Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 11.
Published in final edited form as: J Aging Health. 2019 Nov 29;32(9):1109–1119. doi: 10.1177/0898264319889813

Person–Work Mismatch, Retirement Context, and the Progression of Depressive Symptoms Over Mid-Later Years: A Dyadic Analysis of Couples in Enduring Marriages

Kandauda A S Wickrama 1, Catherine Walker O’Neal 1
PMCID: PMC8356029  NIHMSID: NIHMS1731032  PMID: 31782344

Abstract

Objectives:

This study investigates how person–work mismatch (PWM) and subsequent pre-retirement work circumstances lead to poor mental health in later years for husbands and wives in enduring marriages.

Methods:

Data from 224 dual-earner couples in enduring marriages who participated over 27 years (1991–2015) in their middle to their later years were used to test the conceptual model.

Results:

PWM was related to depressive symptoms in middle years, which continued into later years (a cumulative pathway). In addition, PWM contributed to a stressful pre-retirement work context, which, in turn, influenced depressive symptoms in later years (a stress proliferation pathway). Partner effects were also noted between husbands and wives.

Discussion:

The present study enhances knowledge about how middle-aged couples’ PWM is related to their mental health in later years through their stressful pre-retirement work context with implications for national- and state-level policies.

Keywords: retirement, work stress, depression, life course, mental health, marriage

Research has documented that mental and physical health deteriorate over the middle and later years of adulthood (i.e., mid-later years; Adler & Rehkopf, 2008; Sutin et al., 2013). This may be partly due to persistent exposure to stressful circumstances over the adult years (Lovallo, 2005). One stressful life circumstance is person–work mismatch (PWM), where mismatch refers to the level of incompatibility between an individual and his or her work (Kristof-Brown et al., 2005). Research suggests that over time, PWM may spillover to different aspects of work (Deniz et al., 2015). These stressful work conditions, along with other life stressors, can contribute to individuals’ mental health problems in later years (Griffin et al., 2002).

Much of the work research focused on the later years of life has been fragmented in several important ways. First, some studies have focused on the connection between work stress and health (Karasek & Theorell, 1990; Wickrama et al., 2008), and others focused on the connection between retirement and health (Carr et al., 2016). Thus, less is known about how stressful work experiences, specifically PWM, and a stressful pre-retirement work context (hereafter, stressful pre-retirement work) uniquely and jointly influence mental health outcomes in later years. Second, although husbands and wives have interdependent “linked lives” (Elder & Giele, 2009), previous studies focused on work and health have rarely investigated the association between these constructs in a dyadic context considering working husbands and wives simultaneously.

For a deeper understanding of these potential influences, using a life course framework, the present study addresses three questions: (a) Does PWM in middle years persistently and cumulatively influence mental health over the life course? (b) Irrespective of this direct influence, does PWM in middle years contribute to mental health outcomes in later years through the impact of PWM on stressful pre-retirement work in the later stages of work life? (c) Do spouses influence each other (i.e., partner effects) as it relates to associations between spouses’ PWM, stressful pre-retirement work, and mental health?

For these comprehensive “long view” analyses identifying work and retirement influences over the life course (Elder & Giele, 2009), panel studies of couples with extensive follow-up periods are necessary. Using prospective data collected from 224 working husbands and wives over a period of 25 years (1991–2015), this study examines husbands’ and wives’ PWM over their early middle years (40–50 years), their retirement context in mid-later years (50–65 years), and their mental health in both their middle years (50 years) and later years (67 years).

Two central points of the current study to highlight are the use of a dyadic analytical framework and the measurement of mental health. Utilizing a dyadic analytical framework (Kenny et al., 2006) enables an investigation of both intra-individual (actor) effects and inter-individual (partner) effects between husbands’ and wives’ PWM, retirement context, and mental health. Dyadic analyses are also preferred for couple-level data because they may reduce bias in parameter estimates (Kenny et al., 2006). Regarding the measurement of mental health, depressive symptoms are utilized as a proxy for poor mental health because research suggests that depressive symptoms are a key indicator of overall mental health, and they are often predictive of other internalizing symptoms, such as anxiety and somatization (Ollendick et al., 2005). Moreover, depressive symptoms are also relevant for studying inter-individual associations between working husbands and wives in enduring marriages because they are associated with individuals’ ability to function successfully in marital and parental roles (Rakow et al., 2010).

Findings from the current study will have practical implications for prevention and intervention efforts with older adults. First, in considering PWM as a work stressor with mental health consequences, the findings can inform organizational reform efforts that focus on workers’ performance as well as their well-being. This is an important extension of previous work-health research and intervention efforts, which have primarily focused on traditional work stressors, such as work demand and low control (Carr et al., 2016). Second, the current study assesses PWM and stressful pre-retirement work that may act as “push factors” to drive individuals out of the labor market and encourage early, or premature, retirement (e.g., perceived age discrimination in the work force as a consideration in when to retire). A better understanding of these push factors as antecedents of pre-mature retirement is imperative because pre-mature, or involuntary, retirement has been shown to have adverse economic and health consequences (Bender, 2012). Third, a novel feature of the current study is the dyadic analyses, which provides a better understanding of how spouses influence one another. Although work research has considered spillover and crossover influences between couple members, retirement has largely been viewed as an individual process, rather than a couple, or family, process even though a spouse’s retirement represents a major life transition with implications for both partners. The findings will draw attention to how work and retirement influence both spouses. Thus, the findings of the current study can inform couple therapy programs by shedding light on the importance of considering inter-individual consequences of work and retirement stress in addition to the more commonly addressed intra-individual consequences. Moreover, helping professionals can couple these identified dyadic processes with effective interventions that create and enhance positive interpersonal processes to protect spouses from the mental health influence of adverse life conditions (Gorin et al., 2008).

PWM in Middle Years and Mental Health in Later Years

Drawing from the worker-motivational perspective, PWM reflects the degree to which an organization or environment satisfies the individual’s needs, desires, or preferences (Kristof, 1996), which is thought to be closely related to the worker’s mental and physical well-being (Edwards & Shipp, 2007). Consistent with this previous occupational research, in the current study, we assess PWM using workers’ reports of incompatibility between their day-to-day work and their education, skill/abilities, and experiences along with their general report of disliking work or being unhappy with their work (Scroggins & Benson, 2007). We consider PWM over the middle years as a chronic stressor contributing to depressive symptoms in later years through a direct exposure mechanism (Kristof-Brown et al., 2005). As a direct exposure, PWM can be internalized, resulting in elevated levels of stress (Buchanan & Norko, 2011), which often generate depression and anxiety (Iacovides et al., 2003). In addition, PWM may generate feelings associated with decreased psychological well-being, such as lack of meaningfulness and lack of purpose (Myers & Diener, 1995).

We expect that elevated depressive symptoms in the middle years stemming from PWM may increase and continue into later years (Bryant et al., 2017; Sutin et al., 2013). Consistent with the life course framework (Glymour et al., 2010), we conceptualize this mechanism connecting midlife PWM and depressive symptoms in later years as a cumulative pathway because it represents the buildup of depressive symptoms over time. When an individual is exposed to continuing stress, including PWM, inter-related emotional and physiological stress responses may accumulate and increase the risk for further mental health problems (Ganzel et al., 2010).

Stressful Pre-Retirement Work as a Mechanism Linking PWM and Mental Health

We expect that PWM may influence depressive symptoms in later life indirectly through stressful pre-retirement work.

PWM and Stressful Pre-Retirement Work

The influence of PWM on stressful pre-retirement work in the later stages of work life may be attributed to both organizational changes and intra-individual processes. First, organizational changes in the work environment over time may require more skills for the same position (Lee et al., 2008). This may increase the incompatibility between the person and the work environment, which may precipitate as stressful pre-retirement work over time. Moreover, individuals who already experience PWM may be more vulnerable to institutional changes. This situation may be aggravated due to workers’ common hesitancy to seek out additional educational and skill training in the later stage of their careers (Bender & Heywood, 2011). Second, intra-individual processes stemming from PWM may generate feelings of work inefficiency, loss of control, and inability to meet work demands. These circumstances may contribute to negative work outcomes, such as failing to achieve higher wages or desired recognition/promotions/incentives and a general lack of opportunities, particularly in the later stages of one’s work life (Deniz et al., 2015; Iacovides et al., 2003). Importantly, these feelings may heighten retirement-related stress.

This process is consistent with the stress proliferation principle (Pearlin et al., 2005), which refers to a process in which an initial stressor gives rise to additional stressors (Pearlin et al., 2005). That is, stress can beget more stress within the same or different life domains over the life course because individuals’ life experiences are generally contingent on previous experiences in a sequential manner (Brown, 2009). Accordingly, we posit a stress proliferation pathway, expecting that working husbands and wives who experience PWM in their middle years will develop an increasingly stressful work and retirement context during the later stages of their work life.

Stressful Pre-Retirement Work and Depressive Symptoms

Based on research noting that individuals experiencing a stressful retirement context generally report poorer physical and emotional health and less satisfaction with their retirement and life (Henkens et al., 2008; Shultz et al., 1998), we expect that stressful pre-retirement work has detrimental mental health consequences that continue after retirement.

One primary reason for this continued effect is that stressful pre-retirement work can act as a key determinant in the timing of retirement, driving workers toward an early and/or involuntary, retirement (Negrini et al., 2013). Involuntary retirement due to a stressful work context (i.e., push factors) is more stressful than a motivated retirement where appealing factors, such as a good compensation package and a desire to travel and enjoy life, sway the worker toward retirement (i.e., pull factors; Barnes-Farrell, 2003; Szinovacz & Davey, 2004). Involuntary retirement also increases the likelihood of being unprepared for post-retirement life, leading to a poor adaptation post-retirement in many respects, including socially, emotionally, and financially (Bender, 2012).

Potential Partner Effects

The current study also addresses the question of: “Do spouses’ influence each other as it relates to associations between spouses’ PWM, retirement context, and mental health?” This question draws from both the relational perspective (Berscheid & Ammazzalorso, 2001) and family system theory (Fingerman & Bermann, 2000). These perspectives emphasize that couple members function interdependently, and their experiences occur in a context of mutual influences and interactions forming cross-over effects. That is, spouses have “linked lives,” and their daily life activities are intertwined. One spouse’s mood, behavior, and stresses can affect him/herself as well as their partner (Kiecolt-Glaser & Wilson, 2017). An individual’s stressful work and retirement experiences may transmit and impact their partner’s depressive feelings in addition to their own.

Accordingly, crossover, or partner effects, may exist between spouses in relation to PWM, stressful pre-retirement work, and depressive symptoms. For example, when one spouse perceives his or her work and/or retirement context as stressful, this may be communicated within the interdependent dyadic relationship to the partner whose emotional response, in turn, may include the development of depressive feelings (Berscheid & Ammazzalorso, 2001). Moreover, work stress contributes to negative interactions at home and the decreased provision of support and care toward the partner (Westman, 2001), which can result in more depressive feelings.

In addition, spouses may express similar levels of depressive symptoms even after accounting for actor and partner effects related to PWM and stressful pre-retirement work. Consistent with the emotional contagion perspective, depression can be “induced” and can “spread” contemporaneously from one partner to the other (Cacioppo et al., 2009). This contagion can create dependencies, or shared associations, between husbands’ and wives’ depressive symptoms. We account for these dependencies (non-directional correlations) when examining actor and partner effects.

The Conceptual Framework

The conceptual framework guiding the current study (see Figure 1) illustrates how husbands’ and wives’ PWM may lead to depressive symptoms in later years. Path A depicts a life course cumulative pathway (Glymour et al., 2010) through which PWM in middle years is expected to result in depressive symptoms that accumulate and persist into their later years. Path B depicts a stress proliferation pathway from PWM leading to depressive symptoms in later years through stressful pre-retirement work. This pathway represents a “chain of risk” where one stressor (PWM) may lead to another stressor (retirement context) before ultimately resulting in depressive symptoms (Kuh et al., 2003). In addition, as shown in Figure 1, these two pathways do not operate independently, but are likely inter-related, serving to further intensify the impact of PWM on husbands’ and wives’ mental health in later years.

Figure 1.

Figure 1.

Open in a new tab

A dyadic life course process framework for husbands’ and wives’ work, stressful pre-retirement context, and mental health.

Note. A = a cumulative pathway influence. B = a stress proliferation pathway influence. p = partner influence.

Moreover, we utilize the Actor–Partner Interdependence Model (APIM) approach (Kenny et al., 2006) to examine both intra-individual, actor effects and inter-individual associations (between husbands and wives) concerning the study constructs. In Figure 1, Paths P depict partner effects, while curved and vertical non-directional paths represent contemporaneous associations between husbands and wives in regard to their PWM, stressful pre-retirement work, and mental health.

Method

Participants and Procedures

The data used to evaluate these hypotheses are from the Iowa Youth and Family Project (IYFP, 1989–1994), which was later continued as two panel studies: the Midlife Transitions Project (MTP) in 2001 and the Later Adulthood Study (LAS) in 2015. Together, these projects provide data over 25 years on rural families from a cluster of eight counties in north-central Iowa that closely mirror the economic diversity of the rural Midwest. The IYFP began in 1989 as a study of rural couples with children, at least one of whom was a seventh grader in 1989 (Conger & Elder, 1994). For IYFP, MTP, and LAS, trained field interviewers visited the families in their homes. During the visits, trained interviewers asked each family member to complete a detailed questionnaire about their family life, work, finances, retirement, and mental and physical health. Family members completed the questionnaires independently, so that they could not see one another’s answers. The IYFP involved 451 families from eight counties in Iowa (see Conger & Elder, 1994). The site for the research was determined by interest in rural economic stress (i.e., the farm crisis) and well-being. Families selected to participate had at least two children. Eligible families were identified through contacts with the public and private schools within the eight counties. Median yearly family income in 1989 was US$33,240 (ranged from US$0 to US$259,000). In terms of occupational status, the men in this sample included crafts-men, foremen, and farmers (38.4%); professionals, managers, owners, and officials (23.8%); operatives and kindred workers (16.6%); sales workers, clerical, service workers, private household workers, and military service (14.4%); laborers (3.3%); and other (3.5%). Of the wives, 19% were homemakers. Occupations for the employed women included sales workers, clerical, service workers, and private household workers (46.1%); professionals, managers, owners, and officials (23.7%); operatives and kindred workers (4.2%); crafts-men, foremen, and farmers (2.9%); laborers (0.7%); and other (3.4%). Data collected in 1991, rather than 1989, were used as the first time point of the present study due to the availability of study variables.

We limited our sample for the current study to dual-earner couples in 1991 who also participated in 2015 and were consistently married from 1991 to 2015 (N = 224). Of the individuals, 60% in the current sample were retired in 2015. The attrition rate was 31% from 1991 to 2015. An attrition analysis compared demographic characteristics (i.e., age, education level, economic hardship measured by counts of economic cutbacks, and divorce proneness) and study variables (i.e., PWM and depressive symptoms) in 1991 between the current analytic sample of consistently married dual-earning couples and couples who were excluded from the current analyses. The only significant difference noted was for divorce proneness in 1991, with higher scores reported for couples who were excluded from the current analysis. This is not surprising because couples with poorer quality marriages were more likely to divorce and, therefore, be excluded from the current study of couples in enduring marriages.

In 1991, spouses were in their early middle years. The average ages of husbands and wives were 41.50 (standard deviation [SD] = 4.8) and 39.40 (SD = 3.90) years, respectively, and their ages ranged from 33 to 59 for husbands and 31 to 55 for wives. On average, the couples had been married for 19 years and had three children. The median age of the youngest child was 12 years. In 1989, the average number of years of education for husbands and wives for the current study sample was 13.80 and 13.56 years, respectively. Because there are very few minorities in the rural area studied, all participating families were White.

Measures

PWM.

Thus, drawing from work research (e.g., Scroggins & Benson, 2007), we developed a composite measure of PWM using seven items capturing various mismatch dimensions of work. These items include “My job matches my education and experience,” “My job allows me to use skills/abilities,” “My job matches what I would like to do,” “I can’t use my skills,” “I should have a different job with my experiences,” “I am over-qualified for this job,” and “My education and experience are better for another job.” The items were rated using a 5-point Likert-type scale (1 = strongly agree, 5 = strongly disagree) and reverse scored, when necessary, so that higher scores indicate high work mismatch. Mean scores were computed separately for husbands and wives at each of the four time points (1991, 1992, 1994, and 2001). The internal consistencies for perceived work mismatch ranged from .69 to .72 for husbands and wives across all time points. A composite measure of PWM was created by computing the average level of husbands’ and wives’ PWM over four time points from a decade in their middle years (1991–2001).

Stressful pre-retirement work context.

Seven items captured self-reported stressful pre-retirement work context in 2015. Husbands and wives indicated “the main factors that influenced, or are influencing, your retirement plans.” Items include “I cannot cope with the workload,” “I cannot handle new technology required for my work,” “Age discrimination,” “Lack of promotions or pay-raises for my seniority,” “Decrease in work control/autonomy,” “I cannot meet the physical job demands/poor health,” and “Rigid schedules, such as night shifts.” The number of items selected was summed with higher scores indicating a more stressful pre-retirement work context.

Depressive symptoms.

Nine items from the Symptom Check-list (SCL-90-R; Derogatis & Melisaratos, 1983) were used to capture self-reported ratings of depressive symptoms from the previous week for husbands and wives in 2001 and 2015. Sample items include “thoughts of ending your life,” “feelings of worthlessness,” and “feeling hopeless about the future.” These items were scored on a 5-point Likert-type scale (1 = not at all, 5 = extremely). A mean score was computed for each time point with higher scores indicating more depressive symptoms. The internal consistencies were greater than .90 for both husbands and wives in 2001 and 2015.

Control variables.

Analyses also initially accounted for the influence of age, education, previous job transitions, as well as four components of retirement status (i.e., individuals’ retirement status, partners’ retirement status, couple retirement status, and an age–retirement status interaction). Respondents’ self-reports of their birthdate were utilized to calculate their age in 2015. Respondents also reported their educational attainment in 1991. Life history calendars were utilized over the study period to capture various life events, including job transitions. From this information, we calculated a sum variable identifying the number of job changes experienced from 1991 to 2001. Individuals’ retirement status was assessed in 2015 by respondents reporting whether they were fully retired (2), partially retired (1), or still working (0). Responses were recoded as 1 = fully or partially retired and 0 = working. Because we also considered the possibility that outcomes may vary depending on the couples’ configuration as it relates to retirement status, we used the same binary coding to create a variable capturing partners’ retirement status. Couples’ retirement status distinguished couples comprised of two retired spouses (1) from couples with one or more working spouse (0). Finally, a product term was utilized to assess an age-by-retirement status interaction.

Analytical strategy

We tested the conceptual model using path analysis in a structural equation modeling framework with Mplus statistical software (version 8.0; Muthén & Muthén, 1998/2017, Los Angeles, CA, USA). After fitting the path analysis, most control variables were removed from the final model for parsimony because they did not predict depressive symptoms in later adulthood (i.e., education, previous job disruptions, partners’ retirement status, couple retirement status, and an age–retirement status interaction). The final model included age and retirement status as control variables. The indirect effects were estimated using Bootstrap procedure available in Mplus. Missing data were accounted for using Full Information Maximum Likelihood (FIML) procedures (Enders & Bandalos, 2001). FIML does not impute missing values; rather, it estimates model parameters and standard errors from all available data, which minimizes potential biases that can influence the results (Enders & Bandalos, 2001). Model fit was evaluated using the chi-square divided by degrees of freedom, comparative fit index (CFI > .95), Tucker–Lewis index (TLI > .95), and the root mean square error of approximation (RMSEA < .06).

Results

Univariate and Bivariate Statistics

Table 1 presents descriptive statistics of the study variables. In general, husbands and wives reported low levels of PWM across their middle years (a variable capturing the average of PWM over four occasions from 1991 to 2001; M = 2.25 and 2.24, respectively). Husbands and wives, on average, reported less than one of the seven indicators used to capture stressful pre-retirement work, although the value was slightly higher for wives (M = .97) than husbands (M = .62). Depressive symptoms were relatively infrequent in both later midlife (2001) and later years (2015; M = 1.45 and 1.42 for husbands, respectively; M = 1.56 and 1.60 for wives, respectively).

Table 1.

Univariate and Bivariate Statistics for Study Variables.

Variable Name 1 2 3 4 5 6 7 8
1. Husbands’ PWM (1991–2001) –
2. Wives’ PWM (1991–2001) .032 —
3. Husbands’ stressful pre-retirement work (2015) −.134 −.073 —
4. Wives’ stressful pre-retirement work (2015) .060 .124 .129 —
5. Husbands’ dep. symptoms (2001) .239** .078 .238** −.029 —
6. Wives’ dep. symptoms (2001) .094 .211** .001 .093 .162* —
7. Husbands’ dep. symptoms (2015) .174* −.022 .267** .197** .536** .206** —
8. Wives’ dep. symptoms (2015) .134 .169* −.013 .315** .213** .486** .396** —
M 2.25 2.24 .62 .97 1.45 1.56 1.42 1.60
SD .47 .50 1.07 1.34 .46 .51 .43 .55
Open in a new tab

Note. PWM = person work mismatch.

*

p < .05.

**

p < .01.

Correlations were in the expected directions. Husbands’ depressive symptoms were correlated over time (r = .54, p < .01) as were wives symptoms (r = .49, p < .01). Husbands’ and wives’ symptoms were also significantly correlated at each time point and over time (r ranging from .16 to .40). Husbands with higher average levels of PWM in their mid-later years (1991–2001) reported more depressive symptoms in later midlife (2001; r = .24, p < .05) and later years (2015; r = .17, p < .05). For wives, higher average levels of PWM in the mid-later years (1991–2001) were related to more depressive symptoms in later midlife (2001; r = .24, p < .05) but in later years (2015; r = .17, p < .05). The gender difference between PWM and depressive symptoms in later years was statistically significant (p < .05).

Testing the Hypothesized Model

Consistent with the hypothesized cumulative pathway (see Figure 2), the path analysis indicated husbands’ and wives’ PWM over a decade (four measurement occasions from 1991 to 2001) during their mid-later years was related to their depressive symptoms in later midlife (2001; B = .24 and .16, respectively, p < .05). In turn, depressive symptoms in later midlife (2001) were predictive of depressive symptoms in later years (2015; B = .43 and .57 for husbands and wives, respectively, p < .01).

Figure 2.

Figure 2.

Open in a new tab

Dyadic processes between person–work mismatch, stressful pre-retirement work context, and depressive symptoms from middle to later years.

Note. Unstandardized coefficients are shown with standard errors in parentheses. Lines and numbers in boldface represent statistically significant paths. Age significantly influenced wives’ stressful pre-retirement work context only (b = .07, p < .01). For both husbands and wives, the influence of retirement status on depressive symptoms in 2015 did not vary significantly by respondents’ age (age–retirement interaction). χ2(26) = 24.96; comparative fit index (CFI) = 1.0; Tucker–Lewis index (TLI) = 1.0; root mean square error of approximation (RMSEA) = .00.

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

Consistent with the hypothesized stress proliferation pathway, PWM over the mid-later years (1991–2001) was associated with individuals’ stressful pre-retirement work (assessed in 2015; B = .38 and .43 for husbands and wives, p < .05). In addition, depressive symptoms in later midlife (2001) also contributed to more stressful pre-retirement work (B = .43 and .41 for husbands and wives, p < .05). Even after considering consistency in depressive symptoms from later midlife (2001) to later years (2015), those experiencing more stressful pre-retirement work generally reported more depressive symptoms in later years (B = .06 and .09 for husbands and wives, p < .05).

Recall that the final model also included age and retirement status as control variables. For both husbands and wives, age was included as a control variable explaining variation in the endogenous variables. Age was significantly related to stressful pre-retirement work for wives only (B = .07, p < .05). Retired wives reported more depressive symptoms in 2015 than employed wives (B = .15, p < .05).

Partner effects.

Partner effects between husbands and wives were also estimated along with potential contagion effects, or dependencies, between husbands and wives. Husbands’ and wives’ depressive symptoms in 2001 were significantly correlated (r = .04, p < .05) as were their depressive symptoms in 2015 (r = .04, p < .01). Their reports of PWM were not related, nor were there reports of stressful pre-retirement work. Two partner effects were noted, including a partner effect from wives’ stressful pre-retirement work to their husbands’ depressive symptoms with husbands generally reporting more depressive symptoms in later years when their wives experienced more stressful pre-retirement work (B = .06, p < .05). The second partner effect was for husbands’ depressive symptoms in later midlife (2001), which influenced wives’ depressive symptoms in later years (2015; B = .20, p < .05). Overall, the model explained 34.2% of the variance in husbands’ depressive symptoms in later life and 44.0% of the variance in wives’ depressive symptoms in later life.

Indirect effects through stressful pre-retirement work.

The indirect effects from PWM to depressive symptoms in later years through the retirement context were estimated for husbands and wives. For husbands, the indirect effect from PWM to depressive symptoms in later years through stressful pre-retirement work was significant (B = .11, p < .05) suggesting stressful pre-retirement work mediated the influence of PWM on depressive symptoms in later years. For wives, the indirect effect from PWM to depressive symptoms in later years through stressful pre-retirement work was not significant (B = .05, p > .05) However, bivariate correlations shown in Table 1 provided evidence for the direct association between PWM and depressive symptoms in later years (r = .17 and .17, p < .05 for husbands and wives, respectively).

Discussion

Past research has been fragmented and has not adequately examined how work experiences and retirement circumstances of individuals contribute to husbands’ and wives’ mental health problems in later years. Specifically, less is known about how stressful work experiences, such as PWM. Thus, for a deeper understanding of this stress–health–life course process, the present study addressed three questions: (a) Does PWM in middle years persistently and cumulatively influence depressive symptoms over the life course? (b) Irrespective of this direct influence, does PWM in the middle years contribute to depressive symptoms in later years through the impact of PWM on stressful pre-retirement work? (c) Are there partner effects between husbands’ and wives’ PWM, stressful pre-retirement work, and depressive symptoms?

For both husbands and wives, average levels of PWM measured at four occasions over a decade in the middle years influenced depressive symptoms in the subsequent years, and depressive symptoms generally persisted into their later years. This is consistent with the cumulative pathway of life course model, which posits that the influence of early adversities accumulates over the life course (Glymour et al., 2010). The observed homotypic continuity of depressive symptoms from middle to later years is also consistent with psychopathology research suggesting that when stress exposure is repeated or chronic, emotional stress responses may increase the risk for further mental health problems (Ganzel et al., 2010). In addition to the direct manifestation of negative feelings, the mental health influence of PWM and associated stressful work adversities may operate through several psychosocial mechanisms, such as depleted psychological resources (Orth et al., 2008) and impaired social and family relationships (Menaghan & Parcel, 1990). The results suggest that the mental health of working spouses in later years can be protected by modifying or redirecting their depressive symptoms trajectories through early interventions.

High average levels of PWM over their mid-later years (1991–2001) also contributed to stressful pre-retirement work for husbands and wives in their later years (2015), which is consistent with the stress proliferation pathway also known as a “chain of risk” (Pearlin et al., 2005). This chain of risks may be attributed to both organizational changes as well as intra-individual processes stemming from PWM that may generate negative feelings about the work context surrounding retirement. The influence of PWM on stressful pre-retirement work can also be viewed as a cumulative (dis)advantage process whereby (dis) advantage increases over time representing an increasing social inequality over the life course (Dannefer, 2003). That is, individuals who experienced PWM were prone to experience further deterioration in their work conditions later in the life course. Previous research has documented poor work conditions, such as high demand and low control, influence involuntary or early retirement (Siegrist, 1996). Future research should investigate PWM as a determinant of early and involuntary retirement.

For both retirees and individuals approaching retirement, stressful pre-retirement work generally contributed to increased depressive symptoms in later years regardless of retirement status (i.e., retirement status was incorporated as a control variable). Retirement status influenced subsequent depressive symptoms in later years for wives, but not husbands, with fully or partially retired wives experiencing more depressive symptoms than employed wives. These results suggest that wives’ mental health is more affected by retirement than husbands after taking stressful pre-retirement work into account. Although previous research has examined mental health consequences of different forms of retirement (e.g., voluntary and involuntary, early and late), less is known about the unique mental health influence of a stressful pre-retirement work context independent of retirement status. The findings from the current study highlight the central role that stressful pre-retirement work plays in determining mental health in later years and suggests that the mental health of working spouses in later years can be protected by pre-retirement interventions that improve the context surrounding the transition from work to retirement, rather than postponing interventions until after retirement occurs.

Together, the results illustrated that life course cumulative and stress proliferation pathways operate simultaneously (Glymour et al., 2010), and the delineation of this web of direct and indirect influences advances our understanding of complex life course process.

Moreover, the life course framework emphasizes the “linked lives” of husbands and wives over time; that is, these pathways are parallel and intertwined for husbands and wives (Elder & Geile, 2009). This was supported in the dyadic analyses as partner effects and interdependencies between husbands and wives were found. First, the results identified the influence of husbands’ midlife depressive symptoms on wives’ depressive symptoms in later years, which is consistent with the notion that women tend to engage in interdependent self-construal more than men. That is, husbands’ characteristics, experiences, and emotions appear to impact wives’ well-being and mental health (Foels & Tomcho, 2009). Second, wives’ stressful pre-retirement work influenced husbands’ depressive symptoms in later years. This influence may operate through wives’ provision of decreased spousal support during their own adverse retirement context. This is consonant with research demonstrating that husbands receive, and benefit, more from spousal support than wives (Cutrona, 1996). It seems that these partner effects intensify hypothesized stress-health associations within the dyadic context. Third, husbands’ depressive symptoms were correlated with wives’ depressive symptoms both in their middle and later years, even after accounting for work and retirement stress as determinants of depressive symptoms. This provides evidence for the mental health congruence and synchrony of husbands and wives in enduring marriages (Kiecolt-Glaser, & Wilson, 2017). The utilization of a dyadic analysis also reduced parameter estimate bias compared to single-gender analyses (Kenny et al., 2006).

There are several limitations to the current study that should be noted. The first limitation relates to the temporal order of measurements. Because of data availability at specific measurement occasions, retirement context and depressive symptoms in later years were reported at the same occasion. Thus, reverse causation is a possibility as depressed individuals may be negatively biased in reporting their retirement context. Similarly, the measure of PWM may be biased. It is possible that initially depressed or otherwise-biased individuals reported a high mismatch and dislike their work due to their tendency to view the world unfavorably. In addition, items such as “I am over-qualified for this job” may be conflated with workers’ socioeconomic status (SES), which was not the study focus. We initially controlled for education as a proxy for SES, but more in-depth examinations may identify important differences in how PWM operates depending on SES. Second, the generalizability of the results may be limited due to the unique sample composition (i.e., European American, employed couples who remained married and lived in rural Iowa during the farm crisis of the 1980s). Findings may differ with study samples comprised of African and Latino adults who are more likely to be in lower status jobs and may view PWM differently. Future studies testing similar models with a more diverse population are needed. For instance, future samples should include multiple ethnicities, greater variation in occupations, and other geographic locations. Finally, the current study limited its investigation of mental health problems to depressive symptoms. It would be valuable to investigate other mental health problems, such as anxiety and hostility, as well as age-specific physical health problems, such as physical functioning, bodily pain, and sleep difficulties.

Despite these limitations, the present study contributes to the enhancement of knowledge about the influences of middle-aged couples’ PWM on their mental health in later years through their stressful pre-retirement work. Although previous findings on mental health effects have often conceptualized work stress and retirement stress as separate occurrences, these findings highlight their inter-connectedness. More specifically, the findings emphasize that as workers age, “work stress” often becomes “retirement stress” as a stressful work context for older adults nearing retirement can represent “push factors” that can prematurely drive workers toward retirement even when they are not financially, psychologically, or socially prepared for the transition. In this manner, stressful pre-retirement work may place older adults at risk for later mental health difficulties.

Consequently, this enhanced knowledge can prove useful for various audiences, including policymakers, practitioners, and those who develop and implement programs. These findings provide support for the value and necessity of national- and state-level policies aimed at improving work designs and working conditions particularly during the retirement period, noting that reducing PWM through employee training or continuing education, for example, can facilitate a more positive retirement context with long-term consequences far after the retirement event occurs. Particularly, considering how job requirements have been affected by technology advancements, it may be necessary to reinvest in human capital in work settings. Recent work interventions suggest that improvements to the psychosocial work environment are feasible (Gilbert-Ouimet et al., 2015). From these findings, mental health practitioners that work with older adults or couples can are encouraged to consider the role both partners’ previous work and retirement experiences may play in their mental health difficulties, regardless of current employment (or retirement) status. Although individuals often present for therapy or counseling related to work stress and even mental health difficulties without their partner, the findings highlight couple-level considerations, particularly for couples in long-term enduring marriages. Similarly, interventions and programs should not overlook stressful components of work over the life course as a potential contributor to mental health problems in later years. The results suggest that PWM and stressful pre-retirement work are important, modifiable targets for prevention and intervention efforts as efforts to reduce work stress, particularly PWM, and retirement stress can safe-guard individuals from the negative emotional consequences of stressful work experiences.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is currently supported by a grant from the National Institute on Aging (AG043599, Kandauda A. S. Wickrama, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies. Support for earlier years of the study also came from multiple sources, including the National Institute of Mental Health (MH00567, MH19734, MH43270, MH59355, MH62989, MH48165, MH051361), the National Institute on Drug Abuse (DA05347), the National Institute of Child Health and Human Development (HD027724, HD051746, HD047573, HD064687), the Bureau of Maternal and Child Health (MCJ-109572), and the MacArthur Foundation Research Network on Successful Adolescent Development Among Youth in High-Risk Settings.

Footnotes

Declaration of Conflicting Interests

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

References

  1. Adler NE, & Rehkopf DH (2008). US disparities in health: Descriptions, causes, and mechanisms. Annual Review of Public Health, 29, 235–252. [DOI] [PubMed] [Google Scholar]
  2. Barnes-Farrell JL (2003). Beyond health and wealth: Attitudinal and other influences on retirement decision-making. In Adams GA & Beehr TA (Eds.), Retirement: Reasons, processes, and results (pp. 159–187). Springer. [Google Scholar]
  3. Bender KA (2012). An analysis of well-being in retirement: The role of pensions, health, and “voluntariness” of retirement. The Journal of Socio-Economics, 41(4), 424–433. [Google Scholar]
  4. Bender KA, & Heywood JS (2011). Educational mismatch and the careers of scientists. Education Economics, 19, 253–274. [Google Scholar]
  5. Berscheid E, & Ammazzalorso H (2001). Emotional experience in close relationships. In Fletcher GJ & Clark MS (Eds.), Blackwell handbook of social psychology: Interpersonal process (pp. 308–330). Blackwell. [Google Scholar]
  6. Brown E (2009). Work, retirement, race, and health disparities. Annual Review of Gerontology and Geriatrics, 29(1), 233–249. [Google Scholar]
  7. Bryant V, Wickrama KAS, O’Neal CW, & Lorenz FO (2017). Family hostility and depressive symptoms in middle-aged couples: Moderating effect of marital integration. Journal of Family Psychology, 31(6), 765–774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Buchanan A, & Norko M (2011). The psychiatric report: Principles and practice of forensic writing. Cambridge University Press. [Google Scholar]
  9. Cacioppo JT, Fowler JH, & Christakis NA (2009). Alone in the crowd: The structure and spread of loneliness in a large social network. Journal of Personality and Social Psychology, 97(6), 977–991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carr E, Vahtera J, Goldberg M, Zins M, & Head J (2016). Occupational and educational inequalities in health-related exits from employment at older ages. European Journal of Public Health, 26(Suppl. 1), 142. [Google Scholar]
  11. Conger RD, & Elder GH Jr. (1994). Families in troubled times: Adapting to change in rural America. Gruyter Aldine de. Cutrona CE (1996). Social support in couples. SAGE. [Google Scholar]
  12. Dannefer D (2003). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social science theory. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 58, 327–337. [DOI] [PubMed] [Google Scholar]
  13. Deniz N, Noyan A, & Ertosun ÖG (2015). Linking person-job fit to job stress: The mediating effect of perceived person-organization fit. Procedia-Social and Behavioral Sciences, 207, 369–376. [Google Scholar]
  14. Derogatis LR, & Melisaratos N (1983). The brief symptom inventory: An introductory report. Psychological Medicine, 13(3), 595–605. [PubMed] [Google Scholar]
  15. Edwards JR, & Shipp AJ (2007). The relationship between person-environment fit and outcomes: An integrative theoretical framework. In Ostroff C & Judge TA (Eds.), Perspectives on organizational fit (pp. 209–258). Lawrence Erlbaum. [Google Scholar]
  16. Elder GH, & Giele JZ (2009). The craft of life course research. Guildford Press. [Google Scholar]
  17. Enders CK, & Bandalos DL (2001). The relative performance of full information likelihood estimation for missing data in structural equation models. Structural Equation Modeling, 8, 430–457. [Google Scholar]
  18. Fingerman KL, & Bermann E (2000). Applications of family systems theory to the study of adulthood. The International Journal of Aging and Human Development, 51(1), 5–29. [DOI] [PubMed] [Google Scholar]
  19. Foels R, & Tomcho TJ (2009). Gender differences in interdependent self-construals: It’s not the type of group, it’s the way you see it. Self and Identity, 8, 396–417. [Google Scholar]
  20. Ganzel BL, Morris PA, & Wethington E (2010). Allostasis and the human brain: Integrating models of stress from the social and life sciences. Psychological Review, 117(1), 134–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gilbert-Ouimet M, Baril-Gingras G, Cantin V, Leroux I, Vézina M, Trudel L, & Brisson C (2015). Changes implemented during a workplace psychosocial intervention and their consistency with intervention priorities. Journal of Occupational and Environmental Medicine, 57(3), 251–261. [DOI] [PubMed] [Google Scholar]
  22. Glymour MM, Errel KA, & Berkman L (2010). What can life-course epidemiology tell us about health inequalities in old age? In Antonucci TC & Jackson JS (Eds.), Annual review of gerentology and geriatrics (pp. 77–97). Springer. [Google Scholar]
  23. Gorin AA, Wing RR, Fava JL, Jakicic JM, Jeffery R, West DS, & DiLillo VG (2008). Weight loss treatment influences untreated spouses and the home environment: Evidence of a ripple effect. International Journal of Obesity, 32(11), 1678–1684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Griffin JM, Fuhrer R, Stansfeld SA, & Marmot M (2002). The importance of low control at work and home on depression and anxiety: Do these effects vary by gender and social class? Social Science & Medicine, 54, 783–798. [DOI] [PubMed] [Google Scholar]
  25. Henkens K, van Solinge H, & Gallo WT (2008). Effects of retirement voluntariness on changes in smoking, drinking and physical activity among Dutch older workers. European Journal of Public Health, 18, 644–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Iacovides A, Fountoulakis KN, Kaprinis S, & Kaprinis G (2003). The relationship between job stress, burnout and clinical depression. Journal of Affective Disorders, 75(3), 209–221. [DOI] [PubMed] [Google Scholar]
  27. Karasek RA, & Theorell T (1990). Healthy work. Basic Books. [Google Scholar]
  28. Kenny DA, Kashy DA, & Cook WL (2006). The analysis of dyadic data. Guilford. [Google Scholar]
  29. Kiecolt-Glaser JK, & Wilson SJ (2017). Lovesick: How couples’ relationships influence health. Annual Review of Clinical Psychology, 13, 421–443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kristof-Brown AL, Zimmerman RD, & Johnson EC (2005). Consequences of individuals fit at work: Meta-analysis of person-job, person-organization, person-group, and person-supervisor fit. Personnel Psychology, 58(2), 281–342. [Google Scholar]
  31. Kristof AL (1996). Person-organization fit: An integrative review of its conceptualizations, measurement, and implications. Personnel Psychology, 49, 1–49. [Google Scholar]
  32. Kuh D, Ben-Shlomo Y, & Power JH (2003). Life course epidemiology. Journal of Commmunityhealth, 57, 778–783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lee CC, Czaja SJ, & Sharit J (2008). Training older workers for technology-based employment. Educational Gerontology, 35, 15–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lovallo WR (2005). Stress and health: Biological and psychological interactions (2nd ed.). SAGE. [Google Scholar]
  35. Menaghan EG, & Parcel TL (1990). Parental employment and family life: Research in the 1980s. Journal of Marriage and Family, 52, 1079–1098. [Google Scholar]
  36. Muthén LK, & Muthén BO (2017). Mplus user’s guide (7th ed.). (Originally work published 1998) [Google Scholar]
  37. Myers DG, & Diener E (1995). Who is happy? Psychological Science, 6(1), 10–19. [Google Scholar]
  38. Negrini A, Panari C, Simbula S, & Alcover CM (2013). The push and pull factors related to early retirees’ mental health status: A comparative study between Italy and Spain. Journal of Work and Organizational Psychology, 29(2), 51–58. [Google Scholar]
  39. Ollendick TH, Shortt AL, & Sander JB (2005). Internalizing disorders of childhood and adolescence. In Maddux JE & Winstead BA (Eds.), Psychopathology: Foundations for a contemporary understanding (pp. 353–375). Lawrence Erlbaum. [Google Scholar]
  40. Orth U, Robins RW, & Roberts BW (2008). Low self-esteem prospectively predicts depression in adolescence and young adulthood. Journal of Personality and Social Psychology, 95(3), 695–708. [DOI] [PubMed] [Google Scholar]
  41. Pearlin LI, Schieman S, Fazio EM, & Meersman SC (2005). Stress, health, and the life course: Some conceptual perspectives. Journal of Health and Social Behavior, 46(2), 205–219. [DOI] [PubMed] [Google Scholar]
  42. Rakow A, Forehand R, Haker K, McKee LG, Champion JE, Potts J, Roberts L, & Compas BE (2010). The association of parental depressive symptoms with child internalizing problems: The role of parental guilt induction. Journal of Family Psychology, 25(1), 147–151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Scroggins WA, & Benson PG (2007). Self-concept-job fit: Expanding the person-job fit construct and implications for retention management. Research in Organizational Science, 2, 211–232. [Google Scholar]
  44. Shultz KS, Morton KR, & Weckerle JR (1998). The influence of push and pull factors on voluntary and involuntary early retirees’ retirement decision and adjustment. Journal of Vocational Behavior, 53, 45–57. [Google Scholar]
  45. Siegrist J (1996). Adverse health effects of high-effort/low-reward conditions. Journal of Occupational Health Psychology, 1(1), 27–41. [DOI] [PubMed] [Google Scholar]
  46. Sutin AR, Terracciano A, Milaneschi Y, An Y, Ferrucci L, & Zonderman AB (2013). The trajectory of depressive symptoms across the adult life span. JAMA Psychiatry, 70(8), 803–811. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Szinovacz ME, & Davey A (2004). Retirement transitions and spouse disability: Effects on depressive symptoms. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 59(6), S333–S342. [DOI] [PubMed] [Google Scholar]
  48. Westman M (2001). Stress and strain crossover. Human Relations, 54(6), 717–751. [Google Scholar]
  49. Wickrama KAS, Surjadi FF, Lorenz FO, & Elder GH Jr. (2008). The influence of work control trajectories on men’s mental and physical health during the middle years: Mediational role of personal control. Journal of Gerontology: Social Sciences, 63B, S135–S145. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES