Abstract
Most humans become experts at recognizing faces by adulthood, but a debate arises around how to study the development of this expertise. We propose video game character faces as ecologically valid stimuli to study the development of face expertise. Exposure to these faces can be quantified and video game players typically have hundreds of hours logged for each game. We ran a 3-AFC short-delay recognition task with upright and inverted character faces sourced from The Elder Scrolls V: Skyrim to test if participants show an other race effect (ORE). The two races we tested were Nord (a race based on Nordic Caucasians) and Redguard (a race based on Arabic North Africans). We recruited 49 novice participants who had never played Skyrim from UC Santa Cruz, where African Americans make up 2% of the population (UCSC ODEI, 2016). Results showed a robust face inversion effect: upright faces were correctly recognized significantly more often (M: 0.607, SE: 0.021) than inverted faces (M: 0.426, SE: 0.016; t(48) = 8.68, p< 0.00001). We also found a smaller but reliable ORE: performance for upright Nord faces (M: 0.648, SE: 0.021) was significantly higher than for upright Redguard faces (M: 0.566, SE: 0.029; t(48) = 2.80, p=0.007), whereas there was no difference in performance for inverted faces (p>0.5). However, the size of the inversion effect was only marginally larger for Nord faces than Redguard faces (p=0.10). Our results demonstrate evidence of expertise and other race effects in video game faces in a novice population. In a follow-up study we investigate whether Skyrim experts (participants who have played more than 50 hours of the game) will show an even stronger ORE based on their level of expertise.