The influence of familiarity on sex decisions

Abstract

According to some accounts of face recognition (e.g., Bruce & Young, 1986), gender analysis occurs independently of identity analysis and, as a consequence, no influence of familiarity should be found on the time taken to perform sex decisions. Results of recent behavioural studies cast doubt upon this claim. Two experiments are reported that explore the influence of familiarity on sex decisions to faces (Experiment 1) and surnames (Experiment 2) of different levels of familiarity. In Experiment 1, participants were able to assign sex faster to highly familiar faces than they were to unfamiliar faces. Therefore, familiarity can influence the speed at which sex is analysed from faces. Similarly, in Experiment 2, participants were able to assign sex and familiarity faster to highly familiar surnames than they were to moderately familiar surnames. These findings are discussed in relation to the influence of sex information from identity-specific semantics and an explanation is offered based on the Burton, Bruce, and Johnston (1990) IAC model of face recognition.

Keywords:

• Familiarity
• Gender
• Face recognition

Notes

¹ In order to ensure that the manipulation of familiarity was a valid one, 20 new participants from The University of Birmingham performed speeded familiarity decisions to the same face stimuli used in Experiment 1. Results of t-tests by subjects and items indicated that participants were significantly faster when performing familiarity decisions to faces of high familiarity (mean RT = 862 ms) compared to faces of moderate familiarity (mean RT = 971 ms), t^S(19) = 5.947, p < .001; t^I (46) = 4.817, p < .001.

² In addition, sex decisions were performed faster than familiarity decisions to the same faces. This is the opposite finding to that reported in Experiment 2. To confirm, an analysis was carried out to directly compare the data from the sex decision task and the familiarity decision task employed in Experiment 1. Following log transformation, the data was subjected to a 2 (task: sex or familiarity; between subjects) × 2 (face familiarity level: high or moderate; repeated measure) ANOVA. A main effect of task, F(1, 48) = 175.638, p < .001, familiarity level, F(1, 48) = 74.391, p < .001, and a significant Task × Familiarity level interaction, F(1, 48) = 22.239, p < .001, were reported. Simple main effects analysis reported an effect of task at both high, F(1, 96) = 138.799, p < .001, and moderate, F(1, 96) = 195.677, p < .001, levels of familiarity. Additional effects of familiarity were reported for both the sex, F(1, 48) = 9.551, p < .01, and familiarity, F(1, 48) = 74.158, p < .001, decision task. Responses to highly familiar faces were performed faster than responses to moderately familiar faces in both face tasks. Importantly, sex decisions were always significantly faster than familiarity decisions for all faces.