Breaking face processing tasks apart to improve their predictive value in the real world: A comment on Ramon, Bobak, and White (2019).

In this commentary, I will expand on three aspects suggested by Ramon et al. (2019, British Journal of Psychology, 110, 461) to improve the predictive value of laboratory-based tasks in real-world applications. There are potential benefits that may arise from three interrelated considerations, particularly in terms of predicting agents' susceptibility to errors in operational settings. The first is a proposal to conduct a detailed examination of performance on face processing tests rather than only analysing overall accuracy scores. The second involves considering non-face-related cognitive and meta-cognitive sub-processes involved in face processing tasks. The third highlights the contribution of superior recognisers in creating challenging tests that simulate difficult real-world situations.