One-at-a-time versus grouped presentation of mug book pictures: some surprising results

Eyewitnesses to a simulated crime attempted to identify the perpetrator from a computerized mug book. The 208 mug book pictures were presented either 1 mug shot per page or in groups of 12 mug shots per page. Half of the mug books were arranged by similarity to the perpetrator as determined by a facial recognition algorithm, and half were randomly arranged. In contrast to past findings with photospreads, false-positive identifications were significantly higher using the one-at-a-time procedure than the grouped procedure. Results suggest that the best practice for mug books may be the use of groups of pictures per page rather than the one-at-a-time procedure long advocated by experts for use in lineups and photospreads.