Abstract
An article recently published in this journal (Cox, Devine, Bischmann, & Hyde, 2016) questioned the validity of existing research on the accurate judgment of sexual orientation from photographs of faces. Specifically, those authors reported a confound in their stimuli whereby the photos of sexual minorities (gay men and lesbians) were of higher quality than the photos of heterosexuals. Based on this finding, they concluded that the accuracy in judging sexual orientation from faces demonstrated in the broader literature is an artifact of stimulus quality differences. Here, we addressed this claim by systematically testing the numerous photo sets that we have used in 61 studies documenting accurate judgments of sexual orientation from facial cues published since 2007. Contrary to their claim, the overwhelming majority of studies (93%) showed no significant differences in photo quality according to sexual orientation. Of those that did show differences, most revealed that heterosexual targets’ photos were actually of higher quality than sexual minority targets’ photos – opposite of what Cox et al. found. In addition, we highlight additional research using stimuli equated for quality differences overlooked in the recent article by Cox et al., lending further support to the conclusion that sexual orientation is legible from multiple sensory cues.
Funding
This research was supported in part by a grant from the Social Sciences and Humanities Research Council of Canada to NOR.
Supplemental Material
Supplemental data for this article can be accessed at 10.1080/00224499.2016.1205547
Notes
1. Alternatively, using a multilevel model that accounted for the 13 different photo sets and modeled the time at which the participants made their ratings as a covariate (because we did not collect the data for all photo sets in a single study) showed that photo quality did not differ between the sexual minority and heterosexual targets, B = .022, SE = .029, 95% CI [−.034, .078], t (1271.21) = 0.77, p = .44 (degrees of freedom calculated using Satterthwaite approximation).