In two experiments, a bias in lineups was examined, which has mostly been ignored in previous research. The psychic state of the suspects differs from the state of the fillers, in that they face a situation that can lead to personal loss (if they are identified) or relief (if they are not identified). This state might result in detectable behavioural cues, thus making the lineup unfair. Lineups were conducted with randomly chosen mock suspects who were promised money if they were not identified. Lineup samples were recorded on videotape once before the selection of the mock suspect (serving as a control) and once afterwards. The probability of identification of the mock suspect was significantly increased in the latter compared to the former condition. Copyright © 2004 John Wiley & Sons, Ltd.