Abstract
Face ‘Superrecognisers’ not only recognise faces seen long before, but appear to remember where they met, at least anecdotally. It is not theoretically obvious that good face memory should be associated with good source memory (e.g., for specific locations). We examined this question in two experiments. In Experiment 1, typical perceivers (n=78) were shown 6 face images of the same sex, sequentially, each superimposed on a different background image. At test, they were shown 6 new images of the study faces and 6 unseen faces and asked to make an old/new judgement. If they said old, they were presented with four of the backgrounds and asked to pick the one the face was on at study. They were then asked to click on the specific location they thought the face was on. The process was then repeated, using faces of the opposite sex and new backgrounds. There was a correlation of .53 between d’ (sensitivity) for face memory and the proportion of faces for which the backgrounds were also correctly remembered, and a correlation of -.36 between d’ and error in locating the face, in pixels. In Experiment 2, we removed the cue of the background image at test. Typical perceivers (n=66) were shown 10 faces at study, each presented on one of 10 different backgrounds which could be grouped by type (office, bedroom, bathroom, etc.). For each of the 20 test images reported as old, a word list of the possible location types was presented for a decision about the background at study. The process was again repeated using the opposite sex of faces. The correlation between d’ and the proportion of correct types of background chosen was .62, replicating the result from Experiment 1. Our data suggest a strong correlation between face and source memory in typical perceivers.