People are much better at recognising faces of their own racial group than those of different races, but a new study led by one of our PhD students in the UNSW School of Psychology suggests that drinking alcohol almost eliminates that bias.

When given enough alcohol to be mildly intoxicated (about 0.05 parts per million blood alcohol level), experimental study subjects were no better at recognising faces from their own race than from different races.

The study, conducted by PhD student Ms Kirin Hilliar along with her supervisor Dr Richard Kemp and collaborator Dr Tom Denson was recently published in the journal Law and Human Behaviour.

The study tested about 140 university students of Western European and east-Asian descent and found that recognition of different-race faces was unaffected by alcohol - yet both groups lost their "own-race bias". No such change was observed in a control group given a placebo drink.

The team notes that scientific evidence for the own-race bias is well established and has been found consistently across a variety of racial and age groups, and a variety of recognition tasks. It is, however, unrelated to racism or levels of racial prejudice.

But because it involves recognition it is often raised as an issue when the reliability of identification evidence by a different-race eyewitness is considered in criminal prosecutions.

Yet few studies have explored how that bias might be influenced by other factors known to affect eyewitness identification accuracy, such as alcohol intoxication. Many crimes – particularly violent ones - occur when both victims and witnesses are affected by alcohol.

"Alcohol has a negative effect on people’s memory for information," says Ms Hilliar. "Our results thus have both practical and theoretical implications. They raise potential concerns for eyewitness accuracy in some conditions, and they shed light on the mechanisms underlying the own-race bias.

"Interestingly, intoxication only had a negative effect on participants’ recognition for same-race face, which was significantly worse when intoxicated than when sober. Yet it had no substantial negative effect on recognition for different-race faces.

"In the placebo group we found the normal own-race bias: people are better at recognising same-race faces compared to different-race faces. Under alcohol conditions, however, this bias was significantly reduced to the point of it being practically eliminated.

"Quite counter-intuitively, our accuracy levels for different-race faces is the same when sober and when drunk: the two well-documented negative effects of trying to identify a person or object when drunk and trying to identify a different-race face, do not have an additive effect."

The researchers note that the study conditions did not involve having an intoxicated witness view the faces specifically in an identification parade, so further research would need to address that limitation.