Paper: To freeze or not to freeze – A culture-sensitive motion capture approach to detecting deceit

Scroll this

In 2015 we published our first conference paper on automated motion-based lie detection. Together with Ronald Poppe, Paul J. Taylor and Ross Anderson, I put 180 participants in motion capture suits and studied their behavior during interviews. In these interviews, people told either truths or lies and we measured interviewee and interviewer behavior using Xsens motion capture suits (17 inertial sensors each). The majority of our interviewees moved more across their entire body (from head to toe) when lying compared to truth telling. Within-sample predictions showed an detection accuracy of 82%, which was almost 30% better than the detection accuracy of participant interviewers, who had the opportunity to interview them.

Previously, we wrote a book chapter on cross-cultural deception detection When writing this chapter, we realised that the majority of deception research is conducted in Western countries on Western participants, making it hard to generalize these findings globally. To address this issue, we added a culture manipulation to this research. We tested British and South Asian participants (in both within and cross-cultural interviews) and interestingly, did not find much difference in the way they moved when telling truths and lies. South Asian participants did score much higher on the stereotype threat questionnaire than British participants.

With some updates and cool new insights, this research is now published as a journal article in PLOS ONE! You can check our paper out here.