Social status, facial structure, and assertiveness in brown capuchin monkeys

A recent paper by Lefevre et al. (2014) in PLoS One reported that individual differences in facial structure predicted assertiveness in brown capuchin monkeys (Sapajus spp). Specifically, variation in the facial width-to-height ratio (fWHR) was positively correlated with alpha status and a composite measure of assertiveness. This novel finding adds to a growing body of evidence indicating that variation in facial structure reliably maps onto individual differences in dominance-related phenotypes. Research into fWHR was propelled by an anthropological study of human skulls indicating that fWHR was a sizeindependent sexually dimorphic feature of the human skull that arose around puberty coincident with the rise in pubertal testosterone (Weston et al., 2007). To the extent that variation in the fWHR is under the influence of pubertal testosterone, and that pubertal testosterone organizes the neural circuitry underlying sexually dimorphic behaviors, we tested the hypothesis that this metric would map onto human aggression. Initial support for this hypothesis came from a series of studies in which fWHR was positively correlated with aggressive behavior in men tested in a laboratory task as well as in varsity and professional hockey players (Carré and McCormick, 2008). Since this publication, several studies have found that this metric maps onto other conceptually similar phenotypes. For instance, fWHR is associated with unethical behavior (Haselhuhn and Wong, 2012; Geniole et al., 2014), nonreciprocation of trust (Stirrat and Perrett, 2010), psychopathic traits (Geniole et al., 2014), fighting abilities (Stirrat et al., 2012; Zilioli et al., accepted), explicit prejudice (Hehman et al., 2013) and selfishness (Haselhuhn et al., 2013). Despite these findings, there have been some non-replications. Ozener (2012) first reported that fWHR was not sexually dimorphic and did not predict aggression as assessed using a self-report measure. The lack of a sex difference in fWHR has now been reported in several relatively large-scale studies (e.g., Lefevre et al., 2012). In addition, a study with a larger sample of professional hockey players reported that fWHR was only marginally positively correlated (p = 0.057) with aggression (Deaner et al., 2012). Finally, in a Mexican sample, fWHR did not differ between males convicted of violent vs. non-violent crimes (GómezValdés et al., 2013). What may account for such discrepant findings? Were the original findings Type I errors? I believe this is an unlikely explanation given that several independent laboratories have found associations between fWHR and traits that are conceptually linked to dominance and aggression (see above). Another possibility is that the link between fWHR and dominance behavior is moderated by social context. Consistent with this idea, we recently reported that the relationship between fWHR and aggressive behavior in men was moderated by subjective and objective measures of social status (Goetz et al., 2013). Here, fWHR was positively correlated with aggression, but only among relatively low status men (Goetz et al., 2013). In their paper, Lefevre et al. (2014) reported positive correlations between fWHR, alpha status, and “assertiveness.” The latter construct consisted of traits such as bullying, aggression, dominance, jealousy, and stinginess. This is a novel finding, documenting for the first time a link between fWHR and complex social behavior in a non-human primate. Although the authors reported that alpha status did not significantly moderate the relationship between fWHR and assertiveness (p = 0.09), a careful examination of Figure 4 from Lefevre et al. (2014) certainly suggested that the effects were driven by nonalpha (i.e., low-ranking) monkeys. Indeed, bivariate correlations performed separately for alpha and non-alpha monkeys indicated that the relationship between fWHR and assertiveness was significant in non-alpha [r(23) = 0.54, p = 0.005], but not alpha monkeys [r(16) = 0.02, p = 0.94]. I decided to perform a re-analysis of Lefevre and colleagues’ data which were freely available on the PLoS One website (http://www.plosone.org/article/info %3Adoi%2F10.1371%2Fjournal.pone.009 3369) to investigate the extent to which the link between fWHR and assertiveness was driven by low status monkeys. In this model, assertiveness was the dependent variable and I included fWHR and alpha status on Step 1 and the fWHR-×-alpha status interaction on Step 2. As per Lefevre et al. (2014), I also included sex and age as