The development of preferences for specific body shapes.

In a now-classic study, Fallon and Rozin (1985) demonstrated a discrepancy between how women viewed their own bodies and their preferred body ideals. Subsequent authors demonstrated a link between body dissatisfaction and dieting and/or the development of eating disorders (Baluch, Furnham, & Huszca, 1997; Cash & Deagle, 1997; Probst, Vandereycken, Vanderlinden, & Van-Coppenolle, 1998; Schur, Sanders, & Steiner, 2000; Thelen, Powell, Lawrence, & Kuhnert, 1992). Given the importance of body perceptions to women's and men's mental health, there is renewed interested in determining what human body characteristics are considered attractive. Researchers have investigated various aspects of body shape, including chest and breast sizes (Thompson & Tantleff, 1992) and waist and hip sizes (Furnham, Hester, & Weir, 1990). To date, the majority of studies on body shape preferences have focused on adolescent or adult preferences. This study is one of the few with a developmental approach; the purpose is to chart changes in preferences for specific male and female body shapes, from early childhood to adulthood.

Evidence shows that biases toward certain body builds exist in children as young as 6 years. Even at this early age, children appear to prefer a mesomorphic (average sized and muscular) build over ectomorphic (thin and unmuscled) and endomorphic (overweight and unmuscled) builds, and children of all ages tend to associate certain social stigmas with the endomorphic body build (Jarvie, Lahey, Graziano, & Framer, 1983). Preferences toward thin ideals also appear to be present in children from as early as grade 6. Cohn et al. (1987) replicated Fallon and Rozin's (1985) study with children in grades 6, 7, and 8, finding that girls selected an ideal figure thinner than the one they thought boys would prefer, and boys thought that girls would prefer a thinner male figure than the one they chose as ideal. Although the girls' current and ideal figures did not differ in the younger grades, it was found that as the girls matured and developed, their ideal did not change, which created an increasing discrepancy between their ideal and current figures. This seems to indicate that the thin ideal, as seen in older girls and women, is present from an early age, and that the girls then find themselves moving physically further away from it as they mature and develop.

One clearly visible change in body form at puberty is the waist-to-hip ratio (WHR). WHR is often recorded as an index of human fat deposition, calculated by dividing the circumference of the waist (at the narrowest point around the torso below the iliac crest) by the circumference of the hips (at the point of greatest protrusion of the buttocks). Sex differences in WHR are triggered by hormonal changes experienced during puberty: As boys reach puberty, increases in testosterone levels stimulate fat to be deposited in the abdominal region, while for girls, increases in estrogen stimulate fat to be deposited in the gluteo-femoral region (thighs and buttocks; see Kirschner & Samojilik, 1991). WHR has been found to increase with increasing levels of testosterone (Hauner, Ditschuneit, Pal, Moncayo, & Pfeiffer, 1988). In the Australian population, the average WHR for adult males ages 18 to 49 is 0.87; the average WHR for females of the same age is 0.77 (Abernethy, Olds, Eden, Neill, & Baines, 1996).

Adult preferences for specific WHRs have been studied extensively (Furnham, Tan, & McManus, 1997; Henss, 1995; Singh, 1993a, 1993b, 1994a, 1994b, 1995; Singh & Luis, 1995; Singh & Young, 1995). Singh (1993a) presented White and Hispanic men in the U.S. with 12 cartoon images of identical women, differing only in their WHR and overall body weight. The WHRs ranged from 0.7 through to 1.0, with three weight categories: underweight, normal, and overweight. The men were asked to rank order the figures in terms of attractiveness and healthiness, along with other variables. The normal-weight figure with WHR 0.7 was judged to be the most attractive and healthy and to have the greatest desire and capability for having children. The underweight figure of WHR 0.7 was judged the most youthful, yet was also judged as having the lowest capability for having children. The preferences did not differ between the White and Hispanic men and were also found to be consistent across generations, with the only difference being that older men did not rate the underweight figures as highly as did the college-age men. Using photographs to assess preferences yielded the same results (Singh, 1994b).

Singh (1993a) reasoned that WHR serves an adaptive function of indicating potential mate value. Because high WHR in women is related to several health risks, including cardiovascular disease, cholesterol levels, and diabetes, and low WHR is related to reproductive status and fertility (see Singh, 1993a, 1993b), male preference for low WHR in women is hypothesized to be an adaptive feature of mate selection. According to Singh (1993a), WHR may act as a first pass filter for selection of healthy and viable mates, with men who prefer low WHR increasing their chances of mating with healthy, viable females and thereby maximizing their chances of siring healthy offspring.

Several replications and extensions of Singh's work on WHR have found cross-cultural consistency of ratings of body attractiveness, with Indonesian, Hispanic, British, Australian, and African-American populations also preferring the figures depicting the 0.7 WHR (Connolly, Mealey, & Slaughter, 2000; Furnham et al., 1997; Singh, 1994a; Singh & Luis, 1995). The cross-cultural consistency appears to be limited to Westernized societies, however, with non-Western-exposed indigenous cultures demonstrating a preference for the larger WHRs (Marlowe & Wetsman, 2001; Yu & Shepard, 1998) or exhibiting no differential preference based on WHR (Wetsman & Marlowe, 1999).

Cashdan (2000, June) argues that in cultures where women perform physically demanding roles, such as participating in farming or food gathering, the strength that is consequent to relatively high testosterone/estrogen ratios will be valuable--and perhaps even invaluable in terms of survival of her children. She suggests that in such cultures a higher WHR would therefore be more valued and more attractive than the WHR most preferred by Westerners. This was, indeed, noted in Yu and Shepard's (1998) study: Yomybato and Shipetiari women of child-bearing age have high WHRs, and Yomybato men preferred the high WHR figures; the Shipetiari men also rated the high WHR figures as healthier, although they still found the low WHR figures more attractive. Preferences therefore may be influenced by culturally differential roles of women. Marlowe and Wetsman (2001) similarly argue that larger waist-to-hip ratios should be preferred in cultures where food is hard to come by, and high WHR indicates consistent access to calories.

Singh (1995) also examined female preferences for male body shapes using 12 male figures differing on the same dimensions as the female figures: underweight, average, and overweight figures, and figures with WHRs of 0.7 through 1.0. The average-weight male figure with a WHR of 0.9 was judged to be the most attractive and healthy. Singh's (1993a) argument that WHR may be involved in the early stages of mate selection, acting as a first-pass filter to signal health and reproductive status, raises the questions of at what age such preferences emerge and through what channels they develop.

To date only one other investigation into children's preferences for WHR has been conducted. Singh and Henss (1997) investigated the female waist-to-hip ratio preferences of boys ages 8 to 16 from America, Germany, and India. These children and adolescents, like adults, selected the figure with the lowest waist-to-hip ratio as the most attractive and healthy. Unlike adults, the American and German boys tended to demonstrate a greater a preference for the underweight figures, while the Indian boys demonstrated a preference for the heavier figures. This cultural difference in body weight preference may have reflected the documented tendency for plumpness to be considered attractive in food-scarce areas (Anderson et al., 1992).

Although this unpublished study shows that children and adolescents above age 8 demonstrate the same WHR preferences for females as adults do, it does not provide information as to whether this preference is present in younger children or whether it varies across gender and age groups. Thus, we sought in the present investigation to address these questions by documenting the developmental pattern of changes (or lack of changes) in preferences for WHR across gender and age groups. We hypothesized three possibilities.

The first possibility is that adult-like preferences might be present early in development and apparent from an early age, similar to the early preference for and rapid learning about attractive human faces (Langlois et al., 1987; Slater et al., 1998). A second alternative is that body shape preferences might show punctuated emergence at a specific point in development; for instance, they could be related to children's own physical or hormonal development, perhaps emerging at puberty (Kirshner & Samojilik, 1991; McClintock & Herdt, 1996). Finally, there could be a gradual emergence of preferences for specific body shapes, which would be reflected in a linear association between preferences and age.

METHOD

Participants

A total of 564 students participated in the study. Twenty-four male and 29 female first-year psychology students participated to earn credit for their course. They represented the adult age group of the sample. A total of 285 male and 226 female primary and secondary school children were recruited from private schools in the Brisbane area, with parental consent. Table 1 (see Results) displays the mean age and standard deviation for each age group.

The first-year university students completed the study in two separate same-sex groups. School age children were tested individually at their schools. Given the nature of the developmental hypotheses being investigated, we included age groups from every year around puberty and every 2 years before puberty.

Materials

The stimulus figures were those used by Singh (1993a) to assess adult WHR preferences. They consisted of 12 male and 12 female line drawings depicting four waist-to-hip ratios (0.7, 0.8, 0.9, and 1.0) and three levels of body weight (underweight, average, and overweight). The waist-to-hip ratios had been varied by altering the size of the waist in the female figures, and the hips in the male figures.

Each set of 12 figures was presented on one sheet of typing paper. The figures were numbered 1 through 12 consecutively from the top left stimulus to the bottom right stimulus.

Procedure

We informed participants that we were investigating preferences for male and female body shapes. We counterbalanced presentation of the stimuli so that half of the male and half of the female participants in each age group received the male stimuli first. Participants older than age 12 were asked to select one figure from each set of 12 that they "think looks the most attractive." They were provided with a form for recording their preference along with their date of birth, grade, gender, ethnicity, and country of residence from age 5. The younger children were asked to choose one figure from each set that they thought "looks the best. Which one looks the nicest?" These children either pointed to the figure or named the corresponding number for the experimenter to record.

We deemed differential instructions for the older and younger participants necessary because this study sought to replicate Singh's work with adults and also to extend it to younger children. Following Singh, we used the phrase "looks most attractive" to question the older children and young adults. We felt that this word had sexual connotations which were an appropriate and desirable association for the older participants. For the younger participants, however, we were concerned that the word attractive might not be understood, so we switched to the simpler phrases "looks the best" and "looks the nicest." It is important to note here that the emphasis was placed on the word looks rather than nicest, encouraging a physical rather than attributional interpretation.

Measures of weight, height, and waist and hip circumference were also taken, so that the participants' own body mass index (BMI) and WHR could be calculated and used as indices of the participants' own physical development. The older participants recorded their own weight and waist and hip circumferences to the nearest kilogram and millimetre respectively. Waist circumference was measured around the narrowest point on the torso under the iliac crest. Hip circumference was measured around the point where the buttocks protruded the most. Younger children were manually guided in placing the tape measure in the appropriate places. For all participants, an experimenter measured height to the nearest millimetre using a tape measure attached to the wall. Weight was measured to the nearest kilogram using a portable bathroom scale.

From these measurements we calculated BMI by dividing individual participants' weight in kilograms by their height squared in metres. WHR was calculated by dividing participants' waist circumference in centimetres by their hip circumference in centimetres.

After testing, the university participants were given a written debriefing detailing what the experiment was about and directing them to some relevant reading. The school-aged children received oral debriefings, which included reassurance that all body shapes and sizes are okay and that there are no fight or wrong shapes to be.

RESULTS

We analyzed the data for WHR preferences and weight preferences separately. Results for WHR preferences are presented first.

WHR Preferences

We coded WHR preferences by column and row, with columns specifying which WHR was selected and rows specifying which weight category was selected, so that WHR and weight could be examined independently. The middle, average-weight category is referred to as normal-weight to retain consistency with Singh's (1993a, 1993b) terminology.

A preliminary two-way ANOVA of ethnicity and country of residence from age 5 on the preference measures, collapsed over age, did not reveal any significant effects. These variables were not, therefore, considered in any further analyses. We excluded any missing data from the relevant analysis and focused on addressing the three hypothesized developmental trajectories for the development of WHR preferences (e.g., early presence, punctuated emergence, or gradual emergence). This was achieved primarily through examination of correlation matrices, ANOVAs followed by pairwise tests, linear analysis, and regression analyses.

Descriptive Correlations

The descriptive statistics for each variable for males and females of each age group are depicted in Table 1.

First order correlation matrices for boys and girls revealed some significant (p < .05) patterns. Boys' preferences for male and female WHRs were related to their age, such that as they became older they preferred larger WHRs in the male figures and smaller WHRs in the female figures. Their WHR preferences were also related to their preferences for the female weight categories: As their preferences for larger male WHRs increased, their preferred female weight category became lighter. Their preferences for male and female weight categories were also positively correlated. Preferences for female WHRs were negatively correlated with participants' own body measurements of height, weight, and BMI, such that as these measurements increased, the preferred WHRs decreased.

Girls' preferences for WHRs were also significantly (p < .05) correlated with age, weight, and height, such that as girls grew older, taller, and heavier, they preferred male figures with larger WHRs and female figures with smaller WHRs. Their preferences for the male WHRs were also related to their own WHR, preferring larger ratios in the male figures as their own WHRs decreased. Girls' preferences for male and female weight categories were significantly positively correlated just as boys' preferences had been.

For both boys and girls, age was a significant correlate of all variables except for the male and female weight categories. Since we wanted to examine the role of participants' own physical development on their preferences, we next examined the correlation matrices with the effect of age partialled out. The partial correlation matrices for boys and girls are presented in Tables 2 and 3 respectively.

For boys, once the effect of age was removed all correlations between preferences and body measurements dropped out. For girls, removing the effect of age eliminated all but one of the relationships between male and female WHR preferences and the girls' own body measurements. Girls' own WHR was still correlated with their preferences for male WHRs, such that as their own WHR decreased they preferred larger male WHRs.

Are WHR Preferences Present in Very Young Children?

To address the question of whether the preferences for WHRs of 0.9 in men and 0.7 in women might be present from very early in development, we examined differences between preferences for male and female WHRs at age 6, as well as differences between the preferences at age 6 and at adult level. Chi-square analyses on the frequencies of WHR category selection were also conducted.

A two-way ANOVA on preferences for male and female WHRs by age and sex of participant revealed no effect of participant sex ([F.sub.(1,562)] = .009, p = .926 and F.sub.(1,562) =. 190, p = .663 respectively), so boy's and girl's preferences were combined for the analyses. Paired sample t tests revealed no significant difference between the preferred WHRs of male and female figures at age 6[([t.sub.(69)] = .56,p = .58) but did show a significant difference between these preferences in adults ([t.sub.(52)] = 11.79, p < .001). Independent t tests between age 6 and adult preferences revealed a significant change in preferences for both male WHRs ([t.sub.(121)] = -3.93, p < .001) and female WHRs ([t.sub.(121)] = 3.27, p < .001). In sum, 6-year-olds' preferences for the WHRs of male and female figures did not differ, but adults' preferences did; the adult age group preferred significantly larger male WHRs and significantly smaller female WHRs than did the 6-year-old group.

Chi-square tests on the frequencies with which these two age groups selected each WHR category revealed that the age 6 group did not select differently from chance for the male WHRs ([[chi square].sub.(3)] = 3.94, p = .2.68) or the female WHRs ([[chi square].sub.(3)] = 3.71, p = .294). The adult age group selected significantly differently from chance for both the male WHRs ([[chi square].sub.(3)] = 46.68, p < .01) and the female WHRs ([[chi square].sub.(3)] = 9.32, p < .01).

Do Preferences for Specific Body Shapes Emerge at a Specific Point in Development?

Since it appeared that WHR preferences were not present in adult form in the youngest age group, we next addressed the question of whether body shape preferences emerge at specific points in development, we calculated ANOVAs on preferences by age, followed by exploratory t tests. We also conducted chi-square analyses on the frequencies of WHR category selection at different ages.

First, two-way ANOVAs on preferences for male and female WHRs by age confirmed significant effects of age ([F.sub.(9, 554)] = 3.586, p < .001 and [F.sub.(9,554)] = 2.378, p < .05 respectively). As there were no main effects of participant sex for either male or female WHR preferences, these variables were subsequently examined with boys' and girls' preferences combined. Figure 1 depicts these preferences combined across sex of participant.

[FIGURE 1 OMITTED]

Using Fisher's procedure, once an omnibus F result had been found for age, we conducted independent t tests not exceeding the degrees of freedom in the denominator (Keppel & Zedeck, 1989) to determine the point at which preferences began to differ from those of the youngest age group, which had already been shown to be no different from chance.

Preferences for male WHRs for boys and girls combined. Preferences for the male WHRs first became significantly different from the age 6 preferences at age 10 ([t.sub.(132)] = -2.77, p < .001). Preferences at this age were not significantly different from those of the adult age group ([t.sub.(115)] = -1.19, p = .236), indicating that adult-like preferences first emerged at age 10. A significant decrease in the size of preferred WHR occurred from age 12 to 14 ([t.sub.(112)] = 1.99, p < .05) such that the age 14 preference was once again different from that of the adult age group ([t.sub.(112)] = -2.62, p < .01), but the difference disappeared again at age 15 ([t.sub.(98)] = -1.92, p > .05).

Chi-square analyses of the frequency at which the different figures were chosen revealed that for male WHR preferences, participants began to select significantly differently from chance at age 10 ([[chi square].sub.(3)] = 13.625, p < .05), with preferences directed away from low WHRs. This significant difference continued throughout the rest of the age groups. The percentage of participants selecting each male WHR category is depicted in Figure 2.

[FIGURE 2 OMITTED]

Preferences for female WHRs for boys and girls combined. Preferences for female WHRs first became significantly different from those of the 6-year-olds at age 15 ([t.sub.(115)] = 2.52, p < .05). Preferences at this age were not significantly different from those of the adult group ([t.sub.(98)] = .65, p = .519). Chi-square analyses on the frequency of WHR category selection revealed that preferences first became significantly different from chance at age 12 ([[chi square].sub.(3)] = 8.46, p < .05), such that preferences were directed toward the lower WHR columns significantly more than the high WHR columns. This pattern continued throughout the rest of the age groups. The percentages of each age group selecting each category of female WHR are depicted in Figure 3.

[FIGURE 3 OMITTED]

To assess when preferences for male and female WHRs began to differ from each other, we performed pairwise t tests between preferences for male WHRs and female WHRs at each age level, adjusting alpha to .005 due to the number of tests. The first significant difference occurred at age 10 ([t.sub.(63)] = 3.15, p < .005), and the difference remained significant throughout the rest of the age levels ([t.sub.(46-60)] = 3.55 - 11.79,p < .001).

Relationship between WHR preferences and children's own body shapes. Changes in participants' own WHRs were also examined, as these changes may have been related to when the preferences appeared to emerge. Participants' own WHRs are depicted in Figure 4.

[FIGURE 4 OMITTED]

We calculated a two-way ANOVA on participants' own WHRs, by age and sex of participant. The analysis revealed a significant main effect of age ([F.sub.(9.551)] = 11.837, p < .001) and participant sex ([F.sub.(1,559)] = 196.253, p < .001) with a significant interaction ([F.sub.(9,551)] = 6.150, p < .001).

Following the same procedure as for WHR preferences, we performed a select number of independent t tests. Boys' WHRs decreased significantly from ages 6 to 8 ([t.sub.(65)] = 2.58, p < .05), and ages 8 to 10 ([t.sub.(63)] = 3.21, p < .05). Although a rise in WHR was observed at age 13, this increase was not significant ([t.sub.(54)] = -1.66, p = .102). Boys' WHRs then proceeded to increase gradually, with a significant difference from the age 10 WHR first observed in the adult age group ([t.sub.(57)] = -2.49, p < .05). At adulthood, boys' WHRs were not significantly different than they were at age 6 ([t.sub.(59)] = 1.87, p = .066).

Girls' WHRs decreased significantly from ages 6 to 8 ([t.sub.(64)] = 2.09, p < .05), and 8 to 10 ([t.sub.(59)] = 2.26, p < .05). No decreases between consecutive age levels were significant after age 10, despite a gradual downward trend.

To determine when boys' and girls' own WHRs began to differ, we conducted pairwise t tests, with alpha adjusted to .005, between the WHRs at each age level. Significant differences were found between boys and girls at age 6 ([t.sub.(68)] = 4.26, p < .001), age 8 ([t.sub.(61)] = 3.50, p < .001), age 13 ([t.sub.(50)] = 5.17,p < .001), and age 16 to adult ([t.sub.(46-51)] = 5.42 - 8.46, p < .001). A linear trend analysis of the participants' own WHRs revealed significant trends for both boys ([F.sub.(1,306)] = 7.10, p < .01) and girls ([F.sub.(1,251)] = 155.72, p < .0001), indicating gradual changes over time. The boys' trend deviated significantly from linearity, however, ([F.sub.(8,299)] = 3.24, p < .01), most likely due to the sudden increase at age 13 (see Figure 4).

[FIGURE 4 OMITTED]

Do Preferences for Specific Body Shapes Emerge Gradually?

To address the question of whether preferences emerge gradually, we sought evidence of a smooth linear development by conducting linear regression analyses.

Despite preferences becoming significantly different from chance or becoming significantly similar to the adult preferences at age 10 for male WHRs and age 15 for female WHRs, visual inspection of the data suggested a gradual change over time. The change in male WHR preferences over time was significantly linear ([F.sub.(1,562)] = 19.26, p < .0001) as was the change in female WHR preferences ([F.sub.1,562)] = 17.53, p < .0001).

We used multiple regressions to separate the effects of age and body measurements on preferences. Forcing the covariates of participants' own WHR, height, weight, and BMI into a regression on preferences for male WHRs before age revealed that together these variables significantly contributed to prediction of male WHR preferences ([R.sup.2] = .028, [F.sub.(3,557)] = 5.34, p < .005). However, none of the body measurements alone were significant. After removal of these covariates, age still contributed significantly to prediction ([R.sup.2] = .0374; F [change.sub.(4,556)] = 5.43, p < .05). When we reversed this order and removed the body measurement covariates after removing the significant contribution of age ([R.sup.2] = .0327; [F.sub.(1,559)] = 18.910, p < .0001), we found that the body measurement covariates did not contribute anything further to prediction ([R.sup.2] = .0374; F [change.sub.(4,556)] = .891, p = .445).

We found the same results in the regression of age, own WHR, height, weight, and BMI on preferences for female WHRs. When we entered the body measurement covariates first, we found that together they contributed significantly to prediction ([R.sup.2] = .020; [F.sub.(3,557)] = 3.797, p < .05). However, no single covariate was significant in the equation. Entering age after these covariates had been removed revealed that age still significantly contributed to prediction ([R.sup.2] = .0335; F [change.sub.(4,556)] = 7.714, p < .01). When we entered age into the regression first, the covariates did not contribute anything to the prediction of WHR preferences ([R.sub.2] = .033; F [change.sub.(4.556)] = .611, p = .608) beyond the significant contribution of age ([R.sub.2] = .030; [F.sub.(1,559)] = .17.45, p < .0001).

Weight Preferences

In addition to examining WHR preferences, we also examined the preferences for the three different weight categories of the figures using ANOVA and Chi-square analyses.

Preferences for male weights. A two-way ANOVA revealed a significant effect of sex of participant ([F.sub.(1,562)] = 11.535, p < .001) but not of age ([F.sub.(9,554)] = 1.676, p = .092), with no interaction ([F.sub.(9,554)] = .470, p = .895). No pairwise tests were performed; however, in general, boys preferred heavier male figures than girls did, particularly in the younger age groups. The percentages of boys and girls selecting each weight category of the male figures for each age group are listed in Table 4.

We performed chi-square analyses on the weight category preferences to identify any shift from random choice. Participants first started selecting differently from chance at age 8 ([[chi square].sub.(2)] = 14.000, p < .001), with significant Chi-squares continuing through to the adult group such that the underweight and normal weight categories were selected more often than the overweight category. There was a gradual move away from the underweight figures toward the normal weight figures as age increased, with a switch from more preferences for the underweight figures to more preferences for the normal-weight figures at age 12.

Preferences for female weights. We also calculated a two-way ANOVA on preferences for female weights, which revealed significant effects of age ([F.sub.(9,554)] = 2.771, p < .005) and participant sex ([F.sub.(1,562)] = 6.417, p < .05) with no interaction ([F.sub.(9,554)] = 1.171, p = .311). No pairwise tests between boys' and girls' preferences were conducted; however, in general, as for the male figures, boys tended to prefer slightly heavier figures than did girls, particularly in the younger age groups. As preferences for weight categories appeared quite haphazard, we explored the significant effect of age using Tukey's HSD procedure. Tukey's test of pairwise differences revealed that both boys and girls preferred normal-weight female figures at the adult age level but low-weight figures at age 10. The percentages of boys and girls selecting each weight category of the female figures for each age group are listed in Table 5.

We preformed chi-square analyses on the weight category preferences with an alpha requirement of .005. Selection was significantly different from chance at each age level ([[chi square].sub.(2)] = 9.383 - 49.057, p < .005) except for age 8 ([[chi square].sub.(2)] = 9.524, p = .009). The underweight and normal-weight categories were selected more often than the overweight category, and the same trend toward normal-weight figures over time was observed with a switch away from more preferences for the underweight figures to more preferences for the normal-weight figures at age 14 (see Table 5).

Covariate Analyses

Given the possibility that any observed age effects may have been due to the high correlation between age and participants' own body measurements, we removed WHR, height and weight, and BMI as covariates in ANCOVAs for each preference variable that had an age main effect. The covariates were not significant in the female weight preference ANCOVA and so did not affect the age main effect ([F.sub.(9,551)] = 2.85, p < .01). For the ANCOVA on preferences for male WHRs, the covariates were significant together ([F.sub.(3,557)] = 5.36, p < .001) but not alone. Removing them did not alter the age main effect ([F.sub.(9.551)] = 2.03, p < .05). For the ANCOVA on preferences for female WHRs, the covariates were significant together ([F.sub.(3.557)] = 3.84, p < .01) but not alone. Removing them caused the age main effect to drop out ([F.sub.(9.551)] = 1.19, p = .297).

DISCUSSION

The preference for relatively small WHRs in women and relatively large WHRs in men was clearly replicated with Australian adults in this study, as demonstrated by first-year University students' overwhelming and significant preferences for the female figure with a WHR of 0.7, followed by 0.8 and then 0.9, with no one exhibiting a preference for the WHR of 1.0. Significant and strong preferences for a WHR of 0.9 in the male figures were also observed, followed by 1.0 and then 0.8, with no one exhibiting a preference for the 0.7 figures.

Just as the presence of these definite WHR preferences was clearly evident in the oldest age group, the lack of such definite preferences was clearly evident in the youngest age group, with the responses of the 6-year-old age group being no different from those expected by chance. These results suggest that it is unlikely that preferences for specific body shapes are inborn or learned in early infancy as are preferences for attractive human faces, although of course experiments with infants would be required to rule out the possibility that an early preference for specific body shapes declines in late infancy or early childhood.

We found some evidence for punctuated emergence of preferences for specific body shapes. A significant preference toward the larger WHRs in the male figures emerged at age 10 for both boys and girls and continued to grow stronger and more specifically directed toward male figures with WHRs of 0.9. The first significant difference between preferences for male and female WHRs also occurred at age 10. This was due to a growing preference for larger male WHRs, whereas the preferences for the female figures did not begin to lean toward the smaller WHRs until age 12. It is worth noting that age 10 is approximately the age of adrenarche, when the adrenal glands mature and release hormones related to sexual attraction prior to secondary sexual development at puberty. Thus it is possible that the emergence of preferences for male WHRs is related to hormone secretions driving sexual development, although further studies that measure hormone levels would be required to verify this hypothesis.

Significant preferences toward smaller WHRs in the female figures first occurred at age 15, although responses began to differ from those expected by chance at age 12. Preferences may begin to become more directed at age 12, when boys and girls begin to experience puberty, and may become adult-like at age 15, which is well into puberty for most children. Age 13 is also the age at which the boys' and girls' own WHRs began to diverge. Prior to this age, although there were significant differences between their WHRs at age 6 and 8, boys' and girls' own WHRs changed in the same direction, both decreasing. This may have been a result of the boys losing some residual "baby fat." It isn't until age 13 that boys' and girls' own WHRs begin to actually diverge. While the onset of adrenarche and sexual attraction appeared to coincide with preferences for high WHRs in men, the emergence of more directed preferences for small WHRs in women appeared to coincide with puberty.

These observations suggest that there may be links between sexual maturation and the emergence of preferences for specific body shapes. There are at least two possible developmental scenarios that would explain such a link: First, it could be that the hormonal changes that trigger the onset of secondary sexual development also bring about an increase in children's attention to and interest in the social norms for attractive body shapes. This model postulates a mix of physiological and social influences on the development of preferences for specific WHRs. Second, it could be that the hormones that initiate secondary sexual development affect children's preferences directly, for instance by stimulating maturation of evolutionarily prewired preferences for fit mates. In such a model, physiological influences alone would be responsible for the development of WHR preferences. With this type of physiology-only model, we would expect preferences to correlate with children's own physical developments, as both would be driven by hormonal levels. Preferences for specific body shapes would therefore be expected to be related to the children's own body measurements of WHR, and perhaps height and weight or BMI, more so than age.

Examination of the partial correlations and regression analyses suggest that age is the more important variable in determining preferences. After partialling out age, what few and small correlations there were between body measurements and preferences were no longer present. Similarly, regression analyses revealed that age predicted preferences over and above the influence of body measures, but not vice-versa. It is highly likely that the points at which preferences became significantly more adult-like and directed could simply have been an artefact of the clear linear trend in the data. The observed gradual linear trend and the lack of a convincing pattern of punctuated emergence of preferences suggests that gradual developmental influences, possibly socialization processes, may be the driving force behind the developmental change in WHR preferences observed. It may be that just as children appear to be socialized to prefer attractive peers and slim body builds (Adams, Hicken, & Salehi, 1988), they may also be socialized to prefer certain body shapes. Physiological events are not discounted on such a developmental model, as adrenarche and puberty may increase children's attention to these issues and to the values and ideals that society places on them. Although no direct measures of socialization were included in the study, the pattern of results suggests that the role of socialization the development of WHR preferences should be seriously explored.

In addition to examining preferences for the WHRs of the figures, we also examined preferences for the different weight categories, providing one of the first developmental pictures of weight preferences. Consistent with previous research on children's preferences for body shapes, the bias away from overweight figures appears to be present from an early age for both boys and girls (Jarvie et al., 1983). Initially, there was little discrimination between categories for preferences for male weights. Then, at age 8, preferences became significantly directed away from the overweight category, with the underweight category gaining more preferences than the normal-weight. By age 12, more preferences were directed toward the normal-weight rather than underweight figures. This preference for the normal-weight figures then continued to grow stronger with age. Similar trends were observed for the preferences for female weight, with a significant preference away from the overweight figures occurring as early as age 6. The crossover to preference for normal-weight figures over underweight figures occurred later for the female figures, not becoming obvious until age 15.

These results seem to contradict what most people have come to expect of people's preferences for body weight. Based on the consensus of literature to date, one would expect that the slimmest weight category would be the most favoured, particularly by the older age groups. Such a preference was not observed in this sample, however. The younger age groups preferred the underweight figures over the normal or overweight, with a switch in this trend at around puberty to preferring the normal-weight figures. A possible explanation for this finding is that, as mentioned in the introduction, in previous studies examining WHR preference, the underweight figures were consistently rated as appearing more youthful than any of the other figures (Singh, 1993a). They have been estimated by men and women to be approximately 17-19 years of age, with men rating the figures to be slightly younger than women rate them (Singh, 1993b). Although this is much older than the ages at which preferences for this weight category were strongest, faced with only three weight category choices, younger participants cannot select a figure more youthful looking than the underweight figures. This might explain the early preference for slimmer figures, with younger children attempting to select figures closest to their own age. The gradual tendency of preferences to move away from the underweight figures with age would then reflect a preference for similar-age figures. Participants may have been looking for figures that looked like their own and like their peers.

Another possibility is that there is a difference between what people rate as attractive and what they report as being their ideal. The participants in this study were not asked to rate the figures according to how they would like to look, but according to what they found attractive, or for younger participants, what they thought looked the best. Perhaps in reporting what they would like to look like, or what they believe an ideal shape looks like, people are reporting what they perceive to be the socially accepted body shape, when in actual fact, they find normal-weight figures more attractive. Younger children, being more susceptible to making responses that they believe are "correct," may have been selecting the weight category that they have been socialized to believe is more attractive, thinking they were giving the desired response.

Interestingly, the overweight male figures were not discriminated against as much as the overweight female figures appeared to be. Thinness as an ideal appears to be more important when judging women than when judging men. This trend may be reflected in women's greater dissatisfaction with their bodies, a phenomenon that seems to be increasing over time (Feingold & Mazzella, 1998).

This study demonstrated that the Western preferences for WHRs of 0.7 in females and 0.9 in males are not present in early childhood. The pattern of results indicates that preferences for specific WHRs first emerge in middle childhood and develop gradually to reach the adult norms.

Preferences for specific body weights were shown to develop earlier than those of WHRs, with preferences being directed away from overweight figures as early as 6 years for female figures and 8 years for male figures. Children's preferences for the normal-weight over the underweight figures increased with age. These results confirm that there is an early-developing bias against the endomorphic body build, and suggest that young children prefer youthful (ectomorphic) body builds.

Overall, the results of this study suggest that preferences for specific body shapes develop relatively gradually. Future work examining the roles of sex hormones and socialization practices on the development of body shape preferences may shed light on the specific mechanisms that give rise to these preferences.

APPENDIX

It is worth noting that various criticisms of Singh's figures and his argument that WHR is a key feature when judging attractiveness have been made. Tassinery and Hansen (1998) argued that Singh's figures confounded weight with hip size, and WHR with relative waist size. They created a new set of figures that attempted to unconfound these features and, upon having them ranked for attractiveness, concluded that the independent values of weight and hip size were the more important determinants of attractiveness, with WHR playing only a minor role. However, upon creating these new figures, Tassinery and Hansen (1998) introduced new confounds because "the linear dimensions of the waists and hips of these figures were slightly modified to represent more accurately the anatomical fact that a greater percentage of the hip circumference is due to variation in the saggital plane" (p. 50). This implies that the figures were altered to "account" for the fact that most of the circumference in the hips is due to the protrusion of the buttocks. No evidence was given, however, that the participants actually perceived the figures in the manner intended by these adjustments, and upon physically measuring the actual waist and hip sizes of the new stimulus figures, it is apparent that the adjustments resulted in visual waist-to-hip ratios smaller than those claimed to be depicted. None of these visual WHRs exceeded 0.8, which is a ratio typical of women in their 30s and a ratio found by Singh, and others, to be reasonably highly attractive.

A subset of the children of the current study was also presented with the Tassinery and Hansen (1998) figures after they had selected the figures from Singh's pages. Of the lightweight figures, the first column (small hips) was selected more often than the second or third (medium or large hips), with no clear preference toward any particular row. Of the medium and heavy weight figures, the first columns were selected more often than the second or third, and the second row (medium waist) was selected more often than the first or third (small or large waist). These preferences correspond to the cells with visual WHRs closest to 0.7. Although Tassinery and Hansen argue that this demonstrates a preference for small hips, as opposed to a certain WHR size, their own data as well as the data of the current study demonstrate a preference for small waists relative to hips. Small hips are preferred, hut with a waist size that creates a WHR of 0.7.

It is an anatomical fact that the protrusion of the buttocks accounts for a large proportion of hip size. The best way to take this into account when testing for preferences, however, would be to create three-dimensional figures differing in WHR. This would not only give participants a front view, but also side and back views, allowing WHRs to be depicted accurately.

Another common criticism of Singh's figures is that none depict a WHR below 0.7. The likely reason for this is that ratios below 0.7 are quite rare in women, and also look unnatural. When testing people's preferences for adaptive attractiveness features, it makes little sense to test for preferences that would not be naturally encountered. The goal of the present research was to assess what children find attractive out of the body shapes that they routinely see.

Table 1. Means (above) and Standard Deviations (below) for Each
Variable for Each Age Group

                                    Age group (in years)

                                    6                 8

                                          Boys

N                                 37                30
Age                                6.02              8.12
                                   (.35)             (.52)
Pref. male WHR                      .85               .86
                                   (.12)             (.11)
Pref. female WHR                    .81               .85
                                   (.11)             (.12)
Pref. male weight                  2.00              1.77
                                   (.88)             (.73)
Pref. female weight                1.78              1.80
                                   (.89)             (.81)
Own WHR                             .89               .87
                                   (.04)             (.04)
Height (kg)                      120.85            135.93
                                  (4.48             (4.95)
Weight (cm)                       23.15             31.42
                                  (4.35)            (4.83)
BMI                               15.75             16.93
                                  (2.04)            (1.77)

                                         Girls

N                                 33                33
Age                                5.91              8.13
                                    .41)             (.36)
Pref. male WHR                      .83               .86
                                   (.12)             (.11)
Pref. female WHR                    .85               .80
                                   (.11)             (.12)
Pref. male weight                  1.82              1.49
                                   (.92)             (.67)
Pref. female weight                1.61              1.58
                                   (.79)             (.71)
Own WHIR                            .85               .83
                                   (.04)             (.04)
Height (kg)                      117.68            130.66
                                  (4.62)            (5.42)
Weight (cm)                       21.71             27.14
                                  (4.65)            (4.59)
BMI                               15.56             15.85
                                  (2.14)            (2.23)

                                   Age group (in years)

                                     10                12

                                          Boys

N                                 35                27
Age                               10.24             12.30
                                   (.43)             (.34)
Pref. male WHR                      .89               .91
                                   (.10)             (.10)
Pref. female WHR                    .83               .83
                                   (.11)             (.11)
Pref. male weight                  1.80              1.85
                                   (.46)             (.60)
Pref. female weight                1.54              1.78
                                   (.51)             (.64)
Own WHR                             .84               .83
                                   (.04)             (.04)
Height (kg)                      144.35            157.34
                                  (5.21)            (8.85)
Weight (cm)                       36.09             48.57
                                  (5.67)           (10.55)
BMI                               17.28             19.51
                                  (2.30)            (3.16)

                                         Girls

N                                 29                25
Age                               10.14             12.30
                                   (.37)             (.51)
Pref. male WHR                      .90               .90
                                   (.09)             (.08)
Pref. female WHR                    .83               .78
                                   (.11)             (.10)
Pref. male weight                  1.45              1.56
                                   (.69)             (.65)
Pref. female weight                1.28              1.44
                                   (.46)             (.58)
Own WHIR                            .81               .80
                                   (.04)             (.06)
Height (kg)                      142.02            155.71
                                  (4.51)            (8.24)
Weight (cm)                       35.67             44.21
                                  (5.19)           (10.56)
BMI                               17.64             18.07
                                  (2.07)            (3.15)

                                   Age group (in years)

                                     13                14

                                          Boys

N                                 30                39
Age                               13.03             14.20
                                   (.27)             (.32)
Pref. male WHR                      .89               .87
                                   (.08)             (.09)
Pref. female WHR                    .78               .80
                                   (.10)             (.10)
Pref. male weight                  1.80              1.77
                                   (.71)             (.63)
Pref. female weight                1.47              1.46
                                   (.57)             (.51)
Own WHR                             .86               .83
                                   (.07)             (.06)
Height (kg)                      162.07            171.16
                                  (9.71)            (9.07)
Weight (cm)                       52.59             59.32
                                 (11.66)           (11.75)
BMI                               19.89             20.10
                                  (3.13)            (2.84)

                                         Girls

N                                 23                21
Age                               12.71             13.87
                                   (.38)             (.39)
Pref. male WHR                      .88               .87
                                   (.08)             (.08)
Pref. female WHR                    .81               .81
                                   (.12)             (.12)
Pref. male weight                  1.52              1.64
                                   (.51)             (.49)
Pref. female weight                1.44              1.59
                                   (.51)             (.50)
Own WHIR                            .77               .78
                                   (.04)             (.07)
Height (kg)                      160.32            162.17
                                  (7.95)            (5.70)
Weight (cm)                       47.46             53.52
                                  (6.32)            (6.79)
BMI                               18.46             20.42
                                  (2.09)            (2.93)

                                   Age group (in years)

                                     15                16

                                          Boys

N                                 26                30
Age                               15.48             16.19
                                   (.33)             (.46)
Pref. male WHR                      .87               .90
                                   (.10)             (.07)
Pref. female WHR                    .77               .79
                                   (.08)             (.10)
Pref. male weight                  1.69              1.90
                                   (.55)             (.61)
Pref. female weight                1.81              1.70
                                   (.40)             (.47)
Own WHR                             .84               .85
                                   (.08)             (.07)
Height (kg)                      176.73            179.83
                                  (8.73)            (7.35)
Weight (cm)                       61.52             71.27
                                 (10.53)           (11.76)
BMI                               19.64             22.03
                                  (2.76)            (3.23)

                                         Girls

N                                 23                23
Age                               14.90             15.69
                                   (.29)             (.29)
Pref. male WHR                      .90               .91
                                   (.07)             (.05)
Pref. female WHR                    .79               .80
                                   (.09)             (.09)
Pref. male weight                  1.76              1.83
                                   (.63)             (.58)
Pref. female weight                1.62              1.70
                                   (.50)             (.64)
Own WHIR                            .78               .76
                                   (.05)             (.05)
Height (kg)                      166.15            165.75
                                  (6.96)            (5.73)
Weight (cm)                       56.99             56.61
                                 (14.67)            (9.24)
BMI                               20.54             20.78
                                  (4.74)            (4.14)

                                   Age group (in years)

                                     17              18+

                                          Boys

N                                 31                24
Age                               17.26             20.24
                                   (.41)            (4.35)
Pref. male WHR                      .88               .91
                                   (.07)             (.45)
Pref. female WHR                    .79               .78
                                   (.10)             (.09)
Pref. male weight                  2.00              1.88
                                   (.63)             (.45)
Pref. female weight                1.58              1.96
                                   (.50)             (.36)
Own WHR                             .85               0.87
                                   (.07)             (.06)
Height (kg)                      180.30            179.65
                                 (11.21)            (7.23)
Weight (cm)                       75.40             72.48
                                 (11.80)            (9.21)
BMI                               23.33             22.43
                                  (4.07)            (2.30)

                                         Girls

N                                 17                29
Age                               16.95             19.42
                                   (.34)            (2.57)
Pref. male WHR                      .90               .90
                                   (.06)             (.04)
Pref. female WHR                    .78               .76
                                   (.09)             (.07)
Pref. male weight                  1.88              1.69
                                   (.46)             (.47)
Pref. female weight                1.77              1.69
                                   (.44)             (.47)
Own WHIR                            .73               .74
                                   (.05)             (.05)
Height (kg)                      167.91            168.05
                                  (7.17)             (.60)
Weight (cm)                       55.35             58.64
                                  (6.74)            (8.64)
BMI                               19.60             20.77
                                  (1.63)            (2.95)

Table 2. Partial Correlation Matrix Based on 309 Male Participants,
With Effect of Age Removed

                         Pref for male   Pref for female
                             WHR               WHR

Pref for male WHR           1.000
Pref for female WHR         -.042            1.000
Pref for male weight         .067             .020
Pref for female weight      -.110             .108
Own WHR                      .005            -.024
Height                      -.017             .014
Weight                      -.007            -.029
BMI                          .024            -.067

                         Pref for male   Pref for female
                            weight           weight

Pref for male WHR
Pref for female WHR
Pref for male weight       1.000
Pref for female weight      .150 *           1.000
Own WHR                     .018             -.030
Height                     -.009             -.110
Weight                      .011             -.096
BMI                         .005             -.079

                             Own
                             WHR             Height

Pref for male WHR
Pref for female WHR
Pref for male weight
Pref for female weight
Own WHR                    1.000
Height                     -.126 *          1.000
Weight                      .099             .689 *
BMI                         .182 *           .168 *

                            Weight             BMI

Pref for male WHR
Pref for female WHR
Pref for male weight
Pref for female weight
Own WHR
Height                      1.000
Weight                       .789 *            1.000
BMI

* p < .05.

Table 3. Partial Correlation Matrix for 255 Female Participants.
With Effect of Grade Removed

                          Pref for male    Pref for female
                               WHR               WHR

Pref for male WHR             1.000
Pref for female WHR           -.032            1.000
Pref for male weight           .103             .023
Pref for female weight        -.067             .048
Own WHR                       -.163 *           .001
Height                         .069             .041
Weight                        -.012             .050
BMI                           -.055             .024

                          Pref for male    Pref for female
                             weight            weight

Pref for male WHR
Pref for female WHR
Pref for male weight          1.000
Pref for female weight         .192 *          1.000
Own WHR                        .018             .002
Height                        -.030            -.060
Weight                        -.013            -.017
BMI                           -.034            -.017

                               Own
                               WHR             Height

Pref for male WHR
Pref for female WHR
Pref for male weight
Pref for female weight
Own WHR                       1.000
Height                        -.131 *          1.000
Weight                         .125 *           .615 *
BMI                            .242 *           .148 *

                             Weight              BMI

Pref for male WHR
Pref for female WHR
Pref for male weight
Pref for female weight
Own WHR
Height
Weight                        1.000
BMI                            .849 *          1.000

* p < .05.

Table 4. Percentages of Boys and Girls in Each Age Group
Selecting Each Weight Category for the Male
Figures

Age group       Underweight     Normal-weight     Overweight
(in years)    Boys    Girls    Boys    Girls    Boys    Girls

     6        37.8    51.5     24.3     15.2    37.8     33.3
     8        40.0    60.6     43.3     30.3    16.7      9.1
    10        40.0    65.5     40.0     24.1    20.0     10.3
    12        25.9    52.0     63.0    40.0     11.1      8.0
    13        36.7    47.8     46.7     52.2    16.7      0
    14        33.3    36.4     56.4     63.6    10.3      0
    15        34.6    33.3     64.5     57.1     3.8      9.5
    16        23.3    26.1     63.3     65.2    13.3      8.7
    17        19.4    17.6     61.3     76.5    19.4      5.9
    18+       16.7    31.0     79.2    69.0      4.2      0

Table 5. Percentages of Boys and Girls in Each Age Group
Selecting Each Weight Category of the Female Figures

Age group     Underweight    Normal-weight    Overweight
(in years)   Boys    Girls   Boys    Girls   Boys    Girls

    6        51.4    57.6    18.9    24.2    29.7    18.2
    8        43.3    54.5    33.3    33.3    23.3    12.1
   10        45.7    72.4    54.3    27.6     0       0
   12        33.3    60.0    55.6    36.0    11.1     4.0
   13        56.7    56.5    40.0    43.5     3.3     0
   14        53.8    40.9    46.2    59.1     0       0
   15        19.2    38.1    80.8    61.9     0       0
   16        52.2    39.1    70.0    52.2     0       8.7
   17        41.9    23.5    58.1    76.5     0       0
   18+        8.3    31.0    87.5    69.0     4.2     0

This research was conducted while the first and third authors were located at The University of Queensland.

The authors would like to thank the children, parents, and teachers of all schools that participated and the University students who participated and assisted in a portion of the data collection. Thanks are also extended to everyone who provided input on earlier drafts.

An abbreviated version of this paper was presented at the 2000 meeting of the Across Species Comparisons and Psychopathology Society (ASCAP) in Amherst, MA, by the first author, who was honored by the ASCAP Society with the 2000 Aaron T. Beck Award for Young Investigators.

REFERENCES

Abernethy, P, Olds, T., Eden, B., Neill, M., & Baines, L. (1996). Anthropometry, health and body composition. Sydney: University of New South Wales Press.

Adams, G. R., Hicken, M., & Salehi, M. (1988). Socialisation of the physical attractiveness stereotype: Parental expectations and verbal behaviors. International Journal of Psychology, 23, 137-149.

Anderson, J., Crawford, C., Nadeau, J., & Lindberg, T. (1992). Was the Dutchess of Windsor right? A cross-cultural review of the socioecology of ideals of female body shape. Ethology and Sociobiology, 13, 197-227.

Baluch, B., Furnham, A., & Huszca, A. (1997). Perception of body shapes by anorexics and mature and teenage females. Journal of Clinical Psychology, 53, 167-175.

Cash, T. F., & Deagle, E. A. (1997). The nature and extent of body-image disturbances in anorexia nervosa and bulimia nervosa: A meta-analysis. International Journal of Eating Disorders, 22, 107-125.

Cashdan, E. (2000, June). A woman's view of waist-hip ratios: Trade-offs in androgen and estrogen-dependent traits. Paper presented at the annual meeting of the Human Behavior and Evolution Society, Amherst, MA.

Cohn, L. D., Adler, N. E., Irwin, C. E., Millstein, S. G., Kegeles, S. M., & Stone, G. (1987). Body-figure preferences in male and female adolescents. Journal of Abnormal Psychology, 96, 276-279.

Connolly, J. M., Mealey, L., & Slaughter, V. (2000). The development of body shape preferences. Perspectives in Human Biology, 5, 19-29.

Fallon, A. E., & Rozin, P. (1985). Sex differences in perceptions of desirable body shape. Journal of Abnormal Psychology, 94, 102-105.

Feingold, A., & Mazzella, R. (1998). Gender differences in body image are increasing. Psychological Science, 9, 190-195.

Furnham, A., Hester, C., & Weir, C. (1990). Sex differences in the preferences for specific female body shapes. Sex Roles, 22, 743-754.

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

Hanner, H., Ditschuneit, H. H., Pal, S. B., Moncayo, R., & Pfeiffer, E. F. (1988). Fat distribution, endocrine and metabolic profile in obese women with and without hirsutism. Metabolism, 37, 281-286.

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

Jarvie, G. J., Lahey, B., Graziano, W., & Framer, E. (1983). Childhood obesity and social stigma: What we know and what we don't know. Developmental Review, 3, 237-273.

Keppel, G., & Zedeck, S. (1989). Data analysis for research designs. New York: W. H. Freeman and Co.

Kirschner, M. A., & Samojilik, E. (1991). Sex hormone metabolism in upper and lower body obesity. International Journal of Obesity, 15, 101-108.

Langlois, J., Roggman, L., Casey, R., Ritter, J., Rieser-Danner, L., & Jenkins. V. (1987). Infant preferences for attractive faces: rudiments of a stereotype? Developmental Psychology, 23, 363-369.

Marlowe, F., & Wetsman, A. (2001). Preferred waist-to-hip ratio and ecology. Personality and Individual Differences, 30, 481-489.

McClintock, M. K., & Herdt, G. (1996). Rethinking puberty: The development of sexual attraction. Current Directions in Psychological Science, 5, 178-183.

Probst, M., Vandereycken, W., Vanderlinden, J., & Van-Coppenolle, H. (1998). The significance of body size estimation in eating disorders: Its relationship with clinical and psychological variables. International Journal of Eating Disorders, 24, 167-174.

Schur, E. A., Sanders, M., & Steiner, H. (2000). Body dissatisfaction and dieting in young children. International Journal of Eating Disorders, 27, 74-82.

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

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

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

Singh, D. (1995). Female judgement 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., & Henss, R. (1997). Body weight, waist-to-hip ratio, and female physical attractiveness: A cross-cultural developmental study. Unpublished manuscript.

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 judgements of female attractiveness and desirability for relationships. Ethology and Sociobiology, 16, 483-507.

Slater, A., Von der Schulenburg, C., Brown, E., Badenoch, M., Butterworth, G., Parsons, S., et al. (1998). Newborn infants prefer attractive faces. Infant Behaviour and Development, 21,345-354.

Tassinery, 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.

Thelen, M. H., Powell, A. L., Lawrence, C., & Kuhnert, M. E. (1992). Eating and body image concerns among children. Journal of Clinical Child Psychology, 21, 41-46.

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.

Wetsman, A., & Marlowe, F. (1999). How universal are preferences for female waist-to-hip ratios? Evidence from the Hadza of Tanzania. Evolution and Human Behavior, 20, 219-228.

Yu, D. W., & Shepard, G. H. (1998). Is beauty in the eye of the beholder? Nature, 396, 321-322.

Linda Mealey died on November 5, 2002. She was an inspiring teacher and a valuable colleague who is deeply missed by her coauthors.

Address correspondence to Virginia Slaughter, School of Psychology, University of Queensland, Brisbane, Australia 4072; e-mail: [e-mail].

Manuscript accepted May 1, 2003