A Think Again Media Group publication
True Crime Stumper

Can avoiding eye contact reveal who's lying?

Gaze aversion: what the eye-direction experiments measured

By Phil Springer, Editor · Published · True Crime Stumper

Gaze aversion: what the eye-direction experiments measured

Gaze aversion is a visible movement with several possible causes, rather than an established signal of dishonesty. Looking away, or looking toward a particular upper corner, does not reliably identify a lie. In the FBI Law Enforcement Bulletin in June 2011, David Matsumoto and colleagues reported that 23 of 24 peer-reviewed studies of eye behavior had rejected its use as a reliable indicator. Subsequent experiments tested a more specific directional claim, while research on memory and mental effort documented ordinary reasons for taking one's eyes off another person's face.

Richard Wiseman and colleagues tested the directional claim in PLOS ONE in 2012. The version associated with neuro-linguistic programming linked a right-handed speaker's upward-right look with constructing an event and an upward-left look with remembering one. Speaker and observer perspectives reverse those directions. In the first experiment, 32 right-handed participants gave both truthful and deceptive interviews. Two coders examined the resulting 64 recordings with the audio removed, counting brief glances and longer gazes toward the upper corners. The predicted differences between truthful and deceptive conditions did not emerge.

A second experiment tested whether teaching the proposed pattern helped people detect deception. Fifty participants judged interview clips, with one group given the directional explanation and the other left uninstructed. Training produced no significant advantage in accuracy or confidence. A third experiment examined 52 public appeals for missing relatives, divided equally between apparently genuine and deceptive appeals on the basis of substantial external evidence. Coding those higher-stakes recordings again found no significant directional difference. The experiments therefore tested the proposed eye pattern in controlled interviews, judgments by observers, and real public appeals.

The first experiment used a concrete event that could be described truthfully or falsely. Participants took a mobile telephone into an office. In the truthful condition they placed it in a drawer and later described what they had done. In the deceptive condition they hid it on themselves and claimed to have placed it in the drawer. The interviewer did not know which condition applied. This arrangement gave the researchers a known event against which to classify the account, rather than inferring its truth from the speaker's eyes.

A later experiment changed the demands on eye contact itself. Deborah Riby, Gwyneth Doherty-Sneddon and Lisa Whittle recorded participants completing mental arithmetic with an experimenter. The study included people developing typically and groups with autism spectrum disorder and Williams syndrome. During the first phase, participants could use gaze aversion naturally. During the second, they were asked to maintain eye contact throughout. Task accuracy decreased under the sustained-contact instruction across all groups. The result concerns cognitive performance during arithmetic; the participants were not being assessed for deception.

That study also reported differences in how easily the groups maintained the required gaze. Participants with autism spectrum disorder had difficulty sustaining eye contact throughout the interaction. Those with Williams syndrome found it especially difficult during thinking. The researchers described faces as useful sources of social information that can also impose processing demands. These findings make an individual's visible eye contact difficult to interpret without the activity and the person's circumstances. The same outward behavior can occur in an experiment where the relevant demand is calculating, with no need for a concealed event.

The gaze aversion study tailored its arithmetic questions to participants' abilities after discussion with parents or teachers. Each phase contained six questions. The researchers separated listening, thinking and speaking, recorded answer accuracy, and coded the video frame by frame to check the participants' gaze. The groups with autism spectrum disorder and Williams syndrome were each compared with typically developing participants matched on verbal ability. Matching did not make the two clinical groups equivalent to each other. These design details locate the result in a particular task and comparison, rather than in a general assessment of someone's willingness to cooperate.

The FBI bulletin likewise emphasized that truthful people can be nervous under consequential interviewing conditions. In the research discussed there, outcomes depended on whether an interviewer believed a participant, so truthfulness alone did not remove the possibility of an unpleasant consequence. Its authors also described only weak associations between deception and familiar cues such as fidgeting, posture and voice stress. Treating those observations as reliable signals can misclassify an honest witness or lend unwarranted confidence to a deceptive statement.

The research on gaze aversion differs from Wiseman's directional test. Wiseman and colleagues addressed a limitation of their laboratory interviews: the experimenter had authorised the lies, and the task carried little personal consequence. The public-appeal recordings provided a more consequential setting, yet still showed no significant directional pattern. The researchers suggested that selective memory might help explain the belief's popularity: someone could remember an apparent success while overlooking failures. They presented that possibility as a question for future research. Their experiments tested the proposed eye-direction rule; they did not establish a replacement shortcut for deciding whether an unfamiliar person is honest.

The Question

A suspect avoids eye contact and looks up and to the left during questioning. Textbook portrayals often associate those behaviors with lying. Does the research establish that connection?

Show the answer

The Answer

No scientific evidence establishes eye behavior or gaze aversion as a reliable gauge of truthfulness. Looking away or looking up and to the left does not reliably establish whether someone is lying. Twenty-three of 24 peer-reviewed studies of eye behavior rejected that hypothesis.

Send this question to a friend

Read next

By clicking “Send me a stumper” you agree to True Crime Stumper’s Terms of Use and Privacy Policy.