work, family, and mental health: testing different models of work-family fit
TRANSCRIPT
Journal of Marriage and Family 65 (February 2003): 248–262248
JOSEPH G. GRZYWACZ Wake Forest University School of Medicine
BRENDA L. BASS University of Northern Iowa*
l
Work, Family, and Mental Health: Testing Different
Models of Work-Family Fit
Using family resilience theory, this study exam-ined the effects of work-family conflict and work-family facilitation on mental health among work-ing adults to gain a better understanding ofwork-family fit. Data from the National Survey ofMidlife Development in the United States (MI-DUS) were used to compare different combina-tions of work-family conflict and work-family fa-cilitation. Results suggest that family to workfacilitation is a family protective factor that offsetsand buffers the deleterious effects of work-familyconflict on mental health. The results across theseoutcomes suggest that work-family conflict and fa-cilitation must be considered separately, and thatadult mental health is optimized when family towork facilitation is high and family to work andwork to family conflict is low.
Flexibility and family is one of three major chal-lenges facing workers and employers at the dawnof the 21st century (U.S. Department of Labor,1999). As social and demographic trends createincreased urgency and demand for sufficient fam-ily time among workers, employers are requiringgreater levels of flexibility on the part of employ-
Department of Family and Community Medicine, WakeForest University Health Sciences, Medical Center Boulevard,Winston-Salem, NC 27157-1084 ([email protected]).
*Department of Design, Family, and Consumer Sciences,241 Latham Hall, University of Northern Iowa, Cedar Falls,IA 50614-0332.
Key Words: anxiety, depression, family resilience, mentalhealth, problem drinking, work-family fit.
ees in order to compete in the global marketplace.Active involvement in both work and family areboth widely viewed as signs of a life well-lived,yet the opposing pressures of work and family andthe sequela of work-family conflict have been as-sociated with clear detriments to individuals andfamilies (for recent meta-analysis see Allen, Herst,Bruck, & Sutton, 2000). Clearly how work andfamily intersect in a person’s life, or rather howthey fit, has important ramifications for individu-als and families; therefore, it is important to spec-ify and understand what constitutes this fit.
Historically, fit has not been well defined in thetheoretical and empirical literature. The prepon-derance of work-family research has conceptual-ized fit as the absence of work-family conflict (forrecent reviews of the literature see Barnett, 1998;Frone, in press; Greenhaus & Parasuraman, 1999;Perry-Jenkins, Repetti, & Crouter, 2000); howev-er, workers’ everyday experiences tell us thatwork and family are both sources of growth andsupport as well as burdens and strains (Barnett;Barnett & Hyde, 2001; Crouter, 1984). This evi-dence suggests the possibility that work and fam-ily can benefit each other, and compelling evi-dence suggests that work-family conflict is distinctfrom positive spillover or work-family enhance-ment (Grzywacz & Marks; Kirchmeyer, 1992a).Integrating the conflict and the enhancement per-spectives, work-family fit was recently conceptu-alized as the combination of enhancement andconflict (Barnett, 1998); unfortunately, whichcombination of enhancement and conflict best fa-cilitates individual, work, and/or family-relatedoutcomes remains to be specified. Therefore, the
249Work-Family Fit
primary goal of this study was to further specifythe fit construct by empirically testing three dif-ferent combinations of work-family conflict andwork-family facilitation.
LITERATURE REVIEW
Theoretical Background
Work-family conflict and work-family facilitationare central concepts in emerging perspectives onwork and family dynamics (Barnett, 1998; Frone, inpress). Work-family conflict represents incompati-bilities between work and family responsibilities be-cause of limited resources (Kahn, Wolfe, Quinn,Snoek, & Rosenthal, 1964) and is characterized byexperiences such as missing an important dinnerwith one’s partner because of working late or beingphysically drained at work because the individualwas up the night before with a sick child. Work-family facilitation, including related concepts suchas work-family compatibility (Barnett & Baruch,1985; Barnett & Hyde, 2001) and work-family en-hancement (Greenhaus & Parasuraman, 1999), rep-resents the synergies or complementarities that occurwhen individuals combine work and family (Frone,2003; Grzywacz, 2002). Although much less stud-ied, reports by workers who suggest that the patiencerequired in childrearing helps them interact more ef-fectively with coworkers or clients (Kirchmeyer,1992b), or accounts that paid work provides a need-ed reprieve that helps individuals better parent(Hochschild, 1997), both exemplify the process ofwork-family facilitation. Highlighting these two in-terconnections, Barnett (1998) proposed that work-family fit is the lived experience of combining workand family and the resulting ‘‘multiple dimensionsof compatibility [facilitation] and conflict’’ (p. 167).Barnett (1998, 1999) further contends that work-family fit is a pivotal experience affecting individ-uals and their families, as well as employers.
Although Barnett’s (1998) conceptual definitionof work-family fit provides an important startingpoint for work-family theory and research, it is lim-ited. Perhaps most notably, the fit component of themodel lacks a theoretical basis for explaining howconflict and facilitation operate together in shapingindividual, family, or organizational outcomes. Thistheoretical gap produces operational confusion: Arethe multiple dimensions of work-family experiencebest treated as independent, unique effects, orshould researchers use other mathematical repre-sentations of fit in empirical models? Moreover,how are these representations of fit to be interpret-
ed? Thus a theoretically based exploration of dif-ferent representations of the fit construct would en-rich Barnett’s model and provide a foundation forfuture research.
Family resilience theory, which argues that afamily’s resources or capabilities allow it to thrivein the face of significant risk, provides a valuableperspective for articulating a more precise speci-fication of work-family fit (for a recent discussionof the resilience perspective see Patterson, 2002).First, the processes involved in work-family fitmap nicely onto the central concepts of the familyresilience perspective. Work-family conflict,which is widely accepted as being bidirectional(i.e., work conflicts with family and family con-flicts with work; Crouter, 1984; Frone, Yardley, &Markel, 1997), exemplifies both enduring and ep-isodic forms of the demands concept in familyresilience theory. Additionally, work-family con-flict poses significant theoretical risk to individu-als and families because successful integration ofwork and family is posited to be a primary taskduring young and middle adulthood (Lachman &Boone-James, 1997). Empirically, enduring formsof conflict, such as the difficulty confronted byshift workers in finding adequate child care, aswell as specific instances of conflict such as hav-ing a sick child but not being able to take sickleave, are consistently linked to a variety of neg-ative individual and family outcomes (Allen et al.,2000; Heymann, 2000). Thus not only does work-family conflict map nicely onto the demand con-cept, but theory and empirical evidence suggest itposes significant risk to individuals and families.
The conceptual match between family capabil-ities and work-family facilitation, also posited tobe bidirectional (Grzywacz & Marks, 2000), fur-ther suggests that the family resilience perspectiveis an appropriate framework for considering work-family fit. First, the posited gains from role ac-cumulation such as role privileges, status security,and personality enrichment (Sieber, 1974) are con-ceptually isomorphic with those used to describethe resources required by families to accommo-date or adapt to imposing demands (Patterson,2002). Next, operational indicators of work-familyfacilitation frequently exemplify family behaviorpatterns that have been demonstrated to be im-portant in the face of family stress. For example,an item reflecting the extent to which conversa-tions with family members are helpful for resolv-ing difficulties at work has been used in severalstudies (Grzywacz & Marks; Kirchmeyer, 1992a),and such items are indicative of active coping: an
250 Journal of Marriage and Family
important family capability for resilience (Patter-son). Finally, workers frequently report that thebenefits of combining work and family make thestrains worthwhile (Barnett & Hyde, 2001), sug-gesting that varying levels of work-family facili-tation enable some families to thrive while con-fronting the enduring work-family challenge.
The family resilience framework is also helpfulbecause it articulates possible strategies underly-ing work-family fit that describe how conflict andfacilitation operate together (McCubbin, Mc-Cubbin, Thompson, & Thompson, 1998). Theelimination strategy reflects the extent to whichfamily capabilities, such as work-family facilita-tion, allow families to evaluate imposing demandsas nonthreatening and not requiring further atten-tion. For example, some workers carve out uniquework arrangements or negotiate innovative ar-rangements with their spouses or partners so thatthey can maximize the personal and familial gainsof paid work (e.g., role modeling for children, fi-nancial wherewithal) while minimizing work-fam-ily conflict (Hattery, 2001). The assimilation strat-egy, by contrast, recognizes the potential threat ofthe demand (e.g., work-family conflict); however,family capabilities allow the demand to be viewedas acceptable and manageable without altering afamily’s form or function. Akin to this strategy,Hattery interviewed a series of women who clear-ly recognized that they were making tradeoffs intheir attempts to weave work and family, but whofirmly believed that the benefits outweighed thestresses or conflicts. For example, one workingmother commented:
The costs are that I have absolutely no free time.I don’t do cross-stitch; I don’t read. You know,as far as having a lot of free time to myself, Idon’t have it anymore. To me, that kind of goesalong with it; I can deal with that. (Hattery, 2001,p. 63).
Thus family resilience theory would suggestthat work-family fit represents the extent to whichwork-family facilitation can eliminate experiencesof work-family conflict, or the extent to whichwork-family facilitation creates an environmentthat can tolerate experiences of work-family con-flict. Both of these strategies set up different em-pirical models for evaluating the overall fit be-tween work and family.
Independent effects and interactive effectsmodels capture the essence of the eliminationstrategy, whereas the assimilation strategy can beframed using a relative difference model. The in-
dependent effects model posits that family capa-bilities and family demands would exert additiveeffects on a given outcome. This model assumesthat the magnitude of the effect for each work-family process is different from the others, there-by drawing attention to both the levels of familycapabilities and demands and their salience or im-pact. Thus in this model work-family fit wouldrepresent the extent to which the level and sa-lience of work-family facilitation offset or elimi-nate the effects of work-family conflict. An ex-tension of this model is the interactive effectsmodel, whereby work-family fit would be char-acterized by interaction effects between the dif-ferent forms of work-family conflict and work-family facilitation. Akin to the classic stressbuffering hypothesis (Cohen & Wills, 1985), fa-cilitation would be particularly important whenconflict is high, but may be relatively unimportantwhen work-family conflict is low. In this model,family demands are redefined or their negative po-tential is lessened by the presence of adequatefamily capabilities, thereby minimizing or elimi-nating the consequences of the demand for theindividual or family. Finally, the assimilationstrategy emphasizes the joint relationship of fam-ily capabilities and demands as opposed to theirindependent effects. In this model, the effect of agiven demand and its corresponding capability areposited to be equal but opposite. Therefore, theprimary concern is the extent to which family ca-pabilities exceed the demands, or the degree towhich a demand can be assimilated into a family’sstructures and functions. This ability can be op-erationalized as the relative difference between aspecific family capability (e.g., work to family fa-cilitation) and a family demand (e.g., work to fam-ily conflict).
Empirical Background
Mental health or illness is frequently examined inwork-family studies, and it is also an appropriatedependent variable for this study. First, mental ill-ness can be seen as a crude indicator of a family’sinability to nurture and socialize family members,thereby meeting resilience theory’s requirement ofmodeling competence in core family functions(Patterson, 2002). Next, the widespread use ofmental health indicators in work-family studies(e.g., life satisfaction, depressive symptoms) sug-gest that mental health is sensitive to even minorvariation in work-family experiences (Frone,2000). Moreover, the widespread use of mental
251Work-Family Fit
health provides comparability between this andother studies. Finally, mental health is itself a re-source for individuals and families; thus it is apivotal marker of a family members’ individualand collective capacity to survive and thrive.
Individual and contextual factors are frequentlyimplicated in work-family conflict and facilitation,and by extension, work-family fit. Gender, for ex-ample, is often a central variable in studies ofwork and family. Although differences in themeaning of work- and family-related experiencesare frequently noted (e.g., Larson, Richards, &Perry-Jenkins, 1994), only rarely do gender dif-ferences in levels of work-family experiencesarise. Indeed, several researchers have commentedon the absence or inconsistency of gender differ-ences in feelings of work-family conflict or feel-ings of success in balancing work and family(Frone, 2003; Frone, Russell, & Cooper, 1993;Milkie & Peltola, 1999). Moreover, explicit testsof gender differences in the effects of work-familyexperiences on mental health frequently yield nullfindings, suggesting that the components of work-family fit exert comparable effects on men’s andwomen’s mental health (Frone et al., 1993; Grzy-wacz, 2000). Gender is treated as a control vari-able in this study, however, because work andfamily responsibilities are gendered and the prev-alence of mood and affective disorders is greateramong women than men (Kohn, Dohrenwend, &Mirotznik, 1998).
Family structure is another central variable instudies of work-family experiences. Familiescharacterized by two working parents with youngchildren are typically sampled in work-familystudies because they frequently report the highestlevels of work-family conflict (Ernst Kossek &Ozeki, 1998; Galinsky, Bond, & Friedman, 1993).However, adults in these families also may expe-rience more work-family facilitation than their un-married or childless counterparts (Grzywacz &Marks, 2000). Thus the common practice of fo-cusing on married couples with children creates aunique range of work-family experiences charac-terized by high levels of conflict and high levelsof facilitation that does not allow full explorationof different combinations of these processes.Moreover, restricting work-family studies to mar-ried people with children overlooks the fact non-married and childless individuals frequently takeon enormous family responsibilities such as pro-viding care for aging kin (Allen & Pickett, 1987;Marks, 1996). Finally, evidence suggests thatwork-family conflict and stress are relatively sta-
ble throughout midlife and that work-family fa-cilitation increases across the adult life course, in-dicating that the work-family interface is not onlyan issue for married couples with young children(Grzywacz, Almeida, & McDonald, 2002). Thusto maximize sample variability but yet acknowl-edge established differences in work-family ex-periences by family structure, marital status andparental status are controlled in this study.
For the reasons outlined above, gender, maritalstatus, and having a young child were adjusted forin the multivariate analyses. Additionally, work-family experiences and mental health have beenassociated with life course and socioeconomic lo-cation, as well as a variety of employment circum-stances (Barnett, 1998; Grzywacz, Almeida, &McDonald, 2002; Kessler et al., 1994; Kohn et al.,1998). Therefore, paralleling a recent study byFrone (2000), the analyses reported in this studyadjusted for the effects of age, race (non-White 51), educational attainment, household earnings,number of hours worked per week, and self-em-ployment status.
The primary objective of this study was to com-pare three different conceptual models of work-family fit. Consistent with family resilience theory,the general hypothesis guiding this study is that ahigher level of family demands (i.e., work to familyand family to work conflict) would be associatedwith more risk of mental illness, but that more fam-ily capabilities (i.e., work to family and family towork facilitation) would be associated with less riskof mental illness. However, given the absence oftheory and research regarding the nature and pur-pose of work-family facilitation (Frone, 2003), thisstudy examined two research questions regardingthe most appropriate model and the processes un-derlying work-family fit.
1. Which model of work-family conflict and fa-cilitation (i.e., independent effects, interactiveeffects, or relative difference) best predictsmental health among working adults?
2. Does work-family facilitation promote elimi-nation or assimilation strategies of adjustmentto work-family conflict?
METHOD
Data and Sample
The data for this study are from the National Sur-vey of Midlife Development in the United States(MIDUS) collected by the John D. and Catherine
252 Journal of Marriage and Family
T. MacArthur Foundation Research Network onSuccessful Midlife Development in 1995. MIDUSrespondents are a nationally representative generalpopulation sample of noninstitutionalized personsaged 25–74 who have telephones. The sample wasobtained through random digit dialing, with an ov-ersampling of older respondents and men made toguarantee a good distribution on the cross-classi-fication of age and gender. Sampling weights cor-recting for selection probabilities and nonresponseallow this sample to match the composition of theU.S. population on age, sex, race, and education.
MIDUS respondents first participated in a tele-phone interview lasting approximately 40 min-utes. The response rate for the telephone ques-tionnaire was 70%. Respondents to the telephonesurvey were then asked to complete two self-ad-ministered mail-back questionnaires. The responserate for the mail-back questionnaire was 86.8%.This yielded an overall response rate of 60.8%(0.70 3 0.868) for both parts of the survey. Theanalytic sample used here includes all part-timeand full-time employed respondents aged 24–62(N 5 1,986; 1,038 men and 948 women).
Dependent Variables
Dichotomous depression and anxiety (either gen-eralized anxiety or panic) disorder were opera-tionalized using the Composite International Di-agnostic Interview Short Form Scales (CIDI-SF;Kessler, Andrews, Mroczek, Ustun, & Wittchen,1998) assessed during the telephone interviewportion of the MIDUS. The CIDI-SF generatesscores for psychiatric disorders that representprobability estimates of meeting the full diagnos-tic criteria (DSM-III-R) for these disorders if therespondent were given the complete CompositeInternational Diagnostic Interview (Kessler et al.).A validation study of the short form scales indi-cated greater than 90% sensitivity, specificity, andoverall agreement between the CIDI-SF and thefull CIDI for each major diagnosis (Kessler et al.).On the basis of these results, Kessler et al. contendthat the CIDI-SF scales are ideal for use in gen-eral-purpose epidemiological studies (p. 183).
The depression scale is comprised of one stemquestion (i.e., ‘‘During the past 12 months, was thereever a time when you felt sad, blue, or depressedfor 2 weeks or more in a row?’’), and a series ofprobes to affirmative responses about nine specificsymptoms (e.g., losing interest in things, loss of ap-petite). Additionally, two questions probed the fre-quency within the referenced 2-week period in
which the respondent felt ‘‘sad, blue, or depressed’’and ‘‘had greater difficulty than usual fallingasleep.’’ The respondent was coded 1 for depressionif he or she reported yes to the stem question andscored seven or more on the probes.
A dichotomous anxiety variable was construct-ed with items from two different short form scalesthat used a similar format as that reported for de-pression. The stem question for the general anxi-ety scale asked, ‘‘. . . Thinking about the past 12months, how often did you have each of the fol-lowing reactions because of your worry?’’ andwas followed by a series of 10 probes reflectingspecific somatic responses to worry (e.g., restlessbecause of worry, and sore or aching muscles be-cause of tension). Respondents answering yes tothree or more probes were coded as 1 for generalanxiety disorder. Being coded 1 on the panic scalerequired the respondent to answer affirmatively toone of the stem questions (e.g., ‘‘During the past 12months, did you ever have a spell or an attack whenall of a sudden you felt frightened, anxious, or veryuneasy in a situation when most people would notbe afraid or anxious?’’) and then answer yes tothree or more of six probes about those experi-ences (e.g., ‘‘Did you tremble or shake?’’ or ‘‘Didyou have hot flashes or chills?’’). Anxiety disorderwas coded 1 if the respondent met the criteria forgeneral anxiety disorder or panic disorder.
The short form assessing the 12-month preva-lence of alcohol dependence was intact except forthe pivotal indicator of problem drinking. There-fore, a measure of problem drinking was createdby evaluating the internal consistency of sevensurvey items indicative of alcohol dependence(e.g., ‘‘Were you ever, during the past 12 months,under the effects of alcohol or feeling its after-effects in a situation that increased your chancesof getting hurt?’’ or ‘‘Did you have a period of amonth or more during the past 12 months whenyou spent a great deal of time using alcohol orgetting over its effects?’’ [Chronbach’s a 50.74]). A sum score was constructed from affir-mative responses and then dichotomized using acut-point to generate a similar prevalence rate foralcohol abuse and alcohol dependence that hasbeen obtained from other national estimates(Grant et al., 1994; Kessler et al., 1994). Individ-uals meeting three or more of seven criteria forabuse and dependence were coded 1 for problemdrinking, and 0 otherwise. Additional sensitivityanalyses indicated that the overall pattern of re-sults was uninfluenced by using a more or less
253Work-Family Fit
TABLE 1. DESCRIPTIVE STATISTICS FOR ALL ANALYSIS VARIABLES
M SD Range
Outcome variablesDepressionAnxiety disorderProblem drinking
0.1390.0890.038
0.3460.2860.276
0–10–10–1
Independent variablesWork to family conflictWork to family facilitation
Facilitation—conflict
2.6552.610
20.047
0.7280.8381.134
1–51–5
24–4Family to work conflictFamily to work facilitation
Facilitation—conflict
2.1223.4211.299
0.6660.8321.096
1–51–5
23.33–4
Control variablesAge (years)Gender (female 5 1)Race/ethnicity (non-White 5 1)
40.8590.5170.109
9.8260.4990.312
25–620–10–1
EducationLess than a high school degreeHigh school degree or GEDAssociates degree or some collegeCollege graduate
0.0800.3590.2790.281
0.2720.4800.4490.449
0–10–10–10–1
Annual household earnings (1000s of dol-lars) 47,900 35,371 0–300,000
Marital statusMarried or cohabitingDivorced or separatedNever married
0.7590.1360.106
0.4280.3430.307
0–10–10–1
Parent to a child under the age of 6 0.217 0.412 0–1Self-employed 0.185 0.389 0–1
Note: Weighted estimates using data from the National Survey of Midlife Development in the United States (MIDUS; N5 1,986).
conservative cut-point for categorizing problemdrinking.
Independent Variables
Work to family conflict was assessed with fouritems: ‘‘How often have you experienced each ofthe following in the past year? (a) Your job re-duces the effort you can give to activities at home.(b) Stress at work makes you irritable at home.(c) Your job makes you feel too tired to do thethings that need attention at home. (d) Job worriesor problems distract you when you are at home.’’Response categories for each of these items andeach of the subsequently described work-familyspillover indices ranged from 1 (never) to 5 (allof the time [Chronbach’s a 5 0.82]).
Work to family facilitation was assessed withresponses to three questions: ‘‘How often haveyou experienced each of the following in the pastyear? (a) The things you do at work help you dealwith personal and practical issues at home. (b)The things you do at work make you a more in-
teresting person at home. (c) The skills you useon your job are useful for things you have to doat home’’ (Chronbach’s a 5 0.73).
Family to work conflict was measured by re-sponses to four questions: ‘‘How often have youexperienced each of the following in the pastyear? (a) Responsibilities at home reduce the ef-fort you can devote to your job. (b) Personal orfamily worries and problems distract you whenyou are at work. (c) Activities and chores at homeprevent you from getting the amount of sleep youneed to do your job well. (d) Stress at home makesyou irritable at work’’ (Chronbach’s a 5 0.80).
Family to work facilitation was measured byrespondents’ answers to three questions: ‘‘Howoften have you experienced each of the followingin the past year? (a) Talking with someone athome helps you deal with problems at work. (b)The love and respect you get at home makes youfeel confident about yourself at work. (c) Yourhome life helps you relax and feel ready for thenext day’s work’’ (Chronbach’s a 5 0.70).
254 Journal of Marriage and Family
Analytic Sequence
A series of multivariate logistic regression modelswere specified to examine the effects of differentcombinations of work-family conflict and work-family facilitation on mental disorders and prob-lem drinking. In the first model each of the out-comes was regressed on work to family andfamily to work conflict along with the control var-iables. In the second model, work to family andfamily to work facilitation were added to the mod-el. In the third model, work to family conflict 3work to family facilitation and family to workconflict 3 family to work facilitation interactionterms were included. In this specification, each ofthe work-family spillover measures were centeredon their respective sample mean, and the interac-tion terms were constructed from these centeredmeasures to avoid multicollinearity (Aiken &West, 1991; Jaccard, Turrisi, & Wan, 1990). In thelast model, the outcome variables were regressedon two difference-score variables reflecting thedifference between work to family conflict and fa-cilitation and the difference between family towork conflict and facilitation.
Likelihood difference statistics were calculatedto assess the relative fit of nested models. Themodel specified using the difference scores be-tween various dimensions of work-family conflictand work-family facilitation was not nested withinor an extension of the other models; consequently,likelihood statistics did not provide an appropriatecomparison. McFadden’s rho (McFadden, 1974)was calculated to compare the relative levels ofexplained variance between nonnested models.
RESULTS
Univariate estimates and mean comparisons sug-gest that work to family exchanges were moreproblematic than those from family to work. Asindicated by the difference scores between thework to family and family to work conflict andfacilitation, the average respondent reported thatfamily facilitated work-related obligations morethan it conflicted with work, whereas the work tofamily difference score was essentially 0. Indeed,mean levels of work to family conflict were sta-tistically equivalent to mean levels of work tofamily facilitation (M 5 2.66 and 2.61, respec-tively, t 5 0.80, df 5 1976, p # 0.425). Addi-tionally, mean levels of work to family conflictwere higher than mean levels of family to workconflict (M 5 2.66 vs. 2.12, p # .001), and mean
levels of work to family facilitation were lowerthan mean levels of family to work facilitation (M5 2.61 vs. 3.42, p # .001). Thus first glancessuggest that family benefits work more than workbenefits family, and that family interferes less withwork than work does with family. (Note: The bi-variate associations among the four work-familyexperiences suggest that they are distinct, and thatincluding them in the same multivariate modelshould not undermine the integrity of the param-eter estimates. See Appendix A for a completezero-order correlation matrix.)
Evidence in Table 1 also indicates that theprevalence of psychiatric disorders in this middle-aged (M 5 41, SD 5 9.8), married (75.8%) sam-ple of working adults was modest yet substantial.More than one in seven respondents were esti-mated to meet the DSM-III-R criteria for a majordepressive event during the past 12 months, andone in 11 respondents were estimated to meet di-agnostic criteria for general anxiety disorder orpanic disorder. By specification, approximately8.3% of these working adults were characterizedas problem drinkers.
Supporting the study’s general hypothesis, ini-tial unreported models indicated that higher levelsof work to family conflict and family to work con-flict were independently associated with greaterodds of each outcome. Results reported in Table2, however, illustrate that work-family fit is morethan the absence of conflict. Model 1, for exam-ple, indicates that more family to work facilitationis associated with a lower risk of depression andproblem drinking. Specifically, each unit increasein family to work facilitation is associated with a15% decrease in the odds of reporting depressionand a 38% decrease in the odds of reporting prob-lem drinking. Results reported for the secondmodel for each of the outcomes in Table 2 dem-onstrate no evidence that work-family facilitationbuffers the deleterious effects of work-family con-flict in terms of depression or problem drinking.Finally, results from Model 3 indicate that theodds of meeting diagnostic criteria for depressionis lowered as facilitation exceeds conflict (bothwork to family and family to work), and that thelikelihood of being classified as a problem drinkeris lowered as family to work facilitation surpassesfamily to work conflict.
Model comparisons suggest that the indepen-dent effects model (i.e., Model 1) provides thebest combination of conflict and facilitation forexplaining depression and problem drinking.Model 1 had less unexplained variance than the
255Work-Family Fit
unreported conflict–only model for both depres-sion (LR 5 1444.31 vs. 1451.51 5 7.199 with 2df, p # .05) and problem drinking (LR 5 986.61vs. 1006.53l 5 19.92 with 2 df, p # .001), andModel 1 was clearly more parsimonious than theinteractive models for both outcomes. Althoughthe difference models were more parsimoniousthan the others, they also yielded lower levels ofadditional explained variance above the base mod-el for both depression (r 5 .031) and problemdrinking (r 5 .041) than the independent effects(depression, r 5 .041; problem drinking, r 5.051) or the interaction (depression, r 5 .041;problem drinking, r 5 .054) models.
The pattern of results for anxiety disorder wasnotably different from those reported for depres-sion and problem drinking. Although both typesof work-family conflict were independently asso-ciated with greater odds of anxiety disorder, work-family facilitation was unrelated to this outcome.However, Model 2 indicates that both types of fa-cilitation moderate the effects of their work-familyconflict counterparts, suggesting that work-familyfacilitation buffers the negative consequences ofwork-family conflict. More specifically, whenwork to family facilitation is high, a high level ofwork to family conflict is associated with an 11%increase in the odds of anxiety disorder. By con-trast, when work to family facilitation is low, ahigh level of work to family conflict is associatedwith more than twice the odds of anxiety disorder.Likewise, the reduced odds of anxiety disorder as-sociated with lower work to family conflict areaccentuated when work to family facilitation ishigh. Under conditions of low work to family con-flict, the estimated odds of anxiety disorder is 0.89when work to family facilitation is low but 0.44when facilitation is high. (The conditioning effectof family to work conflict by family to work fa-cilitation was comparable.) Finally, results fromModel 3 suggest that when the level of work-fam-ily facilitation surpasses the level of work-familyconflict, an individual’s odds of meeting casenessfor anxiety disorder decreases.
Comparisons of the anxiety disorder models in-dicate that the interactive model outlines the bestcombination of conflict and facilitation for explain-ing anxiety disorder. The addition of the interactionterms to Model 1 significantly decreases the amountof unexplained variance (LR 1019.85 vs. 1035.87 516.01 with 2 df, p # .001). Estimates of McFadden’srho (1974) also indicated that the difference model(r 5 .028) provides the least amount of additionalexplained variance over the base model in contrast
to the independent effects (r 5 .059) or the inter-action (r 5 .075) models.
DISCUSSION
The overall goal of this study was to specify whichcombination of work-family experiences is optimalfor predicting individual outcomes reflective offamily competence. The most positive outcomes re-sulted almost exclusively from low levels of work-family conflict and high levels of work-family fa-cilitation. However, in some cases this combinationwas manifest in terms of independent, opposite ef-fects (e.g., depression and problem drinking),whereas for anxiety disorders the most health-pro-moting combination manifests itself as an interac-tion effect. This pattern of effects suggests thatwork-family facilitation, as a family capability,may help families adjust to work-family conflict byoffsetting or redefining the meaning of the incom-ing stressor, thereby eliminating its threat.
Unfortunately, why the elimination strategymanifests itself differently for the depression andproblem drinking outcomes compared with theanxiety disorder outcome is not well informed byfamily resilience theory. Perhaps the divergentpattern of results is reflective of the fundamentaldistinction between these disorders made by theDiagnosis and Statistical Manual of Mental Dis-order and the likelihood that they are influencedby different environmental factors (Kendler, Nea-le, Kessler, Heath, & Eaves, 1992). One study ofheart transplant patients, for example, indicatedthat low levels of family support were predictiveof anxiety disorder but not depression, suggestingthat anxiety disorder is more sensitive than de-pression to the absence of family resources (i.e.,support) in the face of family stress (i.e., hearttransplant; Dew et al., 2001). It is also possiblethat families of individuals with depression differfrom families of individuals with anxiety disor-ders in ways separate from work-family processes,and that it is these unspecified differences thatcontribute to the differing pattern of results.
Before proceeding with a discussion of theseresults, it is important to recognize the limitationsof this study. First, the prevalence of each out-come was high in this study in contrast to otherreports using similar measures, such as the Na-tional Comorbidity Study (NCS; Kessler et al.,1994). This difference could reflect the fact thatyoung women were overrepresented in the MI-DUS despite the oversampling of older adults andmen, or that the NCS did not include older men
256 Journal of Marriage and Family
TA
BL
E2.
MU
LT
IVA
RIA
TE
LO
GIS
TIC
RE
GR
ES
SIO
NE
ST
IMA
TE
SO
FT
HE
AS
SO
CIA
TIO
NS
BE
TW
EE
NW
OR
K-F
AM
ILY
EX
PE
RIE
NC
ES
AN
DM
EN
TA
LIL
LN
ES
SE
S
Dep
ress
ion
Mod
el1
Mod
el2
Mod
el3
Prob
lem
Dri
nkin
g
Mod
el1
Mod
el2
Mod
el3
Anx
iety
Dis
orde
r
Mod
el1
Mod
el2
Mod
el3
Wor
kto
fam
ilyco
nflic
t
Wor
kto
fam
ilyfa
cilit
atio
n
Fam
ilyto
wor
kco
nflic
t
.31*
*(.
11)
2.0
4(.
09)
.50*
**(.
12)
.30*
*(.
11)
2.0
4(.
09)
.49*
**(.
12)
.31*
(.14
).2
2(.
12)
.42*
*(.
15)
.32*
(.14
).2
1(.
12)
.42*
*(.
16)
.48*
**(.
13)
2.1
3(.
11)
.63*
**(.
15)
.45*
**(.
14)
2.0
6(.
12)
.72*
**(.
15)
Fam
ilyto
wor
kfa
cilit
atio
n
Wor
kto
fam
ilyba
lanc
ea
Fam
ilyto
wor
kba
lanc
ea
2.2
1**
(.09
)2
.18*
(.09
)2
.15*
(.07
)2
.35†
(.07
)
2.4
8***
(.11
)2
.49*
**(.
11)
2.0
1(.
09)
2.5
0***
(.08
)
.09
(.11
).2
0(.
12)
2.2
7***
(.08
)2
.24*
*(.
08)
Wor
kto
fam
ilyin
tera
ctio
nb
Fam
ilyto
wor
kin
tera
ctio
nb
2.0
4(.
10)
2.1
5(.
11)
.05
(.13
).0
4(.
13)
2.3
5**
(.12
)2
.30*
(.14
)A
ge
Gen
der
(fem
ale
51)
Rac
e(n
on-W
hite
51)
2.0
2**
(.01
).4
8***
(.14
)2
.21
(.28
)
2.0
2**
(.01
).4
8***
(.14
)2
.19
(.28
)
2.0
2**
(.01
).5
1***
(.14
)2
.22
(.28
)
2.0
5***
(.01
)2
.75*
**(.
19)
2.8
5(.
48)
2.0
5***
(.01
)2
.75*
**(.
19)
2.8
5(.
48)
2.0
5***
(.01
)2
.73†
(.18
)2
.85
(.48
)
2.0
1(.
01)
.82*
**(.
18)
21.
51**
(.54
)
2.0
1(.
01)
.81*
**(.
18)
21.
51**
(.54
)
2.0
1(.
01)
.86*
**(.
18)
21.
40**
(.53
)E
duca
tion
Les
sth
anhi
ghsc
hool
Hig
hSc
hool
orG
ED
Som
eC
olle
ge
Ref
2.6
7*(.
29)
2.5
0(.
28)
Ref
2.6
9*(.
29)
2.5
2(.
28)
Ref
2.6
9*(.
29)
2.4
8(.
28)
Ref
.04
(.42
)2
.16
(.42
)
Ref
.05
(.42
)2
.15
(.42
)
Ref
2.0
1(.
42)
2.1
6(.
42)
Ref
2.4
9(.
34)
2.4
3(.
33)
Ref
2.5
4(.
34)
2.4
7(.
33)
Ref
2.5
3(.
33)
2.4
4(.
32)
Col
lege
grad
uate
Hou
seho
ldea
rnin
gsc
Mar
ital
stat
usM
arri
edor
coha
bita
ting
2.4
9(.
29)
2.0
3(.
02)
Ref
d
2.5
2(.
29)
2.0
3(.
02)
Ref
2.4
6(.
29)
2.0
2(.
02)
Ref
2.1
1(.
42)
2.0
3(.
03)
Ref
2.1
0(.
42)
2.2
(.03
)
Ref
2.0
8(.
42)
2.0
2(.
03)
Ref
21.
01**
(.36
).0
2(.
03)
Ref
21.
09**
(.36
).0
2(.
03)
Ref
2.9
8**
(.35
).0
3(.
02)
Ref
Div
orce
dor
sepa
rate
d
Nev
erm
arri
edor
wid
owed
.78*
**(.
18)
.41
(.22
)
.79*
**(.
18)
.44*
(.22
)
.76*
**(.
18)
.34
(.21
)
.43
(.24
).6
2*(.
26)
.24
(.24
).6
1*(.
27)
.43
(.24
).5
6*(.
26)
.64*
*(.
22)
.33
(.27
)
.67*
*(.
22)
.42
(.28
)
.58*
*(.
21)
.21
(.27
)
257Work-Family Fit
TA
BL
E2.
CO
NT
INU
ED
Dep
ress
ion
Mod
el1
Mod
el2
Mod
el3
Prob
lem
Dri
nkin
g
Mod
el1
Mod
el2
Mod
el3
Anx
iety
Dis
orde
r
Mod
el1
Mod
el2
Mod
el3
Pare
ntto
child
,6
year
s2
.21
(.20
)2
.23
(.20
)2
.17
(.20
).0
5(.
23)
.06
(.23
).0
3(.
22)
2.3
2(.
26)
2.3
6(.
26)
2.2
3(.
25)
Self
-em
ploy
ed
22
Log
likel
ihoo
d
2.3
1(.
20)
1444
.31
2.3
1(.
20)
1441
.98
2.2
5(.
19)
1456
.91
2.2
3(.
24)
986.
61
2.2
3(.
24)
986.
29
2.1
3(.
23)
995.
62
.27
(.21
)10
35.8
7
.27
(.21
)10
19.8
5
.35
(.21
)10
67.4
9
Not
e:E
stim
ates
obta
ined
usin
gda
tafr
omth
eN
atio
nal
Surv
eyof
Mid
life
Dev
elop
men
tin
the
Uni
ted
Stat
es(M
IDU
S;N
51,
986)
.a B
alan
ceis
aco
mbi
ned
mea
sure
whe
reby
wor
kto
fam
ilyco
nflic
tw
assu
btra
cted
from
wor
kto
fam
ilyfa
cilit
atio
n,an
dfa
mily
tow
ork
confl
ict
was
subt
ract
edfr
omfa
mily
tow
ork
faci
litat
ion.
b Int
erac
tion
refle
cts
the
prod
uct
term
ofw
ork
tofa
mily
confl
ict
with
wor
kto
fam
ilyfa
cilit
atio
n,an
dfa
mily
tow
ork
confl
ict
with
fam
ilyto
wor
kfa
cilit
atio
n.A
llm
easu
res
wer
ece
nter
edon
thei
rre
spec
tive
sam
ple
mea
npr
ior
toca
lcul
atin
gth
ein
tera
ctio
nte
rms.
c Rep
orte
dea
rnin
gsw
ere
divi
ded
by10
,000
toen
hanc
ein
terp
reta
bilit
yof
the
para
met
eres
timat
es.
d Ref
5R
efer
ence
cate
gory
.*p
#.0
5.**
p#
.01.
***p
,.0
01(t
wo-
taile
d).
(who were also oversampled in the MIDUS). Inboth cases these oversampling strategies couldyield overinflated rates of psychiatric morbidity(E. Wethington, personal communication, Novem-ber 9, 2001). However, it is also possible that theshort form measures employed in the MIDUS per-formed less than optimally. The current analyseswere also conducted using cross-sectional data, socausal inferences cannot be made. This is partic-ularly important given the strong possibility thatmental health could be the primary cause of work-family difficulties (conflict and absence of facili-tation), rather than the outcome, as well as thepossibility that families of individuals with mentalillnesses may differ in other important ways sep-arate from work-family experiences. The biascaused by reverse causality as well as the endo-geneity between mental health and other familyprocesses would likely overinflate parameter es-timates of association between work-family ex-periences and mental illness. Additionally, the re-sults of this study could be a consequence ofcommon method bias, as both the outcomes andthe independent variables were constructed fromself-reported data. Finally, it is possible that thework-family dynamics involved in mental illnessmay be different from the dynamics relevant toother outcomes reflective of family competencesuch as dyadic adjustment or parental involve-ment. Given these limitations as well as the pau-city of previous research using measures of work-family compatibility or facilitation, additionalprospective research is clearly needed.
The results from this study are consistent withthe growing body of studies examining how mul-tiple dimensions of the work-family interface affectindividual workers, and they suggest that work-family fit is more than the absence of conflict. Likeprevious cross-sectional and longitudinal work byFrone and colleagues (1993, 1997, 2000), the re-sults of this study suggest that higher levels of bothwork to family and family to work conflict are as-sociated with poor mental health. Newer to the lit-erature, however, were the repeated protective ef-fects of work-family facilitation, particularly familyto work facilitation. These results parallel a recentstudy indicating that more family to work facilita-tion was associated with better psychological well-being, independent of the effects of work-familyconflict (Grzywacz, 2000). The current study ex-tends this earlier work by linking positive aspectsof combining work and family with more objectiveassessments of mental health.
Echoing a recent observation by Frone (2003),
258 Journal of Marriage and Family
the overall pattern of results from this study sug-gests an asymmetrical preference that consistentlyfavors the work side of the work and family equa-tion. Although not explicitly tested, the magnitudeof the effects suggests that family interfering withwork is more detrimental to mental health thanwas work interfering with family. Moreover, al-though family to work facilitation was consistent-ly associated with lower risk of mental illness,work to family facilitation exerted inconsistent ef-fects on mental health. Thus, at least in terms ofmental health, it appears as though work-familyfit is optimized when work is protected from fam-ily disruptions and when family contributes toproductivity at work.
Although this study’s general pattern of resultsare consistent with earlier work and empirical re-views, several conceptual limitations are apparentand need to be addressed in future research. First,both work-family conflict and work-family facil-itation are best described as bidirectional and mul-tidimensional (Carlson, Kacmar, & Williams,2000; Frone et al., 1997; Greenhaus & Beutell,1985; Grzywacz & Marks, 2000; Small & Riley,1990). Thus the large number of possible combi-nations of work-family conflict and work-familyfacilitation make precise specification of work-family fit difficult, particularly in the absence ofdeveloped theory regarding work-family facilita-tion. Like previous studies (e.g., Barnett & Ba-ruch, 1985; Tiedje et al., 1990), we arbitrarilychose combinations of conflict and compatibilitythat had similar directions (e.g., work to familyconflict with work to family facilitation); however,it is possible that other complex combinationsmore adequately characterize the fit between workand family. For example, if family to work facil-itation represents a family’s responsiveness towork-related demands, then it may be more im-portant than work to family facilitation for defus-ing the meaning of work to family conflict.
Additionally, it is important to acknowledgethat the different types of work-family conflict(i.e., time-, strain-, and behavior-based; Greenhaus& Beutell, 1985) and work-family facilitationhave unique meanings and may influence individ-uals and families in different ways. The resiliencyframework suggests that when family capabilitiesare particularly suited for the imposing demand,more optimal adaptation will occur in contrast towhen capability and demands are matched lesswell. For example, work-family facilitationbrought about by status gains may be importantin the face of time-based conflict because they
may enable the individual to secure high-quality,flexible childcare. By contrast, this type of facil-itation may be unimportant when a person’s phys-ical or psychological resources are drained (i.e.,strain-based conflict). The essential conceptual is-sue that these examples raise is the possibility thatwork-family fit may reflect a generalized combi-nation of conflict and facilitation, but this gener-alized sense may arise from the accumulation ofmore specific matches between distinct types ofconflict and facilitation.
A third area of conceptual development for fu-ture work-family research would explore the dif-ferentiation of work-family experiences that are ep-isodic compared with those that are enduring(Greenhaus & Parasuraman, 1999). For example,even in the best possible work-family arrangement,specific events of work-family conflict can occur,such as a last-minute deadline at work or the strainof caring for an ill family member. Moreover, theseepisodes of conflict or facilitation may be gendered.These examples highlight the possibility that theoverall quality of work-family fit may differ fromspecific events of conflict or facilitation and be ex-perienced differently by members of differentgroups. Additionally, the distinction between en-during qualities of the work-family interface andspecific work-family exchanges highlight the pos-sibility that enduring aspects of fit may buffer orexacerbate specific events when they occur. Con-sistent with this possibility, results from a recentreport suggested that higher levels of family towork facilitation buffered the negative effect of afamily to work stressor on binge drinking (Grzy-wacz, Johnson, Hartwig, & Almeida, 2002). Ad-ditional research examining these nuances of work-family exchanges is clearly needed to betterunderstand the work-family interface and to gen-erate family policy and family life education thatmeets the needs of today’s adults.
The general pattern of results from this studysuggests that work-family facilitation contributesto ‘‘fit’’ by eliminating or offsetting the negativepotential of work-family conflict. If replicated inother research, these results suggest a variety ofdifferent interventions for enhancing work-familyfit. First, individual-level programs could be de-veloped to help employees better understand thepersonal benefits they and their families realizefrom combining work and family. These gains be-come a cognitive resource that can be invokedduring times of conflict to minimize the impact ofthose episodes. Programs could also target familymembers in order to alleviate or eliminate the
259Work-Family Fit
stress that accompanies another family member’sjob. For example, family life education programshave been utilized to decrease the negative impactof deployments on military families by providingsupport- and information-based networks. Finally,organizational interventions could be designed tocreate additional benefits of combining work andfamily that, once realized, could promote enduringstructural changes that facilitate the ability for in-dividuals to balance their work and family roles.
Limitations notwithstanding, the results of thisstudy clearly suggest that work-family fit is morethan the absence of work-family conflict. As theworkforce and the labor market continue address-ing the challenge of the work-family interface, in-creased attention needs to be given to work-familyfacilitation, how it can be cultivated and exploited,and how it operates in conjunction with work-family conflict in shaping desirable individual,family, and work-related outcomes. Such knowl-edge is an essential first step for developing policyand programs that meet the needs of current andfuture workers and their families.
NOTE
The research was supported by a grant from the Na-tional Institute on Alcohol Abuse and Alcoholism (R03-AA-12744-01A1).
REFERENCES
Aiken, L. S., & West, S. G. (1991). Multiple regression:Testing and interpreting interactions. Newbury Park,CA: Sage.
Allen, K. R., & Pickett, R. S. (1987). Forgotten streamsin the family life course: Utilization of qualitative ret-rospective interviews in the analysis of lifelong singlewomen’s family careers. Journal of Marriage and theFamily, 49, 517–526.
Allen, T. D., Herst, D. E., Bruck, C. S., & Sutton, M.(2000). Consequences associated with work-to-familyconflict: A review and agenda for future research. Jour-nal of Occupational Health Psychology, 5, 278–308.
Barnett, R. C. (1998). Toward a review and reconcep-tualization of the work/family literature. Genetic, So-cial, and General Psychology Monographs, 124,125–182.
Barnett, R. C. (1999). A new work-life model for thetwenty-first century. Annals of the American Acade-my of Political and Social Science, 562, 143–158.
Barnett, R. C., & Baruch, G. K. (1985). Women’s in-volvement in multiple roles and psychological dis-tress. Journal of Personality and Social Psychology,49, 135–145.
Barnett, R. C., & Hyde, J. S. (2001). Women, men,work, and family. An expansionist theory. AmericanPsychologist, 56, 781–796.
Carlson, D. S., Kacmar, M. K., & Williams, L. J. (2000).Construction and validation of a multidimensional
measure of work-family conflict. Journal of Voca-tional Behavior, 56, 249–276.
Cohen, S., & Wills, T. A. (1985). Stress, social support,and the buffering hypothesis. Psychological Bulletin,98, 310–357.
Crouter, A. C. (1984). Spillover from family to work:The neglected side of the work-family interface. Hu-man Relations, 37, 425–442.
Dew, M. A., Kormos, R. L., DiMartini, A. F., Switzer,G. E., Schulberg, H. C., Roth, L. H., et al. (2001).Prevalence and risk of depression and anxiety-relateddisorders during the first three years after heart trans-plant. Psychosomatics, 42, 300–313.
Ernst Kossek, E., & Ozeki, C. (1998). Work-family con-flict, policies, and the job-life satisfaction relation-ship: A review and directions for organizational be-havior–human resources research. Journal of AppliedPsychology, 83, 139–149.
Frone, M. R. (2000). Work-family conflict and employ-ee psychiatric disorders: The National ComorbiditySurvey. Journal of Applied Psychology, 85, 888–895.
Frone, M. R. (2003). Work-family balance. In J. C.Quick & L. E. Tetrick (Eds.), Handbook of Occupa-tional Health Psychology (pp. 143–162). Washington,DC: American Psychological Association.
Frone, M. R., Russell, M., & Cooper, M. L. (1993).Relationship of work-family conflict, gender, and al-cohol expectancies to alcohol use/abuse. Journal ofOrganizational Behavior, 14, 545–558.
Frone, M. R., Russell, M., & Cooper, M. L. (1997).Relation of work-family conflict to health outcomes:A four-year longitudinal study of employed parents.Journal of Occupational and Organizational Psy-chology, 70, 325–335.
Frone, M. R., Yardley, J. K., & Markel, K. S. (1997).Developing and testing an integrative model of thework-family interface. Journal of Vocational Behav-ior, 50, 145–167.
Galinsky, E., Bond, J. T., & Friedman, D. F. (1993). Thechanging workforce: Results of the national study.New York: Families and Work Institute.
Grant, B. F., Harford, T. C., Dawson, D. A., Chou, P.,Dufour, M., & Pickering, R. (1994). Prevalence ofDSM-IV alcohol abuse and dependence. AlcoholHealth and Research World, 18, 243–248.
Greenhaus, J. H., & Beutell, N. J. (1985). Sources ofconflict between work and family roles. Academy ofManagement Review, 10, 76–88.
Greenhaus, J. H., & Parasuraman, S. (1999). Researchon work, family, and gender: Current status and fu-ture directions. In G. N. Powell (Ed.), Handbook ofgender and work (pp. 391–412). Thousand Oaks,CA: Sage.
Grzywacz, J. G. (2000). Work-family spillover andhealth during midlife: Is managing conflict every-thing? American Journal of Health Promotion, 14,236–243.
Grzywacz, J. G. (2002, February). Toward a theory ofwork-family facilitation. Paper presented at the Busi-ness and Professional Women’s Foundation, The Cen-ter for Working Families at Purdue University and theAlfred P. Sloan Foundation’s Persons, Processes, andPlaces: Research on Families, Workplaces and Com-munities conference, San Francisco, CA.
Grzywacz, J. G., Almeida, D. M., & McDonald, D. A.(2002). Work-family spillover and daily reports of
260 Journal of Marriage and Family
APPENDIXZERO-ORDER CORRELATION MATRIX OF ALL ANALYSIS VARIABLES
1 2 3 4 5 6 7 8
1. Depression2. Anxiety disorder3. Problem drinking4. Work to family conflict5. Work to family facilitation
—2.36**
.05*
.14**2.01
—.05*.14**
2.01
—.10**.03
—2.04* —
6. Family to work conflict7. Family to work facilitation8. Age9. Gender (female 5 1)
10. Race (non-White 5 1)
.16**2.09**2.07**
.12**2.01
.16**
.01
.01
.12**
.07**
.13**2.11**2.14**2.09**2.04
.50**2.05*2.10**2.032.07**
.13**
.26**
.06**
.022.02
—2.06**2.18**
.022.03
—.02
2.02.05*
—.02
2.0211. Less than high school12. High school or GED13. Some college14. College graduate15. Household earnings
.07**2.05*
.032.022.09**
.06**2.01
.05*2.08**2.07**
2.032.01
.01
.01
.01
2.032.09**
.03
.08**
.10**
.012.09**
.01
.09**
.06**
2.012.07**
.05*
.03
.03
.032.012.02
.01
.05*
.09**
.07**2.11**2.03
.09**16. Married or cohabitating17. Divorced or separated18. Never married19. Parent to child ,6 years20. Self-employed
2.15**.12**.07**
2.022.05*
2.06**.08**.01
2.04.02
2.5*.01.07**.07**
2.03
.022.03
.01
.03
.02
.02
.012.032.05*2.14**
.02
.012.04
.16**
.02
.17**2.12**2.11**
.01
.03
.01
.11**2.13**2.37**2.07**
Note: Weighted estimates from the National Survey of Midlife Development in the United States (MIDUS; n 5 1,986).*p # .05. **p # .01 (two-tailed).
work and family stress in the adult labor-force. Fam-ily Relations, 51, 28–36.
Grzywacz, J. G., Johnson, D., Hartwig, D., & Almeida,D. M. (2002, February). Work-family linkages andproblem drinking among adults: Evidence from alarge daily diary study. Paper presented at the Busi-ness and Professional Women’s Foundation, The Cen-ter for Working Families at Purdue University and theAlfred P. Sloan Foundation’s Persons, Processes, andPlaces: Research on Families, Workplaces and Com-munities conference, San Francisco, CA.
Grzywacz, J. G., & Marks, N. F. (2000). Reconceptu-alizing the work-family interface: An ecological per-spective on the correlates of positive and negativespillover between work and family. Journal of Oc-cupational Health Psychology, 5, 111–126.
Hattery, A. (2001). Women, work, and family: Balanc-ing and weaving. Thousand Oaks, CA: Sage.
Heymann, J. (2000). The widening gap: Why America’sworking families are in jeopardy, and what can bedone about it. New York: Basic Books.
Hochschild, A. R. (1997). The time bind: When workbecomes home and home becomes work. New York:Metropolitan Books.
Jaccard, J., Turrisi, R., & Wan, C. K. (1990). Interactioneffects in multiple regression. Sage University PapersSeries on Quantitative Applications in the Social Sci-ences, 72. Newbury Park, CA: Sage.
Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., &Rosenthal, R. A. (1964). Organizational stress: Studiesin role conflict and ambiguity. New York: Wiley.
Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A.C., & Eaves, L. J. (1992). Major depression and gen-eralized anxiety disorder: Same genes, (partly) dif-ferent environments? Archives of General Psychiatry,49, 716–722.
Kessler, R. C., Andrews, G., Mroczek, D., Ustun, B., &Wittchen, H.-U. (1998). The World Health Organi-zation Composite International Diagnostic InterviewShort Form (CIDI-SF). International Journal ofMethods in Psychiatric Research, 7, 171–185.
Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C.B., Hughes, M., Eshleman, S., Wittchen, H.-U., &Kendler, K. S. (1994). Lifetime and 12-month prev-alence of DSM-III-R psychiatric disorders in theUnited States: Results from the National ComorbiditySurvey. Archives of General Psychiatry, 51, 8–19.
Kirchmeyer, C. (1992a). Nonwork participation and workattitudes: A test of scarcity vs. expansion models ofpersonal resources. Human Relations, 45, 775–795.
Kirchmeyer, C. (1992b). Perceptions of nonwork-to-work spillover: Challenging the common view ofconflict-ridden domain relationships. Basic and Ap-plied Social Psychology, 13, 231–249.
Kohn, R., Dohrenwend, B. P., & Mirotznik, J. (1998).Epidemiological findings on selected psychiatric dis-orders in the general population. In B. P. Dohrenwend(Ed.), Adversity, stress, and psychopathology (pp.235–284). New York: Oxford University Press.
Lachman, M. E., & Boone-James, J. (1997). Chartingthe course of midlife development: An overview. InM. E. Lachman & J. Boone-James (Eds.), Multiplepaths of midlife development (pp. 1–20). Chicago:University of Chicago Press.
Larson, R. W., Richards, M. H., & Perry-Jenkins, M.(1994). Divergent worlds: The daily emotional ex-perience of mothers and fathers in the domestic andpublic spheres. Journal of Personality and SocialPsychology, 67, 1034–1046.
Marks, N. F. (1996). Caregiving across the lifespan: Na-tional prevalence and predictors. Family Relations,45, 1–10.
261Work-Family Fit
APPENDIXEXTENDED
9 10 11 12 13 14 15 16 17 18 19 20
—.06** —
2.02.04*.02
2.06**2.11**
.06**2.03
.022.032.12**
—2.22**2.18**2.18**2.13**
—2.47**2.47**2.12**
—2.39**2.06**
—.28** —
2.11**.10**.04
2.07**2.08**
2.14**.09**.10**
2.012.08**
.01
.012.012.05*
.03
.09**2.022.11**2.012.01
2.05*.04.02.04
2.03
2.05*2.022.10**2.01
.03
.33**2.22**2.22**2.01
.09**
—2.71**2.61**
.16**
.07**
—2.14**2.08**2.04
—2.14**2.05*
—2.02 —
McCubbin, H. I., McCubbin, M. A., Thompson, A. I.,& Thompson, E. A. (1998). Resiliency in ethnic fam-ilies: A conceptual model for predicting family ad-justment and adaptation. H. I. McCubbin, E. A.Thompson, A. I. Thompson, & J. E. Fromer (Eds.),Resiliency in Native American and immigrant fami-lies (Vol. 2, pp. 3–48). Thousand Oaks, CA: Sage.
McFadden, D. (1974). Conditional logit analysis ofqualitative choice behavior. In P. Zarembka (Ed.),Frontiers in econometrics (pp. 105–142). New York:Academic Press.
Milkie, M. A., & Peltola, P. (1999). Playing all the roles:Gender and the work-family balancing act. Journalof Marriage and the Family, 61, 476–490.
Patterson, J. M. (2002). Integrating family resilience andfamily stress theory. Journal of Marriage and Fam-ily, 64, 349–360.
Perry-Jenkins, M., Repetti, R. L., & Crouter, A. C.(2000). Work and family in the 1990s. Journal ofMarriage and the Family, 62, 981–998.
Sieber, S. D. (1974). Toward a theory of role accumu-lation. American Sociological Review, 39, 567–578.
Small, S. A., & Riley, D. (1990). Toward a multidi-mensional assessment of work spillover into familylife. Journal of Marriage and the Family, 52, 51–61.
Tiedje, L. B., Wortman, C. B., Downey, G., Emmons,C., Biernat, M., & Lang, E. (1990). Women with mul-tiple roles: Role-compatibility perceptions, satisfac-tion, and mental health. Journal of Marriage and theFamily, 52, 63–72.
U.S. Department of Labor. (1999). Futurework: Trendsand challenges for work in the 21st century. Wash-ington DC: U.S. Department of Labor.