The current study examined three primary goals within the deception detection literature that have not been previously assessed in combination. First, we assessed if decisions about the veracity of a statement are more accurate in dyads than individuals. Second, we evaluated whether type of lie (mock transgressions versus real transgression [RT]) influenced observers' ability to distinguish a truthful statement from a lie. Third, we assessed if the presentation mode (audio versus video with audio) impacted deception detection accuracy. A sample of college students (n = 282) evaluated eight previously recorded statements for veracity. True and false statements were elicited through a cheating paradigm (i.e. RTs) or a mock transgression paradigm. Participants either viewed and listened to these statements or only listened to an audio recording. Results of an analysis of variance indicated that working in dyads did not improve deception detection accuracy. Participants were significantly more accurate in detecting deception in situations where the speaker believes a RT has occurred. The highest accuracy rates were achieved in the real transgression audio condition. Implications of these results for the practice of detecting deceit are discussed. Copyright © 2014 John Wiley & Sons, Ltd.