IN SEXUALITY

Although there has been much theorizing about gender differences in sexuality, and many scattered empirical studies, there was a need to integrate this theorizing and research in a systematic manner. To meet this need, we conducted a meta-analysis of studies reporting data on gender differences in sexual behaviors and attitudes (Oliver & Hyde, 1993).

We were able to locate 177 usable studies that reported data on gen­der differences in 21 different sexual attitudes and behaviors: attitudes about premarital intercourse (general); attitudes about premarital inter­course in a casual relationship; attitudes about premarital intercourse in a committed relationship; attitudes about premarital intercourse when the couple are engaged but not married; attitudes toward homosexuality1; at­titudes about civil liberties for homosexuals; attitudes about extramarital sex; general sexual permissiveness; anxiety, fear, or guilt about sexuality; sexual satisfaction; double standard attitudes (i. e., beliefs that sexual activ­ity is more acceptable for males than for females); attitudes about mastur­bation; incidence of kissing; incidence of petting; incidence of intercourse; age at first intercourse; number of sexual partners; frequency of intercourse; incidence of masturbation; incidence of homosexual behavior; and incidence of oral sex. (See Oliver & Hyde, 1993, for details on methods of identifying and coding studies.) The studies represented, altogether, the testing of 128,363 respondents (58,553 males and 69,810 females).

For each study, we computed an effect size, to measure the magnitude of the gender difference. We used the formula

^ _ MM — MF

s

where MM is equal to the mean for males, MF is equal to the mean for females, and s is the average within-gender standard deviation. The d sta­tistic essentially measures how far apart the male and female means are, in standard deviation units. A positive value of d therefore indicates that males scored higher than females, whereas a negative value indicates that [1] females scored higher. When means and standard deviations were not avail­able, we used other formulas provided by Hedges and Becker (1986). Effect sizes for a given attitude or behavior were then averaged over studies fol­lowing methods detailed by Hedges and Becker (1986). In interpreting the magnitude of effect sizes, we generally follow the guidelines stated by Co­hen (1969): a d value of.20 is small, .50 is moderate, and.80 is large.

The resulting effect sizes for the 21 sexual attitudes and behaviors are shown in Table 1. Most striking are the two largest gender differences. One is the incidence of masturbation (d = 0.96). Kinsey found that 92% of males, compared with 58% of females, masturbate to orgasm at least once in their lives (Kinsey, Pomeroy, Martin, & Gebhard, 1953). Despite mas­sive changes in patterns of sexual behaviors in the 50 years since Kinsey collected his data, this marked gender difference remains. Relevant data are found in the recent, well-sampled national survey of sexual behavior conducted by the National Opinion Research Center (NORC) (Laumann, Gagnon, Michael, & Michaels, 1994). Although they did not compute lifetime incidence statistics, they did inquire about incidence in the past year. The results indicated that 42% of women, compared with 63% of men, masturbated in the last year. This pattern, however, varies as a func­tion of ethnicity. The large gender gap was found for Whites (67% of men,

44% of women) and Hispanics (67% of men and 34% of women) but was smaller for Blacks (40% of men and 32% of women).

The other large gender difference was in attitudes about premarital sex in a casual relationship (d = 0.81). Males were considerably more ap­proving than females were, as predicted by all the theories reviewed earlier. This gender difference is illustrated by the findings of an innovative study by Clark and Hatfield (1989). An attractive man or woman confederate approached strangers of the other gender on a college campus and posed one of the following three questions: “I have been noticing you around campus. I find you very attractive, (a) Would you go out with me tonight? (b) Would you come over to my apartment tonight? (c) Would you go to bed with me tonight?” Of those approached for a date, roughly 50% of women accepted, as did 50% of men (i. e., there was no gender difference). The question of interest here, though, is the invitation for casual sex; 75% of men agreed to it, compared with only 6% of women. Men evidenced considerably more interest in casual sex. An alternative interpretation of these results is that women responded negatively to the invitation for ca­sual sex because of personal safety issues, namely that it carries with it a potential for rape. These two interpretations are not mutually exclusive, of course; both factors may play a role.

Many other gender differences were moderate in magnitude, accord­ing to the meta-analysis (Oliver & Hyde, 1993). Males were more sexually permissive (d = 0.57), and females reported more anxiety, fear, and guilt about sex (d = —0.35).

Still other behaviors showed no gender difference. We advocate in­terpreting d values of 0.10 or less (in absolute value) as indicating no difference. Using this guideline, there were no gender differences in atti­tudes about civil liberties for homosexuals ( — 0.00), sexual satisfaction (—0.06), or attitudes about masturbation (0.09).

One of our findings of no gender difference was later contested. We found d = —0.01 for attitudes toward homosexuality (Oliver & Hyde, 1993). Whitley and Kite (1995) located a number of additional studies of gender differences in attitudes toward homosexuality, which we had not located for our original meta-analysis. With the addition of these studies, they found d = 0.26; that is males hold more negative attitudes. The mag­nitude of this difference is larger than the near-zero value we found, but it is still small. We were concerned that most of the studies added by Whitley and Kite were based on convenience samples of college students, whereas our original set of studies had contained well-sampled studies of the general population (Oliver & Hyde, 1995). The effect size found by Whitley and Kite may tell us more about college students than anything else. In addi­tion, Whitley and Kite had a number of inflated computations of effect sizes. We concluded that the existence and magnitude of gender differences in general attitudes toward homosexuality is an open question, but is almost certainly small (Oliver & Hyde, 1995).

Whitley and Kite (1995) noted that gender differences in attitudes toward homosexuality may vary as a function of the gender of the person being rated (i. e., whether the study measures attitudes toward lesbians or attitudes toward gay men). Supporting this view, they found d = 0.51 for ratings of gay men, —0.03 for ratings of lesbians, and 0.35 for ratings with gender unspecified. Men hold considerably more negative attitudes about gay men than women do, whereas men and women do not differ in their attitudes about lesbians. This is an illustration of the complex ways in which gender and sexual orientation interact.

We turn now to a consideration of gender differences in some sexual practices that were not reviewed by Oliver and Hyde (1993).