Does Race Matter? Testing Parameters for Collaboration Benefits on Face Identification Accuracy
Date
Authors
item.page.orcid
Journal Title
Journal ISSN
Volume Title
Publisher
item.page.doi
Abstract
The wisdom-of-crowds phenomenon refers to the greater accuracy for judgments made in pairs or groups, compared to individual judgments. These effects have been found also for face identification tasks. Collaborative decision making on face identification tasks improves identification accuracy over individuals acting alone. However, no prior study has considered how the race of the face and the participant impacts collaboration benefits. In the following experiments, I addressed four principal questions to explore the parameters of collaborative decision making. Specifically, the aim of this dissertation was to test if collaboration can be a viable strategy for improving both own- and other-race face identification. In Experiment 1, East Asian and Caucasian participants completed a face-identification task on their own and while working collaboratively as part of a same-race dyad (East Asian/East Asian or Caucasian/Caucasian pairs). Collaborative decisions were completed as part of a social dyad (virtually completing the task together) and as part of a non-social dyad (where individual score were fused independently, item-by-item). Here, I examined the baseline effects of collaboration for East Asian and Caucasian race groups. First, I found that collaboration improved both own- and other-race face identification accuracy. Second, I replicated previous findings that show comparable accuracy across social and non-social collaboration for recognition of own- and other-race faces. In Experiment 2, I explored two potential factors that could directly impact collaboration. First, I examined if the demographic composition of the pair of judges, different-race (two participants of different races) versus same-race (two participants of the same race) pairs, affected the benefits of collaboration. Here, I found that generally collaboration resulted in equal performance across dyad composition, but with a slightly lower performance of East Asian dyads on Caucasian faces. Second, I explored the effect of self-reported individual other-race experience collaborative face identification performance and found no clear relationship between accuracy and self-reported other-race contact as a function of dyads. Taken together, these experiments demonstrate that collaboration can be a potential tool for improving for both own- and other-race face identification accuracy.