ABSTRACT
Participants (N = 106) performed an affective priming task with facial primes that varied in their skin tone and facial physiognomy, and, which were presented either in color or in gray-scale. Participants' racial evaluations were more positive for Eurocentric than for Afrocentric physiognomy faces. Light skin tone faces were evaluated more positively than dark skin tone faces, but the magnitude of this effect depended on the mode of color presentation. The results suggest that in affective priming tasks, faces might not be processed holistically, and instead, visual features of facial priming stimuli independently affect implicit evaluations.
Acknowledgments
Parts of the data in this article were presented at the annual meeting of the Society for Personality and Social Psychology, January 2007, Memphis, TN. The authors thank Chris Loersch for his helpful comments on a draft of this manuscript.
Notes
1. Because participants were forced to respond so quickly, there was relatively little variability in response times. Accordingly, although the patterns of results were similar for response times, some of the effects did not reach significance.
2. We performed all analyses on error rates for both positive and negative targets rather than on a single positivity index (e.g., when error rates on positive trials are subtracted from error rates on negative targets, see ). Although a convenient summary measure, this single index might not be very reflective of the original error rates, as it does not indicate if participants are more positive or less negative towards facial categories.
3. Note that all analyses are collapsed across race of participants; we could not perform racial comparisons, as the numbers of participants for some groups were very low. When analyses were performed with data from White participants only, results were analogous to those reported for the entire sample. Participants' gender did not moderate any of the effects that we describe, so all analyses collapsed across this factor as well.
4. This, of course, is an inherently low power test because it involves the correlation of mean ratings for 6 stimuli rather than the more typical correlation of scores from a larger sample of individuals. It is necessary to take this approach because different samples provided the mean typicality ratings and the implicit evaluations. On the other hand, the separate samples are a strength of this comparison because it reduces biases (e.g., consistency) that might exist if the same individuals provided both sets of means.