Purpose: Few studies have assessed deception during reallife, high-stakes encounters. This study is one of the largest and most geographically diverse to investigate how criminal suspects lie during investigative interviews. It is also one of the most specific; focusing solely on those who committed sexually motivated homicides. Methods: Sections of transcripts from 52 sexually motivated homicide offender interrogations were analysed using Linguistic Inquiry and Word Count software. Truthful (n = 27) and deceptive (n = 25) statements, corroborated through physical evidence, were then compared using the reality monitoring (RM) model of deception. Results: Support for the RM model was mixed. Truthful statements contained more motion and spatial details. There were no significant differences between true and deceptive statements when comparing perceptual, affective, and cognitive process details. Conclusions: The results support the notion that there are verbal cues to deception detectable in high-stakes, real-life situations. It also provides a starting point to assess these cues in special forensic populations.