Lie detection: eye movements and face touching under experimental test
Lie detection cannot turn a hand over the mouth, a touched nose or a glance away into an automatic finding of dishonesty. Those movements can occur while someone is remembering, concentrating or feeling uncomfortable. Experiments have tested particular claims about the eyes, while separate laboratory work has measured spontaneous face touching during a memory task. The results offer a much more specific picture than a catalogue of suspicious gestures: a movement can be measurable without identifying a lie.
In 2012, Richard Wiseman, Caroline Watt and their colleagues tested an eye-movement rule associated with neuro-linguistic programming. The proposed rule linked looking upward in one direction with constructing an event and looking in the other direction with remembering one. Some practitioners had extended that distinction into a test of truthfulness. The researchers recruited 32 right-handed participants because the claimed pattern was supposed to be strongest in right-handed people. Each participant completed both a truthful and a deceptive interview.
The task involved an experimenter's mobile phone and an office drawer. In the truthful condition, participants placed the phone inside the drawer and described what they had actually done and seen. In the deceptive condition, they concealed the phone in a pocket or bag, then claimed to have left it in the drawer. They also invented plausible contents and a layout for that drawer. An interviewer who did not know which condition applied conducted the filmed interview. Truth and falsehood were assigned by the procedure, independently of the participant's facial movements.
The lie detection experiment separated glances lasting less than a second from gazes lasting a second or longer. Coders watched the recordings at normal and slow speeds, counting upward movements to either side; the sound was removed so that speech would not influence their coding. Other eye directions were excluded. The proposed differences between truthful and deceptive conditions did not emerge. This was a test of a particular directional rule, with defined movements and a known task, rather than an assessment of every possible use of the eyes.
A second experiment tested the usefulness of teaching that rule. Fifty participants were assigned either to receive information about the proposed eye pattern or to a control group without that information. Both groups watched clips from the first experiment and classified the statements. The trained group did not achieve significantly greater accuracy, and its confidence ratings also did not differ significantly. Receiving a plausible-sounding interpretation of eye movements did not make the movements a useful diagnostic tool under those conditions.
The researchers then examined 52 public appeals for missing relatives, divided equally between appeals classified as deceptive and genuine on the basis of outside evidence. That evidence included such material as security footage, possession of a murder weapon or a missing person later being found alive with an abductor. The classification did not come from deciding who looked shifty. There were no upward gazes lasting at least a second, and the shorter upward glances did not differ significantly between the groups. The directional rule also lacked support in this higher-stakes setting.
Face touching has its own experimental record. Stephanie Margarete Mueller, Sven Martin and Martin Grunwald published a 2019 study of 60 right-handed volunteers performing a demanding memory task. Participants explored shapes by touch, held them in memory for roughly 14 minutes while unpleasant sounds played, and later drew the shapes. The procedure was repeated with different shapes. The task studied memory and distraction rather than lie detection, making it particularly useful for establishing that spontaneous facial contact can arise without an instruction to deceive anyone.
The Leipzig team combined video with muscle recordings and wrist acceleration sensors. They restricted their analysis to brief, spontaneous facial touches, excluding such actions as scratching, yawning and nose picking. Fifty-four of the 60 volunteers touched their faces during the experiment. Across the experimental phases, 194 touches occurred during the memory-retention intervals, compared with 109 during the other phases combined. The contact durations and movement times also changed with the demands of the task. These were observed differences in ordinary task performance, not a division into guilty and innocent people.
That experiment does not assign an innocent explanation to every gesture in an interview. Its findings are narrower: facial touching occurred in a controlled setting involving sustained memory and distracting sounds, and its frequency and duration varied with that setting. Nor can a nose scratch simply be substituted for one of the carefully defined movements, since scratching was excluded. A physiological measurement, a behavioral category and an inference about someone's honesty are separate steps.
In their 2020 review, Aldert Vrij and Ronald Fisher distinguish feeling nervous from displaying movements that observers label nervous. They also note that truthful people may fear being disbelieved, while deceptive people may try to suppress behaviors they expect to look suspicious. Their synthesis therefore adds a practical limit to lie detection based on gestures: discomfort can belong to either side of an interview. The controlled eye tests and the memory experiment leave no basis for treating a touched mouth, a covered nose or an averted gaze as an automatic confession.
- Richard Wiseman and colleagues, The Eyes Don't Have It: Lie Detection and Neuro-Linguistic Programming, PLOS ONE, 2012 — journals.plos.org
- Stephanie Margarete Mueller, Sven Martin and Martin Grunwald, Self-touch: Contact durations and point of touch of spontaneous facial self-touches differ depending on cognitive and emotional load, PLOS ONE, 2019 — journals.plos.org
- Aldert Vrij and Ronald P. Fisher, Unraveling the Misconception About Deception and Nervous Behavior, Frontiers in Psychology, 2020, full text at PMC — pmc.ncbi.nlm.nih.gov
- Samantha Mann, Aldert Vrij and Ray Bull, Suspects, Lies, and Videotape: An Analysis of Authentic High-Stake Liars, Law and Human Behavior, 2002, original abstract — link.springer.com