Abstract
Participants make multiple social judgments reliably and accurately from briefly-presented face images. In particular, trustworthiness appears as a core dimension in social "face space" and may be an important personality trait that observers estimate from faces. Like other aspects of face perception, perceived trustworthiness is subject to out-group effects for categories like race, but it's unknown whether trustworthiness is impacted by other face categories. We examined whether observers exhibited an other-age bias for facial trustworthiness estimates. Our participants viewed briefly-presented (200ms) young (18-29yrs), middle-aged (30-49yrs), and older adult (50-69yrs) faces and rated trustworthiness on a 6-point Likert scale. We asked two questions about these ratings as a function of age group: (1) Do trustworthiness ratings differ across age groups? (2) Does inter-rater agreement differ across age groups? To address this latter question, we measured within-participant reliability and inter-subject reliability using intra-class coefficients (ICCs) to compare group agreement to individual reliability in all age groups. We recruited 47 participants (24 male) from NDSU's undergraduate study pool. Mean trustworthiness ratings across groups reflected a significant effect of age group (,F(2,45)=15.4, p< 0.001), such that trustworthiness ratings for younger faces (M=4.38) were significantly higher than ratings for older faces (M=3.99) and middle-aged faces (M=4.05). We also found that the relationship between group reliability and internal reliability differed across age groups. ICCs for group reliability were significantly lower than those for internal reliability in younger and middle aged-faces, but this was not true for older faces. That is, young observers rating older faces disagreed with each other less than they agreed with themselves, suggesting differing levels of group consensus as a function of face age. We conclude that face age impacts trustworthiness estimates and discuss this result in the context of other out-group face recognition effects.
Meeting abstract presented at VSS 2016