Deception and communication media
Brundell, Patrick Robert (2013) Deception and communication media. PhD thesis, University of Nottingham.
Much research investigating deception and its detection has focused upon face-to-face communication, but over recent years the variety and extent of new communication media has changed the contexts in which deception might take place. Although work has attempted to characterise communication media, a much smaller body of research exists which has investigated the frequency with which people lie with different media and the detection of deceit under different communication media conditions. Through questionnaires and experimental studies, this work investigated the perceptions that both deceivers (senders) and those attempting to identify lies (receivers) have about communication media and how this relates to their observed behaviour. Results from questionnaire studies suggested that both the characteristics of deception and media influence people's perceived discomfort and believability when lying and the media choices they might make if they are planning to deceive. Some important factors appeared to be the seriousness of the deception, who senders are lying to, and the general frequency with which they use particular means to communicate. Communication media were judged to be similar and dissimilar to each other on a range of characteristics which may impact their appropriateness for deception and lie detection. There was evidence that media used at low frequency in daily life may be more likely to be chosen for deception. In laboratory studies, senders were found to lie more frequently using audio-only media compared to audio-video. There was evidence from experimental studies that detection of deceit was more successful when communication was audio only compared to audio-video. There was little consistent evidence that judgement biases varied between media conditions, but a truth bias was identified in experimental studies. No evidence was identified that interactivity between senders and receivers significantly influenced response biases or lie detection accuracy. A small corpus of messages recorded under audio-video and audio-only conditions were selected for their detectability or believability from two senders, and presented in modified formats to receivers. Stimuli had video removed or introduced, and were presented as audio-only, audio-video, text-only and video-only. The results suggested that detectability of audio-video and audio-only stimuli was dependent upon the condition stimuli were recorded under rather than presented. When messages were only seen and not heard or read, accuracy of lie detection was compromised. There was evidence that judging transcriptions could allow successful detection, but the accuracy of lie detection was typically lower than demonstrated in richer media conditions. These findings may imply that a combination of information channels and/or paraverbal information is important for accurate classification of honesty and lies. Limitations of the studies and directions for further research were discussed.
Actions (Archive Staff Only)