| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Jamila Bookwala, Department of Psychology, Lafayette College, 305 Oechsle Hall, Easton, PA 18042. E-mail: bookwalj{at}lafayette.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Marital processes • Marital quality • Age differences • Depressed affect
It is widely accepted that, as we age, our marital relationships can provide emotional support and act as a valuable source of positive affect. Carstensen's (1992) theory of socioemotional selectivity postulates that an important goal of late life is to enhance emotional closeness in our significant personal relationships. As a result, according to Carstensen, older individuals regulate their emotions such that emotional benefits will accrue. With regard to social relationships in general, older individuals experience a diminished desire for the type of social contact that offers no emotional rewards. Thus, they institute changes in their social networks that will minimize such undesired contact, reducing the breadth of their networks and redistributing their resources such that they can engage in a selected range of intimate relationships that are characterized by meaningful emotional experiences (Carstensen, Gross, & Fung, 1996). We propose that in the case of one's marital relationship, however, such reductions in intimate contact may not be feasible. Instead, the emotional regulation that Carstensen describes as typical across the social relationships of older adults may take a different form within marriage while achieving the same emotional benefits. Specifically, compared with their younger counterparts, older adults may be likely to give less weight to conflictual issues with their spouse and thus are likely to experience higher levels of marital satisfaction and better emotional well-being (conflict and disagreement notwithstanding).
Negative and Positive Marital Processes
Several researchers have highlighted the relevance of disagreement and conflict in marital satisfaction. Christensen and Walczynski (1997) identified conflict to be the most proximal factor influencing marital satisfaction in couples. Levenson and colleagues (1993) found that level of disagreement covaries with marital satisfaction. They observed that middle-aged and older couples who were classified as being in dissatisfying marriages reported more disagreement and conflict than those in satisfied marriages. Clements and colleagues (Clements, Cordova, Markman, & Laurenceau, 1997) pointed out that differences and conflicts are almost inevitable for most married couples and that the manner in which couples handle differences when they emerge is among the most powerful predictors of marital satisfaction. Karney and Bradbury (1997) documented that conflictual interactions and poor conflict resolution strategies are associated with marital dissatisfaction. Strategies for resolving conflict can vary along a continuum of confrontational tactics, ranging from using low levels of confrontation (e.g., using reason) to using high levels (e.g., shouting and hitting; see Straus & Gelles, 1986). Maladaptive or destructive conflict styles (i.e., aggressive or withdrawing responses) tend to be related to lower evaluations of marital quality (Ridley, Wilhelm, & Surra, 2001).
With few exceptions (e.g., Van Laningham, Johnson, & Amato, 2001), studies of age-group comparisons on marital satisfaction have shown a steady increase after middle adulthood, rising to levels close to that experienced by younger newlywed adults (e.g., see Glenn, 1990). In comparison, some researchers have found that marital conflict and negativity generally tend to decline with advancing age (Adelmann, Chadwick, & Baerger, 1996; Carstensen, Gottman, & Levenson, 1995; Keith, 1987; Levenson et al., 1993; Levenson, Carstensen, & Gottman, 1994). Nevertheless, marriages in the second half of life are not free from conflict. Not all long-term marriages realize their potential for enhanced positive affect and emotional support, and some couples remain intact despite experiencing considerable dissatisfaction with their marital relationships (Levenson et al., 1993, 1994). For example, Levenson and colleagues (1993) found that, on a scale of 0–100, middle-aged and older couples reported at least some level of disagreement in each listed source of conflict. Moreover, older couples did not vary significantly from their middle-aged counterparts in reported levels of disagreement for 6 out of the 10 listed sources of conflict (e.g., communication, in-laws, and alcohol and drugs). On the basis of these findings, we can expect that negative marital processes (disagreement, role allocation unfairness, and maladaptive conflict resolution strategies) will be reported by young, middle-aged, and older adult respondents, with evidence for a declining trend across these three age cohorts. Marital satisfaction, in contrast, is likely to be lowest among middle-aged adults. One of the goals of our study was to examine the relevance of negative marital processes to marital satisfaction across age cohorts. On the basis of socioemotional selectivity theory (Carstensen, 1992), we can expect that, just as individuals in late life work to enhance positive affect in their personal relationships and regulate their emotions to be beneficial, they will seek to achieve similar emotional benefits in their marital relationships. As a result of their drive to focus on the positive elements of their marriage, negative marital processes such as disagreement, role allocation unfairness, and poor conflict resolution strategies would be less salient to older individuals' evaluations of their marital relationship. Thus, we hypothesized that these negative marital processes would be less strongly associated with marital satisfaction in older adults and middle-aged adults than in younger adults.
Marital Processes and Depressed Affect
In general, marital processes are known to play a significant role in levels of depressed affect. In a recent meta-analysis, Whisman (2000) found that depressive symptoms were significantly related to marital dissatisfaction, with greater depression severity associated with greater marital dissatisfaction. Keith (1987) pointed out that depression may be the outcome of disagreements and interaction problems within a marriage when the marital system is unable to accommodate the strain. This link between marital processes and depressive symptoms has been observed both in young and older adults (Fincham, Beach, Harold, & Osborne, 1997; Ulrich-Jakubowski, Russell, & O'Hara, 1988). However, to our knowledge, no study has directly compared the strength of the relationships among marital processes and depressed affect across different age cohorts. Our second goal is to address this gap in the literature on marital processes and depressed affect. We propose that, in general, negative marital processes are directly related to depressed affect and indirectly by means of their contribution to lower marital satisfaction. However, extrapolating socioemotional selectivity theory (Carstensen, 1992) to the context of marital relationships, we hypothesize that negative marital processes will play a stronger role in depressed affect among younger married individuals than among middle-aged and older married individuals. Marital satisfaction, in comparison, will contribute more to emotional well-being (i.e., lower depressed affect) among middle-aged and older married individuals than their younger counterparts. This is because with increasing age individuals are likely to regulate their emotional well-being by enhancing the salience of the positive aspects of their intimate relationships and deemphasizing the negative aspects (Carstensen, 1992).
In sum, the goal of this research was to examine the applicability of socioemotional selectivity theory (Carstensen, 1992) to the context of marital relationships. To achieve this, we compared the relative fit of a mediational model linking negative marital processes to depressed affect directly and by means of marital satisfaction across three age cohorts. The age cohorts we used correspond with segments of the life span that are commonly described as young adulthood (20–39 years), middle adulthood (40–59 years), and late adulthood (60 years and older). This study also represents an innovative step toward understanding the role of age in the links between marital processes and mental health. Understanding the role of marital processes in depressed affect over the life course is important because marriage remains a fairly universal institution, and satisfying marriages tend to buffer spouses from psychological distress (Karney & Bradbury, 1997). Given the lack of available longitudinal data on marital relationships that span the years from young adulthood to late life, age-cohort comparisons remain a valuable methodological approach to understanding a presumably longitudinal process (Orbuch, House, Mero, & Webster, 1996). A potential drawback of comparing the relationships among negative marital processes, marital satisfaction, and symptoms of depression across different age cohorts is that individuals whose marriages were characterized by extremely high levels of negative marital processes or who were extremely dissatisfied with their marriages may already have terminated those marriages (see Glenn, 1990). Thus, such individuals would be excluded from a study such as the present one, making selection bias a potential criticism of the current sample. However, experts on marital research do not consider this to be likely because the decline in negative marital processes and the rise in marital quality in later life begins and continues well after the period when most marriages end by divorce (Orbuch et al., 1996).
This study uses data collected during the first wave of the National Survey of Families and Households (NSFH) that included measures of multiple indicators of negative marital processes (including perceived unfairness in the allocation of marital roles, level of disagreement, physical arguments, and conflict resolution strategies), marital satisfaction, and symptoms of depression. In addition to its rich data, the use of a population-based sampling strategy is an important strength of the NSFH (Sweet & Bumpass, 1996). Only individuals in first marriages were included in our study to limit the possible confounding that may occur by combining first and subsequent marriages in terms of marital quality (Kurdek, 1991).
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Measures
For the NSFH, data were collected on a broad range of issues that would permit the holistic analysis of family experience from a wide variety of theoretical perspectives (Sweet & Bumpass, 1996). The current study analyzed a subset of the data related to negative marital processes, marital satisfaction, and depressed affect. The details of each assessment are provided in the paragraphs that follow.
Negative Marital Processes
The NSFH included assessments of multiple indicators of negative marital processes. (Un)fairness in marital role allocation was assessed in four domains (household chores, working for pay, spending money, and child care). Respondents were asked to indicate their feelings of fairness in these domains by using a 5-point scale, ranging from 1 = very unfair to me to 5 = very unfair to my spouse. Because two items may be less relevant to older adults (working for pay and child care), these items were excluded. Scores on the remaining two unfairness items were summed to create a composite of marital unfairness such that higher scores indicate less unfairness to the respondent. Level of disagreement with the spouse was measured in seven different areas (household tasks, money, spending time together, sex, having a[nother] child, in-laws, and the children). Because having another child is likely to be less relevant as an area of disagreement for middle-aged and older adults, this variable was not included in the measure of level of disagreement for the present analyses. The NSFH asked respondents how often, if at all, in the previous year they had had open disagreements with their spouse about each of the areas. A 6-point response scale was used for these items, ranging from 1 = never, to 2 = less than once a month, to 6 = almost every day. Scores on the disagreement items were summed to create a composite of marital disagreement such that higher scores represented greater disagreement with the spouse. We obtained a Cronbach's alpha of.76 for this measure. Conflict resolution strategies were assessed in the NSFH by asking respondents to indicate the extent to which they used four different ways to deal with serious disagreements with their spouse, ranging from low to high in terms of confrontation and representing constructive (problem-solving) and destructive (withdrawal or aggressive) tactics (Ridley et al., 2001; Straus & Gelles, 1986). The specific strategies included 1, just keeping your opinions to yourself (low confrontation; destructive); 2, discussing your disagreements calmly (low confrontation; constructive); 3, arguing heatedly or shouting at each other (high confrontation; destructive); and 4, ending up hitting or throwing things at each other (high confrontation; destructive). A 5-point response scale was used, ranging from 1 = never to 5 = always. Because the four conflict resolution strategies represent qualitatively different ways of resolving conflict, these items were considered individually in the statistical analyses.
Marital Satisfaction
Marital satisfaction, our indicator of positive marital processes, was measured by using a single global item: "Taking things all together, how would you describe your marriage?" Responses were made with a 7-point scale, ranging from 1 = very unhappy to 7 = very happy. Global measures of marital satisfaction have been used with success in past research to assess the quality of marital relationships (e.g., see Glenn, 1990).
Depressed Affect
The NSFH included a 12-item version of the Center for Epidemiological Studies–Depression scale (CES-D). The CES-D is a widely used measure that assesses the severity of depressive symptoms during a 1-week recall period (Radloff, 1977). The revised version of the scale remains an appropriate measure of symptoms of depression (see Kessler, Foster, Webster, & House, 1992). Shorter versions of the CES-D have been used successfully to study depressed affect in other large sample studies (e.g., see Bookwala & Schulz, 1996). The NSFH used an eight-category set of response options to measure the number of days out of the past 7 (0–7) that each depressive symptom occurred. We obtained a Cronbach's alpha of.91 for this scale.
Control Variables
The control variables include measures of respondents' gender, age measured as a continuous variable, and perceived health measured by using a standard single-item assessment of general health (1 = very poor to 5 = excellent).
Data Analysis
Means and standard deviations for the entire sample and for each age cohort were computed for all study variables. Bivariate correlations between variables were computed between indicators of negative marital processes, marital satisfaction, and symptoms of depression (available from J. Bookwala upon request). A multivariate analysis of covariance (MANCOVA) was used to compare young, middle-aged, and older adults on the indicators of negative marital processes. Differences among the age cohorts on marital satisfaction and depressed affect were assessed by using an analysis of covariance (ANCOVA). Because of the considerable age range represented within each age cohort, the continuous variable of age was treated as a covariate in these analyses. Physical health, because of its known association with depressed affect and its expected salience especially for older adults' mental health, also was used as a covariate. Finally, participant gender was used as a control variable because of its known association with depressed affect.
In order to identify age-cohort differences in the mediational model linking negative marital processes, marital satisfaction, and depressed affect, a series of path analyses with observed measures of study variables was performed. First, a path analysis was performed by using EQS (Bentler, 1993) to test the fit of the data to the mediational model presented in Figure 1 for each age cohort. Direct paths from each indicator of negative marital processes to marital satisfaction and depressed affect were estimated. A path from marital satisfaction and depressive symptoms also was estimated to test the hypothesized indirect relationship between negative marital processes and depressed affect by means of marital satisfaction. Age (as a continuous variable), gender, and physical health were included as control variables with paths specified to both marital satisfaction and depressive symptoms. Individual indicators of negative marital processes were permitted to covary. Multiple goodness-of-fit indices for the path model were estimated with acceptable cutoff levels considered to be at.90 or better. When the initial fit indices did not meet acceptable cutoff levels, the Wald and Lagrange diagnostic tests were used to discern the extent to which respecifications in covariance estimates would improve the model fit; however, given their centrality to this research, no modifications of path coefficient estimations were introduced. Estimations of covariance parameters were accordingly modified and the model retested until the fit indices reached acceptable levels. Next, a multiple group analysis was performed to test the equivalence of path coefficients across the three age cohorts. In this analysis, the paths linking the different components of the mediational model were constrained to be equal across the three age cohorts. If specific equality constraints across the three age cohorts were found to be untenable, those constraints were released in subsequent runs of the multiple group analysis until only valid equality constraints remained. This analysis provided information about significant differences that existed in the relations among negative marital processes, marital satisfaction, and depressed affect across young, middle-aged, and older adults.
|
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
2(2) = 373.11, p <.001. In the final model for young adults, we found that, after participants' age, gender, and perceived health were controlled for, less perceived unfairness in marital role allocation and lower level of disagreement were associated with higher marital satisfaction (see Figure 2). Among conflict resolution strategies, keeping one's opinions to oneself, arguing heatedly or shouting, and hitting or throwing things at one's spouse were associated with lower marital satisfaction, whereas a greater tendency to discuss issues calmly was correlated with higher marital satisfaction. With regard to depressed affect, more disagreement, keeping one's opinions to oneself during conflict, and physical aggression were associated with higher depressed affect, whereas calmly discussing issues during conflict was associated with lower depressed affect. Consistent with our expectations, lower marital satisfaction was also significantly linked with higher depressed affect. On examining indirect effects, we found that each negative marital process indicator (role unfairness, disagreement, and the various conflict resolution strategies) had an indirect effect on young adults' depressed affect by means of their contribution to marital satisfaction. Overall, this model explained 18.7% of the variance in marital satisfaction and 20.7% of the variance in depressed affect.
|
2(3) = 86.59, p <.001. Overall, this final model explained less variance in middle-aged adults' marital satisfaction (14.3%) and depressed affect (14.6%) than in younger adults. Although the links between negative marital processes and marital satisfaction that were mostly observed with young adults were replicated in this group (see Figure 3), the association between marital satisfaction and depressed affect was not statistically significant for middle-aged adults, nor were the conflict resolution strategies of calm discussion, heated arguments, and physical aggression significantly associated with depressed affect. Moreover, in this age cohort the link between negative marital processes and depressed affect was not mediated by marital satisfaction.
|
2(6) = 26.51, p <.001. This final model explained 13% of the variance in older adults' marital satisfaction and 18.6% of the variance in their depressed affect. Once again, the association between negative marital processes and marital satisfaction observed in the other two age cohorts was largely replicated. In terms of depressed affect, marital satisfaction contributed significantly to depressed affect (as was seen with young adults) and higher disagreement was associated with higher depressed affect. None of the conflict resolution strategies were directly associated with depressed affect among this age cohort, although these strategies as well as role unfairness and level of disagreement had an indirect relationship with depressed affect by means of marital satisfaction.
Multiple Group Analysis
The initial multiple group analysis in which we constrained all paths to be equal across the three age cohorts indicated that several equality constraints were untenable. Thus, this model was rerun by lifting these constraints. The final results (see Table 2) indicated that the links between negative marital processes and marital satisfaction were uniform across the three age cohorts with one exception: Role unfairness was a stronger predictor of marital satisfaction among older adults than younger adults. When examining the results for depressed affect, we found that the path from perceived health (a control variable) to depressed affect was significantly stronger among middle-aged and older adults than among young adults. The paths from negative marital processes to depressed affect only varied in strength between young and middle-aged adults: Level of disagreement and calm discussion as conflict resolution strategies were stronger predictors of depressed affect among young adults than middle-aged adults. Finally, we found an important difference in the contribution of marital satisfaction to depressed affect between young and older adults: Marital satisfaction was a significantly stronger predictor of depressed affect for older adults compared with younger adults.
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Tests of the mediational model indicated that, in all three age cohorts, negative marital processes played a significant role in marital satisfaction. Unfairness in role allocation, higher levels of disagreement, keeping their opinions to themselves during conflict, avoiding calm discussions, and having heated arguments significantly predicted lower marital satisfaction in the entire sample. These findings are consistent with our expectations and with other research that has identified the powerful role of conflict and disagreement in marital satisfaction (Christensen & Walczynski, 1997; Clements et al., 1997; Karney & Bradbury, 1997). Also consistent with our hypothesis, marital satisfaction was significantly related to depressed affect for young and older adults (but not middle-aged adults). That marital satisfaction appears to play a significant role in the depressed affect of both young and old married individuals has important clinical implications. Thus far, we know that marital dissatisfaction has been linked with the incidence of a major depressive episode in individuals aged 60 years or younger (Whisman & Bruce, 1999). Although we did not assess clinical depression, our results suggest that even older individuals who report lower marital satisfaction may be at risk for higher depressed affect. Thus, clinical interventions that focus on improving marital quality by means of reduced disagreement, improved conflict resolution strategies, and more equitable marital role allocation may serve to lower the risk of more serious depressive symptoms among older adults. Although marital satisfaction did not play a mediating role for middle-aged adults, marital satisfaction mediated the link between negative marital processes and depressed affect for younger and older adults. Interestingly, for older adults the mediation was almost complete—the majority of the indicators of negative marital processes had no direct relationship with depressed affect, only indirect relationships by means of their association with lower marital satisfaction. For younger adults, the mediation was partial because negative marital processes had significant direct and indirect relationships (by means of marital satisfaction) with depressed affect. These findings suggest that the mechanism by which negative marital processes play a role in depressed affect may change over the life span.
Our results from the multigroup analysis revealed that two negative marital processes—higher level of disagreement and less use of calm discussions to resolve conflict—were more strongly associated with depressed affect for young adults than middle-aged adults. More importantly, marital satisfaction was more strongly associated with depressed affect among older adults than young adults. Overall, these findings can be interpreted in the context of the theory of socioemotional selectivity (Carstensen, 1992). At a younger age, married individuals may place greater emphasis on negative marital processes for their emotional well-being because they tend to perceive time as limitless and thus may be less motivated to highlight the positive aspects of the marital relationship. As individuals get older, however, they come to recognize that time is more limited, and they become more focused on emotion-regulation goals (Carstensen, Isaacowitz, & Charles, 1999). Thus, as our results suggest, older adults compared with their younger counterparts may be more motivated to make negative marital processes less salient in directly shaping their emotional well-being and instead give more weight to marital satisfaction per se as a contributor to psychological well-being (i.e., lower depressed affect).
It is important to point out that our findings were discrepant with the findings of some researchers in the area of marital conflict and that not all our hypotheses were confirmed. First, unlike other researchers (e.g., Adelmann et al., 1996; Keith, 1987), we observed no negative decline by age for negative marital processes. Instead, our results indicate that marriages in middle and late adulthood may not differ significantly on negative marital processes from those of young adults. As indicated earlier, our findings are more consistent with those of Levenson and his colleagues (Levenson et al., 1993, 1994), who have found that disagreement and conflict are not uncommon occurrences in longer-term marriages. It also should be noted that we used different indicators of negative marital processes than used in the other studies, controlled for sociodemographic variables and health, and performed a multivariate analysis on these variables. Any or all of these factors may explain the discrepant findings. Second, our hypothesis that negative marital processes would be less strongly related to marital satisfaction among middle-aged and older adults than young adults was not supported. Rather, the magnitude of the path coefficients linking the various negative marital processes to marital satisfaction was equivalent across the three age cohorts with but one exception (greater unfairness in role allocation was more strongly related to lower marital satisfaction among older adults than young adults). These results suggest that negative marital processes generally may remain salient to marital satisfaction across the life span. Both the lack of age-cohort differences in negative marital processes and the lack of differences in their relevance for marital satisfaction suggest that the operation of negative marital processes and the weight individuals assign to them for their psychological well-being may depend more on factors such as personality dispositions than life-course-related changes suggested by socioemotional selectivity theory. For example, Huyck (1995) pointed out that stable patterns evident in negative marital processes may be related to personality attributes such as neuroticism that refers to a tendency for anxiety and a negative perspective on life.
It is also important to point out the limitations of our study. First, our analysis unveils age-cohort differences in the role of marital processes in depressed affect but cannot speak to age-related changes in any definitive measure. However, because of the lack of available longitudinal data on marital relationships that span the years from young adulthood to late life, age-cohort comparisons remain a valuable methodological approach to understanding a presumably longitudinal process (Orbuch et al., 1996). Although the NSFH collected a second wave of data, Waves 1 and 2 were separated by approximately 5 years, which was excessively short to permit an analysis of change from a life-span perspective. Fortunately, however, the NSFH has launched a third wave of data collection with the original sample, which in the near future will offer researchers an opportunity to examine age-related continuity and change in marital processes and their relevance to depressed affect over a span of 15 years in each of our cohorts. Second, it is quite plausible that less satisfactory marriages are underrepresented in the older age cohort because they may have been terminated at an earlier age. However, by limiting our analysis to individuals in first marriages, we can eliminate any confound that could arise from more satisfactory second or third marriages in this age cohort. Moreover, it is important to bear in mind that studies that show a rise in marital satisfaction in late adulthood find that the increase occurs well after the modal age for divorce (Orbuch et al., 1996). Thus, any rise in marital satisfaction that may occur in late adulthood cannot be fully explained by the early dissolution of more contentious marriages. Finally, because our goal was to compare how marital processes and depressive symptoms are linked in different age cohorts, we limited our analysis to individual respondents' evaluations of their marital relationship. A valuable direction for the future would be to use the NSFH data to compare how both members in young and elderly married couples evaluate marital processes and the extent to which couple-level and individual-level evaluations of marital processes contribute to depressed affect in spouses. For now, the present study takes an important first step in studying the relationships among age, marital processes, and depressed affect and the extent to which they can be explained by the theory of socioemotional selectivity.
|
|
Footnotes |
|---|
1 Department of Psychology, Lafayette College, Easton, PA.
Notes: Row means with different superscripts are significantly different at p <.05 or better; n = 2,289, 1,145, and 691 for young, middle-aged, and older adults, respectively. *p <.05 (significant F); **p <.01 (significant F).
Note: YA = young adult; MA = middle-aged adult; OA = older adult.
Decision Editor: Linda S. Noelker, PhD
Received for publication October 7, 2002. Accepted for publication February 3, 2003.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. Choi and N. F. Marks Transition to Caregiving, Marital Disagreement, and Psychological Well-Being: A Prospective U.S. National Study Journal of Family Issues, December 1, 2006; 27(12): 1701 - 1722. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Bookwala and M. M. Franks Moderating Role of Marital Quality in Older Adults' Depressed Affect: Beyond the Main-Effects Model J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2005; 60(6): P338 - P341. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 BMJ, June 26, 2004; 328(7455): 1574 - 1574. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
