| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Department of Sociology, Bowling Green State University, Ohio
b College of Law, Indiana University, Bloomington
c Department of Sociology, University of Connecticut, Storrs
Gary R. Lee, Department of Sociology, Bowling Green State University, Bowling Green, OH 43403-0231 E-mail: grlee{at}bgnet.bgsu.edu.
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
This study documented the stronger adverse effect of widowhood on the psychological well-being of men than that of women and explained why this gender difference in the effect of widowhood exists.
Methods.
Data came from Wave 1 of the National Survey of Families and Households. Married and widowed people aged 65 and older were selected (n = 1,686). The dependent variable was the Center for Epidemiologic Studies—Depression scale (CES-D).
Results.
Widowhood was indeed more depressing for men than women. However, this was due primarily to the fact that married men were much less depressed than married women; widowed men and women were comparably depressed. Other contributors to the stronger effect of widowhood for men included men's shorter average time since widowhood, lower frequency of church attendance, stronger dislike of domestic labor, and lessened ability to assist their children.
Discussion.
Although widowhood has a strong depressive effect for older men, its effect for women is nonsignificant, and it explains a small proportion of the variation in depressive symptomatology. This suggests that most people, particularly women, adapt relatively well in the long run.
MANY studies have shown that widowed persons score lower than married persons on measures of psychological well-being (Bengtson, Rosenthal, and Burton 1990
; Gove and Shin 1989; Lee, Willetts, and Seccombe 1998; Mastekaasa 1994; W. Stroebe and Stroebe 1987, W. Stroebe and Stroebe 1993; Umberson, Wortman, and Kessler 1992). The fact that widowhood is a distressing experience is no mystery. The involuntary loss of a spouse through death, preceded by either serious illness or unanticipated accident, must be at least somewhat traumatic. Most open questions regarding the psychological consequences of widowhood involve what happens in the longer run.
Widowhood is, of course, a much more common experience among women than men. Studies that have compared men and women are somewhat inconsistent. Several reported no gender differences in the psychological consequences of widowhood (Faletti, Gibbs, Clark, Pruchno, and Berman 1989; Gerstel, Riessman, and Rosenfeld 1985; Lieberman 1996; Lund, Caserta, and Dimond 1986, Lund, Caserta, and Dimond 1989; Van Zandt, Mou, and Abbott 1989; Zisook and Schuchter 1991). Others (Farnsworth, Pett, and Lund 1989; Gallagher, Breckenridge, Thompson, and Peterson 1983; Schuster and Butler 1989; Thompson, Gallagher, Cover, Galewski, and Peterson 1989) found widowed women to be more distressed than widowed men. The usual explanation for this is that widowhood is more likely to cause financial difficulties for women than for men, and financial strain reduces psychological well-being (Ross, Mirowsky, and Goldsteen 1991; Umberson et al. 1992).
However, most studies have found widowhood to have a more adverse effect on men than on women (Bowling 1988; Lee et al. 1998; Mastekaasa 1994; Mendes de Leon, Kasl, and Jacobs 1994; Peters and Liefbroer 1997; M.S. Stroebe and Stroebe 1983; W. Stroebe and Stroebe 1987, W. Stroebe and Stroebe 1993; Umberson et al. 1992). Although the preponderance of the evidence favors this pattern, neither the contrary findings of some studies nor the causal processes that may produce a stronger effect for men have been satisfactorily explained. There are at least six reasons why widowhood may be more strongly associated with poor psychological well-being for men than for women, particularly in the long run. The first two of these causes are artifactual or statistical and thus may not produce real differences in the experiences of individuals. The remaining possible causes may be connected to traditional gender roles.
First, men are approximately five times more likely to remarry after widowhood than are women (Mastekaasa 1994; Peters and Liefbroer 1997). Remarriage is likely to be selective of those with the highest levels of well-being, thus lowering the average of those remaining in the widowed population. However, although more common among men than among women, remarriage is still a very infrequent occurrence among widowers, particularly those who were older when their wives died. Because it is rare, the influence of remarriage is likely to be small.
Widowhood also elevates the risk of mortality more for men than for women (Kaprio et al. 1987). Mortality is probably selective of those who are lowest on well-being (Mastekaasa 1994), counterbalancing the effect of remarriage and increasing the average well-being of those remaining in the population of widowers.
Second, the effects of widowhood are moderated by time (Lee et al. 1998; Mastekaasa 1994; Mendes de Leon et al. 1994; Peters and Liefbroer 1997; W. Stroebe and Stroebe 1987; Umberson et al. 1992). Average length of widowhood is shorter for men than for women because of the selection of men out of the status of widower by remarriage and mortality. Therefore, at any given time a higher proportion of widowers than widows is recently widowed, and the recently widowed tend to have lower psychological well-being.
Third, widowhood may be more problematic for men because they have more to lose. Bernard 1973 suggested that marriage is more beneficial for men than for women, at least in psychological terms. Although wives benefit more from marriage financially (Marini 1989), husbands receive more emotional support from their spouses (Gove 1984; Gove, Hughes, and Style 1983; Lee 1988; Lieberman 1996; Peters and Liefbroer 1997; Umberson et al. 1992). If men benefit emotionally from marriage more than do women, then men have more to lose by its involuntary termination.
Fourth, widowhood appears to be associated with subsequent declines in health more strongly for men than for women (Bengtson et al. 1990; Lee et al. 1998; Mastekaasa 1994; W. Stroebe and Stroebe 1987). The reasons for these declines may involve poorer nutrition for widowed men (Umberson et al. 1992) or higher frequencies of unhealthy behaviors such as smoking and drinking among men (Zisook, Schuchter, and Mulvihill 1990). All relevant studies have shown a positive relation between health and psychological well-being.
Fifth, many authors have suggested that widowed men are distressed by the need to assume domestic tasks (Arber and Ginn 1991; Gove and Shin 1989; Lee et al. 1998; Lopata 1996; W. Stroebe and Stroebe 1987). Umberson and colleagues 1992 found that, after widowhood, women's hours of domestic labor decrease but men's hours increase. Further, hours spent on housework increase depression for men but not women. Widowed men, particularly those from older cohorts, may be unprepared by socialization or experience for domestic tasks and find the necessity of performing them daunting.
Sixth, widowed women interact more with, and/or receive more support from, kin and friends than do widowed men (Antonucci 1990; Peters and Liefbroer 1997; W. Stroebe and Stroebe 1987; Wortman, Silver, and Kessler 1993). Silverstein and Bengtson 1994 found that support from children moderates the decline in well-being associated with widowhood. Although widowhood may reduce interaction with and support from married friends, it tends to increase interaction with other widows (Schuster and Butler 1989). Widowers, however, have limited access to other widowers because of their statistical infrequency; at the same time, they are very likely to have experienced a loss of interaction with married friends. This may reduce depression for widowed women relative to widowed men.
In this study we employed cross-sectional data from the first wave of the National Survey of Families and Households to test five of these possible explanations for why widowhood is more distressing for men than for women. The cross-sectional data allowed us to maximize the range of time since widowhood; this was a fairly crucial advantage because, as noted previously, the speed and trajectory of adjustment to widowhood may vary by gender. One disadvantage of cross-sectional data was that we could not assess the effects of the differential probabilities of remarriage and mortality by gender on the effects of widowhood.
We hypothesized that widowhood is more depressing for men than women because (a) men have been widowed for a shorter period of time, (b) widowhood is more strongly associated with poor health for men, (c) widowed men find domestic tasks more distressing, (d) widowhood reduces social interaction and support more for men, and (e) married men are less depressed than married women, so the difference in depression between the married and the widowed should be greater for men. However, widowed women's depression may be exacerbated by their greater financial difficulties.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
), a multistage area probability sample of 13,008 adults aged 19 and older. We selected a subsample consisting of respondents who were either married or widowed. To ensure that these categories were comparable, we restricted our analyses to those aged 65 and older. After adjustments for missing data, our final sample size was 1,686. The sample comprised 355 married women, 729 widowed women, 449 married men, and 153 widowed men.
Measures
The dependent variable was depressive symptomatology, measured by the 12-item version of the Center for Epidemiologic Studies—Depression scale (CES-D), developed from the original 20-item measure (Radloff 1977) by Ross and Mirowsky 1984. In this version, the scale has identical psychometric properties for men and women. According to Ross and Mirowsky 1984(p. 998), "Although the CES-D cannot be used to make a diagnosis of clinical depression, it has proven useful in distinguishing subgroups in community surveys according to their levels of depressive symptoms and in identifying risk factors for depressed mood in the general population." The item stem asks how many days in the past week the respondent felt, for example, "depressed," "fearful," and "that everything you did was an effort." Respondents who answered fewer than nine items (n = 17) were dropped from the analysis. For those who answered 9–11 of the items (n = 42), the mean score for the items they answered was substituted for the missing items. The possible range of the scale is 0–84; the actual range in this sample was 0–82, and the sample mean was 15.21. The reliability of the scale (Cronbach's alpha) is .91.
Gender (male = 1) and marital status (widowed = 1) were dummy variables. For widowed persons, length of widowhood was calculated in years. Married persons were assigned the mean length of widowhood; this insured that the coefficient for length of widowhood applied only to the widowed.
Physical health was measured in two ways. First, respondents were asked to compare their health to others of their own age on a scale ranging from 1 (very poor) to 5 (excellent). Second, we used a measure of functional limitations to ask whether the respondent had any physical or mental condition that limited his or her ability to perform six tasks such as caring for personal needs, climbing stairs, and doing household tasks. One point was assigned for each affirmative response.
There were several measures of social relations. Respondents were asked whether they gave help to or received help from their children in five areas (babysitting, transportation, repairs to homes or cars, other kinds of work around the house, and advice or emotional support). One point was assigned for each affirmative answer; respondents without children were assigned zeroes on these indices.
Our measure of sociability comprises seven items. The first four have a common stem, which asks how often, on a scale ranging from 0 (never) to 4 (several times a week), the respondent spends a social evening with (a) relatives, (b) a neighbor, (c) people he or she works with, and (d) friends who live outside the neighborhood. In addition, respondents were asked how often they (a) attend a social event at their church or synagogue, (b) go to a bar or tavern, and (c) participate in a group recreational activity such as bowling, golf, square dancing, etc. We summed these seven items to create an index of sociability ranging from 0 to 28; the reliability of the scale (alpha) is .57.
A rough measure of the absence of a support network was created from three items asking to whom the respondent would turn for assistance in the event of (a) an emergency in the middle of the night, (b) a need to borrow $200 for an emergency, and (c) a need for advice with a problem. The score was the number of times the respondent answered "no one."
Respondents were asked how often they participated in 15 different types of voluntary associations, on a scale ranging from 0 (never) to 4 (several times a week). Frequency of church attendance was measured as the average number of times per week the respondent attended church.
Two dimensions of domestic labor were included. One was total number of hours per week devoted to each of nine domestic tasks. The second was attitude toward housework, measured by a six-item semantic differential scale. The items, scored 0–7, were interesting–boring, appreciated–unappreciated, overwhelming–manageable, complicated–simple, lonely–sociable, and poorly done–well done. Scores were coded so that high scores reflected negative evaluations of housework (alpha = .67).
Age and education were indexed in years. Race was a dummy variable (non-White = 1). The measure of income was obtained from the appropriate total income indicator for either the individual respondent (for the widowed) or the couple (for married persons). We calculated assets by subtracting the amount owed from the estimated value for homes, other real estate, businesses or farms, and vehicles. For both income and assets, missing values were replaced by the means for the respondent's gender/marital status category.
Analysis
We began by examining mean differences in all variables by gender and marital status to ascertain how married and widowed men and women differed on depressive symptomatology and its hypothesized antecedents. Following this, we performed a regression analysis in which the predictors of depressive symptomatology included gender, marital status, and their multiplicative interaction term. If this term was significant, it meant that widowhood had different effects for men and women. We then added blocks of possible antecedent and intervening variables to the equation to ascertain whether they explained this interaction effect. Like any other covariate, a crossproduct term can be confounded with other covariates that also predict the response variable. The interaction effect can then be explained by the association of the crossproduct term with these covariates.
Inspection of the correlation matrix and the variance inflation factors yielded no evidence of multicollinearity.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
Model 4 added health indicators to gender and marital status. Those with better self-reported health and fewer functional limitations were less depressed. The interaction term regained significance. The health measures explained little of the interaction because differences in health by gender and marital status were minimal (see Table 1 ).
Model 5 added measures of social support to gender and marital status. Having someone to turn to in emergencies (support) was unrelated to depressive symptomatology, as was voluntary association participation. However, receiving help from children was depressing and giving help to children reduced the symptoms of depression. Those who attend church more often were less depressed. Inclusion of these variables reduced the cross-product term, but it remained significant.
Model 6 added the domestic labor variables. Hours of housework were negatively related to depressive symptomatology; those who did more were less depressed. But disliking housework was strongly and positively related to depressive symptomatology. As shown in Table 1 , widowed men reported more dislike of housework than either married men or widowed women. In this model the interaction term was only marginally significant (p < .10).
Model 7 included all variables, thus showing direct effects. The significant effect for gender indicates that married males were significantly less depressed than married females. Length of widowhood retained its negative effect, as did both indicators of health. Those who helped their children were less depressed, as were those who attended church more often. Time spent in housework had no direct effect, but disliking housework was positively related to depressive symptomatology.
More importantly, the gender gap in the widowhood effect, represented by the interaction of widowhood with gender, was no longer significant; it was reduced by 46% from its initial value of 4.29 in Model 2 to 2.33 in the complete model. This reduction was significant (t = 4.04, df = 1666, p < .001; see Clogg, Petkova, and Haritou 1995, for the test). This suggests that the gender gap was partially due to the facts that, compared with widowed women, widowed men had been widowed for a shorter time, provided less assistance to their children, attended church less often, and had a greater dislike of housework. Health did not help to explain the interaction effect because it was not related to gender or marital status in the appropriate directions.
Additional models were run incorporating (a) all possible two-way interactions by gender, (b) all possible two-way interactions by marital status, and (c) all possible three-way interactions by gender and marital status. None of the interaction terms was significant, meaning that the antecedents of depressive symptomatology did not vary by gender or marital status.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
).
Widowhood also has a stronger effect for men in part because their experience of widowhood is more recent (Mastekaasa 1994). Those widowed longer are less depressed, and men have been widowed for an average of nearly 5 years less than women. In addition, widowhood elevates depressive symptomatology more for men than for women because it decreases their church attendance and their provision of assistance to children. Other dimensions of sociability are unrelated to depression.
There is some support in these data for the hypothesis that widowed men find domestic tasks more daunting and therefore depressing (Umberson et al. 1992). Time spent in housework is not the critical variable here, partly because married and widowed men do not differ on it, and partly because, at least without other controls, time spent in housework is negatively related to depressive symptomatology. However, those who most dislike housework are most depressed, and widowed men dislike housework more than anyone else. It is, of course, hazardous to attempt to explain one state of mind with another state of mind; widowed men may dislike housework more than others because they are more depressed. However, it may also be that the burden of domestic labor, or their self-perceived failure to shoulder this burden adequately, is depressing for men, who may have little experience with it.
This study supports previous research showing that widowhood is more depressing for men than women. Unlike prior studies, this analysis shows that the stronger effect of widowhood for men is due largely to the comparatively low levels of depressive symptoms among married men. Widowed men and women are comparably depressed. However, it is also important to point out that, for most people, widowhood does not appear to be a psychologically devastating event in the long run. Our model including widowhood and gender explains only 2% of the variation in depressive symptomatology. Widowhood appears to have a lasting depressive effect, particularly for men, but the long-term effect is relatively small, and most widowed persons adjust quite well over time.
Received for publication September 29, 1999. Accepted for publication October 4, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G.-S. Jeon, S.-N. Jang, S.-J. Rhee, I. Kawachi, and S.-I. Cho Gender Differences in Correlates of Mental Health Among Elderly Koreans J. Gerontol. B. Psychol. Sci. Soc. Sci., September 1, 2007; 62(5): S323 - S329. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. R. Lee and A. DeMaris Widowhood, Gender, and Depression: A Longitudinal Analysis Research on Aging, January 1, 2007; 29(1): 56 - 72. [Abstract] [PDF]
|
|
|
|
|
|
B. M. van Gelder, M. Tijhuis, S. Kalmijn, S. Giampaoli, A. Nissinen, and D. Kromhout Marital Status and Living Situation During a 5-Year Period Are Associated With a Subsequent 10-Year Cognitive Decline in Older Men: The FINE Study. J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2006; 61(4): P213 - P219. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-H. Ha, D. Carr, R. L. Utz, and R. Nesse Older Adults' Perceptions of Intergenerational Support After Widowhood: How Do Men and Women Differ? Journal of Family Issues, January 1, 2006; 27(1): 3 - 30. [Abstract] [PDF]
|
|
|
|
|
|
S.-T. Cheng and A. C. M. Chan Relationship With Others and Life Satisfaction in Later Life: Do Gender and Widowhood Make a Difference? J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2006; 61(1): P46 - P53. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
