Race Bias in Facial Recognition

Hi Josh

Thanks for the interest.
In my most recent study with UK and Chinese participants, I found that people show an own race bias when trying to match 2 images of faces (e.g. like trying to match a person with their passport), where they were more accurate with own race faces as compared to faces that were a different races to them. This own race bias was even more pronounced when removing the hair cues from faces. This is really interesting as many face recognition studies use faces where they remove the hair cues and crop around the faces. It also has implications for when people change their hair, or cover it in some way, like a scarf or a hat.

I moved our convo away from the intros thread since it’s just for intros :slight_smile:

These findings are fascinating. Are there any theories as to why hair cues seem to play such an important role?

And to what extent do you think own race bias is due to biological determinants? Or could it be due to the race-groups that we are regularly exposed to in social settings? For example, if a person is socialised into a different race-group, does own race bias fade?