Asymmetric contextual effects in age perception

How old we think someone is determines how we interact with them [1–3], and most people regularly make age judgements based on facial appearance. For example, when assessing the age of a suspect in crime scenes, or in daily tasks such as determining eligibility to buy alcohol or tobacco (for meta-analysis on age perception, see [3]). Given the importance of making accurate facial age estimations, and how frequently we make them, one might assume our perceptual system is precise in judging age. Indeed, early research claimed that people make reliable age estimates, with errors of ±3–4 years [4,5]. However, those studies were limited in the number and quality of their stimuli. Faces change as we grow older, and ageing of facial skin is driven by factors such as gender, genetics, lifestyle, diet, smoking, and consuming drugs and alcohol [6,7]. Therefore, given the large amount of variability in faces, using a small set of test faces potentially biases age estimates to the specific test identities used. More recent work has addressed this issue by using both a larger number of stimuli and a broader sample of participants. In this case, the authors find that estimating age is less reliable than originally proposed with a reported mean error magnitude between ±6 years [8,9] and ±8 years [10]. Taken together, this reveals that people make errors in age judgements even for single faces viewed individually