| The existence of an own-race face recognition bias has been reliably demonstrated by innumerable experiments. Subsequently, a cross-race effect or deficit for other-race faces must exist. The robustness of these findings is impressive, yet the explanation of the phenomenon remains elusive. This study replicates the own-race bias/cross-race effect, using a yes/no recognition task and Signal Detection analysis (d′), for White, Black, Hispanic and Asian observers and targets. Based on research across domains, e.g., information processing, learning, memory and cognitive processing, a theory utilizing the common denominator of developmental contact is posited. In order to assess such a dimension, observers from the four aforementioned races with multiracial parentage were tested on own- and other-race targets with the same procedures as above. These observers, regardless of self-declared race, demonstrated a multi-own-race bias corresponding to parentage. Predictors of own- and other-race accuracy were investigated in the hopes of establishing some arbitrary measure of eyewitness accuracy. A mock crime and lineup were presented to all observers. Again, for this more generalizable task, an own-race/parentage bias was found. In determining cognitive processing as a function of contact, automatic and rapid decision-making significantly predicted accuracy. Conversely, effortful/process of elimination strategies were less accurate and were accompanied by longer reaction times. Other-race contact from grade school onward was self-reported and was found to not significantly differ across populations. This, combined with own- race expertise/decision-making ability, and the fundamental quality of exposure across the aforementioned domains, support developmental contact as the raison d'être behind the own-race bias/cross-race effect. A model based upon implicit learning through developmental exposure of covariation amongst relational characteristics of own-race faces is proposed. |
