This thesis examines the use of computer-mediated communication for lie-detection and interpersonal deceit. The literature within the fields of lie-detection and mediated communication are reviewed and it is proposed that there is a lack of knowledge surrounding how people use CMC to deceive one another. Qualitative research was carried out in order to address this shortcoming, exploring the self-reported experiences of chat room users who have been exposed to online deceit. Reports were provided that describe the misrepresentation of age, gender, vocation, affection, and appearance. The importance of stereotypes in driving suspicions is also emphasised within the reports. It is suggested that this key characteristic has more dominance in CMC than it would do face-to-face because of the occlusion of the traditional nonstrategic clues to deceit. Evidence for an alternative set of nonstrategic leakage clues was examined further by conducting a variant of the Guilty-Knowledge test within the context of a CMC based crime. It was found that participants exhibited a response time inhibition effect when presented with 'guilty knowledge' and that this effect was detectable through a standard two-button mouse. The use of such nonstrategic cues to deceit was explored further in a study that examined how CMC might be used to add additional control to a Statement Validity Assessment truth-validation test. It was found that the content analysis technique used by SVA was unable in its present form to correctly distinguish between truthful and fabricated statements of participants interviewed using a CMC chat program. In addition, it was found that the deletion-behaviours of participants fabricating a story within CMC provided no quantitative or qualitative evidence that they were lying.