Waist–hip ratio and attractiveness

Abstract

An evolutionary model of mate choice predicts that humans should prefer honest signals of health, youth, and fertility in potential mates. Singh and others have amassed substantial evidence that the waist–hip ratio (WHR) in women is an accurate indicator of these attributes, and proposed that men respond to WHR as an attractiveness cue. In response to a recent study by Tassinary and Hansen [Psychol. Sci. 9 (1998) 150.] that purports to disconfirm Singh's hypothesis, we present evidence showing a clear relationship between WHR and evaluations of attractiveness. We evaluated responses to a range of waist, hip, and chest sizes, spanning the 1st through 99th percentiles of anthropometric data. Waist, hip, and chest sizes were altered independently to give WHRs of 0.5, 0.6, 0.7, 0.9, and 1.2. We replaced line drawings with more realistic computer-manipulated photographs. The preferred WHR was 0.7, concordant with the majority of previous results. By asking participants to estimate weight in each stimulus figure, we were able to statistically control for the effects of weight on attractiveness judgments; the effect of WHR remained.

Introduction

An evolutionary model predicts that humans should prefer honest signals of health, youth, and fertility in potential mates (Buss, 1989). Waist–hip ratio (WHR) in women is an indicator of these attributes (e.g., Singh, 1993a, Wass et al., 1997, Zaadstra et al., 1993). Before puberty, boys and girls both have a WHR of about 1.0. At puberty, female WHR declines to 0.7 and rises again at menopause (Ley, Lees, & Stevenson, 1992). Increased WHR is related to decreased fertility, and can be the result of conditions such as pregnancy, menopause, and polycystic ovary syndrome (PCOS) Pirwany et al., 2001, Wass et al., 1997, Zaadstra et al., 1993. Other factors less directly related to fertility, such as diabetes, Kwashiorkor, and cretinism, can also result in increased WHR with decreased fertility as a correlate.

Many studies have found that both men and women find a female WHR of 0.7 most attractive Furnham et al., 2001, Furnham et al., 1997, Henss, 1995, Henss, 2000, Singh, 1993a, Singh, 1993b, Singh, 1994a, Singh, 1994b, Singh, 1994c, Singh, 1994d, Singh, 1995, Singh & Luis, 1994, Singh & Young, 1995, but these findings have not gone unchallenged. Some have argued that WHR is insignificant in relation to weight, as measured by body mass index (BMI) Tovée & Cornelissen, 1999, Tovée & Cornelissen, 2001, Tovée et al., 1999, Tovée et al., 1997, Tovée et al., 1998, Yu & Shepard, 1998. BMI also predicts health and fertility (Tovée et al., 1999, and references therein). We will address this issue in the Discussion section.

Wetsman and Marlowe (1999) find that the hunter–gatherer Hadza prefer higher WHR than 0.7, and suggest that the Hadza base their attractiveness ratings on weight because starvation risk in this society makes fat stores a better predictor of fecundity than WHR. Singh and Luis (1994), however, found wide cross-cultural agreement on a preference for WHR of 0.7. Although they did include nonwestern populations, none of their data come from hunter–gatherer societies.

Manning, Trivers, Singh, and Thornhill (1999) propose a reason independent of weight for a preference for higher WHR. They found that women with higher WHRs had given birth to more sons than daughters, and propose that in cultures where boy children are preferred, the optimal WHR would be driven slightly higher than 0.7. However, a prospective study of WHR and sex ratio of offspring did not confirm the results of Manning et al. (Tovée, Brown, & Jacobs, 2001).

Tassinary and Hansen (T&H) (1998) challenge the validity of the stimuli used to test the hypothesis in most WHR studies, on the grounds that Singh's stimuli confound WHR with weight. They developed their own set of drawings, supposedly free from such confounds, but as we will discuss later, they did not succeed in unconfounding WHR and weight. T&H conclude that attractiveness actually increases with increases in WHR, contrary to most prior studies, but this conclusion depends entirely on the way they chose to plot their data: when plotting attractiveness as a function of WHR for a set of stimuli that vary both by waist and hip size, one can connect either the data points with the same hip size or those with the same waist size, and by choosing the latter, T&H also chose their conclusion. Fig. 1 shows T&H's data plotted both ways. The dotted lines connect the data points by waist size, as T&H plotted them, and imply that attractiveness increases with higher WHR, but the solid lines show the data plotted by hip size, and indicate that attractiveness decreases with higher WHR, as all others have found. We will discuss later which way of plotting the data is more appropriate. Data we present in this paper, however, show the conventional relationship regardless of which way they are plotted.