Work,Family,and Well-Being at Midlife

Abstract

This study addressed work and family enhancement and conflict among adults at midlife (N = 125). The study included personal well-being measured 25 years earlier. Participants were classified into groups on the basis of latent class analysis. Qualitative analyses of narratives within groups enhanced understanding of the experiences of those in the groups. The first of the three groups identified by latent class analysis consisted of a fairly large number of women who seemed to have achieved balance in roles and enjoyed marital and job satisfaction. The second group, almost exclusively male, was characterized by a low degree of work and family enhancement. In the small third group, a profound lack of marital satisfaction coupled with high work and family conflict and low enhancement identified men and women with seemingly enduring low levels of well-being. Narratives largely validated the quantitative analysis but also provided additional avenues for further research.

Keywords

gender and family life course person-centered qualitative quantitative work and family

Early role theorist William Goode (1960) argued that individuals’ work and family roles are conflicting and, thus, difficult to combine. Indeed, research has shown that work–family conflict has a negative impact on life satisfaction and well-being (e.g., Hill, 2005). One solution to this presumed natural conflict is suggested by the Iowa school of symbolic interactionism, which proposes that a person can resolve such conflict by organizing life roles in a hierarchical manner (Stryker & Burke, 2000). A related solution is that men and women specialize in particular roles. Work, as a “greedy” institution (Coser, 1974), has tended to place demands on breadwinning men, whereas home and parenting have been prioritized for women (LaRossa, Simonds, & Reitzes, 2005). An alternative to these perspectives was set forth by Bowes (2005) and Marks (1977, 2009), who suggested that work and family domains are not necessarily opposing spheres but that researchers should consider how they may operate together in patterns that are adaptive, even enhancing, and how they are associated with well-being. The present study employs Marks’s and Bowes’s views to explore whether and how patterns can be identified about the ways in which midlife adults experience the intersection of work and family lives. This research includes the role of demographic variables, examines the contribution of early social advantage, and identifies the roles of earlier and contemporaneous well-being.

Literature

Work and Family

Work and family roles come into conflict when the demands of work spill over negatively into the family domain (work-to-family conflict) or when the demands of family spill over negatively into work (family-to-work conflict; Grzywacz & Bass, 2003; Grzywacz & Marks, 2000; Steiber, 2009). Work and family conflicts have been associated with (a) individuals’ lower psychological well-being and overall life satisfaction (Rantanen, Mauno, Kinnunen, & Rantanen, 2011), (b) lower satisfaction in marital and work domains (Carlson & Perrewd, 1999; Gareis, Barnett, Ertel, & Berkman, 2009), and (c) financial status and type of occupation (Bianchi, Casper, & King, 2005; Parasuraman & Simmers, 2001). When individuals experience conflict between work and family, marital satisfaction and job satisfaction are compromised and job stress is higher (Carlson, Kacmar, & Williams, 2000; Frone, Yardley, & Markel, 1997).

Family factors are associated differentially with work and family (Grzywacz & Bass, 2003; Grzywacz & Marks, 2000). For example, men and women have been found to allocate time and receive financial remuneration differently in family and paid work (Drago, 2000), reflected in unequal (Han & Moen, 1999), “distinctive life paths of men and women” (Moen, 2011, p. 82). Prioritizing family ties, often done by women (Powell & Greenhaus, 2010), can be problematic to well-being should family relations be strained, yet prioritizing work roles can also be risky in an historical epoch of unstable work (Marks, 1979, 2009; Rothbard, 2001).

Countering the emphasis on work and family conflict, emerging research suggests that the understudied domains of work and family enhancement, facilitation, or enrichment (Bianchi & Milkie, 2010) are also important to consider (Stevens, Minnotte, Mannon, & Kiger, 2007). Work roles that positively spill over into family life define work-to-family enhancement and family roles that spill over positively to work life define family-to-work enhancement (Byron, 2005). Directions of association of work and family with well-being and role satisfaction depend on the domain studied (work-to-family or family-to-work; Gareis et al., 2009; Grzywacz & Bass, 2003).

Marks (1977, 2009) argued that work and family domains can enhance or facilitate one another. Energy created in one role can spill over or be used in other roles (Steiber, 2009). This enhancement view, thus, rejects the assumption that a person cannot construct a satisfying life via equal commitments to work and to family life (Grzywacz & Marks, 2000; Marks & MacDermid, 1996). Instead, it suggests that there are a variety of routes through which individuals can experience well-being and role satisfaction (Marks, 1977, 1979).

According to theorizing by Marks (1977, 1979, 2009), individuals construct different systems of role commitments in their lives that are associated with satisfaction in different roles. The first theorized system is termed sacrilizers or balancers, and consists of people who experience high satisfaction in all aspects of their lives, devoting time and energy to roles in line with their commitment to those roles (Marks, 1979). This balanced system is presumed to be less common. Nevertheless, across time, some individuals, jobs, and contexts of life may allow people to experience their lives in a positive way. A second system contains nihilists or apathetics who get little satisfaction out of any of their roles. Third, Marks posited that the most common type of system would be characterized as invidious, in which people are undercommitted to some roles and overcommitted to others, thus locked in a hierarchical struggle, in which some roles are experienced positively but others are a source of frustration.

Well-Being

Empirically, both marital satisfaction and job satisfaction are linked to greater well-being. Greater well-being is also related to income (Hagerty, 2000); lack of negative affect (Siedlecki, Tucker-Drob, Oishi, & Salthouse, 2008); health, marital, and job satisfaction (Campbell, Converse, & Rodgers, 1976); and lower work–family conflict (Rantanen et al., 2011).

Well-being is considered a form of optimal human functioning (Linley, Joseph, Harrington, & Wood, 2006) and is defined in two ways, as subjective well-being (e.g., life satisfaction, positive affect; Diener, Shuh, Lucas, & Smith, 1999) and as psychological well-being (e.g., actualization, autonomy; Ryan & Deci, 2001; Ryff & Singer, 2006). This study uses the positive affect definition (Fredrickson, 2001). There are at least four conceptual perspectives concerning well-being. The first one (a trait view) sees well-being as an enduring, stable characteristic of persons (Lucas, Diener, & Suh, 1996; Pavot & Diener, 2008). The second perspective (an antecedents/consequences view) suggests that well-being is affected by context, such that economic and social contexts can be a challenge or an advantage to later life (DiPrete & Eirich, 2006; O’Rand, 2002). For example, in the job realm, across 1 and 6 years, higher job exhaustion at baseline was associated with later psychological distress (Rantanen, Kinnunen, Feldt, & Pulkkinen, 2008). The third perspective views well-being as a buffer against negative experiences (Cohn, Fredrickson, Brown, Mikels, & Conway, 2009; Fredrickson, 2001). Finally, the fourth one (called the broaden-and-build theory) argues that positive well-being may set the stage for later positive functioning. This latter theory (Cohn et al., 2009; Fredrickson, 2001) suggests that those with more positive emotions are better able to flourish and counteract the effects of negative emotions as they traverse the life course.

Research Questions and Hypotheses

Research Question 1: Different Midlife Patterns

Based on Marks’s (1997, 1979, 2009) conceptualization and the theory and research of Fredrickson (2001), the first research question for this study addresses whether groups of people with different systems of role commitment in their lives can be identified (i.e., a person-centered approach; Bergman & Magnusson, 1997). We hypothesize three to four groups in the analyses. There should be a group of people whose patterns project a balance of work and family roles and high satisfaction and well-being. Also expected is a group whose members reflect the opposite pattern. Underscoring the nonlinear nature of a person-centered approach, there should be one or two groups with role imbalance and concomitant patterns of satisfaction/dissatisfaction and well-being. Thus:

Research Question 1: Can groups of people with different systems of current role commitments, satisfactions, and demographic variables and past well-being be identified?

Hypothesis 1: There will be three to four identified groups that reflect Marks’s (1977, 1979) theorizing.

Subsidiary Issue 1: Demographic Considerations

There are many variables that have been considered in studies on work–family conflict and enhancement. Central constructs found in many studies (Amstad, Meier, Fasel, Elfering, & Semmer, 2011; Byron, 2005; Hill, 2005; Poelmans, O’Driscolli, & Beham, 2005) are work-to-family conflict, family-to-work conflict, work-to-family enhancement, family-to-work enhancement, job satisfaction, marital satisfaction, type of job/occupation, financial status, individual well-being, and gender (Byron, 2005; Han & Moen, 1999; Steiber, 2009). Several demographic variables have also been considered in work and family research: gender, children, financial status, occupation, education, age. Gender has been implicated in studies of work and family via different social role patterns (Boles, Wood, & Johnson, 2003). Some research suggests that men and women are equivalent on work-to-family conflict and family-to-work conflict but that the association of work involvement with work–family conflict is stronger for men than women (Byron, 2005).

The role of children is mixed. Research indicates that work–family conflict varies by gender across clusters of child, parent, and work demands (Cullen, Hammer, Neal, & Sinclair, 2009), but other research (Cohen, Geron, & Farchi, 2010) suggests that the effects of children are not particularly important, especially among midlife adults whose children may be out of the household (Bowes, 2005). A parallel argument may be made with respect to length of marriage, a variable often considered in family research (e.g., Grote, Frieze, & Stone, 1996). Financial status and occupation have a place in job and marital arrangements and satisfactions of midlife adults (Hagerty, 2000). A subsidiary research question is the following:

Research Question 1a: What roles do demographic variables of gender, financial status, occupation, length of marriage, and the presence of children play in identified groups?

Subsidiary Issue 2: Early Social Background

In an investigation of work and family and well-being, the role of earlier life experience in shaping subsequent life chances must be acknowledged. Social background assets such as parents’ financial and marital status may provide a cumulative advantage in terms of work life trajectories (DiPrete & Eirich, 2006) and well-being. Similarly, being from a small town or a large city is related to subsequent life changes and opportunities (O’Rand, 2002). In the same vein, religious orientation may be a resource on which adults can be seen to build their repertoire of positive responses to life (Fredrickson, 2001). The social background and resource advantages of financial status and marital status in the family of origin, geographic setting growing up, and emerging adult religiosity are potential resources for early and subsequent well-being that may set the stage for advantage or disadvantage carried into midlife work and family domains. Thus, a second subsidiary question is the following:

Research Question 1b: To what extent do earlier social background factors vary by identified groups?

Research Question 2: Contribution of Well-Being

Following Fredrickson’s (2001) theorizing and research, the second research question investigates the role of past and current well-being in current life patterns. In keeping with the broaden-and-build theory of Fredrickson (2001), we hypothesize that those adults with earlier higher well-being should be advantaged across the life course in attaining less work–family conflict and greater work–family enhancement as well as higher marital and job satisfaction. That is, patterns of midlife would involve a mix of higher emerging adult well-being along with midlife lower work–family conflict and greater work–family enhancement as well as a second, mirror opposite pattern illustrating difficulties accompanying lower earlier well-being. Well-being assessed in young adulthood would similarly be expected to vary across groups. Also, if well-being reflects current circumstances, then midlife marital satisfaction and job satisfaction are predicted to be associated with midlife well-being (similar to Rantanen et al., 2008). Enduring and developing characteristics of individuals have been theorized (Fredrickson, 2001; Marks, 1977, 1979, 2009) but not studied in a longitudinal framework of this length.

Research Question 2: What is the role of past and current well-being in midlife patterns?

Hypothesis 2a: Emerging adulthood higher well-being will be copresent in a group with less midlife work–family conflict and greater work–family enhancement as well as higher marital and job satisfaction; and lower earlier well-being will be copresent in a group with higher midlife conflict and lower enhancement, marital and job satisfaction.

Hypothesis 2b: Young adult well-being will vary by identified groups.

Hypothesis 2c: Midlife well-being will vary by identified groups.

Research Question 3: Narratives

Written narratives provided by the participants in midlife allow triangulation of the findings from the quantitative analyses with participants’ written views and thoughts. These narratives can augment the information from the quantitative data (Hsieh & Shannon, 2005) and illustrate similarities and differences in the latent classes. The information from written narratives can provide depth behind the examination of social status factors and the work and personal life concepts that portray the experiences of people inside the groups that are derived via the quantitative approach. Convergent data would suggest validation of the findings whereas divergent data could reveal incomplete interpretations. Thus:

Research Question 3: Can written narratives provided by midlife participants provide validation of the quantitatively based midlife patterns?

The Contribution of the Present Study

A review of the literature on work–family conflict on satisfaction and/or well-being made it evident that among the theoretical perspectives used to examine the topic, Marks’s (1977, 2009) enhancement framework has been employed relatively infrequently (Stevens et al., 2007). Contrary to other frameworks, Marks’s theorizing highlights the importance of patterns and nonlinear approaches to work/family issues. Thus, based on the ideas of Marks (1977, 1979, 2009) and Bowes (2005), the first contribution of this study is to identify types of midlife adults using a person-centered design. Such a design is an alternative to models of antecedents, consequences, and moderators of work–family conflict and enhancement (e.g., Hill, 2005; Poelmans et al., 2005; Steiber, 2009).

There are two main types of person-centered analyses: k-means clustering and mixture modeling (also known as latent class analysis or LCA). An advantage of LCA is the ability to handle categorical variables that are included in this study. Prior person-centered studies have had different aims and variables from those of the current study. For example, Cullen et al. (2009) reported three clusters reflecting variations among those adults who were simultaneously involved with child and parent care and work. Cohen et al. (2010) focused on couple quality (but not on work). Recently, Malinen et al. (2010) used mixture modeling to study dual-earner spouse and parent–child relationships. Han and Moen (1999) identified work trajectories and marital status changes of retired couples. However, none of these simultaneously assessed work and family conflict and enhancement, marital and job satisfaction, and earlier well-being. Thus, the present study contributes to the literature by analyzing these variables together using latent class analysis.

The person-centered approach invites a new view of the role of well-being. Several researchers (Poelmans et al., 2005; Steiber, 2009) noted that work–family conflict should have a negative impact on well-being. In accord with this view, antecedents-consequences models have studied well-being or life satisfaction as outcomes of work–family issues (e.g., Hill, 2005). However, the broaden-and-build life course approach of Fredrickson (2001) considers the contribution of well-being as a correlate to or predictor of midlife patterns. Thus, a second contribution of the study is to investigate the role of past and current well-being to current life patterns. This requires a prospective longitudinal design, in which other influences on historical or contemporaneous well-being are also accounted for. The current study uses data of well-being provided at three strategic times over a 25-year time span—emerging adulthood, young adulthood, and midlife—and is therefore well suited for the purposes of this study. As well, emerging adult social advantage is investigated as possible precursors to midlife patterns. Finally, because quantitative approaches may miss important issues, the present study further contributes to the literature by triangulating data through integration of findings from quantitative and qualitative methods (Hsieh & Shannon, 2005).

Method

Participants and Procedures

A study of adult transitions began in 1982 (cf., Fischer, Sollie, Sorell, & Green, 1989) with randomly selected participants from the 1980 and 1982 graduating cohorts of a large Texas university. Approximately 100 men and 100 women from each graduation cohort were recruited (N = 422; mean age = 23.24 years, SD = 1.91) with characteristics that closely resembled those of the students at that time (Fischer et al., 1989). The research was designed to follow each participant at roughly equal intervals from 1982 to 1984 (Waves 1 through 3). Follow-ups captured participants at later important times in the life course, young adulthood (Wave 4, 1990, N = 324) and midlife (Wave 5, 2007, N = 180). The institutional review board approved procedures at all waves.

Although there were 180 participants at Wave 5, this study focused on 125 individuals who at Wave 5 (a) were married or in committed marriage-like partnerships, (b) had paid employment, and (c) had previous scores on well-being at Wave 1 and Wave 4. Of the 125 who constituted the core sample for this study, at Wave 5, 58% were men (72 men, 53 women), 83% were Protestant/Christian, 98% were White, number of children averaged two, length of current marriage averaged 22.2 years (range 2-30). Eighty-nine percent were in professional or managerial positions and 84% reported that they were financially comfortable or moderately well off. Mean age at Wave 5 was 48 (SD = 2.16). In addition to those without Wave 4 well-being scores (n = 16), homemakers (n = 18), retired persons (n = 9), and singles (n = 12) were not included in the analyses because their circumstances were ambiguous with respect to assessments of work–family conflict/enhancement and many did not fill out the work/family scales. The resulting sample was homogenous with respect to all being in married or marriage-like relationships and all having current paid work outside of the home. At Wave 1, only 18% had both stable employment and marriage, supporting the view that well-being measured at this earlier wave predates these life experiences.

An issue in longitudinal research is attrition. Participants who returned questionnaires after 25 years may differ in meaningful ways from those who did not return questionnaires. As well, there were participants who returned questionnaires in 2007, but did not fit the criteria to be included in the analyses in this study. Therefore, two sets of comparisons were made based on Wave 1 data. First, using Wave 1 responses, comparisons were made between all 180 returnees versus nonreturnees (n = 242) on gender, 1980 or 1982 cohort, married or not, ethnicity, religious affiliation, age, well-being, have children or not, and financial status. With p set at .01, there were no statistically significant differences on any of the variables. Second, comparisons on the same Wave 1 variables were made between the 2007 Wave 5 returnees included in this research (n = 125) and those 55 returnees who did not meet the criteria to be included in the analyses. There were disproportionate numbers of women (58%) not included versus those who were included (43%), possibly reflecting the number of Wave 5 women homemakers who did not have outside paid employment. The second difference was in having had children: More of those included were childless at Wave 1 (94%) than those not included (86%), possibly reflecting retirements of those who started families earlier. Overall, the two groups of returnees (180 vs. 242; 125 vs. 55) were very similar at Wave 1 on important characteristics.

Measures

Conflict Between Work and Family

Five items each on the work-to-family and the family-to-work conflict scales were averaged separately across a 7-point response scale from 1 = Strongly disagree to 7 = Strongly agree (Netemeyer, Boles, & McMurrian, 1996). Cronbach’s αs were .88 and .86. The scores on the family-to-work conflict scale were significantly skewed. A square root transformation was used to successfully reduce skew (from 2.72 to 0.91).

Enhancement Between Work and Family

There were four items averaged on the work-to-family enhancement index and three items on the family-to-work enhancement index (Grzywacz & Marks, 2000). The items were answered from 1 = All the time to 5 = Never and were reverse scored so that higher scores indicated greater enhancement. Cronbach’s αs were .44 and .58. Hill (2005) reported similarly low alpha levels for work-to-family enhancement (α = .55) and used only one item for family-to-work enhancement. The response format of the Grzywacz and Marks (2000) measure asked for how often the events occurred, which could mean they were treated more as a checklist by these midlife adults. A checklist-type scale is helpful for predicting other outcomes (Bollen & Lennox, 1991) and in this spirit we used these scales.

Job Satisfaction

Unlike other research using a one-item measure (e.g., Hill, 2005), four items averaged together indicated job satisfaction. The items were answered on a scale from 1 = A lot to 4 = Not at all (Bond, Galinsky, & Swanberg, 1998). Items were reverse scored so that higher scores indicated greater job satisfaction. Cronbach’s α was .83.

Marital Satisfaction

In contrast to a one-item measure used in other research (e.g., Hill, 2005), the Locke-Wallace Marital Adjustment Test (Locke & Wallace, 1959) assessed marital satisfaction using the original scale coding. According to Birchler, Weiss, and Vincent (1975), marital distress is indicated by a score of less than 100; participants in this study averaged 108.78, indicating satisfaction. Cronbach’s α was .80.

Personal Well-Being

At all waves, personal well-being was measured with a 9-item semantic differential scale adapted from Campbell et al. (1976). The Campbell et al. (1976) well-being scale uses positive emotions in keeping with Fredrickson’s (2001) theory. The Campbell et al. scale was used in the current study when it began in 1982. For continuity and because it captured well the concept developed in the intervening years of positive emotions, it was administered again in 1990 and 2007. Five points separated the bipolar items. For example, 1 = enjoyable anchored one end and 5 = miserable the other end (reverse scored). Instructions were the following: “Here are some words we would like you to use to describe how you feel about your present life.” Thus, the scale captured emotions about one’s life. Cronbach’s α at Waves 1, 4, and 5 were .77, .81, and .83, respectively. Scores were summed and could range from 9 to 45, with higher scores indicating greater well-being.

Demographic Variables

Demographic variables in the latent class analysis were gender (1 = male; 2 = female), Wave 5 financial status (1 = not at all/less than well off to 4 = very well off), occupation (1 = other; 2 = professional/managerial), length of current marriage in years, and number of children. Age and education were not used in the analyses as these had little variation because of the nature of the sample.

Social Background Variables

These were one-item variables. Family of origin financial status asked how well off the family was when growing up (1 = not at all well off to 5 = very well off financially). Parent’s marital status growing up was coded as parents not together = 0; parents together =1. Location growing up was 1 = rural, 2 = small city, 3 = suburb and large city. Cohort was coded 1 = 1980, 2 = 1982 cohort. Finally, religiosity in1982 asked the participant for their degree of religiousness (1 = not at all religious to 4 = strongly religious).

Other than the work-to-family conflict scale, none of the other scales showed evidence of significant skew or the presence of outliers. Raw score means and standard deviations are in Table 1; correlations of the variables of the study are available on request. Prior to analysis, all variables were centered with standardized scores (M = 0, SD = 1).

Table 1.

Raw Score Means (SDs) for Total Sample and Standardized Score Means by Latent Classes.

Scale	Raw score mean (SD), N = 125	Satisfied, n = 62	Low enhancement/low job satisfaction, n = 43	Unhappy, n = 20	F(2, 122)	η²
Work-to-family Conflict (WFC)	4.03 (1.41)	−.27^a	.19	.41^a	5.03**	.08
Family-to-work Conflict (FWC)	2.98 (1.31)	−.31^a	.15	.64^a	8.57****	.12
Work-to-family Enhancement (WFE)	3.24 (0.54)	.52^ab	−.60^b	−.32^a	23.24****	.28
Family-to-work Enhancement (FEW)	3.72 (0.68)	.65^ac	−.48^bc	−.99^ab	49.14****	.45
Wave 1 Well-being (WB)	36.49 (4.52)	.16^a	.01	−.53^a	3.77*	.06
Job Satisfaction (JS)	3.54 (0.53)	.36^a	−.47^a	−.11	10.81****	.14
Marital Satisfaction (MS)	108.78 (28.25)	.56^ac	.09^bc	−1.93^ab	193.88****	.76
Financial Status (F)	2.63 (0.74)	.22^a	−.07	−.52^a	4.47*	.07
Professional Occupation (O)	1.70 (0.46)	−.06	−.12	.43	2.26	.04
Gender female (G)	1.42 (0.50)	.54^ac	−.76^bc	−.05^ab	32.65****	.35

Note. Standardized score means are reported under groups. Standardized score row means with matching superscripts are significantly different across groups, Scheffé, p < .05.

Within column latent class significant differences, p ≤ .003:

Within Satisfied: WFC-WFE; WFC-FWE; WFC-JS; WFC-MS; WFC-G; FWC-WFE; FWC-FWE; FWC-WB; FWC-JS; FWC-MS; FWC-F; FWC-G; WFE-O; FWE-WB; FWE-O; WB-MS; MS-O; O-G.

Within Low Enhancement/Job Satisfaction: WFC-WFE; WFC-FWE; WFC-JS; WFC-G; FWC-WFE; FWC-FWE; FWC-G; WFE-WB; WFE-MS; WFE-F; FWE-MS; WB-G; JS-MS; MS-G; F-JS; O-G.

Within Unhappy: WFC-FWE; WFC-MS; FWC-WFE; FWC-FWE; FWC-MS; FWC-F; WFE-MS; FWE-MS; FWE-O; WB-MS; JS-MS; MS-F; MS-O; MS-G; F-O.

p < .05. **p < .01. ***p < .001. ****p < .0001.

Narratives and Coding of Narratives

Participants were asked to write about their lives:

There have been many changes in society and in the world in the past 25 years. In the blank page on the answer sheet, we would very much appreciate learning first about your changes, important events, challenges, over the past 25 years and second, in your own words learning about how you are doing today.

From this broad prompt, participants’ comments were highlighted inasmuch as they related to work, marital, and family lives. The highlighted segments were then coded according to the directed content analysis method described by Hsieh and Shannon (2005), in which the quantitative latent class groups were the starting points, and all coded segments related to work and family from the narratives were examined within each group. Subcodes under work included changes over time and finances; subcodes under family included marriage relationship, parenthood, family care. Change over time in each of these areas was noted. Each subcode was further annotated as to whether it was positive (“it’s been pretty cool”) or negative (“it’s a stress”) or neutral.

Once a list of these subcodes had been established, the entire open-ended response was read again, following the iterative process recommended by DeSantis and Ugarriza (2000).This iterative process allowed for new information to be placed in the subcodes and for the incorporation of new information not otherwise coded within respondents’ lives to be included in the analysis, particularly as it related to changes in their lives and comments about their well-being (Morse & Richards, 2002). Analysis focused on the ways work, finances, parenthood, family life, and marital life were described by men and by women. Based on the analysis, reporting in the results section generally concerns percentages of participants who mentioned topics that were coded (as subcodes), with quotes included as illustrative examples. Although multiple coders did not interact with the material and therefore interrater reliability is not established, expert qualitative scholars report that rigor is not achieved through such actions (Morse & Richards, 2002; Sandelowski, 1993; Wolcott, 1990). Rigor is enhanced by addressing the questions at hand and capturing the material in an iterative and comprehensive manner (Sandelowski, 1993; Wolcott, 1990), as was done in this study.

Results

Research Question 1: Person-Centered Analysis

Research Question 1 proposed that there are meaningful subgroups, latent classes, of people with identifiable patterns of midlife role commitments, satisfactions, demographics, and emerging adult well-being. Mplus, employing maximum likelihood estimation via the EM algorithm (Muthén & Muthén, 1998-2007), identified latent classes. Assignments of individuals to a latent class occurred on the basis of highest probability of membership in the latent classes. The minimum value of the Bayesian information criteria across 2, 3, 4, and 5 planned latent class solutions helped determine the number of latent classes. Other criteria used to determine number of latent classes included the following: n per latent class should be >5% and the average probability of latent class membership should be .70 or higher (Nagin, 2005).

Preliminary LCA indicated that length of marriage and number of children could be deleted from the latent class analysis. Entered in the latent class analyses were the following: Wave 5 work-to-family conflict, family-to-work conflict, work-to-family enhancement, family-to-work enhancement, job satisfaction, marital satisfaction, financial status, occupation, Wave 1 well-being, and gender of participant.

Based on the criteria, a three-group solution proved best, supporting the hypothesis that there would be three to four groups. The three-group solution had the minimum Bayesian information criteria value (3562.01 for 3-group vs. 3565.50 for 2-group and 3575.23 for 4-group solutions), lowest n in a group was 20 (16%), and latent class probabilities ranged from .90 to .97. Based on their profiles, the latent classes were labeled 1 = satisfied (n = 62), 2 = low enhancement/low job satisfaction (n = 43), and 3 = unhappy (n = 20; Table 1 and Figure 1).

Figure 1.

Patterns of standardized scores within three latent classes.

Comparisons of the standardized mean scores on variables that made up the latent classes were conducted via multivariate and univariate analyses of variance with the conservative Scheffé tests to follow up on significant between-group differences. As well, within-group scores were compared. Taken together, such comparisons provided validation that the latent classes captured unique patterns of participant scores. Overall, there were significant between-group differences ((Wilks’s = .08, p < .0001, η² = .72). One-way analyses of variance results (Table 1) indicated significant main effects of group on all variables except occupation. The follow-up multiple mean comparisons across the three groups supported the inclusion of the selected variables. Not only are between-group differences important to understanding the nature of the latent class solution, so are differences within groups. These analyses used paired t tests with p set at .003 to adjust for the number of comparisons within latent classes. Each variable had nine comparisons with other variables within its latent class. Significant within-group differences are summarized in the footnote to Table 1. Both between and within-group results validated the uniqueness of each group, that patterns were not simply a linear function of the variables.

Using standardized scores (M = 0, SD = 1), Figure 1 provides a graphical representation of the patterns of variable means within and between latent classes. A group with values all at the mean would have a symmetrical figure represented on the graph, plotted at zero for each variable. A figure with below average levels of conflict and higher average levels of marriage and job satisfaction, well-being, and enhancement would resemble the figure drawn for the satisfied group.

Research Question 1a: Role of Demographic Variables

Gender, financial status, and professional job were the demographic variables in the final latent class analyses. A chi-square analysis of gender by latent class indicated gender differences in group membership, χ²(2, N = 125) = 43.58, p < .0001. Men (60%, n = 12) and women (40%, n = 8) were proportionately distributed within the unhappy group. Women (69%, n = 43 vs. 31%, n = 19 men) were disproportionately more frequent in the satisfied latent class. Men were more often found in the low enhancement/low job satisfaction group compared with women (95%, n = 41 vs. 5%, n = 2 women). Overall, 81% of women were in the satisfied group versus 26% of the men. The relative position of financial status and professional job within groups indicated that these were similar within the satisfied and the low enhancement/low job satisfaction groups. However, the unhappy group was characterized by the lowest level of financial status across groups (p < .05) but the highest level of occupational status (ns). The discrepancy between mean scores on these two statuses was uniquely significant among those in the unhappy group (p < .003).

Research Question 1b: Social Background and Resources Analyses

To investigate possible social background and resource variables that could account for the group membership, we tested selected Wave 1 variables for differences across groups (p at .01 to control for number of tests). There were no significant differences across the three midlife groups on these Wave 1 variables (family of origin financial status, parent’s marital status, geographic environment growing up, cohort, and religiosity). In addition, bivariate correlations between measures of well-being in 1982, 1990, and 2007, with these 1982 variables, found no significant associations. Although limited in scope, these results suggest that the social background and resource variables assessed early on were unlikely sources of effects. As noted, few participants were involved in work and partnering in emerging adulthood, ruling out another possible confound. Early well-being as a contributor to midlife patterns seemed to stand alone. Results described below assess the contributions of earlier and later well-being.

Research Question 2: Earlier and Later Well-Being

Hypothesis

2a, Hypothesis 2b, Hypothesis 2c: Earlier and Later Well-Being and Midlife Patterns

Fredrickson (2001) theorized that earlier well-being would contribute to later patterns in a broaden-and-build manner. Thus, Wave 1 well-being was expected to be a significant component of the latent class solution. Wave 1 well-being was significant across latent classes, F(2, 122) = 3.77, p <.05, ^η² = .06, differentiating the satisfied group from the unhappy group (M = .16 vs. −.53, with low enhancement/low job satisfaction group nonsignificantly in between, M = .01). Furthermore, within each of the three latent classes, Wave 1 well-being was significantly different from other variables within the same latent class indicating that emerging adult well-being was not redundant of other variables included in the LCA.

If young adult well-being is important to midlife patterns as was emerging adult well-being, then Wave 4 well-being should also vary across the latent classes. In separate analyses, well-being at Wave 4 was compared across groups with Scheffé tests, p < .05. Similar to the results for Wave 1 well-being, Wave 4 well-being scores, F(2, 122) = 7.73, p = .001, η² = .11, of the unhappy group were significantly lower than those of the satisfied group (Ms = −.63 vs. .29) with the low enhancement group nonsignificantly in between (M = −.13). And if, as hypothesized, concurrent circumstances are important to midlife patterns, then Wave 5 well-being should likewise vary across latent classes. Wave 5 well-being scores, F(2, 122) = 26.29, p < .0001, η² = .30, of the unhappy group were significantly lower than both the low enhancement and satisfied groups (Ms = −1.18, −.02, .39, respectively). Converting the η² to Cohen’s d produced large effect sizes for each wave of well-being assessed (Wave 1, d = .5; Wave 4, d = .7; Wave 5, d = 1.3; Cohen, 1988). In sum, consistent with the hypothesis, there is support for the broaden-and-build view that well-being at earlier waves is an integral part of the identified groups. The Wave 4 and Wave 5 results occurred in spite of the fact that well-being at neither wave was part of the set of variables that were entered into the latent class analysis.

Research Question 3: Triangulation: Narratives Within Latent Classes

Seventy-eight percent of people in the satisfied latent class, 68% in the low enhancement/low job satisfaction group, and 80% in the unhappy group provided narratives at Wave 5. Directed content analysis (Hsieh & Shannon, 2005) examined narratives within each latent class, illuminated various life changes and thoughts experienced by these midlife adults, and provided context to their quantitative data. As described below, these narratives provided validation of the identified midlife patterns. Quotes are verbatim except for omission or slight altering of details designed to protect the identity of the respondents.

Respondents in the satisfied group, by and large, expressed a very positive attitude about life, with several writing “life is good.” It is important to note that narratives within this category were full of life crises. Coded changes included those occurring in work, especially finances, and family, including children and parents. Examples of changes were death of parents, employment setbacks, and challenges from children. One respondent wrote, “Our first years of marriage were framed by Texas’ economy—job layoffs, etc. . . . We then went through a period of poor health for me resulting in infertility.” One commonality was the strength and primacy of the marital relationship: for example, a man who had encountered a devastating financial loss associated with establishing a start-up company with a dishonest partner wrote, “Friends and family members lost money based on their trust in me. . . . My wife stood by me and our relationship was strong.” Another woman wrote about job changes, a 50% pay cut, caring for a relative, and her husband traveling frequently for work, along with children’s activities every evening: “We have many pets and a huge house that rarely gets cleaned. The complexity and craziness of our lives has made us appreciate and care for one another and life in general more passionately.”

Many people in this group (23 of 62, compared with only one person in each of the other groups) mentioned their faith in God and their religious devotion as assisting them in their positive outlook. An example comes from a woman who wrote:

Our marriage was transformed by the grace of God. . . . Despite the sudden death of my father almost ten years ago, and the fact that I am in the final stage of menopause, I must report that I am deeply content and grateful for my life now. I feel good and I share meaningful work with a wonderful husband.

In the second, mostly male, group characterized by low enhancement/low job satisfaction, virtually everyone wrote about work, including strains in their narratives associated with work (present in 52% of the narratives in this group). Even men who recorded positive dimensions to their lives also wrote about work stress:

I’ve been blessed. Everyone is healthy. I make more money than I ever dreamed. I find my work challenging and for the most part enjoyable. Yet my job has a lot of stress and frustration. Why do I put up with it? I’m driven.

Another man in this second group described changes over the years as: “I have less time and more pressure to perform.” As with those in the first group, job uncertainty has been a factor: “Go[ing] through major corporate reorganizations where the chance of losing my job was very high, seeing good friends and coworkers lose their jobs and have to change careers.” Men in this latent class wrote about rewarding home lives and longing to have leisure time and a more active home life, but having work demands that interfered with this wish. As one man wrote,

I’m at the top of my field where I work; I supervise a highly creative staff that is tasked way beyond their capacity on a daily basis. College never taught multi-tasking while making it up on the fly . . . trying to eat healthy while making a paycheck that will cover the bills, that sort of sums up my work life.

This same respondent wrote: “Home life is great—great boys—they’ve turned out exceptional and my wife and I just celebrated 25 years of marriage—pretty cool!” However, when family was mentioned in a less positive tone, it was because of caregiving strain. In the low enhancement group, 35% mentioned caregiving for a family member such as elderly parents or in-laws, a spouse, a child with a disability. In sum, members of this group may lack the necessary resources to achieve enhancement between work and family.

Similar to those in the other groups, respondents who were classified in the unhappy group experienced many life challenges at home and at work, but they expressed different attitudes about these life circumstances. Although respondents in the other latent classes also described health challenges and interpersonal relationship problems, respondents in the unhappy group emphasized their work lives less than those in the other groups—when work was mentioned, it was in the context of financial stress (reported by 32% of the group). For example, a woman wrote,

My husband’s job has always been stressful and finances very tight, loss of contact with family has all taken a toll on our relationships. . . . Our social connections just don’t seem to click or friends have moved away. I think that financial stress is our main problem and just weighing us down.

A major emphasis in many narratives (53%) was marital strain: “My struggles have really only been with [my wife]. I do love her but finding stability with her may never happen.” Similarly, “Relationship with spouse steady but no longer exciting. We tolerate and co-exist together.” In general, respondents in this group described feeling burdened and despondent, an example of which was one respondent who reflected the feelings of many members of this group: “Overall, I feel a sense of burden—I tell my wife life is very hard—it is a struggle to meet all the needs placed upon me.”

Discussion

Research Question 1: Midlife Patterns

The results of this study underscored the interconnections of work and family patterns with role satisfactions, gender, financial status, and occupation in midlife. The latent class results that addressed the first research question were consistent with theory and the hypothesis that there would be three to four identified groups. In the holistic approach suggested by role balance theory (Marks, 1977, 2009; Marks & MacDermid, 1996), the person-centered analyses produced a typological scheme that can be mapped onto Marks’s (1979) theoretical typology. There was a group that evidenced conflict between spheres, reflective of the role patterned systems of over and undercommitment theorized to be prevalent in U.S. society. We termed this group low enhancement/low job satisfaction to reflect the type of struggle evident. As well, when role conflicts were combined with low marital satisfaction, this combination reflected the pattern of equally negative experiences in role commitments that Marks has termed nihilistic or apathetic and that we termed unhappy. The unhappy group had markedly lower midlife marital satisfaction compared with the other two groups and lower well-being at all waves compared with the satisfied group. In comparison, the third group, labeled satisfied, seemed to reflect a system of equally positive commitments to the roles in their lives. The narratives captured the spirit of a sacrilizer or balanced group posited by Marks (1979), in the sense that they found work and family roles, as well as other aspects of their lives, to be sources of inspiration and reward.

In the latent class analysis, enhancement was a key variable in distinguishing between groups. Indeed, latent class membership explained 45% of the variance in family-to-work enhancement. In particular, the unhappy latent class had low family-to-work enhancement whereas the low enhancement/low satisfaction latent class had low enhancement from both work-to-family and family-to-work domains. In part, the narratives allow us to understand what happened to them. The low enhancement/low job satisfaction group was not necessarily disadvantaged nor had they poor resources or skills; they started out lower than the satisfied group in terms of well-being but not significantly lower. However, life circumstances, especially work stress but also caregiving, seemed to prevent them from broadening. In terms of enhancement, the satisfied group had high enhancement across the work and family spheres, the highest levels of the three groups. As Gareis et al. (2009) and others (Frone, 2003) have commented, enhancement has been less investigated than conflict; the findings in the current study support the view that it should be further researched with multiple methods.

Research Question 1a: Demographic Variables

Demographic variables were also important in distinguishing among groups. Within the unhappy group, financial status was significantly lower than occupational status, suggesting a possible struggle between reality and expectations for this well educated group. These findings may reflect the experience of status inconsistency (Bacharach, Bamberger, & Mundell, 1993) where objective or subjective standing in one hierarchy (financial status) is at odds with standing in another hierarchy (occupational status), and is associated with job stress or strain. Moreover, among older workers, such as those in this study, there may be a sense of having achieved a permanent position in the status hierarchy, thus workers in midlife experiencing status inconsistency may evidence higher levels of strain compared with younger workers for whom change is still a possibility. The narratives among those in the unhappy group expressed long-standing struggles with financial issues.

Although the literature suggested possible associations of financial circumstances with measures of satisfaction and well-being, such associations seemed to depend on group membership. For example, the lack of associations between being financially well off and marital and job satisfaction in the satisfied group supports the view that people can be happy in the absence of comparative wealth (Ryan & Deci, 2001). However, the lower level of financial status on the part of people in the unhappy group could mean that financial status is part of a pattern of unhappiness for some people at midlife. As noted by Marks (1979, 2009), there may be societal pressures on men’s earner roles and expectations for job rewards (Marks, 2009; Rothbard, 2001). Unfulfilled expectations about work and work stress may take on a salience in some men’s lives as reflected in the low job satisfaction of those men in the second group.

As well as financial and occupational issues, the study considered the roles of other demographic variables such as children, length of marriage, and gender. Most of the participants were in long-term marriages such that the impact of length of marriage would be minimized by lack of variability; this variable was deleted after preliminary analyses. Unlike the more studied families involving young children (Bianchi & Milkie, 2010), in this midlife sample, having children was unrelated to the variables of interest. As suggested by Bowes (2005), the ages of children may be important. When adults have young children in the home, there may be more impact of children on work and family issues than would be expected for midlife adults whose children are older and often out of the home altogether. This study expanded on earlier research by using participants at a later life stage than those studies that exclusively included participants with children in the home (Hill et al., 2008).

Of great interest is the finding that membership in the latent classes was highly gendered. These findings are consistent with the findings of Han and Moen (1999) on a somewhat older cohort of retired adults where trajectories of career and partnering were heavily gendered. In the current study, the only group with a balanced number of men and women was the group characterized by unhappiness. In the satisfied group there were more women than men. Van Steenbergen, Kluwer, and Karney (2011) indicated that women’s work demands, especially when happy with their jobs, may liberate them from juggling two domains, thereby increasing their satisfaction with marriage. Gordon, Whelan-Berry, and Hamilton’s (2007) study of working women over age 50 found that levels of work/family enhancement were higher than levels of work/family conflict. In addition, the women made intentional choices about their responsibilities and used personalized definitions of success. The authors speculated that such approaches to life may have contributed to the work/family tilt toward enhancement. According to Gordon et al. (2007, p. 361), an enhanced ability to solve conflict in support of a personal definition of success “likely alleviates any potential impact of such conflict on work-related outcomes.” Although Gordon et al.’s research helps explain women appearing more frequently in the study’s satisfied group (81% of women) than the other two groups, it does not explain men’s relative scarcity in this satisfied group (26% of men). Instead, low enhancement and low job satisfaction characterized the situations of most of the men in this study. Apparently, men still experience breadwinning role pressures (LaRossa et al., 2005; Marks, 2009).

Research Question 1b: Social Background Advantage

With a longitudinal design more confidence may be placed in identifying or eliminating alternative causal explanations. For example, it was implausible to suggest that preexisting work and family issues from 25 years previous contributed to the findings because almost none were partnered nor in careers at that early date. However, as predicted by the broaden-and-build theory (Fredrickson, 2001), well-being measured at earlier times was associated with patterns seen many years later.

Third variables, such as those reflecting social background factors at the earlier time, could have accounted for both well-being then and later work and family success. Of note is the lack of significant findings for variables selected for their possible contribution to cumulative advantage (DiPrete & Eirich, 2006; O’Rand, 2002). It may be that in a less homogeneous sample, there would be some advantage to higher standing on these variables. Instead, broaden-and-build contributions came from well-being in emerging adulthood. At this juncture, the findings fit the theory but are not definitive in proving cause and effect.

Research Question 2: Well-Being

The latent class analysis used for examining Research Question 2a is the same as that for Research Question 1. However, the focus shifts to the role of early well-being. Over a 25-year period, well-being was an enduring and broadening characteristic among many in the study, findings consistent with research and theory (Fredrickson, 2001; Pavot & Diener, 2008). For example, Bowling, Eschleman, and Wang’s (2010) meta-analysis of longitudinal studies indicated that the effect of subjective well-being on job satisfaction was stronger than the effect of job satisfaction on well-being. Against changes in marriage and work, earlier well-being was a possible long-term advantage, perhaps offering a protective mechanism (Allen et al., 2012) through a general tendency to find satisfaction in a number of areas of life, including work (Bowling et al., 2010) and family.

The 25-year trajectory of the unhappy group (low marital satisfaction but average job satisfaction) is revealing. This group began adult life with lower well-being, perhaps signaling that experiences were filtered through a lens that saw life’s stresses with pessimism. As the narratives illustrated, the differences between the unhappy and satisfied group were not so much in the actual challenges each faced but in the sense of strain and lack of joy experienced by the unhappy group, especially in their marriages. As noted by Salmela-Aro, Tolvanen, and Nurmi (2011) in their longitudinal study on earlier well-being (optimism) and later job engagement and burnout, those lower in well-being may be more vulnerable to negative feelings and use less effective coping, a process that amplifies over time (Fredrickson, 2001).

The pattern of findings as described above for the satisfied group and the unhappy group fit the hypothesis; however, the second group characterized as low in enhancement and low in job satisfaction offered a pattern that provides nuance to the broaden-and-build hypothesis. That is, the members of this group were nonsignificantly in between the members of the other two groups on earlier well-being but were significantly unlike either group at midlife. Their narratives reflected life circumstances such as work stress and caregiving issues that may have interfered with using earlier well-being to broaden and build.

Research Question 3: Triangulation of Quantitative and Narrative Findings

The integration of quantitative findings with narratives provided insights into the lives of participants. An advantage to using both quantitative and qualitative methods is the opportunity for mutual validation as well as the possibility of identifying aspects of lives that are not captured in the quantitative analyses. Although the narratives largely validated and extended understanding of the quantitative results, there were two possibly missed issues in the quantitative approach. One missed issue appeared to be the importance of religiousness or spirituality in the satisfied group. Literature points to the centrality of religion as a major life domain or social role alongside work, family, community, and leisure (Sharabi & Harpaz, 2011). Religion may (a) influence choices people make about dealing with work and family life (Ammons & Edgell, 2007); (b) serve as a resource, alleviate conflict, and increase facilitation (Davis, 2011); and (c) promote expression of prosocial behavior to others and reduce harmful behaviors in work and family domains (Marks & Dollahite, 2011). Future work on midlife work and family may include issues of religiousness given the mentions of these by members of the satisfied group.

A second missed issue was the crossover effect of a partner’s work and family experiences on the participant’s life choices (Han & Moen, 1999). The quantitative study did not ask about partners’ effects on the participant but the narratives support including these crossover effects. However, limiting crossover effects to the couple, family, or work organization potentially leaves unexplored some important issues. For example, Moen (2011) has argued that work and family are not individual issues but are issues of family, organizations, and communities, and that attention needs to be given to context and the higher-order social forces that affect and change work and family. Moen has asked for a holistic consideration of the fit between work and family. Similarly, Gelfand and Knight (2005) noted the multilevel context of work–family at various levels. They asked researchers to consider how these levels relate to each other and to identify moderators of their associations. The current research can be used as a springboard for consideration of these larger issues.

Limitations, Strengths, and Future Directions

Strengths of the study included a 25-year prospective longitudinal design, theory-based research questions, and quantitative and qualitative methods. Despite these strengths, there are limitations. One limitation is the use of reports from only one relationship partner and the use of measures that did not ask for the extent to which partner’s work affected family life (i.e., partner crossover effects; Gareis et al., 2009; Minnotte, Stevens, Minnotte, & Kiger, 2007). Also, coping mechanisms and strategies would be important to investigate further (Rantanen et al., 2011; Salmela-Aro et al., 2011). More fine-grained analyses could consider the sources of demands (Voydanoff, 2005) and enhancement (Van Steenbergen & Ellemers, 2009), such as time, strain, boundary-spanning, energy, or consider the contributions of emotion work (Minnotte et al., 2007). Research with larger samples would be needed to include such distinctions.

The findings may be limited to those who participate in longitudinal research as well as to this specific well-educated, largely European-American cohort. As long-term longitudinal data sets are analyzed with respect to these issues, the generalizability of the findings to more diverse groups may become testable.

Conclusions

Analysis of heterogeneity in the participants’ patterns found three groups largely consistent with Marks’s (1979) theorizing. A fairly large group of mostly women achieved balance in roles and enjoyed marital and job satisfaction. They seemed to frame even stressful events with optimism and humor. The second group, almost exclusively male, was different from either of the other two groups in important ways revolving around their low degree of work and family enhancement. Job stress did not seem to compromise marital satisfaction or well-being, but the narratives reflected a high degree of job-related concerns. Finally, a profound lack of marital satisfaction coupled with high conflict and low enhancement between work and family domains were characteristic of a small group of men and women who seemed characterized by enduring low levels of well-being and, in midlife narratives, wrote about concerns with finances and feeling burdened. Status inconsistency in occupation and financial roles may have been a contributor to unhappy participants’ midlife job and marital satisfaction. In addition, well-being, measured in emerging (25 years prior) and young adulthood (17 years earlier), was salient to the later work and family considerations.

This study provided support for two major theories, one by Marks (1979) on patterns of work and family and another by Fredrickson (2001) on the role of early well-being to later life outcomes. An integration of these theories suggests that early well-being, however achieved, may be an important contribution to how individuals deal with the issues of work and family in their adult lives, although there is some indication that life events may limit the effectiveness of earlier well-being. Had we measured well-being only at Wave 5, we might have viewed well-being as an outcome of work and family arrangements (Carlson & Perrewd, 1999; Gareis et al., 2009). Instead, well-being may shape these contexts, consistent with the research of Rantanen et al. (2011).

The findings from the narratives include important suggestions for broadening the study of work and family to include faith issues and religiousness (Sharabi & Harpaz, 2011) and crossover effects. Status inconsistency (Bacharach et al., 1993) could be studied further as it may help explain some of the patterns at play in the unhappy group. Prominent in this study was the role of gender. The highly gendered life trajectories reported by Han and Moen (1999) on an older cohort were echoed in the current findings. As younger cohorts are studied, perhaps gender may not be an issue when work and family are considered. Unfortunately, this research did not point to work and family patterns in which men and women were similar except among those who were unhappy. Understanding more about ways to develop family and work enhancement, to raise well-being across the life course, and to strengthen well-being’s ability to broaden-and-build in the face of challenges are important considerations for future research. But these are not simply issues for individuals or families; they are challenges as well for larger contexts such as institutions and social policy (Gelfand & Knight, 2005; Moen, 2011). Finally, additional theory development that integrates work, family, and broaden-and-build concepts should pay dividends.

Footnotes

Acknowledgements

Yingli Li, Brittney Schrick, and Jackie Wiersma helped in various ways as did past and current students. We thank Sylvia Niehuis for her very helpful comments on the article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interests with respect to the authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: NICHD provided the original grant (1982-1985) to Judith Fischer and Donna Sollie (“Network Supports and Coping During Adult Transitions”; R01 HD 15864). Greg Howard conducted the 1990 follow-up. The College of Human Sciences and the Department of Human Development and Family Studies at Texas Tech provided support for the 2007 follow-up.

References

Allen

T. D.

Johnson

R. C.

Saboe

K. N.

Cho

Dumani

Evans

(2012). Dispositional variables and work-family conflict: A meta-analysis. Journal of Vocational Behavior, 80(1), 17-26. doi:10.1016/j/jvb.2011.04.004

Ammons

S. K.

Edgell

(2007). Religious influences on work-family trade-offs. Journal of Family Issues, 28, 704-826. doi:10.1177/0192513X07299682

Amstad

F. T.

Meier

Fasel

Elfering

A. C.

Semmer

N. K.

(2011). A meta-analysis of work-family conflict and various outcomes with a special emphasis on cross-domain versus matching-domain relations. Journal of Occupational Health Psychology, 16, 151-169. doi:10.1037/a0022170

Bacharach

S. B.

Bamberger

Mundell

(1993). Status inconsistency in organizations: From social hierarchy to stress. Journal of Organizational Behavior, 14, 21-36.

Bergman

L. R.

Magnusson

(1997). A person-oriented approach in research on developmental psychopathology. Development and Psychopathology, 9, 291-319.

Bianchi

S. M.

Casper

King

(2005). Workforce/workplace mismatch? Work, family, health, and well-being. Mahwah, NJ: Erlbaum.

Bianchi

S. M.

Milkie

M. A.

(2010). Work and family research in the first decade of the 21st century. Journal of Marriage and Family, 72, 705-725. doi:10.1111/j.1741-3737.2010.00726.x

Birchler

G. R.

Weiss

R. L.

Vincent

J. P.

(1975). Multimethod analysis of social reinforcement exchange between martially distressed and nondistressed spouse and stranger dyads. Journal of Personality and Social Psychology, 31, 349-360. doi:10.1037/h00726280

Boles

J. S.

Wood

J. A.

Johnson

(2003). Interrelationships of role conflict, role ambiguity, and work-family conflict with different facets of job satisfaction and the moderating effects of gender. Journal of Personal Selling and Management, 23, 99-113.

10.

Bollen

Lennox

(1991). Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin, 110, 305-314. doi:10.1037/0033-2909.110.2.305

11.

Bond

J. T.

Galinsky

Swanberg

J. E.

(1998). The 1997 national study of the changing workforce. New York, NY: Families and Work Institute.

12.

Bowes

J. M.

(2005). Emphasizing the family in work-family research: A review of current research and recommendations for future directions. In Poelmans

S. A. Y.

(Ed.), Work and family: International research perspective (pp. 415-438). Mahwah, NJ: Lawrence Erlbaum.

13.

Bowling

N. A.

Eschleman

K. J.

Wang

(2010). A meta-analytic examination of the relationship between job satisfaction and subjective well-being. Journal of Occupational and Organizational Psychology, 83, 915-954. doi:1.10.1348/096317909X478557

14.

Byron

(2005). A meta-analytic review of work-family conflict and its antecedents. Journal of Vocational Behavior, 67, 169-198.

15.

Campbell

Converse

P. E.

Rodgers

W. L.

(1976). The quality of American life. New York, NY: Russell Sage Foundation.

16.

Carlson

D. S.

Kacmar

M. K.

Williams

L. J.

(2000). Construction and initial validation of a multidimensional measure of work-family conflict. Journal of Vocational Behavior, 56, 249-276. doi:10.1006/jvbe.1999.1713

17.

Carlson

D. S.

Perrewd

P. L.

(1999). The role of social support in the stressor-strain relationship: An examination of work-family conflict. Journal of Management, 25, 513-540.

18.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

19.

Cohen

Geron

Farchi

(2010). A typology of martial quality of enduring marriages in Israel. Journal of Family Issues, 31, 727-747. doi:10.1177/0192513X09358566

20.

Cohn

M. A.

Fredrickson

B. L.

Brown

S. L.

Mikels

J. A.

Conway

A. M.

(2009). Happiness unpacked: Positive emotions increase life satisfaction by building resilience. Emotion, 9, 361-368. doi:10.1037/a0015952

21.

Coser

(1974). The family: Its structures and functions. New York, NY: St. Martin’s Press.

22.

Cullen

J. C.

Hammer

L. B.

Neal

M. B.

Sinclair

R. R.

(2009). Development of a typology of dual-earner couples caring for children and aging parents. Journal of Family Issues, 30, 458-483. doi:10.1177/0192513X08326003

23.

Davis

S. N.

(2011). Support, demands, and gender ideology: Exploring work-family facilitation and work-family conflict among older workers. Marriage & Family Review, 47, 363-382. doi:10.1080/01494929.2011.594216

24.

DeSantis

Ugarriza

D. N.

(2000). The concept of theme as used in qualitative nursing research. Western Journal of Nursing Research, 22, 351-372.

25.

Diener

Shuh

E. M.

Lucas

R. E.

Smith

H. L.

(1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276-302. doi:1037/0033-2909.125.2.276

26.

DiPrete

T. A.

Eirich

G. M.

(2006). Cumulative advantage as a mechanism for inequality: A review of theoretical and empirical developments. Annual Review of Sociology, 32, 271-297. doi:0.1146/annurev.soc.32.061604.123127

27.

Drago

(2000). Trends in working time in the U.S.: A policy perspective. Labor Law Journal, 51, 212-218.

28.

Fischer

J. L.

Sollie

D. L.

Sorell

G. T.

Green

(1989). Marital status and career stage influences on social networks of young adults. Journal of Marriage and the Family, 51, 521-534. doi:10.2307/352513

29.

Fredrickson

B. L.

(2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56, 218-226. doi:10.1037/0003-066X.56.3.218

30.

Frone

M. R.

(2003). Work-family balance. In Quick

J. C.

Tetrick

L. E.

(Eds.), Handbook of occupational health psychology (pp. 143-162). Washington, DC: American Psychological Association. doi:10.1006/jvbe.1996.1577

31.

Frone

M. R.

Yardley

J. K.

Markel

K. S.

(1997). Developing and testing an integrative model of the work-family interface. Journal of Vocational Behavior, 50, 145-167.

32.

Gareis

K. C.

Barnett

R. C.

Ertel

K. A.

Berkman

L. F.

(2009). Work-family enrichment and conflict: Additive effects, buffering, or balance. Journal of Marriage and Family, 71, 696-707. doi:10.111/j.1741-3737.2009.00627.x

33.

Gelfand

M. J.

Knight

A. P.

(2005). Cross-cultural perspectives on work-family conflict. In Poelmans

S. A. Y.

(Ed.), Work and family: International research perspective (pp. 401-414). Mahwah, NJ: Lawrence Erlbaum.

34.

Goode

W. J.

(1960). A theory of role strain. American Sociological Review, 25, 483-496.

35.

Gordon

J. R.

Whelan-Berry

K. S.

Hamilton

E. A.

(2007). The relationship among work-family conflict and enhancement, organizational work-family culture, and work outcomes for older working women. Journal of Occupational Health Psychology, 12, 350-364. doi:10.1037/1076-8998.12.4.350

36.

Grote

N. K.

Frieze

I. H.

Stone

C. A.

(1996). Children, traditionalism in the division of family work, and marital satisfaction: “What’s love got to do with it?” Personal Relationships, 3, 211-228. doi:10.1111/j.1475-6811.1996.tb00113.x

37.

Grzywacz

J. G.

Bass

B. L.

(2003). Work, family and mental health: Testing different models of work-family fit. Journal of Marriage and Family, 65, 248-261. doi:10.111/j.1741-3737.2003.00248.x

38.

Grzywacz

J. G.

Marks

N. F.

(2000). Reconceptualizing the work-family interface: An ecological perspective on the correlates of positive and negative spillover between work and family. Journal of Occupational Health Psychology, 5, 111-126. doi:10.1037/1076-8998.5.1.111

39.

Hagerty

M. R.

(2000). Social comparisons of income in one’s community: Evidence from national surveys of income and happiness. Journal of Personality and Social Psychology, 78, 764-771. doi:10.1037/0022-3514.78.4.764

40.

Han

S.-K.

Moen

(1999). Work and family over time: A life course approach. The Annals of the American Academy of Political and Social Science, 562, 98-110. doi:10.1177/000271629956200107

41.

Hill

E. J.

(2005). Work family facilitation and conflict, working fathers and mothers, work-family stressors and support. Journal of Family Issues, 26, 793-819. doi:10.1177/0192513X05277542

42.

Hill

E. J.

Jacob

J. I.

Shannon

L. L.

Brennan

R. T.

Blanchard

V. L.

Martinengo

(2008). Exploring the relationship of workplace flexibility, gender, and life stage to family-to-work conflict, and stress and burnout. Community, Work & Family, 11, 165-181. doi:10.1080/13668800802027564

43.

Hsieh

Shannon

S. E.

(2005). Three approaches to qualitative content analysis. Qualitative Health Research, 15, 1277-1288. doi:10.1177/1049732305276687

44.

LaRossa

Simonds

Reitzes

D. C.

(2005). Culture, cognition, and parenthood. In Bengtson

V. L.

Acock

Allen

Dillworth-Andern

Klein

D. M.

(Eds.), Sourcebook of family theory and research (pp. 423-445). Thousand Oaks, CA: Sage.

45.

Linley

P. A.

Joseph

Harrington

Wood

A. M.

(2006). Positive psychology: Past, present, and (possible) future. Journal of Positive Psychology, 1, 3-16. doi:10.1080/17439760500372796

46.

Locke

H. J.

Wallace

K. M.

(1959). Short marital-adjustment and prediction tests: Their reliability and validity. Marriage and Family Living, 21, 251-255. doi:10.1177/1088868306294590

47.

Lucas

R. E.

Diener

Suh

(1996). Discriminant validity of well-being measures. Journal of Personality and Social Psychology, 71, 616-628. doi:10.1037/0022-3514.71.3.616

48.

Malinen

Kinnunen

Tolvanen

Rönkä

Wierda-Boer

Gerris

(2010). Happy spouses, happy parents? Family relationships among Finnish and Dutch dual earners. Journal of Marriage and Family, 72, 293-306. doi:10.1111/j.1741-3737.2010.00700.x

49.

Marks

L. D.

Dollahite

D. C.

(2011). Mining the meanings and pulling out the processes from psychology of religion’s correlation mountain. Psychology of Religion and Spirituality, 3, 181-193. doi:1.1037/a0022206.

50.

Marks

S. R.

(1977). Multiple roles and role strain. Some notes on human energy, time, and commitment. American Sociological Review, 42, 921-936. doi:10.2307/2094577

51.

Marks

S. R.

(1979). Culture, human energy, and self-actualization: A sociological offering to humanistic psychology. Journal of Humanistic Psychology, 19, 27-42. doi:10.1177/002216787901900308

52.

Marks

S. R.

(2009). Multiple roles and life balance: An intellectual journey. In. Matsuka

Christiansen

(Eds.), Life balance: Multidisciplinary theories and research (pp. 43-58). New York, NY: Slack & Aota Press.

53.

Marks

S. R.

MacDermid

S. M.

(1996). Multiple roles and the self: A theory of role balance. Journal of Marriage and Family, 58, 417-432. doi:10.2307/353506

54.

Minnotte

K. L.

Stevens

D. P.

Minnotte

M. C.

Kiger

(2007). Emotion-work performance among dual-earner couples: Testing four theoretical perspectives. Journal of Family Issues, 28, 773-793. doi:10.1177/0192513X072999676.

55.

Moen

(2011). From “work-family” to the “gendered life course” and “fit”: Five challenges to the field. Community, Work & Family, 14, 81-96.

56.

Morse

J. M.

Richards

(2002). Readme first: A user’s guide to qualitative methods. Thousand Oaks, CA: Sage.

57.

Muthén

L. K.

Muthén

B. O.

(1998-2007). Mplus user’s guide (5th ed.). Los Angeles, CA: Muthén & Muthén.

58.

Nagin

D. S.

(2005). Group-based modeling of development. Cambridge, MA: Harvard University Press.

59.

Netemeyer

R. G.

Boles

J. S.

McMurrian

(1996). Development and validation of work-family conflict scales and family-work conflict scales. Journal of Applied Psychology, 81, 400-410. doi:10.1037/0021-9010.81.4.400

60.

O’Rand

A. M.

(2002). Cumulative advantage theory in life course research. In Schaie

K. W.

(Series Ed.) & Crystal

Shea

(Vol. Eds.), Annual review of gerontology and geriatrics: Vol. 22. Focus on economic outcomes in later life (pp. 14-30). New York, NY: Springer.

61.

Parasuraman

Simmers

(2001). Type of employment, work-family conflict, and well-being: A comparative study. Journal of Organizational Behavior, 22, 551-568. doi:10.10002/job.102

62.

Pavot

Diener

(2008). The Satisfaction With Life Scale and the emerging construct of life satisfaction. Journal of Positive Psychology, 3, 137-152. doi:10.1080/17439760701756946

63.

Poelmans

O’Driscoll

Beham

(2005). An overview of international research on the work-family interface. In Poelmans

S. A. Y.

(Ed.), Work and family: International research perspective (pp. 3-46). Mahwah, NJ: Lawrence Erlbaum.

64.

Powell

G. N.

Greenhaus

J. H.

(2010). Sex, gender, and the work-to-family interface: Exploring negative and positive interdependencies. Academy of Management Journal, 53, 513-534. doi:10.5465/amj.2010.5146847

65.

Rantanen

Kinnunen

Feldt

Pulkkinen

(2008). Work-family conflict and psychological well-being: Stability and cross-lagged relations within one- and six-year follow-ups. Journal of Vocational Behavior, 73, 37-51. doi:10.1016/j.jvb.2008.01.001

66.

Rantanen

Mauno

Kinnunen

Rantanen

(2011). Do individual coping strategies help or harm in the work-family conflict situation? Examining coping as a moderator between work-family conflict and well-being. International Journal of Stress Management, 18, 24-48. doi:10.1037/a0022007

67.

Rothbard

(2001). Enriching or deleting? The dynamics of engagement in work and family roles. Administrative Science Quarterly, 46, 655-684.

68.

Ryan

R. M.

Deci

E. L.

(2001). On happiness and human potentials: A review of research on hedonic and eudaimonic well-being. Annual Review of Psychology, 52, 141-166. doi:10.1146/annurev.psych.52.1.141

69.

Ryff

C. D.

Singer

B. H.

(2006). Commentary: Best news yet on the six-factor model of well-being. Social Science Research, 25, 1103-119. doi:10.1016/j.ssresearch.2006.01.002

70.

Salmela-Aro

Tolvanen

Nurmi

J.-E.

(2011). Social strategies during university studies predict early career work burnout and engagement: 18-year longitudinal study. Journal of Vocational Behavior, 79, 145-157. doi:10.1016/j.jvb.2011.01.002

71.

Sandelowski

(1993). Rigor or rigor mortis: The problem of rigor in qualitative research revisited. Advances in Nursing Science, 16, 1-8.

72.

Sharabi

Harpaz

(2011). Gender and the relative centrality of major life domains: Changes over the course of time. Community, Work & Family, 14, 57-62. doi:10.1080/13668803.2010.506033

73.

Siedlecki

K. L.

Tucker-Drob

E. M.

Oishi

Salthouse

T. A.

(2008). Life satisfaction across adulthood: Different determinants at different ages? Journal of Positive Psychology, 3, 153-164. doi:10.1080/17439760701834602

74.

Steiber

(2009). Reported levels of time-based and strain-based conflict between work and family roles in Europe: A multilevel approach. Social Indicators Research, 93, 469-488. doi:10.1007/s11205-008-9436-z

75.

Stevens

D. P.

Minnotte

K. L.

Mannon

S. E.

Kiger

(2007). Examining the “neglected side of the work-family interface”: Antecedents of positive and negative family-to-work spillover. Journal of Family Issues, 28, 242-262. doi:10.11770192513X07299676

76.

Stryker

Burke

P. J.

(2000). The past, present, and future of an identity theory. Social Psychology Quarterly, 63, 284-297.

77.

Van Steenbergen

E. F.

Ellemers

(2009). Is managing the work-family interface worthwhile? Benefits for employee health and performance. Journal of Occupational Behavior, 30, 617-642. doi:10.1002/job.569

78.

Van Steenbergen

E. F.

Kluwer

E. S.

Karney

B. R.

(2011). Workload and the trajectory of marital satisfaction in newlyweds: Job satisfaction, gender, and parental status as moderators. Journal of Family Psychology, 25, 345-355. doi:10.1037/a0023653

79.

Voydanoff

(2005). Work demands and work-to-family and family-to-work conflict: Direct and indirect relationships. Journal of Family Issues, 26, 707-726. doi:10.1177/0192513X05277516

80.

Wolcott

(1990). On seeking: And rejecting: Validity in qualitative research. In Eisner

Peshkin

(Eds.), Qualitative inquiry in education: The continuing debate. (pp. 121-152). New York, NY: Teacher College Press.