Abstract
The ability to accurately verify facial identity has important forensic implications, but this ability is fallible. Research suggests that anxiety at the time of encoding can impair subsequent recall, but no studies have investigated the effects of anxiety at the time of recall in an experimental paradigm. This study addresses this gap using the carbon dioxide (CO2) model of anxiety induction. Thirty participants completed two inhalations: one of 7.5% CO2-enriched air and one of medical air (i.e., placebo). Prior to each inhalation, participants were presented with 16 facial images (50% own-ethnicity, 50% other-ethnicity). During the inhalation they were required to identify which faces had been seen before from a set of 32 images (16 seen-before and 16 novel images). Identification accuracy was lower during CO2 inhalation compared to air (F[1,29]=5.5, p=.026, ηp(2)=.16), and false alarm rate was higher for other-ethnicity faces compared to own-ethnicity faces (F[1,29]=11.3, p=.002, ηp(2)=.28). There was no evidence of gas by ethnicity interactions for accuracy or false alarms (ps>.34). Ratings of decision confidence did not differ by gas condition, suggesting that participants were unaware of differences in performance. These findings suggest that anxiety, at the point of recognition, impairs facial identification accuracy. This has substantial implications for eyewitness memory situations, and suggests that efforts should be made to attenuate the anxiety in these situations in order to improve the validity of identification.