The Relationship Between Work-to-Family Conflict

and Family-to-Work Conflict: A Longitudinal Study

Yueng-Hsiang Huang
Liberty Mutual Research Institute for Safety

Leslie B. Hammer
Margaret B. Neal
Portland State University

Nancy A. Perrin
Oregon Health & Science University

ABSTRACT: The relationship between work-to-family conflict and family-to-work con-

flict was tested using a longitudinal research design with 234 dual-earner couples car-
ing for both children and aging parents. Two waves of mailed survey data were col-
lected. The mediating effects of role-related satisfaction were hypothesized to link these
two forms of work-family conflict. The analytical steps for determining mediating effects
followed suggestions by Baron and Kenny (1986). The results failed to show significant
mediating effects of role-related satisfaction between the two forms of conflict over time
for both husbands and wives, contrary to suggestions by Frone, Yardley, and Markel
(1997). Instead, positive, direct relationships between the two types of work-family con-
flict were found. Limitations and implications are discussed.

KEY WORDS: longitudinal research; work-family conflict.

Yueng-hsiang E. Huang, Liberty Mutual Research Institute for Safety, 71 Frankland

Road, Hopkinton, MA 01748; e-mail: yueng-hsiang.huang@libertymutual.com.
Leslie B. Hammer, Portland State University, P.O. Box 751, Portland, OR 97207-
0751; e-mail: hammer@pdx.edu.
Margaret B. Neal, Portland State University, P.O. Box 751, Portland, OR 97207-0751;
e-mail: nealm@pdx.edu.
Nancy A. Perrin, Oregon Health & Science University, 3455 SW US Veterans Hospi-
tal Road, Portland, OR 97201-2941; e-mail: perrinn@ohsu.edu.
An earlier version of this paper was presented at the annual conference of the Society
for Industrial and Organizational Psychology in May 2000 in New Orleans, LA. The
results reported here are derived from a larger, three-year study, Dual-Earner Couples
in the Sandwiched Generation, conducted at Portland State University. The study was
supported in part by a grant from the Alfred P. Sloan Foundation (Grant # 96-10-20) to
Margaret Neal and Leslie Hammer, and in part by the Department of Psychology at
Portland State University. The authors would like to thank Wayne Wakeland and Sully
Taylor for their helpful comments on this article.

Journal of Family and Economic Issues, Vol. 25(1), Spring 2004  2004 Human Sciences Press, Inc. 79
80 Journal of Family and Economic Issues

Work and family comprise the two central systems of adult life, and
each offers a unique point from which to study important qualities of
human behavior. It was once believed that work life was separate
from family life; this belief was referred to as the “myth of separate
worlds” (Kanter, 1977). According to this belief, each domain is oper-
ated by its own laws, and thus could be studied separately. This sepa-
rate world argument has now been replaced with the idea that the
relationship between work and family is dynamic and reciprocal. Not
only do factors in the work sphere influence family life, but family
matters also have strong effects on work life (e.g., Crouter, 1984; Near,
Rice, & Hunt, 1980). As a result, changes in the family must be exam-
ined in connection with what happens at work and vice versa.
One way to look at the relationship between work and family sys-
tems is to examine work-family conflict. Kahn, Wolfe, Quinn, Snoek,
and Rosenthal (1964) defined inter-role conflict as the “simultaneous
occurrence of two or more sets of pressures such that compliance with
one would make more difficult compliance with the other” (p. 19). Work-
family conflict (WFC) is a form of inter-role conflict whereby the role
pressures from the work and family domains are mutually incompati-
ble in some respect: one role is made more difficult due to participa-
tion in the other (Greenhaus & Beutell, 1985).
WFC, as a source of stress, has been linked to many negative out-
comes in both work and family life. For example, a number of studies
have found that WFC has negative effects on job, family, and life satis-
faction (Adams, King, & King, 1996; Aryee, 1992; Higgins, Duxbury, &
Irving, 1992; Kossek & Ozeki, 1998). Other outcomes associated with
WFC include absenteeism, lateness to work, intention to quit one’s job,
and poor performance at work (e.g., Aryee, 1992; Burke, 1988; Goff,
Mount, & Jamison, 1990; Thomas & Ganster, 1995; Youngblood &
Chambers-Cook, 1984), as well as mental and physical health prob-
lems (Burke, 1988; Barnett & Rivers, 1996; Frone, Russell, & Cooper,
1992; Frone, Yardley & Markel, 1997; Thomas & Ganster, 1995). The
understanding of this inter-connection between work and family life
has led to a decrease in the traditional belief that work and family are
separate spheres, and the recognition of the need for organizations to
create new, or modify existing, work-family human resource policies
and programs in order to help their employees balance their work and
family lives (e.g., Higgins et al., 1992; Kossek & Ozeki, 1998). In order
to maintain the health and success of individuals and companies,
greater understanding of WFC is important.
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 81

The Relationship Between Work-to-Family Conflict

and Family-to-Work Conflict

Although it was once assumed that WFC was a uni-dimensional

construct (e.g., Kopelman, Greenhaus, & Connolly, 1983), most recent
definitions display a bi-directional conceptualization (Eagle, Miles, &
Icenogle, 1997; Frone et al., 1992; Frone et al., 1997; Gutek, Searle, &
Klepa, 1991; Netemeyer, Boles, & McMurrian, 1996), with WFC hav-
ing two components: work interfering with family (W → F Conflict),
and family interfering with work (F → W Conflict). As recent litera-
ture suggests, W → F Conflict and F → W Conflict are distinct, yet
related, forms of interrole conflict. Antecedents of W → F conflict in-
clude work role stressors and characteristics, such as lack of supervi-
sor support or long work hours, while antecedents of F → W conflict
tend to be family role stressors and characteristics, such as having
young children at home (e.g., Frone et al., 1992; Greenhaus & Beutell,
1985). Likewise, outcomes of W → F conflict include family-related
distress and family-related behaviors, such as less family satisfaction,
while the outcomes of F → W conflict include work-related distress
and work-related behaviors, such as less job satisfaction, more absen-
teeism, lateness to work, and greater intention to quit work (e.g.,
Frone et al., 1992; Frone et al., 1997). Recent research has shown that
the most consistent family characteristic predicting work-life imbal-
ance is being a parent, and the most consistent work characteristic
predicting work-life imbalance is hours worked (Tausig & Fenwick,
2001).
Developing a better understanding of how the two forms of WFC
affect one another is the central purpose of the present study. Al-
though the relationship between W → F conflict and F → W conflict
has been examined in cross-sectional research (e.g., Frone et al., 1992,
1997), little is known about how the two forms of WFC affect one an-
other over time. Thus, the present study conducted empirical research
to test the relationship between the two dimensions of WFC over a
one-year period. Two studies, Frone et al. (1992) and Frone et al.
(1997), were used as the theoretical bases for the current study.
In 1992, Frone et al. developed and tested a comprehensive model
of the work-family interface. That study extended prior research by
explicitly distinguishing between work interfering with family (W →
F) and family interfering with work (F → W). Frone et al. hypothe-
sized a positive, reciprocal and direct relationship between the two
82 Journal of Family and Economic Issues

types of WFC. The rationale was that if one’s work-related problems

start to interfere with the accomplishment of one’s family-related obli-
gations, these unfulfilled family obligations might begin to interfere
with one’s day-to-day functioning at work. A parallel process explain-
ing the direct positive effect between F → W Conflict and W → F Con-
flict was hypothesized.
Frone et al.’s (1992) model was tested with structural equation mod-
eling techniques using cross-sectional data obtained from a random
sample of 613 individuals. The authors concluded that the findings
were strongly supportive of a positive, direct, reciprocal relationship
between these two types of conflict.
In 1997, Frone et al. extended Frone et al’s 1992 model and devel-
oped a revised integrative model of the work-family interface. One of
the significant differences between the two models was that while the
1992 study hypothesized that the relationships between W → F Con-
flict and F → W Conflict were direct, the 1997 study hypothesized that
these relationships were indirect, mediated by role distress/dissatis-
faction and role overload. The rationale for an indirect relationship
between W → F Conflict and F → W Conflict was that high levels of
W → F Conflict would lead to increases in family distress/dissatisfac-
tion because work-related demands reduce one’s available resources
to deal with family responsibilities. Thus, increases in family distress/
dissatisfaction would then lead to increased F → W Conflict. Con-
versely, a parallel process would explain the indirect positive effect of
F → W Conflict on W → F Conflict via increased work distress/dissat-
isfaction. In Frone et al.’s 1997 study, cross-sectional data were ob-
tained from a sample of 372 employed adults who were married and/
or parents. Using structural equation modeling to test their model,
the authors concluded that the results were supportive of their
hypotheses, demonstrating that role-related dissatisfaction was a key
mediating variable which links the work domain with the family do-
main. Therefore, based on both of the above mentioned models, it
appears that understanding the mechanisms that underlie the rela-
tionships between F → W and W → F Conflict is essential in under-
standing key antecedents and outcomes of both forms of WFC.
The major limitation of Frone et al.’s 1992 and 1997 studies is that
although the two models were conceptualized in terms of cause and
effect, cross-sectional data were used, which did not allow the authors
to draw causal inferences from their findings concerning the various
hypothesized relationships. Furthermore, although Frone et al.’s 1997
study hypothesized the mediating processes that linked these two
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 83

types of WFC, the data analysis procedure used did not allow for the
establishment of mediating effects. As Baron and Kenny (1986) clar-
ify, a given variable functions as a mediator to the extent that it ac-
counts for the relationship between the predictor and the outcome
variables. In order to support a mediating effect, the following condi-
tions must hold: (1) variations in levels of the independent variable
significantly account for variations in the presumed mediator; (2) vari-
ations in the mediator significantly account for variations in the de-
pendent variable; and (3) when both the impact of the independent
variable on the mediator and the impact of the mediator on the depen-
dent variable are controlled, a previously significant relationship be-
tween the independent and dependent variables is no longer signifi-
cant (Baron & Kenny, 1986). (Please see Method, data analysis section,
for further discussion.)
In light of these limitations, the present study followed Baron and
Kenny’s (1986) approach to test Frone et al.’s 1997 model of mediating
effects. The purpose of the current study was to re-examine the rela-
tionship between the two forms of WFC. More specifically, the present
study used a longitudinal design to test the mediating effects of job
and family satisfaction variables on the relationship between the two
forms of WFC, W → F Conflict and F → W Conflict, among a sample
of dual-earner couples in the sandwiched generation. This so-called
sandwiched generation consists of middle-aged parents who have com-
mitments to, and responsibilities for, both their children and their
aging parents (Durity, 1991; Raphael & Schlesinger, 1993).
Although the percentage of Americans who, regardless of their work
status, have caregiving responsibilities for both children and aging
parents is undetermined as yet, a few studies have specifically focused
on employees and their multiple caregiving responsibilities. For exam-
ple, Neal, Chapman, Ingersoll-Dayton and Emlen’s (1993) study of
employees found that workers with both types of responsibilities com-
prised 9% of the sample of employees overall. Furthermore, Durity
(1991) summarized published studies on the prevalence of the sand-
wiched generation and found a range of 6–40% of the employees of
corporations were in this population. A study by Nichols and Junk
(1997), who collected data in Idaho, Oregon, Utah, and Michigan from
a random sample of 40–65 year-old respondents, found that 15% of
the population had these dual caregiving responsibilities. Regardless
of the exact number of people in this generation, most experts tend to
agree that the percentages are bound to grow (Durity, 1991; Nichols &
Junk, 1997).
84 Journal of Family and Economic Issues

Previous research has shown that combinations of caregiving roles

can be very stressful for individuals and families, with the most diffi-
cult situation being to provide care for both children and elders (Nich-
ols & Junk, 1997). Furthermore, most people who have both elder and
child care responsibilities also have jobs. Adding the demands of the
work role to these multiple cargiving roles can be even more problem-
atic (Chapman, Ingersoll-Dayton, & Neal, 1994). Research with this
population of people with multiple roles will be especially beneficial
in extending our knowledge about how people balance work and fam-
ily. Thus, this study focuses on dual-earner couples in the sandwiched
generation as its exclusive sample.
As the American workforce becomes more diverse, with more women
engaged in paid employment (Frone et al., 1992; Offermann & Gow-
ing, 1990), knowledge regarding gender differences in WFC dynamics
is very important. In 1977, Pleck hypothesized that work and family
roles were asymmetrically permeable for men and women, with men
experiencing higher levels of work-to-family conflict and women expe-
riencing higher levels of family-to-work conflict. Research findings
testing this hypothesis, however, have been inconsistent. For example,
Gutek et al.’s (1991) study found that gender role expectations muted
the relationship between hours expended and perceived WFC, and
gender interacted with number of hours worked and WFC. However,
Frone et al.’s (1992) study found no gender differences in their model
of WFC. Alternatively, research by Barnett, Davidson, and Marshall
(1991) and Barnett and Marshall (1993), has found that the effects of
work and family role quality differed for men and women such that
higher role quality was related to improved health and well-being out-
comes for women, but not necessarily for men. Thus, the effects of
gender on work and family relationships are not clear. To explore the
effects of gender, in the present study, the relationships between
work-to-family conflict and family-to-work conflict are examined sepa-
rately for husbands and wives and parameter estimates are compared.
The following hypotheses are tested in the present study:

Hypothesis 1: Job Satisfaction mediates the relationship between

F → W Conflict and W → F Conflict, such that as F → W Conflict
increases, Job Satisfaction level decreases, and as Job Satisfac-
tion level decreases, W → F Conflict increases.
Hypothesis 2: Family Satisfaction mediates the relationship be-
tween W → F Conflict and F → W Conflict, such that as W → F
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 85

Conflict increases, Family Satisfaction level decreases, and as

Family Satisfaction level decreases, F → W Conflict increases.

Method

Participants

This study was conducted as part of a larger longitudinal research project

on work and family issues focusing on working couples in the sandwiched
generation. As noted earlier, the term sandwiched generation refers to people
who have commitments to, and responsibilities for, both children and aging
parents (Durity, 1991; Raphael & Schlesinger, 1993). Selection criteria for
participation in the larger study at Wave 1 included: (1) one person in the
couple worked 35 or more hours per week and the other worked at least 20
hours per week, (2) the couple had lived together for at least one year, (3)
they had one or more children aged 18 or under living in the household, (4)
together they spent a total of at least three hours per week caring for one or
more aging parents, (5) both members in the couple were willing to partici-
pate in the study, and (6) the couples earned a minimum of $40,000 per year
(median income for all families) (U.S. Bureau of Labor Statistics, 1995). The
last criterion was established in order to meet the specific interest of the fund-
ing agency in middle- and upper-income families.

Procedure

Sampling was accomplished via Computer-Assisted Telephone Interviewing

(CATI). A list of telephone numbers of households throughout the contiguous
continental U.S. comprised of one or more adults aged 30 through 60 was
purchased. Trained interviewers contacted and screened potential partici-
pants by telephone for eligibility based on the criteria identified above. If a
respondent’s answers to the screening questions indicated that his or her
household met all of the study criteria, except for income, that respondent
was asked if s/he and her/his spouse or partner would be willing to complete
a survey to be sent by mail. Eligible respondents also were told that, as a
token of appreciation, a check for $40 would be mailed to them upon receipt
of completed surveys from both members of the couple. It is important to note
that by using a targeted list of telephone numbers, some members of the pop-
ulation were not represented (e.g., persons with unlisted numbers or without
telephones). This sampling procedure resulted in a sample that was essen-
tially national in scope.
A total of 741 of the households contacted met the larger study’s criteria for
participation (excluding the income requirement). Of these, 624 couples either
agreed outright to participate in the study or to consider participating. Pack-
ages containing two cover letters, two surveys, and two postage-paid enve-
86 Journal of Family and Economic Issues

lopes (one for each member of the couple) were sent to these couples. Surveys
from both members of 360 couples were returned, for a response rate of 57.7%
of the 624 couples who were sent surveys. Each couple was mailed a check for
$40 once completed surveys had been received from both members. Of the
360 couples who responded, 309 met all of the criteria of the study, including
income.
One year later, a second wave of surveys was mailed to these 309 couples
to assess the changes in tested variables over time. Both members of 234 of
the 309 couples returned surveys, for a return rate of 75.6%. Following the
receipt of both partners’ surveys, an appreciation fee of $40 per couple was
again provided. Thus, couples participating in both waves of the survey re-
ceived a total of $80. Analyses for the present study were based on those
couples who returned surveys for both Wave 1 and Wave 2 (N = 234 couples).

Measures

The survey questionnaire was designed to assess, among other variables,

sociodemographic information, W → F Conflict, F → W Conflict, Job Satisfac-
tion, and Family Satisfaction for each participant. The scales used to measure
these variables were adapted from various previous studies, as shown below.
Reliabilities provided are based on Wave 1 data.

Sociodemographic data. Data on gender, race, number of hours worked per

week, gross annual household income, number of children aged 18 years or
under living at home, and number of hours per week spent caring for aging
parents were collected. Number of hours spent caring for children was not
assessed, due to respondent difficulty with estimating this. Instead, child care
demands were assessed by asking respondents how many of their children
aged 18 or younger lived in the household at least three days a week. In addi-
tion, each respondent was asked to indicate whether they, their spouse or
partner, or both equally took the most responsibility for these children.

Job satisfaction. The job satisfaction measure used was adapted from Hack-
man and Oldham’s (1975) Job Diagnostic Survey, which measures “the degree
to which the employee is satisfied and happy with the job” (p. 162). The origi-
nal measure was composed of five items, with a seven-point agree/disagree
response option. Two items were reversed scored. Hackman and Oldham re-
ported an internal consistency reliability of .76. In the present study, the
same five items were used; however, the responses were reported on a 5-point
scale, from strongly disagree (1) to strongly agree (5). The average score of
these five items was used as the indicator of this construct. The internal con-
sistency reliability was .69 for wives and .72 for husbands. A sample job satis-
faction item is, “I am generally satisfied with the kind of work I do in my job.”

Family satisfaction. To measure Family Satisfaction, Kopelman et al.

(1983) used five items adapted from Hackman and Oldham’s (1975) measure
of Job Satisfaction. In their study, the measure had an internal consistency
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 87

reliability of .90. The Family Satisfaction scale used in the present study was
adapted from Kopelman et al.’s measure to be appropriate for use with sand-
wiched generation couples. This scale consisted of three items. One item was
reversed scored. Responses were reported on a five-point scale from strongly
disagree (1) to strongly agree (5). The average score of these three items was
used as the indicator of this construct. The internal consistency reliability
was .65 for wives and .72 for husbands. These three family satisfaction items
are: (1) I am generally satisfied with the role I play in my family; (2) I fre-
quently think I would like to change my family situation; and (3) Generally
speaking, I am very satisfied with my family.

Work-to-family conflict and family-to-work conflict. Work-Family Conflict

was measured using a two-dimensional measure developed by Netemeyer et
al. (1996). This measure consisted of 10 items coded such that 1 = strongly
disagree and 5 = strongly agree. Five interrole conflict items, with an internal
consistency reliability of .86 in Netemeyer et al., measured interference from
work to family (i.e., W → F Conflict). The average score of these five items
was used as the indicator of this construct. The other five items, with an
internal consistency reliability of .88 in Netemeyer et al., reflected the in-
verse: interference from family to work (i.e., F → W Conflict). Similar to W →
F Conflict, the average score of these five items was used as the indicator of
F → W Conflict. In the present study, the W → F Conflict measure had an
internal consistency of .91 for wives and .90 for husbands. The F → W Conflict
measure had internal consistency reliabilities of .88 for both wives and hus-
bands. A sample W → F Conflict item is, “The demands of my work interfere
with my home and family life.” A sample F → W Conflict item is, “The de-
mands of my family or spouse/partner interfere with work-related activities.”

Data Analysis

Data were analyzed separately for wives and husbands for two reasons: (1)
to avoid violating statistical assumptions of independence of data; and (2) to
examine whether gender differences existed in the relationships between
W → F Conflict and F → W Conflict. The findings on gender differences in
WFC have been inconsistent in previous research (e.g., Eagle et al., 1997;
Hammer, Allen, & Grigsby, 1997).
The analytical steps for determining mediating effects in this study fol-
lowed the suggestions of Baron and Kenny (1986). In order to determine that
a mediating effect exists, three regression equations need to be calculated,
and four conditions must hold. First, the mediator is regressed on the inde-
pendent variable; second, the dependent variable is regressed on the indepen-
dent variable; and third, the dependent variable is regressed on both the inde-
pendent variable and on the mediator. The following four conditions must
hold to demonstrate that a variable is a mediator: (1) the independent vari-
ables must affect the mediator; (2) the independent variables must affect the
dependent variable; (3) the mediator must affect the dependent variable in
the third equation; and (4) the effect of the independent variables on the de-
pendent variable in the third equation must no longer be significant.
88 Journal of Family and Economic Issues

In the present study then, three regression equations were estimated in

each mediating effect between the two forms of WFC through satisfaction
variables. Wave 1 data were used for the independent variables (W → F Con-
flict and F → W Conflict, respectively), the means of Wave 1 and Wave 2 data
were used for the mediators (Job Satisfaction, Family Satisfaction), and Wave
2 data were used for the dependent variables (F → W Conflict, W → F Con-
flict). Using the mean of Wave 1 and Wave 2 data for the mediator variables
seemed to be the most approximation to a mid-point of the data from our
perspective.

Results

Descriptive statistics are reported in Table 1 for wives and hus-

bands for all study variables. T-tests revealed that there were no sig-
nificant differences between husbands and wives on any of the study
variables. Table 2 shows the correlation matrices for the major study
variables for both Wave 1 and Wave 2 data.

Additional Descriptive Data for Participants in the Wave 1 Survey

Three hundred and nine eligible couples completed and returned

the first wave of surveys. Wives had a mean age of 41.5 years (SD =
6.0) and husbands averaged 43.5 years of age (SD = 6.3). Ninety-five
percent of husbands and 94% of wives were Caucasian. The average

TABLE 1

Means, Standard Deviations, and Reliabilities for Wives and Husbands

Wives Husbands

Variable M SD n M SD n t p

Wave 1 W → F conflict 2.96 1.00 233 3.12 .93 233 −1.89 .06
Wave 2 W → F conflict 2.76 1.07 217 2.93 1.08 217 −1.81 .07
Wave 1 F → W conflict 2.30 .81 233 2.18 .68 233 1.90 .06
Wave 2 F → W conflict 2.09 .82 217 2.06 .70 217 .46 .65
Wave 1 Job satisfaction 3.48 .68 233 3.44 .71 233 .73 .46
Mean of Wave 1 and 2 Job
satisfaction 3.52 .63 234 3.46 .67 234 .97 .33
Wave 2 Job satisfaction 3.58 .75 216 3.50 .77 216 1.10 .27
Wave 1 Family satisfaction 3.97 .70 233 4.00 .75 233 −.61 .55
Mean of Wave 1 and 2 Family
satisfaction 3.97 .61 234 4.00 .66 234 −.57 .57
Wave 2 Family satisfaction 3.99 .70 216 4.01 .70 216 −.32 .75
 TABLE 2

Correlation Matrix of Major Study Variables for Wives and Husbands

1 2 3 4 5 6 7 8 9 10

Wives
1. Wave 1 W → F conflict —
2. Wave 2 W → F conflict .570** —
3. Wave 1 F → W conflict .596** .502** —
4. Wave 2 F → W conflict .330** .470** .487** —
5. Wave 1 Job Satisfaction −.321** −.277** −.187** −.061 —
6. Mean Job Satisfaction −.302** −.379** −.179** −.091 .874** —
7. Wave 2 Job Satisfaction −.221** −.386** −.122 −.102 .541** .891** —
8. Wave 1 Family Satisfaction −.286** −.192** −.157** −.132* .132* .166* .165* —
9. Mean Family Satisfaction −.328** −.291** −.199** −.170* .157* .200** .203** .870** —
10. Wave 2 Family Satisfaction −.291** −.311** −.170* −.158* .145* .191** .192** .502** .868** —
Husbands
1. Wave 1 W → F conflict —
2. Wave 2 W → F conflict .541** —
3. Wave 1 F → W conflict .370** .311** —
4. Wave 2 F → W conflict .245** .433** .448** —
5. Wave 1 Job Satisfaction −.251** −.222** .039 −.059 —
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin

6. Mean Job Satisfaction −.254** −.290** .007 −.130 .885** —

7. Wave 2 Job Satisfaction −.170* −.291** −.037 −.165* .566** .897** —
8. Wave 1 Family Satisfaction −.208** .265** .259** .172** —
9. Mean Family Satisfaction −.226** −.283** −.316** −.287** .240** −.252** .179** .903** —
10. Wave 2 Family Satisfaction −.195** −.294** −.284** −.302** .127 .155* .146* .573** .882** —

**Correlation is significant at the 0.01 level (2-tailed).

*Correlation is significant at the 0.05 level (2-tailed).
89
90 Journal of Family and Economic Issues

number of children aged 18 or younger living at home was 1.8 (SD =

0.8). Fifty-five percent of the wives identified themselves as the person
most responsible for the care of the children in the household. The
remaining 45% of wives reported that they and their spouses shared
this responsibility equally. For husbands, 2% identified themselves as
primarily child care providers, 51% reported that they and their
spouses shared this responsibility equally, and 47% reported that
their spouses were mostly responsible for child care issues. On aver-
age, wives spent 9.57 hours (SD = 11.5) and husbands spent 7.78
hours (SD = 12.1) per week taking care of aging parents. Wives worked
an average of 38 hours per week (SD = 9.7), while husbands worked
an average of 49 hours per week (SD = 11.4). The average gross an-
nual household income was $71,161; the median gross household in-
come was $62,500.

Additional Descriptive Data for Wave 2 Survey Participants

One year later, 234 couples of the original 309 couples returned sur-
veys. As with the first wave of data, the average number of children
under 18 years of age living at home was 1.8. On average, wives re-
ported spending 9.39 hours and husbands 7.71 hours per week taking
care of aging parents. Wives worked an average of 38 hours per week,
and husbands worked an average of 48 hours per week. The average
annual gross household income was $73,035, with the median income
at $66,000.

Tests of Hypotheses

Hypothesis 1. Hypothesis 1 predicted that Job Satisfaction would

mediate the relationship between F → W Conflict and W → F Conflict.
Results from the regression analyses are reported in Tables 3 (wives)
and 4 (husbands).
For wives, based on Baron and Kenny’s (1986) three-regression pro-
cedure for testing mediator effects, the results were: (1) Wave 1 F →
W Conflict had a significant direct effect (Beta = −.17, p < .05) on
Mean Job Satisfaction; (2) Wave 1 F → W Conflict had a significant
direct effect (Beta = .50, p < .01) and accounted for 25.2% of the vari-
ance in Wave 2 W → F Conflict; and (3) when regressing Wave 2 W →
F Conflict on both Wave 1 F → W Conflict and Mean Job Satisfaction,
Mean Job Satisfaction had a significant effect on Wave 2 W → F Con-
flict (Beta = −.31, p < .01), as did Wave 1 F → W Conflict (Beta = .45,
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 91

TABLE 3

Regression Analyses for the Mediating Effect of Mean Job Satisfaction

on the Relationship Between Wave 1 F → W Conflict and Wave 2 W → F
Conflict for Wives

Beta F-value R square

Regression 1
F → W Conflict → Mean Job Satisfaction −.17* F(1, 222) = 6.52* .03
Regression 2
F → W Conflict → W → F Conflict .50** F(1, 222) = 74.96** .25
Regression 3
Mean Job Satisfaction −.31**
F → W Conflict → W → F Conflict .45** F(2, 221) = 57.85** .34

*p < .05. **p < .01.

p < .01). The results showed that the Beta value between Wave 1 F →
W Conflict and Wave 2 W → F Conflict remained significant, indicat-
ing that Mean Job Satisfaction did not mediate the relationship be-
tween F → W Conflict and W → F Conflict over time (see Table 3).
For husbands, the results were: (1) Wave 1 F → W Conflict did not
have a significant direct effect on Mean Job Satisfaction; (2) Wave 1
F → W Conflict had a significant direct effect (Beta = .31, p < .01) and
accounted for 9.8% of the variance on Wave 2 W → F Conflict; and (3)
when regressing Wave 2 W → F Conflict on both Wave 1 F → W Con-
flict and mean Job Satisfaction, Mean Job Satisfaction had a signifi-

TABLE 4

Regression Analyses for the Mediating Effect of Mean Job Satisfaction

on the Relationship Between Wave 1 F → W Conflict and Wave 2 W → F
Conflict for Husbands

Beta F-value R square

Regression 1
F → W Conflict → Mean Job Satisfaction .01 F(1, 223) = .01 .00
Regression 2
F → W Conflict → W → F Conflict .31** F(1, 223) = 24.29** .10
Regression 3
Mean Job Satisfaction −.30**
F → W Conflict → W → F Conflict .32** F(2, 222) = 25.23** .19

*p < .05. **p < .01.

92 Journal of Family and Economic Issues

cant effect on Wave 2 W → F Conflict (Beta = −.30, p < .01), as did

Wave 1 F → W Conflict (Beta = .32, p < .01). Therefore, there was no
indirect relationship between F → W and W → F Conflict through Job
Satisfaction; that is, job satisfaction did not mediate the relationship
between F → W and W → F Conflict. Only a direct effect between
Wave 1 F → W Conflict and Wave 2 W → F Conflict existed over time
(see Table 4).
To summarize, there was a direct relationship between Wave 1 F →
W Conflict and Wave 2 W → F Conflict over time for both husbands
and wives, with higher conflict at Wave 1 associated with higher con-
flict at Wave 2. Wave 1 F → W Conflict also had a direct impact on
Mean Job Satisfaction for wives but not husbands, with higher conflict
associated with lower job satisfaction. For both husbands and wives,
Mean Job Satisfaction explained a unique amount of variance in Wave
2 W → F Conflict above and beyond Wave 1 F → W Conflict, with
higher job satisfaction associated with lower conflict at Wave 2. How-
ever, Mean Job Satisfaction could not be established as a mediator of
the relationship between Wave 1 F → W Conflict and Wave 2 W → F
Conflict for either husbands and or wives, contrary to H1.

Hypothesis 2. Hypothesis 2 predicted that Family Satisfaction

would mediate the relationship between W → F Conflict and F → W
Conflict. Results from the regression analyses are reported in Tables
5 and 6.

TABLE 5

Regression Analyses for the Mediating Effect of Mean Family Satisfaction

on the Relationship Between Wave 1 W → F Conflict and Wave 2 F → W
Conflict for Wives

Beta F-value R square

Regression 1
W → F Conflict → Mean Family Satisfaction −.34** F(1, 222) = 28.45** .11
Regression 2
W → F Conflict → F → W Conflict .33** F(1, 222) = 27.22** .11
Regression 3
Mean Family
Satisfaction −.06
W → F Conflict → F → W Conflict .31** F(2, 221) = 14.01** .11

*p < .05. **p < .01.

Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 93

TABLE 6

Regression Analyses for the Mediating Effect of Mean Family Satisfaction

on the Relationship Between Wave 1 W → F Conflict and Wave 2 F → W
Conflict for Husbands

Beta F-value R square

Regression 1
W → F Conflict → Mean Family Satisfaction −.26** F(1, 223) = 11.94** .05
Regression 2
W → F Conflict → F → W Conflict .25** F(1, 223) = 14.18** .06
Regression 3
Mean Family
Satisfaction −.24**
W → F Conflict → F → W Conflict .19** F(2, 222) = 14.45** .12

*p < .05. **p < .01.

For wives, using Baron and Kenny’s (1986) three-regression proce-

dure for testing mediator effects, the results were: (1) Wave 1 W → F
Conflict had a significant direct effect (Beta = −.34, p < .01 ) on Mean
Family Satisfaction; (2) Wave 1 W → F Conflict had a significant di-
rect effect (Beta = .33, p < .01) and accounted for 10.9% of the variance
in Wave 2 F → W Conflict; and (3) when regressing Wave 2 F → W
Conflict on both Wave 1 W → F Conflict and Mean Family Satisfac-
tion, Mean Family Satisfaction did not have a significant effect on
Wave 2 F → W Conflict, but Wave 1 W → F Conflict did (Beta = .31,
p < .01). In order to demonstrate a mediating effect, Mean Family Sat-
isfaction must affect Wave 2 F → W Conflict. Furthermore, the effects
of Wave 1 W → F Conflict on Wave 2 F → W Conflict must no longer
be significant or should show a significant reduction from the second
equation. Contrary to expectations, the findings did not support those
conditions. Thus, Mean Family Satisfaction did not mediate the rela-
tionship between W → F Conflict and F → W Conflict for wives.
For husbands, the results were: (1) Wave 1 W → F Conflict had a
significant direct effect (Beta = −.26, p < .01) on Mean Family Satisfac-
tion; (2) Wave 1 W → F Conflict had a significant direct effect (Beta =
.25, p < .01) and accounted for 6% of the variance in Wave 2 F → W
Conflict; and (3) when regressing Wave 2 F → W Conflict on both
Wave 1 W → F Conflict and Mean Family Satisfaction, Mean Family
Satisfaction had a significant effect on Wave 2 F → W Conflict (Beta =
−.24, p < .01), as did Wave 1 W → F Conflict (Beta = .19, p < .01). In
94 Journal of Family and Economic Issues

order to show the mediating effect existed, the effect of Wave 1 W →

F Conflict in the third regression should no longer be significant or be
significantly reduced. These results showed that the Beta value be-
tween Wave 1 W → F Conflict and Wave 2 F → W Conflict remained
significant. Therefore, there was no indirect relationship between W →
F and F → W Conflict through Mean Family Satisfaction. Only a di-
rect effect between Wave 1 W → F Conflict and Wave 2 F → W Con-
flict existed over time for husbands.
To summarize this set of analyses, there was a direct positive rela-
tionship between Wave 1 W → F Conflict and Wave 2 F → W Conflict
for both husbands and wives over time: Higher conflict at Wave 1 was
associated with greater conflict at Wave 2. Wave 1 W → F Conflict
significantly predicted Mean Family Satisfaction for both husbands
and wives, with greater conflict associated with lower family satisfac-
tion. In addition, for husbands but not for wives, Mean Family Sat-
isfaction had a significant unique contribution to Wave 2 F → W
Conflict beyond Wave 1 W → F Conflict. However, despite previous
empirical and theoretical support for a mediating effect of Family Sat-
isfaction on the relationship between W → F and F → W Conflict, con-
trary to H2, no mediating effect could be established in this study.

Discussion

The primary goal of this study was to test the mediating effects of
Job and Family Satisfaction on the relationship between W → F and
F → W Conflict. These effects were suggested by Frone et al. (1997)
but could not be tested by them, given the cross-sectional nature of
their data. Contrary to expectations, the present study failed to find
significant mediating effects of either Job or Family Satisfaction on
the relationship between W → F and F → W Conflict for husbands or
wives. Only direct relationships were evident, such that as W → F
Conflict increased, F → W Conflict increased, and vice versa, and this
relationship held up over time. Furthermore, for both husbands and
wives, Job Satisfaction had a significant effect on Wave 2 W → F Con-
flict, and for husbands, Family Satisfaction had a significant effect on
Wave 2 F → W conflict.
The failure to demonstrate mediating effects of family role-related
satisfaction between the two forms of WFC may be due to the non-
specific nature of the measure. Future research should examine the
effects of satisfaction with each of the family roles (i.e., child care,
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 95

parent care, and spousal roles) as mediators of the relationships be-

tween the two forms of WFC. In addition, the family role satisfaction
measure used in this study did not achieve a high level of reliability.
For example, the original five-item family satisfaction measure used
by Hackman and Oldham (1975) had an internal consistency reliabil-
ity of .90, but in the current study, the three-item scale had an inter-
nal consistency reliability of .65 for wives and .72 for husbands. The
reduced number of items may well have negatively affected the scales’
reliability.
For both husbands and wives, there were significant direct effects
between the two forms of WFC overtime. Moreover, again for both
husbands and wives, Wave 1 F → W Conflict accounted for more vari-
ance in Wave 2 W → F Conflict than did Wave 1 W → F Conflict in
Wave 2 F → W Conflict. Furthermore, compared to those for hus-
bands, the reciprocal relationships between the two forms of WFC
were stronger for wives. Specifically, the present study found that for
wives, Wave 1 W → F Conflict accounted for 10.9% of the variance in
Wave 2 F → W Conflict, and Wave 1 F → W Conflict accounted for
25.2% of the variance in Wave 2 W → F Conflict. For husbands, Wave
1 W → F Conflict accounted for 6% of the variance in Wave 2 F → W
Conflict, and Wave 1 F → W Conflict accounted for 9.8% of the vari-
ance in Wave 2 W → F Conflict. These findings suggest that future
research on WFC should consider the direct effects of both types of
WFC on one another, and also that future studies of the work-family
interface should take potential gender differences into account.

Implications

The findings from this study have both theoretical and practical im-
plications. First, the present study focused on WFC in dual-earner
couples in the sandwiched generation; prior research has devoted in-
sufficient attention to this group of employees, given that prevalence
estimates from this larger study’s national screening data indicate
that anywhere from 9%–13% of the dual-earner aged 30+ working
population is caring for both children and aging parents (Neal, Ham-
mer, Rickard, Isgrigg, & Brockwood, 1999). Second, by using a longitu-
dinal research design to test the effects between W → F Conflict and
F → W Conflict, the present study went beyond previous cross-
sectional research to more conclusively determining the effects of this
relationship over time between the two forms of WFC. This design
was responsive to suggestions by Frone et al. (1997) that future re-
96 Journal of Family and Economic Issues

search should examine parts of their model longitudinally to deter-

mine the aspects that are supported over time. Third, compared to
previous research, the present study demonstrated an improved ana-
lytical approach for testing the mediating effects between the two
forms of WFC (Baron and Kenny, 1986). Kossek and Ozeki (1999) note
that more research is needed on the effects of implementing family-
friendly supports on both forms of WFC and on the resulting work
and family outcomes (e.g., withdrawal behaviors and job perfor-
mance). The findings of the present study further suggest that organi-
zational interventions such as family-friendly supportive programs
(e.g., Kossek & Ozeki, 1999) that help to alleviate W → F Conflict will
likely help to alleviate F → W Conflict over time. In other words, orga-
nizational efforts to decrease W → F Conflict may have the additional
resulting benefit of decreasing F → W conflict over time, leading to
improved attitudes and behaviors on the job. In addition, the findings
suggest that the longitudinal effects of both forms of WFC may be
stronger for women, as compared to men. Thus, attempts to reduce
WFC may be particularly beneficial for women. Moverover, the find-
ing that the effects of F → W Conflict on W → F Conflict may be
stronger over time than the effects of W → F Conflict on F → W Con-
flict, for both husbands and wives, suggests that organizational sup-
ports that decrease F → W Conflict will have added salutary and long
term effects for family life, in the form of less family distress and more
family satisfaction (Frone et al., 1992, 1997). These findings indicate
that future research should be longitudinal in nature, should explore
the differential effects of family-friendly programs on men and women,
and should include measures both of Family-to-Work and Work-to-
Family Conflict at each measurement point.

Limitations

Despite its contributions, this research also has limitations. Two

important ones concern the comparability of the sample and measures
used in this study and these used by Frone et al. (1997). First, the
present study used a sample of dual-earner couples in the sandwiched
generation, while the Frone et al. study examined members of dual-
earner couples, regardless of their dependent care responsibilities.
Therefore, the unique characteristics of the present study’s sample,
with their multiple family roles, may have contributed to the in-
creased likelihood of direct, rather than indirect, effects between the
two dimensions of WFC. The individuals in this sample were dealing
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 97

with extensive work- and family-related demands and may not have
been as readily affected by these levels of satisfaction with either their
job or their family, compared to the sample used by Frone et al. Fur-
thermore, the two studies used different measures, suggesting that
the findings of either study may simply be an artifact of the measures
used. For example, the reliabilities of the satisfaction measures in the
present study were low and could have contributed to the failure to
detect mediating effects. At the same time, however, both job and fam-
ily satisfaction had direct effects on WFC in the expected directions.
Likewise, the reliabilities of the WFC measures in the present study
were considerably higher than those for the measures used by Frone
et al. Thus, it is possible that differences in the findings between the
two studies may be due to differences in the characteristics of the
samples and/or differences in measurement of key variables.
Besides the unique characteristics of the study’s sample and the
non-specific nature of the satisfaction measures, along with the low
reliability particularly of the measure of family satisfaction, several
other potential limitations of the study should be acknowledged, as
well. First, racial and ethnic differences cannot be tested because al-
most all of the respondents were Caucasian. Thus, future research
should replicate the present study with participants from diverse cul-
tural backgrounds. Second, because participation in this study was
limited to couples with a gross annual household income of at least
$40,000 a year, the results of this study cannot be generalized to cou-
ples whose household income was below $40,000 a year. Previous re-
search has demonstrated that financial resources help individuals bet-
ter manage work and family, leading to decreased WFC (Chapman et
al., 1994). Third, compared to previous research, this study addressed
the question of long-term effects in the WFC systems; nonetheless,
having only two data points restricted the contribution this study
could make to understanding the relationships between the trajecto-
ries of changes in predictors over time and the trajectory of changes
in outcomes over time. Having three data points better enables a
study to estimate change-over-time effects (Barnett & Brennan, 1997).
Furthermore, with three waves of data, a study could use Wave 1 data
for the measurement of predictors, Wave 2 data for the mediators, and
Wave 3 data for the criteria, for a more effective understanding of the
longitudinal effects of W → F Conflict on F → W Conflict and vice
versa. Lastly, another possible explanation for the non-significant me-
diator relationships could be that the one-year interval between Wave
1 and Wave 2 was not appropriate to test these particular relation-
98 Journal of Family and Economic Issues

ships. A study design that incorporates multiple waves of data with

relatively short time lags might be more fruitful.

Conclusions

At a time when the number of dual earner couples in the workforce

continues to increase and roles in the family are constantly changing,
it is crucial to understand the dynamics of the work-family interface
and the implications of the simultaneous impact of the work and fam-
ily systems on one another. This study integrates theory and research
in the examination of the reciprocal relationships between the two
forms of WFC for dual earner couples in the sandwiched generation.
Expanding upon Frone et al.’s (1997) cross-sectional research, this
study uses longitudinal data to test the mediating effects of job and
family satisfaction on the relationships between W → F Conflict and
F → W Conflict. The present study, therefore, provides a more strin-
gent test of Forne et al.’s (1997) model of the work-family interface.
Contrary to the indirect relationships suggested by Frone et al. (1997),
the present study’s results indicate no mediating effects of job satisfac-
tion or family satisfaction on the relationships between the two forms
of WFC for either husbands or wives over time. Instead, the results
provide support for the notion that if one’s work-related problems be-
gin to interfere with the completion of one’s family-related obligations,
these unfulfilled family obligations will begin to interfere with one’s
day-to-day functioning at work and vice versa, as suggested by Frone
et al. (1992). These findings suggest that future research on WFC
should consider the direct relationships of both types of WFC on one
another. They also demonstrate that future research on the work-
family interface must further explore the dynamic relationships of the
work-family system over time.

References

Adams, G. A., King, L. A., & King, D. W. (1996). Relationships of job and family involve-
ment, family social support, and work-family conflict with job and life satisfaction.
Journal of Applied Psychology, 81(4), 411–420.
Aryee, S. (1992). Antecedents and outcomes of work-family conflict among married pro-
fessional women: Evidence from Singapore. Human Relations, 24, 813–837.
Barnett, R. C., & Brennan, R. T. (1997). Change in job conditions, change in psychologi-
cal distress, and gender: A longitudinal study of dual-earner couples. Journal of
Organizational Behavior, 18, 253–274.
Y. H. Huang, L. B. Hammer, M. B. Neal, and N. A. Perrin 99

Barnett, R. C., Davidson, H., & Marshall, N. L. (1991). Physical symptoms and the
interplay of work and family roles. Health Psychology, 10(2), 94–101.
Barnett, R. C., & Marshall, N. L. (1993). Men, family-role quality, job-role quality, and
physical health. Health Psychology, 12(1), 48–55.
Barnett, R. C., & Rivers, C. (1996). She works he works: How two-income families are
happier, healthier, and better off. San Francisco: Harper Collins.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173–1182.
Burke, R. J. (1988). Some antecedents and consequences of work-family conflict. Jour-
nal of Social Behavior and Personality, 3(4), 287–302.
Chapman, N. J., Ingersoll-Dayton, B., & Neal, M. B. (1994). Balancing the multiple roles
of work and caregiving for children, adults, and elders. Washington, DC: American
Psychological Association.
Crouter, A. (1984). Spillover from family to work: The neglected side of the work-family
interface. Human Relations, 37, 425–552.
Durity, A. (1991). The sandwich generation feels the squeeze. Management Review,
80(12), 38–41.
Eagle, B. W., Miles, E. W., & Icenogle, M. L. (1997). Interrole conflicts and the perme-
ability of work and family domains: Are there gender differences? Journal of Voca-
tional Behavior, 50, 168–184.
Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents of outcomes of work-
family conflict: Testing a model of the work-family interface. Journal of Applied
Psychology, 77(1), 65–78.
Frone, M. R., Yardley, J. K., & Markel, K. S. (1997). Developing and testing an integra-
tive model of the work-family interface. Journal of Vocational Behavior, 51(2), 145–
167.
Goff, S. J., Mount, M. K., & Jamison, R. L. (1990). Employer supported child care, work/
family conflict, and absenteeism: A field study. Personnel Psychology, 43(4), 793–
809.
Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family
roles. Academic of Management Review, 10, 76–88.
Gutek, B. A., Searle, S., & Klepa, L. (1991). Rational versus gender role explanations
for work-family conflict. Journal of Applied Psychology, 76, 560–568.
Hackman, J. R., & Oldham, G. R. (1975). Development of the Job Diagnostic Survey.
Journal of Applied Psychology, 60, 159–170.
Hammer, L., Allen, E., & Grigsby, T. (1997). Work-family conflict in dual-earner cou-
ples: Within-individual and crossover effects of work and family. Journal of Voca-
tional Behavior, 50, 185–203.
Higgins, C. A., Duxbury, L. E., & Irving, R. H. (1992). Work-family conflict in the dual-
earner family. Organization Behavior and Human Decision Processes, 51, 51–75.
Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organi-
zational stress: Studies in role conflict and ambiguity. New York: John Wiley.
Kanter, R. M. (1977). Work and family in the United States: A critical review and
agenda of research and policy. New York: Russell Sage Foundation.
Kopelman, R. E., Greenhaus, J. H., & Connolly, T. F. (1983). A model of work, family,
and interrole conflict: A construct validation study. Organizational Behavior and
Human Performance, 32, 198–215.
Kossek, E. E., & Ozeki, C. (1999). Work-family conflict, policies, and the job-life satisfac-
tion relationship: A review and directions for organizational behavior-human re-
sources research. Journal of Applied Psychology, 83(2), 139–149.
Neal, M. B., Chapman, N. J., Ingersoll-Dayton, B., & Emlen, A. C. (1993). Balancing
work and care loom is giving for children, adults, and elders. Newbury Park, CA:
Sage.
Neal, M. B., Hammer, L. B., Rickard, A., Isgrigg, J. L., & Brockwood, K. J. (1999, No-
100 Journal of Family and Economic Issues

vember). Dual-earner couples in the sandwiched generation: Who they are, what
they do, how they manage. Paper presented at the 52nd annual scientific meeting of
the Grontological Society of America, San Francisco, CA.
Near, J., Rice, R., & Hunt, R. (1980). The relationship between work and nonwork do-
mains: A review of empirical research. Academy of Management Review, 5, 415–
429.
Netemeyer, R. G., Boles, J. S., & McMurrian R. (1996). Development and validation of
work-family conflict scales. Journal of Applied Psychology, 81, 400–410.
Nichols, L. S., & Junk, V. W. (1997). The sandwich generation: Dependency, proximity,
and task assistance needs of parents. Journal of Family & Economic Issues, 18(3),
299–326.
Offermann, L. R., & Gowing, M. K. (1990). Organizations of the future: Changes and
challenges. American Psychologist, 45(2), 95–108.
Pleck, J. H. (1977). The work-family role system. Social Problems, 24(4), 417–427.
Raphael, D., & Schlesinger, B. (1993). Caring for elderly parents and adult children
living at home: Interactions of the sandwich generation family. Social Work Re-
search and Abstracts, 29(1), 3–8.
Tausig, M., & Fenwick, R. (2001). Unbinding time: Alternate work schedules and work-
life balance. Journal of Family & Economic Issues, 22(2), 101–119.
Thomas, L. T., & Ganster, D. C. (1995). Impact of family-supportive work variables on
work-family conflict and strain: A control perspective. Journal of Applied Psychol-
ogy, 80(1), 6–15.
U. S. Bureau of Labor Statistics (1995). Women in the workforce: An overview. Washing-
ton, DC: Government Printing Office.
Youngblood, S. A., & Chambers-Cook, K. (1984). Child care assistance can improve em-
ployee attitude and behavior. Personnel Administrator, 29(2), 45–46, 93–95.