Dominant Voices and Attractive Faces: The Contribution of Visual and Auditory Information to Integrated Person Impressions

Abstract

Faces and voices, in isolation, prompt consistent social evaluations. However, most human interactions involve both seeing and talking with another person. Our main goal was to investigate how facial and vocal information are combined to reach an integrated person impression. In Study 1, we asked participants to rate faces and voices separately for perceived trustworthiness, attractiveness, and dominance. Most previous studies relied on stimuli in which extra-vocal information (e.g., verbal content, prosody) may have confounded voice-based effects; to prevent these unwanted influences, we used brief, neutral vowel sounds. Voices, like faces, led to the formation of highly reliable impressions. Voice trustworthiness correlated with voice attractiveness, mirroring the relation between face trustworthiness and attractiveness, but did not correlate with voice dominance. Inconsistent with the possibility that face and voice evaluations are indicative of real character traits, we found no positive correlations between judgments of trustworthiness or dominance based on faces and the same judgments based on voices (there was also no correlation between face attractiveness and voice attractiveness). In Study 2, we asked participants to evaluate male targets after seeing their faces and hearing their voices. Faces and voices contributed equally to judgments of trustworthiness and combined to produce a significant interaction effect. For attractiveness, faces were given more weight than voices, possibly due to the predominantly visual character of the attractiveness concept (there was no interaction effect). For dominance, the reverse pattern was true, with voices having a larger effect than faces on final judgments. In this case the auditory cues may be perceived to be more reliable because of the strong links between voice pitch, masculinity, and dominance.

Audiovisual integration in social evaluation.

“…Dominance of both faces and voices independently contributed to the impression formed when stimuli were presented multimodally. However, consistent with Rezlescu et al (2015) we found that voices had the larger effect on overall judgements.…”

“…Our approach therefore differs from those based on presentations of neutrally-worded reading passages or non-verbal vocalisations (e.g. vowels sounds), both of which are very commonly used techniques in this field (Berry, 1991;Rezlescu, et al, 2015).…”

“…All had normal or corrected-to-normal vision, reported no hearing impairments and received payment or course credit for their participation. Sample sizes were chosen following Study 2, in which effects were larger than those reported in Rezlescu, et al (2015). A post hoc power analysis was conducted using GPower (Erdfelder, Faul & Buchner, 1996).…”

“…Rezlescu, et al (2015) report that when participants were required to make dominance judgements to multimodal stimuli (face-voice), their judgements were more influenced by the voices than the faces (a pattern which was reversed for ratings of attractiveness). Our study therefore builds on this finding, but with the following differences.…”

“…While some argue that the use of brief, neutral vowel sounds mitigates the influence of aspects of voice such as prosody and semantic content (Rezlescu, 2015), the extent to which using non-verbal vocalisations replicates real everyday speech has been the topic of debate (Apple et al, 1979). Social evaluations are clearly multi-faceted in everyday life, and so there is value in studying them using contentful utterances.…”