Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Nov 27.
Published in final edited form as: J Occup Health Psychol. 2011 Apr;16(2):230–246. doi: 10.1037/a0022334

Psychological Well-Being in Retirement:The Effects of Personal and Gendered Contextual Resources

Bettina Kubicek 1, Christian Korunka 1, James M Raymo 2, Peter Hoonakker 3
PMCID: PMC6258024  NIHMSID: NIHMS992545  PMID: 21463050

Abstract

Although prior research points to the gendered nature of work and private routines, surprisingly few studies have explored the influence of gender on the sources of psychological well-being in retirement. Drawing on resource theories and theories on the gendered division of labor, this study examines how preretirement resources relate to retirees’ psychological well-being by using data from the Wisconsin Longitudinal Study. It is hypothesized that possessing key resources prior to retirement as well as losing or gaining resources in the transition to retirement influence retirees’ well-being and that these effects are partially conditioned by gender. Results indicate that preretirement physical health, tenacity in goal pursuit, and flexibility in goal adjustment are beneficial for men’s and women’s well-being alike. By contrast, financial assets and job dissatisfaction are more strongly related to men’s psychological well-being in retirement and preretirement social contacts to that of women. Thus, the study underscores the importance of considering gendered resources in retirement research.

Keywords: retirement, psychological well-being, preretirement resources, gender

Retirement constitutes a major transition in older workers’ lives. It ushers in a new stage in the life course, which requires the restructuring of daily routines and social contacts. Factors helping people to cope with these new requirements are personal, social, and financial resources (Szinovacz, 2003). Contingent on the resources available, people differ not only in their ability to adjust to retirement (Wang, 2007), but also in their attitudes toward retirement (Reitzes & Mutran, 2004) and in their well-being after retirement (Richardson & Kilty, 1991). When looking more thoroughly at the effects of resources on retirees’ well-being, two different perspectives can be distinguished. One focuses on the current context, asking whether and which resources and activities available to retirees promote their well-being (see Szinovacz, 2003, for an overview). This type of research showed that being healthy and wealthy in retirement (e.g., Kim & Moen, 2002; Richardson & Kilty, 1991; Szinovacz, 2003), being married, identifying with the marital role, and joint leisure activities of spouses during retirement (Price & Joo, 2005) all contribute to well-being in retirement. The second, though less-often taken perspective, focuses on past experiences, status, and roles as predictors of retirees’ well-being. Researchers following this perspective commonly pursue a life course approach (see, e.g., Elder, 1995; Moen, 1996). They showed that retirees’ former work role, their pre-retirement social background, and their preretirement self-esteem and self-efficacy all exert an influence on postretirement well-being (Esteban, Haverstick, & Sass, 2009; Reitzes & Mutran, 2004; van Solinge & Henkens, 2005, 2008). Yet, despite the fact that well-being in retirement is now more widely recognized as a phenomenon that is influenced by preretirement resources, there is still ambiguity regarding the nature of important resources in promoting retirees’ well-being. Specifically, personal resources, such as goal-directed behavior, have received limited attention in the retirement literature (Taylor-Carter & Cook, 1995). The effects of losing or gaining resources in the course of retirement also remain unclear.

Furthermore, gendered experiences throughout the life course are only partially reflected in current retirement research (Calasanti, 1996; Slevin & Wing-rove, 1995). Although an increasing number of studies have explored gender differences in subjective well-being, which suggested that women report slightly lower levels of well-being than do men (Pin-quart & Sörensen, 2001), far less research has examined how gender conditions the sources of well-being in retirement (see Calasanti, 1996; Kim & Moen, 2002; Quick & Moen, 1998, for exceptions). This means that the role gender plays in determining pre-retirement sources of retirees’ psychological well-being is still unclear.

The present study thus aims to narrow these two research gaps by investigating the relationship between psychological well-being in retirement and personal, social, and financial preretirement resources, and by examining how gender sets the context in which particular resources relate to retirees’ well-being using data from the Wisconsin Longitudinal Study (WLS). To give consideration to the multifaceted nature of psychological well-being (Diener, Suh, Lucas, & Smith, 1999), we examined two different outcome variables: positive psychological functioning (Ryff & Keyes, 1995) and depressive symptoms. We argue that by using positive and negative indicators of psychological well-being, broader and more diverse components of well-being are captured than by a single satisfaction measure.

Theories Linking Preretirement Resources and Psychological Well-Being in Retirement

An important factor common in studies on postretirement well-being is an emphasis on resources in the retirement transition and adaptation process (e.g., van Solinge & Henkens, 2005; Wang, 2007). In fact, numerous studies refer to the importance of an increasingly diverse set of resources for retirees’ well-being. Integrating these different approaches, resources can be defined as material, social, or personal characteristics or conditions that are valued by the individual or that are used as a means to achieve personal goals (see Hobfoll, 1989, for this definition of resources). For example, in the retirement context, close relationships or social group involvement may be valued goals in their own right, whereas socioeconomic status or the degree of goal pursuit may be the means to obtain valued goals, because they may help in maintaining previous lifestyles or in taking up new activities.

Given the significance ascribed to resources, it is, however, astonishing that resource theories only play a marginal role in theoretical considerations on post-retirement well-being. To reduce this discrepancy between, on the one hand, empirical findings pointing to the relevance of resources in the retirement process and on the other, the lack of theoretical foundations, a promising approach seems to be the combination of key resource theories (e.g., Carver & Scheier, 1999; Skinner, 1996) with conservation of resources theory (Hobfoll, 1989).

Resource theories often focus on one single resource for managing stressful situations and maintaining psychological well-being (see, e.g., Carver & Scheier, 1999; Skinner, 1996). These approaches can, according to Hobfoll (2002), be subsumed under the rubric “key resource theories” (p. 308), because they share the assumption that possessing one specific resource—be it social support, control, or optimism, to name but a few—enables people to immediately cope with a problem or grants access to what is needed in meeting situational requirements. Key resources might therefore be viewed as “management resources” (Thoits, 1994). Among the social and personal factors identified as key resources so far, social networks, material resources, and goal-directedness seem of primary importance in the retirement context.

Conservation of resources theory shares the assumption with key resource theories that resources are central to conditioning individual’s well-being. But instead of focusing on one specific resource, this theory looks at the processes in the general use of resources (Hobfoll, 2002). More precisely, conservation of resources theory proposes that individuals seek to obtain and preserve resources, and that well-being declines when resources are threatened, lost, or when an expected gain in resources after a significant resource investment is missing (Hobfoll, 2002). We can therefore identify two mechanisms whereby resources influence postretirement well-being: resource loss and resource gain.

Preretirement Resources and Gender

If resources help people achieve their personal goals, the importance of resources is likely to vary across individuals, depending on their personal strivings (Diener & Fujita, 1995). For example, individuals who gain joy from social activities and from sharing their spare time with friends and family are likely to regard their social network prior to retirement as a highly valued resource. Such individuals may concurrently perceive material resources as a less important source of well-being than do individuals who value social status and thus desire status goods. Consequently, we may expect the influence that resources have on psychological well-being to be contingent on the person’s own values and goals.

One important arena in which individuals’ world views, values, and goals are formed is that of their work experiences (Calasanti, 1996; Kohn & Schooler, 1983). Given the gendered allocation of occupations and industries, the job conditions for, and tasks engaged in by women and men differ greatly. Although men’s jobs in general involve more supervisory behavior, greater autonomy, and more task variety, women’s occupational mobility is often more restricted (Bottero, 2000) and they receive fewer intrinsic rewards from their work in terms of task diversity and challenge. Another important factor in shaping men’s and women’s world views and values is their divergent commitment to household and family tasks. In comparison with men, women experience the demands of caring for others, such as spouses, children, or elderly relatives more frequently. The gendered division of labor both on and off the job leads to different experiences for men and women. These are presumed to influence their identification of personal goals and the resources valuable to obtaining them. Following from this line of argument, we hypothesize that the relationship between preretirement resources and postretirement psychological well-being is partially conditioned by gender.

Key Preretirement Resources

Key resources are social networks, material resources, and goal-directed behavior; these help manage the requirements associated with the transition to retirement. Social networks can be viewed as interactions and ties among individuals that are based on different types of interdependency such as kinship, friendships, sexual relationships, or common interests. In addition to the number of people in a network, the quality of the relationships is also an important consideration. Granovetter (1973), for example, showed that the weak ties characterizing acquaintances are even more important for receiving information, news, and other resources than are the strong ties of close friendships. In order to capture this spectrum of social relations, we consider closeness to one’s spouse, social contact with friends and family, and wider social group involvement as con stituting an individual’s social network. We assume that social networks prior to retirement increase the likelihood of retirees receiving instrumental or emotional support during, and subsequent to, the transition process and that they facilitate the adoption of new roles and leisure activities. As a result they enhance postretirement well-being (e.g., Szinovacz, 1992). Given that cultural and social constructions of gender associate femininity with relatedness, the importance of social networks should differ for men and women. In line with this reasoning, Barnes and Parry (2004) found, in their interviews with female and male retirees, that women’s close relationships tend to be more intimate and supportive, and that loneliness was particularly problematic for women in retirement. Similarly, marital status and marital quality were shown to contribute only to women’s and not to men’s retirement well-being (Kim & Moen, 2002; Reitzes & Mutran, 2006). We therefore expect pre-retirement social networks to be of greater importance for psychological well-being in retirement for women than for men.

In addition to social networks, preretirement finances can be seen as a key resource in the retirement context, for they aid in acquiring or gaining other kinds of resources. This substitutive function of money may account for the consistently positive effect of income and wealth on postretirement well-being (Shultz, Morton, & Weckerle, 1998; Taylor & Doverspike, 2003; Taylor & Shore, 1995). Men, in concordance with the societal expectations tied to their status as breadwinner, may attach greater importance to finance than do women. Preretirement financial resources may therefore have a greater bearing on the psychological well-being of men than of women.

The proposition that key resources stimulate the development and use of other resources is not limited to financial or social factors but also holds for preretirement personal resources such as goal-directed behavior. Strategies supporting goal achievement as well as the ability to cope with failure are crucial to maintaining psychological well-being in retirement, for this life stage is associated with inevitable gains (e.g., freedom and leisure time; Rosenmayr, 1983) and losses (e.g., loss of the worker role, declining health), which have to be balanced. Two strategies for keeping such a balance have been identified (Brandsta¨dter & Renner, 1990). One tackles problems or losses by actively altering life circumstances in order to achieve personal aims. People who habitually use this strategy are likely to pursue their goals tenaciously despite hindrances. The other seeks to maintain balance by adjusting personal preferences to the life circumstances at hand. This coping strategy is characterized by flexible goal adjustment. People using it are likely to dismiss unreachable goals and to adapt their personal strivings when confronted with insurmountable obstacles. Both strategies, though almost independent of each other, seem to be crucial personal resources in determining the responses of individuals’ to change. In line with this argument, tenacious goal pursuit and flexible goal adjustment were both found to relate positively to optimism in general (Brandsta¨dter & Renner, 1990) and to life satisfaction and self-esteem in retirement (Trépanier, Lapierre, & Baillargeon, 2001). As retirement requires adopting new roles and setting new goals, we assume that flexibility in goal adjustment and persistence in goal pursuit are important resources for all retirees, regardless of gender.

  • Hypothesis 1a: Preretirement social networks, as indicated by social contacts, social group involvement, and closeness to one’s spouse, are more positively related to women’s psychological well-being in retirement than to men’s.

  • Hypothesis 1b: Preretirement financial resources are more positively related to men’s psychological well-being in retirement than to women’s.

  • Hypothesis 1c: Flexibility in goal adjustment and persistence in goal pursuit prior to retirement are positively related to men’s and women’s psychological well-being in retirement.

Resource Change in the Transition to Retirement

Resource Loss

On the basis of conservation of resources theory (Hobfoll, 1989), one could argue that significant declines in resources are associated with lower levels of psychological well-being in retirement. Losing one’s work role reflects such a loss, because one forfeits an important source of developing and holding a positive self-identity (George, 1993). Thus, retirees who considered their jobs as important or satisfying life domains may experience retirement as a loss. In support of this assumption, greater intrinsic job values were found to result in lower retirement satisfaction (van Solinge & Henkens, 2008), and bridge employment, which helps retirees maintain their work role, was found to relate to better physical and mental health in comparison with full retirement (Zhan, Wang, Liu, & Shultz, 2009). Therefore, those female and male workers who perceived their preretirement jobs as important or satisfying life domains are expected to report less psychological well-being in retirement.

Resource loss may be experienced not only if one abandons an important and satisfying job but also if physical resources are lacking. Poor health prior to retirement prevents retirees from successfully maintaining previous lifestyles and from establishing new activities in retirement, thus reducing their opportunities and well-being (e.g., Calasanti, 1996; Kim & Moen, 2002; Wong & Earl, 2009). This negative association between one’s own preretirement health status and psychological well-being should hold for both male and female retirees.

Spousal health problems may also relate to lower levels of well-being, be it because the partner’s impaired health restricts opportunities to take up new roles in retirement or because demanding caring responsibilities strain the relationship. One might even argue that because of persisting gendered expectations and the prevailing gendered division of labor among heterosexual couples (Gildemeister & Robert, 2008), men and women who retired within the past decade differ with regard to the effect of spousal health status on individual well-being (Allmendinger, 1990). Men of the pre-baby-boom cohort expected to be breadwinners and leave most household and caring responsibilities to their spouses, women expected to exit the labor force whenever caring for children, relatives, or spouses was required (Bottero, 2000; Crompton, 2006). Reflecting this gendered ethic of caring, women are more inclined to view caregiving for spouses as an obligation and tend to spend more time with caregiving tasks than do men (Spitze & Ward, 2000). Perceiving care as obligatory and investing longer care hours has, in turn, been linked to increased stress (Chappell & Reid, 2002). In line with this reasoning, the chronic condition of a spouse was shown to affect women’s but not men’s adaptation to retirement (Haug, Belgrave, & Jones, 1992). We therefore expect a spouse’s preretirement health problems to be a more important source of reduced psychological well-being for women than for men.

  • Hypothesis 2a: Female and male workers who attach greater importance to their preretirement jobs report less psychological well-being in retirement.

  • Hypothesis 2b: Female and male workers’ pre-retirement job satisfaction is negatively related to psychological well-being in retirement.

  • Hypothesis 2c: Poor health prior to retirement is negatively related to men’s and women’s psychological well-being in retirement.

  • Hypothesis 2d: Poor health of a spouse prior to retirement is more negatively related to women’s than to men’s psychological well-being in retirement.

Resource Gain

Aside from the negative effect of resource loss, it can be determined from conservation of resources theory that gaining resources in the transition to, or in the course of, retirement will be positively associated with higher levels of psychological well-being. People who gain resources by leaving unpleasant jobs may experience retirement as a relief and may benefit from abandoning the worker role (Adams, Prescher, Beehr, & Lepisto, 2002). Thus, we hypothesize that

  • Hypothesis 3: Male and female workers who perceived their preretirement jobs as monotonous report higher levels of psychological well-being in retirement.

Method

Data

To assess these hypotheses, we used data from the Wisconsin Longitudinal Study (WLS), a long-term cohort study of a random sample of 10,317 men and women who graduated from Wisconsin high schools in 1957. Since then graduates have been reinter-viewed in 1975, 1993, and 2004, providing extensive information on work histories, job characteristics, health complaints, and psychological well-being, among others. By 2004 the respondents, who all belong to the pre-baby-boom generation, were 65 years of age (SD = 0.66) on average.

The analyses presented here focus on those WLS individuals who completed both the telephone and the mail surveys in 1993 and in 2004 (N = 5528), who retired between 1993 and 2004 (N = 3388) and who had a spouse in 1993. The selection resulted in a reduced sample size of 2899 participants. The final analytical sample consisted of those WLS individuals who had no missing data on any of the study variables. Because of differences in the amount of missing data for the two outcome variables, this left 1728 cases for the analysis of positive psychological functioning and 1609 cases for the analysis of depressive symptoms. Comparing the full sample of married retirees with the respective analytical samples, we found only 2 out of 32 means to differ significantly. In particular, participants in the final sample for the analysis of depressive symptoms reported higher hourly wages and higher net assets, on average, than did participants of the full sample (Cohen’s ds = .11 and .03, respectively).

As a result of focusing on high school graduates, the WLS sample does not represent the U.S. population at large. Respondents tend to be better educated on average than is the general population. Furthermore—mirroring the demographic composition of Wisconsin in the 1950s—the large majority of respondents are White. Our results may therefore not be generalized across the entire population of similarly aged Americans. Yet, despite this limitation, the WLS is particularly well suited for examining psychological well-being in retirement. First and foremost, its long-term longitudinal design allows us to assess the relationship between preretirement factors and retirees’ psychological well-being. The utilization of theory-guided and well-established measures of psychological well-being constitutes the second strength of the WLS. Its third strength lies in the use of different data-gathering methods (telephone interviews and mail questionnaires). This reduces one source of common method bias, namely, common measurement context (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Finally, previous research on retirement preferences (Raymo & Sweeney, 2006) and emotional well-being in retirement (Coursolle, Sweeney, Raymo, & Ho, 2010) has demonstrated the adequacy of the WLS for retirement research.

Measures

Outcomes.

Given the multifaceted nature of psychological well-being and the presumably divergent consequences of retirement on positive and negative psychological functioning, we used two distinct indicators to measure psychological well-being. Our first indicator is a short version of Ryff’s (1989) psychological well-being scale, which comprises six subscales. Respondents indicated their level of positive psychological functioning by agreeing or disagreeing on a 6-point rating scale to questions pertaining to self-acceptance, personal growth, purpose in life, positive relations with others, environmental mastery, and autonomy. A sample item from the self-acceptance subscale reads as follows: “In general, I feel confident and positive about myself.” After recoding the 20 items that were enclosed in both the 1993 and the 2004 mail questionnaires so that higher values indicate better psychological functioning, confirmatory factor analyses were conducted. Their results supported the assumption that each of the six subscales belongs to a second-order well-being factor (see also Ryff & Keyes, 1995; results not shown). Therefore, it seemed reasonable to sum the ratings from the subscales and to form an overall psychological functioning score—as recently suggested by Springer, Hauser, and Freese (2006).

Our second indicator is the Center for Epidemio-logic Studies Depression Scale (CES–D; Radloff, 1977), which consists of 20 items to assess depressive symptoms. For each item, respondents indicated how many days in the past week (0–7) they had experienced the respective feelings. For example, respondents stated on how many days they felt bothered by things that usually do not bother them. Because the 8-point rating response format used in the WLS deviates from the original 4-point format, we recoded the answers to assure concordance with the original scale. The recoding seems feasible because the original response scale not only contains frequency adverbs but also specifies the corresponding number of days (e.g., “Most or all of the time (5–7 days)”; Radloff, 1977, p. 387). Hence, the number of days was grouped into categories of “0 days [H11005] less than 1 day,” “1–2 days,” “3–4 days,” and “5–7 days.” Responses were then summed into an overall depression score, with higher values indicating more frequent symptoms of depression.1

Predictor variables.

Predictor variables were drawn from the 1993 telephone and questionnaire survey. At that time, respondents were 53 years of age, on average, and approaching retirement. The coding algorithms and wordings of the predictor variables are presented in Table 1. As can be seen in Table 1, some variables are represented by single items. Although effort has been made to include multi-item measures, well-established measures of constructs were not always available, a limitation that has been described as a common disadvantage of using archival data (Kosloski, Ekerdt, & DeViney, 2001). Nonetheless, previous research points to the utility of the WLS for retirement research in particular (e.g., Coursolle et al., 2010) and for occupational health psychology in general (e.g., Warren, Hoonakker, Carayon, & Brand, 2004).

Table 1.

Coding Algorithms and Wording of Survey Questions for the Independent Variables

Variable Coding algorithm Source: Wording of question
Gender (female) Dummy variable: 1 = female, 0 = male
Time since retirement Century month (CM) of interview minus CM of exit from labor force Phone 2004
Closeness to spouse Metric variable ranging from 1 = not at all close to 4 = very close Phone 1993: How close would you say you are to your (husband or wife)?
Social contacts Metric variable ranging from 0 to 50; the variable was constructed by summing responses to two items concerning the frequency of social contacts. Mail 1993: How many times, if at all, during the past 4 weeks have you gotten together with friends? How many times, if at all, during the past 4 weeks have you gotten together socially with relatives?
Social group involvement Metric variable ranging from 0 = least involvement to 64 = strongest involvement; the variable was constructed by summing responses to 16 items asking to assess the extent to which respondents participate in various groups or clubs (0 = not involved, 4 = a great deal). Mail 1993: What is your level of involvement with church-connected groups, labor unions, veterans’ organizations, fraternal organizations or lodges, business or civic groups, parent-teachers’ associations, community centers, organizations of people of the same nationality, sport teams, country clubs, youth groups, professional groups, political clubs or organizations, neighborhood improvement organizations, charity or welfare organizations, hobby groups?
Net assets (ln $) To allow natural log transformation, a small start value ($1) was added for those reporting zero assets. Phone 1993: Respondent’s total asset net worth, in 1992 dollars.
Hourly wage (ln $) To allow natural log transformation, a small start value ($1) was added for those reporting wages of $0. Phone 1993: Base hourly wage rate, in 1992 dollars.
Flexible goal adjustment (Brandstadter & Renner, 1990) Metric variable ranging from 5 = least flexibility to 25 = highest flexibility; the variable was constructed by summing responses to five items. For each item, respondents were asked to assess the extent to which they agree or disagree (1 = strongly disagree, 5 = strongly agree) to the presented statement. Mail 1993: If I don’t get something I want, I take it with patience. It is very difficult for me to accept a setback or defeat.a I find it easy to see something positive even in a serious mishap. When everything seems to be going wrong, I can usually find a bright side to a situation. In general, I am not upset very long about an opportunity passed up.
Tenacious goal pursuit (Brandstadter & Renner, 1990) Multiple-item measure (ranging from 5 = least tenacity to 25 = highest tenacity); the variable was constructed by summing responses to five items. For each item, respondents were asked to assess the extent to which they agree or disagree (1 = strongly disagree, 5 = strongly agree) to the presented statement. Mail 1993: Even when things seem hopeless, I keep on fighting to reach my goals. I stick to my goals and projects even in the face of great adversity. The harder a goal is to achieve, the more appeal it has to me. I can be very stubborn in pursuing my goals. To avoid disappointment, I don’t set my goals too high.a
Importance of work Metric variable ranging from 1 = not at all important to 4 = very important Phone 1993: How important is your work to you?
Job satisfaction Metric variable ranging from 1 = very dissatisfied to 4 = very satisfied Phone 1993: All things considered, how satisfied are you with your job as a whole?
Poor health Metric variable ranging from 1 = excellent to 5 = very poor Mail 1993: How would you rate your health at the present time?
Spouse’s poor health Metric variable ranging from 1 = excellent to 5 = very poor Phone 1993: How would you describe your spouse’s health?
Monotonous work Metric variable ranging from 0 hr/wk to 96 hr/wk Phone 1993: How many hours per week do you do the same things over and over?
Open in a new tab

Note. Phone = telephone survey; Mail = mail-in questionnaire survey.

a

Items were recoded.

Control variables.

By 2004, respondents had spent divergent amounts of time in retirement; time elapsed since retirement was therefore included as a control variable, in addition to well-being in 1993.

Results

Means, standard deviations, internal consistencies, and correlations among study variables are presented in Table 2. As can be seen in Table 2, internal consistencies for multi-item measures, as indicated by Cronbach’s alpha, are satisfactory, with values higher than .65 (Nunnally, 1978).

Table 2.

Means, Standard Deviations, Correlations, and Internal Consistencies

Variable M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1. Time since retirement 59.14 34.68 —
2. Closeness to spouse 3.80 0.44 .03 —
3. Social contacts 6.45 4.97 .02 .03 —
4. Social group involvement 7.11 5.94 .04 .07 .16 —
5. Hourly wage 2.67 0.64 .04 .00 −.06 .14 —
6. Net assets 11.81 1.94 .09 .02 .04 .07 .24 —
7. Flexible goal adjustment 17.83 2.89 .02 .12 .09 .12 −.07 −.04 (.65)
8. Tenacious goal pursuit 18.06 2.99 .03 .11 .04 .26 .15 .08 .26 (.72)
9. Importance of work 3.82 0.42 −.09 .02 −.02 .05 .11 .00 .03 .06 —
10. Job satisfaction 3.46 0.67 −.07 .11 .06 .08 .05 .04 .16 .09 .16 —
11. Poor health 1.81 0.61 .01 −.06 −.01 −.08 −.10 −.08 −.15 −.13 −.00 −.13 —
12. Spouse’s poor health 1.63 0.71 .00 −.16 −.01 −.08 −.10 −.14 −.08 −.12 −.02 −.09 −.24 —
13. Monotonous work 20.18 16.52 −.01 .01 .01 −.15 −.25 −.12 −.04 −.10 .02 −.04 −.09 −.04 —
14. Depressive symptoms 1993 9.62 7.75 .01 −.21 −.10 −.07 −.10 −.07 −.32 −.24 −.05 −.21 −.27 −.18 .05 (.87)
15. Depressive symptoms 2004 7.68 7.07 .02 −.17 −.06 −.02 −.11 −.07 −.26 −.21 .00 −.14 −.24 −.15 .05 .59 (.86)
16. Psychological functioning 1993 4.92 0.64 .04 .20 .16 .26 .09 .08 .43 .52 .07 .26 .26 .20 −.11 −.57 −.41 (.88)
17. Psychological functioning 2004 4.86 0.63 .06 .19 .14 .22 .10 .08 .35 .45 .04 .17 .26 .18 −.14 −.43 −.50 .69 (.89)
Open in a new tab

Note. Correlations >|.05| are statistically significant at p < .05, correlations >|.07| are statistically significant at p < .01. N = 1532. Correlations are based on listwise deletion. Internal consistency estimates (Cronbach’s alpha) are presented in parentheses on the diagonal.

Gender Differences in Psychological Well-Being

Looking first at psychological well-being in general, we see that on average, retirees report moderate to high levels of positive psychological functioning and low levels of depressive symptoms (see Table 2). Sample averages for 1993 and 2004 indicate a decline in psychological functioning, t(1727) =−5.19, p < .01, d = 0.10, and in depressive symptoms from working life to retirement, t(1608) =−11.46, p < .01, d = 0.26. Mean-level analysis revealed signifi-cant, but small, gender differences, with women scoring slightly higher on psychological functioning, t(1726) = 3.54, p < .01, d = 0.17, Ms = 4.91, and4.81, respectively, and reporting more frequent symptoms of depression than did men, t(1607) =4.68, p < .01, d = 0.23, Ms = 8.57 and 6.92, respectively. With regard to psychological functioning, gender differences are attributable to differences in five of the six subscales, namely, environmental mastery, personal growth, positive relations with others, purpose in life, and autonomy. Whereas women score slightly higher on the first four scales, autonomy scores are higher among men.

General and Gendered Resources of Retirees’ Psychological Well-Being

The adequacy of our hypotheses was ascertained by estimating a set of hierarchical ordinary least squares (OLS) regressions. In a first step, well-being in 1993 was entered in the regression equation as a control variable. In a second step, time elapsed since retirement, gender and all predictor variables were included simultaneously in the regression model. We chose this procedure to evaluate whether preretirement resources contribute to psychological well-being in 2004 above and beyond the effect of psychological well-being in 1993. And indeed, the inclusion of preretirement resources yields significant improvements in model fit (see Tables 3 and 4). As a set of predictors, preretirement resources explain an additional 3% of the variance in positive psychological functioning and in depressive symptoms, respectively.

Table 3.

Regression Analyses Predicting Psychological Functioning in Retirement

General Women Men
Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Difference
t value
Variable B β B β B β B β B β B β
Constant 1.44 1.25 1.48 1.40 1.44 1.04
Control variables
 Psychological functioning 1993 0.69*** .70 0.56*** .56 0.69*** .71 0.52*** .54 0.69*** .68 0.58*** .57 −1.26
 Time since retirement 0.00 .01 0.00 −.01 0.00 .03 −1.22
 Gender 0.10*** .08
Key resources
 Social networks
  Closeness to spouse 0.07** .05 0.10** .07 0.04 .03 1.14
  Social contacts 0.00 .02 0.00 .02 0.00 .03 −0.27
  Social group involvement 0.00 .01 0.00 .03 0.00 .01 0.80
 Financial resources
  Hourly wage (ln $) 0.05** .06 0.08** .07 0.02 .02 1.42
  Net assets (ln $) 0.00 .01 −0.00 −.02 0.03* .06 −2.31*
 Personal resources
  Flexible goal adjustment 0.01** .06 0.01* .06 0.01* .06 −0.12
  Tenacious goal pursuit 0.02*** .11 0.03*** .14 0.02** .09 1.21
Resource loss
 Importance of work −0.02 −.01 −0.06 −.04 0.03 .02 −1.64
 Job satisfaction −0.02 −.02 0.02 .02 −0.05* −.06 1.96*
 Poor health −0.08*** −.08 −0.10*** −.10 − 0.07** −.07 0.89
 Spouse’s poor health −0.02 −.02 −0.02 −.03 −0.02 −.02 0.18
Resource gain
 Monotonous work −0.00* −.04 0.00 .00 −0.00** −.07 1.73
ΔF 1645.51*** 8.07*** 813.02*** 5.43*** 811.45*** 4.58***
dfs 1, 1726 14, 1712 1, 800 13, 787 1, 924 13, 911
ΔR2 .49 .03 .50 .04 .47 .03
Open in a new tab
*

p < .05.

**

p < .01.

***

p < .001.

Table 4.

Regression Analyses Predicting Depressive Symptoms in Retirement

General Women Men
Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Difference
t value
Variable B β B β B β B β B β B β
Constant 2.54 5.88 2.88 4.61 2.39 12.26
Control variables
 Depressive symptoms 1993 0.54*** .59 0.47*** .51 0.56*** .61 0.47*** .51 0.50*** .56 0.44*** .50 0.75
 Time since retirement 0.00 .02 0.00 .01 0.01 .03 −0.21
 Gender 1.19*** .08
Key resources
 Social networks
  Closeness to spouse −0.64 −.04 −0.76 −.04 −0.54 −.04 −0.32
  Social contacts −0.03 −.02 −0.09* −.06 0.04 .03 −2.30*
  Social group involvement 0.07** .06 −0.01 −.01 0.10*** .10 −2.06*
 Financial resources
  Hourly wage (ln $) −0.22 −.02 −0.36 −.03 −0.10 −.01 −0.48
  Net assets (ln $) −0.04 −.01 0.05 .02 −0.27* −.06 1.98*
 Personal resources
  Flexible goal adjustment −0.17** −.07 −0.07 −.03 −0.26*** −.12 1.75
  Tenacious goal pursuit −0.11* −.05 −0.08 −.03 −0.16* −.07 0.83
Resource loss
 Importance of work 0.74* .04 1.19* .07 0.09 .01 1.56
 Job satisfaction −0.06 −.01 −0.57 −.05 0.33 .03 −2.00*
 Poor health 0.85*** .07 1.27** .10 0.48 .04 −1.57
 Spouse’s poor health 0.32 .03 0.62 .06 0.01 .00 −1.45
Resource gain
 Monotonous work 0.00 .01 −0.02 −.04 0.01 .02 −1.37
ΔF 839.68*** 4.82*** 428.94*** 2.86*** 388.48*** 3.41***
dfs 1, 1607 14, 1593 1, 743 13, 730 1, 862 13, 849
ΔR2 .34 .03 .37 .03 .31 .03
Open in a new tab
*

p < .05.

**

p < .01.

***

p < .001.

In order to obtain evidence as to whether gender affects the context in which preretirement resources influence retirees’ psychological well-being, aggregated (gender invariant) regression models were estimated first. Then, gender-specific models were conducted. Comparisons in model fit between the general models and the gender-specific models were obtained using a Chow test (Chow, 1960). The significant Chow statistics—F(14, 1700) = 2.56; p < .05 for psychological functioning; F(14, 1581) = 3.17; p < .01, for depressive symptoms—suggest that the gender-specific models do indeed differ from the respective general models. These results lend initial support to the assumption that the influence of preretirement resources is contingent on gender. To assess whether specific regression coefficients differ among men and women, we calculated difference scores between the respective t values (von Eye & Schuster, 1998). Significant t value differences indicate that the respective resource is qualified by gender.2

Key preretirement resources.

We first examined whether key resources, which are assumed to stimulate the availability and use of other resources, foster psychological well-being in retirement, as is indicated by high levels of positive psychological functioning and low levels of depressive symptoms. Three such key resources were considered: social networks, finances, and goal-directed behavior. In contrast to our expectations, social networks predict psychological well-being in retirement only in part. Although social contacts and social group involvement are not related to psychological functioning (see Table 3), they do play a significant role in determining depressive symptoms in retirement (see Table 4). As is suggested by the significant difference terms, these associations are clearly qualified by gender, in that the effect of preretirement social contacts holds only for female retirees (η2 = .01), whereas the effect of preretirement social group involvement holds only for male retirees (η2 = .01). In congruence with our hypothesis, social contacts seem to be more important a resource for women than for men. However, contradicting our hypothesis, men who were more frequently engaged in organizational activities prior to retirement report higher levels of depressive symptoms in retirement. Being close to one’s spouse, our third social networks indicator, proves to be an important preretirement resource for women’s psychological functioning (η2 = .01), but is not associated with CES–D scores in 2004. Thus, Hypothesis 1a is only partially supported.

In conformity with our expectations, financial resources are more strongly related to men’s than to women’s psychological well-being in retirement. Although hourly wage is only positively associated with women’s psychological functioning in retirement, no significant gender difference was found. The effect of net assets on psychological functioning is, however, clearly qualified by gender, as is indicated by the significant t value difference. Thus, having higher assets at one’s disposal is conducive only to men’s psychological functioning (η2 = .01) but not to women’s. We were able to replicate this finding for depressive symptoms. Again, only for male retirees, having higher preretirement assets is associated with higher levels of well-being, that is, with lower levels of depressive symptoms (η2 = .01). These results lend support to Hypothesis 1b.

Flexibility in goal adjustment and tenacity in goal pursuit prior to retirement contribute, as was hypothesized, to women and men’s psychological functioning in retirement (η2 = .01 and .03 for women, and .01 for men). With regard to depressive symptoms, the beneficial effect of goal-directed behavior could only be replicated for male retirees (η2 = .02 for flexibility and η2 = .01 for tenacity). Thus, Hypothesis 1c is fully supported for positive psychological functioning, but only in part for depression.

Resource loss.

With regard to our second hypothesis, stating that resource loss over the course of retirement reduces psychological well-being, we found that leaving an important job is not related to psychological functioning but to depressive symptoms among female retirees (η2 = .01). The more importance women attached to their preretirement job, the more frequently they reported depressive symptoms in retirement. Although the regression coefficients for male retirees point in the same direction, they are not significantly different from zero. Thus, Hypotheses 2a is only partially supported.

The relationship between job satisfaction and psychological well-being in retirement is contingent on gender, as is indicated by the significant difference scores. Only for men, but not for women, leaving a satisfying job is detrimental to their psychological functioning (η2 = .01). This gender effect could also be replicated with regard to depression in retirement. Although the main effects are not statistically significant, the t value difference points to the gendered nature of leaving a satisfying job, with men showing a positive association, and women a negative association, between job satisfaction and depressive symptoms. Thus, Hypothesis 2b on the gender-invariant effect of leaving a satisfying job on psychological well-being in retirement is not supported.

Being worried by health problems does also predict psychological well-being in retirement. As we hypothesized, men and women with impaired health prior to retirement report lower levels of psychological functioning (η2 = .02 for women and .01 for men). In addition, the poorer women’s health was prior to retirement, the more frequently they reported depressive symptoms in retirement (η2 = .01). Thus, our hypothesis on the detrimental effect of resource loss over the course of retirement in terms of poor health is fully supported in the case of positive psychological functioning and partly supported in the case of depression.

By contrast, spouse’s preretirement health status is not related to psychological well-being in retirement, neither to positive psychological functioning nor to depressive symptoms. And the assumed gender-differentiated effect is not supported by the data.

Resource gain.

Turning to our third hypothesis, stating that resource gain over the course of retirement, as manifested in abandoning an unpleasant job, shows a positive relationship with psychological well-being in retirement, we found no empirical evidence. Contrary to our expectations, leaving a monotonous job does not relate to higher but to lower levels of psychological functioning among retired men (η2 = .01). In addition, leaving a monotonous job is not associated with depression in retirement. Thus, our hypothesis on the beneficial effect of gaining resources in the course of retirement because of abandoning an unpleasant job is not supported by the data.

Discussion

On the basis of resource theories and theories on the gendered division of domestic and paid labor, we derived hypotheses to assess general and gendered preretirement resources of psychological well-being in retirement. Resource theories propose that possessing key resources is crucial for retirees’ well-being and that changes in psychological well-being will occur if resources are lost or gained in the transition to retirement. Gendered experiences in the work and the family domains, on the other hand, suggest spouse’s health status and social networks are more important as sources of women’s psychological well-being, and financial resources are more important as a source of men’s psychological well-being in retirement. In summary, our results provided partial support for these hypotheses. We were able to show that personal resources—namely, goal-directed behavior—and resource loss, in terms of health problems, are related to positive psychological functioning, regardless of retirees’ gender. Furthermore, we found some resources to be contingent on gender. First, fewer preretirement social contacts increased depressive symptoms among female retirees but not among male retirees. Second, greater social group involvement prior to retirement showed a detrimental effect on men’s depressive symptoms in retirement, but not on women’s. Third, preretirement job satisfaction had opposing effects on men’s and women’s depressive symptoms and psychological functioning; men’s psychological well-being was negatively affected and women’s was positively affected. Finally, men’s, but not women’s, psychological well-being in retirement was associated with higher net assets prior to retirement.

Predicting Psychological Well-Being in Retirement

Our findings indicate that key resources, resource loss, and resource gain in the course of retirement contribute to the understanding of psychological well-being in retirement above and beyond the effect of preretirement well-being. Although most of the variance in the outcome variables was explained by the corresponding baseline well-being scores, the amount of variance explained by the resource variables is comparable to that in previous research findings (e.g., Coursolle et al., 2010; Zhan et al., 2009). It should further be noted that controlling for individual baseline well-being (i.e., psychological well-being prior to retirement) represents a rigorous test of the influence of preretirement resources on psychological well-being in retirement.

Looking at the effects of preretirement resources more thoroughly, we found that possessing key resources prior to retirement plays an important role in fostering well-being in retirement. Having personal, social, and to a lesser extent, financial resources prior to retirement contributed to retirees’ well-being. In contrast to earlier findings, showing financial considerations to be of particular importance in the retirement context, our study points to the dwindling influence of finances in the retirement process—as did a study by Gall and Evans (2000). Goal-directed behavior turned out to be a more important determinant of retirees’ positive psychological functioning and depressive symptoms than did net assets or hourly wage prior to retirement. Actually, flexibility in goal adjustment and tenacity in goal pursuit proved to be the most consistent predictors of well-being. These findings support prior evidence that retirees’ well-being is influenced by personal resources (e.g., Kim & Moen, 2002; Taylor-Carter & Cook, 1995; Trépanier et al., 2001; van Solinge & Henkens, 2005, 2008).

Contrary to our expectations, losing and gaining resources in the context of abandoning the worker role did not yield consistent results. On the one hand, job satisfaction prior to retirement showed opposing effects for male and female retirees. Whereas men’s psychological well-being in retirement was, as hypothesized, negatively associated with job satisfaction, women’s psychological well-being in retirement was positively associated with job satisfaction prior to retirement. This finding questions our assumption that abandoning a satisfying job in the transition to retirement always creates feelings of loss. Rather, female workers seem to appreciate the experiences made in working life and benefit from the esteem and approval associated with their former work roles. Therefore, in order to assess the influence of resource loss or gain, more information about the meaning people ascribe to these experiences is needed. Instead of placing emphasis solely on objective loss or gain, people’s appraisal and perception of resource changes are important aspects to be considered in future research (see, e.g., Lazarus & Folkman, 1984).

On the other hand, leaving a monotonous job was negatively (instead of positively) related to psychological functioning among male retirees. One viable explanation for the unexpected detrimental effect of preretirement job monotony on men’s well-being is that work experiences spill over to retirement. As was indicated primarily by research on the work–home interface, moods, skills, and behaviors associated with one domain can spill over to another (Grzywacz & Marks, 2000). Thus, experiences made in the work context, such as restricted task variety, can carry over to leisure activities and overshadow employees’ lives outside the workplace. Our results suggest that experiences during working life can even foster or impinge on well-being after leaving one’s job because of retirement. This is in concordance with Reitzes and Mutran’s (2006) finding that previous work roles continue to shape retirees’ identities. A second explanation for the unexpected finding is that positive or negative changes in well-being after retirement are only short-lived, as is suggested by Atchley’s (1976) phases of retirement adjustment. In that sense, the effects of resource gains or losses in the transition to retirement—in terms of job characteristics or attitudes toward one’s job—are only transient in nature. After phases of heightened or reduced well-being immediately after the transition, retirees’ adjustment to life in retirement may help them return to initial levels of well-being. By studying the effect of temporally somewhat distal experiences on well-being over the course of retirement, we were, however, not able to assess such short-term effects of job characteristics.

Psychological Well-Being in Retirement and Gender

An unexpected finding of our study is the contradictory gender difference in psychological well-being, with women, on average, reporting higher levels of positive psychological functioning and more depressive symptoms than men. According to a recent cross-national comparison of the prevalence of depression in 25 European countries, based on the CES–D scale, women’s tendency to report more complaints of depression are not attributable to measurement bias; rather, the authors discuss psychobiological and social conditions, ranging from role expectations and gender beliefs to gendered welfare state regimes, as causes of gender differences in depressive symptoms (Van de Velde, Bracke, Levecque, & Meuleman, 2010). One might expect that these results would also hold for the United States. Unfortunately, studies on the measurement invariance among subgroups do, to our knowledge, not exist for Ryff’s (1989) psychological well-being scale. Yet prior research supports the gender differences in psychological functioning found in this study (Lindfors, Berntsson, & Lundberg, 2006; Ryff, 1989; Ryff & Keyes, 1995). Now, how can the inconsistent gender differences in psychological well-being be explained other than by referring to measurement errors? The overall pattern of gender differences (women scoring higher on depressive symptoms, positive relations with others, purpose in life, environmental mastery, and personal growth; men scoring higher on autonomy) seems consistent with gendered expectations and may reflect unequal social and economic roles and opportunities. Relatedness and caring for others has been culturally constructed and coded as female throughout Western tradition, whereas carrying a sense of autonomy and agency “is decisively coded as male” (Bordo, 2003, p. 205).

In addition, the gender differences found suggest that positive psychological functioning and depressive symptoms are independent, though negatively related, indicators of psychological well-being, rather than the ends of a continuum. This reasoning is in line with evidence showing that positive and negative affect are not opposite poles but independent unipolar dimensions (see, e.g., Watson & Clark, 1997). And it highlights the importance of considering and comparing positive as well as negative outcome variables.

Another important issue raised in this study is the gendered pattern of resources. Divergences among men and women revealed in this study concerned financial, social, and job resources. Congruent with our expectations, men’s psychological well-being in retirement was more strongly related to financial resources than was women’s well-being (see Kim & Moen, 2002, for a similar pattern of findings), and depressive symptoms were more strongly associated with social contacts among women than those among men. Women’s stronger reliance on other people, together with men’s reliance on financial resources, reflect general patterns of gendered expectations among the pre-baby-boom generation. Women and men of this cohort show different patterns of lifetime employment and caring roles in the family, with men predominantly being breadwinners and women predominantly being responsible for most caring and household responsibilities. Moreover, men and women differ in their employment experiences (Calasanti, 1996) as well as in their friendships and social contacts (Barnes & Parry, 2004). These gendered experiences throughout the life course seem to translate into the found pattern of resources for men’s and women’s psychological well-being.

One finding that deviates from this general pattern is the positive association between men’s social group involvement prior to retirement and their levels of depressive symptoms in retirement. To understand this unexpected and at first sight counterintuitive result, we took a closer look at men’s and women’s social group involvement. What we found was that the level of involvement decreased more strongly among men than among women. Whereas both male and female retirees reported less involvement in professional or business groups, women were able to substitute these losses by increasing their engagement in community centers or charity groups. Men, by contrast, were not able to compensate for their losses. It may be that the men who reported high levels of involvement in 1993 experienced a particular decline in their social networks and were therefore more susceptible to depressive symptoms. This could have been due to decreased involvement in professional groups or sports teams, the former because of retirement; the latter because of impaired health.

Practical Implications

From a practical perspective, the current findings have implications for retirement preparation programs. As it is suggested that psychological functioning and depressive symptoms are independent dimensions of psychological well-being, they should be addressed separately. And the fact that they have partially different predictors means retirement preparation programs need to consider divergent sources of psychological well-being. In other words, reducing depressive symptoms and fostering positive psychological functioning may not be achievable by improving the same resources. In addition, retirement preparation programs should address potential predictors of psychological well-being beyond financial planning. Increasing prospective retirees’ knowledge of the beneficial effects of personal and situational resources, such as goal-directed behavior, and training them to use goal setting and goal pursuit in everyday life, should contribute to workers’ adjustment to retirement.

Finally, gender differences in the predictors of psychological well-being in retirement should be taken into account when developing retirement preparation programs. For example, it should be considered that male retirees are less likely than are their female counterparts to compensate for losses in social networks and that those with high levels of social group involvement are particularly inclined to face depressive symptoms in retirement. Therefore, apart from general information on finances, and general support in preparing future retirees for potential changes in daily routines and activities, gender-related issues could be addressed in homogeneous groups.

In summary, supporting people in the transition to retirement may help to reduce health costs in the long run—an important issue given the increasing prevalence of mental health problems such as depression in the Western world.

Study Limitations and Directions for Future Research

In evaluating the findings of the present study there are some limitations to be considered. Because the initial purpose of the WLS was not to assess the association between preretirement resources and retirees’ psychological well-being, the measures available from the dataset were limited. Future studies could include more sophisticated measures of the former work role and of resource gain in the transition to retirement. For example, we expect other stressful working conditions besides limited task variety to have potential implications for psychological well-being in retirement. Working long hours and experiencing time pressure during work may contribute to workers’ view of retirement as a relief. This assumption concurs with recent findings by Coursolle et al. (2010), showing that workers who previously experienced a spillover from stressful work experiences to family life do benefit from retirement and report higher levels of emotional well-being in retirement. Future research could be expected to elaborate further on the relationship between preretirement working conditions, resource gain in the course of retirement, and psychological well-being in retirement. In addition, although one important personality variable, namely, the degree of goal-oriented behavior, was included in the current study, future work could assess other personal resources of relevance to the retirement context, such as openness to change.

With regard to the generalizability of our results, one has to keep in mind that, on average, the respondents were better educated, and probably also better off in terms of their socioeconomic status, than was the general population. This fact may partly account for the marginal role financial resources played within our sample of pre–baby boomers.

These limitations to generalizability notwithstanding, our study suggests that well-being in retirement is embedded in prior personal as well as gendered contextual resources. In our view therefore, resource theories in combination with theories on the gendered division of domestic and paid labor herald the advent of promising theoretical approaches to the future study of retirees’ psychological well-being. In fact, looking more closely at gendered experiences in midlife seems warranted in order to better understand the gendered effects of preretirement resources on psychological well-being in retirement. However, whether the gendered patterning of resources found in this study of pre–baby boomers holds for future cohorts remains an open question. As the traditional gendered division of labor and the gender-specific expectations of femininity and masculinity lose importance, so too should gendered resources in this sphere.

Acknowledgments

The data cited from the Wisconsin Longitudinal Study are available to the public at http://www.ssc.wisc.edu/wlsresearch. This research was partly funded by National Institute on Aging grants No. R01-AG09775–10 and No. P01-AG21079–01 and was supported by Center grants from the National Institute of Child Health and Human Development (R24 HD047873) and the National Institute on Aging (P30 AG17266).

Footnotes

1

In order to assure our results, we conducted supplementary analyses with the 8-point rating format originally applied in the WLS. The results of these analyses were virtually the same as those presented in the following. The effect sizes in terms of adjusted R2 were either slightly lower or higher than those in the 4-point version, and the standardized regression weights changed only at the second decimal place.

2

We also conducted moderated multiple regression analyses using gender-related cross-product terms. For positive psychological functioning, these analyses showed that including the interaction between gender and each of the predictors did yield a marginally significant improvement in model fit, (ΔF(12, 1700) = 1.47, p = .13, ΔR2 = .01. This result may partly be due to the great number of interaction terms being included in the regression equation. Inspection of the regression coefficients revealed the effect of net assets on positive psychological functioning to be qualified by gender (β = − .09, p < .05). In accordance with results obtained by calculating difference values, the effect of job satisfaction was only marginally significant (β = .05, p = .06). For depressive symptoms, moderated regression analysis showed that the inclusion of interaction terms improved model fit, ΔF(12, 1581) = 2.55, p < .01, ΔR2 = .01. Single and joint testing of the interaction terms revealed social contacts (β = −.07, p < .05), social group involvement (β = −.06, p < .05), and job satisfaction (β = −.06, p = .05) to be qualified by gender. The effect of net assets was, however, only marginally significant (β = .07, p = .057). Given the fact that product terms are more prone to multi-collinearity than are difference values (Cortina, 1993; von Eye & Schuster, 1998) and that for interaction terms to be used, error variances need to be the same (Clogg, Petkova, & Haritou, 1995), calculating difference values of regression coefficients retrieved from separate models for men and women seems a more appropriate way of testing dichotomous moderators.

Contributor Information

James M. Raymo, Department of Sociology, University of Wisconsin—Madison

Peter Hoonakker, Center for Quality and Productivity Improvement, University of Wisconsin—Madison.

References

  1. Adams GA, Prescher J, Beehr TA, & Lepisto L (2002). Applying work-role attachment theory to retirement decision-making. International Journal of Aging and Human Development, 54, 125–137. [DOI] [PubMed] [Google Scholar]
  2. Allmendinger J (1990). Der Übergang in den Ruhestand von Ehepaaren: Auswirkungen individueller und familiärer Lebensläufe. Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft, 31, 272–303. [Google Scholar]
  3. Atchley RC (1976). The sociology of retirement. Cambridge, MA: Schenkman. [Google Scholar]
  4. Barnes H, & Parry J (2004). Renegotiating identity and relationships: Men and women’s adjustments to retirement. Ageing & Society, 24, 213–233. [Google Scholar]
  5. Bordo S (2003). Unbearable weight: Feminism, western culture, and the body. Berkeley, CA: University of California Press. [Google Scholar]
  6. Bottero W (2000). Gender and the labour market at the turn of the century: Complexity, ambiguity and change. Work, Employment and Society, 14, 781–791. [Google Scholar]
  7. Brandstädter J, & Renner G (1990). Tenacious goal pursuit and flexible goal adjustment: Explication and age-related analysis of assimilative and accommodative strategies of coping. Psychology and Aging, 5, 58–67. [DOI] [PubMed] [Google Scholar]
  8. Calasanti TM (1996). Gender and life satisfaction in retirement: An assessment of the male model. Journal of Gerontology: Social Sciences, 51B, S18–S29. [DOI] [PubMed] [Google Scholar]
  9. Carver CS, & Scheier MF (1999). Optimism In Snyder CR (Ed.), Coping. The psychology of what works (pp. 182–204). New York, NY: Oxford University Press. [Google Scholar]
  10. Chappell NL, & Reid CR (2002). Burden and well-being among caregivers: Examining the distinction. Gerontologist, 42, 772–780. [DOI] [PubMed] [Google Scholar]
  11. Chow GC (1960). Tests of equality between sets of coefficients in two linear regressions. Econometrica, 28, 591–605. [Google Scholar]
  12. Clogg CC, Petkova E, & Haritou A (1995). Statistical methods for comparing regression coefficients between models. American Journal of Sociology, 100, 1261–1293. [Google Scholar]
  13. Cortina JM (1993). Interaction, nonlinearity, and multi-collinearity: Implications for multiple regression. Journal of Management, 19, 915–922. [Google Scholar]
  14. Coursolle KM, Sweeney MM, Raymo JM, & Ho J-H (2010). The association between retirement and emotional well-being: Does prior work–family conflict matter? Journal of Gerontology: Social Sciences. doi:10.1093/geronb/gbp116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Crompton R (2006). Employment and the family The reconfiguration of work and family life in contemporary societies. New York, NY: Cambridge University Press. [Google Scholar]
  16. Diener E, & Fujita F (1995). Resources, personal strivings, and subjective well-being: A nomothetic and idio-graphic approach. Journal of Personality and Social Psychology, 68, 926–935. [DOI] [PubMed] [Google Scholar]
  17. Diener E, Suh EM, Lucas RE, & Smith HL (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276–302. [Google Scholar]
  18. Elder GH Jr. (1995). The life course paradigm: Social change and individual development In Moen P, Elder GH Jr, & Lüscher K (Eds.), Examining lives in context: Perspectives on the ecology of human development (pp. 101–139). Washington, DC: APA Press. [Google Scholar]
  19. Esteban C, Haverstick K, & Sass SA (2009). Gradual retirement, sense of control, and retirees’ happiness. Research on Aging, 31, 112–135. [Google Scholar]
  20. Gall TL, & Evans DR (2000). Preretirement expectations and the quality of life of male retirees in later retirement. Canadian Journal of Behavioural Sciences, 32, 187–197. [Google Scholar]
  21. George LK (1993). Sociological perspectives on life transitions. Annual Review of Sociology, 19, 353–373. [Google Scholar]
  22. Gildemeister R, & Robert G (2008). Geschlechterdifferenzierungen in lebenszeitlicher Perspektive Interaktion -Institution - Biographie. Wiesbaden, Germany: Verlag für Sozialwissenschaften. [Google Scholar]
  23. Granovetter MS (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380. [Google Scholar]
  24. Grzywacz JG, & Marks NF (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] [PubMed] [Google Scholar]
  25. Haug MR, Belgrave LL, & Jones S (1992). Partners’ health and retirement adaptation of women and their husbands. Journal of Women and Aging, 4, 5–29.29710449 [Google Scholar]
  26. Hobfoll SE (1989). Conservation of resources: A new attempt at conceptionalizing stress. American Psychologist, 44, 513–524. [DOI] [PubMed] [Google Scholar]
  27. Hobfoll SE (2002). Social and psychological resources and adaptation. Review of General Psychology, 6, 307–324. [Google Scholar]
  28. Kim JE, & Moen P (2002). Retirement transitions, gender and psychological well-being: A life-course, ecological model. Journal of Gerontology: Psychological Sciences, 57B, P212–P222. [DOI] [PubMed] [Google Scholar]
  29. Kohn ML, & Schooler C (1983). Stratification, occupation, and orientation In Kohn ML & Schooler C (Eds.), Work and personality (pp. 5–33). Norwood, NJ: Ablex. [Google Scholar]
  30. Kosloski K, Ekerdt DJ, & DeViney S (2001). The role of job-related rewards in retirement planning. Journal of Gerontology: Psychological and Social Sciences, 56B, P160–P169. [DOI] [PubMed] [Google Scholar]
  31. Lazarus RS, & Folkman S (1984). Stress, appraisal and coping. New York, NY: Springer. [Google Scholar]
  32. Lindfors P, Berntsson L, & Lundberg U (2006). Factor structure of Ryff’s psychological well-being scales in Swedish female and male white-collar workers. Personality and Individual Differences, 40, 1213–1222. [Google Scholar]
  33. Moen P (1996). A life course perspective on retirement, gender, and well-being. Journal of Occupational Health Psychology, 1, 131–144. [DOI] [PubMed] [Google Scholar]
  34. Nunnally JC (1978). Psychometric theory. New York,NY: McGraw-Hill Book. [Google Scholar]
  35. Pinquart M, & Sörensen S (2001). Gender differences in self-concept and psychological well-being in old age: A meta-analysis. Journal of Gerontology: Psychological Sciences, 56B, P195–P213. [DOI] [PubMed] [Google Scholar]
  36. Podsakoff PM, MacKenzie SB, Lee J-Y, & Podsakoff NP (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903. [DOI] [PubMed] [Google Scholar]
  37. Price CA, & Joo E (2005). Exploring the relationship between marital status and women’s retirement satisfaction. International Journal of Aging and Human Development, 61, 37–55. [DOI] [PubMed] [Google Scholar]
  38. Quick HE, & Moen P (1998). Gender, employment, and retirement quality: A life course approach to the differential experiences of men and women. Journal of Occupational Health Psychology, 3, 44–64. [PubMed] [Google Scholar]
  39. Radloff LS (1977). The CES–D scale: A self-report depression scale for research in general population. Applied Psychological Measurement, 1, 385–401. [Google Scholar]
  40. Raymo JM, & Sweeney MM (2006). Work–family conflict and retirement preferences. Journal of Gerontology: Social Sciences, 61B, S161–S169. [DOI] [PubMed] [Google Scholar]
  41. Reitzes DC, & Mutran EJ (2004). The transition to retirement: Stages and factors that influence retirement adjustment. International Journal of Aging and Human Development, 59, 63–84. [DOI] [PubMed] [Google Scholar]
  42. Reitzes DC, & Mutran EJ (2006). Lingering identities in retirement. Sociological Quarterly, 47, 333–359. [Google Scholar]
  43. Richardson V, & Kilty KM (1991). Adjustment to retirement: Continuity vs. discontinuity. International Journal of Aging and Human Development, 33, 151–169. [DOI] [PubMed] [Google Scholar]
  44. Rosenmayr L (1983). Die späte Freiheit Das Alter—Ein Stück bewußt gelebten Lebens. Berlin, Germany: Siedler. [Google Scholar]
  45. Ryff CD (1989). Happiness is everything, or is it? Explanations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57, 1069–1081. [Google Scholar]
  46. Ryff CD, & Keyes CLM (1995). The structure of psychological well-being revisited. Journal of Personality and Social Psychology, 69, 719–727. [DOI] [PubMed] [Google Scholar]
  47. Shultz KS, Morton KR, & Weckerle JR (1998). The influence of push and pull factors on voluntary early retirees’ retirement decision and adjustment. Journal of Vocational Behavior, 53, 45–57. [Google Scholar]
  48. Skinner EA (1996). A guide to the constructs of control. Journal of Personality and Social Psychology, 71, 549–570. [DOI] [PubMed] [Google Scholar]
  49. Slevin KF, & Wingrove CR (1995). Women in retirement: A review and critique of empirical research since 1976. Social Inquiry, 65, 1–21. [Google Scholar]
  50. Spitze G, & Ward R (2000). Gender, marriage, and expectations for personal care. Research on Aging, 22, 451–469. [Google Scholar]
  51. Springer KW, Hauser RM, & Freese J (2006). Bad news indeed for Ryff’s six-factor model of well-being. Social Science Research, 35, 1120–1131. [Google Scholar]
  52. Szinovacz ME (1992). Social activities and retirement adaptation. Variations by gender, marital status and household composition In Szinovacz ME, Ekerdt DJ, & Vinick BH (Eds.), Families and retirement (pp. 236–253). Newbury Park, CA: Sage. [Google Scholar]
  53. Szinovacz ME (2003). Contexts and pathways: Retirement as institution, process, and experience In Adams GA & Beehr TA (Eds.), Retirement: Reasons, processes, and results (pp. 6–52). New York, NY: Springer. [Google Scholar]
  54. Taylor MA, & Doverspike D (2003). Retirement planning and preparation In Adams GA & Beehr TA (Eds.), Retirement: Reasons, processes, and results (pp. 53–82). New York, NY: Springer. [Google Scholar]
  55. Taylor MA, & Shore LM (1995). Predictors of planned retirement age: An application of Beehr’s model. Psychology and Aging, 10, 76–83. [DOI] [PubMed] [Google Scholar]
  56. Taylor-Carter MA, & Cook K (1995). Adaptation to retirement: Role changes and psychological resources. Career Development Quarterly, 44, 67–83. [Google Scholar]
  57. Thoits PA (1994). Stressors and problem-solving: The individual as psychological activist. Journal of Health and Social Behavior, 35, 143–160. [PubMed] [Google Scholar]
  58. Trépanier L, Lapierre S, & Baillargeon J (2001). Ténacité et flexibilité dans la poursuite de projets personnels: Impact sur le bien-être à la retraite. Canadian Journal of Aging, 20, 557–576. [Google Scholar]
  59. Van de Velde S, Bracke P, Levecque K, & Meuleman B (2010). Gender differences in depression in 25 European countries after eliminating measurement bias in the CES–D8. Social Science Research, 39, 396–404. [Google Scholar]
  60. van Solinge H, & Henkens K (2005). Couples’ adjustment to retirement: A multi-actor panel study. Journal of Gerontology: Social Sciences, 60B, S11–S20. [DOI] [PubMed] [Google Scholar]
  61. van Solinge H, & Henkens K (2008). Adjustment to and satisfaction with retirement: Two of a kind? Psychology and Aging, 23, 422–434. [DOI] [PubMed] [Google Scholar]
  62. von Eye A, & Schuster C (1998). Regression analysis for social sciences. San Diego, CA: Academic Press. [Google Scholar]
  63. Wang M (2007). Profiling retirees in the retirement transition and adjustment process: Examining the longitudinal change patterns of retirees’ psychological well-being. Journal of Applied Psychology, 92, 455–474. [DOI] [PubMed] [Google Scholar]
  64. Warren JR, Hoonakker P, Carayon P, & Brand J (2004). Job characteristics as mediators in SES–health relationships. Social Science & Medicine, 59, 1367–1378. [DOI] [PubMed] [Google Scholar]
  65. Watson D, & Clark LA (1997). Measurement and mismeasurement of mood: Recurrent and emergent issues. Journal of Personality Assessment, 68, 267–296. [DOI] [PubMed] [Google Scholar]
  66. Wong JY, & Earl JK (2009). Towards an integrated model of individual, psychosocial, and organizational predictors of retirement adjustment. Journal of Vocational Behavior, 75, 1–13. [Google Scholar]
  67. Zhan Y, Wang M, Liu S, & Shultz K (2009). Bridge employment and retirees’ health: A longitudinal investigation. Journal of Occupational Health Psychology, 14, 374–389. [DOI] [PubMed] [Google Scholar]

RESOURCES