The present experiment examined the apparently incongruent findings that police officers (i) believe that suspects show nervous behaviour when they lie but (ii) seem able to detect deceit in suspects who do not show nervous behaviours. It was hypothesized that police officers’ judgements about whether a suspect is lying would not be correlated with their judgements regarding whether the suspect is tense, but, instead, would be correlated with their judgements regarding whether suspects are having to think hard or attempting to control their behaviour. Each of 84 police officers saw seven truths and seven lies told by suspects during their police interviews. Participants in Condition 1 were asked to indicate after each clip whether the suspect was lying, whereas participants in Condition 2 were asked to note after each clip to what extent the suspect appeared to be tense, having to think hard or attempting to control their behaviour. Condition 3 was a combination of Conditions 1 and 2 and participants were asked all the above questions. The findings support the hypotheses and therefore negate the assumption that police officers primarily look for cues of nervousness when they attempt to detect deceit.