PROGRESS ON GOALS AND METRICS:
The Cost to Working Parents of the Mismatch Between Their Work
Schedules and the Schedules of the Systems on Which They Depend
(The Family Schedule Coordination Study)
In general, given the exploratory nature of this study and the small sample size, we were
nevertheless able to find a pattern of results that formed a coherent picture to be further examined
in future hypothesis-testing studies.
Goal 1: To quantify the relationship between the stress resulting from the mismatch
(mismatch stress) between the work schedules of parents of school age children and the
multiple delivery systems (i.e. schools, after school care arrangements, transportation system)
on which they depend and negative outcomes.
We developed a 36-item psychometrically sound measure of community resource fit that covers
the domains of work, public transportation, school, school transportation, after-school programs,
and after-school transportation. As shown in the Appendix, the psychometrics for the overall
scale were excellent for both mothers and fathers, and ranged from adequate to excellent for the
six subscales for both mothers and fathers.
We then conducted a series of regression analyses estimating the relationship between resource
fit and outcomes. Because all participants in this initial study lived in Waltham, there was
insufficient variance on several of the community resource fit subscales to include them in our
analyses; therefore, we focus on work resource fit and school resource fit. Also, given the small
sample size and the restricted range on resource fit in this sample, it is difficult to demonstrate
any but the strongest of effects, so our findings should be viewed as a quite conservative estimate
of the real relationship between community resource fit and outcomes.
We conducted parallel series of simultaneous regression analyses for mothers and for fathers, one
set with school resource fit as the predictor (n = 86 for mothers and 59 for fathers) and the other
with work resource fit (n = 66 for mothers and 54 for fathers) as the predictor. In all cases, we
used the same set of controls: work hours, negative affectivity, logged per capita household
income, and number of children at home. The outcome variables were psychological distress, job
disruptions, work interfering with family, family interfering with work, life satisfaction, and
marital-role, parent-role, and job-role quality.
For mothers, the only finding linking school resource fit to any of the eight outcome variables
was a trend for high school resource fit to predict low job disruptions. In a larger, multicommunity sample with greater variance on school resource fit, we would expect the findings to
be stronger. With that caveat in mind, it is nevertheless interesting to note that the school
schedule, including start and end times, extracurricular events, and parent meetings predict
parental job disruptions.
There were several findings linking work schedule fit to our outcomes among the mothers. High
work resource fit was significantly related to low psychological distress, few job disruptions, low
work interfering with family and family interfering with work, and high job-role quality. There
was also a trend for high work resource fit to predict high life satisfaction. Thus, even in this
small and restricted sample, we found several strong and interesting results that tell a coherent
story. It is noteworthy that a number of the items in this subscale reflect workplace flexibility.
For the fathers, there were fewer significant findings, some of which replicated those for the
mothers. High school resource fit was significantly related to low psychological distress and high
job-role quality. High work resource fit was significantly related to low work interfering with
family and high job-role quality.
Thus, we have a measure with good psychometrics. Preliminary findings suggest that this
measure has important relationships to several outcomes of interest to work-family researchers
and warrants further study. Specifically, the measure should be expanded to include resources
relevant to working families with caregiving responsibilities for preschool-aged children and
elders and disabled adults. In addition, to further elucidate the linkages between community
resource fit and outcomes of interest, larger multi-community samples are needed.
Goal 2: To identify the conditions under which mismatch stress is especially high.
We identified several conditions – specifically, parent gender, mothers’ marital status, and
household income – that might moderate the relationship between community resource fit and
outcomes. The preliminary findings described below suggest that the pattern of results is similar
for mothers and fathers. We conducted a series of moderator regression analyses to estimate
whether the relationship between school and work resource fit and our outcomes differed by
marital status and per capita household income. We also examined more informally the impact of
gender on the relationships described in Goal #1.
We tested whether mothers’ marital status (married [n = 58] vs. single [n = 28]) moderated the
relationship between school and work resource fit and our set of outcome variables. There was
only one single father in our sample, so we were unable to test whether marital status affected
relationships between community resource fit and outcomes among fathers. Among the mothers,
we found that marital status was a significant moderator of the relationship between work
resource fit and work interfering with family such that the ameliorating effect of work resource
fit on work interfering with family is much stronger for the married mothers. There was also a
significant moderating effect of marital status on the relationship between school resource fit and
job disruptions such that married mothers benefit from school resource fit whereas single
mothers do not. Other analyses with this sample suggest that one explanation for both of these
effects may be that single mothers in our sample have a larger network of outside sources of
child care help and have older children than do the married mothers in our sample. Thus, they
seem to be relatively unaffected by these aspects of community resource fit. Clearly, the small
size of this sample requires caution in generalizing these intriguing findings.
We also tested the moderating effect of per capita household income on the relationships
between school and work resource fit and our set of outcomes. We found no moderating effects
of income on any of the outcomes among mothers. Among fathers, there was one significant
moderating effect of income on the relationship between school resource fit and job-role quality
such that high income buffers fathers from the effects of school resource fit on job-role quality.
In contrast, low-income fathers benefit from high school resource fit in terms of job-role quality.
We found two significant interaction effects involving income and work resource fit. First,
income moderated the relationship between work resource fit and distress such that when work
resource fit is poor, high-income fathers report significantly higher distress. In this instance,
instead of being buffered by their high income, they are more vulnerable to poor work resource
fit. Speculatively, their expectations for high work resource fit may be greater than the reality,
leading them to be more distressed than their low-income counterparts. We also found that
income significantly moderated the relationship between work resource fit and job-role quality
such that at low levels of work resource fit, high-income fathers report lower job-role quality
than their low-income counterparts. Once again, we see that the high-income fathers are more
sensitive to work resource fit than are low-income fathers.
Goal 3: To identify the processes by which mismatch stress leads to negative outcomes.
We conducted a series of mediator regression analyses to estimate whether the 3-item NSCW
measure of workplace flexibility mediated the significant relationships between school and work
resource fit and our outcomes. We found no evidence that workplace flexibility, as assessed by
the 3-item NSCW measure, mediated any of the significant relationships reported above. It is
entirely possible that in this sample, one of the other flexibility indicators in our data set might
reveal a mediation effect.
Goal 4: To identify the range of flexibility options that different working parents believe might
ease their mismatch stress.
Employed participants were asked to indicate whether each of 18 flexibility and family-friendly
policies and practices is available to them and, if so, how much it reduces stress. For mothers,
having informal flexibility (i.e., flexibility to come to work later, leave work earlier, or leave and
then return to work to deal with family matters when necessary) received a mean rating of 3.42,
halfway between “considerably” and “extremely” in terms of how much it would reduce stress,
while formal flextime (i.e., the ability to change regular starting and quitting times) received a
mean rating of 3.19. For fathers, informal flexibility received a mean rating of 3.33 and formal
flextime received a mean rating of 3.35. For fathers, these two forms of flexibility were rated
among the top three, whereas for mothers, informal flexibility was also rated among the top
three.
Metric 1: Understand the relationship between degree of mismatch stress and outcomes such
as psychological distress, stress related health problems, self reported job productivity, and
quality of life indicators such as work-family conflict, life satisfaction, and marital-, parent-,
and job-role quality.
As described above, to this end, we conducted a series of regression analyses separately for
mothers and fathers. We determined that for mothers, school resource fit showed a trend to
predict low job disruptions and work resource fit was significantly related to low psychological
distress, few job disruptions, low work interfering with family and family interfering with work,
and high job-role quality. There was also a trend for high work resource fit to predict high life
satisfaction. For the fathers school resource fit was significantly related to low psychological
distress and high job-role quality, and work resource fit was significantly related to low work
interfering with family and high job-role quality.
Metric 2: Identify the conditions (e.g., family structure, exempt status, income, parent(s) and
child(ren)s gender and age, size of workplace) under which mismatch stress is especially high.
As described above, among the mothers, marital status was a significant moderator of the
relationship between work resource fit and work interfering with family and between school
resource fit and job disruptions. We found no moderating effects of per capita household income
on any of the outcomes among mothers. Among fathers, income was a significant moderator of
the relationship between school resource fit and job-role quality, between work resource fit and
psychological distress, and between work resource fit and job-role quality.
Metric 3: Identify the current workplace flexibility policies or practices to ascertain whether
they reduce the relationship between mismatch stress and negative outcomes.
As described above, we found that both mothers and fathers rated informal workplace flexibility
and formal flextime policies as extremely helpful in reducing stress.
Metric 4: Identify the range of flexibility options parents suggest as potentially helpful in
reducing mismatch stress and to determine whether these suggested policies or practices vary
by such factors as the family structure, income, exempt status, parent(s) and child(ren)s
gender and age.
There were no major differences in the helpfulness ratings that parents in different types of
families gave to the various workplace supports. Parents in two-earner families, families with
one main breadwinner, and single-parent families all rated informal flexibility among their top
three most helpful workplace supports. Parents in all family types also rated having paid time off
to deal with family matters and being able to make and take telephone calls at work to deal with
family matters among their top three most helpful workplace supports.
To examine whether children’s age was related to which workplace supports parents rated as
most helpful, we estimated correlations between the average age of the children in a family and
each of the 18 workplace supports. Overall, parents of younger children, especially fathers, rated
a number of workplace supports as especially effective in reducing stress. Among mothers, those
with younger children particularly appreciated backup child care or after-school care for
emergencies ( r = -.27, p < .05) and showed a trend to rate being able to leave work at a regular
time more highly ( r = -.22, p = .10) than did their counterparts with older children. Among
fathers, those with younger children gave significantly higher ratings to a supervisor or manager
who is understanding about family matters ( r = -.29, p < .05), formal flextime ( r = -.36, p < .05),
formal telecommuting ( r = -.44, p < .05), employer-subsidized after-school programs ( r = -.38, p
< .05), and parenting education ( r = -.30, p < .05) than did their counterparts with older children.
Fathers of younger children also gave marginally higher ratings to being able to telecommute if
necessary ( r = -.27, p < .10) and access to information about or referrals to local after-school
programs ( r = -.28, p < .10) than did their counterparts with older children. Thus, fathers of
younger children appear to be even more appreciative of a number or workplace supports than
are mothers of younger children. Given this finding, companies should work to ensure that men
are as well-informed about and feel equally welcome to take advantage of available workplace
supports as women.
Finally, we examined whether income was related to helpfulness ratings by estimating
correlations between the family’s household income per capita and each of the 18 workplace
supports. Contrary to expectations, parents with lower incomes did not differ from their higherincome counterparts in their ratings. The only significant or marginally significant correlations
were that mothers with higher incomes rated informal flexibility significantly higher than their
lower-income counterparts ( r = .27, p < .05) and fathers with higher incomes gave marginally
higher ratings to informal flexibility ( r = .25, p < .10) and informal telecommuting ( r = .28, p <
.10). This may be due to the fact that higher-income parents are more likely to have the kinds of
jobs that are amenable to flex-time and flex-place policies.
Metric 5: Publish four to six articles in refereed academic journals, present findings to
various professional and lay audiences, and publicize our results through the resources of the
Community, Families & Work Program, Brandeis University.
Clearly, we have enough findings to support the preparation of four to six academic articles
which we plan to work on this fall. We have already presented the development of the instrument
and selected findings to the 23rd Annual Claremont Symposium on Applied Social Psychology
entitled “Work and Families: Changing Realities.” This talk will be elaborated as a chapter in a
forthcoming book edited by Diane Halpern and Amy Marcus-Newhall to be published by
Erlbaum. In addition, we will be presenting this material at the UCLA Sloan Center on Everyday
Lives of Families in January. We have also prepared an academic article that will be submitted to
Community, Work and Family. We have also submitted two abstracts for presentation at the 2nd
International Conference on Community, Work and Families, Lisbon, Portugal, and we have two
articles on those findings under review at the Journal of Occupational Health Psychology.
Finally, we are preparing an academic article reporting on HLM analyses predicting
psychological distress from work resource fit and school resource fit among mothers and fathers
in two-parent families. These documents are appended. We will also be conducting a series of
additional within-couples analysis with the same community resource fit predictors but different
outcomes (i.e., job-role quality, marital-role quality, parent-role quality, and work-family
conflict) to be published in various journals.
APPENDIX:
Community Resource Fit
As you know, we’re interested in how families with school-age children coordinate the work,
school, after-school, and transportation schedules of all household members. The following
items are statements others have made about things that make it harder or easier for them to do
that.
When you think about the way things are now, how satisfied are you with each of the following?
Completely
Dissatisfied
Mostly
Dissatisfied
Slightly
Dissatisfied
Neither
Satisfied
Nor
Dissatisfied
Slightly
Satisfied
Mostly
Satisfied
Completely
Satisfied
1
2
3
4
5
6
7
(Overall alphas: Mothers = .86 / Fathers = .90)
CRF: Work (Mothers = .77 / Fathers = .80)
1. The way your work schedule fits with your child(ren)’s schedule(s)
2. The flexibility available at your workplace to handle emergencies
3. The flexibility available at your workplace to attend to family needs
4. Your ability to work at home if necessary
5. Your ability to bring child(ren) to work if necessary
CRF: Public Transportation (Mothers = .91 / Fathers = .81)
6. The way the public transportation schedule fits with your own travel needs
7. The way the public transportation schedule fits with your child(ren)’s travel needs
8. The convenience of your own access to public transportation
9. The convenience of your child(ren)’s access to public transportation
10. The way that the available public transportation routes meet your own travel needs
11. The way that the available public transportation routes meet your child(ren)’s travel needs
CRF: School (Mothers = .73 / Fathers = .69)
12. The time(s) your child(ren)’s school(s) start(s) in the morning
13. The time(s) your child(ren)’s school(s) let(s) out in the afternoon
14. The way different schools in the community coordinate their schedules with each other
15. The scheduling of school meetings, parent conferences, and events
16. Communication between the school(s) and parents
17. Scheduling of extracurricular activities
CRF: School Transportation (Mothers = .77 / Fathers = .73)
18. Where the children wait to be picked up by the school bus in the morning
19. Where the children wait to be picked up by the school bus in the afternoon
20. The reliability of school bus transportation to and from school
21. The availability and scheduling of late buses
CRF: After-school Programs (Mothers = .88 / Fathers = .85)
22. The availability of after-school programs
23. The cost of after-school programs
24. The location of after-school programs
25. The scheduling of after-school programs
28. Communication between after-school program providers and parents
29. The expectations of after-school program providers for parental involvement
CRF: After-school Transportation (Mothers = .95 / Fathers = .91)
30. The availability of transportation to and from after-school activities
31. The scheduling of transportation to and from after-school activities
32. The reliability of transportation to and from after-school activities
36. The cost of transportation to and/or from after-school activities
Community: The missing link in
the work-family literature
Rosalind Chait Barnett
Karen C. Gareis
Women’s Studies Research Center
Brandeis University
March 25, 2006
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Community is an important
missing link in the work-family
literature.
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Community
Well-Being
Family
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Work
What do we mean by community?
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
A community is a
• group of people who live in a common territory, have a
common history and shared values, participate together in
various activities, and have a high degree of solidarity
• group of people who are socially interdependent, who
participate together in discussion and decision making, and
who share certain practices.
• sharing association involving two or more human beings.
(cited by Voydanoff, 2001, p, 138).
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
RESIDENTIAL COMMUNITY
Real geographical community that shapes
family life and work (Bookman, 2005, p.
144).
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
What do we mean by community
resources?
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
COMMUNITY RESOURCE FIT
is the extent to which resources are available
and well-matched to the needs of working
families
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Community Resources
Work
School
After-school arrangements
Transportation
Public transportation
School transportation
After-school
transportation
Community
Resource
Fit
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Quality of Life
Our approach focuses on cognitive
appraisals
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
THREE FAMILY FORMS
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Two-Phase Research Project
Phase One
Develop and evaluate an overall measure of
community resource fit
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Development of the Community
Resource Fit (CRF) Scale
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
• Waltham has a broad range on SES.
• Waltham has a median household income of
$54,000.
• 7.0% of the population live below the poverty
line, and 1.9% receive public assistance and
17.1% are minorities
• Waltham is more racially diverse than many
towns in the Greater Boston Metropolitan Area
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
SAMPLE CHARACTERISTICS
Mothers
Fathers
Work Hours
Dual-earner
One main breadwinner
37.5 (11.6)
10.6 (16.29)
47.6 (9.4)
41.2 (24.4)
Single-parent
43.0
50.0
Per capita household income
Dual-earner
One main breadwinner
Single-parent
(8.8)
24,371
19,044
15,956
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
(9,889)
(7,149)
(10,261)
(--)
Mothers
Number of children
Dual-earner
2.1
(0.8)
One main breadwinner
2.4
(0.7)
Single-parent
1.8
(0.7)
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Here a few of the questions that we asked:
• Where (do/does) your school-age child(ren)
spend (his/her/their) time after school?
• What are (his/her) usual weekly child-care
after-school arrangements?
• For your school-age child(ren), how (do/does)
(he/she/they) get to and from school and to and
from their after-school arrangements?
• Do your child(ren)’s transportation
arrangements vary from day to day or at
different times of year?
™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
• How does your school-age child(ren)’s schedule
– start times and end times, early release days,
holidays and so on – work with your family’s
scheduling needs?
• How does the local transportation schedules work
with your family’s other scheduling needs?
• What features of your community affect your
ability to coordinate family schedules? Here we
mean the schools, the after-school programs, the
transportation system, and so on?
™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
SCALE DEVELOPMENT
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
SCALE PSYCHOMETRICS
Mothers
Alpha
Mean
Fathers
.856
4.90
.896
4.97
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
SIX CATEGORIES OF
RESOURCE FIT
ALPHA
Work
Public transportation
School
School Transportation
After-School Programs
After-school Transportation
Mothers
.77
.91
.73
.77
.88
.95
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Fathers
.80
.81
.69
.73
.85
.91
Because all study participants lived in
Waltham, there was insufficient
variance on several of the community
resource fit subscales to include them
in our analyses.
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
WORK RESOURCE FIT
SCHOOL RESOURCE FIT
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
WORK RESOURCE FIT
How satisfied are you with the:
way your work schedule fits with your child’s
schedule
the flexibility available at your workplace to
handle emergencies
your ability to work at home if necessary
your ability to bring your child to work if
necessary
™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
SCHOOL RESOURCE FIT
How satisfied are you with the:
time your child’s school starts in the morning
time your child’s school lets out in the afternoon
scheduling of extracurricular and other events
communication between school and parents ™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Two-Phase Research Project
Phase Two
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
• How does community resource fit affect
mothers’ and fathers’ quality-of -life (QOL)
indicators?
–
–
–
–
–
–
psychological distress
job disruptions
work interfering with family
family interfering with work
life satisfaction
marital-role, parent-role, and job-role quality™
• Do these relationships differ by parent sex? Or
by family form?
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
METHOD
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
RESULTS
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Family Structure:
Two Parent: Full-time employed
Two Parent: One primary breadwinner
Single Parent
Community Resources
Work
School
After-school arrangements
Transportation
Public transportation
School transportation
After-school
transportation
Community
Resource
Fit
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Quality of Life
Mothers
• All mothers in two-parent families regardless
of employment status (n = 58)
• Employed mothers in two-parent families (n =
38)
• Single mothers (n = 25)
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
All mothers in two-parent families
regardless of employment status (n = 58).
Married mothers who are satisfied with the
extent to which their child’s school is meeting
their needs, also tend to be also more positive
about their family relationships – their marital
relationship as well as their relationship with
their school-age child.
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Employed Mothers in Two-Parent Families (n = 38)
Distress WFC
Work
-.535***
Resource Fit
School
Resource Fit
FWC JRQ
-.738*** -.337*
JobD
MRQ
-.337*
.330*
.511**
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Among single mothers (n = 25), there were few
significant correlations with either resource fit
subscale.
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
For mothers of school-age children, the effect
of resource fit depends primarily on their
marital status
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Most fathers were married and employed
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
All Fathers
Distress WFC
Work
Resource Fit
School
Resource Fit
-.398**
-.429**
FWC
JRQ
.693**
.329*
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
JobD
MRQ
CONCLUSIONS
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
1. We have developed a 36-item community-resource
fit measure that has excellent psychometric properties
and warrants further study.
2. This measure appears to be comprised of 6
subscales which are moderately intercorrelated. Each
subscale had excellent psychometric properties.
3. Even with very small samples we found interesting
patterns of results with multiple QOL indicators. ™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
4. In this sample we found that community resource fit
is most strongly related to QOL indicators among
married mothers.
5. The results for fathers and mothers were similar for
work resource fit, but not for school resource fit.
6. Future research should include a large sample of
participants who reside in range of communities (e.g.,
inner city, exurban, suburban, rural). With a large and
more heterogeneous sample, we will be better able to
detect relationships between the full range of
community resource fit subscales and outcomes of
interest. ™
Work & Family: Changing Realities Conference
Claremont Colleges
March 25, 2006
Critical Missing Link
1
Community: The Critical Missing Link in Work-Family Research1
Rosalind Chait Barnett and Karen C. Gareis
Brandeis University
Introduction
Work-family research is a relatively new and interdisciplinary field; to date, much workfamily research has focused on workplace stressors (e.g., Kossek & Ozeki, 1998; Fox & Dwyer,
1999). This focus has yielded important insights and has even led to changes in policies and
benefits. For example, flexible work arrangements and parental leave policies derived from
research on work hours; research on supervisor and co-worker support prompted the
development of EAP services; and on-site day care and family referral services derived from
research on and advocacy for workers’ dependent care needs.
As with all emerging fields, some norms and assumptions have influenced the direction of
theory and research. In U.S. work-family research, there is a growing consensus that
corporations alone cannot meet the many needs of working families, especially those with
children. It is time to explore the role of community – not only as a context outside of work, but
as a provider of resources essential to the well-being of working families.
Community resources are key to the well-being of resident families, regardless of
socioeconomic class or geographic location. The presence of adequate resources can facilitate
the lives of working families, whereas inadequate resources constitutes a hindrance. We argue
that the distress that working parents experience in striving to meet family needs contributes to
their overall distress, net of the effect of such well-studied stressors as job demands and job
control (Karasek & Theorell, 1990). These newly identified stressors are important because they
may significantly increase employees’ health risks and decrease their productivity on the job.
Family Structure and Community Life
The two-earner family is now the modal American family. As of 2005, 61.3% of U.S.
couples with children were dual-earner families (U.S. Bureau of Labor Statistics, 2006a). Parents
of minor children constituted 35.2% of the labor force – close to 50 million employees – in 2005
(percentage computed by the authors from data in U.S. Bureau of Labor Statistics, 2006a,
2006b). Yet major aspects of work and community life remain organized as though all
households were families with an adult available to participate in household-sustaining but nonearning activities. The breadwinner-homemaker family no longer prevails, but much local
commerce and many public services do not accommodate this reality. This disjuncture burdens
dual-earner families, single parents with substantial custodianship, and even those without
1
Data for this analysis were gathered under a grant from the Alfred P. Sloan Foundation
to the first author. We gratefully acknowledge the contributions of Claudia Morgan, Ph.D., to an
earlier draft of this chapter. Correspondence concerning this analysis should be addressed to
Rosalind Chait Barnett, Brandeis University, Women's Studies Research Center, Mailstop 079,
515 South Street, Waltham, MA 02453-2720, rbarnett@brandeis.edu, (781) 736-2287, FAX
(781) 736-4881.
Critical Missing Link
2
dependents who work full-time.
Workers and their families reside in communities, and their lives are structured in part by the
resources that are available to them in those communities. Full-day kindergarten is unavailable
in most locales. School days end between 2:00 p.m. and 3:00 p.m.; school conferences and
parent-teacher meetings are often scheduled during the workday. Most doctors and dentists see
patients only during regular work hours. Retail and local service businesses often do not open
until 10:00 a.m., and many are closed on Sundays. Public transit authorities run limited holiday
schedules on many workdays each year. Home repairs and deliveries can usually be scheduled
only during regular work hours. Although there has been some movement toward aligning
working adults’ needs and community resources – for example, post offices and banks are now
routinely open half-days on Saturdays – the needs of working adults (and their children) are
often left unfulfilled or are inadequately met by their communities.
Of course, communities vary in the resources they provide to their residents, thereby
affecting the ability of resident families to thrive. “Community resources” may encompass a
broad array of assets – good schools, libraries, well-lit playgrounds, sidewalks, bike trails,
community facilities for teens and elders, accessible healthcare services, safe and adequate
transportation, preschool programs, after-school programs, cultural events sponsorship, and
accessible retail business zones.
Workers residing in communities with abundant resources are likely to report better qualityof-life and well-being outcomes than those who reside in communities with meager community
resources. Community resources can function as supports for working families in a number of
ways. For example, if a state or the federal government does not mandate the provision of
maternity leave, women may be forced to leave their jobs when they give birth unless their
community has infant day care resources that they can easily use. Workers who lose their jobs in
a poor economy may be better able to find new employment if their communities provide
commuting options to areas with more employment opportunities. The impact of work schedule
changes on the employee’s work-social system will vary depending on how well communities
meet the needs of the family. Thus, community resources have both direct and indirect effects on
work-family outcomes. When the resources of a community are well matched to the needs of a
working family, that family’s “community resource fit” is good. When the resources are poorly
matched to a family’s needs, their community resource fit is poor.
Theoretical Background
“Community” in Work-Family Thought
In contrast to fields such as child development and crime prevention, until recently,
“community” has not had a conceptual presence in the work-family literature. However, despite
growing interest in community among work-family scholars (e.g., Bookman, 2004, 2005; Lewis
& Cooper, 1999; Voydanoff, 2001a, 2001b), research to date has been scattered and noncumulative. One important reason for the current state of affairs is the absence of an agreed-upon
definition of community.
What do we mean by community? Most of us are members of several communities:
communities of shared interests, religious faith, political leanings, professional identification,
employment, interpersonal commitments, common history, shared values, shared practices, and
Critical Missing Link
3
common territory. Each of these communities has its own resources (e.g., a holy book, a set of
traditions, a political platform). The focus of this chapter is on residential communities and the
resources available within them, building on Bookman’s definition of community as a “real
geographical community that shape[s] family life and work” (2005, p. 144). We also build on
Voydanoff’s (2001a) analysis of community. She identifies six different aspects: (1) community
social organization, (2) social networks, (3) social capital, (4) formal volunteering and informal
helping, (5) sense of community, and (6) community satisfaction. Our approach is linked to the
satisfaction aspect of community. In contrast to researchers who focus on community-level
resources such as crime rate, poverty level, academic performance of the school system, (e.g.,
Sampson & Groves, 1989), we assess individual community members’ perceptions of how well
community resources meet their needs using a newly developed quantitative measure.
The Community-Families-Work Model
In the Community-Families-Work Model, decisions about one’s work and family
arrangements are influenced by such factors as (1) regional economic welfare, social structures,
and cultural norms; (2) the availability of community resources that support family life (for
example, after-school arrangements and transportation to and from those arrangements); (3)
workplace policies and practices; and (4) objective job characteristics and actual work conditions
(e.g., flexibility, expectations about productivity, and task discretion). This model is dynamic
and recursive. When economic conditions are good, the job environment is favorable, and
workers may need fewer services from the community. At the same time, a healthy tax base
arising from high rates of employment means that communities have more resources with which
to provide working families with the services they need. When economic times are bad, job
conditions are less favorable, while communities concomitantly have fewer resources to allocate.
Additionally, and independently of the state of the economy, workers’ needs for community
resources vary as a function of the life stage of members of their social system. For example,
families with young children seek day care services; families with older children need afterschool activities and programs; families with older or otherwise vulnerable adult dependents
may need visiting nurses, physical therapists, or social workers. When families’ communities fail
to meet their needs – because the needed resource is either unavailable or inaccessible –
individuals may have to alter their labor force participation in order to provide more direct care.
The resources that communities direct toward working families are not only a reflection of
socioeconomic factors; social philosophy informs priorities at many levels. In tight economic
times, some communities raise taxes to preserve such services as full-day kindergarten and afterschool programs. Other communities choose to cut services. Thus, communities differ in the
extent to which the needs of working families are a priority, and they allocate their resources
accordingly.
Unavailable or inaccessible community resources may result in increased distress and in
disruptions and ultimately, lower productivity at work. Reduced productivity, especially reduced
work hours, generally means fewer economic resources, making it even harder for working
parents to obtain the community resources they need. Again, these effects are complex and
recursive – inadequate community resources create strains for working parents (Maguire, 2003),
and distressed working parents create demands for community resources.
Critical Missing Link
4
The Family Schedule Coordination Study
Our recently completed Family Schedule Coordination Study had two major aims: (1) to
develop a psychometrically sound measure of community resource fit, and (2) to discover how
community resource fit is related to quality-of-life and well-being outcomes in a sample of
working families with at least one school-aged child. For the purposes of this study, we defined
working families as those in which at least one parent is employed full-time.
Developing a Measure of Community Resource Fit
For the first phase of the study, we conducted and audiotaped open-ended telephone
interviews with 17 parents and guardians of at least one school-aged (i.e., in grades K-12) child
about each family member's work or school schedule and transportation needs and asked them
about what factors in the community, the children’s schools, and the parents’ workplaces made it
easier or more difficult to coordinate family schedules and transportation. The audiotapes were
transcribed and analyzed to develop a measure of community resource fit.
Content analysis of the responses to the open-ended questions yielded a 36-item community
resource fit measure (see Appendix). The items ask participants to rate their level of satisfaction
with 36 aspects of community resources clustering into six categories of resources: work, public
transportation, school, school transportation, after-school programs, and after-school
transportation. (As shown in the Appendix, we suggest cutting the measure to 31 items by
combining several of the after-school transportation items and dropping two of the less relevant
after-school program items.)
The overall community resource fit scale has excellent psychometric properties (see Table
1), with Cronbach’s alphas of .86 for mothers and .90 for fathers. Cronbach’s alphas for the
subscales were within the moderate to high range (alphas ranged from .73 to .95 for mothers and
from.69 to .91 for fathers). There were no significant differences in global community resource
fit scores by gender; t(143) = 0.65, p = .515 or by family type; F(2,79) = 0.05, p = .955. Overall,
the means for mothers (Mean = 4.90, SD = 0.73) and fathers (Mean = 4.97, SD = 0.64) were
almost one point above the midpoint of the scale, close to “slightly satisfied.”
Preliminary Findings Linking Community Resource Fit to Outcomes
For the second phase of the study, we conducted face-to-face quantitative interviews with
parents in three different types of families defined by parent marital status and employment
pattern: (1) dual-earner families had one parent who was employed full-time and a second parent
who was employed at least 20 hours per week (n = 29 families, or 58 individual parents); (2)
one-main-breadwinner families had one parent who was employed full-time and a second parent
who was not employed or who was employed for fewer than 20 hours per week (n = 29 families,
or 58 individuals); and (3) single-parent families had one parent who was employed full-time (n
= 29 families, or 29 individuals). Thus, we interviewed a total of 145 parents from 87 families.
All of the families have at least one school-aged (i.e., in grades K-12) child and reside in an
urban community near Boston. The city has a broad range on socioeconomic status, with a
median household income of $54,000. It has a fairly small poverty population: 7.0% of the
population live below the poverty line, and 1.9% receive public assistance. Almost one-fifth
(17.1%) of residents are minorities. Table 2 shows descriptive data on this sample.
In this preliminary analysis, we focus on work resource fit and school resource fit because,
Critical Missing Link
5
as the places where parents and children spend the largest chunk of their weekdays, these seem
to be the most crucial aspects of community resource fit for working families of school-aged
children. In addition, because employed parents work in all types of settings and under all types
of conditions, and because the community we studied has nine different public schools as well as
several private schools, there should be substantial variability on these aspects of community
resource fit. We predict that high levels of work and school resource fit should be associated
with positive quality-of-life and well-being outcomes for employed parents.
There were few significant correlations between the two community resource fit subscales
and the parents’ demographics (i.e., number of work hours, size of employing company, number
of children at home, parents’ education, and household income). However, not unexpectedly,
mothers who worked longer hours reported lower school resource fit and fathers with higher
household incomes reported higher work resource fit. Surprisingly, for both mothers and fathers,
having more children at home was associated with reporting higher school resource fit. Future
research is needed to determine whether this unexpected finding replicates and, if so, to
determine what the mechanism of this effect might be.
In preliminary analyses, we tested whether work and school resource fit were related to
seven quality-of-life and well-being outcomes: psychological distress, work-to-family conflict,
family-to-work conflict, job disruptions, job-role quality, marital-role quality, and parent-role
quality. We computed partial correlations controlling for negative affectivity, a mooddispositional trait to view the world negatively that is thought to account for spuriously high
correlations between self-report measures of predictor and outcome variables, especially in
cross-sectional analyses (Brennan & Barnett, 1998).
Because work resource fit and four of the seven outcome variables are only available for
parents who are employed, we focus on three groups of parents in these analyses: employed
mothers in two-parent families (n = 40), employed fathers in two-parent families (n = 53), and
employed single mothers (n = 28; there was only one single father in the sample). It is important
to note that this is an exploratory study: Our sample is located in a single community, and our
sample size is fairly small, especially when further subdivided by gender, employment status,
and single-parent vs. two-parent families. Therefore, results should be interpreted with caution.
Partial correlations for the three groups of parents are shown in Table 3. Overall, stronger
results were found for married mothers than for married fathers. This pattern of findings may
reflect the fact that mothers, who tend to take more responsibility for child care, may be more
reliant on their community’s resources than are fathers. Specifically, we found that when married
mothers are satisfied with the extent to which their workplaces are meeting their needs, they
report significantly lower work-to-family and family-to-work conflict and psychological distress
and significantly higher job-role quality. If they are satisfied with the extent to which their
child’s school is meeting their needs, they report significantly fewer job disruptions,
significantly higher marital-role quality, and marginally higher parent-role quality.
For married fathers, those with high work resource fit report significantly lower work-tofamily conflict and significantly higher job-role quality. To the extent that married fathers are
satisfied with the way their children’s schools meet their needs, the report significantly lower
psychological distress and significantly higher job-role quality. Thus, for work resource fit,
married fathers show some of the same patterns of outcomes as married mothers, but school
resource fit appears to be related to different outcomes for them than for the married mothers.
Critical Missing Link
6
It is easy to imagine that single mothers might be more dependent on community resources
than are their married counterparts. Surprisingly, however, there were few relationships linking
resource fit to quality-of-life and well-being outcomes among the single mothers. Like the
married mothers and fathers, those with high work resource fit reported significantly higher jobrole quality. They also reported marginally lower levels of job disruptions. Inspecting the
magnitude of the correlation coefficients in Table 3 suggests that some of the relationships
between work resource fit and other outcomes found among the married mothers might be
replicated among the single mothers if the sample size were larger. However, this does not
appear to be the case for school resource fit, where there is little evidence of links to the
outcomes we assessed.
These counterintuitive findings appear to be at least partially due to the fact that single
mothers report that their children and unspecified “others” take more responsibility for
household and child-care labor and for planning, coordinating, and keeping track of family
members’ day-to-day schedules than is reported by mothers in the other two family types. Even
after correcting for the average age of the children in the household, single mothers report
getting significantly more help from others with child care; F(2,82) = 3.53, p = .034 and
marginally more help from their children with household tasks; F(2,82) = 2.50, p = .089 than do
married mothers.
Conclusions
To summarize, we have developed a 36-item measure of community resource fit with good
psychometric properties that warrants further study. The measure is comprised of six moderately
intercorrelated subscales assessing resource fit in the areas of work, public transportation,
school, school transportation, after-school program, and after-school transportation resources.
Even with very small samples, we found interesting patterns of results linking two aspects of
community resource fit, work and school resource fit, to a variety of quality-of-life and wellbeing outcomes among employed parents. We plan to conduct further, more sophisticated
regression analyses to look more closely at the process by which various aspects of community
resource fit may act as safety valves for working parents. We also plan to look at couple-level
effects, including crossover effects from one spouse to the other, among the families that are
headed by married couples.
Directions for Future Research
Previous models of the work-family relationship should be expanded to include the direct
and indirect effects of the availability and adequacy of community resources. Our findings
suggest the need to more fully identify the range of community resources that impact worker
distress. The scale could easily be modified to add or subtract modules assessing resource fit in
areas of concern to different types of families or to families in different stages of the life cycle.
For example, families with preschoolers could be asked about community resource fit in the area
of day care, while families with adult- or elder-care responsibilities could be asked about similar
aspects of community-based resources for adults and elders.
In addition, our findings suggest that community-level policies and practices can act as
resources that alleviate stress for working parents. Given the pattern of results we found, such
community resources appear to have consequences for individuals (e.g., psychological distress),
Critical Missing Link
7
for families (e.g., marital-role quality, parent-role quality), and for workplaces (e.g., job
disruptions, job-role quality). Such community-level policies and practices might include, for
example, coordinated start and end times for elementary, middle, and high schools in the
community; engaging, high-quality after-school programs that are located in schools to avoid the
need for transportation, or in central locations with safe, reliable transportation provided; and
workplaces that allow for flex-time and flex-place scheduling, such as programs allowing
parents to work in the office during school hours and then from home during the after-school
hours.
Further research into the impact of community resources on worker outcomes is clearly
needed. Future research should be longitudinal in design and include employees from a wider
range of workplaces and residential communities – inner city, exurban, suburban, and rural –
with a broader range of community resources. With a larger and more heterogeneous sample, we
will be better able to detect relationships between the full range of community resource fit
subscales and outcomes of interest.
Finally, the community resource fit measure can provide information useful to community
leaders as they decide how to allocate tax revenues. The subscales can help decision makers
determine which community aspects need to be better designed to meet the needs of the working
families the town wants to attract. The measure could also be used to evaluate the success of
community initiatives aimed at attracting working families. Finally, communities and businesses
in communities can use scores on the community resource fit measure to recruit residents and
workers. A town that scores high on meeting the needs of working families has a powerful
marketing tool. In a competitive world, being able to justifiably claim that your community has
designed its services to make the lives of working families easier to manage might prove to be a
winning strategy.
Critical Missing Link
8
References
Bookman, A. (2004). Starting in our own backyards: How working families can build
community and survive the new economy. New York: Routledge.
Bookman, A. (2005). Can employers be good neighbors? Redesigning the workplace-community
interface. In S. M. Bianchi, L. M. Casper, & R. B. King (Eds.), Work, family, health, and
well-being (pp. 141-156). Mahwah, NJ: Lawrence Erlbaum Associates.
Brennan, R. T., & Barnett, R. C. (1998). Negative affectivity: How serious a threat to self-report
studies of psychological distress? Women's Health: Research on Gender, Behavior, and
Policy, 4, 369-384.
Fox, M. L., & Dwyer, D. J. (1999). An investigation of the effects of time and involvement in
the relationship between stressors and work-family conflict. Journal of Occupational Health
Psychology, 4(2), 164-174.
Kossek, E. E., & Ozeki, C. (1998). Work-family conflict, policies, and the job-life satisfaction
relationship: A review and directions for organizational behavior--human resources research.
Journal of Applied Psychology, 83(2), 139-149.
Karasek, R., & Theorell, T. (1990). Healthy work: Stress, productivity, and the reconstruction of
working life. New York: Basic Books, Inc.
Lewis, S., & Cooper, C. L. (1999). The work-family research agenda in changing contexts.
Journal of Occupational Health Psychology, 4(4), 382-393.
Maguire, K. (2003, April 16). Bus fees eyed for k-6 school kids. Associated Press State and
Local Wire.
Sampson, R. J., & Groves, W. B. (1989). Community structure and crime: Testing socialdisorganization theory. American Journal of Sociology, 94, 774-802.
Simons, R. L., Lin, K.-H., Gordon, L. C., Brody, G. H., Murry, V., & Conger, R. D. (2002).
Community differences in the association between parenting practices and child conduct
problems. Journal of Marriage and Family, 64, 331-345.
U. S. Bureau of Labor Statistics. (2006a). Employment characteristics of families in 2005 (News
Release, April 27, 2006). Washington, DC: U. S. Department of Labor,
http://www.bls.gov/news.release/pdf/famee.pdf.
U. S. Bureau of Labor Statistics. (2006b). State and regional unemployment, 2005 annual
averages (News Release, March 1, 2006). Washington, DC: U. S. Department of Labor,
(http://www.bls.gov/news.release/pdf/srgune.pdf.
Voydanoff, P. (2001a). Conceptualizing community in the context of work and family.
Community, Work & Family, 4(2), 133-156.
Voydanoff, P. (2001b). Incorporating community into work and family research: A review of
basic relationships. Human Relations, 54(12), 1609-1637.
Critical Missing Link
Appendix
Community Resource Fit Measure2
Work Resources
1. The way your work schedule fits with your child(ren)’s schedule(s)
2. The flexibility available at your workplace to handle emergencies
3. The flexibility available at your workplace to attend to family needs
4. Your ability to work at home if necessary
5. Your ability to bring child(ren) to work if necessary
Public Transportation Resources
6. The way the public transportation schedule fits with your own travel needs
7. The way the public transportation schedule fits with your child(ren)’s travel needs
8. The convenience of your own access to public transportation
9. The convenience of your child(ren)’s access to public transportation
10. The way that the available public transportation routes meet your own travel needs
11. The way that the available public transportation routes meet your child(ren)’s travel needs
School Resources
12. The time(s) your child(ren)’s school(s) start(s) in the morning
13. The time(s) your child(ren)’s school(s) let(s) out in the afternoon
14. The way different schools in the community coordinate their schedules with each other
15. The scheduling of school meetings, parent conferences, and events
16. Communication between the school(s) and parents
17. Scheduling of extracurricular activities
School Transportation Resources
18. Where the children wait to be picked up by the school bus in the morning
19. Where the children wait to be picked up by the school bus in the afternoon
20. The reliability of school bus transportation to and from school
21. The availability and scheduling of late buses
After-School Program Resources 3
22. The availability of after-school programs
23. The cost of after-school programs
24. The location of after-school programs
25. The scheduling of after-school programs
26. The availability of supervised programs for children on early release days
27. The availability of supervised programs for children during school vacations
28. Communication between after-school program providers and parents
29. The expectations of after-school program providers for parental involvement
2
Response scale ranged from 1 (completely dissatisfied) to 7 (completely satisfied).
3
We suggest dropping items 26 and 27 in the interests of shortening the measure.
9
Critical Missing Link
10
After-School Transportation Resources 4
30. The availability of transportation between school and after-school activities
31. The scheduling of transportation between school and after-school activities
32. The reliability of transportation between school and after-school activities
33. The availability of transportation between after-school activities and home
34. The scheduling of transportation between after-school activities and home
35. The reliability of transportation between after-school activities and home
36. The cost of transportation to and/or from after-school activities
4
We suggest combining items 30 and 33, items 31 and 34, and items 32 and 35 into three
items about the availability, scheduling, and reliability of transportation “to and from afterschool activities” in the interests of shortening the measure.
Critical Missing Link
Table 1: Cronbach’s Alphas For Community Resource Fit Scale and Subscales
Mothers
(n = 86)
Fathers
(n = 59)
Global Score
.86
.90
Work
.77
.80
Public Transportation
.91
.81
School
.73
.69
School Transportation
.77
.73
After-School Programs
.88
.85
After-School Transportation
.95
.91
11
Critical Missing Link
Table 2: Descriptive Characteristics of Sample
Dual-Earnera
(n = 29)
One Main
Breadwinnerb
(n = 29)
Single Parentc
(n = 29)
Mean (SD)
Mean (SD)
Mean (SD)
Mother’s Work Hours
37.5 (11.6)
10.6 (16.2)
43.0 (8.8)
Father’s Work Hours
47.6 (9.3)
41.2 (24.4)
50.0 (—)
$24,371 (9,889)
$19,044 (7,149)
$15,956 (10,261)
Per Capita Household Income
Number of Children at Home
2.1 (0.8)
2.4 (0.7)
1.8 (0.7)
In dual-earner families, 20 mothers were full-time employed and 9 mothers were part-time
employed (at least 20 hours per week). All 29 fathers were full-time employed.
b
In one-main-breadwinner families, 6 mothers were full-time employed, 5 mothers were parttime employed (fewer than 20 hours per week), and 18 mothers were not employed. In these
same families, 23 husbands were full-time employed, 1 father was part-time employed (fewer
than 20 hours per week), and 5 fathers were not employed.
c
Of the single-parent families, 28 were headed by mothers and 1 was headed by a father; all
single parents were full-time employed.
a
12
Table 3: Partial Correlations Linking Work and School Resource Fit to Outcomes among Employed Parents
Married Mothers
(n = 40)
Married Fathers
(n = 53)
Single Mothers
(n = 28)
Resource Fit
Work
School
Work
School
Work
School
Distress
-.54*
.03
-.16
-.42*
-.26
.19
Work-to-Family Conflict
-.74*
-.04
-.40*
-.01
-.10
-.05
Family-to-Work Conflict
-.34*
-.24
-.06
-.04
-.29
-.04
Job Disruptions
-.23
-.34*
.00
-.00
-.34†
-.07
Job-Role Quality
.51*
-.08
.69*
.33*
.61*
.17
Marital-Role Quality
.11
.33*
-.12
.09
--
--
Parent-Role Quality
-.02
.31†
Note. All correlations controlled for negative affectivity.
* p < .05. † p < .10.
-.16
.11
.02
.05
Note. GLOBAL = global resource fit, WORK = work resource fit, PUBTRAN = public transportation resource fit, SCHOOL = school
resource fit, SCHTRAN = school transportation resource fit, AFTSCH = after-school resource fit, and AFTRAN = after-school
transportation resource fit. N = 145 except for work resource fit (only 122 parents were employed). All correlations controlled for
negative affectivity and logged per capita household income.
* p < .05.
† p < .10.
Community Resource Fit
1
Running head: COMMUNITY RESOURCE FIT
The Development of a New Measure for
Work-Family Research: Community Resource Fit
Karen C. Gareis
Brandeis University
Rosalind Chait Barnett
Brandeis University
Data for this analysis were gathered under a grant from the Alfred P. Sloan Foundation
(2003-12-1) to the second author. We gratefully acknowledge the contributions of our
interviewers, Sarah Anderson, Joyce Buni, Connie Festo, Carol Genovese, Eleanor Jacobs, and
Heidi La Bash. Correspondence concerning this analysis should be addressed to Rosalind Chait
Barnett, Brandeis University, Women's Studies Research Center, Mailstop 079, 515 South Street,
Waltham, MA 02453-2720, rbarnett@brandeis.edu, (781) 736-2287, FAX (781) 736-4881.
Community Resource Fit
2
Abstract
We describe the development and validation of a quantitative measure of community resource fit.
The measure has good psychometric properties, and preliminary results suggest that the measure
warrants further study. The measure is comprised of six moderately intercorrelated subscales
assessing resource fit in the areas of work, public transportation, school, school transportation,
after-school program, and after-school transportation resources. We found interesting patterns of
results linking community resource fit, especially in the areas of work and school resource fit, to
a variety of quality-of-life and well-being outcomes among employed parents of school-aged
children. These outcomes include work-to-family and family-to-work conflict and enhancement,
psychological distress, job-role quality, likelihood of losing or leaving one’s job, and likelihood
of leaving one’s line of work.
KEYWORDS: community, work-family, school-aged children, psychological distress, job-role
quality
Community Resource Fit
3
The Development of a New Measure for
Work-Family Research: Community Resource Fit
It is widely agreed that resources available to workers in their residential communities might
facilitate their ability to meet their many work and family demands. Well-run schools, responsive
workplaces, safe streets, and adequate transportation are resources that, if matched to the needs
of working families, can potentially relieve some of the distress associated with combining work
and family. However, despite growing interest in community among work-family scholars (e.g.,
Bookman, 2005; Voydanoff, 2001), research has been scattered and non-cumulative. One
important reason for the current state of affairs is the absence of a consensus definition of
community and, relatedly, the lack of a reliable quantitative measure of community resource fit
(Barnett & Gareis, 2006).
Based on Bookman’s definition of community as a “real geographical community that shapes
family life and work” (2005, p. 144), the focus of this paper is on the development of a measure
assessing the resources available in employed families’ residential communities. Voydanoff
(2001) identifies six different aspects of community, including : (1) community social
organization, (2) social networks, (3) social capital, (4) formal volunteering and informal
helping, (5) sense of community, and (6) community satisfaction. We build on Voydanoff’s
community satisfaction aspect in developing our quantitative measure assessing individual
community members’ perceptions of how well community resources meet their needs. We are
concerned with subjective evaluations of community resources relative to individual needs.
Therefore, rather than assessing the objective resources available to working families in their
Community Resource Fit
4
residential communities, we focus on perceived fit, or satisfaction with the match between the
community’s resources and the person’s values, desires, or goals.
Our conceptualization of the role of community resource fit in managing work and family
responsibilities builds on Emlen’s (1997) notion of a “flexibility solution.” In his work with
employed parents of preschoolers, Emlen argues that employees who have “safety valves” in
their families, at work, or at their child-care provider are less likely to suffer negative effects of
managing work and family. Such safety valves provide parents with a kind of insurance against
the crises that crop up all too often in working parents’ lives. In our work, we extend Emlen’s
insights to families with school-aged children (Barnett & Gareis, 2006; Barnett & Gareis, under
review a; under review b). For these parents, in addition to family and work flexibility, we posit
that the critical areas in which flexibility is needed are school, after-school care, and
transportation – all aspects of community resources. Given how much empirical attention has
been paid to family-friendly workplace policies and practices, it seems timely to explore the role
that community institutions such as these can play in facilitating the positive adjustment of
working families.
In the larger study of which this analysis is a part, the focus is on how well the workplace, the
local schools, the after-school programs, and the local transportation systems – including public,
school, and after-school transportation – meet the needs of individuals in working families who
have at least one child who is in primary or secondary school (i.e., in U.S. grades K through 12).
Although the specific components of community resources we assess are those of special
importance to families of school-aged children, it would be a simple matter to tailor the measure
Community Resource Fit
5
to include modules assessing resources that are important to different types of families; for
example, those with infant- or elder-care responsibilities.
Method
Stage I: Measure Development
For the first phase of the study, we conducted and audiotaped open-ended telephone
interviews lasting approximately 30 minutes with 17 parents and guardians of at least one schoolaged (i.e., in grades K-12) child about each family member's work or school schedule and
transportation needs and asked them about what factors in the community, the children’s schools,
and the parents’ or guardians’ workplaces made it easier or more difficult to coordinate family
schedules and transportation. The sample was developed through a combination of random
sampling from the household census of the city in which Stage II of the study was to be
conducted and publicizing the study and asking interested volunteers to contact us. Each
interviewee received $25 for participating. The audiotapes were transcribed and analyzed to
develop a measure of community resource fit.
Stage II: Measure Validation
In the second phase of the study, we collected data on the psychometric properties of the
measure and validated it against theoretically related outcomes. As part of a larger study of how
parents of school-aged (i.e., in grades K-12) children in one Boston-area city coordinate their
work schedules with their children’s school, after-school, and transportation schedules, we
conducted private 45-minute face-to-face interviews with all parents in 29 dual-earner families,
29 families with one main breadwinner and one stay-at-home or part-time employed parent, and
Community Resource Fit
6
29 single-parent families (total N = 145). In the interviews, parents completed the new
community resource fit measure and established measures of a number of theoretically related
outcomes, including work-to-family and family-to-work conflict and enhancement (MacDermid
et al., 2000), psychological distress (Derogatis, 1975), job-role quality (Barnett & Brennan,
1995), and likelihood of losing a job one wants to keep within the next 12 months, voluntarily
leaving one’s job within the next 12 months, and leaving one’s line of work and going into
another line of work entirely within the next five years. The sample was drawn randomly from
the city’s household census. Each parent received $50 for participating.
Results
Content analysis of the responses to the open-ended questions yielded a 36-item community
resource fit measure (see Appendix). The items ask participants to rate their level of satisfaction
with 36 aspects of community resources clustering into six categories of resources: work, public
transportation, school, school transportation, after-school programs, and after-school
transportation. (As shown in the Appendix, we suggest cutting the measure to 31 items by
combining several of the after-school transportation items and dropping two of the less relevant
after-school program items.)
The overall community resource fit scale has excellent psychometric properties (see Table 1),
with a Cronbach’s alpha of .88 in the sample as a whole. Cronbach’s alphas for the subscales
were within the moderate to high range, with alphas ranging from .72 to .94. There were no
significant differences in global community resource fit scores by gender; t(143) = 0.65, p = .515
or by family type; F(2,79) = 0.05, p = .955. As shown in Table 1, the mean for the global score
Community Resource Fit
7
was 4.93 (SD = 0.69) were almost one full point above the midpoint of the scale, close to
“slightly satisfied.” The subscale means ranged from a low of 4.54 (SD = 1.26) for after-school
transportation resource fit and a high of 5.39 (SD = 0.86) for school resource fit. As shown in
Table 2, subscale scores were moderately intercorrelated, with correlations ranging from .045 (p
= .630) to .382 (p = .000).
Table 3 shows the correlations between resource fit scores and outcomes, controlling for
logged per capita household income and for negative affectivity (Spielberger, 1983), a mooddispositional trait to view the world negatively that is thought to account for spuriously high
correlations between self-report measures of predictor and outcome variables, especially in crosssectional analyses (Brennan & Barnett, 1998). As shown in the table, parents who report higher
global resource fit report significantly higher levels of work-to-family enhancement. Of the
subscale scores, work-to-family enhancement is specifically associated with higher levels of
work and school resource fit. School resource fit is also a significant predictor of family-to-work
enhancement, and work resource fit is a significant predictor of work-to-family and family-towork conflict. These results provide support for the idea that good community resource fit,
particularly in the areas of work and school, may act as a safety valve buffering working parents
of school-age children from the stress they might otherwise experience as they try to meet the
demands associated with work and family obligations.
As shown in Table 3, good work resource fit is also significantly associated with lower levels
of psychological distress, higher levels of job-role quality, and lower reported likelihood of
losing one’s job or voluntarily leaving one’s job in the next 12 months. Good school resource fit
Community Resource Fit
8
is significantly associated with higher levels of job-role quality and lower reported likelihood of
leaving one’s line of work in the next five years.
Public transportation and school transportation resource fit were not significantly associated
with any outcomes, and after-school program and after-school transportation resource fit showed
only a single marginal association each with outcomes (after-school program resource fit with
greater family-to-work enhancement and after-school transportation resource fit with lower
reported likelihood of leaving one’s line of work in the next five years).
Discussion and Conclusions
To summarize, we have developed a quantitative measure of community resource fit with
good psychometric properties that warrants further study. The measure is comprised of six
moderately intercorrelated subscales assessing resource fit in the areas of work, public
transportation, school, school transportation, after-school program, and after-school
transportation resources. Even with a fairly small sample, we found interesting patterns of results
linking community resource fit, especially in the areas of work and school resource fit, to a
variety of quality-of-life and well-being outcomes among employed parents of school-aged
children.
Previous models of the work-family relationship should be expanded to include the direct and
indirect effects of community resource fit. Our findings suggest the need to more fully identify
the range of community resources that impact worker distress. The scale could easily be modified
to add or subtract modules assessing resource fit in areas of concern to different types of families
or to families in different stages of the life cycle. For example, families with preschoolers could
Community Resource Fit
9
be asked about community resource fit in the area of day care, while families with adult- or eldercare responsibilities could be asked about similar aspects of community-based resources for
adults and elders.
Our findings also suggest that community-level policies and practices can act as resources
that alleviate stress for working parents. Given the pattern of results we found, such community
resources appear to have consequences for families (e.g., work-to-family and family-to-work
conflict and enhancement, psychological distress) and for workplaces (e.g., job-role quality,
likelihood of losing or leaving one’s job and leaving one’s line of work). Such community-level
policies and practices might include, for example, coordinated start and end times for elementary,
middle, and high schools in the community; engaging, high-quality after-school programs that
are located in schools to avoid the need for transportation, or in central locations with safe,
reliable transportation provided; and workplaces that allow for flex-time and flex-place
scheduling, such as programs allowing parents to work in the office during school hours and then
from home during the after-school hours.
Further research into the impact of community resources on worker outcomes is clearly
needed. Ideally, such research should be longitudinal in design, and include employees from a
wider range of workplaces and residential communities internationally – inner city, exurban,
suburban, and rural – with a broader range of community resources. With a larger and more
heterogeneous sample, we will be better able to detect relationships between the full range of
community resource fit subscales and outcomes of interest.
Community Resource Fit
10
References
Barnett, R. C., & Brennan, R. T. (1995). The relationship between job experiences and
psychological distress: A structural equation approach. Journal of Organizational Behavior,
16, 259-276.
Barnett, R. C., & Gareis, K. C. (2006, March). Community: The missing link in the work-family
literature. Paper presented at the 23rd Annual Claremont Symposium on Applied Social
Psychology, Work and Families: Changing Realities, Claremont Graduate University,
Claremont, CA.
Barnett, R. C., & Gareis, K. C. (under review a). Does community resource fit matter to fathers?
A study of employed fathers, school resource fit, and well-being. Manuscript submitted.
Barnett, R. C., & Gareis, K. C. (under review b). When your work arrangements work for you: A
study of employed women with school-aged children. Manuscript submitted.
Bookman, A. (2005). Can employers be good neighbors? Redesigning the workplace-community
interface. In S. M. Bianchi, L. M. Casper, & R. B. King (Eds.), Work, family, health, and
well-being (pp. 141-156). Mahwah, NJ: Lawrence Erlbaum Associates.
Brennan, R. T., & Barnett, R. C. (1998). Negative affectivity: How serious a threat to self-report
studies of psychological distress? Women's Health: Research on Gender, Behavior, and
Policy, 4, 369-384.
Derogatis, L. R. (1975). The SCL-90-R. Baltimore: Clinical Psychometrics.
Emlen, A.C. (1997, May). Quality of childcare and special needs of children who have
emotional or behavioral problems. Paper presented at the Building on Family Strengths: A
Community Resource Fit
11
National Conference on Research and Services in Support of Children and Families,
Portland, OR.
MacDermid, S. M., Barnett, R., Crosby, F., Greenhaus, J., Koblenz, M., Marks, S., et al. (2000).
The measurement of work/life tension: Recommendations of a virtual think tank. Boston,
MA: Alfred P Sloan Foundation.
Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y ed.). Palo Alto,
CA: Consulting Psychologists Press.
Voydanoff, P. (2001). Conceptualizing community in the context of work and family.
Community, Work & Family, 4, 133-156.
Community Resource Fit
Appendix
Community Resource Fit Measure1
Work Resources
1. The way your work schedule fits with your child(ren)’s schedule(s)
2. The flexibility available at your workplace to handle emergencies
3. The flexibility available at your workplace to attend to family needs
4. Your ability to work at home if necessary
5. Your ability to bring child(ren) to work if necessary
Public Transportation Resources
6. The way the public transportation schedule fits with your own travel needs
7. The way the public transportation schedule fits with your child(ren)’s travel needs
8. The convenience of your own access to public transportation
9. The convenience of your child(ren)’s access to public transportation
10. The way that the available public transportation routes meet your own travel needs
11. The way that the available public transportation routes meet your child(ren)’s travel needs
School Resources
12. The time(s) your child(ren)’s school(s) start(s) in the morning
13. The time(s) your child(ren)’s school(s) let(s) out in the afternoon
14. The way different schools in the community coordinate their schedules with each other
15. The scheduling of school meetings, parent conferences, and events
1
Response scale ranged from 1 (completely dissatisfied) to 7 (completely satisfied).
12
Community Resource Fit
13
16. Communication between the school(s) and parents
17. Scheduling of extracurricular activities
School Transportation Resources
18. Where the children wait to be picked up by the school bus in the morning
19. Where the children wait to be picked up by the school bus in the afternoon
20. The reliability of school bus transportation to and from school
21. The availability and scheduling of late buses
After-School Program Resources 2
22. The availability of after-school programs
23. The cost of after-school programs
24. The location of after-school programs
25. The scheduling of after-school programs
26. The availability of supervised programs for children on early release days
27. The availability of supervised programs for children during school vacations
28. Communication between after-school program providers and parents
29. The expectations of after-school program providers for parental involvement
After-School Transportation Resources 3
30. The availability of transportation between school and after-school activities
2
3
We suggest dropping items 26 and 27 in the interests of shortening the measure.
We suggest combining items 30 and 33, items 31 and 34, and items 32 and 35 into three
items about the availability, scheduling, and reliability of transportation “to and from after-school
activities” in the interests of shortening the measure.
Community Resource Fit
31. The scheduling of transportation between school and after-school activities
32. The reliability of transportation between school and after-school activities
33. The availability of transportation between after-school activities and home
34. The scheduling of transportation between after-school activities and home
35. The reliability of transportation between after-school activities and home
36. The cost of transportation to and/or from after-school activities
14
Community Resource Fit
15
Table 1
Cronbach’s Alphas and Descriptive Statistics for Community Resource Fit Scale and Subscales
Alpha
Mean (SD)
Global Score
.88
4.93 (0.69)
Work
.78
5.17 (1.23)
Public Transportation
.87
4.75 (1.25)
School
.72
5.39 (0.86)
School Transportation
.75
5.07 (1.07)
After-School Programs
.87
4.65 (1.11)
After-School Transportation
.94
4.54 (1.26)
Note. N = 145 except for work resource fit (only 122 parents were employed). Scores on the
scales ranged from 1 (completely dissatisfied) to 7 (completely satisfied).
Table 2
Intercorrelations Among Resource Fit Subscale Scores
Resource Fit
1
2
3
1. Work
--
2. Public Transportation
.12
--
3. School
.18*
.38*
--
4. School Transportation
.29*
.31*
.29*
--
5. After-School Programs
.20*
.16†
.33*
.05
--
6. After-School Transportation
.15
.30*
.30*
.33*
.29*
Note. N = 145 except for work resource fit (only 122 parents were employed).
* p < .05.
† p < .10.
4
5
6
--
Table 3
Partial Correlations Linking Community Resource Fit to Outcomes
GLOBAL
WORK
PUBTRAN
SCHOOL
SCHTRAN
AFTSCH
AFTRAN
Work-to-Family Conflict
-.08
-.42*
.04
-.02
.00
-.01
.01
Family-to-Work Conflict
-.03
-.22*
.06
-.16†
.11
.03
-.01
Work-to-Family Enhancement
.24*
.46*
.01
.21*
.17
.10
.03
Family-to-Work Enhancement
.14
.04
.02
.23*
-.03
.16†
.10
Psychological Distress
-.10
-.32*
-.01
Job-Role Quality
.31*
.64*
Lose Job
-.08
Leave Job
Leave Occupation
-.11
.07
-.04
.02
.04
.19*
.09
.11
.06
-.37*
.01
-.06
.05
-.01
.05
-.10
-.30*
-.07
-.03
-.13
.06
-.07
-.20*
-.18†
-.10
-.19*
-.09
-.05
-.18†
Note. GLOBAL = global resource fit, WORK = work resource fit, PUBTRAN = public transportation resource fit, SCHOOL = school
resource fit, SCHTRAN = school transportation resource fit, AFTSCH = after-school resource fit, and AFTRAN = after-school
transportation resource fit. N = 145 except for work resource fit (only 122 parents were employed). All correlations controlled for
negative affectivity and logged per capita household income.
* p < .05.
† p < .10.
The Importance of Satisfactory Work Arrangements: A Study of Employed Women with
School-Aged Children
*Rosalind Chait Barnett, Ph.D.
Brandeis University
Women's Studies Research Center, MS 079
515 South Street
Waltham, MA 02453-2720 USA
rbarnett@brandeis.edu
(781) 736-2287
FAX (781) 736-4881
Karen C. Gareis, Ph.D.
Brandeis University
Women's Studies Research Center, MS 079
515 South Street
Waltham, MA 02453-2720 USA
gareis@brandeis.edu
(781) 736-4886
FAX (781) 736-4881
Preferred Presentation Type: oral
Themes: Well-Being and Quality of Life; Community, Work, and Family
Abstract
Studies suggest that because employed mothers tend to take primary responsibility for
their children’s welfare, they are more likely than employed fathers to adjust their work
schedules (i.e., cut hours, drop out) in response such family demands as the birth of a child or the
needs of school-aged children. The dilemmas of employed mothers of school-aged children have
received much less empirical attention than those of their counterparts with preschoolers. Yet
several reports suggest that school-aged children are at risk for many negative outcomes that may
require mothers to be available during the workday.
Perhaps women’s employment decisions have more to do with the lack of work resource
fit (WRF; i.e., satisfaction with the way one’s work schedule fits children’s schedules, workplace
flexibility for emergencies or family needs, ability to work at home if necessary, and ability to
bring children to work if necessary) than the demands of their children per se. (Our WRF
measure is a subscale of a new measure of Community Resource Fit. WRF is closely related to
an earlier measure, work schedule fit, but has been tailored to the needs of employed parents of
school-aged children.)
Stated differently, high WRF provides a safety valve for working mothers that might ease
the stress they would otherwise experience. In this study we focus on the relationship linking
WRF to job disruptions, intention to leave one’s job within a year, and job-role quality (JRQ) in
a sample of working mothers of school-aged children who vary in marital status.
Method
For the present paper, we report on data from a random sample of employed mothers (N =
68; 40 married and 28 single) who have at least one school-aged child and were part of a larger
interview study. Each mother completed a 45-minute interview and a brief mailed questionnaire
and was compensated $50. The husbands of the married women vary in employment status (34
full-time, 1 part-time, 5 not employed).
Results and Discussion
Work-resource fit (WRF) was a significant predictor of job disruptions, intention to leave
one’s job, and JRQ: When the timing and scheduling of their work works for them, mothers
report low job disruptions and intentions to leave and high JRQ.
Further, WRF and intentions to leave are more tightly linked for married women, who
have a safety valve at home (i.e., a second income, a husband who can provide instrumental and
emotional support). Importantly, the combination of low WRF and long work hours is
particularly toxic for JRQ. In contrast, when WRF is high, mothers who work long hours report
levels of JRQ as high as those of mothers with low work hours. Thus, high WRF acts as a safety
valve protecting mothers from the negative effects of long work hours on JRQ.
In sum, even when they are employed and responsible for school-aged children, women’s
experiences of job disruptions, intentions to leave their jobs, and JRQ are not necessarily
negative; instead, they depend in part on access to safety valves, one of which is WRF.
Abstract Length: 499 words
Fathers Benefit When Their Children’s School Schedules Meet Their Needs: The
Importance of Community Resource Fit
*Rosalind Chait Barnett, Ph.D.
Brandeis University
Women's Studies Research Center, MS 079
515 South Street
Waltham, MA 02453-2720 USA
rbarnett@brandeis.edu
(781) 736-2287
FAX (781) 736-4881
Karen C. Gareis, Ph.D.
Brandeis University
Women's Studies Research Center, MS 079
515 South Street
Waltham, MA 02453-2720 USA
gareis@brandeis.edu
(781) 736-4886
FAX (781) 736-4881
Preferred Presentation Type: oral
Themes: Well-Being and Quality of Life; Community, Work, and Family
Abstract
It is widely agreed that resources available to workers in their communities might
facilitate their ability to meet work and family demands. However, only recently has a
quantitative measure of community resource fit been developed (Barnett, 2006; Barnett & Gareis,
in preparation). This measure has six subscales, each assessing an important aspect of community
(i.e., work, school, after-school activities, public transportation, school transportation, and afterschool transportation). In this analysis, we focus on the relationship linking school resource fit
(i.e., the degree to which your child’s school schedule meets your needs) to psychological
distress and job-role quality among married employed fathers of school-aged children.
According to Emlen (1977), employed parents with young children seek to increase
work-family fit through a “flexibility solution.” High fit occurs when there is flexibility in work,
family, and child care. Low flexibility in any domain reduces work/family fit and stresses the
family system. We expand Emlen’s definition of flexibility to include the fit between employed
parents’ needs and school resource fit. In this paper, we examine whether school resource fit is
beneficial to job-role quality and psychological distress in its own right and test the hypothesis
that good school resource fit will be especially helpful to fathers who do not have other resources
at work or at home; e.g., fathers who have low income or little job flexibility.
Our focus on school resource fit reflects the growing understanding that community
assets impact the lives of working parents. Given the attention paid to family-friendly workplace
policies, it seems timely to explore the role that other community institutions can play in
facilitating the adjustment of working families. We focus on employed fathers because they tend
to work longer hours on less flexible schedules than employed mothers. Thus, they may be more
reactive to the safety valve provided by good school resource fit.
Method
For the present paper, we report on data from a random sample of married, employed
fathers (N = 53) whose wives varied in employment status and who were part of a larger
interview study. Data were collected during the school year. Each father completed a 45-minute
interview and a brief mailed questionnaire and was compensated $50.
Results and Discussion
Results suggest that employed fathers of school-aged children benefit if their children’s
schools provide them with a safety valve. Fathers with high school resource fit report low
psychological distress and high job-role quality. In addition, high school resource fit was
particularly beneficial to fathers with few work and family resources (i.e., fathers who have low
income or low job flexibility). Thus, if school schedules (e.g., schedule of parent conferences and
school events) can better accommodate the schedules of working parents, they may become a
community resource that ameliorates some of the stress such parents would otherwise
experience.
Abstract Length: 460 words
Does Community
1
Running head: DOES COMMUNITY RESOURCE FIT MATTER TO FATHERS?
Does Community Resource Fit Matter to Fathers?
A Study of Employed Fathers, School Resource Fit, and Well-Being1
Rosalind Chait Barnett and Karen C. Gareis
Brandeis University
Abstract
Several scholars have noted that community resources might facilitate or hinder employees’
ability to meet their many work and family demands. However, this is the first study to estimate
these relationships using a newly developed quantitative measure of community resource fit
assessing the satisfaction of employed parents of school-aged children with key community
resources (i.e., work, school, after-school, and transportation resources). In this analysis, we
focused on the relationships linking school resource fit to well-being (i.e., job-role quality,
psychological distress) in a sample of 53 employed married fathers. Good school resource fit
was associated with low psychological distress and high job-role quality. In addition, fathers
with fewer resources in terms of income and job flexibility benefitted most in terms of their jobrole quality.
KEYWORDS: community, work-family, school-aged children, job-role quality, psychological
distress
It is widely agreed that resources available to workers in their residential communities might
facilitate employees’ ability to meet their many work and family demands. Well-run schools,
responsive workplaces, safe streets, and adequate transportation are resources that, if matched to
the needs of working families, can potentially relieve some of the distress associated with
combining work and family. Such community resources might be especially helpful at this time
when employees’ access to workplace flexibility options is declining (Levin-Epstein, 2006).
Despite growing interest in community among work-family scholars (e.g., Bookman, 2005;
Voydanoff, 2001), research has been scattered and non-cumulative. One important reason for the
current state of affairs is the absence of an agreed-upon definition of community and, relatedly,
the lack of a reliable quantitative measure of community resource fit (Barnett, 2006).
We focus on employed fathers because they tend to work longer hours than employed
mothers and to have less flexible work schedules (Glass & Estes, 1997; Jacobs & Gerson, 2004;
1
Data for this analysis were gathered under a grant from the Alfred P. Sloan Foundation
(2003-12-1) to the first author. We gratefully acknowledge the contributions of our interviewers,
Sarah Anderson, Joyce Buni, Connie Festo, Carol Genovese, Eleanor Jacobs, and Heidi La Bash.
Correspondence concerning this analysis should be addressed to Rosalind Chait Barnett,
Brandeis University, Women's Studies Research Center, Mailstop 079, 515 South Street,
Waltham, MA 02453-2720, rbarnett@brandeis.edu, (781) 736-2287, FAX (781) 736-4881.
Does Community
2
Judiesch & Lyness, 1999). Thus, they may be more reactive to the safety valve provided by good
school resource fit. Moreover, several studies indicate that fathers are assuming a larger share of
the child care burden, especially in two-earner families (Becker & Moen, 1999; Bond, Galinsky,
& Swanberg,1998), that the role of father is just as salient to them as it is to mothers (Thoits,
1983), and that when their relationships with their children are good, they derive the same mental
health benefit as do mothers (Barnett, Brennan, & Marshall, 1994). Yet far less is known about
the ways in which employed fathers manage their work and family demands.
What do we mean by community? Most of us are members of several communities:
communities of shared interests, religious faith, political leanings, professional identification,
employment, interpersonal commitments, common history, shared values, shared practices, and
common territory (Voydanoff, 2001). Each of these communities has its own resources (e.g., a
holy book, a set of traditions, a political platform). The focus of this paper is on the communities
in which employees reside and on the resources (e.g., schools, public parks, libraries) within
those communities. The notion of community defined as a residential space builds on Bookman’s
definition of community as a “Real geographical community that shapes family life and work”
(2005, p. 144).
Some researchers (e.g., Sampson & Groves, 1989) focus on objective features of
communities (e.g., crime rate, poverty level, or residential stability) and assess resident
awareness of those features. In contrast, Voydanoff (2001) identifies six different aspects of
community: (1) community social organization, (2) social networks, (3) social capital, (4) formal
volunteering and informal helping, (5) sense of community, and (6) community satisfaction. We
build on Voydanoff’s community satisfaction aspect, using a newly developed quantitative
measure to assess individual community members’ perceptions of how well community
resources meet their needs.
In the larger study of which this analysis is a part, the focus is on how well the workplace, the
local schools, the after-school programs, and the local transportation systems meet the needs of
individuals in working families who have at least one school-aged (i.e., in grades K-12) child.
Although the specific components of community resources we assess are those of special
importance to families of school-aged children, it would certainly be possible to tailor the
measure to include modules assessing resources that are important to different types of families.
For example, families with preschoolers could be asked about the availability, reliability, quality,
and affordability of child-care facilities in the community, while families with elder-care
responsibilities could be asked about similar aspects of community-based resources for elders.
We are concerned with subjective evaluations of community resources relative to individual
needs, values, or goals. Therefore, rather than assessing the objective resources available to
working families in their residential communities, we focus on the perceived fit (i.e., satisfaction
with the degree of match or mismatch) between the community resources and the person’s
values, desires, or goals. Only recently has a quantitative measure of perceived community
resource fit been developed (Barnett, 2006; Barnett & Gareis, in preparation). This measure is
comprised of six subscales, each assessing an important aspect of community for families of
school-aged children (i.e., work, school, after-school activities, public transportation,
transportation to and from school, and transportation to and from after-school activities).
Does Community
3
According to Emlen (1997), employed parents with young children seek to increase
work-family fit through a “flexibility solution.” High fit occurs when there is flexibility in work,
family, and child care. Low flexibility in any of these domains reduces work/family fit and
increases stress in the family system. We expand Emlen’s definition of flexibility to include the
fit between employed parents’ needs and the community resources available to them. Emlen’s
contribution was in conceptualizing community resources as potential buffers to offset the stress
working families experience as they try to meet the demands associated with work and family
obligations. For employed parents of school-aged children, we posit that the four critical areas in
which flexibility is needed are school, work, transportation, and family.
Given how much empirical attention has been paid to family-friendly workplace policies and
practices, it seems timely to explore the role that other community institutions can play in
facilitating the positive adjustment of working families. In this paper, we test a modified version
of Emlen’s hypothesis by focusing on only one source of community resource fit, school resource
fit, or the degree to which one’s child’s school schedule meets one’s own needs. We test the
relationship linking school resource fit to well-being outcomes – specifically, psychological
distress and job-role quality – in a sample of married, employed fathers who have at least one
school-aged child.
Multiple Safety Valves
As Emlen noted, families may have multiple safety valves in various domains of their lives.
Safety valves may include such factors as higher household income, greater job flexibility, and
other resources that make it easier for employed parents to manage their work and family lives.
The implication is that the more safety valves working families have, the greater the likelihood of
positive outcomes.
However, there are at least two mechanisms by which multiple safety valves might benefit
employees. First, there is the additive effect of having more than one safety valve on an outcome.
Second, there is the indirect effect of a second safety valve on the relationship between the first
safety valve and an outcome. In this process, the presence of a second safety valve might have a
synergistic effect such that when both safety valves are present, the ameliorative effect of the first
is heightened. Conversely, in the absence of a second safety valve, the influence of the first safety
valve is unaffected.
Therefore, in addition to testing the main effect of one safety valve, school resource fit, on
outcomes, we also test the direct and indirect effects of several secondary safety valves –
household income and job flexibility – on the relationship between school resource fit and
outcomes.
Moderators
We expect that good school resource fit will be especially helpful to fathers who do not have
other resources at work or at home; specifically, fathers who have low income or little job
flexibility.
Household income. Fathers with high income may be less affected by their level of school
resource fit than their low-income counterparts, perhaps because they are more easily able to
Does Community
4
purchase after-school child care and other services that could increase the likelihood of their
being able to handle work and family demands without undue negative effects on their job
experiences or well-being. Stated differently, adequate family financial resources can constitute a
“safety valve” that protects working families from untoward consequences.
Job flexibility. Fathers with greater flexibility on the job may be less affected by their level of
school resource fit than their counterparts with less job flexibility. There is a growing consensus
in the research literature that workplace flexibility is key to successfully managing parental and
job demands (Bond et al., 1998; Galinsky, Bond, & Friedman, 1993; Glass, 2000). Scholars
agree that individuals can better manage long work hours and the unpredictable demands of
dependent care when they are given a measure of control over when and where their work is done
(Barnett, 1994; Glass, 2000; Schor, 1991).
Finally, we include negative affectivity as a covariate. Negative affectivity is a mooddispositional trait to view the world negatively that is thought to account for spuriously high
correlations between self-report measures of predictor and outcome variables, especially in crosssectional analyses (Brennan & Barnett, 1998).
In sum, in this random sample of 53 employed married fathers of at least one school-aged
child, we test the following hypotheses:
1. School resource fit will be negatively related to psychological distress.
2. School resource fit will be positively related to job-role quality.
3. School resource fit will be more closely linked to outcomes for low-income fathers than
for high-income fathers.
4. School resource fit will be more closely linked to outcomes for fathers with less job
flexibility than for their counterparts who have more job flexibility.
Method
Participants
As part of a larger study of how parents of school-aged (i.e., in grades K-12) children in one
Boston-area city coordinate their work schedules with their children’s school, after-school, and
transportation schedules, we interviewed all parents in 29 dual-earner families, 29 families with
one main breadwinner and one stay-at-home or part-time employed parent, and 29 single-parent
families (total N = 145). For the present paper, we report only on the data from the married,
employed fathers (n = 53). Table 1 shows descriptive data on work hours, annual family income,
and other variables for this subsample. In this subsample, the fathers worked 48.6 hours per week
(SD = 13.0), on average. Annual household income ranged from $43,250 to $175,000, with a
median of $87,500. The families had from one to four children, with most having two (39.6%) or
three (41.5%) children. The children in the household ranged in age from infancy to 22 years of
age, although at least one child in the household had to be school-aged in order to meet the
eligibility criteria for participation in the study.
Procedures
The sample was drawn randomly from the city’s household census. Residents received letters
Does Community
5
describing the study and then screeners followed up by telephone to determine whether they were
eligible and willing to participate.
As with other studies relying on public databases to develop their random samples, it is very
difficult to determine a response rate. Many people contacted refused to give demographic
information, so we were not able to determine how many who did not respond were actually
eligible to participate, nor does the household census provide information on presence of schoolaged children.
Data were collected during the 2003-2004 and 2004-2005 school years. Trained interviewers
arranged 45-minute face-to-face quantitative interviews with parents at a time and place
convenient to the participants. In the case of two-parent families, mothers and fathers were
interviewed privately. Parents also received a 15-minute mailed questionnaire to be completed in
advance and returned at the time of the interview. Each parent received $50 for participating.
Measures
School resource fit was measured using a 6-item subscale (see Appendix) of a global
community resource fit measure developed for this study through a series of open-ended
interviews with parents and guardians of school-aged children. In the full measure, participants
rated their satisfaction with 36 components of community resource fit in the areas of work,
school, after-school activities, public transportation, transportation to and from school, and
transportation to and from after-school activities on a scale ranging from 1 (completely
dissatisfied) to 7 (completely satisfied). Internal consistency in the full sample is excellent for the
scale as a whole (Mothers: " = .86; Fathers: " = .90) and acceptable for the school resource fit
subscale (Mothers: " = .73; Fathers: " = .69).
Psychological distress was assessed using a state measure asking participants to indicate on a
scale ranging from 0 (not at all) to 4 (extremely) how often in the past week they were bothered
by each of 10 symptoms of anxiety and 14 symptoms of depression (Derogatis, 1975). Sample
depression items include “feeling low in energy or slowed down” and “feeling lonely.” Sample
anxiety items include “feeling fearful” and “feeling tense or keyed up.” In the present sample, the
depression and anxiety subscales are extremely highly correlated ( r = .89, p = .000) and show
similar relationships to other study variables. Therefore, as in our previous work (e.g., Barnett,
Marshall, Raudenbush, & Brennan, 1993), anxiety and depression scores have been combined to
create a measure of psychological distress. The combined score has excellent internal consistency
in the full sample (Mothers: " = .95; Fathers: " = .91), further supporting the decision to
combine these subscales.
Job-role quality was assessed with a 28-item measure on which participants were instructed
to rate on a scale ranging from 1 (not at all) to 4 (considerably) the extent to which each item
was rewarding or of concern (Barnett & Brennan, 1995, 1997). Items covered job conditions in
the areas of skill discretion, decision authority, job demands, schedule control, pay adequacy, job
security, and supervisor relations. Sample reward items include “challenging or stimulating
work” and “your supervisor’s respect for your abilities.” Sample concern items include “the job’s
dullness, monotony, lack of variety” and “having to juggle conflicting tasks or duties.” Concerns
were negatively weighted and rewards positively weighted in constructing the score, which was
Does Community
6
the weighted average of item scores (Barnett et al., 1993). Internal consistency is excellent in the
full sample (Mothers: " = .91; Fathers: " = .89).
Job flexibility. Job flexibility was assessed using a three-item measure from the 1997
National Study of the Changing Workforce (Bond et al., 1998) with questions about whether the
participant can choose starting and quitting times within a certain range of hours, whether
starting and quitting times can be changed daily, and how difficult it is to take time off during the
workday to take care of personal or family matters. Internal consistency is acceptable in the full
sample (Mothers: " = .72; Fathers: " = .63).
Work hours was assessed by asking participants how many hours they work per week, on
average, including overtime, at all jobs. Logged per capita household income was assessed by
asking participants to indicate their yearly household income from all sources before taxes. To
obtain per capita household income, we divided this figure by the number of people in the
household. The distribution of this variable is skewed, so we used the natural log of per capita
income in the analyses.
Negative affectivity was assessed with the 10-item Trait Anxiety Scale (Spielberger, 1983) on
which participants indicated on a 4-point scale from 1 (almost never) to 4 (almost always) how
characteristic traits were of them; items include “I am a steady person” and “I have selfconfidence.” The overall negative affectivity score is the mean response to all items answered.
Internal consistency is very good in the full sample (Mothers: " = .83; Fathers: " = .81).
Results
As shown in Table 1, there are significant bivariate correlations between school resource fit
and both well-being outcomes (psychological distress and job-role quality). In order to determine
whether this relationship held up while controlling for important covariates, we estimated
stepwise regression models using school resource fit to predict psychological well-being and jobrole quality, with covariates of work hours and negative affectivity.
The results for psychological distress are shown in Table 2, Step 1: School resource fit is
significantly negatively related to psychological distress. As predicted in Hypothesis 1, the better
the father’s school resource fit, the lower his psychological distress. As an indication of the size
of the main effect of school resource fit on psychological distress, Cohen’s d is -0.90, which is
classified as a large effect (Cohen, 1988).
Table 3 shows the results for job-role quality. As shown in Step 1, school resource fit is
significantly related to job-role quality: The better the father’s school resource fit, the higher his
job-role quality, as predicted in Hypothesis 2. As an indication of the size of the main effect of
school resource fit on job-role quality, Cohen’s d is +0.67, which is again classified as a large
effect.
In Step 2, we added the secondary safety valves of household income and job flexibility to
both of the main-effects models described above. As shown in Step 2 of Table 2, there were no
main effects of these safety valves on psychological distress, nor did the addition of this pair of
variables contribute significantly to R2. However, for job-role quality, the addition of the
secondary safety valve variables resulted in a significant increment to R2 over that associated
with the main-effects model for job-role quality; Fchange(2,47) = 3.22, p = .049. Specifically, there
Does Community
7
was a main effect of job flexibility (but not household income) on job-role quality, with greater
job flexibility being associated with higher job-role quality (see Table 3, Step 2). Cohen’s d for
this effect was +0.59, classified as a moderate-sized effect. The effect of school resource fit on
job-role quality remained significant, with a Cohen’s d of +0.76, classified as a large effect.
In Step 3, we tested whether the relationships between school resource fit and the two
outcomes were moderated by household income or by job flexibility by adding the interaction
terms SCHOOL RESOURCE FIT x HOUSEHOLD INCOME, SCHOOL RESOURCE FIT x
JOB FLEXIBILITY, and HOUSEHOLD INCOME x JOB FLEXIBILITY to the main-effects
models described above. For psychological distress, there was no significant increment to R2 with
the addition of this set of variables, nor were any of the individual interaction terms significant
(see Table 2, Step 3), offering no evidence of moderation by the secondary safety valves.
However, the addition of the set of interaction terms resulted in a significant increment to R2 over
that associated with the main-effects model; Fchange(3,44) = 4.71, p = .006. Specifically, as shown
in Step 3 of Table 3, household income and job flexibility to moderate the relationship between
school resource fit and job-role quality, offering partial support for Hypotheses 3 and 4.
As shown in Figure 1, the relationship between school resource fit and job-role quality is
significantly stronger for low-income fathers, while there is essentially no relationship between
school resource fit and job-role quality for high-income fathers. In other words, under conditions
of low school resource fit, it is the low-income fathers whose job-role quality suffers, whereas
under conditions of high school resource fit, the low-income fathers report slightly higher jobrole quality than do their high-income counterparts. Cohen’s d for this interaction is -.80,
classified as a large effect.
Figure 2 shows that, as with household income, job flexibility moderates the relationship
between school resource fit and job-role quality, with a stronger relationship between school
resource fit and job-role quality for fathers with low vs. high job flexibility. Under conditions of
low school resource fit, it is the fathers with less flexible jobs whose job-role quality is lowest,
whereas under conditions of high school resource fit, the low-flexibility fathers are
indistinguishable from the high-flexibility fathers in their ratings of job-role quality. Cohen’s d
for this interaction is -.62, classified as a moderate-sized effect.
Discussion and Conclusion
In this sample of married employed fathers with at least one school-aged child, we found that
fathers benefit if their children’s schools provide them with a safety valve; that is, if they have
good school resource fit. Our findings support a modified version of the Emlen hypothesis;
namely, that fathers with high school resource fit report low psychological distress and high jobrole quality. In addition, high school resource fit was particularly beneficial to the job-role quality
of fathers with few work and family resources (i.e., fathers who have low income or low job
flexibility). Thus, if school schedules (e.g., start and end times, schedule of parent conferences
and school events) can better accommodate the schedules of full-time working fathers (and most
likely mothers), they may become a community resource that ameliorates some of the stress such
parents would otherwise experience. A full treatment of the impact of community resource fit on
married employed mothers’ outcomes as well as the within-couple effects of community resource
Does Community
8
fit on both partners’ outcomes are important topics that need to be addressed in future studies.
Safety valves include community, family, and personal resources. For example, high school
resource fit, a community resource, may act as a safety valve when work resource fit is low. At
the personal level, employees with high income may be less affected by the level of school
resource fit than employees with low income. high school resource fit may act as a safety valve
releasing the strain that might otherwise diminish the job-role quality of low-income and lowflexibility fathers. Similarly, employed fathers who enjoy the safety valves of high income and
high job flexibility may be less affected by school resource fit, whereas those without such safety
valves may be especially benefitted when school resource fit is high.
This study, like other studies, has limitations. Notably, the sample was quite small and all
participants resided in one Boston-area city. Consequently, issue of generalizability arise. In
addition, there were too few single fathers in the sample to allow us to analyze their data. Future
research should include a larger sample drawn from multiple communities that differ with respect
to the various aspects of community resource fit and family composition. Nevertheless, the study
yielded several significant findings that strongly suggest the utility of incorporating the notion of
community resource fit into future studies of work-family issues.
Does Community
9
References
Barnett, R. C. (1994). Home-to-work spillover revisited: A study of full-time employed women
in dual-earner couples. Journal of Marriage and the Family, 56, 647-656.
Barnett, R. (2006, March). Community: The missing link in the work-family literature. Paper
presented at the Work & Families: Changing Realities Conference, Claremont McKenna
College, CA.
Barnett, R. C., & Brennan, R. T. (1995). The relationship between job experiences and
psychological distress: A structural equation approach. Journal of Organizational Behavior,
16, 259-276.
Barnett, R. C., & Brennan, R. T. (1997). Change in job conditions, change in psychological
distress, and gender: A longitudinal study of dual-earner couples. Journal of Organizational
Behavior, 18, 253-274.
Barnett, R. C., Brennan, R. T., & Marshall, N. L. (1994). Gender and the relationship between
parent-role quality and psychological distress. Journal of Family Issues, 15, 229-252.
Barnett, R. C., & Gareis, K. C. The critical missing link in work-family research: Community.
Manuscript in preparation.
Barnett, R. C., Marshall, N. L., Raudenbush, S., & Brennan, R. (1993). Gender and the
relationship between job experiences and psychological distress: A study of dual-earner
couples. Journal of Personality and Social Psychology, 64, 794-806.
Becker, P. E., & Moen, P. (1999). Scaling back: Dual-earner couples' work-family strategies.
Journal of Marriage and the Family, 61, 995-1007.
Bond, J. T., Galinsky, E., & Swanberg, J. E. (1998). The 1997 National Study of the Changing
Workforce. New York: Families and Work Institute.
Bookman, A. (2005). Can employers be good neighbors? Redesigning the workplace-community
interface. In S. M. Bianchi, L. M. Casper, & R. B. King (Eds.), Work, family, health, and
well-being (pp. 141-156). Mahwah, NJ: Lawrence Erlbaum Associates.
Brennan, R. T., & Barnett, R. C. (1998). Negative affectivity: How serious a threat to self-report
studies of psychological distress? Women's Health: Research on Gender, Behavior, and
Policy, 4, 369-384.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Derogatis, L. R. (1975). The SCL-90-R. Baltimore: Clinical Psychometrics.
Emlen, A.C. (1997, May). Quality of childcare and special needs of children who have
emotional or behavioral problems. Paper presented at the Building on Family Strengths: A
National Conference on Research and Services in Support of Children and Families,
Portland, OR.
Galinsky, E., Bond, J. T., & Friedman, D. E. (1993). The changing workforce: Highlights of the
national study. New York: Families and Work Institute.
Glass, J. (2000). Envisioning the integration of family and work: Toward a kinder, gentler
workplace. Contemporary Sociology, 29(1), 129-143.
Glass, J. L., & Estes, S. B. (1997). The family responsive workplace. Annual Review of
Sociology, 23, 289-313.
Does Community
10
Jacobs, J. A., & Gerson, K. (2004). The time divide: Work, family and gender inequality.
Cambridge, MA: Harvard University Press.
Judiesch, M. K., & Lyness, K. S. (1999). Left behind? The impact of leaves of absence on
managers' career success. American Educational Research Journal, Academy of Management
Journal,, 42, 641-651.
Levin-Epstein, J. (2006). Getting punched: The job and family clock, It’s time for flexible work
for workers of all wages. Washington, DC: Center for Law and Social Policy.
Sampson, R. J., & Groves, W. B. (1989). Community structure and crime: Testing socialdisorganization theory. American Journal of Sociology, 94, 774-802.
Schor, J. B. (1991). The overworked American: The unexpected decline of leisure. New York:
Basic Books.
Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y ed.). Palo Alto,
CA: Consulting Psychologists Press.
Thoits, P. A. (1983). Multiple identities and psychological well-being: A reformulation and test
of the social isolation hypothesis. American Sociological Review, 48, 174-187.
Voydanoff, P. (2001). Conceptualizing community in the context of work and family.
Community, Work & Family, 4, 133-156.
Does Community
Appendix
School Resource Fit Subscale
1. The time(s) your child(ren)’s school(s) start(s) in the morning
2. The time(s) your child(ren)’s school(s) let(s) out in the afternoon
3. The way different schools in the community coordinate their schedules with each other
4. The scheduling of school meetings, parent conferences, and events
5. Communication between the school(s) and parents
6. Scheduling of extracurricular activities
Note. Response scale ranged from 1 (completely dissatisfied) to 7 (completely satisfied).
11
Table 1: Intercorrelations among Measures
Mean (SD)
1
2
3
4
5
6
1. Work Hours
48.62 (13.04)
2. Household Income
$88,674 (27,673)
-.04
---
3. Negative Affectivity
1.69 (0.42)
-.35*
-.06
---
4. Job Flexibility
0.01 (0.76)
-.04
.20
-.19
---
5. School Resource Fit
5.45 (0.74)
-.12
-.28*
.09
.02
---
6. Job-Role Quality
1.13 (0.81)
-.24†
.13
-.23
.35*
.30*
---
10.29 (9.96)
-.04
.21
-.30*
-29*
7. Psychological Distress
Note. N = 53.
† p < .10. * p < .05.
7
---
.53*
-.14
---
Table 2: Relationship Between School Resource Fit and Psychological Distress
Psychological Distress
School Resource Fit
Work Hours
Negative Affectivity
Step 1
Step 2
Step 3
B
$
(SE)
B
$
(SE)
B
$
(SE)
-4.63*
-.34
(1.47)
-3.92*
-.29
(1.53)
-1.76
-.13
(2.13)
0.10
.13
(0.09)
0.11
.15
(0.09)
0.14
.19
(0.10)
14.45*
.61
(2.75)
14.42*
.61
(2.80)
14.81*
.62
(2.92)
Household Income
--
--
--
10.35
.18
(6.56)
9.74
.17
(6.75)
Job Flexibility
--
--
--
-0.74
-.06
(1.47)
-0.35
-.03
(1.65)
INCOME x JOBFLEX
--
--
--
--
--
--
-7.00
-.09
(10.42)
SRF x INCOME
--
--
--
--
--
--
-15.13
-.17
(13.48)
SRF x JOBFLEX
--
--
--
--
--
--
2.20
.12
(2.31)
R2
Adjusted R2
Note. N = 53.
† p < .10. * p < .05.
.42
.45
.49
.39
.39
.39
Table 3: Relationship Between School Resource Fit and Job-Role Quality
Job-Role Quality
Step 1
Step 2
$
(SE)
B
0.32*
.29
(0.14)
Work Hours
-0.02*
-.34
Negative Affectivity
-0.72*
-.37
B
School Resource Fit
Step 3
$
(SE)
B
$
(SE)
0.35*
.32
(0.14)
0.56*
.52
(0.17)
(0.01)
-0.02*
-.29
(0.01)
-0.03*
-.44
(0.01)
(0.26)
-0.59*
-.31
(0.25)
-0.56*
-.29
(0.23)
Household Income
--
--
--
0.63
.14
(0.58)
0.88
.19
(0.54)
Job Flexibility
--
--
--
0.27*
.25
(0.13)
0.34*
.32
(0.13)
INCOME x JOBFLEX
--
--
--
--
--
--
0.70
.11
(0.84)
SRF x INCOME
--
--
--
--
--
--
-2.86*
-.40
(1.08)
SRF x JOBFLEX
--
--
--
--
--
--
-0.38*
-.26
(0.19)
R2
Adjusted R2
Note. N = 53.
† p < .10. * p < .05.
.25
.34
.50
.21
.27
.41
Figure Captions
Figure 1. The Relationship Between School Resource Fit and Job-Role Quality Is Moderated by
Household Income.
Figure 2. The Relationship Between School Resource Fit and Job-Role Quality Is Moderated by
Job Flexibility.
Does Community
16
Does Community
17
When Your Work
1
Running head: WHEN YOUR WORK ARRANGEMENTS WORK FOR YOU
When Your Work Arrangements Work for You:
A Study of Employed Women with School-Aged Children1
Rosalind Chait Barnett and Karen C. Gareis
Brandeis University
Abstract
Most employed mothers of school-aged children have primary responsibility for childcare, yet
the vast majority remain employed. Building on Emlen (1997), we estimated the direct and
indirect relationship between the presence of a “safety valve” at work (i.e., work resource fit)
and three work-related outcomes (i.e., job disruptions, turnover intentions, and job-role quality)
in a sample of mothers who had at least one school-aged child (K-12th grade) and who varied in
marital status, family income, and work hours. Work resource fit was significantly associated
with each outcome in the predicted direction. In addition, marital status moderated the
relationship between work resource fit and turnover intentions, work hours moderated the
relationship between work resource fit and job-role quality, and there was a trend for family
income to moderate the relationship between work resource fit and job-role quality.
KEYWORDS: work resource fit, job disruptions, turnover intentions, job-role quality
There is general agreement that women have primary responsibility for their young children’s
care and nurture (U.S. Census Bureau, 2003), and that employed women are more likely than are
employed men to adapt their work lives to accommodate child-related concerns (Becker & Moen,
1999; Bianchi, 2000; Glass & Estes, 1997). Specifically, employed mothers are highly likely to
adjust their work schedules (i.e., work part-time, drop out of the workforce) in response to childcare and other family needs. One argument is that women are by nature more nurturant and
relational than men and therefore want to spend more time with their young children (Chodorow,
1978; Gilligan, 1982). The strength of this argument is called into question by the high
percentage of women with preschool children who are in the paid labor force; as of 2005, 52.9%
of women with a children under 3 years of age were employed (Bureau of Labor Statistics, 2005),
as were about half (49.5%) of women with children under 1 year of age (Bureau of Labor
Statistics, 2006, Table 6). A second argument is that inflexible workplace policies and practices,
1
Data for this analysis were gathered under a grant from the Alfred P. Sloan Foundation
(2003-12-1) to the first author. We gratefully acknowledge the contributions of our interviewers,
Sarah Anderson, Joyce Buni, Connie Festo, Carol Genovese, Eleanor Jacobs, and Heidi La Bash.
Correspondence concerning this analysis should be addressed to Rosalind Chait Barnett,
Brandeis University, Women's Studies Research Center, Mailstop 079, 515 South Street,
Waltham, MA 02453-2720, rbarnett@brandeis.edu, (781) 736-2287, FAX (781) 736-4881.
When Your Work
2
not a maternal instinct per se, is the primary reason that women with young children tend to cut
back at work or leave the workforce (Boushey, 2005, 2006; Gerson, 1986). The aim of this paper
is to shed some light on the second line of argument.
In spite of difficulties, most mothers of minor children do not leave the workforce. We argue
that employed mothers who experience high levels of satisfaction with the workplace policies
and practices at their place of employment, especially concerning flexibility (i.e., mothers with
high work resource fit), will report fewer job disruptions, lower intentions to leave their job, and
more positive experiences at work (i.e., higher job-role quality) than their counterparts with low
work resource fit. With respect to intention to leave one’s employer within a year, married
mothers (i.e., those with a second source of income and more help with child care and support)
are more likely than single mothers to contemplate turnover when work resource fit is poor. With
respect to job-role quality, it is likely that the mothers who will benefit most from high work
resource fit are those who work longer compared to shorter work hours and those with low
compared to high family income.
Most of the literature on the employment patterns of working mothers has focused on
mothers of preschoolers, even though more employed mothers have school-aged (i.e., 5-18 years
of age) than preschool (i.e., 0-5 years of age) children. Moreover, there is reason to believe that
the dilemmas of employed mothers of school-aged children are at least as severe as those of their
counterparts with preschoolers (e.g., Heymann, 2000). For example, several reports suggest that
school-aged children are at considerable risk for many negative outcomes including both schoolrelated (e.g., truancy) and after-school related problems (e.g., engaging in risky and antisocial
behaviors) (Cohen, Farley, Taylor, Martin, & Schuster, 2002; Kurz, 2002; National Center for
Schools & Communities, 1999; Newman, Fox, Flynn, & Christeson, 2000) that may cause
employed mothers to be anxious and distracted at work.
We estimate the direct effects of work resource fit on three work-related outcomes (i.e., job
disruptions, intention to leave one’s job within a year, and job-role quality) in a random sample
of 68 mothers who have at least one school-aged child and who vary in marital status, family
income, and work hours. We also estimate the indirect effects of marital status, family income,
and work hours on these relationships. The data for this paper come from a larger study of the
relationship between community resource fit and work and family outcomes in a sample of
married and single mothers and fathers who vary in work hours and work status and who have at
least one school-aged child.
For the larger study, we developed a new quantitative measure of community resource fit
(Barnett & Gareis, 2006a; Barnett & Gareis, under review; Gareis & Barnett, in preparation) that
is comprised of six subscales, each assessing an important aspect of community for families of
school-aged children (i.e., work, school, after-school activities, public transportation,
transportation to and from school, and transportation to and from after-school activities). In this
study we focus on one aspect of community resource fit, work resource fit. This new measure of
work resource fit is closely related to an earlier measure, work schedule fit (Barnett, Gareis, &
Brennan, 1999; Gareis, Barnett & Brennan, 2003), but has been tailored to the needs of employed
parents of school-aged children.
Literature Review
When Your Work
3
Increasing numbers of mothers of minor children are in the labor force, and the majority have
school-aged children. As of 2004, three-quarters (73.7%) of women who only had school-aged
children (i.e., 6-17 years of age) were in the civilian labor force, compared to two-thirds (57.8%)
of those with children under 6 years of age. In absolute numbers, 15 million women with schoolaged children were employed, compared to roughly 9.5 million women with preschool children
(Bureau of Labor Statistics, Table 5, 2005). Clearly, more working mothers have school-aged
than preschool children.
Because of a serious mismatch between most mothers’ work schedules and those of their
children’s schools and the paucity of community resources available for school-aged children
(Heymann, 2000), mothers often have to cobble together after-school child-care arrangements.
When these arrangements break down, as they frequently do (Heymann, 2000), mothers often
have to choose between leaving work at some cost to their future job security, or leaving their
children to fend for themselves, a dangerous and anxiety-producing option.
In contrast, community resources for preschoolers, when available, have hours of care that
typically match well with the work schedules of their parents. Not so for school-aged children.
Schools typically dismiss their students between 2:00 p.m. and 3:00 p.m., whereas parents may
be at work until 5:00 p.m., 6:00 p.m., or later. Many parents also have long commutes home from
work, adding to the time they are away from their children and, perhaps, to the time those
children may be unsupervised by an adult.
In spite of being so widespread, the child-care concerns of employed parents of school-aged
children have until recently received little empirical attention. Previous research with parents in
dual-earner families focused on work schedule fit and found that the relationship between work
hours and burnout was mediated by the extent to which the parents’ work schedules met their
needs as well as the needs of their partners and those of their children and elder dependents
(Barnett, Gareis, & Brennan, 1999). Other studies indicate that employed parents who have high
levels of concern about their children’s after-school time report high levels of job disruptions and
low levels of personal well-being (Barnett & Gareis, 2006b, 2006c). The present study builds on
these previous studies.
Work Resource Fit
There is a growing consensus in the research literature that workplace flexibility is key to
successfully managing parental and job demands (Bond, Galinsky, & Swanberg, 1998; Galinsky,
Bond, & Friedman, 1993). Scholars agree that individuals can better manage long work hours
and the unpredictable demands of dependent care when they are given a measure of control over
when and where their work is done (Barnett, 1994; Schor, 1991). Within the work-family
literature, it is often assumed that workplace flexibility is most important when parents have
preschool children. This assumption is not surprising, given the intensive and time-consuming
nature of early child care. Yet, Jacobs and Gerson (2004) found that age of child had little impact
on work-family conflict, which appeared to be related to the mere presence of children and not to
their age; as they note, “[t]he nature of parental stresses may change as children age, but they do
not necessarily diminish, as many parents of teenagers will attest” (p. 92). Despite this finding,
the care issues confronted by employed parents of school-aged children receive relatively little
systematic attention (Heymann, 2000).
When Your Work
4
Emlen (1997), in his work with employed parents of preschoolers, argues that employees
who have “safety valves” at work, in their families, or at their child-care provider are less likely
to suffer negative effects of managing work and family responsibilities. In our work, we extend
his insights to working parents with school-aged children (Barnett & Gareis, 2006a; Barnett &
Gareis, under review; Gareis & Barnett, in preparation). Safety valves provide parents with
insurance against the “crises” that all too often crop up in the lives of working parents of schoolaged children.
We expect that under conditions of poor work resource fit, employed women will have higher
levels of job disruptions. For example, in a large national sample of employed mothers, job
disruptions associated with breakdowns in child-care arrangements included being late, leaving
early, or missing a day of work in the preceding month due to child-care arrangement failures; in
the same survey, almost twenty percent (17.9%) of mothers reported missing a day of work to
stay home with a sick child in the preceding month (Hofferth, Brayfield, Deich, & Holcomb,
1991). Although job interruptions due to family responsibilities decline as children grow older,
39% of mothers and 17% of fathers of children aged as old as 15 to 18 reported such
interruptions (Fernandez, 1986). Among parents with school-aged children, disruptions may be
due to events that occur during the school day (e.g., truancy, sickness, disobedience) or to events
that occur after school (e.g., getting into an accident, smoking, drinking).
Thus, poor work resource fit is likely to be associated with increased work-family conflict,
which has been shown in previous research to predict intentions to leave one’s job (e.g.,
Greenhaus, Parasuraman, & Collins, 2001). Therefore, we expect that when work resource fit is
poor, employed women will be more likely to consider leaving their place of employment.
With respect to job-role quality, previous research demonstrated that work schedule fit was
related to job-role quality (Gareis & Barnett, 2001). Parents whose work schedules were a good
fit to their needs were more likely to report high job-role quality than were those whose work
schedules were a poor fit. Previous studies of job-role quality have not looked specifically at
employed mothers of school-aged children, nor have they estimated the linkage between work
resource fit and job-role quality. We expect that poor work resource fit will be associated with
low job-role quality in the present sample. Specifically, if parents at work are worried about their
children, they may be distracted and anxious on the job, thereby affecting their sense of job
security and their overall job quality. Even if their jobs are otherwise rewarding to them, high
levels of job concerns may result in low job-role quality.
Multiple Safety Valves
As Emlen noted, families may have multiple safety valves in various domains of their lives.
Safety valves may include such factors as the presence of a spouse, higher family income, more
available non-work time (i.e., shorter work hours), and other resources that make it easier for
employed parents to manage their work and family lives. The implication is that the more safety
valves working families have, the greater the likelihood of positive outcomes.
However, there are at least two mechanisms by which multiple safety valves might benefit
employees. First, there is the additive effect of having more than one safety valve on an outcome.
Second, there is the indirect effect of a second safety valve on the relationship between the first
safety valve and an outcome. In this process, the presence of a second safety valve might have a
When Your Work
5
synergistic effect such that when both safety valves are present, the ameliorative effect of the first
is heightened. Conversely, in the absence of a second safety valve, the influence of the first safety
valve is unaffected.
In addition to testing the main effect of one safety valve, work resource fit, on outcomes, we
also test the direct and indirect effects of several secondary safety valves – marital status, family
income, and work hours – on the relationship between work resource fit and outcomes.
Moderators
Marital status. With respect to marital status, surprisingly little empirical research has
focused on the effects of family life on any of the three outcome variables in this study. In one
exception, Lee and Maurer (1999) studied voluntary intentions to leave or to stay in the Navy
until retirement among a sample of U.S. Navy officers. They conclude that family structure
affected voluntary turnover and that considerably more scholarly attention should be directed
toward understanding the association between family variables and work behaviors in general.
We heed this advice in our analyses. Clearly the turnover option may not be equally realistic for
all employed mothers of school-aged children. Issues of support (financial and emotional) may
be particularly critical. Married employed mothers who have a safety valve at home, in the form
of a husband who can help out with child care, may be in a better position to contemplate leaving
their place of employment if work resource fit is poor than single mothers who do not have this
family safety valve. As Carr (2002) notes, it is women with greater financial means and spousal
support who are most able to cut back on paid employment in response to family demands.
Family income. In order to estimate the safety valve effects of marital status per se one needs
to control for household income. Arguably, single mothers are disadvantaged compared to
married mothers primarily because of their lower household income, not because of the absence
of a marital partner (Barnett & Marshall, 1992). Moreover, family income needs to be controlled
to disentangle the effects of household income from the effects of work hours per se
(Stolzenberg, 2001). It may be that employed mothers with high income are less affected by their
level of work resource fit than employed mothers with low income. In other words, high work
resource fit may act as a safety valve releasing the strain that might otherwise diminish job-role
quality among low-income mothers.
Work hours. In addition to the direct effect of work resource fit on job disruptions, intention
to leave one’s job within a year, and job-role quality, we hypothesize that work hours will
moderate at least some of these relationships. For example, the job-role quality of mothers who
work long as compared to short hours may be especially positively affected by the safety valve of
high work resource fit.
We also include negative affectivity as a covariate. Negative affectivity is a mooddispositional trait to view the world negatively that is thought to account for spuriously high
correlations between self-report measures of predictor and outcome variables, especially in crosssectional analyses (Brennan & Barnett, 1998).
In sum, in this random sample of 68 employed mothers of at least one school-aged child, we
test the following hypotheses:
1. High work resource fit will be related to low job disruptions.
2. High work resource fit will be related to low intentions to leave one’s job within a year.
When Your Work
6
3. High work resource fit will be related to high job-role quality.
We do not hypothesize a relationship between each of these three moderators and each of the
three outcome measures. However, because of the paucity of literature in this area, we estimate
the indirect effect of these moderators on the relationships linking work resource fit to each of
the three outcomes. Thus, we tested the following research questions:
4. Marital status may moderate the relationships between work resource fit and outcomes.
5. Family income may moderate the relationships between work resource fit and outcomes.
6. Work hours may moderate the relationships between work resource fit and outcomes.
Method
Participants
As part of a larger study of how parents of school-aged (i.e., in grades K-12) children in one
Boston-area city coordinate their work schedules with their children’s school, after-school, and
transportation schedules, we interviewed all parents in 29 dual-earner families, 29 families with
one main breadwinner and one stay-at-home or part-time employed parent, and 29 single-parent
families (total N = 145). For the present paper, we report only on the data from the employed
mothers (n = 68). Table 1 shows descriptive data on work hours, annual family income, and other
variables for this subsample. The families had from one to four children, with most having one
(27.9%) or two (45.6%) children. The children in the household ranged in age from 2 to 22 years
of age, although at least one child in the household had to be school-aged in order to meet the
eligibility criteria for participation in the study.
Procedures
The sample was drawn randomly from the city’s household census. Residents received letters
describing the study and then screeners followed up by telephone to determine whether they were
eligible and willing to participate.
As with other studies relying on public databases to develop their random samples, it is very
difficult to determine a response rate. Many people contacted refused to give demographic
information, so we were not able to determine how many who did not respond were actually
eligible to participate, nor does the household census provide information on presence of schoolaged children.
Data were collected during the 2003-2004 and 2004-2005 school years. Trained interviewers
arranged 45-minute face-to-face quantitative interviews with parents at a time and place
convenient to the participants. In the case of two-parent families, mothers and fathers were
interviewed privately. Parents also received a 15-minute mailed questionnaire to be completed in
advance and returned at the time of the interview. Each parent received $50 for participating.
Measures
Work resource fit was measured using a 5-item subscale (see Appendix) of a global
community resource fit measure developed for this study through a series of open-ended
interviews with parents and guardians of school-aged children. In the full measure, participants
rated their satisfaction with 36 components of community resource fit in the areas of work,
school, after-school activities, public transportation, transportation to and from school, and
transportation to and from after-school activities on a scale ranging from 1 (completely
dissatisfied) to 7 (completely satisfied). Internal consistency in the full sample is excellent for the
When Your Work
7
scale as a whole (Mothers: " = .86; Fathers: " = .90) and very good for the work resource fit
subscale (Mothers: " = .77; Fathers: " = .80).
Job disruptions was assessed using a measure developed by the authors (Barnett & Gareis,
2006b). Respondents answered 12 questions about job disruptions in the areas of missed work,
distractions on the job, not meeting expectations, and quality of work. Sample items include
“During the past three months, not counting vacation days, personal days, or holidays that were
scheduled in advance, how many times have you missed a full day of work for any reason,”
“How often have you felt distracted at work by non-work issues,” rated on a scale ranging from 1
(never) to 5 (very often), “How often did you miss a deadline at work due to non-work issues,”
rated on a scale ranging from 1 (never) to 5 (very often), and “How would you rate the quality of
the work you did,” rated on a scale ranging from 1 (poor) to 4 (excellent). Because individual job
disruptions items were rated on different response scales, participants’ responses to each item
were converted to standard or z-scores and then averaged. Thus, overall job disruptions scores
can be interpreted in terms of standard deviations around a mean of zero. The measure has
acceptable internal consistency in the full sample (Mothers: " = .67; Fathers: " = .76).
Intention to leave one’s job within a year was assessed by asking participants to rate on a
scale ranging from 1 (extremely unlikely) to 7 (extremely likely) the likelihood that they would
voluntarily terminate their employment with their organization within the next 12 months.
Job-role quality was assessed with a 28-item measure on which participants were instructed
to rate on a scale ranging from 1 (not at all) to 4 (considerably) the extent to which each item
was rewarding or of concern (Barnett & Brennan, 1995, 1997). Items covered job conditions in
the areas of skill discretion, decision authority, job demands, schedule control, pay adequacy, job
security, and supervisor relations. Sample reward items include “challenging or stimulating
work” and “your supervisor’s respect for your abilities.” Sample concern items include “the job’s
dullness, monotony, lack of variety” and “having to juggle conflicting tasks or duties.” Concerns
were negatively weighted and rewards positively weighted in constructing the score, which was
the weighted average of item scores (Barnett et al., 1993). Internal consistency is excellent in the
full sample (Mothers: " = .91; Fathers: " = .89).
Marital status was coded as a dichotomous variable (0 = single, 1 = married). Logged per
capita family income was assessed by asking participants to indicate their yearly household
income from all sources before taxes. To obtain per capita household income, we divided this
figure by the number of people in the household. The distribution of this variable is skewed, so
we used the natural log of per capita income in the analyses. Work hours was assessed by asking
participants how many hours they work per week, on average, including overtime, at all jobs.
Negative affectivity was assessed with the 10-item Trait Anxiety Scale (Spielberger, 1983) on
which participants indicated on a 4-point scale from 1 (almost never) to 4 (almost always) how
characteristic traits were of them; items include “I am a steady person” and “I have selfconfidence.” The overall negative affectivity score is the mean response to all items answered.
Internal consistency is very good in the full sample (Mothers: " = .83; Fathers: " = .81).
Results
As shown in Table 1, there are significant bivariate correlations between work resource fit
and all three outcomes (job disruptions, intention to leave one’s job, and job-role quality). In
When Your Work
8
order to determine whether these relationships held up while controlling for important covariates,
we estimated a simultaneous regression model using work resource fit to predict the three
outcomes, with covariates of marital status, family income, work hours, and negative affectivity.
As shown in Table 2, work resource fit is significantly negatively related to job disruptions.
That is, as predicted in Hypothesis 1, the better the mother’s work resource fit, the lower the
level of job disruptions she reports. As an indication of the size of the main effect of work
resource fit on job disruptions, Cohen’s d is -0.59, which is classified as a moderate-sized effect
(Cohen, 1988).
Table 3 shows the results for intention to leave one’s job. Again, work resource fit is
significantly related to intention to leave one’s job: As predicted in Hypothesis 2, the better the
mother’s work resource fit, the less likely she is to report intending to leave her job within the
next year. As an indication of the size of the main effect of work resource fit on intention to
leave, Cohen’s d is -0.90, which is classified as a large effect.
The results for job-role quality are shown in Table 4. Once again, work resource fit is
significantly related to job-role quality, this time in a positive direction: As predicted in
Hypothesis 3, the better the mother’s work resource fit, the higher she rates her job-role quality.
This effect is quite substantial, with a Cohen’s d of +1.29, which is classified as a very large
effect.
To examine the research questions, we also tested whether the relationships between work
resource fit and the three outcomes were moderated by marital status, by family income, or by
work hours by conducting three more sets of regression analyses, one set adding the interaction
term WORK RESOURCE FIT x MARITAL STATUS, one set adding the interaction term
WORK RESOURCE FIT x FAMILY INCOME, and one set adding the interaction term WORK
RESOURCE FIT x WORK HOURS to the main-effects models described above.
For Research Question 4, we found that marital status moderated the relationship between
work resource fit and intention to leave one’s job (see Table 3): The addition of the interaction
term resulted in a significant increment to R2 over that associated with the main-effects model;
Fchange(1,59) = 5.25, p = .025. Specifically, as shown in Figure 1, the relationship between work
resource fit and intention to leave is significantly stronger for married mothers than for single
mothers. Cohen’s d for this interaction is -0.64, classified as a moderate-sized effect.
With regard to Research Question 5, we found that there was a trend for family income to
moderate the relationship between work resource fit and job-role quality (see Table 4);
Fchange(1,59) = 2.88, p = .095. As shown in Figure 2, the relationship between work hours and
job-role quality is marginally stronger for mothers whose income is lower. Cohen’s d for this
interaction is -0.42, classified as a moderate-sized effect.
Finally, for Research Question 6, we found that work hours moderated the relationship
between work resource fit and job-role quality (see Table 4); Fchange(1,59) = 5.39, p = .024. As
shown in Figure 3, the relationship between work hours and job-role quality is significantly
stronger for mothers who work longer hours, whereas for mothers who work shorter hours, there
is essentially no relationship between work resource fit and job-role quality. Under conditions of
high work resource fit, mothers who work long hours have the same high levels of job-role
When Your Work
9
quality as do their counterparts who work shorter hours. Cohen’s d for this interaction is +0.62,
classified as a moderate-sized effect.
Discussion and Conclusions
In this random sample of 68 employed mothers with at least one school-aged child, work
resource fit was a significant predictor of three work-related outcomes: job disruptions,
intentions to leave one’s job within a year, and job-role quality. Specifically, mothers with high
work resource fit reported fewer job disruptions, lower intention to leave their jobs, and higher
job-role quality than did mothers with low work resource fit. Thus, in this sample, work resource
fit seems to serve as a safety valve with positive effects on three important work-related
variables.
In addition, marital status moderated the relationship between work resource fit and intention
to leave one’s employer within a year. There is a strong relationship between work resource fit
and intention to leave among married mothers, whereas there is essentially no relationship
between these two variables among single mothers. In other words, married employed women
seem to have more flexibility to react to change in work resource fit, at least with respect to
turnover intention: They are much more likely to report low intention to leave their jobs when
work resource fit is high and a somewhat greater likelihood of leaving their jobs when work
resource fit is low. Stated differently, with regard to this work-related outcome, married women,
seem to benefit from having two safety valves: high resource fit at work and a marital partner at
home who can provide instrumental and emotional support. In addition, preliminary analyses
indicated that single mothers rely more heavily than married mothers on a network of “others” to
help them handle their family demands. This dependence may tie them to their community in
ways that make it less likely that they will leave their place of employment even when work
resource fit is low.
Although there were no direct effects of family income on any of the three outcomes, there
was a trend for family income to moderate the relationship between work resource fit and jobrole quality. The job-role quality of low-income employed mothers is marginally more
responsive to work resource fit than is the job-role quality of their high-income counterparts.
Specifically, when work resource fit is low, low-income women report lower job-role quality
than high-income women. In contrast, when work resource fit is high, the difference between the
two groups of mothers is negligible. Thus, it appears that high work resource fit is a safety valve
that protects employed mothers from low job-role quality when family income is low.
Work hours moderated the relationship between work resource fit and job-role quality.
Mothers who worked longer hours benefitted more than those who worked shorter hours from
the safety valve of high work resource fit. In other words, high work resource fit may act as a
safety valve releasing the strain that might otherwise diminish the job-role quality of mothers
who work long hours; instead, when work resource fit is high, their job-role quality is
indistinguishable from that of their counterparts who work fewer hours. These findings lend
support to Emlen’s insight that working families benefit when they have a safety valve at work,
at home, or at their child-care provider. They also support the corollary that the more safety
valves working families have, the greater the likelihood of positive outcomes (Barnett & Gareis,
under review).
When Your Work
10
With respect to marital status, it might be assumed that single-parent families would be more
reactive than dual-earner families to low work resource fit (particularly low flexibility) because
they have fewer resources (e.g., another responsible adult to help with child care). It might be
further assumed that the lack of family safety valves would increase single mothers’ likelihood of
job disruptions, of leaving their jobs, and of poor job-role quality. However, we found no
evidence that single mothers were more vulnerable to poor work resource fit. This result may be
due to the fact that our analyses controlled for family income. Previous research (Barnett &
Marshall, 1992) indicates that when family income was controlled, there was no significant
difference in psychological distress between married and single employed mothers. Thus, the
higher psychological distress of employed single mothers was due to their relatively lower family
income. In light of these findings, the general assumption that employed single mothers are
especially vulnerable to job-related stressors needs to be revisited.
It is of some interest that the annual family income was not directly related to any of the three
work-related outcome variables. The lack of significant findings may be due to the fact that the
median family income was $72,500 (the range was from $20,000 to $175,000). A number of
these families may not be experiencing severe financial pressures which could result in higher
levels of job-related negative performance indicators. Moreover, they may be able to purchase
after-school child care and other services that could increase the likelihood of their being able to
handle work and family demands without undue negative effects on their job performance. Stated
differently, adequate family financial resources can constitute a “safety valve” that protects
working families from untoward consequences. (However, it should be noted that almost onethird of the sample – 21 families, or 30.9% – has a total family income of less than $50,000.)
In this sample, 50 mothers (73.5%) worked full time (i.e., 35 or more hours per week), and of
those, five worked at least 60 hours per week. However, work hours was significantly and
negatively related only to job-role quality, not to job disruptions or to intentions to leave one’s
job. Moreover, work hours did not moderate the relationships between work resource fit and
intention to quit or job disruptions. It may be that in times of high job insecurity and economic
uncertainty, employees, especially if they have school-aged children, will tolerate considerable
distress before thinking seriously about giving up “the devil they know” for an uncertain future
employment. As discussed above, work hours moderated the relationship between work resource
fit and job-role quality, such that when work resource fit was high, there was no difference in
job-role quality between employed mothers who worked long compared to short hours. In other
words, when employed mothers have the safety valve of high work resource fit they can work
long hours without reporting the negative effects on job-role quality they might otherwise
experience.
Thus, while we find main effects of work resource fit on all of the work-related outcomes
we examined (i.e., job disruptions, turnover intentions, and job-role quality), there was only one
main effect of a single secondary safety valve, work hours, on a single outcome, job-role quality.
There were no main effects of work hours on the other outcomes, and no main effects of the
other secondary safety valves (i.e., marital status and family income) on any of the three
outcomes. Thus, there is little evidence in the present study that these secondary safety valves
have additive effects on outcomes in combination with work resource fit. Instead, we found
When Your Work
11
evidence in the form of two significant interactions and one marginally significant interaction
that the secondary safety valves we examined operate in a synergistic fashion with work resource
fit on outcomes. Further research is needed in order to determine whether these findings hold for
different safety valves and for different outcomes.
As with any study, this one has limitations. First, the sample size was limited, and all
participants resided in a single city in the Boston area, raising issues of generalizability. In future
research, larger samples should be drawn from a variety of communities that differ with respect
to the resources they offer to working families. Nevertheless, the present research yielded
interesting findings supporting the utility of including the community resource fit construct in
future studies of work-family issues.
When Your Work
12
References
Barnett, R. C. (1994). Home-to-work spillover revisited: A study of full-time employed women
in dual-earner couples. Journal of Marriage and the Family, 56, 647-656.
Barnett, R. C., & Brennan, R. T. (1995). The relationship between job experiences and
psychological distress: A structural equation approach. Journal of Organizational Behavior,
16, 259-276.
Barnett, R. C., & Brennan, R. T. (1997). Change in job conditions, change in psychological
distress, and gender: A longitudinal study of dual-earner couples. Journal of Organizational
Behavior, 18, 253-274.
Barnett, R. C., & Gareis, K. C. (2006a, March). Community: The missing link in the work-family
literature. Paper presented at the 23rd Annual Claremont Symposium on Applied Social
Psychology, Work and Families: Changing Realities, Claremont Graduate University,
Claremont, CA.
Barnett, R. C., & Gareis, K. C. (2006b). Antecedents and correlates of parental after-school
stress: Exploring a newly identified work-family stressor. American Behavioral Scientist, 49,
1382-1399.
Barnett, R. C., & Gareis, K. C. (2006c). Parental after-school stress and psychological wellbeing. Journal of Marriage and Family, 68, 101-108.
Barnett, R. C., & Gareis, K. C. (under review). Does community resource fit matter to fathers? A
study of employed fathers of school-aged children.
Barnett, R. C., Gareis, K. C., & Brennan, R. T. (1999). Fit as a mediator of the relationship
between work hours and burnout. Journal of Occupational Health Psychology, 4, 307-317.
Barnett, R. C., & Marshall, N. L. (1992). Worker and mother roles, spillover effects, and
psychological distress. Women & Health, 18, 9-40.
Barnett, R. C., Marshall, N. L., Raudenbush, S. W., & Brennan, R. T. (1993). Gender and the
relationship between job experiences and psychological distress: A study of dual-earner
couples. Journal of Personality and Social Psychology, 65, 794-806.
Becker, P. E., & Moen, P. (1999). Scaling back: Dual-earner couples' work-family strategies.
Journal of Marriage and the Family, 61, 995-1007.
Bianchi, S. M. (2000). Maternal employment and time with children: Dramatic change or
surprising continuity? Demography, 37, 401-414.
Bond, J. T., Galinsky, E., & Swanberg, J. E. (1998). The 1997 National Study of the Changing
Workforce. New York: Families and Work Institute.
Boushey, H. (2005). Are women opting out? Debunking the myth. Washington DC: Center for
Economic and Policy Research.
Boushey, H. (2006). Are mothers really leaving the workplace? Chicago, IL: Council on
Contemporary Families and the Center for Economic and Policy Research.
Brennan, R. T., & Barnett, R. C. (1998). Negative affectivity: How serious a threat to self-report
studies of psychological distress? Women's Health: Research on Gender, Behavior, and
Policy, 4, 369-384.
Bureau of Labor Statistics. (2005). Women in the labor force: A databook (Report 985).
Retrieved September 26, 2005, from http://www.bls.gov/cps/wlf-databook-2005.pdf
When Your Work
13
Bureau of Labor Statistics. (2006). Employment characteristics of families in 2005 (News
Release). Washington, DC: U.S. Department of Labor.
Carr, D. (2002). The psychological consequences of work-family trade-offs for three cohorts of
men and women. Social Psychology Quarterly, 65, 103-124.
Chodorow, N. (1978). The reproduction of mothering: Psychoanalysis and the sociology of
gender. Berkeley, CA: University of California Press.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Cohen, D. A., Farley, T. A., Taylor, S. N., Martin, D. H., & Schuster, M. A. (2002). When and
where do youths have sex? The potential role of adult supervision. Pediatrics, 6, 66.
Emlen, A. C. (1997, May). Quality of childcare and special needs of children who have
emotional or behavioral problems. Paper presented at the Building on Family Strengths: A
National Conference on Research and Services in Support of Children and Families,
Portland, OR.
Fernandez, J. P. (1986). Child care and corporate productivity. Lexington, MA: Lexington
Books.
Galinsky, E., Bond, J. T., & Friedman, D. E. (1993). The changing workforce: Highlights of the
national study. New York: Families and Work Institute.
Gareis, K. C., & Barnett, R. C. (2001, August). Schedule fit and stress-related outcomes among
women doctors with families. Paper presented at the annual meeting of the American
Psychological Association, San Francisco, CA.
Gareis, K. C., & Barnett, R. C. The development of a new measure for work-family research:
Community Resource Fit. Manuscript in preparation.
Gareis, K. C., Barnett, R. C., & Brennan, R. T. (2003). Individual and crossover effects of work
schedule fit: A within-couple analysis. Journal of Marriage and Family, 65, 1041-1054.
Gerson, K. (1986). Hard choices: How women decide about work, career and motherhood
(California series on social choice & political economy) (reprint ed.). Berkeley: University of
California Press.
Gilligan, C. (1982). In a different voice: Psychological theory and women's development.
Cambridge, MA: Harvard University Press.
Glass, J. L., & Estes, S. B. (1997). The family responsive workplace. Annual Review of
Sociology, 23, 289-313.
Greenhaus, J. H., Parasuraman, S., & Collins, K. M. (2001). Career involvement and family
involvement as moderators of relationships between work-family conflict and withdrawal
from a profession. Journal of Occupational Health Psychology, 6, 91-100.
Heymann, J. (2000). The widening gap: Why America's working families are in jeopardy--and
what can be done about it. New York: Basic Books.
Hofferth, S. L., Brayfield, A. A., Deich, S. G., & Holcomb, P. A. (1991). National Child Care
Survey, 1990: A National Association for the Education of Young Children (NAEYC) study.
Lanham, MD: Urban Institute Press.
Jacobs, J. A., & Gerson, K. (2004). The time divide: Work, family and gender inequality.
Cambridge, MA: Harvard University Press.
When Your Work
14
Kurz, D. (2002). Caring for teenage children. Journal of Family Issues, 23, 748-767.
Lee, T. W., & Maurer, S. D. (1999). The effects of family structure on organizational
commitment, intention to leave and voluntary turnover. Journal of Managerial Issues, 11,
493-513.
National Center for Schools and Communities. (1999). After-school programs: An analysis of
need, current research, and public opinion. New York: Fordham University.
Newman, S. A., Fox, J. A., Flynn, E. A., & Christeson, W. (2000).America's after-school choice:
The prime time for juvenile crime, or youth enrichment and achievement. Retrieved
November 19th, 2004, from http://www.fightcrime.org/reports/as2000.pdf
Schor, J. B. (1991). The overworked American: The unexpected decline of leisure. New York:
Basic Books.
Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y ed.). Palo Alto,
CA: Consulting Psychologists Press.
Stolzenberg, R. M. (2001). It's about time and gender: Spousal employment and health. American
Journal of Sociology, 107, 61-100.
U.S. Census Bureau. (2003). Table H3. Fertility indicators for women in their thirties: Selected
years 1976 to present (Current Population Survey). Washington, DC: Author.
When Your Work
Appendix
Work Resource Fit Subscale
1. The way your work schedule fits with your child(ren)’s schedule(s)
2. The flexibility available at your workplace to handle emergencies
3. The flexibility available at your workplace to attend to family needs
4. Your ability to work at home if necessary
5. Your ability to bring child(ren) to work if necessary
Note. Response scale ranged from 1 (completely dissatisfied) to 7 (completely satisfied).
15
Table 1: Descriptive Statistics and Intercorrelations among Variables
Mean (SD)
1
2
3
4
5
1. Work Hours
38.18 (12.07)
2. Marital Status
0.59 (0.50)
-.33*
---
3. Family Income
$72,664 (37,507)
-.23†
.67*
---
4. Neg Affectivity
1.83 (0.46)
-.04
.02
-.03
---
5. Work Resource Fit 5.18 (1.24)
-.10
.14
.06
-.35*
---
6
7
8
---
6. Job Disruptions
0.00 (0.47)
.07
-.09
.03
.46*
-.39*
---
7. Intention to Leave
1.94 (1.58)
.00
-.09
.03
.11
-.41*
.14
---
-.49*
.64*
-.51*
-.41*
8. Job-Role Quality
1.14 (0.92)
-.26*
.18
.11
Note. N = 68. For marital status, 0 = single (n = 28), 1 = married (n = 40).
† p < .10. * p < .05.
---
Table 2: Relationship Between Work Resource Fit and Job Disruptions
Main Effects
B
$
(SE)
Work Resource Fit
-0.09*
-.24
(0.05)
Marital Status
-0.13
-.14
(0.12)
Family Income
0.37
.18
(0.24)
Work Hours
0.00
.03
(0.01)
Negative Affectivity
0.39*
.38
(0.12)
Job Disruptions
R2
.30
Adjusted R2
.25
Note. N = 68. For marital status, 0 = single (n = 28), 1 = married (n = 40).
† p < .10. * p < .05.
When Your Work
18
Table 3: Relationship Between Work Resource Fit and Intentions to Leave One’s Job with A Year
Turnover Intentions
Main Effects
Moderation By Marital Status
B
$
(SE)
B
$
(SE)
Work Resource Fit
-0.53*
-.42
(0.16)
-0.70*
.58
(0.18)
Marital Status
-0.34
-.11
(0.44)
-0.46
-.35
(0.01)
Family Income
0.85
.12
(0.88)
1.28
.22
(0.58)
Work Hours
-0.01
-.07
(0.02)
-0.01
-.35
(0.01)
Negative Affectivity
-0.13
-.04
(0.43)
-0.17
-.36
(0.24)
WRF x MAR STAT
--
--
--
-0.69*
-.34
(1.16)
R2
.18
Adjusted R2
.12
Note. N = 68. For marital status, 0 = single (n = 28), 1 = married (n = 40).
† p < .10. * p < .05.
.25
.17
Table 4: Relationship Between Work Resource Fit and Job-Role Quality
Job-Role Quality
Main Effects
B
Moderation by Work Hours
$
(SE)
B
Moderation by Income
$
(SE)
B
$
(SE)
Work Resource Fit
0.37*
.50
(0.07)
0.26*
.35
(0.08)
.35*
.47
(0.07)
Marital Status
0.06
.03
(0.19)
0.04
.02
(0.19)
.12
.06
(0.19)
Family Income
0.14
.04
(0.38)
0.13
.03
(0.37)
.20
.05
(0.38)
Work Hours
-0.02*
-.21
(0.01)
-0.02*
-.26
(0.01)
-.02*
-.20
(0.01)
Negative Affectivity
-0.65*
-.33
(0.19)
-0.65*
-.33
(0.18)
-.65*
-.32
(0.19)
.25
(0.01)
--
--
-.60†
WRF x WORK HRS
--
--
--
0.01*
WRF x FAM INCOME
--
--
--
--
R2
.54
Adjusted R2
.51
Note. N = 68. For marital status, 0 = single (n = 28), 1 = married (n = 40).
† p < .10. * p < .05.
--
--.15
.58
.57
.54
.52
-(0.36)
Figure Captions
Figure 1. The Relationship Between Work Resource Fit and Intentions to Leave One’s Job with
A Year Is Moderated by Marital Status.
Figure 2. The Relationship Between Work Resource Fit and Job-Role Quality Shows A Trend
To Be Moderated by Family Income.
Figure 3. The Relationship Between Work Resource Fit and Job-Role Quality Is Moderated by
Work Hours.
When Your Work
21
When Your Work
22
When Your Work
23
When Your Community Works
1
Running head: WHEN YOUR COMMUNITY WORKS FOR YOU
When Your Community Works for You: A Within-Couples Analysis of
Community Resource Fit and Psychological Distress among Parents of School-Aged
Children1
Rosalind Chait Barnett and Karen C. Gareis
Brandeis University
Abstract
[XXX]
KEYWORDS: [XXX]
Several studies suggest that the degree of fit between community resources and the needs of
working parents with school-aged children may affect parents’ well-being. These studies were
informed by the work of Emlen and his colleagues (Emlen & Koren, 1984; Emlen, Koren, &
Schultze, 2000), who suggest that the well-being of employed parents depends on their access to
“safety valves” (i.e., sources of flexibility), be they at home, at work, or, expanding on his
original concern with very young children, on such community institutions as schools. Previous
studies focused on two community resources: school resource fit and work resource fit (Barnett
& Gareis, 2006a, 2006b). The focus on school resource fit reflects the growing understanding
that community assets, especially schools, can play an important role in facilitating or impeding
the adjustment of working families. For example, if school schedules mesh well with parents’
work schedules, one potential source of distress is reduced. The focus on work resource fit grows
out of the literature indicating that employed parents, especially mothers, tend to adjust their
work schedules (i.e., cut back at work, drop out of the workforce) in response such family needs
as the birth of a child or the needs of school-aged children (Glass & Estes, 1997). Thus, flexible
work schedules ought to be related to mothers’ levels of distress.
In the present analysis, we build on and extend these previous studies by taking the couple, in
contrast to the individual parent, as the unit of analysis, and by asking whether work and school
resource fit are related similarly to mothers’ and fathers’ psychological distress within couples.
We also ask whether the number of hours each partner works affects the relationship between
school and work resource fit on the one hand and psychological distress on the other. We address
these questions in a random sample of 58 married couples (N = 116 individuals), who have at
1
Data for this analysis were gathered under a grant from the Alfred P. Sloan Foundation
(2003-12-1) to the first author. We gratefully acknowledge the contributions of our interviewers,
Sarah Anderson, Joyce Buni, Connie Festo, Carol Genovese, Eleanor Jacobs, and Heidi La Bash.
Correspondence concerning this analysis should be addressed to Rosalind Chait Barnett,
Brandeis University, Women's Studies Research Center, Mailstop 079, 515 South Street,
Waltham, MA 02453-2720, rbarnett@brandeis.edu, (781) 736-2287, FAX (781) 736-4881.
When Your Community Works
2
least one school-aged child (i.e., K though 12th grade) and who vary in employment status.
Literature Review
According to Emlen (1997), employed parents with young children seek to increase
work-family fit through a “flexibility solution.” High fit occurs when there is flexibility in work,
family, and child care. Low flexibility in any domain reduces work/family fit and stresses the
family system. We expand Emlen’s definition of flexibility to include the fit between employed
parents’ needs and school resource fit. In a previous analysis with data from the same sample as
is used in the present analysis, employed fathers (but not mothers) of school-aged children
benefitted if their children’s schools provided them with a safety valve (Barnett & Gareis,
2006a). Fathers with high school resource fit reported low psychological distress and high jobrole quality. Based on these results, we expect that school resource fit will again be related to
fathers’, but not mothers’, psychological distress and that the gender difference will be
significant.
H1:
School resource fit will be significantly related to psychological distress for fathers,
but not mothers.
Although we also test the interaction term school resource fit X work hours, there is insufficient
literature upon which to formulate a specific hypothesis about the joint effect of these variables.
With respect to work resource fit, when their work schedules do not enable them to meet
their family responsibilities, women are more likely than men to cut back their work hours (Glass
& Estes, 1997; Hill, Hawkins, Ferris, & Weitzman, 2001). It is also true that because employed
mothers work, on average, fewer hours than employed fathers, they have more flexibility in their
work schedules. For example, women are more likely than men to work part-time (Negrey,
1983). In contrast to employed mothers, employed fathers are more likely to take paid sick or
vacation time rather than reduce their work hours if they need to handle family matters. Thus,
compared to the employed fathers in this sample, the employed mothers’ work resource fit may
be more closely related to their psychological distress. (It is important to note that not all fathers
or mothers in this sample were employed; see Participants, below). Stated differently, high work
resource fit provides a safety valve for working mothers that might ease the stress they would
otherwise experience.
In a previous analysis (Barnett & Gareis, 2006b) with all the mothers in the original sample
(i.e., married and single) employed mothers who reported high work resource fit also reported
high job-role quality. Furthermore, among low work hours mothers, job-role quality was high
regardless of work resource fit. In contrast, among high work-hours mothers, job-role quality
increased as work resource fit increased. And, job-role quality has consistently been shown to be
a strong predictor of psychological distress (Barnett, Marshall, Raudenbush, & Brennan, 1993;
Barnett & Brennan, 1995, 1998). Based on this body of literature, we expect work resource fit to
be related to psychological distress for employed mothers, but not fathers. Moreover, we expect
the gender difference to be significant.
When Your Community Works
H2:
3
Work resource fit will be significantly related to psychological distress for mothers,
but not fathers.
Here, too, we are unable to propose a specific hypothesis concerning the interaction effect of
school resource fit x work hours.
We include negative affectivity as a covariate. Negative affectivity is a mood-dispositional
trait to view the world negatively that is thought to account for spuriously high correlations
between self-report measures of predictor and outcome variables, especially in cross-sectional
analyses (Brennan & Barnett, 1998). We also include as covariates logged per capita household
income, number of children at home, average age of children at home, and number of hours
worked per week.
Method
Participants
As part of a larger study of how parents of school-aged (i.e., in grades K-12) children in one
Boston-area community coordinate their work schedules with their children’s school, afterschool, and transportation schedules, we interviewed all parents in 29 dual-earner families, 29
families with one main breadwinner and one stay-at-home or part-time employed parent, and 29
single-parent families (total N = 145). For the present paper, we report only on the data from the
58 married couples (n = 116 individuals).
Half of the couples (n = 29) are two-earner couples, defined as having one parent who is
employed full-time and one parent who works 20 or more hours per week. The other half of the
couples (n = 29) are families with one main breadwinner, defined as having one parent who is
employed full-time and one parent who works 0 to 19 hours per week. In the two-earner couples,
all of the fathers and 20 of the mothers are employed full-time and 9 of the mothers are employed
part-time. In the couples with one main breadwinner, 23 of the fathers and 6 of the mothers are
employed full-time, 1 of the fathers and 5 of the mothers are employed part-time, and 5 of the
fathers and 18 of the mothers are not employed outside the home.
Procedures
The sample was drawn randomly from the community’s household census. Residents
received letters describing the study and then screeners followed up by telephone to determine
whether they were eligible and willing to participate.
Data were collected during the 2003-2004 and 2004-2005 school years. Trained interviewers
arranged private 45-minute face-to-face quantitative interviews with each parent at a time and
place convenient to the participants. Parents also received a 15-minute mailed questionnaire to be
completed in advance and returned at the time of the interview. Each parent received $50 for
participating.
As with other studies relying on public databases to develop their random samples, it is very
difficult to determine a response rate. Many people contacted refused to give demographic
When Your Community Works
4
information, so we were not able to determine how many who did not respond were actually
eligible to participate, nor does the household census provide information on presence of schoolaged children.
Measures
School resource fit was measured using a 6-item subscale (Barnett & Gareis, 2006a) of a
global community resource fit measure developed for this study through a series of open-ended
interviews with parents and guardians of school-aged children. In the full measure, participants
rated their satisfaction with 36 components of community resource fit in the areas of work,
school, after-school activities, public transportation, transportation to and from school, and
transportation to and from after-school activities on a scale ranging from 1 (completely
dissatisfied) to 7 (completely satisfied). Internal consistency in the full sample is acceptable for
the school resource fit subscale (Mothers: " = .73; Fathers: " = .69).
Work resource fit was measured using a 5-item subscale (Barnett & Gareis, 2006b) of the
above mentioned global community resource fit measure. Internal consistency in the full sample
is very good for the work resource fit subscale (Mothers: " = .77; Fathers: " = .80).
Psychological distress was assessed using a state measure asking participants to indicate on a
scale ranging from 0 (not at all) to 4 (extremely) how often in the past week they were bothered
by each of 10 symptoms of anxiety and 14 symptoms of depression (Derogatis, 1975). Anxiety
and depression scores were combined to create a measure of psychological distress (Barnett,
Marshall, Raudenbush, & Brennan, 1993). The combined score has excellent internal consistency
in the full sample (Mothers: " = .95; Fathers: " = .91).
Work hours was assessed by asking participants how many hours they work per week, on
average, including overtime, at all jobs. Logged per capita household income was assessed by
asking participants to indicate their yearly household income from all sources before taxes. To
obtain per capita household income, we divided this figure by the number of people in the
household. The distribution of this variable is skewed, so we used the natural log of per capita
income in the analyses. Number of children at home and average age of children at home are
self-explanatory.
Negative affectivity was assessed with the 10-item Trait Anxiety Scale (Spielberger, 1983) on
which participants indicated on a 4-point scale from 1 (almost never) to 4 (almost always) how
characteristic traits were of them; items include “I am a steady person” and “I have selfconfidence.” The overall negative affectivity score is the mean response to all items answered.
Internal consistency is very good in the full sample (Mothers: " = .83; Fathers: " = .81).
Analytic Overview
[XXX]
Results
Descriptive Results
Among the married couples in this study, on average, the fathers worked 44.4 hours per week
When Your Community Works
5
(SD = 18.6) and the mothers worked 24.0 hours per week (SD = 19.5). Annual household income
ranged from $28,500 to $175,000, with a median of $87,500. The families had from one to four
children, with most having two (37.9%) or three (41.4%) children. The children in the household
ranged in age from infancy to 22 years of age, although at least one child in the household had to
be school-aged in order to meet the eligibility criteria for participation in the study.
Results of Hypothesis Testing
As shown in Table 1, school resource fit is significantly negatively related to psychological
distress for fathers, but not mothers. That is, when a father’s school resource fit is high, his
psychological distress is low. In contrast, for mothers, there is no significant effect of school
resource fit on distress. These results support Hypothesis 1. There is a marginally significant
gender difference in the association between school resource fit and psychological distress (P2 =
3.46, p = .060), offering some support, albeit weak, for the second part of Hypothesis 1.
Unexpectedly, work hours is significantly positively related to psychological distress for mothers;
that is, the longer a mother’s work hours, the greater her psychological distress.
Importantly, inclusion of the interaction term school resource fit X psychological distress for
each spouse significantly improves model fit as compared to the main-effects model (P2 = 7.01, p
= .029). Inspection of the interaction terms shows a significant moderating effect of work hours
on school resource for mothers and a marginally significant moderating effect of work hours on
school resource fit for fathers.
Specifically, when mothers work fewer hours, high school resource fit is associated with low
distress. That positive effect is lost when her work hours are long. In other words, married
women with at least one school-aged child experience some reduction in psychological distress
when their school resource fit is high and they work fewer hours, but mothers who work longer
hours do not enjoy this mental health benefit.
In contrast, among fathers, there is a trend for the beneficial effect of school resource fit to be
greater when the fathers work longer hours. Not surprisingly, there is a significant gender
difference in the moderating effect of work hours on the relationship between school resource fit
and psychological distress (P2 = 7.70, p = .006).
As shown in Table 2, work resource fit is significantly negatively related to psychological
distress for mothers, but not fathers, supporting Hypothesis 2. When a mother’s work resource fit
is high, her distress is low. For fathers, there is no significant relationship between work resource
fit and distress. Moreover, there is a significant gender difference in the association between
work resource fit and psychological distress (P2 = 4.65, p = .029), supporting the second part of
Hypothesis 2.
Adding the interaction term work resource fit X work hours did not significantly improve
model fit as compared to the main-effects model (P2 = 2.07, p = .357). In other words, regardless
of the number of hours they work, mothers derive a mental-health benefit from high work
resource fit that fathers do not. Although this analysis controlled for number of hours worked, the
results may have been affected by the fact that mothers were far more likely than fathers to work
less than full-time. The percentages not employed, part-time employed, and full-time employed
are 31.0%, 27.6%, and 41.4% for mothers and 8.6%, 1.7%, and 89.7% for fathers.
When Your Community Works
6
Discussion and Conclusions
In this random sample of 58 Boston-area married couples (N = 116) with at least one schoolaged child, work resource fit and school resource fit had beneficial associations with
psychological distress for mothers and fathers, but in different ways. These findings support the
study’s two hypotheses and are consistent with the results of previous studies conducted
separately on the mothers and fathers in this sample. Specifically, school resource fit had a
beneficial association with employed fathers’ psychological distress, but had no direct effect
among employed mothers, again controlling for number of hours worked. In contrast, work
resource fit benefitted mothers’ mental health, but not fathers’. Furthermore, work hours
moderated the relationship between school resource fit and psychological distress in opposite
ways for mothers and fathers, with school resource fit beneficial for mothers who work fewer
hours and (marginally) for fathers who work more hours.
Employed fathers, because they tend to work longer hours than employed mothers and to
have less flexible work schedules (Glass & Estes, 1997; Jacobs & Gerson, 2004; Judiesch &
Lyness, 1999), may be especially benefitted with respect to their psychological distress when
they have a safety valve in the form of high school resource fit. It is also possible that fathers may
have fewer expectations than mothers for involvement with their children’s schools and may
therefore be more easily satisfied. Expecting little, they may react more positively when there is
some match. In contrast, employed mothers, who are more likely than fathers to work less than
full-time or not at all, may be less reactive to the safety valve provided by good school resource
fit.
Interestingly, mothers, but not fathers, benefitted from high school resource fit when they
worked fewer rather than longer hours. When school resource fit was high and mothers worked
comparatively few hours, they reported low distress. Thus for mothers who work part-time or not
at all and who are involved with their children’s school, high school resource fit acts as safety
valve, decreasing the likelihood that they will report high psychological distress. In contrast, for
their counterparts who work full-time, high school resource fit does not provide an effective
safety valve, at least with respect to psychological distress. For fathers, the strong relationship
between high school resource fit and low distress was not affected by the number of hours they
worked. Thus, in dual-earner families, in which both parents work full-time, she derives no
mental health benefit from school resource fit, but he does.
One possible explanation of these interaction results is that mothers have historically felt, and
may still feel, more responsible than fathers for managing their children’s school-related issues. It
is also likely that schools expect more involvement from mothers than fathers. Mothers who
work shorter hours may be better able to accommodate school personnel’s desire for them to be
actively engaged with their children during the school day. Such involvement may cushion
employed mothers from the distress they might otherwise feel if their work schedules prevented
them from such participation. However, many mothers may not be able or willing to cut back
their work hours for financial or long-term career reasons (Lyness & Judiesch, 2001).
In addition, mothers who work short hours may be better able to attend such after-school
activities as parent-teacher conferences and school performances. In contrast, mothers who work
long hours may feel distressed because of their inability to “be there” for their children or
When Your Community Works
7
because of their inability to meet the expectations schools have for them. These mothers are
unlikely to reap as great a mental health benefit from high school resource fit as their peers who
work shorter hours. Thus, it behooves schools to modify their schedules and expectations to
better match the needs of full-time employed mothers. So doing would also likely improve
fathers’ school resource fit and thereby have a beneficial effect on his psychological distress.
If school schedules (e.g., schedule of parent conferences and school events) can better
accommodate the schedules of full-time working mothers and fathers, schools may become a
community resource that ameliorates some of the distress such parents would otherwise
experience. Thus, communities, if they tailor their resources to the needs of working parents with
school-aged children, can provide parents with safety valves (i.e., resources that add some
flexibility to their lives) that may have salutary effects on their well-being.
With respect to work resource fit, mothers derived a greater mental health benefit than
fathers. Moreover, this gender difference was significant, even controlling for the number of
hours worked. Previous research (Barnett & Gareis, 2006a) found that mothers with high work
resource fit reported fewer job disruptions, lower intention to leave their jobs, and higher job-role
quality than did mothers with low work resource fit. The findings from this study, suggest that, in
this sample, work resource fit seems to serve as a safety valve for mothers with positive effects
on yet another important outcome, namely, psychological distress. Finally, the relationship
between work resource fit and psychological distress did not vary by number of hours worked for
fathers or mothers.
This study, like others, has several limitations. Most importantly, the data are cross-sectional.
Longitudinal data are needed to better understand the causal relationship between community
resource fit and psychological distress. In addition, the data were collected from families in only
one community. Although there were differences between schools in scheduling, and differences
among the workplaces in which the parents were employed, it would be desirable for future
research to collect data from families in several communities with more variability in public
transportation, after-school programs, school transportation, and after-school transportation in
order to examine the effects of community resource fit in those areas. Nevertheless, this study
adds to the growing literature that highlights the important role that community resources can
play in easing the task of managing work and family demands for today’s growing number of
working families