Abstract
Judgments about the top halves of two faces are influenced by irrelevant differences in the bottom halves of the faces. For example, subjects typically take longer to correctly determine that the top halves are the same if the bottom halves differ. Interestingly, this so-called composite face effect (CFE) is greatly diminished if the top and bottom halves of the face are misaligned. These results have been attributed to automatic, holistic processing of aligned faces, which presumably aids recognition in naturalistic contexts. However, to our knowledge there has been no measurement of the association between the CFE and face identification. In the current experiment, we measured the correlation between the CFE and accuracy in a face identification task in a group of 24 young adults.
We used the stimuli and methods of Le Grand et al. (2004) to measure the CFE. Each trial consisted of two successively-presented faces, and subjects determined if the top halves differed. (The bottom halves always differed and were irrelevant to the task). The top and bottom halves of each face were aligned or misaligned in separate blocks of trials. For each subject, the CFE was defined as the difference between RT on “same” trials in the aligned and misaligned conditions. Subjects also participated in an identification task (Bruce et al., 1999). On each trial, a target face and a lineup of 10 faces were presented. The task was to determine if the target was present in the lineup and, if so, to select the face that matched the target. The dependent variable was percent correct.
There were large individual differences in both tasks. However, performance in the two tasks was not correlated (r=-0.2, p=0.34). This result suggests that the mechanisms that produce the chimeric face effect may not facilitate accurate face identification.
This research was supported by NSERC grants 42133, and 105494, and the Canada Research Chair programme.