Lesbian and bisexual women's judgments of the attractiveness of different body types.

Because little is known about what features lesbian and bisexual women are attracted to (Bailey, Kim, Hills, & Linsenmeier, 1997), the ways in which certain physical traits affect attractiveness ratings by lesbian and bisexual women is a theoretically important issue. Interestingly, the preferences of lesbians seem to be similar in certain ways and different in certain ways from those of heterosexual men and women.

In one study of age preferences, for instance, lesbians were shown to differ somewhat both from heterosexual men and from heterosexual women, in that when they were older they began to be interested in younger women while continuing to be interested in women their own ages and older (Kenrick, Keefe, Bryan, Barr, & Brown, 1995). Further, Bailey, Gaulin, Agyei, and Gladue (1994) found that lesbians were similar to heterosexual women in many areas of sexuality, such as lack of interest in uncommitted sex, emotional (rather than sexual)jealousy, and preference for older partners, but that lesbians were less concerned than heterosexual women with a partner's financial status, and lesbians were more interested in visual sexual stimuli than heterosexual women.

We believe that an important topic for study is the physical factors which influence attraction among nonheterosexual women. We chose to focus on physical traits because for both male and female figures, women and men rate body build as the single most important trait in physical attractiveness (Pederson, Markee, & Salusso, 1994). Given the current dearth of research on attraction in nonheterosexual populations, such an investigation may be illuminating regarding mate preferences in general, and it may bridge topics in sex research related to sexual orientation and attraction. In general, our investigation is not so much theory driven as empirically driven, given the early state of research intersecting these topics.

Waist-to-hip Ratio

Waist-to-hip ratio, which is an indication of the pattern of fat deposition on the body, is perhaps the best-researched characteristic in terms of its relation to attractiveness ratings. Female figures with low, typically feminine waist-to-hip ratios (around 0.7) are rated more attractive than those with typically masculine waist-to-hip ratios (around 1.0). The similarity in ratings between men and women, and across cultures and ethnic groups, is quite striking (Furnham & Baguma, 1994; Furnham, Dias, & McClelland, 1998; Furnham, Tan, & McManus, 1997; Henns, 1995; Singh, 1993a, 1993b, 1994a, 1994b; Singh & Luis, 1995; Singh & Young, 1995; Tassinary & Hansen, 1998).

If one were to attempt to predict lesbian and bisexual women's judgments of figures varying in waist-to-hip ratio, one might first turn to the evolutionary biology literature just cited, and predict that, as women may have evolved to be attracted to figures with a typically masculine, high waist-to-hip ratio, lesbian and bisexual women would be attracted to female figures with such a waist-to-hip ratio. However, we did not consider this a likely outcome because, as mentioned above, lesbians' preferences do not seem to be able to be predicted from the preferences of heterosexual women or from those of heterosexual men. Moreover, women have been shown to rate female figures with a low waist-to-hip ratio, large breasts, and slender body weight in the same ways that men do. Therefore, based on the robustness of the findings across gender and cultures that females with a low waist-to-hip ratio are rated more attractive than those with a high waist-to-hip ratio (especially Singh & Luis, 1995), we felt it was a stronger prediction that lesbian and bisexual women would prefer women with a low waist-to-hip ratio.

Breast Size

The literature on how breast size relates to preferences for male and female raters is not very consistent. Some studies suggest that females prefer small breasts to larger ones and that males prefer larger breasts than females do, although a few studies have shown little difference between men and women in ratings of different breast sizes (Beck, Ward-Hull, & McLear, 1976; Gitter, Lomranz, Saxe, & Bar-Tal, 1983; Horvath, 1981; Wiggins, Wiggins, & Conger, 1968). We note that as small and large breast size is relative, the particular stimuli used may account in large part for differences between studies. Interestingly, women across cultures seem to want larger breasts than they have, and women seem to be dissatisfied if they perceive themselves as having small breasts (Arkoff & Weaver, 1966; Jourard & Secord, 1955; Koff & Benavage, 1998; Lee, Leung, Lee, Yu, & Leung, 1996; Thompson & Tantleff, 1992). Based on this latter set of findings, and the fact that women in general were shown to rate figures with large breasts more attractive than similar figures with small breasts (Singh & Young, 1995), we predicted that lesbian and bisexual women would prefer figures with large breasts to figures with small breasts.

Body Weight

Some of the same studies that examined the role of waist-to-hip ratio in attractiveness have also looked at the role of body weight in attractiveness, and have consistently found that male and female figures of normal body weight are rated more attractive than underweight or overweight figures when controlling for waist-to-hip ratio. Importantly, cultural differences have been shown in attractiveness of figures of different body weights (Furnham & Alibhai, 1983; Furnham & Baguma, 1994), but using figures that control for waist-to-hip ratio, Singh (1994c; Singh & Luis, 1995) found that Black males and females and Indonesian students show the typical pattern of responses: normal body weight female figures with low waist-to-hip ratio are rated attractive, feminine, and healthy, but less kind and understanding than heavier figures.

We believe that lesbian and bisexual women represent another group in which a preference for heavier figures may be observed, and that, therefore, investigating body type preferences among such women is a theoretically interesting and important issue. We base the prediction of a preference for heavy figures on two related types of findings. First, it is probably safe to say that the majority of prior research, while not entirely consistent, suggests that lesbians are heavier than heterosexual women, have higher ideal weights, and are more comfortable with their bodies (French, Story, Remafedi, Resnick, & Blum, 1996; Gettelman & Thompson, 1993; Herzog, et al., 1992; Kenyon, 1968; LaTorre & Wendenburg, 1983; Perkins, 1981; Siever, 1994). It might be noted that lesbians have been shown by Beren, Hayden, Wilfley, and Striegel-Moore (1997) to consider thinness and fitness important in attraction, and Bailey et al. (1997) also showed a preference for thinness among lesbian women. Furthermore, some research suggests that lesbians are similar to heterosexual women in terms of body satisfaction (Beren, Hayden, Wilfley, & Grilo, 1996; Heffernan, 1996; Striegel-Moore, Tucker, & Hsu, 1990). Doubtless these inconsistent findings are due in part to the different measures and sample studies.

The second type of finding that makes us predict that lesbian and bisexual women would prefer heavier figures has to do with the lesser extent of internalization of societal standards of attractiveness on the part of lesbians as compared to heterosexual women. Bergeron and Senn (1998) gave participants the Sociocultural Attitudes Toward Appearance Questionnaire (Heinberg, Thompson, & Stormer, 1995), which has two subscales: awareness (assesses acknowledgment of societal emphasis on appearance) and internalization (assesses acceptance of these standards). No differences between lesbian and heterosexual women on awareness were observed, though lesbians had lower internalization scores. Higher internalization was associated with more negative attitudes toward one's own body for both lesbian and heterosexual women. Such findings may reflect a feeling on the part of lesbians that the denigration of fat women is similar to the denigration experienced by lesbians, and further that both fat women and lesbians are violating rules and resisting patriarchal control (Brown, 1987).

In these two studies, we investigated the role of three physical variables (waist-to-hip ratio, breast size, and body weight) in attraction among lesbian and bisexual women. Based on the literatures reviewed above, we predicted that lesbian and bisexual women would prefer figures with a low waist-to-hip ratio, large breasts, and heavy body weight. In the first study, we explored which figures would be preferred by lesbian and bisexual women. In the second study, we replicated the effect of Study 1 and further examined whether gender conformity related to such preferences. The effect of gender conformity on figure preferences is important because there appear to be meaningful differences between lesbians who are higher in gender nonconformity ("butch" lesbians) and those who are relatively more gender conforming ("femme" lesbians; Rosenzweig & Lebow, 1992; Tripp, 1975). Self-described butch lesbians have higher waist-to-hip ratios, higher saliva testosterone levels, and less desire to give birth as compared to self-described femme lesbians (Singh, 1999). In addition, butch lesbians reported being more gender nonconforming than femme lesbians in childhood and adulthood (Bailey & Zucker, 1995).

Both studies were conducted online for two reasons. First, a higher proportion of online participants report being homosexual or bisexual than participants from psychology classes (Bailey, Foote, & Throckmorton, 2000). Second, we reasoned that this method would tend to protect participants' anonymity to a greater extent than other methods of participant recruitment. Previous research has shown that online rating of figures produces valid results (Krantz, Ballard, & Scher, 1997).

STUDY

Methods

Questionnaire. We based the methods on Singh and Young's (1995) Study 1. The stimuli used in these studies were based on those used by Singh and Young. These stimuli consist of eight line drawings of female figures representing every possible combination of two categories of body weight (slender and heavy), two categories of breast size (small or large), and two waist-to-hip ratios (0.7 or 1.0). The figures have identical postures, facial expressions, hairstyles, and attire (a bathing suit). These figures are widely used in the literatures reviewed above. The figures from Singh and Young's paper were scanned into a computer graphics package, edited to ensure identical appearance on the characteristics not being studied (hair style, etc.), and posted on a web site. On the website, all figures were visible simultaneously (this is typical, as reviewed above) and respondents used a similar rating scale to that of Singh and Young's study; each figure was rated on a 0 (least) to 20 (most) scale for attributes of attractiveness, healthiness, feminine looking, kindness and understanding, desirability for short-term relationships, and desirability for long-term relationships. The figures were presented in the following order to all respondents (breast size/waist-to-hip ratioPoody weight)--1: small/0.7/heavy; 2: small/1.0/heavy; 3: small/1.0/slender; 4: small/0.7/slender; 5: large/1.0/slender; 6: large/1.0/heavy; 7: large/0.7/heavy; 8: large/0.7/slender.

After obtaining IRB approval, the website containing the questionnaire was e-mailed to several local gay and lesbian college organizations. The second author, an undergraduate at the University of Pennsylvania, sent the e-mail with a short description that the study was part of her undergraduate thesis and was focused on attraction among lesbian and bisexual women. When the questionnaire was completed, the data were e-mailed back to us. We made sure that each subject submitted only one set of data by examining the demographic characteristics of the respondent and the internet provider address from which the data were submitted. This did not compromise the privacy or anonymity of the participant, as this address consists entirely of numbers. The data were analyzed in Systat.

Participants. Participants self-reported their sexual orientation in an open-ended format. One hundred twenty-four subjects self-identified as homosexual, 76 as bisexual, 2 as homosexual or bisexual, 4 as queer, 2 as other, and 1 as undecided. We considered it important to investigate the preferences of bisexual women as well as those of lesbian women because bisexuality does appear to be a meaningful category apart from homosexuality (Loewenstein, 1985), though there also may be much overlap between bisexual and homosexual women in sexual experience and attraction (Rust, 1992).

Ages ranged from 17 to 57, with a mean of 25.7 and a standard deviation of 6.96. Eighty-two percent of the subjects self-reported ethnicity as White, 5% as Hispanic or Latin, 5% as Black, 4% as Asian or Asian-American, 1% as Native American, and 0.5% as Eskimo. Subjects also self-reported education level. Thirty-eight percent had a high school diploma, 1% had an associate's degree, 35% had a bachelor's degree, 22% had a master's degree, and 4% had doctorates.

Data reduction. In order to simplify data analysis we performed a principle components factor analysis with a minimum Eigenvalue of 1, Varimax rotation, and casewise deletion for each figure. The factor analysis results are in Table 1. They show one factor solutions for every figure, with the factor accounting for 52 to 60% of the variance. For every figure, attractiveness, healthiness, femininity, kindness, and desirability for short- and long-term relationships all loaded positively on the same factor.

A general preference scale for each figure was produced by summing the ratings of attractiveness, healthiness, femininity, kindness, desirability for short-term relationships, and desirability for long-term relationships. Cronbach's alphas for these scales are also in Table 1, and they show good internal consistency. Since responses on all scales could range from 0 to 20, the theoretical range for each scale was 0 to 120, with higher scores indicating high preference (i.e., higher ratings of attractiveness, healthiness, femininity, kindness, desirability for short-term relationships, and desirability for long-term relationships).

Results

As the figures were presented in the same order to all respondents, our first concern was that the order of the figures affected ratings. We calculated a correlation for each participant between the number of the figure and the preference score associated with that figure. The average within-subject correlation was very small (M = .06, SD = 0.35). We thus concluded that order did not affect ratings.

Next, we explored whether it was appropriate to analyze data collapsed across sexual orientation. We limited analysis to homosexual and bisexual participants, and performed a MANOVA with sexual orientation (homosexual or bisexual) as a between-subjects factor; breast size (large or small), weight (heavy or slender), and waist-to-hip ratio (0.7 or 1.0) as within-subjects factors; and preference scores as the dependent variables. Sexual orientation did not have a significant main effect on ratings, and did not interact with breast size, body weight, or waist-to-hip ratio, or their interactions. Therefore, aU analyses reported below are for all participants, collapsed across sexual orientation, and not limited to those participants who reported their orientation as either homosexual or as bisexual.

There was a significant effect of waist-to-hip ratio (F(1,160) = 82.681, p < .001), body weight, (F(1,160) = 16.125, p < .001) and breast size, (F(1,160) = 15.71, p < .001), as well as a two-way interaction between breast size and weight (F(1,160) = 27.253, p < .001) and between waist-to-hip ratio and weight (F(1,160) = 78.242, p < .001). These effects were modified by a significant three-way interaction between breast size, waist-to-hip ratio, and weight on preference ratings (F(1,160) = 19.287, p < .001). Because of the large number of comparisons inherent in the design used, we only report results which were highly significant (p's < .01).

Table 2 contains the mean preference scores for each figure. We considered the most striking finding to be that the figure rated highest overall, the heavy figure with a waist-to-hip ratio of 0.7 and large breasts, was significantly preferred to all other figures (Fs ranging from 46 to 166, all ps < .001). The second highest rated figure, the heavy figure with a waist-to-hip ratio of 0.7 and small breasts, was also significantly preferred to every other figure (Fs ranging from 19 to 119, all ps < .001).

To further explore the different ratings of the figures, we calculated an effect size for each comparison of two figures by taking the difference in mean ratings for each figure and dividing this difference by the pooled standard deviation for those two figures. For the most preferred figure (the one with large breasts, a waist-to-hip ratio of 0.7, and heavy body weight), the effect sizes of the comparisons to all other figures ranged from 0.8 to 1.1, except for the second most preferred figure, for which the effect size was 0.4. For the second most preferred figure (the similar figure with small breasts), comparing that figure to all others (except the most preferred figure) yielded effect sizes from 0.5 to 0.7. These are quite substantial effects.

STUDY 2

In Study 2 we investigated whether extent of gender conformity or gender nonconformity related to figure ratings.

Methods

Questionnaire. The questionnaire was similar to that in Study 1 but we also solicited information about childhood and adult gender-typicality, based on items adapted from Bailey and Zucker (1995). There were seven childhood gender-typicality items (sample item: as a child I preferred playing with boys rather than girls) and eight adult gender-typicality items (sample item: I often wonder what it would be like to be a man). Subjects indicated their strong disagreement (coded as a-3), somewhat disagreement (-2), slight disagreement (-1), uncertainty (0), slight agreement (+1), somewhat agreement (+2), or strong agreement (+3) with each item. The internal reliability of the total scale was good (Cronbach's alpha = .86). These items were after the figure ratings so they would not influence responses on figure ratings. The exploratory nature of the study was stressed throughout the questionnaire because of the sensitive nature of the items. The order of the figures was changed in order to ensure that the results in Study 1 were not due to order effects. Figures were presented in the following order (breast size/waist-to-hip ratio/body weight)--1: large/0.7/slender; 2: small/1.0/heavy; 3: small/0.7/slender; 4: large/1.0/heavy; 5: large/1.0/slender; 6: small/0.7/heavy; 7: small/1.0/slender; 8: large/0.7/heavy.

As in Study 1, the questionnaire was web based. The questionnaire was accessible through the web page of the first author, and a solicitation to participate was e-mailed to lesbian organizations and clubs on the World Wide Web. Respondents were compensated through a lottery in which four randomly-selected participants were each paid $100 for participation, provided they had voluntarily provided some contact information in order to be included in the lottery. Contact information was removed from each questionnaire so that no individual's data could be traced back to her.

Participants. Participants (n = 141) ranged in age from 16 to 65, with a mean age of 28.4, a standard deviation of 9.90, and a median age of 24. Nine percent of the participants had a high school diploma, 49% had some college education, 25% had a college degree, and 17% had graduate degrees. Thirty percent of the participants identified as bisexual and the rest as homosexual.

Data reduction. As in Study 1, we factor analyzed the ratings of attractiveness, femininity, healthiness, kindness/understanding, and desirability for short and long-term relationships for each figure, and we obtained similar results. Also as in Study 1, all of the ratings were positively intercorrelated for each figure, with Cronbach's alphas ranging from .83 to .90. We thus summed ratings into general preference scores as in Study 1. The childhood and adult gender-nonconformity items were significantly intercorrelated (Cronbach's alpha = .86) and so they were summed into a gender-nonconformity scale, reverse scoring the appropriate items such that higher scores would indicate higher childhood and adult gender-nonconformity.

Results

As in Study 1, out first concern had to do with order effects, so we calculated the average within-subjects correlation between figure number and general preference score. The average correlation was, as in Study 1, small and nonsignificant (M = -.10, SD = 0.37). Furthermore, as the figures were presented in a different order in Study 2 than in Study 1, we combined the data from Studies 1 and 2 and included a between-subjects variable coding for the study. This variable did not have a main effect and did not modify the within-subjects three-way interaction reported below (F = 0.03). It was thus concluded that the order of the figures did not significantly affect the results.

Childhood and adult gender-nonconformity scale score was entered in as between-subjects factors in a repeated-measures MANOVA, similar to the one used in Study 1, with three within-subjects variables: breast size (small or large), body weight (slender or heavy), and waist-to-hip ratio (0.7 or 1.0). The gender-nonconformity scale had no between-subjects effect or any interaction with any other variable (ail Fs < 1), indicating that those participants who were high in gender nonconformity were similar to those participants who were low in gender nonconformity in their pattern of preferences for the eight figures.

As in Study 1, there was a significant main effect of breast size (F(I, 127) = 24.82, p < .001), a significant main effect of body weight (F(1,127) = 32.13, p < .001), a significant interaction between breast size and waist-to-hip ratio (F(1,127) = 36.91, p < .001), a significant interaction between breast size and weight (F(1,127) = 30.70, p < .001), and an interaction between waist-to-hip ratio and weight (F(1,127) = 46.09, p < .001). These effects were contained with a significant three-way interaction between breast size, waist-to-hip ratio, and body weight (F(1,120) = 8.87, p = .004).

Table 2 shows the mean ratings for each figure. As in Study 1, the most preferred figure was the one with a waist-to-hip ratio of 0.7, a heavy body weight, and large breasts. This figure was rated significantly higher than every other (Fs from 9 to 19, ps < .001), and was rated somewhat higher than the heavy figure with a waist-to-hip ratio of 0.7 and small breasts (F(4,120) = 2.30, p = .06). The next most preferred figures were the one with a heavy body weight, waist-to-hip ratio of 0.7, and small breasts, and the one with a waist-to-hip ratio of 1.0, heavy body weight, and large breasts. These two figures were not rated differently from each other (F < 1).

The second most preferred figure was the heavy figure with a waist-to-hip ratio of 0.7 and small breasts. This figure was rated significantly higher than four other figures: the figure with small breasts, a 1.0 waist-to-hip ratio, and slender body weight; the figure with small breasts, a 1.0 waist-to-hip ratio, and heavy body weight; the figure with large breasts, 0.7 waist-to-hip ratio, and slender body weight; and the figure with large breasts, 1.0 waist-to-hip ratio, and slender body weight (Fs ranging from 3.0 to 10.4, ps < .001).

As in Study 1, we calculated effect sizes to quantify the differences in ratings of the different figures. Comparing the most preferred figure (large breasts, low waist-to-hip ratio, and heavy body weight) to all other figures yielded effect sizes ranging from 0.4 (for the second most preferred figure, the one with small breasts, 0.7 waist-to-hip ratio, and heavy body weight) to 1.0 (for the figure with small breasts, waist-to-hip ratio of 1.0, and slender body weight). For the second most preferred figure, the effect sizes ranged from 0.0 (for the figure with large breasts, waist-to-hip ratio of 1.0, and heavy body weight) to 0.6 (for the figure with small breasts, waist-to-hip ratio of 1.0, and slender body weight).

In all, participants who were high and low in childhood and adult gender-nonconformity rated figures similarly, most preferring the same figure as in study 1, the heavy figure with a waist-to-hip ratio of 0.7, and large breasts.

DISCUSSION

Investigating the physical factors that influence attraction among nonheterosexual women is important because there is little research at the intersection of mating preferences and sexual orientation, and thus this research brings together studies on topics that are too often kept separate. The purpose of these two studies was to add to the small literature on attraction in nonheterosexual populations relying on a set of stimuli used extensively to investigate the effects of breast size, body weight, and waist-to-hip ratio among heterosexual men and women.

Two studies were conducted to investigate the roles of body weight, breast size, and waist-to-hip ratio in attractiveness for lesbian and bisexual women with similar results. In both studies, a heavy figure with a low waist-to-hip ratio was most preferred. Study 2 showed that gender nonconformity did not relate to preferences. The preference for heavy figures is quite striking given the robust preferences seen in heterosexual men and women for female figures with a slender body weight, as reviewed above. Given that the methods of this study were similar on relevant dimensions to previous studies using heterosexual men and women as participants, these findings are particularly interesting. Further, the fact that lesbian and bisexual women may have different body type preferences than heterosexual women may be considered surprising given that prior research suggests largely that lesbians' sexuality is most powerfully shaped by their gender, not their sexual orientation (Kingdon, 1979; Schreurs, 1993).

One notable departure from the methods of previous studies was the fact that we conducted our studies online in order to have access to a larger sample of nonheterosexual women and to maximize anonymity. However, we consider our results to be accurate because previous studies have demonstrated the validity of online figure ratings (Bailey et al., 2000).

We think there is one particularly serious limitation of the current design. It has been argued that research on attractiveness of different body types is in poor shape because of over-reliance on Singh's stimulus set of line drawings of different figures (Henss, 2000; Tassinary & Hansen, 1998). Naturally some advantages to using something other than line drawings includes the possibility of converging results and ecological validity, but the line drawings do provide good control over extraneous parameters that would vary using other kinds of stimuli (Singh, 1995). For example, if one used photographs of women, these would vary in skin tone, musculature, hair color and deposition, and so forth. Singh (1994a, Study 2) and Henss (2000) have shown that using retouched photographs that manipulate the waist size in photographs of the same woman produces similar results to past research; that is, lower waist-to-hip ratios were rated more attractive than figures with higher waist-to-hip ratios.

In all, this study adds to the small literature on what lesbian and bisexual women find attractive in body shape. While heterosexual males and females, as reviewed above, find a female figure with a low waist-to-hip ratio, large breasts, and slender body weight most attractive, feminine, and desirable for relationships, we found that our respondents most preferred a figure with a heavy body weight. These findings are particularly interesting because they are, to the best of our knowledge, the first demonstration of a preference for heavy figures with waist-to-hip ratio controlled.

We believe that it is likely that a preference on the part of lesbians for heavier women may be the result of several factors. First, lesbians are heavier on average than heterosexual women and are more comfortable with their body weights, although the literature is not entirely consistent in this conclusion. (Beren et al., 1996; French et al., 1996; Gettelman & Thompson, 1993; Heffernan, 1996; Herzog, et al., 1992; Kenyon, 1968; LaTorre & Wendenburg, 1983; Perkins, 1981; Siever, 1994; Striegel-Moore et al., 1990). It is a well-established finding that people mate assortatively based on a wide variety of traits (the "matching hypothesis"), and so the participants in the current studies may have been choosing figures that more closely resembled their own body types. A future study could correlate preference ratings for figures with the physical characteristics of the rater in order to explore this hypothesis.

A related explanation for the current findings is that the findings reflect a rejection by our participants of what may be seen as an inappropriate societal fixation on thinness. As mentioned above, lesbians have show less internalization of societal standards of attractiveness than heterosexual women (Heinberg et al., 1995). Therefore, the participants in these studies may have given lower scores to the slender figures to reflect a belief that those figures are too skinny to be healthy or attractive. In a future study, it would be informative to correlate participants' preferences with their attitudes about societal standards of attractiveness and with participants' current body satisfaction.

Table 1. Factor Analysis Results for Study 1

                                      Figure

Factor loadings (a)             1     2     3     4

Item
  Attractiveness               .84   .80   .83   .83
  Healthiness                  .85   .75   .82   .73
  Femininity                   .70   .71   .72   .59
  Kindness                     .62   .63   .63   .61
  Desirability for
    short-term relationships   .77   .77   .78   .78
  Desirability for
    long-term relationships    .86   .82   .79   .81
  Percent of variance
    explained by factor         60%   56%   58%   53%
  Cronbach's alpha             .87   .84   .86   .82

                                      Figure

Factor loadings (a)             5     6     7     8

Item
  Attractiveness               .79   .80   .76   .81
  Healthiness                  .80   .81   .81   .74
  Femininity                   .71   .58   .69   .63
  Kindness                     .47   .64   .64   .54
  Desirability for
    short-term relationships   .75   .73   .75   .79
  Desirability for
    long-term relationships    .77   .82   .81   .80
  Percent of variance
    explained by factor         52%   54%   56%   53%
  Cronbach's alpha             .81   .82   .84   .82

(a) Factor analysis was a PCA with minimum
Eigenvalue of 1. Casewise deletion was used.
Table 2. General Preference Scale Scores for each Figure in
Study 1 and Study 2

                     Study 1 (b)   Study 2 (c)

Figure type (a)      M      SD     M      SD

small/0.7/slender   70.0   26.2   68.1   26.9
small/0.7/heavy     81.8   21.7   74.7   23.5
small/1.0/slender   65.7   25.8   57.6   31.1
small/1.0/heavy     65.8   21.6   66.1   26.3
large/0.7/slender   67.0   24.4   65.7   27.9
large/0.7/heavy     89.5   21.7   85.3   25.7
large/1.0/slender   68.6   23.8   61.7   26.3
large/1.0/heavy     69.7   21.5   73.8   23.9

(a) Figure types = breast size/waist-to-hip ratio/body weight.
(b) N = 161. (c) N = 141.

REFERENCES

Arkoff, A., & Weaver, H. B. (1966). Body image and body dissatisfaction in Japanese-Americans. Journal of Social Psychology, 68, 323-330.

Bailey, J. M., Gaulin, S., Agyei, Y., & Gladue, B. A. (1994). Effects of gender and sexual orientation on evolutionarily relevant aspects of human mating psychology. Journal of Personality and Social Psychology, 66, 1081-1093.

Bailey, J. M., Kim, P. Y., Hills, A., & Linsenmeier, J. A. W. (1997). Butch, femme or straight acting? Partner preferences of gay men and lesbians. Journal of Personality and Social Psychology, 73, 960-973.

Bailey, J. M., & Zucker, K. J. (1995). Childhood sex-types behavior and sexual orientation: A conceptual analysis and quantitative review. Developmental Psychology, 31, 41-54.

Bailey, R. D., Foote, W. E., & Throckmorton, B. (2000). Human sexual behavior: A comparison of college and internet surveys. In M. H. Birnbaum (Ed.), Psychological experiments on the internet (pp. 141-168). San Diego: Academic Press.

Beck, S. B., Ward-Hull, C. I., & McLear, P. M. (1976). Variables related to women's somatic preferences of the male and female body. Journal of Personality and Social Psychology, 34, 1200-1210.

Beren, S. E., Hayden, H. A., Wilfley, D. E., & Grilo, C. M. (1996). The influence of sexual orientation on body dissatisfaction in adult men and women. International Journal of Eating Disorders, 20, 135-141.

Beren, S. E., Hayden, H. A., Wilfley, D. E., & Striegel-Moore, R. H. (1997). Body dissatisfaction among lesbian college students: The conflict of straddling mainstream and lesbian cultures. Psychology of Women Quarterly, 21, 431-445.

Bergeron, S. M., & Senn, C. Y. (1998). Body image and sociocultural norms: A comparison of heterosexual and lesbian women. Psychology of Women Quarterly, 22, 385-401.

Brown, L. S. (1987). Lesbians, weight and eating: New analyses and perspectives. In Boston Lesbian Psychologies Collective (Eds.), Lesbian psychologies: Explorations and challenges (pp. 294-309). Chicago: University of Illinois Press.

French, S. A., Story, M., Remafedi, G., Resnick, M. D., & Blum, R. W. (1996). Sexual orientation and prevalence of body dissatisfaction and eating disordered behaviors: A population-based study of adolescents. International Journal of Eating Disorders, 2, 119-126.

Furnham, A., & Alibhai, N. (1983). Cross-cultural differences in the perception of female body shapes. Psychological Medicine, 13, 829-837.

Furnham, A., & Baguma, P. (1994). Cross-cultural differences in the evaluation of male and female body shapes. International Journal of Eating Disorders, 15, 81-89.

Furnham, A., Dias, M., & McClelland, A. (1998). The role of body weight, waist-to-hip ratio, and breast size in judgments of female attractiveness. Sex Roles, 3/4, 311-326.

Furnham, A., Tan, T., & McManus, C. (1997). Waist-to-hip ratio preferences for body shape: A replication and extension. Personality & Individual Differences, 22, 539-549.

Gettelman, T. E., & Thompson, J. K. (1993). Actual differences and stereotypical perceptions in body image and eating disturbance: A comparison of male and female heterosexual and homosexual samples. Sex Roles, 29, 545-562.

Gitter, A., Lomranz, J., Saxe, L., & Bar-Tal, D. (1983). Perception of female physique characteristics by American and Israeli students. Journal of Social Psychology, 121, 7-13.

Heinberg, L. J., Thompson, J. K., & Stormer, S. (1995). Development and validation of the Sociocultural Attitudes Toward Appearance Questionnaire. International Journal of Eating Disorders, 17, 81-89.

Heffernan, K. (1996). Eating disorders and weight concern among lesbians. International Journal of Eating Disorders, 19, 127-138.

Henss, R. (1995). Waist-to-hip ratio and attractiveness: Replication and extension. Personality & Individual Differences, 19, 479-488.

Henss, R. (2000). Waist-to-hip ratio and female attractiveness: Evidence from photographic stimuli and methodological considerations. Personality and Individual Differences, 28, 501-513.

Herzog, D. B., Newman, K. L., Yeh, C. J., & Warshaw, M. (1992). Body image satisfaction in homosexual and heterosexual women. International Journal of Eating Disorders, 11, 391-396.

Horvath, T. (1981). Physical attractiveness: The influence of selected torso parameters. Archives of Sexual Behavior, 10, 21-24.

Jourard, S. M., & Secord, P. F. (1955). Body-cathexis and the ideal female figure. Journal of Abnormal and Social Psychology, 50, 243-244.

Kenrick, D. T., Keefe, R. C., Bryan, A., Barr, A., & Brown, S. (1995). Age preferences and mate choice among homosexuals and heterosexuals: A case for modular psychological mechanisms. Journal of Personality and Social Psychology, 69, 1166-1172.

Kenyon, F. E. (1968). Studies in female homosexuality. VI: The exclusively homosexual group. Acta Psychiatrica Scandinavia, 44, 224-237.

Kingdon, M. A. (1979). Lesbians. Counseling Psychologist, 8, 44-45.

Koff, E., & Benavage, A. (1998). Breast size perception and satisfaction, body image, and psychological functioning in Caucasian and Asian American college women. Sex Roles, 38, 655-673.

Krantz, J. H., Ballard, J., & Scher, J. (1997). Comparing the results of laboratory and World-Wide-Web samples on the determinants of female attractiveness. Behavior Research Methods, Instruments & Computers, 29, 264-269.

LaTorre, R. A., & Wendenburg, K. (1983). Psychological characteristics of bisexual, heterosexual and homosexual women. Journal of Homosexuality, 9, 87-97.

Lee, S., Leung, T., Lee, A. M., Yu, H., & Leung, C. M. (1996). Body dissatisfaction among Chinese undergraduates and its implications for eating disorders in Hong Kong. International Journal of Eating Disorders, 20, 77-84.

Loewenstein, S. F. (1985). On the diversity of love object orientations among women. Journal of Social Work & Human Sexuality, 3, 7-24.

Pederson, E. L., Markee, N. L., & Salusso, C. J. (1994). Gender differences in characteristics reported to be important features of physical attractiveness. Perceptual & Motor Skills, 79, 1539-1544.

Perkins, M. W. (1981). Female homosexuality and body build. Archives of Sexual Behavior, 10, 337-345.

Rosenzweig, J. M., & Leboew, W. C. (1992). Femme on the streets, butch in the sheets? Lesbian sex-roles, dyadic adjustment and sexual satisfaction. Journal of Homosexuality, 23, 1-20.

Rust, P. (1992). The politics of sexuality: Sexual attraction and behavior among lesbian and bisexual women. Social Problems, 39, 366-386.

Schreurs, K. M. G. (1993). Sexuality in lesbian couples: The importance of gender. Annual Review of Sex Research, 4, 49-66.

Siever, M. D. (1994). Sexual orientation and gender as factors in socioculturally acquired vulnerability to body dissatisfaction and eating disorders. Journal of Consulting and Clinical Psychology, 62, 252-260.

Singh, D. (1993a). Body shape and women's attractiveness: The critical role of waist-to-hip ratio. Human Nature, 4, 297-321.

Singh, D. (1993b). Adaptive significance of female physical attractiveness: Role of waist-to-hip ratio. Journal of Personality and Social Psychology, 65, 293-307.

Singh, D. (1994a). Ideal female body shape: Role of body weight and waist-to-hip ratio. International Journal of Eating Disorders, 16, 283-288.

Singh, D. (1994b). Is thin really beautiful and good? Relationship between waist-to-hip ratio (WHR) and female attractiveness. Personality & Individual Differences, 16, 123-132.

Singh, D. (1994c). Body fat distribution and perception of desirable female body shape by young Black men and women. International Journal of Eating Disorders, 16, 289-294.

Singh, D. (1995). Female judgment of male attractiveness and desirability for relationships: Role of waist-to-hip ratio and financial status. Journal of Personality and Social Psychology, 69, 1089-1101.

Singh, D. (1999). Lesbian erotic role identification: Behavioral, morphological, and hormonal correlates. Journal of Personality and Social Psychology, 76, 1035-1049.

Singh, D., & Luis, S. (1995). Ethnic and gender consensus for the effect of waist-to-hip ratio on judgment of women's attractiveness. Human Nature, 6, 51-65.

Singh, D., & Young, R. K. (1995). Body weight, waist-to-hip ratio, breasts, and hips: Role in judgments of female attractiveness and desirability for relationships. Ethology & Sociobiology, 16, 483-507.

Striegel-Moore, R. H., Tucker, N., & Hsu, J. (1990). Body image dissatisfaction and disordered eating in lesbian college students. International Journal of Eating Disorders, 9, 493-500.

Tassinary, L. G., & Hansen, K. A. (1998). A critical test of the waist-to-hip ratio hypothesis of female physical attractiveness. Psychological Science, 9, 150-155.

Thompson, J. K, & Tantleff, S. (1992). Female and male ratings of upper torso: Actual, ideal, and stereotypical conceptions. Journal of Social Behavior and Personality, 7, 345-354.

Tripp, C. A. (1975). The homosexual matrix. New York: McGraw-Hill.

Wiggins, J. S., Wiggins, N., & Conger, J. (1968). Correlates of heterosexual somatic preference. Journal of Personality and Social Psychology, 10, 82-90.

Manuscript accepted July 28, 2001

Adam B. Cohen and Ilana J. Tannenbaum University of Pennsylvania

Data from Study 1 were used as part of the senior undergraduate thesis of second author. This research was supported by an undergraduate research grant to the second author from the University of Pennsylvania Department of Psychology (Study 1), and a graduate student research grant from the University of Pennsylvania Department of Psychology to the first author (Study 2). Thanks to Dr. J. Michael Bailey for comments and for providing the gender nonconformity scales used in Study 2 and to Dr. John Sabini for statistical advice. We are also very grateful to two anonymous reviewers for excellent comments on this paper.

Address correspondence to Adam B. Cohen, Department of Psychology, Dickinson College, Carlisle, PA 17013; e-mail: [e-mail].