Faces carry intertwined cues to multiple social dimensions, which are often examined in isolation. Here we investigated the categorisation of sex and age from human faces to determine whether these dimensions are analysed in parallel or rely on shared processes. Three experiments with large sample sizes tested age or sex classification tasks: a Garner speeded classification task ( Garner, 1974 ), a word-priming task, and a face-priming task. Garner interference was found between sex and age but was stronger in the age task than the sex task. Both word and face priming effects were found, but these varied as a function of stimulus set. These effects cannot be explained by differences in task difficulty nor by carryover of response mappings. We also replicated findings that young females and old males were categorised more efficiently than young males and old females. Using three different sets of faces provided evidence about whether interference and priming effects originated at the stimulus level or a more abstract level. We propose that the asymmetric relationship between the processes that categorise faces by age and by sex reflects a hierarchical organisation of social dimensions in the face, such that age coding is nested within sex coding. One underlying driver of this scheme may be that age judgements typically rely more on featural than configural processing, whereas sex judgements show the opposite tendency. Our findings illustrate how considering joint coding of social dimensions sheds new light on face perception mechanisms. • Asymmetrical integrality between sex and age found with multiple experiments and stimulus sets. • Sex information affects age processing more than vice versa when viewing photographs of faces. • Processing of old male and young female faces is facilitated compared to old females and young males. • Results accounted for in terms of differential reliance upon configural and featural processing.
Aitken et al. (2026) studied this question.