According to the texts discussion of gender and interaction quantity/quality,

Abstract

A persistent question in marriage research is whether men and women experience similar levels of marital satisfaction and whether similar components of marital interactions contribute to their global marital satisfaction (Jackson, Miller, Oka, & Henry, 2014; Spotts, Prescott, & Kendler, 2006). This is an important question because most studies conclude that women have lower levels of marital satisfaction than men and that marital appraisals have a more powerful effect on the health and well-being of older women versus men (Cohen, Geron, & Farchi, 2009; Faulkner, Davey, & Davey, 2005; Proulx, Helms, & Buehler, 2007). However, these conclusions presume that general measures of marital satisfaction, widely assessed in large-scale sample surveys, have the same meaning for men and women. This assumption invites an empirical investigation of the particular positive and negative marital interactions that are most central to men’s and women’s global marital satisfaction.

Gender differences in global marital satisfaction are fairly well established, with women typically rating their marriages less favorably than men (Jackson et al., 2014; Umberson & Williams, 2005; Windsor & Butterworth, 2010). For example, one analysis of nationally representative survey data showed that older women report lower levels of marital happiness and power than do their male counterparts (Bulanda, 2011). Similarly, results from laboratory-based observational studies of older adults reveal that wives tend to be more confrontational and direct in acknowledging marital problems, whereas husbands evade or ignore conflict (Carstensen, Gottman, & Levenson, 1995). The literature is less conclusive, however, regarding the specific positive (e.g., partner listens to me) and negative (e.g., partner makes me upset) aspects of marital interaction that may contribute to men’s and women’s responses on general marital satisfaction items. This is an important line of inquiry, given the well-documented linkage between global marital quality and health, especially for older women (Birditt, Newton, & Hope, 2014; Kiecolt-Glaser & Newton, 2001; Proulx et al., 2007). Identifying the specific aspects of marriage that contribute to global marital quality may help researchers and practitioners to target the specific types of marital interactions that may compromise (or enhance) the health and well-being of older spouses.

The few empirical studies that have examined gender differences in associations between specific marital appraisals and global marital satisfaction are inconclusive. Some have found striking similarity in what men and women consider critical to marital success (Lauer, Lauer, & Kerr, 1990), whereas others find more differences than similarities (Acitelli & Antonucci, 1994; Rhyne, 1981). However, these studies focused primarily on positive factors such as perceived social support (Acitelli & Antonucci, 1994) and satisfaction with marital sex, love, help around the house, and relationships with in-laws (Rhyne, 1981). Conflictual or critical interactions also may be particularly important for older spouses, given that health changes, decrements in cognitive functioning, caregiving, retirement, and other stressors associated with aging may create strains that compromise marital well-being, even among couples with historically strong relationships (Bookwala, 2005; Roberto, McCann, & Blieszner, 2013).

We seek to enhance our understanding of “his” and “her” marriage by examining gender differences in both positive and negative marital appraisals and in the associations between these specific appraisals and global evaluations of one’s marriage. Importantly, our assessments of positive and negative marital appraisals incorporate one’s perceptions of both how one treats and is treated by one’s spouse. Attention to one’s own contributions to the marital climate carries important implications for understanding gender differences in marital perceptions. We use data from the Changing Lives of Older Couples (CLOC) study, a large sample of married men and women aged 65 and older who have been married an average of 42 years.

Background

Gender Perspectives on Marriage: His and Hers Experiences?

That men and women experience marriage differently is a truism, dating back to the classic writings of Jessie Bernard (1973) and earlier. Contemporary writings, adopting a gender relations framework, suggest ways that gender shapes interpersonal relations, including marital interactions. These perspectives argue that gender is not a trait of an individual, per se, but rather it is a social relation shaped by socialization processes and power inequalities (Ferree, 2010). Even dyadic interactions between spouses are shaped directly and indirectly by social institutions, including economic and political systems that afford different privileges, opportunities, and resources to men and women. For example, cultural and structural support for the “male breadwinner” and “female caregiver” model through much of the 20th century created a context where women, especially those raised in the early half of the century, may be viewed as primarily responsible for the successful functioning of a marriage (Sullivan, 2006). Their male counterparts, by contrast, are socialized to specialize in paid employment, at least during one’s prime working years. Part of the traditional marital bargain is that bread-winning men have a reprieve from home responsibilities such as caregiving and instead contribute to the family by providing financially (Becker, 1991).

The gendered allocation of social roles has powerful implications for one’s expectations regarding men and women in general, and one’s spouse specifically. Two social psychological models have been proposed to describe the ways that gendered expectations shape social interactions. Role congruity theory suggests that members of a social group will be evaluated positively when their characteristics or behaviors are viewed as consistent with that group’s “typical” social roles (Eagly & Diekman, 2005). Similarly, expectation states theory is based on the assumption that individuals impute traits and skills to others on the basis of their group membership, such as gender (Ridgeway, Backor, Li, Tinkler, & Erickson, 2009). Although both theories were initially developed to understand gender roles and expectations in the workplace, they suggest ways that men and women may use gender as a frame for interpreting and responding to the words and actions of their spouses. Women may expect their spouses to be “strong and silent” and consequently may not register marital discontent when their own emotional needs go unfulfilled. Conversely, men who expect that their wife (and women, more generally) should provide emotional support and encouragement may enjoy particularly high levels of overall marital satisfaction when their wives provide them the support that they expect.

Gender-based expectations within marriage vary across social contexts, yet traditional expectations regarding “women as nurturers” were particularly common for members of the CLOC sample, most of whom were born in the 1920s, and who married and started families in the 1940s and 1950s. As such, we expect that specific aspects of marital functioning will be more strongly linked to global marital satisfaction among women than men. Because the nurturer role is particularly salient to women’s identities, their global marital satisfaction may be powerfully enhanced by positive yet undermined by negative interactions with their husbands.

For example, women raised in the early half of the 20th century were socialized to value communion, or the tendency to focus one’s attentions on significant others, whereas men were socialized to value and display agency, a trait that motivates a person to focus on self-directed actions and personal achievement (Helgeson, 1994). Women’s overall marital well-being may be more vulnerable to relationship stressors yet may be bolstered by rewarding relationships, given women’s tendency to derive their identity, in part, from the quality of their interpersonal relationships. At the same time, women’s high levels of investment in the nurturer role may also contribute to their husband’s greater satisfaction (Proulx et al., 2007). For example, the wife’s communion-oriented practices may enhance emotional intimacy and caregiving and thus promote overall marital happiness for their husbands (Bradbury & Fincham, 1988; Kiecolt-Glaser & Newton, 2001; Thompson, 1993).

Positive and Negative Characteristics of Marriage

Most research on marital quality reveals that even happily married persons hold both positive and negative feelings about their relationship (Fincham & Linfield, 1997; Rook 1998), and that formal conceptualizations and measures of marital satisfaction should recognize the contribution of positive and negative attributes (Kiecolt-Glaser & Newton, 2001). The relative balance of positive and negative traits varies widely across marriages, however. For example, Windsor and Butterworth (2010) detected four relationship types among older and midlife married adults: Supportive, aversive, ambivalent, and indifferent. However, most studies exploring positive and negative facets of marital interaction are limited to relatively minor negative exchanges, such as small disagreements or nagging (e.g., “How much tension is there between you and your partner?”; Windsor & Butterworth, 2010), yet do not encompass more overt negative exchanges (e.g., serious disagreements) that may not necessarily co-occur with positive attributes such as warmth, social support, and love (Newsom, Nishishiba, Morgan, & Rook, 2003). Given that more extreme aspects of marital discord are typically perpetuated by men (Archer, 2000), neglect of these more strongly worded dimensions of negative marital interactions may lead to an understatement in the gender gap in marital appraisals.

Prior studies generally show that the psychological benefits associated with positive characteristics of marriage are weaker than the psychological costs imposed by negative characteristics (Bookwala, 2005; Umberson, Williams, Powers, Liu, & Needham, 2006; Whisman, Uebelacker, Tolejko, Chatav, & McKelvie, 2006). Furthermore, particular combinations of positive and negative characteristics, ranging from strongly positive to strongly negative, may contribute to global marital satisfaction (Gottman & Notarius, 2000). Considering how combinations of specific marital characteristics naturally co-occur in individuals requires a person-centered approach, which can augment traditional, variable-level approaches. Person-centered approaches can identify different styles of evaluating marriage that individuals may have (i.e., profiles), and how these relate to global marital satisfaction in men and women. Variable-centered approaches, by contrast, analyze the sample as a whole and may fail to capture the nuanced experiences of specific groups within a particular sample (von Eye & Bergman 2003).

Study Aims

Our study had three aims. First, we evaluated gender differences in four positive and nine negative marital appraisals as well as global marital satisfaction. Second, we identified conceptually and statistically distinct profiles of marriages, based on the co-occurrence of specific positive (e.g., feeling loved, listened to, and supported) and negative appraisals of one’s relationship (e.g., frequency of conflict, criticism, and demands). Finally, we evaluated gender differences in the strength of association between each marital profile and one’s global marital satisfaction (after adjusting for potential confounds of health and depressive symptoms). This analysis allowed us to identify which specific aspects of marital interactions and combinations therein contribute to one’s overall level of marital satisfaction, and how their relative contributions differ by gender. Importantly, we considered specific marital items that capture perceptions of both one’s treatment of and treatment by one’s spouse. This is an important distinction which is typically neglected in research on marital quality. Recognizing the distinctive contributions to marital quality of both one’s treatment toward and of one’s spouse carries important implications for understanding gender differences in marital experiences.

Theoretical and empirical work suggests that women typically feel a greater responsibility to contribute to the emotional climate and stability of a marriage than do men (Beach, Katz, Kim, & Brody, 2003), and that men may expect women to do so (Eagly & Diekman, 2005; Ridgeway et al., 2009). Conversely, men may offer less emotional support and contribute less to the interpersonal climate of the marriage, especially cohorts of men socialized to prioritize work obligations and personal attributes such as achievement (Helgeson, 1994). Women’s appraisals of their spouses, consequently, may reflect expectations regarding what husbands can and should do (Eagly & Diekman, 2005).

In sum, drawing on gender theories outlined earlier, we expected that (H1) men will offer more favorable specific and global marital assessments than women. We also expected that (H2) both men and women will evaluate the wives’ behavior toward the husband more positively than the husband’s behavior towards the wife, a hypothesis consistent with prior writings on women’s communal orientation and gender-typed expectations. To test these hypotheses, we contrasted men’s and women’s responses on specific marital appraisal items and global marital satisfaction. We then identified statistically defensible profiles of positive and negative marital appraisals separately for men and women and documented associations between these profiles and global marital satisfaction. Our final objective was to evaluate the extent to which the marital profiles contribute to and explain the variance in men’s and women’s global marital assessments. We expected that (H3) specific marital profiles will be more strongly linked to global marital satisfaction among women than men, drawing on prior work regarding the salience of marriage to women’s identity and well-being.

Method

Sample

The CLOC study is a prospective study of a two-stage area probability sample of 1,532 married individuals from the Detroit SMSA (standardized metropolitan statistical area). To be eligible for the study, respondents had to be English-speaking members of a married couple in which the husband was aged 65 or older. All sample members were noninstitutionalized and capable of participating in a 2-hr long interview. Approximately 65% of those persons contacted for an interview participated, which is consistent with response rates from other Detroit area studies; response rates did not differ significantly by gender. Baseline face-to-face interviews were conducted from June 1987 through April 1988. Because the CLOC was designed as a prospective study of spousal loss, women were oversampled at baseline, to increase the likelihood that sample members would become bereaved during the study period (see Carr, 2006, for detailed description of CLOC study).

Although the CLOC is a multiwave study, the follow-up interviews target bereaved persons, and thus, the number of married persons interviewed at subsequent waves is too small for adequately powered analyses of gender differences in marital quality appraisals among currently married persons. Members of the baseline sample were randomly chosen to be interviewed as either the sole reporter on their marriage (n = 686) or as part of a marital dyad where each would independently participate in the CLOC interview (n = 846 persons or 423 couples). Our analytic sample included the 686 individuals who took part in the study without their partner and one randomly selected partner from each of the 423 couples, resulting in a total sample of 1,110 participants (263 men and 847 women). We focus on just one partner per marital dyad to eliminate dependencies between partner responses. The selected study sample did not differ in any meaningful or patterned ways from the rest of the sample. The average age in the sample was 69.5 years old (SD = 6.43), and the average marital duration was 42 years (SD = 12.48). Nearly all were in their first marriage, 15% of the sample was African American, and the mean level of educational attainment was 11.5 years. Slightly more than 90% owned their homes at baseline. All analyses are weighted to adjust for sample design, including the oversampling of women at baseline.

Measures

The CLOC assessed marital quality using items from a widely used inventory of positive and negative marital interactions (Schuster, Kessler, & Aseltine, 1990).

Global marital satisfaction.

We created a two-item measure of global marital satisfaction (α = .84) by summing responses to the questions: “How happy are you with your marriage?” and “How satisfied are you with your marriage.” The five response categories ranged from 1 (not at all) to 5 (completely).

Positive and negative marital appraisals.

We measured 13 positive and negative marital characteristics which capture both emotional responses to and cognitive appraisals of the marriage (Glenn, 2003). Respondents were asked how frequent (ranging from (1) never to (5) always) or how true (ranging from (1) not at all to (4) very true) a specific description would be for their marriage. Positive characteristics (α = .77) comprised four items: (a) Spouse loves/cares for me, (b) spouse is willing to listen to me, (c) I make spouse feel loved, and (d) I am willing to listen to spouse. Negative characteristics (α = .70) comprised nine items: (a) I am upset with marriage, (b) spouse is critical of me, (c) I am critical of spouse, (d) disagreements/conflicts in marriage, (e) I feel close yet sometimes upset, (f) spouse makes too many demands, (g) serious difficulties in marriage, (h) spouse does not treat me well, and (i) I don’t treat spouse well.

Control variables.

We controlled for depressed affect and poor self-rated health because each is associated with making more negative appraisals of one’s self and one’s marriage (Carr & Boerner, 2009). Women are more susceptible than men to depressive symptoms (Takkinin et al., 2004), which could contribute to more critical appraisals. Depressive symptoms (α = .83) were assessed at baseline with a subset of nine negative items (e.g., felt depressed, felt sad) from the 20-item Center for Epidemiologic Studies depression (CES-D) scale (Radloff, 1977). Respondents indicated how often they experienced each symptom in the week prior to interview. Response categories were hardly ever, some of the time, or most of the time. Physical health was assessed with the question: “How would you rate your health at the present time? Would you say it is excellent, very good, good, fair, or poor?” Responses of “fair” or “poor” are coded as 1. We also controlled for age (in years).

Analytic Plan

We first evaluated mean-level gender differences using t-tests. Given that the subgroups of men and women differed substantially in size (i.e., 263 men vs. 847 women), and given poor robustness of t-tests with very different group sizes, we tested the statistical significance of gender differences using t′, which assumes lack of homogeneity of variance. We adjusted the alpha level using a Bonferroni correction (p < .05/14 = .004) to prevent alpha-error inflation due to multiple testing. To identify frequent combinations of specific marital characteristics, we used an innovative method called profile analysis via principal component analysis (PAPCA; Kim, 2010). Although PAPCA uses a principal component analysis (PCA) as a tool to identify components of interest, this approach is different from PCA because its main goal is classifying individuals, rather than variables, by interpreting arrays of individuals’ response patterns as latent profile patterns (most typical patterns of participants’ item responses).

The strongest profiles (i.e., explaining most of the variance in the 13 marital characteristics variables) indicate the combination of positive and/or negative items that occur most frequently in the sample and can be interpreted as the most typical profiles. The criteria used to determine the number of profiles considered as meaningful are identical to those used in PCA analysis (i.e., eigenvalue, explained variance, and interpretability). The correlations of each individual variable with each profile can be used to interpret the profile. We consider an item as characteristic for a certain profile if it has a minimum correlation with the profile of .35 (absolute value) because .352 = .12 meaning that the item is accounting for about 12% of the variance in a profile. We used this conservative criterion to ensure that our exploratory approach may be replicated in other data sets (10% explained variance is typically considered sufficient to be seen as meaningful; Sosinsky & Kim, 2013). Interpretation of the profiles is based on the correlations of the single items with each profile. Each profile’s pattern of characteristic variables can be visually represented as “peaks” and “valleys” and used to label the profile. A positive correlation of an individual variable with the profile indicates a positive relationship, and a negative value suggests a negative relationship between the item and the latent profile.

PAPCA has important advantages relative to cluster analysis. Specifically, individuals are not forced into a cluster when their characteristics do not fit completely to one cluster or if they show a mix of characteristics that makes them similar to more than one cluster. This is the case because person weights extracted from the PCA indicate the degree of closeness between each individual’s observed variable pattern and the latent profile, and do therefore not classify individuals into distinct categories, but indicate the degree of relatedness to each profile. A person’s observed score profile pattern can be most closely related to one core profile (high person weights), yet can also possess features associated with the other profiles (lower person weights). That is, an individual’s responses can vary around each of several prototypical persons. Strong positive (>1.0 in the z-score metric) person weights indicate a strong match between a person’s observed variables and a specific latent profile, and a strong negative (<−1.0) person weight indicates a strong match between the person’s observed variables and the mirror-image latent profile. Because latent profiles are orthogonal, the characteristics of an individual can be accounted for by a single profile or by linear combinations of some of the profiles.

We identified profiles of specific marital characteristics for men and women and then explored their level of overlap (or profile pattern similarity) by computing correlations between male and female profiles. A significant correlation indicates that the profiles do not differ substantially from each other. Profile person weights determine the extent to which a specific profile accounted for interindividual differences in general marital satisfaction for men and women. The PAPCA and subsequent regression analyses allow us to identify whether men and women evidence similar profiles of positive and negative interactions within their marriages, and the extent to which each marital profile is associated with global appraisal of marital satisfaction. To further confirm whether the predictive validity of the profiles differed between men and women, we conducted an additional regression analysis testing gender × profiles interactions.

Results

Gender Differences in Mean Levels of Specific Marital Appraisals and Global Marital Satisfaction

Table 1 presents gender comparisons for all marital quality assessments, including specific appraisals and global marital satisfaction. We detected a greater number of differences than similarities. As hypothesized, men offered more positive marital assessments than women, yielding higher scores on positive items and lower scores on negative items. Also consistent with our predictions, men offered consistently more positive assessments on items referring to spouse’s treatment of them, rather than their treatment of the spouse. Men reported significantly higher scores than women on indicators such as “My spouse loves and cares for me” (Mmale = 4.56 vs. Mfemale = 4.28) and “My spouse is willing to listen to me,” (Mmale = 4.02 vs. Mfemale = 3.62), and significantly higher scores on items reflecting their negative treatment of the spouse such as “I don’t treat my spouse well” (Mmale = 1.95 vs. Mfemale = 1.63, respectively). These gender gaps persisted net of age, depressive symptoms, and health.

Table 1.

Mean Levels of Global Marital Satisfaction and Specific Characteristics of Marital Quality in Total Sample and by Gender

Notes. We estimated two-tailed t-tests to evaluate significant gender differences. Given poor robustness of t-tests with very different group sizes, we used t′ assuming lack of homogeneity of variance; control analysis was tested with general linear model (GLM) controlling for age, depressive symptoms, and self-rated health (df = 1). Adjusted alpha correction after Bonferroni (p = .05/14 = .004). Bold values highlight significant findings.

Table 1.

Mean Levels of Global Marital Satisfaction and Specific Characteristics of Marital Quality in Total Sample and by Gender

Notes. We estimated two-tailed t-tests to evaluate significant gender differences. Given poor robustness of t-tests with very different group sizes, we used t′ assuming lack of homogeneity of variance; control analysis was tested with general linear model (GLM) controlling for age, depressive symptoms, and self-rated health (df = 1). Adjusted alpha correction after Bonferroni (p = .05/14 = .004). Bold values highlight significant findings.

Profiles of Specific Marital Characteristics

We next identified the distinctive co-occurring combinations of 13 positive and negative marital characteristics for men and women, using PAPCA (Kim, 2010; see Method section for more detail). We identified three core profiles with eigenvalues over or near 1.00, explaining more than half of the variance in the 13 marital items; 54.2% and 58.2% for men and women, respectively. Figure 1 displays these profiles for men and women by plotting the correlations between each marital characteristic and each profile.

Figure 1.

Profiles of specific marital characteristics. Dotted line indicates which specific characteristic is considered a marker for the profile (r > |.35|).

The three profiles were easily interpretable: The Positive profile (top panel) was characterized by high values on all positive items and very low values on the negative items. Men’s and women’s response patterns were nearly identical and yielded a near perfect correlation (r = .99). The Positive–Negative profile (middle panel) characterized persons whose marriages encompassed positive characteristics such as feeling cared for, as well as tensions (e.g., being critical of each other). We label this profile “Positive–Negative” to reflect the coexistence of strain and support. Men’s and women’s Positive–Negative profiles overlapped considerably (r = .76). Both men and women had high values on “spouse loves/cares for me,” and on negative indicators including “disagreement/conflict in marriage.” However, men scored noticeably higher than women on the item reflecting positive treatment by rather than treatment of one’s partner (“spouse makes me feel loved and cared for”).

The Negative profile (bottom panel) was characterized by high values on two items indicating the perception of poor treatment (“spouse does not treat me well” and “I don’t treat spouse well”) and low endorsements of positive items (with the exception of “I am willing to listen to spouse” reported by women). The correlation between men and women was somewhat weaker for the Negative profile than the other two profiles (r = .62). We detected notable gender differences with respect to the most influential markers of the profile (i.e., those that correlated at least .35 with the profile). Specifically, men had characteristically low values on the items indicating demands from and criticism by spouse (i.e., “spouse makes too many demands” and “spouse is critical of me”) and own criticism (“I am critical of my spouse”), and relatively high values on “I feel close yet sometimes upset.” Women, by contrast, indicated high values on the item highly characteristic of communality (“I am willing to listen to my spouse”).

We next determined the percentage of cases with strong person weights (larger than or equal to |1.00|) on one specific core profile to further illustrate the profile results at the individual level. (A person weight indicates one’s degree of closeness to each of the three profiles.) A total of 484 (57%) women and 159 (60%) men had substantial person weights on the Positive profile, 267 (32%) women and 87 (33%) men on the Positive–Negative profile, and 223 (26%) women and 77 (29%) men on the Negative profile. That these percentages add up to 115% (women) and 122% (men) indicate that a small proportion of individuals had strong person weights on more than one profile. A traditional cluster analysis approach would have required us to force these individuals into a single cluster, thereby capturing only a portion of their marital experience.

To provide further support for the three profiles generated here, we also replicated all analyses using the marital partner in each dyad who was not randomly selected for inclusion in the analytic sample used in this study. The replication analysis yielded the exact same three profiles (complete results available from authors).

Linkages Between Profiles and Global Marital Satisfaction

We next investigated the extent to which the three profiles relate to global marital satisfaction. We estimated separate hierarchical regression models for men and women, using the person weights of each profile (those of men and women, respectively) as predictors of global marital satisfaction. We also controlled for age, depressive symptoms, and self-rated health (Table 2). The first regression model for men explained 57% of individual differences in global marital satisfaction, which was mostly attributable to the three profiles explaining 42% of independent (i.e., unique) variance. Age (B = 0.02, ns), depressive symptoms (B = −0.12, p = .05), and self-rated health (B = −0.05, ns) contributed little to explaining variance in the outcome. All profile variables had statistically significant beta weights. As expected, the Positive profile was the strongest predictor of global marital satisfaction (B = 0.81, p < .001), explaining 37% of unique variance above and beyond the other two profiles and control variables.

Table 2.

Regression Analysis Predicting Global Marital Satisfaction With Profiles

Notes. Runique2: Variance explained by specific predictor variable independent from all other predictors. Control variables included in both models were age (men: B = 0.02, SE = 0.01, β = .07, p = .12; women: B = 0.00, SE = 0.01, β = −.01, p = .59), depressive symptoms (men: B = −0.12, SE = 0.06, β = −.09, p = .05; women: B = −0.10, SE = 0.04, β = −.07, p = .007), subjective health (men: B = −0.05, SE = 0.05, β = −.04, p = .34; women: B = 0.01, SE = 0.03, β = .01, p = .74).

Table 2.

Regression Analysis Predicting Global Marital Satisfaction With Profiles

Notes. Runique2: Variance explained by specific predictor variable independent from all other predictors. Control variables included in both models were age (men: B = 0.02, SE = 0.01, β = .07, p = .12; women: B = 0.00, SE = 0.01, β = −.01, p = .59), depressive symptoms (men: B = −0.12, SE = 0.06, β = −.09, p = .05; women: B = −0.10, SE = 0.04, β = −.07, p = .007), subjective health (men: B = −0.05, SE = 0.05, β = −.04, p = .34; women: B = 0.01, SE = 0.03, β = .01, p = .74).

The Positive–Negative profile also had a positive B value (B = 0.29, p < .001) and explained 5% of unique variance. Higher person weights on men’s Positive–Negative profiles were associated with higher marital satisfaction, whereas a Negative Profile was associated with significantly lower global marital satisfaction (B = −0.14, p < .007). Although statistically significant, this effect was relatively weak, explaining only 1% of unique variance.

A second identical regression was conducted with the female participants. The overall amount of variance in marital satisfaction explained by the profiles was slightly greater (63%) among women than men. Among women, the three profiles explained 50% of independent variance in global marital satisfaction versus just 42% among men. Depressive symptoms had a small but statistically significant effect (B = −0.10, p < .01), yet age and health were not significant predictors. The Positive profile had a high positive B value (B = 1.17, p < .001) and explained the greatest amount of unique variance (48%). The explained variance was substantially higher for women relative to men (48% vs. 37%). The Positive–Negative profile had a positive B value (B = 0.26, p < .001) but explained less unique variance (2%). The Negative profile was not a significant predictor of global marital satisfaction among women (B = −0.05, ns), even though it had a small but statistically significant effect among men.

To confirm differences in the predictive power of the profiles between men and women, we conducted an additional regression analysis with the full sample, including gender, all profile variables, as well as gender by profiles interaction terms (plus all control variables). We found that the interaction between gender and the Positive profile was significant (β = −.40, p < .001). Thus, although the Positive profile was a predictor in both genders, it was a significantly stronger predictor of marital satisfaction among women. For the Positive–Negative and the Negative profiles, this regression analysis did not provide further evidence for a differential predictive value (p = .60 and p = .18, respectively).

Discussion

Our overarching goal was to develop a multifaceted understanding of marriage in later life. We evaluated gender differences in older adults’ endorsement of specific positive and negative aspects of marriage, encompassing items that reflected both how one treats and is treated by one’s spouse. We then identified the extent to which men’s and women’s perceptions of specific positive and negative interactions related to their overall assessments of global marital satisfaction. We used innovative statistical methods to develop distinctive profiles of marital characteristics and evaluated the relative explanatory power of these profiles for understanding global marital satisfaction. We briefly summarize our key findings and discuss their implications for understanding gender differences in late-life marital processes.

Gendered Differences in Later-Life Marriage

Consistent with most prior studies of older adults (Bulanda, 2011; Cohen et al., 2009; Windsor & Butterworth, 2010), we found higher levels of global marital satisfaction among men than women. We also explored gender differences in positive and negative characteristics of marital quality, encompassing both treatment of and by one’s spouse. Using this more nuanced set of measures, we found that the gender differences in marital satisfaction documented elsewhere did not exist uniformly across all positive and negative marital characteristics. Specifically, three key gender differences emerged.

First, we found strong gender differences based on whether a particular item assessed spouse’s treatment of the respondent or respondent’s treatment by the spouse. Men offered consistently more positive assessments of their spouse’s treatment of them than did women. By contrast, less pronounced gender differences emerged in evaluations of one’s own treatment of one’s spouse. In other words, men and women rated themselves as equally good spouses although men rated their wives more positively than women evaluated their husbands. Men’s superior ratings of their spouse’s treatment may reflect the fact that current cohorts of older women were likely to be socialized into the role of caregiver; as such, men may accurately perceive their wives as highly nurturing, giving, and communal (Helgeson, 1994; Thompson, 1993). Men also may impute to their wives those traits and behaviors that they expect women to have, consistent with the themes of expectation-based theories that describe how individuals view members of particular social groups (such as women) favorably when they conform to expectations for the group (Eagly & Diekman, 2005). The lack of gender difference in appraisals of one’s own treatment toward the spouse, by contrast, may reflect the widely documented finding that men tend to offer excessively positive evaluations of themselves relative to those offered by their partner on dimensions ranging from sexual attractiveness to household labor contributions (Perilloux, Easton, & Buss, 2012; Press & Townsley, 1998).

Second, we found evidence for gender differences in marital quality profiles by investigating naturally co-occurring combinations of marital characteristics (person-centered approach). The most pronounced differences emerged for the Negative Profile; different items were core markers of this profile in men and women. By contrast, the Positive and the Positive–Negative profiles showed greater similarities as indicated by high correlations between male and female profiles and few differences in core markers. These results are consistent with studies detecting gender differences in emotional responses to unhappy marriages; women are more likely than men to acknowledge and respond to negative interactions (Carstensen et al., 1995).

Third, our person-centered approach of marital profiles allowed us to examine the interrelatedness of specific characteristics for men versus women. For instance, in the Negative Profile, men had low levels of endorsing items such as “spouse makes too many demands” and “spouse is critical of me.” We suspect that these responses may reflect men’s expectations for what they believe wives “should” do. For current cohorts of older men, wives are an important source of help in both maintaining health behaviors (Umberson et al., 2006) and in providing direct physical care (Kaufman & Taniguchi, 2006). Men who receive frequent health reminders from their wives, for instance, may perceive these “nudges” as a normal part of marriage rather than an indication of criticism. Consistent with this interpretation, we found that men in the Negative profile did not report high levels of criticism or demands from their wives and instead endorsed items like “I feel close yet sometimes upset.”

Interestingly, we found that for men and women, the two most negative items were similarly related with the Negative profiles (i.e., “serious difficulty in marriage” and “spouse does not treat me well”). This suggests that more serious negative experiences in marriage may not be as susceptible to gender differences, at least among older adults in long-term marriages. Appraisals of negative exchanges may be most accurate and least susceptible to recall bias because negative experiences are less common and thus more salient than positive ones (Tversky & Kahneman, 1981). Negative events or interactions tend to be more severe and thus easier to recall (Elliott, 1997). Consistent with this perspective, studies of spousal concordance in marital appraisals document substantially higher correlations between husband and wife reports of conflict relative to positive marital characteristics (Carr & Boerner, 2009).

Overall, the three profiles explained more variance in marital satisfaction of women compared with men, consistent with our hypothesis that the profiles would weigh more heavily into global marriage evaluations for women. The Positive profile contributed most to the global satisfaction of both women and men, underscoring the important role of positive interactions for global marital satisfaction. However, it also explained more variance in the global marital satisfaction of women versus men. By contrast, the Negative and Positive–Negative profiles were slightly more important for predicting men’s global marital satisfaction.

These patterns suggest that when men do perceive difficulties in their marriage, it might have a more sizeable and salient effect on their global marital satisfaction assessments. Given prior studies documenting that men are less likely than women to both notice and respond to marital difficulties (Carstensen et al., 1995), those who perceive and acknowledge such difficulties may be particularly sensitive to their marriage’s emotional climate, which carries implications for their global marital satisfaction. Moreover, men who report difficulties in their marriage may feel that their wives are failing to live up to gendered expectations (Eagly & Diekman, 2005) and thus may register more profound feelings of marital discontent. Spouses are an important source of well-being (or strain) for older men, when their broader networks of friends and family decrease in size, thus intensifying the salience of marital relations (Schwarzer & Gutierrez-Dona, 2005). Women, by contrast, may have lower expectations for their husbands and may expect at least some negative exchanges. As such, negative exchanges may not factor into their overall marital satisfaction as profoundly as they do for men. Further, women may have a richer pool of social relationships and may receive support from friends, siblings, and children. These positive relations may compensate for problematic relations with a spouse.

Limitations and Future Directions

We focused on older adults’ individual assessments of their marriage, rather than both own and partner’s assessments. The primary reason was that we wanted to use the more representative full sample of the CLOC to establish the relationships between specific marital characteristics and global marital satisfaction in older men and women. Because enrolling two members of an older couple in a research study is more challenging than recruiting a single spouse only, couples who agree to participate may be positively selected in terms of marital quality and health. Our goal was to have a sample that encompassed the range from very happy to quite problematic marriages. However, we recognize the importance of adopting an explicitly dyadic approach to explore how husband and wife perceive their marriage and will explore these issues in future work.

Second, we could not assess the cognitive processes through which men and women arrive at their marital quality assessments. Future studies should explore the processes through which men and women factor specific marital interactions into their global assessments, by using in-depth interviews, focus groups, or cognitive interviewing in tandem with closed-ended survey measures of marital quality.

Third, the CLOC includes a single cohort of men and women who were born in the early 20th century and came of age in an era marked by traditional gender-role socialization and gender-typed allocation of social roles in the family. It is uncertain whether our findings can be generalized to future cohorts of older adults. For members of the Baby Boom cohort, both men and women are more likely than prior generations to have a somewhat more equitable allocation of household labor, and both men and women typically worked for pay (Pruchno, 2012). Gender-typed expectations in marriage may be shifting toward a context where both men and women expect that their partner will be nurturing and will contribute to the overall emotional functioning of the relationship (Sullivan, 2006). We look forward to future studies that examine the gendered dynamics among older married couples belonging to the large Baby Boom cohort.

Despite these limitations, our findings may prove useful to researchers studying later-life marriage. We encourage researchers to develop marital quality scales that explicitly differentiate treatment toward spouse and treatment by spouse and evaluate the extent to which each of these dimensions affects important outcomes such as health, mortality, and caregiver strain. In addition, given the differential predictive value of the specific marital evaluations in men and women, an open question remains which aspects other than the specific behaviors, support exchanges, or emotions evaluated in this study may help to better understand men’s global marital satisfaction.

Funding

The Changing Lives of Older Couples (CLOC) Study was supported by the National Institute on Aging Grants R01-AG610757 (Camille B. Wortman, Principal Investigator), R01_AG15948-01 A1 (Randolph M. Nesse, Principal Investigator), and R01 AG05561-01 (James S. House, Principal Investigator).

Acknowledgments

K. Boerner, D. Jopp, and D. Carr collaborated closely in conceptualizing and writing the paper. S. Kim and D. Jopp performed all statistical analyses. L. Sosinsky cowrote the paper.

References

.

(1994)

.

Gender differences in the link between marital support and satisfaction in older couples

.

Journal of Personality and Social Psychology

,

67

,

688

–

698

. doi:10.1037/0022-3514.67.4.688

.

(2000)

.

Sex differences in aggression between heterosexual partners: A meta-analytic review

.

Psychological Bulletin

,

126

,

651

–

680

. doi:10.1037/0033-2909.126.5.651

.

(2003)

.

Prospective effects of marital satisfaction on depressive symptoms in established marriages: A dyadic model

.

Journal of Social and Personal Relationships

,

20

,

355

–

371

. doi:10.1177/0265407503020003005

.

(1991)

.

A treatise on the family

.

Cambridge, MA

:

Harvard University Press

.

.

(2014)

.

Implications of marital/partner relationship quality and perceived stress for blood pressure among older adults

.

Journal of Gerontology: Psychological Sciences

,

69

,

188

–

198

. doi:

10.1093/geronb/gbs123

.

(2005)

.

The role of marital quality in physical health during the mature years

.

Journal of Aging and Health

,

17

,

85

–

104

. doi:10.1177/0898264304272794

.

(1988)

.

Individual difference variables in close relationships: A contextual model of marriage as an integrative framework

.

Journal of Personality and Social Psychology

,

54

,

713

–

721

.

.

(2011)

.

Gender, marital power, and marital quality in later life

.

Journal of Women & Aging

,

23

,

3

–

22

. doi:

10.1080/08952841.2011.540481

.

(2006)

.

Methodological issues in studying bereavement

. In (Eds.),

Spousal bereavement in late life

(pp.

19

–

47

).

New York, NY

:

Springer

.

.

(2009)

.

Do spousal discrepancies in marital quality assessments affect psychological adjustment to widowhood?

Journal of Marriage and Family

,

71

,

495

–

509

. doi:

10.1111/j.1741-3737.2009.00615

.

(1995)

.

Emotional behavior in long-term marriage

.

Psychology and Aging

,

10

,

140

–

149

. doi:10.1037/0882-7974.10.1.140

.

(2009)

.

Marital quality and global well-being among older adult Israeli couples in enduring marriages

.

The American Journal of Family Therapy

,

37

,

299

–

317

. doi:

10.1080/01926180802405968

.

.

(2005)

.

What is the problem? Prejudice as an attitude-in-context

. In (Eds.),

On the nature of prejudice: Fifty years after Allport

(pp.

19

–

35

).

Oxford, UK

:

Blackwell Publishing

.

.

(1997)

.

Traumatic events: Prevalence and delayed recall in the general population

.

Journal of Consulting and Clinical Psychology

,

65

,

811

–

820

. doi:10.1037/0022-006X.65.5.811

.

(2005)

.

Gender-related predictors of change in marital satisfaction and marital conflict

.

American Journal of Family Therapy

,

33

,

61

–

83

. doi:10.1080/01926180590889211

.

(2010)

.

Filling the glass: Gender perspectives on families

.

Journal of Marriage and Family

,

72

,

420

–

439

. doi:

10.1111/j.1741-3737.2010.00711

.

(1997)

.

A new look at marital quality: Can spouses feel positive and negative about their marriage?

Journal of Family Psychology

,

11

,

489

–

502

. doi:10.1037/0893-3200.11.4.489-502

.

(2003)

.

Marital quality. International Encyclopedia of Marriage and Family

. Encyclopedia.com. Retrieved

March 7, 2014

, from http://www.encyclopedia.com

.

(2000)

.

Decade review: Observing marital interaction

.

Journal of Marriage and Family

,

62

,

927

–

947

. doi:10.1111/j.1741-3737.2000.00927.x

.

(1994)

.

Relation of agency and communion to well-being: Evidence and potential explanations

.

Psychological Bulletin

,

116

,

416

–

428

. doi:10.1037/0033-2909.116.3.412

.

(2014)

.

Gender differences in marital satisfaction: A meta-analysis

.

Journal of Marriage and Family

,

76

,

105

–

129

. doi:

10.1111/jomf.12077

.

(2006)

.

Gender and marital happiness in later life

.

Journal of Family Issues

,

27

,

735

–

757

. doi:

10.1177/0192513X05285293

.

(2001)

.

Marriage and health: His and hers

.

Psychological Bulletin

,

127

,

472

–

503

. doi:10.1037/0033-2909.127.4.472

.

(2010)

.

Evaluating the invariance of cognitive profile patterns derived from Profile Analysis via Multidimensional Scaling (PAMS): A bootstrapping approach

.

Journal of International Testing

,

10

,

33

–

46

. doi:10.1080/15305050903454549

.

(1990)

.

The long-term marriage: Perceptions of stability and satisfaction

.

International Journal of Aging & Human Development

,

31

,

189

–

195

. doi:10.2190/H4X7-9DVX-W2N1-D3BF

.

(2003)

.

The relative importance of three domains of positive and negative social exchanges: A longitudinal model with comparable measures

.

Psychology and Aging

,

18

,

746

–

754

. doi:10.1037/0882-7974.18.4.746

.

(2012)

.

The misperception of sexual interest

.

Psychological Science

,

23

,

146

–

151

. doi:

10.1177/0956797611424162

.

(1998)

.

Wives’ and husbands’ housework reporting: Gender, class, and social desirability

.

Gender & Society

,

12

,

188

–

218

. doi:

10.1177/089124398012002005

.

(2007)

.

Marital quality and personal well-being: A meta-analysis

.

Journal of Marriage and Family

,

69

,

576

–

593

. doi:10.1111/j.1741-3737.2007.00393.x

.

(2012)

.

Not your mother’s old age: The Baby Boomers at age 65

.

The Gerontologist

,

52

,

149

–

152

. doi:

10.1093/geront/gns038

.

(1977)

.

The CES-D scale: A self-report depression scale for research in the general population

.

Applied Psychological Measurement

,

1

,

385

–

401

. doi:10.1177/014662167700100306

.

(1981)

.

Bases of marital satisfaction among men and women

.

Journal of Marriage and the Family

,

43

,

941

–

955

. doi:10.2307/351350

.

(2009)

.

How easily does a social difference become a status distinction? Gender matters

.

American Sociological Review

,

74

,

44

–

62

. doi:10.1177/000312240907400103

.

(2013)

.

Trajectories of care: Spouses coping with changes related to mild cognitive impairment

.

Dementia

,

12

,

45

–

62

.

.

(1998)

.

Investigating the positive and negative sides of personal relationships: Through a lens darkly?

In (Eds.),

The dark side of close relationships

(pp.

369

–

393

).

Mahwah, NJ

:

Erlbaum

.

.

(1990)

.

Supportive interactions, negative interactions, and depressed mood

.

American Journal of Community Psychology

,

18

,

423

–

438

. doi:10.1007/BF00938116

.

(2005)

.

More spousal support for men than for women: A comparison of sources and types of support

.

Sex Roles

,

52

,

523

–

532

. doi:10.1007/s11199-005-3718-6

.

(2013)

.

A profile approach to child care quality, quantity, and type of setting: Parent selection of infant child care arrangements

.

Applied Developmental Science

,

17

,

39

–

56

. doi:10.1080/10888691.2013.750196

.

(2006)

.

Examining the origins of gender differences in marital quality: A behavior genetic analysis

.

Journal of Family Psychology: JFP

,

20

,

605

–

613

. doi:10.1037/0893-3200.20.4.605

.

(2006)

.

Changing gender relations, changing families: Tracing the pace of change over time (Gender Lens Series)

.

New York, NY

:

Rowman & Littlefield

.

.

(2004)

.

Gender differences in depression: A study of older unlike-sex twins

.

Aging and Mental Health

,

8

,

187

–

194

. doi:10.1080/13607860410001669714

.

(1993)

.

Conceptualizing gender in marriage: The case of marital care

.

Journal of Marriage and Family

,

55

,

557

–

569

.

.

(1981)

.

The framing of decisions and the psychology of choice

.

Science

,

211

,

453

–

458

. doi:10.1126/science.7455683

.

(2005)

.

Marital quality, health, and aging: Gender equity?

Journal of Gerontology: Psychological Sciences

,

60

(Special Issue II),

109

–

112

. doi:10.1093/geronb/60.Special_Issue_2

.

(2006)

.

You make me sick: Marital quality and health over the life course

.

Journal of Health and Social Behavior

,

47

,

1

–

16

. doi:10.1177/002214650604700101

.

(2003)

.

Research strategies in developmental psychopathy: Dimensional identity and the person-oriented approach

.

Development and Psychopathy

,

15

,

553

–

580

. Retrieved from http://dx.doi.org/10.1017/S0954579403000294

.

(2010)

.

Supportive, aversive, ambivalent, and indifferent partner evaluations in midlife and young-old adulthood

.

Journal of Gerontology: Psychological Sciences

,

65

,

287

–

295

, doi:

10.1093/geronb/gbq016

.

(2006)

.

Marital discord and well-being in older adults: Is the association confounded by personality?

Psychology and Aging

,

21

,

626

–

631

. doi:10.1037/0882-7974.21.3.626

Author notes

Decision Editor: Merril Silverstein, PhD

© The Author 2014. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: .

Which of the following is true of the gender differences in nonverbal behavior group of answer choices?

What is the attribution of motives to someone?

What is the totality of the beliefs and attitudes people have about themselves?

Which two messages are especially likely to lead to a power struggle rather than to effective communication?