The Performance of Untrained Humans Verifying Children and Adults in a Face Matching Task

Abstract

Identifying children is a high priority in numerous government agencies not only in border checks, but to aid against child exploitation. However, research suggests children are hard to identify due to childhood facial development. Face matching is a common form of identification in areas such as border checks and investigative applications. A recent study demonstrated that trained facial practitioners found it more difficult to verify child identities compared to adult identities in a one-to-one unfamiliar face matching task, but there is a significant gap in research on whether it is naturally challenging to verify child identities. Thus, the present study primarily aimed to determine whether people with no training or experience find it naturally harder to match child compared to adult faces. The study secondly aimed to find if these people performed better determining whether two faces belonged to the same person, or different people. Students (N = 35) were asked to perform 200 one-to-one face matching trials, determining whether pairs of faces belonged to the same or different people and rate their confidence. The results demonstrated novices were significantly less accurate, confident and slower comparing child images compared to adult images, although performance when image pairs were the same versus different people had mixed results. The findings indicate people are naturally worse at verifying children, providing an argument for child specific training for human facial practitioners. Future research should compare people with and without training and experience, for an extensive analysis on whether this effects human face matching performance.