Social Comparisons and Satisfaction With the Division of Housework:
Implications for Men's and Women's Role Strain
AMY J. HIMSEL WENDY A. GOLDBERG
University of California, Irvine
Contemporary parents lack clear guidelines for the fair and equitable allocation of family work. According to social comparison theory, under conditions of uncertainty, individuals often compare themselves to others to gain a sense of what is "normal." The authors applied social comparison theory to the examination of satisfaction with the division of housework and the experience of role strain. Results of covariance structure analysis indicated that women reported higher levels of satisfaction when they did less housework than their female friends and greater satisfaction and less role strain when their husbands did more than other male comparison referents. In contrast, men were more satisfied when their wives did more housework than their own mothers did. Satisfaction mediated the link between social comparisons and role strain. Interviews with 25 fathers revealed that some men invoke an image of the "generalized other" to make their own contributions to housework seem more noteworthy.
Keywords:    dual-earner couples; social comparisons; fathers; division of labor; gender roles
The landscape for American families has changed dramatically over the past 25 years, with the rise in maternal employment among middle-class families emerging as one of the mostprominent social changes of our time (Hoffman & Youngblade, 1999). With the increase in the number of dual-earner families has come change in family life, notably the redefinition of fatherhood to include expectations for active involvement in child rearing (Marsiglio, 1995; Pleck, 1997). Despite greater involvement by men in
Authors' 'Note: Amy J. Himsel, Department of Psychology and Social Behavior; Wendy A. Goldberg, Department of Psychology and Social Behavior. We gratefully acknowledge Rachelle Strauss for facilitating data collection andDr. Chuansheng Chen for providing us with statistical consultation. Correspondence concerning this article should be addressed to Professor Wendy A. Goldberg, Department ofPsychology and Social Behavior, SchoolofSocial Ecology, University of California, Irvine, CA 92697.
JOURNAL OF FAMILY ISSUES, Vol. 24 No. 7, October 2003 843-866 DOI: 10.1177/0192513X03255323 O 2003 Sage Publications
843
844      JOURNAL OF FAMILY ISSUES / October 2003
parenting activities, the frontier of housework remains stalled in gender-typed patterns (Hochschild, 1989). Indoor household labor is still unequally divided between most couples, with wives doing almost twice as much housework as husbands (Lennon & Rosenfield, 1994; Wilkie, Ferree, & Ratcliff, 1998). The division of housework persists as a source of conflict in contemporary marriages: Men and women often hold different views on how housework should be allocated, and disagreements over the division of labor contribute to marital dissatisfaction (Wilkie et al, 1998; Yogev & Brett, 1985). In a recent study, Wilkie and colleagues (1998) found that perceived fairness (and perceived empathy) mediated the effects of division of labor on marital satisfaction.
The problem of the unequal division of housework extends beyond marital satisfaction to affect individual well-being. Several studies have demonstrated a link between an unequal division of labor and women's levels of depressive symptoms (e.g., Bird, 1999; Golding, 1990; Kessler & McRae, 1982). More than the actual allocation of tasks between couples, perceptions that the division of labor is unfair are associated with higher levels of depression and distress (Glass & Fujimoto, 1994; Lennon & Rosenfield, 1994; Robinson & Spitze, 1992).
Because of the associations with indicators of individual and couple well-being, it is important for family researchers to gain an understanding of the factors that influence adults' evaluations of their division of labor situation. Family sociologists have produced a large body of work addressing this issue, revealing the symbolic, gendered ways couples come to terms with struggles over housework allocation (Gager, 1998; Hawkins, Marshall, & Allen, 1998; Hawkins, Marshall, & Meiners, 1995; Hochschild, 1989; Thompson, 1991). Implicit in some of these studies is the notion that couples may assess their own housework arrangements by comparing their division of labor to that of others. For example, Gager's (1998) in-depth interview study of dual-earner couples revealed that women obtain social comparison information by talking with their friends about how much housework their husbands perform. Favorable comparisons (e.g., when wives' husbands are doing more housework than their friends' husbands) correspond to perceptions that the division of labor is fair and contribute to "feeling better" about their own domestic arrangements.
SOCIAL COMPARISON THEORY
In presenting his now-classic social comparison theory, Festinger (1954) postulated that social behaviors were predicated on the assumptions that individuals seek a sense of normalcy and accuracy about their
Himsel, Goldberg / DIVISION OF HOUSEWORK      845
world and that individuals affiliate more with others when they desire others ' views about their own thoughts and behaviors. Except when potential embarrassment is involved, conditions of high anxiety and uncertainty motivate affiliative behavior and social comparisons (Sarnoff & Zimbardo, 1961).
Recent research has extended social comparison theory to the realm of psychological well-being, demonstrating that one's relative standing in comparison to similar others has an impact on satisfaction with life (see Diener & Fujita, 1997). Based on variations in laboratory study results, some social psychologists suggest that social comparison processes are rather flexible and that individuals may pick and choose with whom to compare to self-regulate emotions and maintain positive well-being (Taylor, Wayment, & Carrillo, 1996).
Within the past decade, social psychologists have examined the role of social comparisons in global marital functioning, finding that uncertainty within a relationship motivated an interest in comparing one's situation with others. In particular, research has shown that adults with egalitarian attitudes felt more uncertain about how well their relationship was going and that this uncertainty was reduced by information gained through social comparisons with same-sex others (VanYperen & Buunk, 1991). The quality of the marital relationship also related to social comparisons: Individuals with better functioning marriages (i.e., low in marital dissatisfaction) preferred contact with similar others, whereas those with unsatisfactory marriages preferred to affiliate with others who had better marriages (Buunk, VanYperen, Taylor, & Collins, 1991). In particular, women who were dissatisfied with their marriages or uncertain about how things are going expressed a desire to talk with others about their marriage. Connecting social comparisons to emotional outcomes, Buunk and colleagues (1991) found that individuals who were high in marital dissatisfaction and rather uncertain about their marriages experienced negative affect from upward and downward social comparisons, that is, comparisons with those who are doing better and with those in worse shape were associated with negative affect (Buunk, Collins, Taylor, VanYperen, &Dakof, 1990). These studies have highlighted the importance of social comparison for self-evaluation among adults who are dissatisfied or uncertain about their marriages.
Despite its prominence as an area of uncertainty and conflict within marriages, to our knowledge social comparison theory has not yet been applied to a study of adults' division of housework and psychological well-being. In the absence of experts and concrete standards forbehavior, social comparison theory suggests that we look to others in our social
846      JOURNAL OF FAMILY ISSUES / October 2003
world for information on "how we are doing." In other words, it is through social comparison processes that we create social reality and affirm our own normalcy. Today, dual-earner families are the norm (Bureau of Labor Statistics, 2003), but most adults grew up in far more traditional households where fathers were the sole breadwinners and mothers were the primary caregivers. Contemporary husbands and wives lack objective standards and role models for how to manage both work and family obligations successfully. As put succinctly by Hawkins and Crouter (1991), contemporary dual-earner families must find their own way "without map or compass."
Social comparison theory would predict that dual-earner couples are motivated to affiliate and seek the views of others in domains of anxiety and uncertainty. As dual-earner couples forge their own work and family arrangements, division of household labor persists as an uncertain and often contentious issue for husbands and wives. Looking at the situations of others may help couples gauge how well they are managing their division of housework and multiple role arrangements.
SOCIOLOGICAL PERSPECTIVES ON THE DIVISION OF HOUSEWORK
Although social comparison theory has not been applied specifically to division of labor, family sociologists have revealed a more general role of comparison to peers in couples' assessments of their division of labor. Comparison with others has been found to be an important aspect of adults', especially wives', sense of fairness. Hochschild's (1989) seminal work on the "second shift" demonstrated that gender ideology does not clearly predict participation in household labor. Instead, wives make comparisons to establish a "going rate" for husband participation against which to compare their own husbands. Some men compare their family work contributions to a manufactured "do-nothing average dad" to establish their own high level of involvement (Gager, 1998). Thompson's (1991) influential theory of distributive justice includes suchwithin-gender comparisons as well. She suggests that wives feel better about their own division of labor situation by comparing their own husbands to other men they know. In contrast, when wives compare themselves to their husbands (between-gender), they are likely to assess their situation as unfair. To date, comparisons have been linked to perceptions of fairness in the division of labor. However, social comparison theory prompts us to investigate whether well-being is enhanced under conditions of favorable social comparisons.
Himsel, Goldberg / DIVISION OF HOUSEWORK      847
Providing an empirical test of Thompson's (1991) theory of distributive justice, Hawkins et al. (1995) examined division of labor and fairness issues in a sample of dual-earner wives. They found that between-gender comparisons were directly, negatively associated with perceptions of fairness, but indirectly, wives who made such comparisons had a more equitable division of labor, which was perceived as fairer. More recent work by Hawkins and colleagues (1998) has shown that dual-earner wives who believe their financial contribution is equally important as their husbands' contribution were less likely to describe their division of labor as fair. Gager's (1998) in-depth interviews with dual-earner couples indicated that the most frequently mentioned comparison referents were to peers and within gender, that is, wives comparing themselves to other female referents and comparing their husbands to other male referents; men made comparable within-gender comparisons as well.
THE CURRENT STUDY
This study aims to bridge disciplinary inquiry into social comparisons and psychological health, on one hand, and the gendered sense of fairness and satisfaction in the division of housework, on the other hand. Our empirical objective is to study the extent of one's own involvement in housework in comparison to others and to examine whether satisfaction and well-being are enhanced under conditions of favorable social comparisons. Hypotheses are examined in a sample of couples who are parents of school-aged children, a segment of society for whomfamily tasks, role arrangements, and the potential for role strain are particularly salient. The quantitative component of the current study focuses explicitly on the social comparison environments of dual-earner parents and the connections to satisfaction with division of labor and psychological well-being. The constructs of satisfaction and fairness are addressed in a qualitative component with a sample of men in dual- and single-earner households. This study marks an important extension of previous research because it addresses not only the current division of labor but also considers how adults' division of labor situations compare to others with whom they are likely to interact on a regular basis (e.g., their friends and spouses of their friends). Comparisons to current conceptualizations of past family models (e.g., how much housework done by one's mother) are also included.
Family sociologists have shown how comparisons are linked to perceptions of fairness, and social psychologists have demonstrated a link between social comparisons and psychological well-being. This study will build on previous work by bringing together these areas of study to exam-
848      JOURNAL OF FAMILY ISSUES / October 2003
ine the contribution of social comparisons to satisfaction with division of labor and role strain among adults in dual-earner marriages. Our consideration of social comparison is more elaborated than past work, includes more comparison referents, and allows us to understand the impact of individuals' "relative standing" on satisfaction and well-being. We measure the social comparison context by asking for information about four potential within-gender comparisons—self/parent, self/friend, spouse/parent, and spouse/friend. We investigate whether the social comparison environment around the division of labor is related to role strain and which comparison referents provide the most salient social comparison information. Interviews with men provide information from males' perspectives, which has been overlooked in previous research on social comparisons around the division of labor. Indeed, men have been presumed to be oblivious to these issues or have been characterized as satisfied even when the division of labor is unfair (e.g., Hochschild, 1989).
Research on well-being and division of labor thus far has focused on depression as the mental health variable of interest. We examine associations with role strain, an aspect of psychological well-being that deals specifically with the stress and conflict due to managing work and family responsibilities. Role strain in this study is operationalized as strain within roles and conflict between roles (Greenberger, Goldberg, Hamill, O'Neil, & Payne, 1989; Greenberger & O'Neil, 1993). Typically, other studies have measured role strain by a single item (e.g., Barnett & Baruch, 1985; Kessler & McRae, 1982), or they have examined only one aspect of the construct, namely, either strain within roles (e.g., Kandel, Davies, & Raveis, 1985) or conflict between roles (Kessler & McRae, 1982).
The current study is multimethod, relying on both quantitative, questionnaire-derived data and qualitative interviews. We expect favorable social comparison information to be positively associated with satisfaction with division of labor. For women, we hypothesize that higher levels of satisfaction and lower role strain will occur when women either do less than other women they know or their husbands do more housework than other men they know. Similarly, for men in our study, we expect comparisons where men do more than other men they know, and their wives do less than other women they know to be related to higher levels of satisfaction and less role strain. We hypothesize that favorable social comparisons will be associated with greater satisfaction with division of labor and less role strain. Whether the path from social comparisons to role strain is direct or mediated through satisfaction with the division of labor will be examined
Himsel, Goldberg / DIVISION OF HOUSEWORK      849
separately for men and women. The interviews, conducted with fathers in dual- and single-earner marriages, are expected to confirm that men also make social comparisons about the division of labor. The content of the interviews, which covered both satisfaction and fairness, should inform our understanding of the nature (within or between gender) and the direction (upward or downward) of these comparisons.
METHOD
SAMPLE
The sample forthe questionnaire study consisted of 172 adults, 91 men and 81 women, who were employed parents of school-aged children residing in two Southern California cities. Adults were selected from a larger pool of 256 respondents; criteria for selection into the present study were the marital and employment statuses of the parents. Adults needed to be married (89%) or living with a partner (11%). Criteria for inclusion also stipulated that the respondents have dual-earner marriages (or partnerships). Thus, all parents were employed for pay outside the home: Men worked an average of 45.8 hours per week (SD = 9.2) and women worked an average of 31.7 hours per week (SD = 13.6).
Families had an average of two children (M = 2.3, SD = .9). Approximately three quarters of the men and women were White, and the remaining one quarter were Latino (8%), Asian (5%), and other/unreported (11%). The mean age was 40.6 years (SD = 5.5) for men and 38.7 years (SD = 4.6) for women. The average parent had attended some college and the yearly family income for these dual-earner couples was between $60,000 and $80,000. The mean and mode of families' socioeconomic status was 4 (range = 1 -5) on the Hollingshead (1975) Four-Factor Index.
A subset of men (n = 25) from the whole study sample participated in the interview portion of the study. All but two of the men were married; these two men shared physical custody of their children. Nearly half (44%) of the men were in dual-earner marriages. The majority of the men were White (88%); one was Asian, one was Latino, and one self-described as "other." On average, these fathers were 41.4 years old (SD = 4.9), worked 45.0 hours per week (SD = 11.3), and had 2.2 children (SD = .62). About 40% of the fathers held a 4-year college degree and the mean and mode was 4 on the Hollingshead (1975) Four-Factor Index.
850      JOURNAL OF FAMILY ISSUES / October 2003 PROCEDURE
Parents completed survey packets that were distributed to them at their child's public elementary school. Childrenbrought the packets home with them, and parents were instructed to complete the surveys individually. Parents mailed completed surveys and consent forms to the investigators in postage-prepaid envelopes. Precise response rates were difficult to determine because we could not ascertain whether surveys sent home with children actually reached the parents; however, more women than men returned completed surveys. Demographics of those who did return surveys showed that our sample included more White adults and fewer Latino adults than the larger community in which the parents resided. Participant remuneration consisted of a coupon for a free item at a local fast-food restaurant. A subset of men who were active in their children's school were asked to participate in a 30-minute interview, which was scheduled for a time and place at their convenience. Fathers were interviewed by a graduate student interviewer without others present. Remuneration was a $5 gift certificate to a local eatery. The semistructured interviews were audio-tape-recorded and transcribed verbatim. Interviews and surveys were completed during the Spring of 2000.
MEASURES
Demographic information. Respondents reported their total weekly work hours, educational attainment, occupation, and the number of children living in the home. Occupational prestige was calculated for each participant using a combination of educational attainment and level of complexity of work (Hollingshead, 1975).
Division of housework. Men's and women's reports of their participation in feminine-type household tasks were captured on a 7-point response scale, ranging from -3 (wife does it all) to +3 (husband does it all), with the midpoint of 0 representing an even distribution of labor (response scale adapted from Cowan & Cowan, 1979/1988). Higher scores indicate greater participation by husbands in feminine housework tasks. The four feminine housework items were meal preparation, dishes, laundry, and housecleaning and represent tasks that are demanding on a regular basis. Labeling these tasks "feminine typed" is consistent with the approach used in other studies of household labor and gender (e.g., Arrighi & Maume, 2000; Lennon & Rosenfield, 1994; Presser, 1994). Item scores were averaged to create an index where higher scores indicated greater
Himsel, Goldberg / DIVISION OF HOUSEWORK      851
participation by men. Reliability was acceptable (Cronbach's alpha = .63).
Evaluations of the division of labor. In the questionnaire, respondents were asked to report the extent of their own satisfaction with the division of labor: "In general, how satisfied are you with the way you and your spouse divide child care/household tasks?" Responses ranged from 1 {very dissatisfied) to 5 {very satisfied). In the interviews, men were asked to reflect on the extent to which their own division of labor was fair or satisfactory in comparison to others (see social comparisonmeasuresbelow).
Psychological well-being. Seven items were selected from Greenberger's (Greenberger, 1988; Greenberger et al., 1989) Role Strain scale, which was the measure of well-being in the present study. Men and women rated how strongly they agreed or disagreed with statements concerning strain or overload within roles and conflict or negative spillover between roles. Responses were marked on a 6-point scale, where 1 = strongly disagree and 6 = strongly agree. Sample items include "Responsibilities at home are putting me under some strain" and "The quality of my everyday family life would be better if I were less involved in my work." The scale demonstrated strong reliability in the current study (Cronbach's alpha = .82). The validity of the original 32-item scale is indicated by its moderate correla-tion(rs= .47-.54) with the Center for Epidemiological Studies Depression Scale (CES-D) (Radloff, 1975) checklist and the Speilberger (1983) Trait Anxiety Scale for employed men and women (Greenberger, 1988).
Gender role ideology. Attitudes about work and gender roles were assessed using six items from the General Social Survey (GSS) (Davis & Smith, 1996). On a 5-point Likert-type scale (1 = strongly agree to 5 = strongly disagree), fathers and mothers indicated the extent of their agreement/disagreement with statements such as "It is not good if the man stays home and cares for the children and the woman goes out to work." The gender role ideology score was created by averaging the scores of these six items. The scale demonstrated good reliability (Cronbach's alpha=.71).
Social comparison: Referents and processes. Using qualitative information gleaned from Gager (1998) regarding comparison referents in perceptions of fairness among dual-earner couples, four items were developed to measure adults' sense of their own level of participation in household tasks in comparison to key referents. Lateral referents and referents to past family models were included. Men and women were asked
852      JOURNAL OF FAMILY ISSUES / October 2003
to choose among three statements to indicate whether they do (a) less, (b) about the same amount, or (c) more of household tasks as their same-sex peers and parent. They also were asked to compare their spouse's level of task participation to their spouse's same-sex peers and parent.
The interview was conducted to understand whether and how husbands make comparisons to others in their social world. The following question was asked: "When you think about how you and your spouse divide up household and child care in terms of how much you do, how much your spouse does and whether that division seems fair or satisfactory, do you ever compare yourself to other people you know?" If the father indicated that he did make comparisons, the interviewer then probed for information about to whom the father compared himself and how he thought his situation was better or worse than others'. A grounded theory approach was adopted to analyze the interview responses. Using the constant comparative method (Glaser & Strauss, 1967), different patterns of social comparisons emerged.
Prior to data collection, a pilot phase was conducted with a different sample of 37 dual-earner parents (20 men, 17 women) to establish the psychometric properties of the social comparison items. The pilot phase also produced significant correlations among the measures of social comparisons, role strain, and division of labor, suggesting that this was a fruitful area for further study.
PLAN OF ANALYSIS
The AMOS statistical software package (Arbuckle, 1997) was used to analyze covariance structures. Correlations of demographics with other key variables were examined to determine which control variables should be included in the model. Because previous research has found issues of fairness and satisfaction related to division of labor to differ by sex, the model was tested on men and women separately.
At least two indices of fit should be used when testing the accuracy of a proposed structural equation model (Reise, Widaman, & Pugh, 1993). In this study, the following five goodness-of-fit measures were calculated: (a) The likelihood ratio chi-square statistic (%2). This statistic is best used to assess the general fit of the model; however, due to the relationship be-tween%2 and sample size, it is not an accurate statistical test of the data, (b) The minimum value of chi-square/degrees of freedom (CMIN/df) statistic. This index of fit adjusts for degrees of freedom. Values of 2.0 and below are considered a good fit. (c) The comparative fit index (CFI). This statistic also adjusts for degrees of freedom. Value range is 0.00 to 1.00,
Himsel, Goldberg / DIVISION OF HOUSEWORK      853
with values of .90 and above considered an acceptable fit. (d) The goodness of fit index (GFI). This absolute fit index represents a ratio of the sums of squares of the observed model and a model of maximum possible fit. Similartothe CFI, the GFI ranges fromO.OO to 1.00, with values of .90 and above indicating acceptable fit. And (e) the root mean square error of approximation (RMSEA). This goodness-of-fit statistic is a reverse model fit indicator. It indicates looseness of fit per degree of freedom of the model. A perfect fit of the data to the model would be reflected in a value of 0.00. Values of .05 and below are generally considered acceptable, whereas a value of .08 indicates the upper boundary and reflects a questionable fit.
RESULTS
DESCRIPTIVE STATISTICS
Division of labor. Men and women agree that women are doing more of the feminine-type tasks, but women report that they do a significantly greater share of the housework than men say they do (men, M = -1.1; women, M= -1.5), F(l, 171) = 3.89, p < .05. Men report significantly higher levels of satisfaction with the division of labor than do women (men,M= 4.17; women,M= 3.60), F(\, 170) = 12.10,/? < .001. Women are more satisfied with the division of housework when men participate more in feminine tasks (r = .31, p < .01). Division of feminine tasks and satisfaction with division of housework are unrelated for men (r = -.14, as.).
Social comparison. Examination of responses to the social comparison items indicates that men and women see their division of labor situation as different than others, that is, they rate themselves as doing more or less than key referents. Men and women offer different portrayals of how their division of labor situation compares to those of others in their social world. Both men and women report that they do more housework than do their same-sex peers; men perceive this difference to be greater than do women (men, M = 2.71; women, M = 2.20). Men perceive their level of housework participation to be greater than that of their own fathers, whereas women see themselves doing somewhat less housework than their mothers did (men,M= 2.88; women,M= 1.93). Women think their husbands are doing more housework than their own fathers did and men
854      JOURNAL OF FAMILY ISSUES / October 2003
report their wives to be doing slightly more housework than their own mothers did (men, M= 2.12; women, M = 2.60). Women reported their husbands doing more than their female friends' husbands and men reported their wives doing more in comparison to their male friends' wives (men,M= 2.43; women, M =2.45).
Among women, reports of the division of labor are associated with perceptions of how one's division of labor compares to friends. Women who report their husbands doing more housework also report doing less housework themselves compared to female friends (r = -33,p<.0l). Similarly, women whose husbands are highly involved in housework report their husbands doing more housework than the husbands of their female friends (r = 30,p < .05). These findings provide partial evidence of the validity of the social comparison items constructed for this study.
Gender role ideology. Men and women were very similar in reports of gender role ideology; in general, this sample holds moderate views on appropriate roles for men and women (men, M= 2.81; women, M= 2.78), F{\, 169) = .17, as.
Role strain. Reports of role strain indicate that this sample of men and women experiences moderate levels of stress related to occupying roles of parent and worker (men,M= 3.28; womeiLM= 3.21), 1^(1, 171) = .26, as. Women who hold more egalitarian attitudes about gender report more role strain (r = .36, p < .01), providing partial support for the uncertainty hypothesis.
COVARIANCE STRUCTURE ANALYSIS
Using the AMOS program (Arbuckle, 1997), a model composed of key observed variables was tested controlling for work hours and gender role ideology.1 Indices of model fit were good: (4, TV = 172) = 3.46, p = .48; CMIN/df = .87; CFI = 1.00; GFI = .99; RMSEA =.00.
For women, spouse/parent comparisons and spouse/friend comparisons were positively associated with satisfaction (b = .40,/? < .05; b = .91, p < .001), whereas self/friend comparisons were negatively associated with satisfaction (b = -.48,/» < .05). The more husbands did in comparison to other male referents, the more satisfied wives were with division of labor. Conversely, the more wives did in comparison to their own female friends, the less satisfied wives were with the division of labor. Spouse/ friend comparisons were negatively related to wives' reports of role strain
Himsel, Goldberg / DIVISION OF HOUSEWORK      855
spouse/ friend
-.42**
spouse/ parent
self/ friend
self/ parent
role strain
Figure 1: Results for Women: Covariance Structure Analyses of the Effects of Social Comparison Information on Satisfaction With Division of Labor and Role Strain
*p < .05. **p < .01. ***p < .001.
(b = -.45,/? < .01). For wives, the more husbands did in comparison to their female friends' husbands, the lower levels of role strain reported. Satisfaction was not significantly associated with role strain. Variables in the model accounted for 34% of the explained variance in role strain among women.
Among men, spouse/parent comparisons were positively associated with satisfaction (b = 32,p< .05), and satisfaction was directly associated with role strain (b = -.38,/? < .001). The more their wives did in comparison to their own mothers, the more satisfied husbands were with the current division of labor. Higher satisfaction with division of labor was associated with lower levels of role strain. Results for men suggest that spouse/ parent comparisons have an indirect impact on role strain through satisfaction with division of labor. Taken together, variables included in the model accounted for 21% of the explained variance in men's role strain.
856      JOURNAL OF FAMILY ISSUES / October 2003
spouse/ friend
self/ friend
self/ parent
spouse/ parent
)   role strain
Figure 2: Results for Men: Covariance Structure Analyses of the Effects of Social Comparison Information on Satisfaction With Division of Labor and Role Strain
NOTE: Model was tested controlling for work hours and gender role ideology. For men. work hours and gender role ideology were not significantly related to role strain in the covariance structure analysis. Error terms were included for endogenous variables. *p<.05. ***/><.001.
SOCIAL COMPARISONS AROUND THE DIVISION OF FAMILY: RESULTS FROM THE QUALITATIVE INTERVIEWS
From the constant comparison method applied to the 25 interviews of fathers, categories emerged concerning whether social comparisons on division of labor issues were made, the object of the comparisons (i.e., to whom the men compared themselves), and the direction (upward, downward) of their comparisons. The marriages of the men in the interview sample varied in division of labor behavior, representing many points on the continuum of traditional-to-nontraditional, from the conventional "he does the outside, she does the inside" to the transitional "she does a little
Himsel, Goldberg / DIVISION OF HOUSEWORK      857
more" to the nontraditional format of "he assumes 90% to 95% of the responsibility for cooking and shopping."
The interviews established links between the comparison environment and the generation of social comparisons. Nearly all of the men claimed that they did make social comparisons, if not specifically to their spouse or parents, then to a "generalized other" or "average dad." Many respondents supplied more than one social referent, for example, spouse and other men or other families in general. Social referents often were selected that affirmed each household's particular division of labor and cast the respondent's contribution in a flattering light. For example, comparing himself to "other guys," one father remarked,
I think I'm significantly more involved in housework and whatnot than most fathers are. It's no scientific study, but I've always gotten the impression that most guys when given their choice come home, crank up the TV, pop the top on a beer and kick back and veg for the evening. Susy will regularly brag about how much I try to do like when she was pregnant: cook the meals and get up early so I'd have breakfast ready and waiting for her— little things like that.
About a third of the men responded to the interview questions by making a between-gender comparison and contrasting their level of participation in family work to that of their wives. Typically, husbands observed that their wives did more around the house and with the children, but they underscored their specific areas of involvement. A husband in a transitional dual-earner marriage noted his areas of participation:
She drops Jason off at school butlget Bobby off to school.. . .We're pretty even cooking, doing dishes, I try to help with the laundry, you know, even though she probably does the bulk of it.
Men who were the sole breadwinner acknowledged their wives' primary responsibility for child care but also pointed to their nontrivial role in the division of housework. As one sole wage earner commented,
She's a nonworking mom so she'll take the kids to school and pick them up and she'll watch them when she comes home. As far as household duties, I handle everything outside the house, like the yards and all that. On the inside, I think we each do it, whether it be the laundry or housecleaning or what have you.
Another primary wage earner brought up the issue of fairness:
858      JOURNAL OF FAMILY ISSUES / October 2003
Gina is the front end manager and I'm more of the back end, or back of the house support, doing some of the fixing, cooking, cleaning, and stuff like that. The division is different, but it's fair.
About a third of the fathers made within-gender comparisons, considering their household division of labor in comparison to that of other men their age. Dual-earner fathers were especially likely to speak generally of "other guys" or "other dads," what Gager (1998) termed the "average dad." Usually the image of other dads was invoked to make an implied or explicit downward comparison that placed one's own level of participation at a more active and involved level. A full-time employed father asked rhetorically, "How many people do you know where the husband does the majority of the cooking?" A man in a dual-earner marriage who was home one afternoon to care for his sick daughter, also asked rhetorically, "How many guys are doing that?" Another husband whose wife also worked assumed that other dads did less than he did: "I think probably for the most part I do a little more than most dads and ľ m probably more domesticated."
When generating social comparisons, a few men took the point of view of their own wives and their friends' wives to underscore their own high level of involvement: "In talking to other moms/wives, I have the distinct impression that I'm somewhere on the highest part of the curve."
Several men made cross-generational comparisons as well, contrasting their own level of involvement to current conceptualizations of past family models. These fathers tended to see the men of past generations and their own fathers as underinvolved, which motivated them to create more equality in their marriages. One man who was employed part-time, very involved in household and child care tasks, and married to a full-time employed wife, commented on the great difference between the generations:
When I grew up, the man was the man and he didn't do nothing. He went to work. He came home and his dinner better have been put on the plate. That's the way I was brought up. So yes, it is a big difference.
Another man in a dual-earner marriage was inspired to take a greater part in family work because he felt that his mother was "overtaxed" and his father "didn't do anything to help her."
A theme that echoed in a number of interviews was the utility of making social comparisons to confirm that the couple was on the right track. The absence of rules and guidelines for dual-earner marriages, especially those that strove to be egalitarian, seems to have prompted some men to assess how they and their families measure up compared to neighbors or
Himsel, Goldberg / DIVISION OF HOUSEWORK      859
others in general. A husband in a dual-earner marriage who had some uncertainty about the wisdom of his wife's working commented,
We do compare to others mostly to kind of reflect upon whether we're doing things right or we should go in this direction.... We will compare ourselves just to kind of stand back and see if we are going in the right direction or if there is a better way.
A few of the men denied making social comparisons, noting instead that you do what you have to do for your own situation. As one man put it: "I don't compare myself to any other people . . . somebody has to do things, and you just go from there." Echoing this sentiment, another man commented," [Our division of labor] has no relation to what other friends or relatives are doing. We got different schedules from them."
A self-employed father who said he did not make comparisons still invoked an understanding of social norms as exemplified by other men in general:
So our situation is really unique. Usually the men are typically gone. There are of course some stay-at-home dads and the women work at our school, but probably for 90% the fathers are gone and the stay-at-home moms handle the kids, and the babies, and stuff like that.
In summary, most of the fathers in this study did generate social comparisons, usually to a generalized other but also to specific referents in their immediate social sphere such as their spouses and their own parents. In making between-gender comparisons, men typically acknowledged the greater load placed on the wives, but men judged their own contributions to family work in a favorable light. Within-gender comparisons often were made to showcase the greater involvement of these dads as compared to other men.
DISCUSSION
Social comparisons provide a window into the paradox of why unequal divisions of labor are perceived as fair. The current study applied social comparison theory to an examination of satisfaction with division of housework and to the experience of role strain. The main findings of this study are that (a) social comparisons inform our understanding of satisfaction with the division of labor and psychological well-being among dual-earner parents and (b) the patterns of relations among social comparisons
860      JOURNAL OF FAMILY ISSUES / October 2003
and satisfaction that predict role strain differ for men and women. Interviews with 25 fathers supported our assumption that contemporary parents do make social comparisons about the division of housework, and they pick and choose social referents that help affirm the normalcy and accuracy of their own situation.
Common characteristics of employed parents reported from national random studies were mirrored in our nonrandom sample of husbands and wives. On average, women did more housework than did their husbands and were less satisfied with the division of labor than were their husbands. Satisfaction with the division of labor was higher among women whose husbands participated in a greater amount of housework. Men and women in our sample reported similar levels of role strain and egalitarian gender role ideology; however, the multivariate associations between the division of labor, satisfaction, and role strain differed by gender.
SATISFACTION WITH DIVISION OF LABOR
Drawing from the literature on social comparisons (Diener & Fujita, 1997; Gager, 1998), we hypothesized that social comparison information that suggests wives are doing less than other female referents and husbands are doing more than other male referents would produce higher levels of satisfaction for both men and women. Results for women supported our hypothesis. Information from several social referents, especially peers, had a strong bearing on women's level of satisfaction. In particular, women who perceived their husbands to be doing more housework than their friends' husbands reported higher levels of satisfaction than those who considered their husbands to be doing less than other male peers. Furthermore, women who believed they did less housework than their female friends reported higher levels of satisfaction with division of labor than women who felt they did more housework than their friends. Also making a significant contribution were comparisons of their husbands to their own fathers. Again, when their husbands compared favorably, women reported higher satisfaction. Results lend support to a growing literature suggesting that women's assessments of their own division of labor situations are much more complex than a simple input/output tabulation between husband and wife (Robinson & Spitze, 1992; Sanchez & Kane, 1996; Thompson, 1991). Information gleaned from the social comparison environment, as evidenced in this study, also may play an important part in these appraisals.
For men, results of the covariance structure analysis indicated that husbands experienced lower levels of satisfaction when their wives were do-
Himsel, Goldberg / DIVISION OF HOUSEWORK      861
ing less housework than their own mothers did. Although contrary to our study hypotheses, this finding may reflect the persistence of social roles (Eagly, 1987), the continued centrality of the breadwinner role for men, and the transitional state of the paternal role today (Meyers, 1993). Men's lower satisfaction with the division of labor when their wives are doing less housework may signal men's negative response to their wives' relinquishment of a primary domestic role as well as their own desire to keep a stronghold on the provider role. Indeed, men whose attitudes reflect strong preferences to be primary breadwinners are less likely to participate in housework (Perry-Jenkins & Crouter, 1990). Furthermore, having a more conventional division of labor increases men's perceptions of justice within their marriages (Wilkie et al., 1998), which may be consistent with our finding that men's greater satisfaction was linked to a preference for wives' greater domestic role.
ROLE STRAIN
Women and men reported similar levels of role strain, but the associations between social comparisons, satisfaction with division of labor, and role strain differed by sex. We turn now to a discussion of sex differences in the predictors of parental well-being.
Social comparisons, but not satisfaction with division of labor, directly related to role strain for women in our study. Interestingly, the social comparison referent most strongly associated with satisfaction also was the strongest predictor for role strain, highlighting the importance of peer referents for women. Wives who perceived their husbands to be doing more housework than their female friends' husbands reported significantly less role strain than other women whose husbands did the same amount or less than their own husbands. In contrast to women, analyses with dual-earner men revealed that social comparisons were linked to role strain through satisfaction with division of labor. It was not men's own level of participation, but their spouses' level, that was related to their satisfaction and well-being.
Our results point to gender differences in the importance of support from spouse and peers for well-being. Women are more likely than men to stay in close contact with their female friends and to seek support for their multiple roles from the larger social network, whereas men tend to rely on their wives as their main source of social support (Greenberger & O'Neil, 1993; Rubin, 1985). The importance of information from peers for the well-being of women in our study underscores the positive mental health outcomes that accrue from the social support of close female friendships
862      JOURNAL OF FAMILY ISSUES / October 2003
as dual-earner wives cope with division of labor issues. Men's results demonstrate that information from peers is not critical for their well-being; moreover, social comparison information of any kind relates to role strain only indirectly through satisfaction with the couples' division of housework. These findings highlight the centrality of the marital relationship for men's well-being. Furthermore, the closer link between satisfaction with the division of housework and role strain for men than for women in our study is consistent with a recent observation that it is men's preferences that determine the nature of the couples' division of housework (Wilkie et al., 1998). Social comparison information that women gain through their larger social networks may ameliorate their experience of role strain and explain why, in this sample, men and women report similar levels of role strain despite women's doing more housework.
These findings also demonstrate that satisfaction with division of labor has somewhat different implications when the outcome of interest is role strain rather than depression. The concept of role strain is well suited for the study of psychological outcomes related to dual-earner family processes because it is a measure of stress from "trying to do it all." Although it has been predictive of depression in other studies (Kessler & McRae, 1982;Lennon&Rosenfield, 1994), satisfaction with division oflaborwas not associated with role strain for the dual-earner wives in this study. As Crosby (1991) noted, there is a clear limit to the benefit of multiple roles for women; after a certain point, women simply become impoverished in time. Mothers may be satisfied with the division of labor at home but overwhelmed in other areas of life. Moreover, those who have an egalitarian orientation may feel pressure to be capable of doing it all by participating fully in provider and domestic roles. In the current study, egalitarian gender role ideology was associated with higher levels of role strain for women.
"THE GENERALIZED OTHER"
From the qualitative component of our study, we learned that dual- and single-earner fathers compare their family work to that of their wives, their ownfathers, and their friends. A number of men, particularly those in dual-earner marriages, invoked a generalized other (Blumer, 1969) who did less family work than they did. When they contrasted themselves to their wives, they often acknowledged that their wives assumed a greater share of family work, but their wives' level of involvement was not a standard to which the men aspired. Rather, in making lateral comparisons to others their own age, or comparisons back a generation to their own fa-
Himsel, Goldberg / DIVISION OF HOUSEWORK      863
thers, the men in our interview sample engaged in "social downgrading" (Heckhausen & Brim, 1997). These men described an image of the generalized other whose meager contributions to family work enhanced their own relative involvement. This exaggerated negative view of the typical dad around the house is consistent with the "do-nothing dad" image generated by men in Gager's (1998) study. The process of making downward social comparisons, suggest Heckhausen and Brim (1997), serves to "maximize and protect motivational and emotional resources of the individual" (p. 610).
Social downgrading is thought to be most likely to occur in domains in which individuals have experienced problems. As division of labor issues remain a thorny topic for many couples, especially those in dual-earner marriages, we can expect that individuals accrue mental health benefits through the negative characterization of the workload assumed by others their age. A promising avenue for further study would be the full examination of the explicit association between the extent of social downgrading and the level of conflict around division of labor issues for men and women in dual- and single-earner marriages.
Most (nonclinical) adults prefer to view their own situations as positive; thus, they will choose to attend to information from the social world that casts their own situation in the most favorable light (Taylor & Brown, 1999). Emphasis in future studies also should be directed toward identifying social referents who make one's own situation seem favorable. Our interviews suggest that it would be promising to probe directionality and intensity of social comparisons in future studies and ask directly about the extent of marital conflict over division of labor issues.
CONCLUSIONS
Limitations of the current study must be considered when interpreting results. This convenience sample was largely White and, on average, rather well educated and middle class. Stress and strain are likely more intense among dual-earner families with less income because financial realities and occupational constraints exert additional pressures on individuals and families. Furthermore, the study focuses on social comparisons related to the time-consuming, tedious, feminine-type household tasks. Because some feminine-type tasks were not included, and as men were employed an average of 14 hours a week more than women, this study does not address the full complement of time and labor allocation issues in contemporary households.
864      JOURNAL OF FAMILY ISSUES / October 2003
Strengths of this study include the research design, the very recent data collection, and the consideration of peers, spouses, and past family models in the social comparisons. Other strengths of our study are the fuller empirical assessment of social comparison beyond the typical single-item approach and the inclusion of another dimension of well-being in research on social comparisons and the division of labor.
In general, social comparison processes are most likely to occur when individuals feel uncertain about some domain in their lives (Festinger, 1954). Social comparisons inform our understanding of complex family processes because they provide a link between adults and their social worlds. Among contemporary parents, division of labor issues tap into insecurities and concerns about the roles of parent, spouse, and worker. Social comparisons may actually help parents cope and find their way as they negotiate the many challenges of contemporary family life.
NOTE
1. Preliminary analyses revealed that several potential influences on parental psychological well-being were in fact unrelated to role strain: education, income, occupational prestige, number of children living in the home, and whether the respondent was one of a husband-wife pair who participated in the study. Consequently, these variables were not controlled in the covariance structure analysis. The final model includes work hours and gender role ideology as control variables.
REFERENCES
Arbuckle, J. L. (1997). AMOS user's guide. Chicago: SmallWaters.
Arrighi, B. A., & Maume, D. J. (2000). Workplace subordination and men's avoidance of housework. Journal of Family Issues, 21, 464-487.
Barnett, R. C, &Baruch, G. K. (1985). Women's involvement in multiple roles and psychological distress. Journal of Personality & Social Psychology, 49, 135-145.
Bird, C E. (1999). Gender, household labor, and psychological distress: The impact of the amount and division of housework. Journal of Health and Social Behavior, 40, 32-45.
Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs, NJ: Prentice Hall.
Bureau of Labor Statistics. (2003). Employment characteristics of families in 2002. Retrieved from http://www.bls.gov/news.release/famee.nrO.htm
Buunk, B. P., Collins, R. L., Taylor, S. E., VanYperen, N. W., & Dakof, G. A. (1990). The affective consequences of social comparison: Either direction has its ups and downs. Journal of Personality and Social Psychology, 59, 1238-1249.
Himsel, Goldberg / DIVISION OF HOUSEWORK      865
Buunk, B. P., VanYperen, N. W., Taylor, S. E., & Collins, R. L. (1991). Social comparison and the drive upward revisited: Affiliation as aresponse to marital stress. European Journal of Social Psychology, 21, 529-546.
Cowan, C. P., & Cowan, P. A. (1988). Who does what when partners become parents: Implications for men, women, and marriage. Marriage and Family Review, 12, 105-131. (Original work published 1979)
Crosby, F. J. (1991). Juggling: The unexpected advantages of balancing career and home for women and their families. New York: Free Press.
Davis, J. A., & Smith, T. W. (1996). NORC—General Social Survey. Chicago: National Opinion Research Center.
Diener, E., & Fujita, R. (1997). Social comparisons and subjective well-being. In B. P. Buunk & F X. Gibbons (Eds.), Health, coping, and well-being: Perspectives from social comparison theory. Mahwah, NJ: Lawrence Erlbaum.
Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Lawrence Erlbaum.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.
Gager, C. T. (1998). The role of valued outcomes, justifications, and comparison referents in perceptions of fairness among dual-earner couples. Journal of Family Issues, 19, 622-648.
Glaser, B., & Strauss, A. (1967). Discovery of grounded theory. Chicago: Aldine.
Glass, J., & Fujimoto, T (1994). Housework, paid work, and depression among husbands and wives. Journal of Health and Social Behavior, 35, 179-191.
Golding, J. M. (1990). Division of household labor, strain, and depressive symptoms among Mexican-Americans and non-Hispanic Whites. Psychology of Women Quarterly, 14, 103-117.
Greenberger, E. (1988). New measures for research on work, parenting, and the socialization of children. Irvine: University of California, Program in Social Ecology.
Greenberger, E., Goldberg, W. A., Hamill, S., O'Neil, R., & Payne, C. K. (1989). Contributions of a supportive work environment to parents' well-being and orientation to work. American Journal of Community Psychology, 17, 755-783.
Greenberger, E., & O'Neil, R. (1993). Spouse, parent, worker: Role commitments and role-related experiences in the construction of adults' well-being. Developmental Psychology, 29, 181-197.
Hawkins, A. J., & Crouter, A. C. (1991). Without map or compass: Finding the way in contemporary dual-earner marriages. In J. V. Lerner & N. L. Galambos (Eds.), Employed mothers and their children (pp. 211-235). New York: Garland.
Hawkins, A. J., Marshall, C. M., & Allen, S. M. (1998). The Orientation Toward Domestic Labor Questionnaire: Exploring dual-earner wives' sense of fairness about family work. Journal of Family Psychology, 12, 244-258.
Hawkins, A. J., Marshall, C, & Meiners, K. (1995). Exploring wives'sense of fairness about family work. Journal of Family Issues, 16, 693-721.
Heckhausen, J., & Brim, O. G. (1997). Perceived problems for self and others: Self-protection by social downgrading throughout adulthood. Psychology & Aging, 12, 610-619.
Hochschild, A. (with Machung, A.). (1989). The second shift: Working parents and the revolution at home. New York: Viking.
Hoffman, L. W., & Youngblade, L. M. (1999). Mothers at work: Effects on children's well-being. New York: Cambridge University Press.
866      JOURNAL OF FAMILY ISSUES / October 2003
Hollingshead, A. B. (1975). Four factor index of social status. New Haven, CT: Yale University, Department of Sociology.
Kandel, D. B., Davies, M., & Raveis, V. H. (1985). The stressfulness of daily social roles for women: Marital, occupational and household roles. Journal of Health & Social Behavior, 26, 64-78.
Kessler, R. C, & McRae, J. E. (1982). The effects of wives' employment on the mental health of married men and women. American Sociological Review, 47, 216-227.
Lennon, M. C, & Rosenfield, S. (1994). Relative fairness and the division of housework: The importance of options. American Journal of Sociology, 100, 506-531.
Marsiglio, W. (1995). Fathers' diverse life course patterns and roles: Theory and social interventions. In W. Marsiglio (Ed.), Fatherhood: Contemporary theory, research, and social policy. Thousand Oaks, CA: Sage.
Meyers, S. A. (1993). Adapting parent education programs to meet the needs of fathers: An ecological perspective. Family Relations, 42, 447-452.
Perry-Jenkins, M., & Crouter, A. C. (1990). Men's provider-role attitudes: Implications for household work and marital satisfaction. Journal of Family Issues, 11, 136-156.
Pleck, J. H. (1997). Paternal involvement: Levels, sources, and consequences. In M. E. Lamb (Ed.), The role ofthe father in child development (3rd ed., pp. 66-103). New York: John Wiley.
Presser, H. B. (1994). Employment schedules among dual-earner spouses and the division of household labor by gender. American Sociological Review, 59, 348-364.
Radloff, L. S. (1975). Sex differences in depression: The effects of occupation and marital status. SexRoles, 1, 249-265.
Reise, S. P., Widaman, K. F., & Pugh, R. H. (1993). Confirmatory factor analysis and item response theory: Two approaches for exploring measurement invariance. Psychological Bulletin, 114, 552-566.
Robinson, J., & Spitze, G. (1992). Whistle while you work? The effect of household task performance on women's and men's well-being. Social Science Quarterly, 73, 844-861.
Rubin, L. B. (1985). Just friends: The role of friendship in our lives. New York: Harper & Row.
Sanchez, L., & Kane, E. W. (1996). Women's and men's constructions of perceptions of housework fairness. Journal of Family Issues, 17, 358-387.
Sarnoff, I., & Zimbardo, P. G. (1961). Anxiety, fear, and social facilitation. Journal of Abnormal and Social Psychology, 62, 597-605.
Speilberger, C D. (1983). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.
Taylor, S. E., & Brown, J. D. (1999). Illusion and well-being: A social psychological perspective on mental health. In R. F. Baumeister (Ed.), The self in social psychology (pp. 43-68). Philadelphia: Psychology Press/Taylor & Francis.
Taylor, S. E., Wayment, H. A., & Carrillo, M. (1996). Social comparison, self-regulation, and motivation. In R. M. Sorrentino & E. T Higgins (Eds.), Handbook ofmotivation and cognition (pp. 3-27). New York: Guilford.
Thompson, L. (1991). Family work: Women's sense of fairness. Journal of Family Issues, 12, 181-196.
VanYperen, N. W., & Buunk, B. P. (1991). Sex-role attitudes, social comparison, and satisfaction with relationships. Social psychology quarterly, 54, 169-180.
Wilkie, J. R., Ferree, M. M., & Ratcliff, K. S. (1998). Gender and fairness: Marital satisfaction in two earner couples. Journal ofMarriage and the Family, 60, 577-594.
Yogev, S., & Brett, J. (1985). Perceptions of the division of housework and child care and marital satisfaction. Journal ofMarriage and the Family, 47, 609-618.