Introduction

Explorations into the extent and meaning of sex differences continue, in an environment of energetic debate over causes stemming from nature or nurture (Archer, 2006; Eagly, 1995; Hyde, 2005; Mealey, 2000). Males and females differ because of distinct selection pressures throughout human evolution, pressures for defense, foraging, and reproductive investment, among others. At the same time, cultures provide additional pressures which may accentuate or attenuate these tendencies for the sexes to appear and behave somewhat differently (Friedl, 1975). While sex differences apparent at the point of initial mate choice in humans are well documented (Buss, 2003; Furnham, 2009; Lippa, 2010), less is known about sex differences in the latter stages of the human life cycle. The literature on couples in the United States (U.S.) has suggested that, with the arrival of children, parents may drift into more traditional gender roles as defined by culture, and so sex differences may become more pronounced (e.g., Cowan, Cowan, Coie, & Coie, 1978). Gutmann (1972) suggested that couples may revert to the earlier, less-differentiated behavioral patterns once their “parental imperative” has been fulfilled. Such shifting in the patterns of sex differences has led some (e.g., Canary & Emmers-Sommer, 1997) to suggest that humans display marked plasticity in gender roles, indeed “doing gender” or “creating gender” over the life course. One way to explore this hypothesized plasticity is to examine gender differences cross-culturally; if very different social structural systems produce similar patterns in males and females, one would suspect that there are some underlying predispositions not amenable in significant ways to cultural influence, even if they are not actually evolved (for examples of literature reviews synthesizing biocultural influences, see Lippa, 2007; Rohner, 1976). On the other hand, if sex differences are not found consistently across cultures, then those sex differences that do occur are likely to be essentially cultural.

Previous work on British couples (Russell, Wells, Weisfeld, & Weisfeld, 1992) explored the notion of life course plasticity and found no evidence that sex differences were influenced by the presence or absence of children, beyond employment/income differences. In a subsequent exploration of the interplay between gender and culture (Weisfeld & Karana, 2003), differences between husbands and wives were wide-ranging and fairly consistent across cultures, with or without children in the home. These findings raised further questions concerning the extent of sex differences throughout the life cycle, and the role that culture may or may not play. This article will focus on sex differences in adults in a variety of cultures, in an effort to clarify some of these issues. As Lippa (2007) put it, the goal is to enable us to “make more precise and nuanced predictions about the patterning of sex differences across cultures” (p. 631).

Most adults in most cultures are married at some point in their lives. Married couples exhibit a high degree of similarity, or homogamy, in age, socioeconomic status, ethnicity, education, and religion. The data on personality variables are mixed, with results showing a low-level match for some traits such as neuroticism, but less of a clear trend for other traits such as extraversion (Jenkins, 2006; Russell & Wells, 1991; Slatcher & Vazire, 2009). As such, married couples offer a convenient and well-matched sample of men and women for scientists who are interested in studying sex differences in adults. In this study presented here, we examined the patterns of sex differences in men and women married to each other in five cultures (China, Russia, Turkey, UK, and the U.S.) to look for universal patterns in physical and behavioral dimorphism, and cultural modifications of those patterns. All couples had completed appropriately translated versions of the Marriage and Relationship Questionnaire (MARQ) developed by Russell and Wells (1986). The MARQ’s scales have been shown to have good invariance across genders and cultures (Lucas et al., 2008; Wendorf, Lucas, İmamoğlu, Weisfeld, & Weisfeld, 2011).

The MARQ is a general-purpose relationship questionnaire consisting of 235 items inquiring about perceptions of one’s life with a long-term partner. We anticipated that we might find sex differences in response to questions related to these six categories: female choosiness, labor performed, emotional expressiveness, physical attractiveness, interest in sexual activity, and jealousy. These expectations deserve further explanation. The psychology literature on sex and gender differences (Hoyenga & Hoyenga, 1979; Lippa, 2010; Mealey, 2000) documents fairly well the tendency for males to desire more frequent and more varied sex, on average, than females do; that literature also suggests that, while females may be more expressive of most emotions than males are, males tend to become more easily and more sexually jealous about their partners. This latter finding has been studied in detail by evolutionary psychologists (e.g., Buss, 2003) who find that males may be more jealous of sexual rivals, while females are more bothered by emotional or resource infidelity. The cross-cultural data suggest that division of labor by sex is fairly universal although responsibility for a specific task may vary from culture to culture (Canary & Emmers-Sommer, 1997; Daly & Wilson, 1983; Friedl, 1975; Murdock, 1965). The evolutionary psychology literature suggests that both men and women value physical attractiveness in short and long-term partners, but cross-culturally men put a greater emphasis on this characteristic than women do (Buss, 1989; Lippa, 2010; Singh, 1993). In a study in which American adults had to make tradeoffs among characteristics in a potential partner, even with such constraints women focused more on social status and men more on physical attractiveness in a mate (Li, Bailey, Kenrick, & Linsenmeier, 2002). Lastly, because sexual selection is characterized by females being the more choosy or discriminating sex (Darwin, 1871; Mealey, 2000; Miller, 2001), we anticipated that females might continue to be more judgmental and less accepting of the status quo, beyond the initial point of mate choice.

Method

Participants

Over 2000 married couples from largely urban areas in five racially, religiously, economically, and geographically dissimilar countries participated in the study. Characteristics of these samples and how they were recruited are described in Lucas et al. (2008). Table 1 provides a summary of the demographic variables of interest in this study. Importantly, it should be noted that, in each of the samples, husbands were slightly older than wives, that couples had been married on average for more than 11 years, and that couples had, on average, more than one child. Table 1 also shows the degree to which males and females within each sample matched each other on demographic variables, thus providing matched samples for sex differences research.

Table 1 Sample demographics
Full size table

Measures and Procedure

All couples completed the MARQFootnote 1 (described in Russell & Wells, 1993), an empirically-developed, multi-purpose marriage assessment instrument originally normed on the British sample. Turkish couples completed appropriately translated and back-translated versions of the MARQ, to which questions of local interest were added (İmamoğlu & Yasak, 1997; Russell, Wells, & Imamoğlu, 1989). The American version contained minor changes from the British version, reflecting vocabulary and spelling differences between the countries. The U.S. sample included Caucasian and African-American couples (the sample is about 80% Caucasian.) The Russian and Chinese versions (Russell, Wells & Butovskaya, 2003; Russell, Wells & Shen, 1999) underwent the same translation process, and questions of local interest were added as well. All versions of the MARQ included a set of demographic questions which asked participants to indicate, among other things, the length or duration of their marriage (in years), their ages (in years), and the number of children they have.

The MARQ does not contain a scale relating to sex differences; consequently, the U.S.-based members of the research team examined the MARQ to find questions related to the sex categories identified above. Those members of the research team who reside outside the U.S. provided input as items were selected, and they also provided assistance with data analysis and interpretation. As a result of this collaborative effort, six specific questions were selected relating to each of three categories (female choice, division of labor, and interest in sex) and four specific questions were selected relating to each of the remaining three (emotional expressiveness, physical attractiveness, and jealousy). In sum, then, the design anticipated the emergence of sex differences in response to 30 questions. The 30 questions are listed in the “Results” section of this article, Tables 2–7, and each question was re-worded to reflect the expected sex difference. For example, the first female choice question is, “Have you thought of divorcing your spouse?” with five response choices in a Likert-type scale, ranging from “Never” to “I intend to.” The prediction was that women would answer more frequently in the affirmative; thus, the question was re-worded in Table 2 to read, “Wife has thought of divorcing him.”

Table 2 Effect sizes of mean differences between husbands and wives on female choosiness items
Full size table
Table 3 Effect sizes of mean differences between husbands and wives for division of labor items
Full size table
Table 4 Effect sizes of mean differences between husbands and wives for emotional expressiveness items
Full size table
Table 5 Effect sizes of mean differences between husbands and wives for physical attractiveness items
Full size table
Table 6 Effect sizes of mean differences between husbands and wives for interest in sex items
Full size table
Table 7 Effect sizes of mean differences between husbands and wives for jealousy items
Full size table

Data Analysis

For each individual question in each culture, male and female responses were averaged separately and compared by a t-test for matched pairs (two-tailed). Thus, a total of 150 t-tests were conducted (5 cultures by 30 questions) in order to provide profiles of sex differences and cultural patterns. With such a large number of tests, one would anticipate findings that might be seen as artifacts, and so it is more helpful to examine effect sizes of the differences. Consequently, where the t-test resulted in a significant difference (p < .05, two-tailed test for matched pairs), the effect size was calculated, using Cohen’s d statistic. In each table, then, the effect size is noted at the intersection of the cultural group and the specific question. (Use of Cohen’s d also addresses concerns about differences in sample size in studies such as this one.) As Cohen (1977) suggested, effect sizes of .20 are small, .50 is a moderate effect, and .80 is a large effect. Finally, the general patterns in sex differences were examined to see if culture seemed to generate an exception to the pattern.

Results

Tables 2–7 show the results of 150 t-tests, in six categories of questions where sex differences in these married respondents were anticipated. A straight line in one of the tables numbered 2–7 indicates no significant sex difference, and a negative sign before the effect size indicates a reversal of the expected direction of significant difference. Thus, these analyses also attempt to address common criticisms of biased or incomplete findings on sex differences (Canary & Emmers-Sommer, 1997; Hyde, 2005). As the tables indicate, some significant sex differences (p < .05, two-tailed) emerged in all six areas examined, although cultural differences were seen in the patterns.

Table 2 shows items related to Female Choosiness. Women in all groups were more likely than their husbands to endorse the statement: “I have thought of divorcing my spouse.” The wife’s parents played a role in choosing the husband in every group more than the husband’s parents in choosing the wife, and in two of these groups the wife’s family did not like the man whom the wife chose. In Great Britain and Russia, wives were more likely to say, “I would not choose him again.” Husbands, on the other hand, in four groups (all but the U.S. group) felt fortunate to have this wife and said it was love at first sight. Most effect sizes were small.

Table 3 shows items related to Division of Labor, with large effect sizes, and with clear demarcation by gender: Men were making more money (although full-time labor did not always sort by gender), while wives did more housework and seemed to enjoy it more. Interestingly, wives everywhere helped their husbands to choose their clothes.

Table 4 shows items related to Emotional Expressiveness. In five of six groups, wives seemed to worry more and to be more moody. These differences tended to show small to medium effect sizes. “Bottling up one’s feelings” did not show a consistent sex difference. The Chinese sample had the lowest degree of sexual dimorphism in emotional expressiveness.

Table 5 shows results demonstrating that wives everywhere agreed that their husbands attended less to their own appearance. Husbands in five of six groups said they found other women attractive. Husbands in three of six groups were significantly more likely to find the spouse attractive than the other way around, and the wives judged themselves more attractive than husbands found themselves to be, in three of six groups. Again, these differences showed small to medium effect sizes.

Table 6 shows a fairly consistent pattern of differences in Interest in Sex. Husbands saw sex as more important, wanted their wives to be more responsive, enjoyed sexy books and videos more than wives did, and said they found sexual fulfillment outside of marriage. Wives, on the other hand, outside of Turkey, were more likely to have had sex against their will. These differences also showed small to medium effect sizes.

Table 7 shows few consistent findings regarding Jealousy. Wives in three groups (not China and Russia) worried more about husbands being unfaithful, and wives saw the husband’s prior marriage as problematic. Possessiveness showed no consistency.

While the consistency in patterns of differences was quite striking in the results reported above, some exceptions are noteworthy. Chinese couples showed only one of four expected differences in emotional expressiveness, with the wife more likely to be a worrier. Husbands and wives were similar to each other in moodiness, loneliness, and bottling up one’s feelings. Scores in these four categories were uniformly low, which may have brought all scores low enough to eliminate differences between the sexes.

Similarly, few sex differences emerged for the Chinese and Russian couples on items related to jealousy. One reversal occurred, with wives feeling more possessive about their husbands than the other way around in China. Scores on questions about prior marriage for the Chinese couples yielded very high scores, which may reflect the cultural taboo about divorce in China, and the general sense that one who is divorced has serious character flaws.

Discussion

One way to understand these data is to look at the sheer volume of differences. In the five cultural groups included here, a total of 150 male–female comparisons were made. Of these, 116 comparisons produced male–female differences in the expected direction that were statistically significant. This amounts to 77% of comparisons demonstrating sex differences in the anticipated direction. In 30 cases, there were no significant differences and, in 4 cases (2.6%), the anticipated difference was reversed.

The patterns suggest a possible underlying logic. It may be useful to consider to what extent behavioral niches are described here, niches in which the behavior of one sex influences the behavior of the other in a synergistic manner. Some of these niches are familiar to behavioral scientists, such as the categories “Division of Labor” and “Interest in Sex.” But it is suggested that some of the differences reported here were rather surprising and would not have surfaced outside the context of a multi-purpose questionnaire such as the MARQ. One example is the finding that wives everywhere help husbands to choose their clothes. Were their husbands color-blind and in need of assistance with color coordination? Were husbands not attending to cleanliness to the same extent as their wives, because of the latter’s greater involvement in infant care, or because wives are held responsible for the appearance of the family members? Were husbands dressing down so that wives would not suspect infidelity? Husbands may neglect their personal appearance and concede clothing selection to their wives partly out of fear of appearing overly fastidious. On the other hand, in many societies men often maintain highly decorative appearances indicative of social status, such as a royal or priestly class, while women’s dress typically indicates marital status. Alternatively, it may be that these wives were simply being more critical, in one more way, as wives were in the items in the “Female Choice” category. Again, in the “Division of Labor” niche, wives everywhere did more housework, and husbands believed that their wives enjoy doing housework, more than the other way around, in all six groups. Surveying the cross-cultural literature, van den Berghe (1980) reported that women predominate at these tasks. Not being responsible for as much housework and child care, men in most cultures are expected to contribute more to subsistence. But what is surprising is that wives in all six groups reported that they do indeed enjoy doing housework, more than their husbands do. So the husbands’ assessment of their wives’ feelings about housework may come less out of laziness than out of observation. To test that possibility, within-couple correlations were run for responses in each culture, looking at the correlation between “How much does your wife enjoy doing things about the house?” and the wives’ question “How much do you enjoy doing things about the house?” In each culture, the correlations were significant at p < .01, effect sizes small to moderate, suggesting that husbands were describing their wives’ feelings fairly accurately.

The finding that women worried more than men everywhere is consistent with the greater fearfulness of females in most situations (e.g., Ekehammer, 1974). This is not just an artifact of greater acceptability for women to acknowledge fear, since females show more sympathetic arousal in fear-inducing situations than men (Ekehammer, Magnusson, & Ricklander, 1974). Also, androgens lower anxiety (Ellis, 1986; Jacklin, Maccoby & Doering, 1983).

The behavioral niche related to “Female Choice” is of particular interest. Because female choosiness is such a basic evolutionary concept, first described by Darwin (1871) and only more recently regaining a central place in scientific inquiry (Miller, 2001), it offers great heuristic promise. The results documented in Table 2 showed a consistent pattern across cultures, in which females are moderately more likely than their spouses are to consider the possibility of divorce. This is consistent with the finding that, universally, women are more likely to initiate divorce, although men are more likely to simply desert a spouse (Mealey, 2000). While it is not surprising that parents played a role in spousal choice in Turkey and China (in view of the close-knit family traditions there), one may be skeptical that parents still play a role in spousal choice in Russia, Great Britain, and the U.S. What is important here is that wives in those countries reported it more often than husbands did, making the finding consistent with greater female parental investment (Trivers, 1972), even though in all likelihood parents and daughters are not conscious of this parental role until asked about it. It is striking that wives in Russia and Great Britain were more likely to say no to such an extreme statement as “If you could choose, would you marry the same person again?” This profile strongly suggests that female choosiness continues right through the course of marriage—that wives continue to evaluate, judge, compare, and consider eventual rejection of the man they married.

Husbands, for their part, seemed to reflect a relative lack of choosiness. In four of the five countries (outside the U.S.) husbands reported “love at first sight” and more often endorsed the statement, “I feel fortunate to have married my wife.” Similarly, in other studies, U.S. men reported falling in love earlier and more often than young women (Kanin, Davidson, & Scheck, 1970), and reported more unhappiness after a dating breakup than women (Hill, Rubin, & Peplau, 1976). U.S. women, on the other hand, were more likely to marry someone they did not love (Kephart, 1967). Perhaps the lower male threshold for amorousness, as Buss (2003) and Mealey (2000) both pointed out, is because men are more focused on physical attractiveness and less on behavioral characteristics—or because of their greater eagerness for sex. In functional terms, men may have evolved a tendency toward passionate love in order to assure a reluctant woman of their commitment. What is striking is that, for both husbands and wives, these initial differences in choosiness apparently persist through the life course, with differences in tendencies to be judgmental being maintained.

The findings in Table 7 deserve some comment, as the results were not consistent and showed two reversals. It is possible that questions about prior marriages and past relationships have little relevance to feelings of jealousy in China and Russia. Those concepts may be more likely to evoke a sense of character weakness in someone who was not able to make a relationship work, despite support from an extended family. A similar problem may arise with use of the concept of possessiveness, and it may be that refinement of vocabulary is required to investigate these ideas further.

The finding of greater jealousy of males in response to the spouse’s sexual infidelity has also been reported in many other countries and cultures (Brase, Caprar, & Voracek, 2004; Buunk, Angleitner, Oubaid, & Buss, 1996; de Souza, Pereira, Taira, & Otta, 2006; Wiederman, & Kendall, 1999). Self-report data have been corroborated by documentation of sex differences in physiological arousal and emotional expressiveness in response to imagining sexual infidelity by the mate (Buss, Larsen, Westen, & Semmelroth, 1992; Sagarin, 2005).

Limitations

In this study we have not controlled for possible mediating variables such as education, nor have we examined other variables such as United Nations indices of nations’ gender equality. Differences observed between men and women often seem to be parallel to those of power relations or status differences, and they may change in time or with differential experiences. We have tried to address some of these difficulties by examining different cultural groups, but the challenges are large and complex, and it is difficult to make generalizations with a great degree of confidence. In terms of data collection, our respondents were self-selected. Although most were volunteers and not compensated for volunteering, both husband and wife were willing to complete a lengthy questionnaire that included many personal questions. Lastly, in this study, we have used single-item measures rather than scales—although the practice is defended by Wanous, Reichers, and Hudy (1997).

Conclusion

These findings were seen in couples who have been together an average of 12 years and who were already matched on many factors such as age and socioeconomic status. Long-term marriage may be characterized by a balance between homogamy and dimorphism serving reproductive interests. Women retain high standards which men struggle to meet, while men seem more easily satisfied, focusing on physical attractiveness and maximizing sexual enjoyment opportunities. Men continue to provide more monetary resources, while women work more at home and have more emotional challenges. Cultural exceptions are few (e.g., low emotionality in general in China), suggesting that these aspects of sexual dimorphism may serve adaptive functions in human marriage.

While some (e.g., Canary & Emmers-Sommer, 1997) suggest that it is time to stop examining sex differences, these results suggest that we have just begun to examine sex differences. The MARQ has revealed other interesting sex differences, as well, reported elsewhere, such as sex differences in humor (Weisfeld et al., in press). We concur with others that it is important to document and understand sex differences—without gainsaying the fact that considerable behavioral variability exists within each sex. It is particularly important to document cultural patterns in sexual dimorphism, as they reveal biological and social structural influences on human behavior. Moreover, continuities in the lifespan offer additional clues about evolutionary adaptations in males and females. Understanding sex differences can only enhance the skills of human service professionals, including healthcare professionals, mental health professionals, and educators. Ultimately, acknowledgment of sex differences may facilitate intersexual cooperation and reduce conflict.