Collaboration to improve cross-race face identification: Wisdom of the multi-racial crowd?

Area of Science:

• Cognitive Psychology
• Social Psychology
• Forensic Science

Background:

• The other-race effect (ORE) describes reduced accuracy in identifying faces of a different race compared to one's own race.
• Existing research highlights the challenges in cross-race face recognition, impacting legal and eyewitness contexts.

Purpose of the Study:

• To investigate whether collaborative decision-making improves face identification accuracy, particularly for other-race faces.
• To compare the effectiveness of social versus computational collaboration in face identification tasks.

Main Methods:

• Two online experiments were conducted with East Asian and White participants.
• Participants identified own- and other-race faces both individually and as part of a collaborative dyad.
• Computational aggregation of individual decisions was compared to dyadic collaboration.

Main Results:

• Collaborative decision-making significantly improved face identification accuracy for both own- and other-race faces compared to individual performance.
• Computational collaboration yielded benefits comparable to or exceeding social collaboration.
• The racial diversity of collaborators did not affect face identification accuracy.

Conclusions:

• Collaborative decision-making, especially non-social aggregation, is a robust strategy for enhancing face identification accuracy.
• This approach shows potential for improving cross-race identification in forensic and eyewitness applications.
• The benefits of collaboration may arise from aggregating information while avoiding potential group conflict.