Você está na página 1de 34

The Relationships of Informal High Performance

Work Practices to Job Satisfaction and


Workplace Protability*
YOSHIO YANADORI and DANIELLE D. VAN JAARSVELD
Recent empirical evidence reveals considerable divergence between management
reports and employee reports regarding organizational high performance work
practices (HPWPs). This divergence implies that employees may not participate in
some HPWPs that are formally present in their organizations, but also, that
employees may participate in HPWPs that are not formally present in their organizations. In this study, we examine the implication of the latter case (i.e.,
employee participation in informal HPWPs) for employee-level and organization-level outcomes. Our analyses, using data from the Statistics Canada Workplace and Employee Survey, suggest that employee participation in informal
HPWPs is associated with enhanced job satisfaction and workplace protability in
a similar way as employee participation in formal HPWPs is associated with these
outcomes.

Introduction
Over the past two decades, many researchers in the human resource (HR)
eld have devoted considerable effort to understanding the relationship
between HR practices and organizational performance (Blasi and Kruse 2006).
These research endeavors helped to identify high performance work practices
(HPWPs), which consist of HR practices that improve employee human capital, motivate employees to use their human capital for organizational goals,
and provide employees with opportunities to do so (Batt 2002; Huselid 1995).
Empirical studies typically capture these practices as a set and develop an
index that measures the extensiveness of HPWPs. These studies consistently
* The authors afliations are, respectively, School of Management, The University of South Australia,
Adelaide, Australia. Email: yoshio.yanadori@unisa.edu.au; Sauder School of Business, University of British
Columbia, Vancouver, British Columbia, Canada. Email: vanjaarsveld@sauder.ubc.ca.
The authors would like to thank Ariel Avgar, John Benson, Alexander Colvin, and Carol Kulik for helpful
comments on earlier versions of this paper, and also Tina Morganella for her editorial assistance. While the
research and analysis are based on data from Statistics Canada, the opinions expressed do not represent the
views of Statistics Canada. Lee Grenon of Statistics Canada provided valuable assistance on the project.
INDUSTRIAL RELATIONS, Vol. 53, No. 3 (July 2014). 2014 Regents of the University of California
Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington
Road, Oxford, OX4 2DQ, UK.

501

502 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

nd that a higher score on these HPWP indices is associated with better organizational performance (Combs et al. 2006), and with positive individual-level
outcomes such as organizational commitment (Kehoe and Wright 2013), citizenship behavior (Godard 2001), and job satisfaction (Mohr and Zoghi 2008).
Empirical research on HPWPs generally relies on two information sources:
management and employees. The studies that use information collected from
management commonly ask management to report whether or not a specic
practice is present in their organizations (e.g., Huselid 1995; Walsworth and
Verma 2007). Other studies use information collected from employees. In
these studies, employees are asked to report the extent to which the practices
they participate in are aligned with high performance HR policies (e.g., Avgar,
Givan, and Liu 2011; Godard 2001; Liao et al. 2009). These two approaches
capture the extensiveness of HPWPs in organizations from different perspectives, examining the presence of formal practices from the view of management or examining employee participation in these practices.
Several researchers have acknowledged the divergence between the presence
of formal practices and employee participation in them (Arthur and Boyles
2007); i.e., the presence of formal practices does not necessarily lead to
employee participation in these practices. Some practices may not be implemented successfully and remain inaccessible to employees (Diez, Wilkinson,
and Redman 2009; Glew et al. 1995). These practices, if implemented, may
only be accessible to some employees because organizations often develop distinct HR practices for different employee groups (Lepak and Snell 1999).
We contend that divergence can also occur when employees participate in
HPWPs on an informal basis. That is, employees participate in workplace
activities even though their organizations do not formally adopt these activities
as their formal HR practices (Eaton 2003). Some activities may be spontaneous in the sense that line managers or employees initiate the practices without
their organizations formal endorsement. Indeed, by examining the effects of
employee participation in HPWPs on job satisfaction, Mohr and Zoghi (2008)
found that employees often report participating in HPWPs that do not formally
exist in their organizations.
Recognition that employees participate in HPWPs both on formal and informal bases raises the following question: Is employee participation in informal
HPWPs (i.e., HPWPs that are not formally present in organizations) associated
with positive employee-level and organization-level outcomes in a similar
manner as employee participation in formal HPWPs (i.e., HPWPs that are formally present in organizations) is associated with these outcomes? On the one
hand, despite their informal nature, employees still participate in practices that
are instrumental to the mechanisms by which positive employee-level and organization-level outcomes are stimulated. On the other hand, employee reactions

Informal HPWPs

/ 503

to HR practices may differ depending on how they interpret their organizations HR policies (Bowen and Ostroff 2004). An ambiguous link between
organizations formal policies and actual workplace practices may weaken the
positive impact of informal HPWPs on employee-level and organization-level
outcomes. To address this theoretical conict, this study investigates the implications of employee participation in informal HPWPs by exploring relationships to job satisfaction and workplace protability. In so doing, we compare
these relationships to the relationships of employee participation in formal
HPWPs to the same outcomes.
To achieve our goal, we undertake two analyses with data from the Statistics Canada Workplace and Employee Survey (WES). The WES consists of
two surveys with matching reports from management and employees located
in the same workplace: (1) the Workplace Survey asks management about
their workplaces with a primary emphasis on their work practices, and (2)
the Employee Survey asks employees about their degree of participation in
these practices. The matched design of WES facilitates an investigation of
divergence between management and employee reports. Comparing management reports to the WES Workplace Survey with employee reports to the
WES Employee Survey enables us to distinguish among different types of
HPWPs. Our particular interest is in distinguishing between (1) employee
participation in formal HPWPs (i.e., employees report participating in HPWPs
that their management reports are present in the workplace), and (2)
employee participation in informal HPWPs (i.e., employees report participating in HPWPs that their management reports are not formally present in the
workplace).
Our rst analysis investigates the relationship between informal HPWPs and
employee job satisfaction. Job satisfaction is a key employee-level reaction to
their work experiences that predicts turnover (Tett and Meyer 1993), absenteeism (Scott and Taylor 1985), and performance (Judge et al. 2001), as well as
workers subjective well-being (Bowling, Eschleman, and Wang 2010).
Although Mohr and Zoghi (2008) have already demonstrated that employee
participation in HPWPs is associated with higher job satisfaction using the
WES, they did not examine formal HPWPs and informal HPWPs separately,
obscuring the extent to which employee participation in informal HPWPs contributed to enhanced job satisfaction. The second analysis aggregates the
reports from employees to the workplace level, and examines the relationship
between informal HPWPs and workplace protability. Since the seminal work
of Huselid (1995) and MacDufe (1995), the impact of HPWPs on organizational performance has been the central focus of HPWP research. Yet, to our
knowledge, no research has considered the implications of employee participation in informal HPWPs for organizational performance.

504 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

Employee Participation in Informal HPWPs


Reviewing the HPWP literature reveals a reliance on two key informants:
management and employees. Many studies collecting information from management focus on the presence of HPWPs in organizations (e.g., Kato and
Morishima 2002; Walsworth and Verma 2007; Zatzick and Iverson 2006) and
develop a HPWP index by counting the number of practices that are present
in organizations. Studies based on information from employees typically ask
the degree to which they participate in various HPWPs in their workplaces
(e.g., Avgar, Givan, and Liu 2011; Bartel 2004; Preuss 2003). These studies
develop a HPWP index by averaging the scores of individual practices. While
both approaches have successfully demonstrated the positive impact of
HPWPs on individual-level and organization-level outcomes, recent evidence
reveals a divergence between management and employee reports collected
from the same workplace. In studying the effects of HPWPs on organizational
performance, a few recent studies (Liao et al. 2009; Messersmith, Patel, and
Lepak 2011; Takeuchi et al. 2007) derive separate HPWP indices based on
management reports and employee reports. The correlations between the two
resulting HPWP indices vary across studies (r = .39 in Liao et al. 2009; r =
.59 in Messersmith, Patel, and Lepak 2011; and r = .41 in Takeuchi et al.
2007) as does the treatment of the two types of HPWP indices. Takeuchi
et al. (2007) use the two HPWP indices separately when examining the effect
of HPWPs on organizational performance. They nd that the positive relationship between HPWPs and organizational performance remains consistent
regardless of the HPWP index used in the model. Liao et al. (2009) propose
that the management perspective about organizational work practices shapes
the employee perspective, which subsequently inuences individual performance. Finally, Messersmith, Patel, and Lepak (2011) use the relatively strong
correlation (r = .59) as evidence of consistency across management and
employee reports. Messersmith, Patel, and Lepak (2011) eventually rely on a
HPWP index based on management reports when examining the effects of
HPWPs.
Arguably, divergence between management and employee reports addressed
in these studies stems from the difference in the way management and
employees understand and evaluate their organizational work practices. The
management reports focus on the practices in the formal organizational structure, whereas the employee reports reect actual participation in these practices in their workplaces (Avgar, Givan, and Liu 2011; Kehoe and Wright
2013). Comparing the management and employee reports enables us to make
an inference about how a particular HPWP operates in organizations. Figure 1

Informal HPWPs

/ 505

illustrates four different types of HPWPs that emerge from our comparison of
management and employee reports.1
In Figure 1, the management reports are on the horizontal axis, and the
employee reports are on the vertical axis. Both the management and employee
reports consist of two response options: Yes and No. For the management
reports, Yes (or No) indicates that management recognizes the practice is present
(or not present) in the organization. For the employee reports, Yes (or No) indicates that employees participate in (or do not participate in) the practice. Simultaneous consideration of the management and employee reports on these questions
yields four categories. In Category 1, the practice is formally present in organizations and employees participate in the practice. In Category 2, the practice is not
formally present in organizations but employees somehow participate in the practice. We refer to these types of HPWPs as formal HPWPs and informal HPWPs,
respectively. In Category 3, the practice is formally present in organizations but
employees do not participate in the practice. We refer to this type of HPWPs as
unused HPWPs. Finally in Category 4, the practice is not formally present in the
workplace and employees do not participate in the practice. In this category, the
FIGURE 1
Management reports

Yes
(Employees participate in
the practice)
No
(Employees do not
participate in the practice)

Employee reports

Yes
(The practice is present in the
organization)

No
(The practice is not present in the
organization)

Category 1: Formal HPWPs

Category 2: Informal HPWPs

The practice is present in the organization and


employees participate in the practice.

The practice is not present in the organization


but employees somehow participate in the
practice.

Category 3: Unused HPWPs

Category 4: No HPWPs

The practice is present in the organization but


employees do not participate in the practice.

The practice is not present in the organization


and employees do not participate in the
practice.

We thank an anonymous reviewer for suggesting that we develop this gure.

506 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

practice does not exist formally or informally. We refer to this case as no


HPWPs. While management and employee reports agree in Category 1 (i.e., formal HPWPs) and Category 4 (i.e., no HPWPs), they diverge in both Category 2
(i.e., informal HPWPs) and Category 3 (i.e., unused HPWPs).
The focus of our study is on informal HPWPs; the divergence in Category 2.
This case occurs when employees participate in activities that are not structured
by senior management as formal organizational HR policies. Some line managers may voluntarily establish HPWPs that apply to their subordinates. For
example, in an organization without an ofcial information-sharing program,
some managers may frequently communicate with their employees regarding
the status of their organization, leading employees to perceive that they participate in an information-sharing program. Alternatively, some employees may
voluntarily initiate HPWPs. For example, employees may meet and discuss
improvements to work processes even if their organizations do not formally
adopt problem-solving teams or task teams. In some workplaces, production
workers often improve their skills by engaging in their tasks and obtaining
feedback from senior colleagues or supervisors. Such workers may recognize
that they are receiving on-the-job training (OJT) even in an organization that
lacks a formal OJT program. In sum, informal HPWPs are practices that stem
from the initiatives that exist outside of formal organizational HR policies.2
There are two opposing arguments about the implications of informal HPWPs
for employee-level and organization-level outcomes. On the one hand, empirical
evidence shows that employee experience with HR practices inuences their attitudes and behaviors (e.g., Meyer et al. 2002). With respect to the effects of
HPWPs on job satisfaction, Heckman and Oldhams (1976) job characteristics
model posits that what employees experience in their job has a direct impact on
their psychological state. Even on an informal basis, participating in various
HPWPs, particularly those that involve employee participation (e.g., suggestion
programs, self-directed teams), will be associated with employees positive evaluation of their work experiences. Furthermore, informal practices may be as inuential as formal practices in shaping employee interpretations of organizational

One may argue that informal HPWPs should also include practices that go above and beyond formal
HPWPs (e.g., line managers communicate about their organizations more frequently than what the formal
information-sharing program originally expects of them) because this component of practices is informal in
the sense that it is not prescribed by formal organizational policies. However, this study treats this case as
part of formal HPWPs following the specication in Figure 1 (i.e., the practice is present in organizations
and employees participate in the practice). From an empirical standpoint, we expect that this case (i.e., an
extended use of formal HPWPs) is positively associated with job satisfaction and workplace protability as
it takes full advantage of the positive effects of formally structured HPWPs. Our narrow denition of informal HPWPs offers a more conservative test for their relationships to job satisfaction and workplace
protability.

Informal HPWPs

/ 507

HR policies as these practices often reect organizations culture (Schneider,


Gunnarson, and Niles-Jolly 1994). Because HPWPs are thought to reect
managements willingness to invest in its employees knowledge and skills
(Takeuchi et al. 2007), employees may form a positive evaluation of their workplaces, leading to improved job satisfaction. With respect to the effects on organizational performance, to the extent that employees participate in activities,
informal HPWPs will still activate the mechanisms that positively impact organizational performance (i.e., improving employee human capital, motivating
employees to use their human capital for organizational goals, and providing
employees with opportunities to do so). Consequently, employee participation in
informal HPWPs may inuence job satisfaction and organizational performance
in a similar manner as employee participation in formal HPWPs.
On the other hand, some researchers claim that the way employees understand organizational HR policies may inuence their attitudinal and behavioral
reactions (Nishii, Lepak, and Schneider 2008). In theorizing the mechanisms
through which organizational HR systems enhance organizational performance,
Bowen and Ostroff (2004) argue that when HR systems communicate an
unambiguous message to their organizational members about expected workplace behaviors, these systems have a strong impact on members attitudes and
behaviors. Informal HPWPs may be inferior to formal HPWPs in this regard
because informal practices have relatively less visibility due to their informal
nature, and thus convey a more ambiguous message to employees about organizational HR policies. If line managers initiate informal HPWPs, the use of
these practices hinges on the discretion of these managers, affecting the consistency of the message employees receive. The positive relationship between
informal HPWPs and job satisfaction may be reduced by the inconsistency surrounding the organizations HR policy. At the organization-level, even if informal HPWPs may still stimulate positive employee behaviors, these behaviors
may not be adequately coordinated among employees to enhance organizations overall performance because employees have inconsistent understandings
about their organizations HR policies. Overall, the positive relationships of
employee participation in informal HPWPs to job satisfaction and organizational performance may be smaller in magnitude than the relationships of
employee participation in formal HPWPs. Recognizing these opposing views,
our study empirically examines the relationships of employee participation in
informal HPWPs to job satisfaction and workplace protability by comparing
them with the relationships of employee participation in formal HPWPs.
It is important to note that management and employee reports also diverge
in Category 3, when employees do not participate in HPWPs that are formally
present in organizations, leaving HPWPs unused. Arthur and Boyles (2007:
80) view this divergence as the gap between HR programs, which refer to

508 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

formal activities, and HR practices, which refer to the implementation and


actual employee participation. They argue that HR programs precede HR practices and the latter may diverge from the former depending on the success of program implementation. It is also possible that management strategically
differentiates HPWPs among employee groups considering the differences in the
expected contributions of different employee groups (Lepak and Snell 1999). As
a result, while a practice is present in the workplace and actually participated in
by some employees, it remains inaccessible to other employees. Furthermore,
even if the practices are accessible to employees, the degree of employee participation will also vary depending on employee abilities and motivation (Cooke,
Zeytinoglu, and Chowhan 2009; Pendleton and Robinson 2010).
The implications of these unused HPWPs for employee-level and organization-level outcomes will vary depending on the mechanisms through which this
type of HPWPs is developed. When a practice is not implemented successfully
in the workplace, employees do not participate in the practice. Consequently, the
expected mechanisms that positively stimulate employee job satisfaction (e.g.,
improving psychological states) or organizational performance (e.g., improving
employee human capital, motivating employees to use their human capital for
organizational goals, and providing employees with opportunities to do so) are
unlikely to be activated. Recognizing the disconnect between organizational formal HR policies and actual practices, employees may even lose trust in their
organization. However, when some employees not having access to HPWPs
reects the organizations strategic differentiation of HR practices among
employee groups, this arrangement may maximize the returns from HPWPs at
the organization-level by customizing organizational HR practices according to
employee groups, leading to improvements in organizational performance. Collectively, the effects of unused HPWPs on job satisfaction and organizational
performance involve multiple, possibly conicting, effects. Because the WES
Employee Survey does not allow us to identify the reason why employees do not
participate in a HPWP that is formally present in organizations, the analyses in
this study do not examine these effects separately. Rather we include a variable
that reects unused HPWPs in the statistical models as a control variable.

Relationship of Informal HPWPs to Job Satisfaction


We rst examine the relationship between employee participation in informal HPWPs and employee-level job satisfaction.
Data. We analyze WES, a national-level survey produced by Statistics Canada that explores a broad range of issues surrounding workplaces and employees
(Statistics Canada 2008). Its target population (in terms of workplaces) includes

Informal HPWPs

/ 509

all business locations operating in Canada.3 To ensure that the survey reects its
target population accurately, Statistics Canada uses stratied random sampling
(by industry, region, and workplace size) to build a sample of workplaces from
Canadas Business Register (Statistics Canada 2008). From this sample, Statistics
Canada contacts employees using lists provided by the workplaces. HR researchers have published several studies capitalizing on this high-quality data (e.g., Haines, Jalette, and Larose 2010; Mohr and Zoghi 2008; Zatzick and Iverson 2006).
The most recent wave of the WES was conducted in 2005 for the Employee
Survey (relevant to the employee participation in HPWPs) and in 2006 for the
Workplace Survey (relevant to the presence of formal HPWPs). To examine
the most recent information about employee participation in informal HPWPs
and its implications, our analysis matched the 2005 WES Workplace Survey
and Employee Survey.4 The 2005 wave of WES includes 24,197 employees
from 6693 workplaces. Response rates are 77.7 percent for the workplace and
81.2 percent for employees. We excluded employees from workplaces with ten
or fewer employees because the WES Workplace Survey does not collect
information about some work practices (e.g., information sharing, self-directed
teams) from these workplaces (Haines, Jalette, and Larose 2010; Mohr and
Zoghi 2008). We also excluded the reports from managerial employees to the
WES Employee Survey because HPWPs are primarily applied to nonmanagerial employees (Colvin, Batt, and Katz 2001). After removing observations with
unusable values (e.g., some employees answered no opinion regarding their
job satisfaction), the sample in this study consists of 17,697 nonmanagerial
employees from 4000 workplaces. Mean workplace size (measured by the
number of employees) is 47.720, with a median of 22.
Measures. Our independent variable is employee participation in informal
HPWPs. Mohr and Zoghis (2008) study, which builds on the WES Employee

The following workplaces are excluded: (1) those located in the Yukon, Nunavut, and the Northwest
Territories, and (2) those operating in crop production and animal production; shing, hunting and trapping,
private households, religious organizations and public administration (Statistics Canada 2008).
4
The analyses in this article employ a cross-sectional analysis. While the Statistics Canada WES administers multiple waves of surveys, construction of longitudinal data suffers from signicant attrition. For
instance, approximately 25 percent of employees who participated in the 2003 Employee Survey were not
included in the 2004 Employee Survey. With respect to the Workplace Survey, information relevant to
HPWPs is collected every other year (i.e., 1999, 2001, 2003, and 2005), and approximately 13 percent of
workplaces that reported to the 2003 Workplace Survey were not included in the 2005 Workplace Survey.
Our analysis reveals that this attrition is systematic. Employees with lower job satisfaction are more likely to
drop out from the next wave of the Employee Survey, and workplaces with lower prot are more likely to
drop out from the next wave of the Workplace Survey. Due to concerns that the analyses of these panel data
might produce biased estimates, we decided to focus on a cross-sectional analysis drawing on the most
recent WES data.

510 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

Survey, and Zatzick and Iversons (2006) study, which builds on the WES
Workplace Survey, guide our choice of HPWPs because both studies use WES
data and examine the effects of HPWPs. In reviewing these two studies, we
found disagreement regarding the selection of practices considered to be part of
the HPWP index. Mohr and Zoghi (2008) analyzed seven participatory practices
(i.e., employee surveys, suggestion programs, exible job design, information
sharing, task teams, problem-solving teams, and self-directed teams). Zatzick
and Iverson (2006) examined four participatory practices (i.e., exible job
designs, information sharing, problem-solving teams, and self-directed teams),
gainsharing, and off-the-job training (off-JT). Because we found that the practices included in these two studies are all aligned with high performance policies,
we combined the HR practices from both the Mohr and Zoghi (2008) HPWP
index and the Zatzick and Iverson (2006) HPWP index. Following the specication in the WES Workplace Survey, we regard employee surveys as part of suggestion programs. In addition, we included OJT because it aims to improve
employee human capital, which is one aspect of a high performance work policy
(Crook et al. 2011). Similarly, we combined gainsharing, prot sharing, and
stock awards to create a group of nancial incentive plans whose payment is tied
to group- or organization-wide performance, referred to as productivity-related
bonuses in this article. All of these nancial incentives induce employees to work
effectively for the success of organizations by aligning their interests with organizational goals. As a result, HPWPs in our analysis consist of the following nine
practices: (1) suggestion programs, (2) exible job designs, (3) information sharing, (4) task teams, (5) problem-solving teams, (6) self-directed teams, (7)
productivity-related bonuses, (8) off-JT, and (9) OJT.
Table 1 compares specic questions for these nine HPWPs in the 2005
WES Workplace Survey and Employee Survey. Note that the WES Employee
Survey asks about employee participation in employee surveys and suggestion
programs separately, although these two practices collectively constitute suggestion programs in the WES Workplace Survey. In our two analyses, we
combined employee reports of employee surveys and suggestion programs into
one item to make employee reports and management reports comparable to
each other. For other practices, questions in the WES Workplace Survey closely match questions in the WES Employee Survey.
Before distinguishing among the different types of HPWPs we describe in
Figure 1, we created HPWP indices that reect the presence of HPWPs (from
the management reports) and employee participation in HPWPs (from the
employee report), respectively. For the index that reects the presence of
HPWPs, we counted the number of HPWPs that are present according to the
2005 WES Workplace Survey. For suggestion programs, exible job design,
information sharing, task teams, problem-solving teams, and self-directed

18 A.

18 B.

18 C.

18 D.

Suggestion program

Flexible job design

Information sharing

Problem-solving teams

Practice

OF

SURVEY QUESTIONS USED


TO

Flexible job design: Includes job rotation, job


enrichment/redesign (broad job denitions), job
enrichment (increased skills, variety or autonomy of
work).
Information sharing with employees: For example,
with respect to rms performance, colleagues
wages, technological or organizational changes, etc.
This implies that employees can provide feedback on
policies.
Problem-solving teams: Responsibilities of teams are
limited to specic areas such as quality or work ow
(i.e., narrower range of responsibilities than F).

For non-managerial employees, which of the following


practices exist on a formal basis in your workplace?
(This applies to 18. A-F)
Employees suggestion program: Includes employee
survey feedback.

31 (f)

31 (d)

31 (c)

31 (b)

31 (a)

How frequently are you informed (through meetings,


newsletters, email, or Internet) about overall
workplace performance, changes to workplace
organization, or the implementation of new
technology?
How frequently do you participate in a team or circle
concerned with quality or work ow issues?

How frequently are you asked to complete employee


surveys?
How frequently do you participate in an employee
suggestion program or regular meetings in which you
offer suggestions to your superior regarding areas of
work that may need improvement?
How frequently do you participate in a job rotation or
cross-training program where you work or are trained
on a job with different duties that your regular jobs?

Employee participation in HPWPs


(Employee reports to the 2005 WES Employee
Survey)

DEFINE HIGH PERFORMANCE WORK PRACTICES

Presence of formal HPWPs


(Management reports to the 2005 WES Workplace
Survey)

WORDING

TABLE 1

Informal HPWPs
/ 511

18 F.

6 (a)

14 (a)

16 (c)

Self-directed teams

Productivity-related
bonuses

Off-the-job training
(Off-JT)

On-the-job training
(OJT)

Does your compensation system include the following


incentives? (A. Individual incentive systems, B.
Group incentive systems, C. Prot sharing plan, D.
Merit pay or skill-based pay, E. Employee stock
plans)
Between April 1, 2004 and March 31, 2005, did this
workplace pay for or provide any of the following
types of classroom job-related training? (1. No
classroom training, 2. Orientation for new employees,
3. Managerial/supervisory training, . . . 13. Literacy or
numeracy, 14 Other training.)
Between April 1, 2004 and March 31, 2005, did this
workplace pay for or provide any of the following
types of on-the-job training? (1. No on-the-job
training, 2. Orientation for new employees, 3.
Managerial/supervisory training, . . . 13. Literacy or
numeracy, 14. Other training.)

Joint labor-management committees: Include


nonlegislated joint labormanagement committees and
task teams that generally cover a broad range of
issues, yet tend to be consultative in nature.
Self-directed work groups: Semi-autonomous work
groups or mini-enterprise work groups that have a
high level of responsibility for a wide range of
decisions/issues.

In the past 12 months, have you received any informal


training related to your job (this is on-the-job
training)?

VAN

25 (d)

In the past 12 months, have you received any


classroom training related to your job?

How frequently are you part of a self-directed work


group (or semi-autonomous work group or minienterprise group) that has a high level of
responsibility for a particular product or service area?
In such systems, part of your pay is normally related
to group performance.
Did you receive any productivity-related bonuses,
prot-sharing or prot-related bonuses for that
period?

How frequently do you participate in a task team or


labor management committee that is concerned with a
broad range of workplace issues?

DANIELLE D.

25

36 (d)

31 (g)

31 (e)

Employee participation in HPWPs


(Employee reports to the 2005 WES Employee
Survey)

AND

NOTE: Numbers are the question numbers in the survey.

18 E.

Task teams

Practice

Presence of formal HPWPs


(Management reports to the 2005 WES Workplace
Survey)

TABLE 1 (Cont.)

512 / YOSHIO YANADORI,


JAARSVELD

Informal HPWPs

/ 513

teams, the survey asks management whether or not respective practices exist
in their workplaces on a formal basis. For productivity-related bonuses, the
survey asks management whether or not their workplaces have different types
of nancial incentive plans, and we considered the workplace as having productivity-related bonuses if management indicated that they adopt at least one
of the following plans: (1) group incentive systems including gainsharing, (2)
prot-sharing plans, and (3) employee stock plans. For off-JT and OJT, the
survey asks management whether or not their workplaces offer any types of
off-JT and OJT. We classied workplaces as having off-JT and OJT unless
management answered no classroom training and no on-the-job training,
respectively. Because HPWPs consist of nine practices, this variable can range
in value from zero to nine, and we divided this score by nine. We refer to this
index as management-reported HPWP index.
For the HPWP index that reects employee participation, we similarly counted
the number of practices each employee reports participating in. The 2005
Employee Survey asks employees to report their frequency of participation in
certain practices in the previous 12 months, except for productivity-related
bonuses, off-JT, and OJT. With regard to employee surveys, suggestion programs, exible job design, and information sharing, employees have three
response options: (1) never, (2) occasionally, and (3) frequently. For task teams,
problem-solving teams, and self-directed teams, employees have four response
options: (1) never, (2) occasionally, (3) frequently, and (4) always. We considered employees as participating in each practice unless they answered never.
Regarding the employee survey and suggestion program questions, which collectively constitute suggestion programs in the WES Workplace Survey, we considered employees as participating in suggestion programs unless they answered
never to both employee survey and suggestion program questions. For productivity-related bonuses, employees reported whether they received any productivity-related bonuses in the previous 12 months. We considered that employees
participated in productivity-related bonuses if they received any productivityrelated bonuses irrespective of the amount.5 For off-JT and OJT, employees

We acknowledge that management reports and employee reports regarding productivity-related bonuses
might not match well for two reasons. First, unlike the questions for other HPWPs, the WES Employee Survey does not provide a clear denition for productivity-related bonuses. Consequently, employee reports
regarding the receipt of productivity-related bonuses may exclude any of the practices we dene as productivity-related bonuses or include other incentives (e.g., individual incentives). Second, the fact that employees do not receive any productivity-related bonuses does not necessarily mean that these bonuses are not
applied to the employees. It may mean that workplaces apply productivity-related bonuses, but their employees do not receive payment because the workplaces (or employees) do not achieve performance goals. Based
on this concern, we ran analyses with HPWP indices that excluded productivity-related bonuses. The results
are essentially the same with our original analyses for job satisfaction and workplace protability.

514 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

reported whether or not they participated in the respective types of training programs in the previous 12 months. The number of practices an employee participated in ranges from zero to nine, and we divided the number by nine. We
hereinafter refer to this HPWP index as employee-reported HPWP index.
We then derived indices that reect different types of practices following
the model presented in Figure 1. We created a HPWP index that reects formal HPWPs (i.e., Category 1) by counting the number of practices which
management reports are present in their workplaces (according to the 2005
WES Workplace Survey) and the employee reports participating in (according
to the 2005 WES Employee Survey). This variable can also range in value
from zero to nine, and we divided the number by nine. We refer to this index
as formal HPWP index. Next, we created a HPWP index that reects informal
HPWPs (Category 2) by counting the number of practices which management
reports are not present in their workplaces, but that the employee reports participating in, and dividing the number by nine. We refer to this index as informal HPWP index. Finally, we also created a HPWP index that reects unused
HPWPs (i.e., Category 3) by counting the number of HPWPs the employee
reports not participating in although the management reports are present and
dividing the number by nine. We refer to this index as unused HPWP index.
We did not create an index for no HPWPs (Category 4) because this is the
base condition. The relationships among different HPWP indices we derived
are described as follows:
management-reported HPWP index formal HPWP index
unused HPWP index
employee-reported HPWP index formal HPWP index
informal HPWP index
We acknowledge that both management and employee reports to the surveys
involve errors and the discrepancy between management reports (i.e., presence
of the HPWP) and employee reports (i.e., participation in a HPWP) in part
may be attributable to these measurement errors; e.g., management is unfamiliar with their organizations HPWPs or employees misunderstand activities
involved in HPWPs. Yet, given the detailed explanations of HPWPs in respective questions, as shown in Table 1, it is unlikely that either employee or management reports are misguided. Statistics Canada also endeavors to ensure the
accuracy of management reports by taking various steps such as checking
irregular answers (Statistics Canada 2008). Collectively, we contend that
comparison between the management and employee reports provides us with
reasonably accurate information about employee participation in formal,
informal, and unused HPWPs.

Informal HPWPs

/ 515

Our dependent variable in this analysis is employee job satisfaction. The


2005 WES Employee Survey asks employees to report their overall job satisfaction (i.e., Considering all aspects of this job, how satised are you with
this job?). Employees were asked to indicate their job satisfaction level using
four response alternatives: (1) very satised, (2) satised, (3) dissatised, and
(4) very dissatised. We reverse-coded employee responses to this question so
that a higher value reects a higher level of job satisfaction. Research shows
that a single-item job satisfaction scale has good convergent validity to a multiple-item job satisfaction scale, particularly the one that focuses on overall job
satisfaction (Wanous, Reichers, and Hudy 1997). While bipolar scales like job
satisfaction typically include a neutral answer and two or three response
options on either side (Dillman, Smyth, and Christian 2009), the job satisfaction scale in the WES Employee Survey has four response options. Recognizing the possibility that our job satisfaction measure does not satisfy the
equidistant condition, we treated it as an ordered categorical variable.
In this analysis, we collected employee participation in HPWPs and job satisfaction information from the same survey wave (i.e., the 2005 WES
Employee Survey) because the framing of questions in the 2005 WES
Employee Survey already incorporates a temporal sequence. The job satisfaction question asks employees at the moment when they completed the survey,
while the HPWP questions ask employees to answer their participation in
HPWPs in the previous 12 months. Given our expectation that employee participation in HPWPs has a positive impact on job satisfaction, it is reasonable
to expect that employee participation in HPWPs in the previous 12 months is
associated with current job satisfaction. Guided by the lead of previous job satisfaction studies (e.g., Heywood, Siebert, and Wei 2002; Mohr and Zoghi
2008), we controlled for several variables representing employee demographic
characteristics (e.g., gender, education level, disability), job characteristics
(e.g., tenure, occupation, work hour), and workplace size. Table 2 describes
how we measured these variables.
Analysis. Because our job satisfaction variable is an ordered categorical
variable, we used ordinal logistic regression, rather than linear regression.6 We
used the employee survey weight provided in the WES Employee Survey.
Given the survey design (i.e., employees were clustered in workplaces that
were drawn from the stratied random sampling), we used the proc surveylogistic function in SAS 9.2 (Heeringa, West, and Berglund 2010; SAS Institute
2008). By specifying workplace as the cluster from which multiple employees
We also ran ordinary least squares regression using proc surveyreg. The results are substantively the
same.
6

516 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

TABLE 2
VARIABLE DESCRIPTION
Variables
Dependent variables
Job satisfaction
HPWP indices
Management-reported
HPWP index
Employee-reported
HPWP index
Formal HPWP index
Informal HPWP index
Unused HPWP index
Control variables
Female
Married
Have disabilities
University degree or
higher
Tenure
Work hours per
week
Union member
Wage rate
Professional workers
Technical workers
Clerical workers
Production workers
Workplace sizeb

FOR

ANALYSIS

ON

JOB SATISFACTION

Description
Job satisfaction (4 = very satised, 1 = very unsatised)a

Mean
3.191

Number of HPWPs management answers are present in its workplaces

.420

Number of HPWPs employees answer they engage in

.493

Number of HPWPs that management answers are present in its


workplaces and employees answer they engage in
Number of HPWPs that management answers are not present in its
workplaces but employees answer they engage in
Number of HPWPs that management answers are present in its
workplaces but employees answer they do not engage in

.228

Gender of the employee (1 = female; 0 = male)


Marital status of the employee (1 = legally married; 0 = otherwise)
Disability of the employee (1 = has disability; 0 = otherwise)
Education of the employee (1 = university degree or higher; 0 =
otherwise)
Length of tenure (in years) since the employee joined the workplace
Work hours per week
Union dummy (1 if the employee is a member of a union or covered by
a collective bargaining agreement; 0 otherwise)
Hourly wage before tax deduction
Professional job dummy (1 if the employee holds a professional job; 0
otherwise)
Technical job dummy (1 if the employee holds a technical job; 0
otherwise)
Clerical job dummy (1 if the employee holds a clerical job; 0 otherwise)
Production job dummy (1 if the employee holds a production job; 0
otherwise)
The number of employees in the workplace

.265
.192

.540
.516
.103
.198
8.795
35.909
.339
20.615
.218
.455
.169
.079
47.720

NOTE: N = 17,697 employees from 4000 workplaces. Calculated using survey employee weight except for workplace size,
which used workplace weight. The mean and median number of employee reports per workplace is 3.297 and 3, respectively.
a
Reverse-coded original responses where (4 = very dissatised, 1 = very satised).
b
Mean workplace size of 4000 workplaces used in this analysis. Calculated before applying log transformation. Median
workplace size is 22.

reports are drawn, this procedure calculates parameter estimates and standard
errors accounting for the hierarchical structure of the data (Cassell 2006).
While our theoretical argument predicts that employee participation in various
HPWPs inuences job satisfaction, reverse causality is also possible; that is,
employee job satisfaction inuences the level of employee participation in
HPWPs (Mohr and Zoghi 2008). We contend that reverse causality is more relevant to the distinction between formal HPWPs and unused HPWPs (i.e., whether

Informal HPWPs

/ 517

or not employees participate in a practice that is present in their organizations).


We posit that informal HPWPs that are initiated by line managers are unlikely to
be inuenced by employee job satisfaction; however, those that are initiated by
employees voluntary activities may be inuenced by employee satisfaction in
the sense that more satised employees may be willing to initiate such voluntary
activities and participate in them. Readers must be aware of the potential for
reverse causality problem when interpreting the results of this analysis.
Results. Table 2 presents the mean of the variables we used in this analysis. Note that the mean value of the informal HPWP index is not negligible
(.265), supporting our view that employees participate in HPWPs that are not
formally present in the workplace. Indeed, the mean value of the informal
HPWP index is larger than the mean value for the formal HPWP index (.228).
Thus, employees in our sample report participating in informal HPWPs to a
greater extent than formal HPWPs. This nding is not surprising when compared with Eaton (2003), whose respondents reported a greater number of
informal work-family practices than formal work-family practices in their
workplaces. Our preliminary analysis also revealed that these employees did
not participate in some HPWPs that were present in their workplaces, represented by the unused HPWP index (.192). As a result, the mean value of the
employee-reported HPWP index (.493) is slightly greater than the mean value
of the management-reported HPWP index (.420). Although not reported here
in the table, the correlation between the management-reported HPWP index
and employee-reported HPWP index is r = .22, lower than the correlations
reported in previous studies (i.e., r = .39 in Liao et al. 2009; r = .59 in Messersmith, Patel, and Lepak 2011; and r = .41 in Takeuchi et al. 2007).
We also ran an analysis of variance (ANOVA) on our HPWP indices,
and that in job satisfaction to see the degree of variation in these variables
between and across workplaces (Appendix A). As the results demonstrate,
workplace accounts for a large proportion of the variance of these HPWP indices. The results of the ANOVA are converted to the intraclass correlation
(ICC), which indicates the appropriateness of aggregating employee reports at
the workplace level (Klein et al. 2000; Wright et al. 2005).7 For all HPWP
indices, ICC(2) is higher than .60, the threshold recommended by Glick
(1985). These results justify our subsequent analysis on the workplace-level
implications of various HPWPs. In contrast, while job satisfaction is systematically different across workplaces (p <.001), its ICC (2) is .45, which is lower

ICC(2) = (MSB MWS)/MSB, where MSB is the mean square between groups (i.e., workplace in this
case), and MSW is the mean square within groups (i.e., employee in this case).
7

518 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

TABLE 3
RELATIONSHIP
Variables
Control variables
Female
Married
Have disabilities
University degree or higher
Tenure
Work hours per week
Union member
Wage rate
Professional workers
Technical workers
Clerical workers
Production workers
Workplace sizea

OF

HPWP INDICES

Model 1
.163*
(.074)
.266***
(.065)
.434**
(.142)
.144
(.087)
.005
(.003)
.001
(.004)
.043
(.081)
.023***
(.004)
.133
(.150)
.022
(.139)
.234
(.151)
.115
(.171)
.001
(.024)

HPWP indices
Management-reported
HPWP index
Employee-reported
HPWP index
Formal HPWP index

TO

JOB SATISFACTION

Model 2
.161*
(.074)
.266***
(.065)
.436**
(.142)
.143
(.087)
.005
(.003)
.001
(.004)
.047
(.081)
.022***
(.004)
.129
(.149)
.019
(.138)
.231
(.150)
.117
(.170)
.002
(.025)

Model 3
.137
(.075)
.257***
(.066)
.438**
(.143)
.145
(.086)
.005
(.003)
.004
(.004)
.006
(.081)
.019***
(.004)
.040
(.147)
.053
(.136)
.194
(.149)
.095
(.170)
.008
(.025)

.133
(.074)
.261***
(.066)
.440**
(.143)
.147
(.086)
.006
(.003)
.003
(.004)
.007
(.082)
.019***
(.004)
.042
(.146)
.042
(.136)
.205
(.149)
.091
(.170)
.002
(.026)

.070
(.161)
.981***
(.150)

Informal HPWP index


Unused HPWP index
Log likelihood ratio

Model 4

279343***

279844***

372147***

.922***
(.209)
.706***
(.211)
.574*
(.269)
384420***

NOTE: Standard errors are in parentheses. N = 17,697 employees from 4000 workplaces.

< 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001.


a
In natural logarithm.

than .60, supporting our analytical strategy to treat job satisfaction as an individual-level variable.
Table 3 displays the results of our ordinal logistic regression analyses. To
highlight how different HPWP indices are associated with job satisfaction, we

Informal HPWPs

/ 519

ran four models. In Model 1, we used only control variables. In Model 2, we


added the management-reported HPWP index to Model 1. In Model 3, we
substituted the employee-reported HPWP index for the management-reported
HPWP index. Finally, in Model 4, we substituted the formal HPWP index,
informal HPWP index, and unused HPWP index for the employee-reported
HPWP index.
Results show that the employee-reported HPWP index is positively and signicantly associated with job satisfaction (b = .981, p < .001, Model 3),
whereas the management-reported HPWP index is not (b = .070, p = .665,
Model 2). The signicant relationship between the employee-reported HPWP
index and job satisfaction is consistent with Mohr and Zoghis (2008) ndings.
We then turned to the relationship of decomposed HPWP indices to job satisfaction. As shown in Model 4, both the formal HPWP index and informal
HPWP index are positively associated with job satisfaction. Beta coefcients
are comparable between the formal HPWP index (b = .922, p < .001) and
informal HPWP index (b = .706, p < .001), and these two beta coefcients are
not statistically different (p = .321). The unused HPWP index is negatively
associated with job satisfaction. This explains the nonsignicant effect of the
management-reported HPWP index on job satisfaction in Model 2. Because
the management-reported HPWP index consists of the formal HPWP index
and the unused HPWP index, it appears that the negative effect of unused
HPWP index cancelled out the positive effect of the formal HPWP index. The
negative relationship between unused HPWP index and job satisfaction is
examined in the Discussion section.
Overall, we interpret these results as showing that HPWPs are likely to be
associated with increased job satisfaction when employees participate in these
practices, regardless of their formality. Participation in informal HPWPs is
associated with enhanced job satisfaction in a similar manner as participation
in formal HPWPs is associated with enhanced job satisfaction.

Relationship of Informal HPWPs to Workplace Protability


Next, we shift our unit of analysis from individual to workplace by examining the organization-level implications of employee participation in informal
HPWPs.
Data. Unlike the previous analysis on job satisfaction, workplace protability is captured for the entire survey year, so our model in this analysis
included a temporal sequence (i.e., HPWPs in year t inuences workplace
protability in year t + 1) to augment our causal inferences. As a result, the

520 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

information about HPWPs came from the 2005 WES Workplace Survey and
Employee Survey, while the information about workplace protability came
from the 2006 WES Workplace Survey. Consistent with the previous analysis,
we focused on workplaces that (1) have more than ten employees, and (2)
have a report to the Employee Survey from at least one nonmanagerial
employee. In addition, workplaces must (3) be in the private sector because
the current analysis examined protability, and (4) have participated in both
the 2005 and 2006 waves of the Workplace Survey.
Workplace is the unit of analysis in this analysis, and therefore, we aggregated individual employees reports to the 2005 WES Employee Survey to
derive HPWP indices that reect employee participation in HPWPs at the
workplace level. Because previous HPWP studies do not provide the denitive
guideline regarding the minimum number of employee reports needed for
aggregation, we created two samples with different criteria used in previously
published HPWP studies: (1) Sample 1 includes workplaces with reports from
two or more nonmanagerial employees (e.g., Takeuchi et al. 2007), and (2)
Sample 2 includes reports from three or more nonmanagerial employees (e.g.,
Jensen, Patel, and Messersmith 2013). The sample size is 3125 workplaces
and 2665 workplaces, respectively. Mean workplace size is 43.928 with a
median of 24 for Sample 1, and 49.159 with a median of 27 for Sample 2.
The mean number of employee reports per workplace is 3.631 with a median
of 3 for Sample 1, and 4.154 with a median of 4 for Sample 2.
Measures. We summarize the variables we used in this analysis in
Table 4. We used the same procedure to develop the management-reported
HPWP index as we did in the previous job satisfaction analysis. For the other
HPWP indices that incorporate employee participation (i.e., employee-reported
HPWP index, formal HPWP index, informal HPWP index, and unused HPWP
index), we aggregated these indices calculated at the individual level in the
previous analysis by averaging them for each workplace (Gardner, Wright, and
Moynihan 2011; Wright et al. 2005). Appendix A summarizes the ANOVA
results and corresponding ICC(2). ICC(2) for our HPWP indices in both samples is sufciently high to support our aggregation at the workplace level: .66
(Sample 1) and .65 (Sample 2) for employee-reported HPWP index, .93 (Sample 1) and .93 (Sample 2) for the formal HPWP index, .73 (Sample 1) and .72
(Sample 2) for the informal HPWP index, and .86 (Sample 1) and .85 (Sample
2) for the unused HPWP index. These statistics support our aggregation of
individual-level HPWP indices to the workplace level in this analysis.
Our dependent variable is workplace protability. To derive this measure,
we divide workplace prot (revenue expenditures) by the number of employees (Zatzick and Iverson 2006). For both revenue and expenditure, manage-

Informal HPWPs

/ 521

TABLE 4
VARIABLE DESCRIPTION

Variables
Dependent variables
Organizational protability
in 2006a
HPWP indices
Management-reported
HPWP index
Employee-reported
HPWP index
Formal HPWP index

Informal HPWP index

Unused HPWP index

Control variables
Proportion of female
employees
Proportion of part-time
employees
Union density
Workplace sizeb
Workplace age
Cost leadership strategy

Average wage levelc

AND

SUMMARY STATISTICS

FOR

ANALYSIS

ON

WORKPLACE PROFITABILITY

Description
Prot (= operating revenue operating
expenditure) divided by the number of
employees in 2006
Average number of HPWPs management
answers are present in its workplaces
Average number of HPWPs employees
answer they participate in
Average number of HPWPs that
management answers are present in its
workplaces and employees answer they
participate in
Average number of HPWPs that
management answers are not present in its
workplaces but employees answer they
participate in
Average number of HPWPs that
management answers are present in its
workplaces but employees answer they do
not participate in
Proportion of female employees in the
workplace
Proportion of part-time employees in the
workplace
Proportion of employees that are covered
by a collective agreement
Number of employees in the workplace
Length of years the workplace has been
located at the current address
Average of the answers to the questions
that ask the organizations strategic
orientation: (1) reducing labor costs and
(2) reducing other operating costs, using a
5-point scale (Cronbachs alpha = .85)
Average wage of the employees in the
workplaces (total payroll expenditure
divided by the number of employees)

Sample 1
Workplaces
with 2
or more
employee
reports

Sample 2
Workplaces
with 3
or more
employee
reports

Mean

Mean

6.803

6.847

.319

.325

.458

.463

.157

.163

.301

.300

.162

.162

.451

.463

.238

.241

.112

.131

43.928
18.096

49.159
19.292

3.997

4.014

10.291

10.305

522 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

TABLE 4 (Cont.)
Sample 1
Workplaces
with 2
or more
employee
reports

Sample 2
Workplaces
with 3
or more
employee
reports

Variables

Description

Mean

Mean

Organizational protability
(2005)a

Prot (= operating revenue operating


expenditure) divided by the number of
employees in 2005
13 industry dummy variables (e.g., labor
intensive manufacturing, construction,
communication and other utilities, nance
and insurance, business services)

7.678

7.579

Industry dummies

NOTE: N = 3125 for Sample 1 and 2665 for Sample 2. Calculated using survey workplace weight. The mean and median
number of employee reports per workplace is 3.631 and 3 for Sample 1 and 4.154 and 4 for Sample 2.
a
Hyperbolic sine function.
b
Calculated before applying log transformation. Median workplace size is 24 and 27, respectively.
c
In natural logarithm.

ment is asked to report respective dollar gures for the location. WES is a
workplace-level survey, and a workplace is often part of a larger corporation
(e.g., a manufacturing plant). We excluded observations if managements
answer about their workplaces revenue or expenditure included that for other
locations.8 While HPWPs may increase output by employees, they also
increase expenditures for these practices (e.g., training, bonus payment). Thus,
to determine whether organizations truly benet from HPWPs, performance
measures should take these expenditures into account (Chadwick, Ahn, and
Kwon 2012). Our preliminary analysis of workplace protability suggested
that it is highly skewed. Because protability can be positive or negative, we
applied a hyperbolic sine function (Burbidge, Magee, and Robb 1988). Similar
to a logarithmic transformation, the use of a hyperbolic sine function in regression models informs the percent change in the variable in question. Several
researchers have argued that this transformation is more appropriate when the
variable takes a negative value, which is problematic for logarithmic transformations (see Nyberg et al. [2010] for more discussion).9 Following the lead of
previous HPWP studies (e.g., Datta, Guthrie, and Wright 2005; Zatzick and
Iverson 2006), we controlled for several workplace-level variables including
workplace size, union density, and industry.
8
The WES Workplace Survey includes questions that ask whether the amount reported represents only
the gure for the workplace (e.g., 29(b) Does this amount represent the revenues for this location only?).
9
It is represented by the following equation: sinh-1(x) = log [x + (x2+1)1/2]

NOTE: Standard errors are in parentheses.

R2
Adjusted R2
N

Unused HPWP index

Informal HPWP index

Formal HPWP index

Employee-reported HPWP index

HPWP indices
Management-reported HPWP index

Organizational protability (2005)

Average wage level

Cost leadership strategy

Workplace age

Workplace size

Union density

Control variables
Proportion of female
employees
Proportion of part-time employees

Variables

< 0.10;

.307
.302
3,125
p < 0.05;

1.553**
(.568)
1.068
(.690)
1.510***
(.447)
.961***
(.157)
.005
(.007)
0.110
(.099)
1.233***
(.261)
.588***
(.017)

Model 1

OF

**

p < 0.01;

.307
.302

.767
(.537)

1.618**
(.570)
.994
(.692)
1.545***
(.448)
.900***
(.162)
.006
(.007)
.127
(.100)
1.185***
(.263)
.588***
(.017)

Model 2

***

1.875*
(.887)
1.633
(.954)
.787
(1.141)
.308
.303

1.640**
(.571)
1.024
(.693)
1.439**
(.451)
.883***
(.163)
.005
(.007)
.137
(.100)
1.098***
(.267)
.585***
(.017)

Model 4

.410
.406
2,665

.704
(.601)
.455
(.712)
2.257***
(.417)
1.238***
(.151)
.016*
(.007)
.043
(.105)
1.050***
(.274)
.663***
(.017)

Model 1

WORKPLACE PROFITABILITY

.412
.407

1.476**
(.546)

.792
(.601)
.530
(.711)
2.318***
(.417)
1.120***
(.157)
.015*
(.007)
.010
(.105)
.967***
(.276)
.662***
(.017)

Model 2

.414
.409

3.174***
(.818)

.858
(.600)
.256
(.712)
2.016***
(.421)
1.139***
(.153)
.018**
(.007)
.033
(.104)
.832**
(.279)
.661***
(.017)

Model 3

Sample 2 (Workplaces
with 3 or more employee reports)

p < 0.001. Industry dummies are included but not reported.

.308
.303

1.656*
(.757)

1.614**
(0.569)
1.077
(.690)
1.410**
(.449)
.920***
(.158)
.004
(.007)
.126
(.099)
1.123***
(.265)
.585***
(.017)

Model 3

TO

TABLE 5
HPWP INDICES

Sample 1 (Workplaces with


2 or more employee reports)

RELATIONSHIP

3.123***
(.899)
3.500***
(1.056)
2.275
(1.260)
.414
.409

.835
(.601)
.290
(.714)
2.072
(.424)
1.081
(.157)
.017*
(.007)
.009
(.105)
.815
(.279)
.658
(.017)

Model 4

Informal HPWPs
/ 523

524 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

Analysis. We used the ordinary least squares regression for this analysis.
When running the regression, we used the workplace survey weight provided
in the WES Workplace Survey. We recognized that some independent variables might be correlated with each other (e.g., proportion of part-time
employees and proportion of female employees). To assess whether multicollinearity problems were an issue in our analysis, we calculated the variance
ination factor (VIF) when we ran our regressions. The VIF in our regression
is lower than 3.0 for all independent variables, well below the threshold value
of 10.0 suggested by Neter, Wasserman, and Kutner (1985). Thus, multicollinearity is unlikely to be a concern in our analysis.
Similar to the analysis on job satisfaction, reverse causality is possible in
this analysis; that is, more protable workplaces may offer greater opportunities for their employees to participate in various HPWPs. We posit that the
reverse causality problem is more serious for the relationship between formal
HPWPs and workplace protability because more protable workplaces may
be willing to invest in various HPWPs. Yet, it may also be possible that
employees in more protable workplaces exhibit higher job satisfaction
(Schneider et al. 2003), which may promote these employees voluntary participation in informal HPWPs. To address the potential reverse causality, we controlled for the 2005 workplace protability. In so doing, our analysis offers a
better causal inference by estimating the 2006 workplace protability that is
unexplained by the previous years protability (Wright et al. 2005).
Results. Table 4 presents the mean of the variables we used in this analysis. The relationships among various HPWP indices are consistent with those
in our previous analysis on job satisfaction. Employees participated in informal
HPWPs (.301 in Sample 1 and .300 in Sample 2) to a greater extent than formal HPWPs (.157 in Sample 1 and .163 in Sample 2). Although employees
did not participate in some HPWPs that were present in the workplace, as represented by the unused HPWP index (.162 in Sample 1 and .162 in Sample 2),
the employee-reported HPWP index (.458 in Sample 1 and .463 in Sample 2)
is larger than the management-reported HPWP index (.319 in Sample 1 and
.325 in Sample 2). Although not reported in this manuscript, the correlation
between the management-reported HPWP index and employee-reported HPWP
index is moderate (r = .32 in Sample 1 and r = .40 in Sample 2). These gures
are comparable to r = .39 in Liao et al. (2009) and r = .41 in Takeuchi et al.
(2007), although lower than r = .59 in Messersmith, Patel, and Lepak (2011).
Table 5 displays the results of our regression analyses. We ran separate
models with different HPWP indices. Results show that both the managementreported HPWP index and employee-reported HPWP index are positively associated with subsequent workplace protability in Sample 2, but only

Informal HPWPs

/ 525

employee-reported HPWP index is positively associated with workplace protability in Sample 1. The insignicant effect of management-reported HPWP
index in Sample 1 appears to be due to the insignicant effect of unused
HPWP index considering that the management-reported HPWP index consists
of formal HPWP index and unused HPWP index.
For the analysis of the decomposed HPWP indices (Model 4), the relationship between the formal HPWP index and workplace protability is consistently positive and signicant in both samples. The relationship between the
informal HPWP index and workplace protability appears to be signicantly
positive. It is marginally signicant in Sample 1 (p = .087) and highly signicant in Sample 2 (p < .001). Beta coefcients are comparable between the formal HPWP index (b = 1.875 in Sample 1 and b = 3.123 in Sample 2) and
informal HPWP index (b = 1.633 in Sample 1 and b = 3.500 in Sample 2) in
both samples, and their differences are not statistically signicant (p = .811 in
Sample 1 and p = .716 in Sample 2). Overall, we surmise that employee participation in informal HPWPs is associated with better subsequent workplace
performance in a similar manner as employee participation in formal HPWPs
is associated with better subsequent workplace performance. The effect of
unused HPWPs on workplace protability is ambiguous. Their effect is insignicant in Sample 1, whereas it is marginally signicant in Sample 2.
The criteria we used to dene our samples, that is, two or more employee
reports from a workplace (Sample 1, N = 3125) and three or more employee
reports from a workplace (Sample 2, N = 2665), are based on the criteria used
in existing research (Jensen, Patel, and Messersmith 2013; Takeuchi et al.
2007). We ran analyses with samples that used alternative criteria. Specically
we ran the analysis with a sample with a lower threshold consisting of workplaces with employee reports from one or more employees (Sample 3, N =
3357) and a sample with a higher threshold consisting of workplaces with
employee reports from four or more employees (Sample 4, N = 2022). We
refer to the former sample as Sample 3 and the latter sample as Sample 4.
Mean workplace size is 41.040 with a median of 23 for Sample 3, and 64.667
with a median of 34 for Sample 4. The mean number of employee reports per
workplace is 3.268 with a median of 3 for Sample 3, and 4.986 with a median
of 5 for Sample 4.
We summarize our results of this additional analysis in Appendix B. Results
of the analysis with Sample 3 are highly consistent with the results of the
analysis with Samples 1 and 2. Both formal and informal HPWP indices are
positively and signicantly associated with workplace protability. In contrast,
none of the HPWP indices are signicant in our analysis using Sample 4.
Informal HPWP index is not signicantly associated with workplace protability although the sign of its beta coefcient is positive. Neither is formal HPWP

526 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

index, management-reported HPWP index, nor employee-reported HPWP


index. These results do not support our analyses with Samples 1 through 3, as
well as prior HPWP studies, which have reported positive effects of HPWPs
on organizational performance (Combs et al. 2006).
One possible explanation for the insignicant ndings with Sample 4 is a
sampling bias. The sample size decreased considerably from Sample 2 (workplaces with reports from three or more employees, N = 2665) to Sample 4
(workplaces with reports from four or more employees, N = 2002), suggesting
that almost 25 percent of the workplaces in Sample 2 were excluded from
Sample 4 (2665 2022 / 2655 = 24.1 percent). Considering that more
employee reports were collected from large workplaces, increasing the threshold for the number of employee reports leads to oversampling of large workplaces. Indeed, the mean workplace size of Sample 4 (66.667) is larger than
that of our original samples (43.928 in Sample 1 and 49.159 in Sample 2), as
well as that in other HPWP studies that examined the WES dataset (e.g.,
45.70 in Haines, Jalette, and Larose 2010).
As we noted, the target population of the WES is all business locations in Canada. To ensure that its survey reects this target population accurately, the WES
used random stratied sampling by region, industry, and workplace size. Thus,
increasing a threshold for the number of employee reports departs from the target
population of the WES by limiting the sample to large workplaces. While we
acknowledge that it is useful to examine whether or not our models are robust
for workplaces with a greater number of employee reports, our original samples
(Samples 1 and 2) reect the target population of WES more accurately, and are
consistent with other publications that analyze WES data (Haines, Jalette, and
Larose 2010; Zatzick and Iverson 2006). For these reasons, we contend that the
results of the analyses with these samples should carry more weight when drawing an inference about the implications of informal HPWPs. We acknowledge
that the insignicant ndings of the analysis with Sample 4 point to an issue that
empirical HPWP studies need to consider. We subsequently discuss this issue as
an important limitation of our study.

Discussion and Conclusion


Extant HR research is primarily concerned with the HR practices organizations establish, and there has been little effort devoted to understanding how
employees actually participate in these HR practices. Some researchers are
beginning to recognize the divergence between formal practices and employee
participation in them (e.g., Arthur and Boyles 2007). These researchers typically focus on the case in which a formal practice is not used by employees,

Informal HPWPs

/ 527

explaining that such a practice is not successfully implemented. We contend


that divergence can also occur when employees participate in a practice that is
not formally introduced by their organizations. Employees may participate in a
HPWP that is initiated by line managers or employees without formal endorsement from their organizations. Our analyses of the WES support this argument. A close comparison of management reports and employee reports from
WES reveals the case in which employees participate in a HPWP that is not
formally present in the organization, as well as the case in which employees
do not participate in a HPWP that is formally present in the organization.
Recognizing the reality that employees participate in HPWPs on both formal
and informal bases, we examined the implications of employee participation in
informal HPWPs for employees and organizations. Our empirical analysis
demonstrates that employee participation in informal HPWPs is positively
related to job satisfaction. These relationships are not statistically different
from the relationships of employee participation in formal HPWPs to these
outcomes. Thus, to the extent that employees participate in HPWPs, the
formality of these practices appears to make little difference. Regardless of formality, employee participation in HPWPs stimulates mechanisms that positively impact employee-level and organizational-level outcomes. Although we
posited that the ambiguity of organizational HR policies surrounding informal
HPWPs may weaken the positive effects, our analyses did not support this
argument. Interestingly, ICC(2) for the informal HPWP index is high (i.e., .73
in Sample 1 and .72 in Sample 2), suggesting that the extent to which employees participate in informal HPWP is similar among employees in the same
workplaces. As a result of participating in informal HPWPs in a similar manner, employees in the same workplaces may share similar perceptions about
organizational HR policies. This lower level of ambiguity might contribute to
stronger organizational culture (Schneider, Gunnarson, and Niles-Jolly 1994)
or organizational climate (Bowen and Ostroff 2004), promoting employee
behaviors that positively impact employees and organizations.
Methodologically, by demonstrating that employee participation in formal
and informal HPWPs is likely to be associated with employee-level and workplace-level outcomes, our study also highlights the advantage of obtaining
employee reports on HPWPs, compared to the management report on HPWPs,
in predicting the positive employee-level and workplace-level outcomes. From
these results, we contend that simply focusing on the presence of formal practices obscures the effect of HPWPs. Admittedly, HPWP research that relies on
management reports still demonstrates the positive effects of HPWPs on organizational performance. We suspect that this is because employees actually
participate in some of these HPWPs and these practices exert a positive impact
on performance. Solely relying on the information about the presence of

528 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

formal practices overlooks employee participation in informal HPWPs, which


our analyses suggested has a positive impact on job satisfaction and workplace
protability. This is not to say that HPWP research should focus exclusively
on employee reports. As our analyses demonstrate, comparison of the management and employee reports enables us to derive multiple HPWPs that have
varying implications for employee-level and organization-level outcomes. Collecting reports from both management and employees helps us develop a more
sophisticated understanding about the effects of HPWPs. We also acknowledge
the merit of management reports for HR research. Arguably, collecting information from employees is more costly than collecting information from management. Considering that researchers undertake research projects with
available resources, if they can ensure a reasonable level of accuracy of information, relying on management reports could be a more practical option
(Huselid and Becker 2000).
We expect that our nding about informal HPWPs will develop a new
research direction. While our study identied employee participation in informal HPWPs by matching management reports to the WES Workplace Survey
and employee reports to the WES Employee Survey, we were unable to obtain
the information about how these informal HPWPs became accessible to these
employees. As we elaborate, line managers may initiate some practices without
formal endorsement from senior management (e.g., teams, information sharing); others may be initiated by employees. Future researchers are encouraged
to examine how these informal HPWPs emerge in workplaces, and whether or
not the effects on job satisfaction and organizational performance vary according to these mechanisms.
It is also worth mentioning the implications of unused HPWPs. As we
explained, unused HPWPs could involve multiple, possibly conicting, effects
on job satisfaction and organizational performance. Our rst analysis showed
that unused HPWPs are negatively associated with job satisfaction. One possible interpretation is that employees nd reduced job satisfaction if there are
practices that are formally present in their organizations, but are not really
accessible to them. Alternatively, the opposite causal direction is possible such
that employees with lower job satisfaction are less likely to participate in
HPWPs even though they are accessible (Mohr and Zoghi 2008). For example,
those with lower job satisfaction may be reluctant to participate in various
team practices (e.g., problem-solving teams, self-directed teams) or training
programs. In contrast, our second analysis seems to indicate that unused
HPWPs are positively related to workplace protability. If unused HPWPs are
due to the strategic differentiation of HR practices among employee groups
(Lepak and Snell 1999), this arrangement could capitalize on the positive
effects of HPWPs. Admittedly, these explanations are post hoc. Research is

Informal HPWPs

/ 529

needed to delve into the effects of unused HPWPs. Such research will need to
consider multiple mechanisms through which unused HPWPs affect employee
attitudes and behaviors.
There are several important limitations of this study. First, and most important, the results of the analysis on workplace protability changed when we
experimented with various thresholds for the number of employee reports collected. When we increased the threshold to four or more employee reports
(i.e., the analysis with Sample 4), the effects of informal HPWP index, as well
as those of other HPWP indices, became insignicant. While we suspect that
increasing the threshold excludes a large number of workplaces and departs
from the target sampling population of the WES Workplace Survey, we
acknowledge that this is an important limitation of our study. Considering that
the number of employee reports is positively associated with workplace size,
future studies should investigate whether our results regarding the effect of
informal HPWPs are more relevant to small and medium workplaces. While
we are not aware of any statistics, empirical research suggests that employers
apply distinct HR practices according to employee groups (Lepak and Snell
2002). Ideally, aggregation of HPWP indices at an organization level in large
workplaces should consider the possibility of this differentiation of HR practices across employee groups. The proportion of employees sampled in our
analyses is relatively small (i.e., 1315 percent of employees),10 and thus the
coverage of employee reports may not be as extensive as we would like it to
be. Although ICC(2) for our HPWP indices supports our aggregation of individual-level measures at the workplace level, our results are likely to be inuenced by which employees, as well as which occupational groups, are
included in the employee sample. We call for future research that compares
the effects of various types of HPWPs on workplace performance by incorporating the possible within-organizational differences in HR practices.
Second, the survey question wording in the WES Workplace Survey does
not perfectly match that in the Employee Survey. We submit that it is evident
from Table 1 that corresponding questions attempt to capture the same practices. Yet, it is still possible that part of the difference between management
reports about the presence of formal practices and employee reports about participation in HPWPs is attributable to the slight wording differences. Third,
our measure of informal HPWPs is derived from the comparison between
employee reports to the WES Employee Survey about their participation in
HPWPs and management reports to the WES Workplace Survey about the
presence of formal practices. In this sense, our measure does not directly ask

10

Calculation is made using the median values of workplace size and the number of employee reports.

530 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

employees about participation in informal practices. We contend that our


approach is appropriate to compare the effects of formal HPWPs and informal
HPWPs because directly asking employees about the formality of practices
could inuence their perceptions about formal and informal practices. That
said, to understand the psychological mechanisms through which informal
HPWPs stimulate employee reactions, asking about employees perceptions
regarding the formality of practices might be useful.
Despite these limitations, our study makes a unique contribution to HPWP
research by highlighting the potential positive implications of employee participation in informal HPWPs for employee job satisfaction and workplace protability. Exploring the mechanism through which these informal practices
emerge in the workplace and how employees perceive and react to these practices will further promote our understanding of the impact of work practices
on employee-level and organization-level outcomes.

REFERENCES
Avgar, Ariel C., Rebecca Kolins Givan, and MingWei Liu. 2011. Patient-Centered but Employee Delivered: Patient Care Innovation, Turnover Intentions, and Organizational Outcomes in Hospitals. Industrial & Labor Relations Review 64(3): 42340.
Arthur, Jeffrey B., and Trish Boyles. 2007. Validating the Human Resource System Structure: A LevelsBased Strategic HRM Approach. Human Resource Management Review 17(1): 7792.
Bartel, Ann P.. 2004. Human Resource Management and Organizational Performance: Evidence from Retail
Banking. Industrial & Labor Relations Review 57(2): 181203.
Batt, Rosemary. 2002. Managing Customer Services: Human Resource Practices, Quit Rates, and Sales
Growth. Academy of Management Journal 45(3): 58797.
Blasi, Joseph R., and Douglas L. Kruse. 2006. U.S. High-Performance Work Practices at Centurys End.
Industrial Relations 45(4): 54778.
Bowen, David E., and Cheri Ostroff. 2004. Understanding HRM-Firm Performance Linkages: The Role of
the Strength of the HRM System. Academy of Management Review 29(2): 20321.
Bowling, Nathan A., Kevin J. Eschleman, and Qiang Wang. 2010. A Meta-Analytic Examination of the
Relationship between Job Satisfaction and Subjective Well-Being. Journal of Occupational and
Organizational Psychology 83(4): 91534.
Burbidge, John B., Lonnie Magee, and A. Leslie Robb. 1988. Alternative Transformations to Handle
Extreme Values of the Dependent Variable. Journal of the American Statistical Association 83(401):
12327.
Cassell, David L. 2006. Wait, Wait, Dont Tell Me . . . Youre Using the Wrong Proc! Proceedings of the
Thirty-rst Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc.
Chadwick, Clint, Ji-Young Ahn, and Kiwook Kwon. 2012. Human Resource Managements Effects on
Firm-Level Relative Efciency. Industrial Relations 51(3): 70430.
Colvin, Alexander J. S, Rosemary Batt, and Harry C. Katz. 2001. How High Performance Human Resource
Practices and Workforce Unionization Affect Managerial Pay. Personnel Psychology 54(4): 90334.
Combs, James, YongMei Liu, Angela Hall, and David Ketchen. 2006. How Much Do High Performance
Work Practices Matter? A Meta-Analysis of Their Effects on Organizational Performance. Personnel
Psychology 59(3): 50128.
Cooke, Gordon B., Isik U. Zeytinoglu, and James Chowhan. 2009. Barriers to Training Access. Perspectives on Labour and Income 10(7). http://www.statcan.gc.ca/pub/75-001-x/2009107/article/10907-eng.
htm (accessed November 13, 2012).

Informal HPWPs

/ 531

Crook, T. Russell, Samuel Y. Todd, James G. Combs, and David J. Woehr. 2011. Does Human Capital
Matter? A Meta-Analysis of the Relationship between Human Capital and Firm Performance. Journal
of Applied Psychology 96(3): 44356.
Datta, Deepak K., James P. Guthrie, and Patrick M. Wright. 2005. Human Resource Management and
Labor Productivity: Does Industry Matter? Academy of Management Journal 48(1): 13545.
Diez, Graham, Adrian Wilkinson, and Tom Redman. 2009. Involvement and Participation. In The Sage
Handbook of Human Resource Management, edited by Adrian Wilkinson, Nicolas Bacon, Tom Redman, and Scott Snell, pp. 24568. Thousand Oaks, CA: Sage.
Dillman, Don A., Jolene D. Smyth, and Leah Melani Christian. 2009. Internet, Mail, and Mixed-Mode Surveys: The Tailored Design Method, 3rd ed. Hoboken, NJ: Wiley.
Eaton, Susan C. 2003. If You Can Use Them: Flexibility Policies, Organizational Commitment, and Perceived Performance. Industrial Relations 42(2): 14567.
Gardner, Timothy M., Patrick M. Wright, and Lisa M. Moynihan. 2011. The Impact of Motivation,
Empowerment, and Skill-Enhancing Practices on Aggregate Voluntary Turnover: The Mediating Effect
of Collective Affective Commitment. Personnel Psychology 64(2): 31550.
Glew, David J., Anne M. OLeary-Kelly, Ricky W. Grifn, and David D. Van Fleet. 1995. Participation in
Organizations: A Preview of the Issues and Proposed Framework for Future Analysis. Journal of
Management 21(3): 395421.
Glick, W. H. 1985. Conceptualizing and Measuring Organizational and Psychological Climate: Pitfalls in
Multilevel Research. Academy of Management Review 10(3): 60116.
Godard, John. 2001. High Performance and the Transformation of Work? The Implications of Alternative
Work Practices for the Experience and Outcomes of Work. Industrial & Labor Relations Review 54
(4): 776805.
Haines, III, Victor Y., Patrice Jalette, and Karine Larose. 2010. The Inuence of Human Resource Management Practices on Employee Voluntary Turnover Rates in the Canadian Non-Governmental Sector.
Industrial & Labor Relations Review 63(2): 22846.
Heckman, J. Richard, and Greg R. Oldham. 1976. Motivation through the Design of Work: Test of Theory. Organizational Behavior and Human Performance 16(2): 25079.
Heeringa, Steven G., Brady T. West, and Patricia A. Berglund. 2010. Applied Survey Data Analysis. Boca
Raton, FL: Chapman & Hall/CRC.
Heywood, John S., W. S. Siebert, and Xiangdong Wei. 2002. Worker Sorting and Job Satisfaction: The
Case of Union and Government Jobs. Industrial & Labor Relations Review 55(4): 595609.
Huselid, Mark A. 1995. The Impact of Human Resource Management Practices on Turnover, Productivity,
and Corporate Financial Performance. Academy of Management Journal 38(3): 63572.
, and Brian E. Becker 2000. Comment on Measurement Error in Research on Human Resources
and Firm Performance: How Much Error is There and How Does It Inuence effect Size Estimates?
by Gerhart, Wright, McMahan, and Snell. Personnel Psychology 53(4): 83554.
Jensen, Jaclyn M., Pankaj C. Patel, and Jake G. Messersmith. 2013. High-Performance Work Systems and
Job Control: Consequences for Anxiety, Role Overload, and Turnover Intentions. Journal of Management 39(6): 1699724.
Judge, Timothy A., Carl J. Thoresen, Joyce E. Bono, and Gregory K. Patton. 2001. The Job SatisfactionJob Performance Relationship: A Qualitative and Quantitative Review. Psychological Bulletin 127(3):
376407.
Kato, Takao, and Motohiro Morishima. 2002. The Productivity Effects of Participatory Employment Practices: Evidence from New Japanese Panel Data. Industrial Relations 41(4): 487520.
Kehoe, Rebecca R., and Patrick M. Wright. 2013. The Impact of High-Performance Human Resource Practices on Employees Attitudes and Behaviors. Journal of Management 39(2): 36691.
Klein, Katherine J., Mark A. Grifn, Paul D. Bliese, David A. Hofmann, Steve W. J. Kozolowski, Lawrence
R. James, Fred Dansereau, Francis J. Yammarino, Mark B. Gavin, and Michelle C. Bligh. 2000.
Multilevel Analysis Techniques: Commonalities, Differences, and Continuing Questions. In Multilevel Theory, Research, and Methods in Organizations: Foundations, Extension, and New Directions,
edited by Katherine J. Klein, and Steve W. J. Kozlowski, pp. 51253. San Francisco: Jossey-Bass.

532 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

Lepak, David P., and Scott A Snell. 1999. The Human Resource Architecture: Toward a Theory of Human
Capital Allocation and Development. Academy of Management Review 24(1): 3148.
, and Scott A. Snell. 2002. Examining the Human Resource Architecture: The Relationships among
Human Capital, Employment, and Human Resource Congurations. Journal of Management 28(4):
51743.
Liao, Hui, Keio Toya, David P. Lepak, and Ying Hong. 2009. Do They See Eye to Eye? Management and
Employee Perspectives of High-Performance Work Systems and Inuence Processes on Service Quality. Journal of Applied Psychology 94(2): 37198.
MacDufe, John P. 1995. Human Resource Bundles and Manufacturing Performance: Organizational Logic
and Flexible Production Systems in the World Auto Industry. Industrial & Labor Relations Review
48(2): 197221.
Messersmith, Jake G., Pankaj C. Patel, and David P. Lepak. 2011. Unlocking the Black Box: Exploring the
Link between High-Performance Work Systems and Performance. Journal of Applied Psychology 96
(6): 110518.
Meyer, John P., David J. Stanley, Lynne Herscovitch, and Laryssa Topolnytsky. 2002. Affective, Continuance, and Normative Commitment to the Organization: A Meta-Analysis of Antecedents, Correlates,
and Consequences. Journal of Vocational Behavior 61(1): 2052.
Mohr, Robert D., and Cindy X. Zoghi. 2008. High-Involvement Work Design and Job Satisfaction. Industrial & Labor Relations Review 61(3): 27596.
Neter, John, William Wasserman, and Michael H. Kutner. 1985. Applied Linear Statistical Models. Homewood, IL: Irwin.
Nishii, Lisa H., David P. Lepak, and Benjamin Schneider. 2008. Employee Attributions of the Why of
HR Practices: Their Effects on Employee Attitudes and Behaviors, and Customer Satisfaction. Personnel Psychology 61(3): 50345.
Nyberg, Anthony J., Ingrid Smithey Fulmer, Barry Gerhart, and Mason A. Carpenter. 2010. Agency Theory
Revisited: CEO Return and Shareholder Interest Alignment. Academy of Management Journal 53(5):
102949.
Pendleton, Andrew, and Andrew Robinson. 2010. Employee Stock Ownership, Involvement, and Productivity: An Interaction-Based Approach. Industrial & Labor Relations Review 64(1): 329.
Preuss, Gil A. 2003. High Performance Work Systems and Organizational Outcomes: The Mediating Role
of Information Quality. Industrial & Labor Relations Review 56(4): 590605.
SAS Institute. 2008. SAS/STAT 9.2 Users Guide. Cary, NC: SAS Publishing.
Schneider, Benjamin, Sara K. Gunnarson, and Kathryn Niles-Jolly. 1994. Creating the Climate and Culture
of Success. Organizational Dynamics 23(1): 1729.
, Paul J. Hanges, Brent Smith, and Amy N. Salvaggio. 2003. Which Comes First: Employee Attitudes or Organizational Financial or Market Performance? Journal of Applied Psychology 88(5):
83651.
Scott, K. Dow, and G. Stephen Taylor. 1985. An Examination of Conicting Findings on the Relationship
between Job Satisfaction and Absenteeism: A Meta-Analysis. Academy of Management Journal 28
(3): 599612.
Statistics Canada. 2008. Workplace and Employee Survey (WES). http://www.statcan.gc.ca/cgi-bin/imdb/
p2SV.pl?Function=getSurvey&SurvId=2615&SurvVer=1&InstaId=13978&InstaVer=8&SDDS=2615&lang=en&db=imdb&adm=8&dis=2 (accessed November 13, 2012).
Takeuchi, Riki, David P. Lepak, Heli Wang, and Kazuo Takeuchi. 2007. An Empirical Examination of the
Mechanisms Mediating between High Performance Work Systems and the Performance of Japanese
Organizations. Journal of Applied Psychology 92(4): 106983.
Tett, Robert T., and John P. Meyer. 1993. Job Satisfaction, Organizational Commitment, Turnover Intention, and Turnover: Path Analyses based on Meta-Analytic Findings. Personnel Psychology 46(2):
25993.
Walsworth, Scott, and Anil Verma. 2007. Globalization, Human Resource Practices and Innovation: Recent
Evidence from the Canadian Workplace and Employee Survey. Industrial Relations 46(2): 22240.
Wanous, John P., Arnon E. Reichers, and Michael J. Hudy. 1997. Overall Job Satisfaction: How Good Are
Single-Item Measures? Journal of Applied Psychology 82(2): 24752.

Informal HPWPs

/ 533

Wright, Patrick M., Timothy Gardner, Lisa M. Moynihan, and Mathew R. Allen. 2005. The Relationship
between HR Practices and Firm Performance: Examining Causal Order. Personnel Psychology 58(2):
40946.
Zatzick, Christopher D., and Roderick D. Iverson. 2006. High-Involvement Management and Workforce
Reduction: Competitive Advantage or Disadvantage? Academy of Management Journal 49(5): 999
1015.
APPENDIX A

RESULTS

SAMPLE USED

IN THE

ANALYSIS

ON

Variables

Component

DF

Employee-reported
HPWP index
Formal HPWP index

Workplace
Employee
Workplace
Employee
Workplace
Employee
Workplace
Employee
Workplace
Employee

3,999
13,697
3,999
13,697
3,999
13,697
3,999
13,697
3,999
13,697

Informal HPWP index


Unused HPWP index
Job satisfaction

ANALYSIS

OF

OF

VARIANCE

JOB SATISFACTION
Sum of Squares

Mean Square

F-value

R2

200,330
242,022
254,741
74,557
197,745
126,621
113,968
74,557
1,308,945
2,462,396

50.095
17.670
63.701
5.443
49.449
9.244
28.499
5.443
327.318
179.776

2.84***

.453

.65

11.70***

.774

.91

5.35***

.610

.81

5.24***

.605

.81

1.82***

.347

.45

ICC(2)

NOTE: N = 17,697 employees from 4000 workplaces. *** p < .001.

SAMPLE USED

IN THE

ANALYSIS

ON

Variables

Component

DF

Employee-reported
HPWP index
Formal HPWP index

Workplace
Employee
Workplace
Employee
Workplace
Employee
Workplace
Employee

3,124
10,429
3,124
10,429
3,124
10,429
3,124
10,429

Informal HPWP index


Unused HPWP index

WORKPLACE PROFITABILITY (SAMPLE 1)


Sum of Squares

Mean Square

F-value

R2

12,247
14,092
12,464
2,855
10,393
9,287
6,240
2,855

3.920
1.351
3.99
.274
3.327
.891
1.998
.274

2.90***

.465

.66

14.57***

.814

.93

3.74***

.528

.73

7.30***

.686

.86

ICC(2)

NOTE: N = 13,554 employees from 3125 workplaces. *** p < .001.

SAMPLE USED
variables
Employee-reported
HPWP index
Formal HPWP index
Informal HPWP index
Unused HPWP index

IN THE

ANALYSIS

Component
Workplace
Employee
Workplace
Employee
Workplace
Employee
Workplace
Employee

ON

DF
2,664
9,969
2,664
9,969
2,664
9,969
2,664
9,969

WORKPLACE PROFITABILITY (SAMPLE 2)

Sum of Squares
9,877
13,030
10,842
2,658
8,100
8,509
4,803
2,658

NOTE: N = 12,634 employees from 2665 workplaces. ***p < .001.

Mean Square
3.708
1.307
4.070
.267
3.040
.854
1.803
.267

F-value

R2

ICC(2)

***

.431

.65

15.27***

.803

.93

3.56***

.488

.72

6.76***

.644

.85

2.84

534 / YOSHIO YANADORI,

AND

DANIELLE D.

VAN

JAARSVELD

APPENDIX B

RELATIONSHIP

OF

HPWP INDICES

TO

WORKPLACE PROFITABILITY

Sample 3 (Workplaces with 1 or More


Employee Reports)
Variables
HPWP indices
Managementreported
HPWP index
Employee-reported
HPWP index
Formal HPWP
index
Informal HPWP
index
Unused HPWP
index
R2
Adjusted R2
N

Model 1

Model 2

Model 3

Model 4

Sample 4 (Workplaces with 4 or More


Employee Reports)
Model 1

1.954***
(.542)

Model 2

.284
.280

Model 4

0.726
(.528)
2.449***
(.709)

.282
.277
3,357

Model 3

.284
.279

0.431
(.816)
3.432***
(.851)
2.048*
(.874)
1.872
(1.091)
.286
.281

.484
.479
2,665

.485
.479

.484
.479

.378
(.866)
.795
(1.095)
2.196
(1.198)
.485
.479

NOTE: Standard errors are in parentheses. < 0.10; *p < 0.05; *** p < 0.001. Controls are included but not reported. Mean
and median workplace size is 41.040 and 23 for Sample 3 and 64.667 and 34 for Sample 4. The mean and median number of employee reports per workplace is 3.268 and 3 for Sample 3 and 4.986 and 5 for Sample 4.

Você também pode gostar