaviors may
197
have common antecedents, but they have different functions and consequences for the employees. March and Simon argued that dissatisfaction,
or motivation to withdraw, is "general and holds for both absences and
voluntary turnover" (p. 93). Absences are hypothesized to reduce negative job attitudes by removing employees temporarily from the stresses of
their day-to-day tasks. Quitting reduces stress from a specific job by
removing the individual permanently from the organization and required
work tasks. There are few empirical results that support this model
(Mowday & Spencer, 1981; Weiner, 1980) and it is generally not considered a likely candidate for the relationship among withdrawal behaviors
(Dalton & Todor, 1993).
Compensatory Behaviors Withdrawal Model. The Compensatory
Behaviors Withdrawal Model, first described by Hill and Trist (1955),
assumes that enacting one withdrawal behavior, in varying amounts (e.g.,
once, several times), in response to negative organizational attitudes or
work-related stress is sufficient to reduce job stress and dissatisfaction. All
withdrawal behaviors appear to have the same psychological function in
this model: Relieving job dissatisfaction and job stress. Any withdrawal
behavior, once enacted, compensates for other withdrawal behaviors
making them unnecessary. Hill and Trist (1955) observed that absences
were a characteristic of "stayers" rather than "leavers." Those who were
not absent were more likely to quit than those who had been absent in the
past. From this observation they concluded that absenteeism compensated for any job dissatisfaction that may have been experienced by the
"absence-prone" employees making it unnecessary for them to quit.
Dalton and Todor (1993) argue that one of the two prevailing views
regarding the relation between absenteeism and turnover is the adjustment model. This model suggests that absenteeism is a means to cope
with an unsatisfactory job and that it substitutes for turnover. Thus, being
absent makes one's job manageable and negates having to quit. This view
would predict the same negative relation between absenteeism and
turnover as the Compensatory Behaviors Model.
Progression of Withdrawal Model. Several articles contain elements of the Progression of Withdrawal Model. Baruch (1944) described
absences from work as "short-term quits." Herzberg, Mausner, Peterson,
and Capwell (1957) argued that the problems of absenteeism and
turnover should be considered jointly because "the small decision which
is taken when the worker absents himself is a miniature version of the
important decision he makes when he quits his job." Melbin (1961)
argued that leaving a job is the outcome of a chain of experiences leading
198
Kathy A. Hanisch
to a final break with an organization. "High absenteeism (lateness and

absence) appears to be an earlier sign, and turnover (quitting and being
fired) the dying stage of a long and lively process of leaving" (p. 15).
The Progression of Withdrawal Model specifies that individuals
enact, in response to negative job attitudes, a stochastically ordered
sequence of withdrawal responses. These responses range from minor
acts, such as daydreaming or frequent trips to water fountains and restrooms, to lateness or leaving work early, to absences from meetings, to
abuse of sick and personal leave, and finally to more major withdrawal
behaviors such as unexcused absences, quitting, or early retirement.
A generic ordering of these withdrawal behaviors in terms of severity is
implied in this list. The order would likely vary depending on organizational differences that summarize technological and work force differences, job classifications within organizations, and perhaps individual
differences such as gender, age, or years of organizational service. As with
the other withdrawal models, it is assumed that each of the withdrawal
behaviors enacted by employees has a feedback effect onto their job
attitudes and withdrawal tendencies. However, the Progression of
Withdrawal Model also assumes that the effects of the feedbackamounts by which job dissatisfaction and withdrawal tendencies are
decreased-will vary as a function of the degree of withdrawal or seriousness of the withdrawal behavior. Taking frequent trips to the water
fountain or long coffee breaks represents a less severe degree of withdrawal than being absent from work; enacting these behaviors should
have minor effects on job dissatisfaction or withdrawal tendencies compared to more serious forms of withdrawal such as absenteeism.
Dalton and Todor (1993) state the second of the two prevailing views
with regard to only absenteeism and turnover as the pain-avoidance view.
This view is that there is a positive relation between absenteeism and
turnover such that high absenteeism will lead to turnover. This view is
very similar to that espoused by the proponents of the Progression of
Withdrawal Model. The cause(s) of employees' absenteeism will eventually cause them to quit.
Spillover Model of Withdrawal. Beehr and Gupta (1978) argued

that withdrawal behaviors should be positively intercorrelated-there
should be a nucleus of withdrawal behaviors that occur in concert.
Aversive work situations should generate nonspecific negative attitudes
and nonspecific withdrawal tendencies that allow employees to avoid the
aversive parts of their work. The nonspecific avoidance tendencies
according to the Spillover Model then "spill over" from one withdrawal
behavior to another. Thus, an aversive work situation may generate
199
several different withdrawal behaviors at the same time or several

occurrences of the same behavior. The behaviors enacted depend on the
behavior-one can only quit a job once, for example-and situational
constraints imposed on one or some subset of the withdrawal behaviors.
This withdrawal model accounts for any spread of withdrawal
behaviors from the one initially enacted by mechanisms that function
independently of attitudinal antecedents that triggered the initial behavior. This process may reflect habituation or a spread of behavior effect.
Another descriptive metaphor for the Spillover Model is "behavioral diffusion" from one withdrawal behavior, once enacted, to other behaviors,
which occur without any necessary feedback effects on job attitudes.
Alternate Forms of Withdrawal Model. The Alternate Forms of

Withdrawal Model (Mobley, 1977; Rice & Trist, 1952) hypothesizes that
different withdrawal behaviors are substitutable. This model assumes
aversive work conditions, through their effects on job dissatisfaction, will
lead to organizational withdrawal. However, which specific withdrawal
behavior is enacted will be determined largely by external economic conditions, organizational sanctions for different behaviors, and reinforcement histories of individuals who have enacted different withdrawal
behaviors (Hanisch, 1995a). According to this model, when constraints are
placed on one withdrawal behavior, the probability of an alternate form
of withdrawal increases. The assumption of substitutability of behaviors
in the Alternate Forms Model is likely to be important only when a set of
events constrains an employee from enacting one or more withdrawal
behaviors that might otherwise have been selected.
For example, some jobs have severe sanctions on tardiness because
the output of one employee is the input for the next employee. These
long-linked technologies (e.g., assembly lines) require all employees to be
present during the times they are scheduled to work or face harsh consequences (e.g., termination). Under such conditions, tardiness, as a response
to aversive work conditions or job stress, might be enacted infrequently
even among those with very negative job attitudes. The negative sanctions
for tardiness may be greater than the relief the behavior provides from job
stress. Other forms of organizational withdrawal, as alternates to tardiness, such as missing meetings, absenteeism, or quitting, should increase
according to the Alternate Forms Model. Other organizations may place
severe sanctions on absenteeism suppressing it as an alternative or may
have benefit packages that encourage "early" retirement and increase the
likelihood of this form of organizational withdrawal. Some jobs, because of
their structure and constraints, make it virtually impossible to wander
around and look busy or miss meetings. In these jobs, some subsets of
200
Kathy A. Hanisch
withdrawal behaviors are unlikely, but other forms of withdrawal are

likely to be substituted for the constrained withdrawal behavior(s). This
model can be extremely difficult to evaluate because extensive details on
the organization and external environment are needed.
Summary of Withdrawal Models
All of these models have static and dynamic structural relations

among different withdrawal behaviors as their core outcomes. These
models have different implications for organizational practice in addition
to theoretical! structural differences. Empirical research on these models
is limited; there is little research on multiple forms of withdrawal within
single studies. Even fewer studies consider the role of time in the evaluation of withdrawal or whether and how time might impact the veracity of
the different models for individuals. The analyses completed here are
a first step toward considering and evaluating the role of time in the
withdrawal decision process.
EMPIRICAL RESEARCH RELEVANT TO THE
WITHDRAWAL MODELS
Several articles discuss the relationship among withdrawal behaviors

including Beehr and Gupta (1978), Benson and Pond (1987), Gupta and
Jenkins (1982), Keller (1984), and Lyons (1972). Waters and Roach (1971)
evaluating the relationship between frequency of absences and termination found a correlation of - .32 indicating that those who terminated had
more absences than those who did not terminate, providing limited support for the Progression of Withdrawal or Spillover Model. Both absences
( - .28) and termination (.24) were significantly correlated with work satisfaction; co-worker satisfaction (- .18) was significantly correlated with
absences in Waters and Roach's sample of 160 nonsupervisory female
employees. Keller (1984) found that low performance and high absenteeism were the most important variables used to explain turnover in his
study providing limited support for the Progression of Withdrawal or
Spillover Model. Fitzgibbons and Moch (1980) and Williams, Levy,
Silverstone, and Adams (1979) found absenteeism and turnover to be negatively related lending support to the Compensatory Behaviors Model.
Focusing on individual-level studies, Mitra, Jenkins, Douglas, and
Gupta (1992) in a meta-analysis with over 5300 data points present an
average corrected correlation of .33 between absence and turnover.
Lyons (1972) in a review of turnover and absenteeism found that in
201
10 samples, all of the people who left had significantly higher absenteeism
than those who stayed. Porter, Crampon, and Smith (1976) using a
matched group of recently hired stayers and leavers found no significant
differences with the attitudes of leavers and stayers from 2 to 6 months
prior to the time the leavers left the organization. Significant differences
were found for organizational commitment in the last month of employment of leavers. Job satisfaction was found to be one of the most important discriminators between stayers and leavers in a study conducted by
Gaertner and Nollen (1992). Gupta and Jenkins (1982) found that overall,
leavers had higher absence levels than stayers. Finally, Wolpin and Burke
(1985) evaluating voluntary turnover and voluntary absenteeism found
them to be positively correlated suggesting that voluntary leavers had
higher rates of voluntary absenteeism.
Current Study Variables
Based on previous calls for research on the withdrawal process, this

study systematically evaluates when individuals first proactively thought
about leaving their organization and whether timing from those thoughts
until termination date is related to their job attitudes and behaviors.
Specifically, the attitude and behavior dimensions for those who had been
thinking about quitting less than six months and those who had been
thinking about quitting more than six months were completed. This evaluation was done using discriminant function and canonical analyses
because the interest was in exploring the attitude and behavior variables
that differentiate between the two groups.
METHOD
Subjects
The participants in this study were individuals who had voluntarily

resigned from staff positions at a Midwestern, state university based on
information obtained from personnel records. Exit surveys were mailed in
two consecutive years with an overall response rate of 43% (42 and 43%
for each year). Most individuals completed their surveys within 1 to
2 months of their resignation. Of the 123 returned surveys, a total of
120 individuals provided usable data on the questionnaire for this study.
The sample consisted of 66 females and 54 males who were between
the ages of 24 and 60, and had been thinking about quitting their jobs for
less than 1 month to more than 8 years (mean time = 12.5 months). The
202
Kathy A. Hanisch
employees had worked for the university on average 4.7 years with a
range of 3 months to 25 years. Eighteen (15%) participants had worked
less than 1 year with 8 (8.3%) of those working less than 6 months. The
average salary was approximately $35,000 and 97% were Caucasian. The
individuals in this study had been employed in professional and scientific
positions (e.g., research associate, program coordinator, associate scientist) at the university.
Procedure and Questionnaire
The names of respondents were obtained from the university human
resources office. A letter and an exit survey with a preaddressed, postagepaid return envelope were sent to each individual on the list who had a
complete mailing address. The letter described the study and encouraged
participation. The author only contacted individuals; information shared
with the human resources office was in aggregate form only. Relevant sections of the exit survey included the demographic items described earlier
and the five scales from the Job Descriptive Index 001; Smith, Kendall, &
Hulin, 1969) assessing work satisfaction (18 items; e.g., "fascinating"),
supervision satisfaction (18 items; e.g., "hard to please"), satisfaction with
promotion (9 items; e.g., "opportunities somewhat limited"), co-worker
satisfaction (18 items; e.g., "stimulating"), and pay satisfaction (9 items;
e.g., "barely live on income"); all items have response scales of Yes, No,
or ? (i.e., cannot decide). The JOI has been used numerous times in past
research with excellent reliabilities and validities (e.g., Hanisch & Hulin,
1990, 1991); the developmental, psychometric properties, and scoring of
the index have been studied and the validity is well documented
(Hanisch, 1992; Smith et a1., 1969).
In addition to the satisfaction scales, the participants also reported
the frequency with which they engaged in 17 withdrawal behaviors (e.g.,
late to work, surfing the Web on work time) in their last year of employment at the university. Many of the behaviors used in this study were first
generated and reported in a study by Hanisch and Hulin (1990) and have
been used successfully in later research. Employees rated each behavior
using an 8-point scale (1 = Never in the past year, 8 = Once or more a day
in the past year) to indicate how often they had engaged in each behavior
in the past year.
RESULTS
Table 1 reports the correlations among all variables of interest in this

study. Reliabilities for the attitude and withdrawal scales are inserted in
203

Table 1. Correlations among Demographic, Attitude, and Withdrawal
Behavior Scales
(1) Gender"
(2) Age
(3) Salary
(4) Months employed
(5) Months thinking
about quitting
(6) Work satisfaction
(7) Pay satisfaction
(8) Co-worker
satisfaction
(9) Supervision
satisfaction
(10) Promotion
satisfaction
(11) Withdrawal
behaviors
.08
.08
.15
.01 -.08
.07 -.05
.05 -.02 -.25
-.02 -.17 -.18 -.08 -.19
.90
.19
.28
.86
.05
.91
-.14 -.05 -.37
.39
.21
.44
.88
.14 -.19 -.04 -.12 -.16
.26
0.26
.33
.19
.06
-.22
-.04
-.13
.37
.31
.21
.38
.28
.48
.07 -.10
-.01
-.02 -.04
.09
10
11
.81
.01 -.23 -.02 -.12 -.24 -.17 .79
Note. Correlations statistically significant (p < .05) are bold. Coefficient alpha underlined
and inserted in the main diagonal.
a Male = 1, female = 2.
the main diagonal and range from .79 to .91 indicating homogeneous and
consistent scales.
Many of the correlations shown in Table 1 are consistent with expectations. For example, gender is significantly correlated with salary indicating that men have higher salaries than women. Age, as expected, is
positively correlated with salary and months employed. It is also positively correlated with the number of months individuals thought about
resigning and negatively correlated with promotion satisfaction indicating that older individuals report being less satisfied with their promotion
opportunities and spend more time (in months) thinking about quitting
than younger individuals. Total months employed and months thinking
about resigning are both positively correlated with salary; an expected
finding. In addition, those who spent more time thinking about quitting
are also employed longer in this study.
Number of months thinking about quitting is negatively correlated
with pay satisfaction, supervision satisfaction, and co-worker satisfaction
indicating that these dissatisfactions were related to their thinking of quitting; its correlation with promotion satisfaction approached statistical significance (p < .10). Several of the attitude scales are intercorrelated (e.g.,
work satisfaction is positively correlated with all other satisfaction scales).
Work satisfaction and withdrawal behaviors are negatively correlated
Kathy A. Hanisch
204
(- .23) and supervision satisfaction is negatively correlated with withdrawal behaviors ( - .24). These results suggest that those who are less satisfied with their work and co-workers are more likely to engage in
withdrawal behaviors replicating the work of others discussed earlier
(e.g., George & Jones, 1996; Waters & Roach, 1971) and when absenteeism,
intent to quit, or quitting was used singularly as the withdrawal behavior.
Discriminant Analysis
The job attitudes and behaviors that might differentiate those who
had thought about their decision to quit longer than those whose cognitive processes about quitting were more recent were examined. Decision
time was determined based on each employee's response to the question,
"What date did you start thinking you might like to resign from your job
(e.g., such as looked at advertisements, sent out a resume, made phone
calls to friends at other organizations inquiring about jobs)?" The median
length of time reported was 6 months in this study; individuals with
values above the median were classified as longer decision, those with
scores below the median were classified as shorter decision. Discriminant
function analyses were completed using the five JDI job attitude scales
and the withdrawal behavior scale as the variables to classify the individuals into their decision time lag group. The discriminant function results
are shown in Table 2.
The attitudes and behaviors successfully discriminated between
those who had made their decision 6 months or more and less than
6 months prior to quitting as shown in the main diagonal of Table 2; individuals were classified correctly into their respective decision time lag
Table 2. Attitudes and Behaviors by Decision Time
Classification Results
Predicted group membership
Actual group membership
Less than 6 months
No. classified
% classified
Six months or greater
No. classified
% classified
Prior probability
Total
Less than
6 months
6 months
or greater
38
71.7
15
28.3
53
16
26.5
.51
54
35
73.5
.49
50
51
Total
104
205
group significantly more often than would be expected based on their

prior probabilities. The overall percentage of cases correctly classified
using the attitudes and behaviors was approximately 70% with correct
classifications by group equal to 71.7% (less than 6 months) and 73.5%
(6 months or more). Comparing these values to the chance value that
would maximize prediction (51%) and examining the corresponding
z-test (4.23, p < .05) indicates that the accuracy in classification was significantly greater than chance (Brown & Tinsley, 1983).
A canonical discriminant analysis was also completed for the classification of individuals based on their job attitudes and behaviors (see
Table 3). Function I was significant (p < .05) with a canonical correlation
of .34. ui for the discriminant analysis was.11 which indicates that 11 % of
the variance was accounted for by differences between the two groups on
Function I (Brown & Tinsley, 1983).
Table 3 presents the structure coefficients, standardized weights, standardized means, as well as variable means for the two groups on Function
I for the five attitudes and withdrawal behaviors. They are listed according
to the absolute size of the Function I structure coefficients; they range from
- .81 to .25. The largest structure coefficient for Function I is for withdrawal
behaviors indicating that withdrawal behaviors had the strongest relation
to the discriminating function. The other two large structure coefficients are
promotion satisfaction (.61) and co-worker satisfaction (.55). The standardized means and variables are higher (more positive) on all of the attitudinal
variables for those who had thought about quitting for less than 6 months
compared to those who had been thinking about it for 6 months or more.
Table 3. Structure Coefficients, Standardized Weights, Standardized Class
Means, and Variable Means for the Job Attitudes and Behaviors
Standardized means
Variable
Withdrawal
behaviors
Promotion
satisfaction
Co-worker
satisfaction
Supervision
satisfaction
Pay satisfaction
Work satisfaction
Variable means
Structure Standardized Less than 6 months Less than 6 months

coefficient
weight
6 months or greater 6 months or greater
-.81
-.71
.08
-.09
34.42
41.63
.61
.49
.21
-.22
7.17
4.84
.55
.41
.19
-.19
42.36
37.57
.35
-.15
.12
-.12
33.36
29.96
.33
.25
.17
-.28
.11
-.28
-.11
.29
14.15
40.64
12.43
38.45
206
Kathy A. Hanisch
The opposite was true for the withdrawal behaviors; the mean was higher
for those with a longer rather than shorter time lag indicating they reported
engaging in more withdrawal behaviors prior to quitting.
DISCUSSION
The discriminant function analysis completed for this study separated

those who had been thinking about quitting for 6 months or more from
those who had been thinking about it for less than 6 months based on their
frequency of engaging in withdrawal behaviors as well as promotion and
co-worker satisfaction. Those who had been thinking about quitting for
less than 6 months engaged in fewer withdrawal behaviors, were more satisfied with their promotion opportunities as well as their co-workers. They
were also higher on pay satisfaction and work satisfaction than those who
had been thinking about resigning for 6 months or more.
The research presented here suggests that the longer individuals
think about quitting before making the break from the organization has
detrimental consequences compared to those who leave within 6 months
of their first thoughts of leaving. Employees who thought about leaving
more than 6 months before their resignation date reported lower satisfactions on all measured attitudes compared to those who had thought about
leaving for less than 6 months. Individuals with the longer time lag were
also more likely to engage in withdrawal behaviors in their last year of
employment than those with a shorter time lag. In fact, withdrawal
behaviors were the best discriminator between the two groups.
These findings provide limited support for the Progression of
Withdrawal Model (or pain avoidance view) or the Spillover Model. Both
of these models predict a positive relation between absenteeism and quitting or in this case between withdrawal behaviors and quitting. This is
true in the last year of employment for both groups evaluated here.
Given previous as well as confirmed support in this study of the relation between job dissatisfaction and quitting, unhappy employees who
have their first thoughts of quitting longer than 6 months before they
leave the organization have negative attitudes plus frequently engage in
negative behaviors on the job such as missing meetings, being late to
work, or leaving work early. Thus, individuals who are unhappy longer
in the organization before they leave are costing the organization more
in their withdrawal behaviors as well as their negative attitudes than
those who leave within 6 months of their first thoughts of quitting.
Interestingly, when individuals in this study were asked if they behaved
differently on the job after their first thoughts of quitting, the typical
207
response was no and several said that they worked harder in an effort to
make the transition for their replacement easier. This was true regardless
of when they had their first thoughts of leaving.
Given this is the first study to address this issue, additional research
on this topic is needed. Specifically, evaluating individuals' thoughts of
leaving with their attitudes and behaviors over time in a longitudinal
evaluation would be useful to more fully explicate the process of leaving.
The average length of time employees in this study had thought about
leaving was about 12 months; it was serendipity that the question about
frequency of engaging in the withdrawal behaviors was framed with a
time reference of the last year of employment. However, several individuals had thought about leaving for longer than 12 months so the question
of their attitudes and behaviors since their first thoughts of quitting
would add to our understanding of the withdrawal process.
Additional research is needed on the models of withdrawal particularly to begin to discern which model is the most accurate description
under specific circumstances. As Wolpin and Burke (1985) have noted,
research supports at least three of the models with regard to the relation
between absenteeism and turnover. They posit this may be due to a
number of reasons: (1) measurement error, (2) third variables causing the
relationship, (3) population type, and (4) conceptual and operational definitions. All of these issues need to be addressed and carefully evaluated
in order to give us a better understanding of the withdrawal process and
the relations among the various withdrawal behaviors. The research area
becomes difficult to pursue once withdrawal behaviors are expanded to
include more than just the typically evaluated absenteeism and turnover.
However, as noted by Thurstone (1931) over 70 years ago, it is possible
and probable that two individuals with the same attitude may enact quite
different behaviors. This, in effect, argues for studying the multiple withdrawal behaviors employees have available to them and understanding
the relations and parameters that define the relations among the withdrawal behaviors. One model may work well for one subset of employees
while another may be better suited for another group. First thoughts of
quitting may be an important parameter to consider when conducting
further tests of the withdrawal models.
This study focused only on voluntary turnover decisions and as others have noted, it is important to distinguish between voluntary turnover
and voluntary forms of withdrawal versus involuntary forms of withdrawal (Benson & Pond, 1987; Wolpin & Burke, 1985). An employee who
is absent due to factors such as jury duty or a funeral would not be
viewed as withdrawing from the organization. An interesting extension
of this research would be to examine the timing of layoff decisions and
208
Kathy A. Hanisch
their effect on employees' organizational behaviors. Would the results be

similar to those found here or, because the decision to leave was not initiated by the employee, vary greatly from person to person?
Clearly the role of time is important in the turnover process and we,
as researchers, need to place more emphasis on time in our theories and
data collection (Katzell, 1994). When we consider time we must also
account for it. For example, further research to explicate whether the positive relation between withdrawal and first thoughts of quitting is, in fact,
supportive of the Progression of Withdrawal Model or whether it is an
artifact of time itself-with an increase in time, one can engage in more
withdrawal behaviors prior to quitting-would be useful. In addition,
those who leave shortly after their first thoughts of quitting may reflect
turnover as "shocks to the system" instead of as an orderly and progressive process.
This study evaluated employees' first thoughts of quitting and the
impact of this variable in the withdrawal process. Voluntary turnover has
and continues to be important for organizational managers in their pursuit of a successful organization. Regardless of the type of organization,
many individuals are leaving resulting in expensive costs to organizations
(Branham, 2001). Based on the results of this paper, the costs organizations must bear are even greater tkm normally considered and include
withdrawal behaviors (along with negative attitudes) that occur prior to
employees exiting the organization.
REFERENCES
Atkinson, T. J., & Lefferts, E. A. (1972). The prediction of turnover using
Herzberg's job satisfaction technique. Personnel Psychology, 25, 53-64.
Baruch, D. W. (1944). Why they terminate. Journal of Consulting Psychology,B,35-46.
Beehr, T. A., & Gupta, N. (1978). A note on the structure of employee withdrawal.

Benson, P. G., & Pond S. B., III. (1987). An investigation of the process of employee
withdrawal. Journal of Business and Psychology, I, 218-229.
Branham, L. (2001). Keeping the people who keep you in business. New York: American
Management Association.
Brown, M. T., & Tinsley, H. E. A. (1983). Discriminant analysis. Journal of Leisure
Research, 15(4), 290-310.

Dalton, D. R., & Todor, W. D. (1993). Turnover, transfer, absenteeism: An interdependent perspective. Journal of Management, 19(2), 193-219.
Eberhadt, B. J., Pooyan, A., & Moser, S. B. (1995). Moderators of the relationship
between job satisfaction and nurses' intention to quit. International Journal of
Organizational Analysis, 3, 394-406.
209
Fisher, C. D., & Locke, E. A (1992). The new look in job satisfaction research and theory. In C. J. Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction (pp. 165-194).
New York: Lexington.
Fitzgibbons, D., & Moch, M. (1980). Employee absenteeism: A multiple analysis
with replication. Organizational Behavior and Human Performance, 26, 349-372.
Gaertner, K. N., & Nollen, S. D. (1992). Turnover intentions and desire among
executives. Human Relations, 45, 447-465.
George, J. M., & Jones, G. R. (1996). The experience of work and turnover intentions: Interactive effects of value attainment, job satisfaction, and positive
mood. Journal of Applied Psychology, 81, 318-325.
Good, L. K., Page, T. J., & Young, C. E. (1996). Assessing hierarchical differences in
job-related attitudes and turnover among retail managers. Journal of Academy
of Marketing Science, 24, 148-156.
Gupta, N., & Jenkins, G. D., Jr. (1982). Absenteeism and turnover: Is there a progression? Journal of Management Studies, 19, 395-412.
Hanisch, K. A. (1992). The Job Descriptive Index revisited: Questions about the
question mark. Journal of Applied Psychology, 77(3), 377-382.
Hanisch, K. A. (1995a). Behavioral families and multiple causes: Matching the
complexity of responses to the complexity of antecedents. Current Directions
in Psychological Science, 4, 156-162.
Hanisch, K. A (1995b). Organizational withdrawal. In N. Nicholson (Ed.), The
Blackwell dictionary of organizational behavior (p. 604). Oxford: Blackwell.
Hanisch, K. A, & Hulin, C. L. (1990). Job attitudes and organizational withdrawal:
Hanisch, K. A., & Hulin, C. L. (1991). General attitudes and organizational
withdrawal: An evaluation of a causal model. Journal of Vocational Behavior, 39,
110-128.
Hanisch, K. A, Hulin, C. L., & Roznowski, M. (1998). The importance of individuals' repertoires of behaviors: The scientific appropriateness of studying multiple
behaviors and general attitudes. Journal of Organizational Behavior, 19,463-480.
Hanisch, K. A., Hulin, C. L., & Seitz, S. T. (1996). Mathematical/computational
modeling of organizational withdrawal processes: Benefits, methods, and
results. In G. R. Ferris (Ed.), Research in personnel and human resources management (Vo1.14, pp. 91-142). Greenwich, CT: JAI Press.
Herzberg, F., Mausner, B., Peterson, B., & Capwell, D. (1957). Job attitudes: Review
of research and opinion. Pittsburgh: Psychological Services of Pittsburgh.
Hill, J. M., & Trist, E. L. (1955). Changes in accidents and other absences with
length of service: A further study of their incidence and relation to each other
in an iron and steel works. Human Relations, 8,121-152.
Katzell, R. A (1994). Contemporary meta-trends in industrial and organizational
psychology. In H. C. Triandis, M. D. Dunnette, & L. M. Hough (Eds.),
Handbook of industrial and organizational psychology (2nd ed., Vol. 4, pp. 1-89).
Palo Alto, CA: Consulting Psychologists Press.
Keller, R. T. (1984). The role of performance and absenteeism in the prediction of
turnover. Academy of Management Journal, 27,176-183.
Kathy A. Hanisch
210
Lee, T. w., & Mitchell, T. R (1994). An alternative approach: The unfolding model
of employee turnover. Academy of Management Review, 19, 51-89.
Lee, T. w., Mitchell, T. R, Wise, L., & Fireman, S. (1996). An unfolding model of
voluntary employee turnover. Academy of Management Journal, 39, 5-36.
Lyons, T. F. (1972). Turnover and absenteeism: A review of relationships and
shared correlates. Personnel Psychology, 25, 271-281.
Melbin, M. (1961). Organizational practice and individual behavior: Absenteeism
among psychiatric aids. American Sociological Review, 26, 14-23.
Mitra, A., Jenkins, G., Douglas, & Gupta, N. (1992). A meta-analytic review of the
relationship between absence and turnover. Journal of Applied Psychology, 77,
879-889.
Mobley, W. H., Horner, S. 0., & Hollingsworth, A. T. (1978). An evaluation of precursors of hospital employee turnover. Journal of Applied Psycholof;>Y, 63,408-414.
Mowday, R T., & Spencer, D. G. (1981). The influence of task and personality characteristics on employee turnover and absenteeism incidents. Academy of
Management Journal, 24(3), 634-642.
Porter, L. W., Crampon, W. J., & Smith, F. J. (1976). Organizational commitment
and turnover: A longitudinal study. Organizational Behavior and Human
Performallce, 15, 87-98.
Porter, L. W., & Steers, R M (1973). Organizational work and personal factors in
employee turnover and absenteeism. Psychological Bulletin, 80,151-176.
Rice, A. K, & Trist, E. L. (1952). Institutional and subinstitutional determinants of
change in labor turnover. Human Relations,S, 347-372.
Rosin, H., & Korabik, K (1995). Organizational experiences and propensity to
leave: A multivariate investigation of men and women managers. Journal of
Vocational Behavior, 46,1-16.
Smith, P. c., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in
work and retirement: A strategy for the study of attitudes. Skokie, IL: Rand-McNally.
Somers, M. J. (1996). Modelling employee withdrawal behaviour over time: A
study of turnover using survival analysis. Journal of Occupational and
Organizational Psychology, 69, 315-326.
Taft, R, & Mullins, A. (1946). Who quits and why. Personnel Journal, 24, 300-307.
Tett, R P., & Meyer, J. P. (1993). Job satisfaction, organizational commitment,
turnover intention, and turnover: Path analyses based on meta-analytic findings. Personnel Psychology, 46, 259-293.
Thurstone, L. L. (1931). The measurement of social attitudes. Journal of Abnormal
Waters, L. K, & Roach, D. (1971). Relationship between job attitudes and two
forms of withdrawal from the work situation. Journal of Applied Psychology, 55,
92-94.
Weiner, N. (1980). Determinants and behavior consequences of pay satisfaction: A

comparison of two models. Personnel Psychology, 33, 341-357.
211
Williams, A., Livy, B., Silverstone, R., & Adams, P. (1979). Factors associated with
labour turnover among ancillary staff in two London hospitals. Journal of
Occupational Psychology, 52(1), 1-16.
Wolpin, J., & Burke, R. J. (1985). Relationships between absenteeism and turnover:
A function of the measures. Personnel Psychology, 38, 57-74.
Youngblood, S. A., Mobley, W. H., & Meglino, B. M. (1983). A longitudinal analysis of the turnover process. Journal of Applied Psychology, 68, 507-516.
Index
Absence
definitions
constitutive, 98-100
integrative, 120
operational, 100-104
frequency, 100, 186, 188
perceptions of one's, 102
Absence behaviors, recall of, 102-103
Absence legitimacy, at cultural and
individual levels, 24-25
Absenteeism, 22, 167-168; see also Smoking
voluntary, 147, 148
measures of, 142-143
Absenteeism questions, encoding of, 102
Accidents: see Illness-accident, personal
Alternate Forms of Withdrawal Model,
199-200
Appraisal review, formal, 88-89
Attendance behaviors for superiors and
subordinates, protocol of, 5-6
Attendance rules, 5-6
Attitudes, job, 144, 195, 203, 205
Compensatory Behaviors Withdrawal

Model, 197
Compressed workweek, 31
Computerized performance monitoring
(CPM),121-122
Constraint theory, 179, 181; see also Job
involvement and absence
Contract violation, 59
Control, internal-external locus of, 34-35
Cross-cultural differences: see National
culture
Cultural dimensions, 12; see also National
culture
Cultural persistence and adaptation
hypothesis, 43
Benefits, employee, 38-39

Burnout contagion, 60-61
Early return, 106-107

Employee assistance programs, 38-39
Employment prospects, alternative, 13; see
also Mobility
Empowerment: see also Powerlessness
determinants of, 77-79
in management practice, 77-80
Challenge (C), 182, 185-190

Chinese culture, 23-24
Co-worker satisfaction, 205
Coaching, on-the-job, 87-88
Collectivism, 12, 23-24
Commitment, organizational
and lateness, 147-154
types of, 76
and withdrawal behaviors, 76-77
Common method variance, 113-117
Companies, high-tech vs. low-tech, 10-11
Demographics
of composition of workforce, 36
and lateness, 149, 152
and withdrawal behavior, 203
Discipline, 9, 199
Distributed work, 119
Family-centered work ethic, 27

Family-work conflict, 136, 144, 147-154
Flexibility
of boundaries between work and
nonwork time, 33
organizational, 64-66
213
214
Flexible organization-home boundaries,
companies with, 17
Flexplace, 119-120
Flextime, 30-31, 120
Gender differences: see Demographics
Gender role differentiation, 35-38
Group cohesiveness, 9
Group-level lateness behavior, 8-10
Hazard rate, 101
Identification
in-group, 63
with organization, 86
Illness-accident, personal
and lateness, 145, 147-154
Illnesses, 39-40
perceived legitimacy of various, 24-25
Independent Forms of Withdrawal Model
196-197
'
Individualism, 12, 23
Information technology (IT), 119; see also
under Withdrawal behaviors,
measurement of
"Jarring events," 59
Job alienation, 75-77
Job challenge (C), 182, 185-189
Job design, 80-81, 83
Job involvement (JI), 26
and lateness, 147-154
Job involvement and absence, model of
with constraints as moderators, 180, 182,
185-190
Job satisfaction, 144, 147-151, 203-206; see
also Turnover
and job alternatives
in the context of progression models,
56-57
differential importance of, 57
Lateness
causes identified by managers, 13-15
costs, 1-2
definitions
constitutive, 104
integrative, 120
operational, 104-105
Index
Lateness (cont.)
empirical research on, since Blau, 1994,
162-163
increasing chronic, 135-136,147-156
new conceptualizations of, since Blau,
1994, 161-164
nonwork, 143, 147-154
sanctions on, 199
stable periodic, 136-137, 147-156
unavoidable/involuntary, 137, 154, 164
Lateness behavior, 133
measures, 139-142
multiple level antecedents, 15-18
extraorganizationallevel, 4, 10-13
group and organizational levels, 4, 810
individual level, 2-8
multiple-level model of, 2-3
taxonomy of, 134-135, 145-156
Lateness proneness, 104
Lateness propensity, 104
Lateness records, 120
Layoffs, 12, 106
Leaving work early, 143, 147-154
Leisure-income-trade-off, 136, 144
Locus of control, internal-external, 34-35
Mobility: see also Employment prospects
perceived ease of, viii
within-organization, 64-66
Modified Organizational Adaptation/
Withdrawal model, 73
Monitoring performance, 87, 121-122; see
also Performance Management
Process
Moves, internal/within-organization: see
Transfers
National culture, 11-12
effects on absenteeism, 21-24, 42-43
future theory and research, 40-42
work centrality and, 25-28
Native Americans, 23
"New Economy," vii
Organization-home boundaries, companies
with flexible, 17
Organizational flexibility, 64-66
Organizational records, 100-101, 120
Index
Pace of life, 33
Pay satisfaction, 205
Performance Management Process, 90-91
as antidote to withdrawal behavior, 8391
essential preconditions for, 85
and new work environment, 89-90
procedural steps, 84-89
Performance monitoring, computerized,
121-122
Performance reviews, 88-89
Performance standards, setting, 86
Personality traits and factors
and attendance behavior, 4
and turnover, 58
Political climate, 13
Power distance, 12
Powerlessness
contextual factors leading to, 80-83
organizational factors, 81-82
job alienation and, 75-77
Problem-solving approach, 88
Progression of Withdrawal Model, 197-198,
206
Promotion satisfaction, 205
Punctuality, culture and, 11-12
Quitting, 106; see also Turnover, voluntary;
Withdrawal behaviors, models of
timing of thinking about, 193-195, 203208
Relocation, 61-62
Resonance, 60
Retirement, 107
Reward systems, 9, 80, 82-83
Rewards, setting, 86
Satisfaction: see Job satisfaction
Saudi Arabia, 27
Schedules, work, 9-10, 29-32
Self-determination, 79
Self-efficacy, 79, 88
Sick pay, 39-40
Sickness: see Illnesses
Sickness insurance, 39-40
Smoking and absenteeism, 168-169, 171-176
international studies, of 172-173
Smoking cessation and absenteeism, 172173
215
Social support systems, 38-40
Societies; see also National culture
sequential vs. synchronic, 12
Spillover Model of Withdrawal, 198-199,
206
Stimulus event, 59
Stress (S), 182, 185-190
Student absenteeism, academic
achievement and, 27-28
Supervision satisfaction, 205
Supervisor's impression (SI), 185-190
Supervisory style, 80, 82
Teams, 14
Telecommuting, 122-123
Temporary help agency workers, 55
Termination, 106
Time loss, 100, 186
Time orientation, 28-34
Time pressure behavior (TB), 185-190
Time pressure cognition (TC), 185-190
Time urgency, 4-5, 33
Transfers, within-organization, 65, 66, 106
Transportation concerns, 144
Turnover, employee, 53-54; see also Job
satisfaction
changing meaning of, 54-56
constitutive definition, 106-107
difficulty disentangling voluntary from
involuntary, 61-62
without dissatisfaction, 58-62
dissatisfaction without, 58
impulsive / non -sa tisfa ction-dependen t,
58-59
incited by others' behavior, 59-61
and initial intention to leave, 63-64
intra- vs. inter-organizational, 65
job attitudes and, 195, 203, 205
material and psychological aspects, 5456
measures of, ]43
model(s) of, 194
unfolding, 59, 62-63, 194
need for multilevel outlook on, 66-67
operational definition, 107-108
organizational flexibility vs., 64-66
without prior search for alternatives, 6263
socioemotional implications, 65
voluntary, ]47-154; see also Quitting
216
Turnover, employee (cont.)
imposed, 61-62
Turnover propensity, short- and long-term,
106
Type A personality, and attendance
behavior, 4
Unemployment, 12
Vietnamese culture, 23
Weather, inclement, 144-145, 147-154
Withdrawal
as behavior, 113
as construct vs. process, 116
constructs overlapping with, 109-110
job, 23, 112, 113; see also Quitting
organizational,196-200
defined, 195-196
Withdrawal behaviors, 57, 71-72, 84, 205;
see also Empowerment; Performance
Management Process; Powerlessness
categories of, 73-74, 116
changes in constitutive meaning, 119-121
job involvement, job alienation, and, 7476
lateness, 1
measurement of, 95, 113-117, 123-124
changes in information technology
and, 118-124
outcome-based measures, 122-123
process-based measures, 121-122
self- and social-report measures, 115
Index
Withdrawal behaviors (cont.)
measurement of (cont.)
and time, 117-118
models of, 196-200, 206
empirical research relevant to, 200-201
multiple, 112-116
broad constructs as preferable to
narrow, 108-109
constitutive definition, 111-112
different goals, 110-111
narrow constructs as preferable to
broad, 109-110
operational definitions of, 112-116
changes in, 121-123
organizational commitment and, 76-77
single, 98, ]20; see also Absence;
Lateness; Turnover
temporary vs. permanent, 116
unified meaning, 116
Withdrawal research
constitutive definitions in, 96-97, 119-121
operational definitions in, 97-98
Work and nonwork time and activities,
28-33
flexibility of boundaries between, 33
Work attitudes, 144, 195, 203, 205
Work centrality, 25-28
Work-family conflict, 136,144,147-154
Work group attendance (WGA), 185-190
Work role, reducing/withholding inputs
from, 116
Work role withdrawal: see Withdrawal
Work satisfaction: see Job satisfaction

Voluntary Employee Withdrawal and Inattendance - A

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Voluntary Employee Withdrawal and Inattendance - A

Enviado por

Direitos autorais:

Formatos disponíveis

Voluntary Employee

INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY:

Springer Science+Business Media, LLC

2002 Springer Science+Business Media New York

When we started editing this volume, the world seemed to be a worker's

less frequently and less conspicuously than before so as not to jeopardize

turnover as part of an individual's preplanned career program, regardless

and tested a lateness taxonomy consisting of three types-increasing

Finally, in Chapter 10, Hanisch presents an empirical study that

National Culture and Perceptions of Absence Legitimacy

The Many Faces of Voluntary Employee Turnover:

Employee Withdrawal Behavior: Role of the Performance

Meaning and Measurement of Work Role Withdrawal:

Developing and Testing a Taxonomy of Lateness Behavior

New Conceptualizations of Lateness since Blau, 1994

Smoking and Absence from Work: A Quantitative Review

The Timing of Thinking about Quitting: The Effect on

Lateness behavior can be described as arriving at work after the scheduled

Abraham Sagie et al.

Figure 1. A multiple-level model of lateness behavior.

attributes, work-related situational factors, nonwork situational factors,

Abraham Sagie et al.

Table 1. Multiple-Level Antecedents of Employee Lateness

Rules and norms

Industry type, ownership

The subordinate's time is money; hence,

The superior's time costs more money;

In the case of a pending appointment,

The superior's right to initiate

Abraham Sagie et al.

relations, unfair supervision, low salary, boredom, chronic stress,

Abraham Sagie et at.

While discussing the antecedents of absence, Johns (1997) reviewed

As displayed in Fig. 1, the immediate level beyond that of the individual

more similar than dissimilar with regard to their lateness behaviors.

Abraham Sagie et al.

generally provides greater motivation for effective functioning at work and

high-tech industrial firms, tardiness is less tolerated in the low-tech

Abraham Sagie et at.

illustrate this point. A meeting of managers working for a multinational

Proposition 5: Extraorganizational factors, like industry type, ownership type,

Abraham Sagie et al.

1. Think of two employees in a team. What are

2. Think of employees in two different teams.

3. Think of employees in two different

4. Think of employees in two different

dimensions (e.g., punctuality) or situational variables (e.g., commuting) at

Up to this point the chapter has discussed parallel tracks of influence

Abraham Sagie et al.

The fact that multiple determinants at different levels of analysis may

This idea was suggested by an anonymous reviewer.

subsidiary than in neighboring local organizations, they would be more

Abraham Sagie et al.

including variables at the individual, group, organizational, and extra organizationallevels.

Abraham Sagie et al.

Helena Mensah Addae and Gary Johns

National Culture and Absence Legitimacy

Helena Mensah Addae and Gary Johns

the collective orientation of the Chinese was particularly conducive to the

A key dependent variable in the model of absence to be presented is

National Culture and Absence Legitimacy

There is a general recognition that work occupies a great deal of an

Helena Mensah Addae and Gary Johns