Understanding facial impressions between and within identities

Abstract

A paradoxical finding from recent studies of face perception is that observers are error-prone and inconsistent when judging the identity of unfamiliar faces, but nevertheless reasonably consistent when judging traits. Our aim is to understand this difference. Using everyday ambient images of faces, we show that visual image statistics can predict observers' consensual impressions of trustworthiness, attractiveness and dominance, which represent key dimensions of evaluation in leading theoretical accounts of trait judgement. In Study 1, image statistics derived from ambient images of multiple face identities were able to account for 51% of the variance in consensual impressions of entirely novel ambient images. Shape properties were more effective predictors than surface properties, but a combination of both achieved the best results. In Study 2 and Study 3, statistics derived from multiple images of a particular face achieved the best generalisation to new images of that face, but there was nonetheless significant generalisation between images of the faces of different individuals. Hence, whereas idiosyncratic variability across different images of the same face is sufficient to cause substantial problems in judging the identities of unfamiliar faces, there are consistencies between faces which are sufficient to support (to some extent) consensual trait judgements. Furthermore, much of this consistency can be captured in simple operational models based on image statistics. Facial impressions 3 Facial impressions 4 surprisingly wide range of performance across different observers in the normal population (Burton, White & McNeill, 2010). By analysing the statistical properties of ambient images of faces, Burton, Kramer, Ritchie and Jenkins (2016) showed that image variability is to some extent idiosyncratic-that is, the ways in which one person's face varies across images can be different for someone else's face. Learning to recognise a familiar face thus involves learning how that face can vary, through seeing it in many settings-in effect becoming sufficiently expert with that face identity to be able to recognise new photos of the same person. But this form of perceptual expertise is identity-specific and may not generalise to another person's face, because that face varies in different ways. For this reason, unfamiliar face recognition is often poor because the range of variability of an unfamiliar face is unknown (Burton et al., 2016; Kramer, Young & Burton, 2018; Young & Burton, 2018). Computational work has shown the utility of this approach by simulating a range of well-known properties of familiar and unfamiliar face recognition (Kramer, Young, Day & Burton, 2017; Kramer et al., 2018). Although perception of the identities of unfamiliar faces can be problematic, many other characteristics are more easily seen. These include relatively objective properties determined by structural cues such as apparent gender, age and ethnicity (Bruce & Young, 2012) and more subjective impressions of social dispositions su...

Trait evaluations of faces and voices: Comparing within- and between-person variability.

“…Recent studies have already demonstrated that the variability in social trait evaluations attributed to different face photographs of the same person is indeed substantial, to the extent that it matches or sometimes even exceeds the variability in social ratings attributed to images of different people (Jenkins et al, 2011; Todorov & Porter, 2014). This pattern of results has been reported for many different social traits, including attractiveness, trustworthiness, dominance, competence, creativity, and others, in both natural and more controlled face stimuli sets (Mileva, Young, et al, 2019; Sutherland et al, 2017). Estimating the relative proportion of within- and between-person variability in social ratings has important implications for the existing literature, which often takes a single (usually highly standardized) image as a veridical representation of a person.…”

“…Different groups of participants rated 20 face or voice stimuli of 20 unfamiliar identities (10 female) for trustworthiness, dominance, and attractiveness on a 9-point Likert scale. For the trait ratings of faces, we describe a reanalysis of existing data of social evaluations of faces reported in Mileva, Young, et al (2019, Study 1). Details of the participants recruited and methods used are described again below.…”

Face and Voice Perception: Understanding Commonalities and Differences

“…One reason why the role of familiarity for interpreting highly changeable signals is small (relative to its prominent role for recognition) seems to be that whilst there are identity-specific differences that can to some extent limit the universality of characteristics that underlie social signals, these are relatively small compared to the identity-specificity of perceptual signals of personal identity [133]. This has the useful consequence of facilitating the many interactions with strangers that characterise much of modern life (or simply watching television).…”

Familiar faces as islands of expertise

“…In fact the lines of best fit have identical slopes of 0.20 increase in hit confidence for each step in familiarity. Univariate ANOVA confirms a significant linear trend for both the same image (F(1,129) However, precisely because different pictures look different (Mileva et al, 2019), the new one may not 'look like' Fred in which case there will be no advantage. Since it is unlikely that such competing effects would so exactly cancel out, a simpler explanation may be that, in this context, the same image engages 'face processing' to the same extent as a different image.…”