Catching the bad guy: morphing composite faces helps.

Abstract

Composite faces built by eyewitnesses commonly are poor likenesses of the target face. When there are multiple witnesses, however, an opportunity exists to morph the composites. Morphs were rated as more similar to the target face than were the mean ratings of the individual composites. However, as hypothesized, the morph also came to resemble non-target faces more than the individual composites did (a prototype effect). This prototype effect was so strong that the morphs resembled non-targets more than the individual composites resembled the targets. In addition, morphing composites produced an attractiveness bias, which made the morphing of composites less effective for less attractive targets. Even when the prototype effect and the attractiveness bias were controlled for, however, a true morph-superiority effect continued to exist.