Cognitive approach to lie detection rendered useless by made-up alibi

GettyImages-531979628.jpgBy Alex Fradera

The desire to catch people in a lie has led to the development of techniques that are meant to detect the physical markers of dishonesty – from the polygraph to brain scans. However, these methods are often found wanting. The insights of cognitive psychologists have arguably fared better, based on the idea that lying is more mentally demanding than telling the truth – real knowledge is automatically called to mind when we are questioned, and this needs to be inhibited  before we answer, leading to slower responses. Unfortunately new research in the Journal of Experimental Psychology: Applied seems to pour cold water on the idea of using these subtle reaction-time differences to develop objective (and cheap) measures to get at the truth. The findings suggest that all it takes to render this cognitive approach ineffective is a prepared false alibi.

The University of Würzburg team led by Anna Foerster set up a quasi-realistic situation, sending each of their 36 participants into a room – containing various items like a computer, pen and paper and USB stick – to read instructions in a sealed envelope that sent them on a mission. The mission involved drawing shapes on a piece of paper, tearing it and hiding the pieces in different parts of the room. Once they’d finished, the participants discovered they were going to be questioned about their actions, and that they must keep the mission secret by giving a false account of what they’d done. Their alibi was that they’d used the USB stick to email a file out of the room.

After familiarising themselves with their alibi, the participants went to another room where they answered questions on computer about the mission, mixed in with “routine” questions with predictable answers (e.g. Did you cross the street today? Did you win the lottery?).

The participants’ task was to use Yes/No responses to answer the routine questions honestly, while responding to the mission questions according to their false alibi. At other times (denoted with an onscreen cue), they had to do the reverse, lying in response to the true routine questions while giving a true account of their mission.

The participants answered more slowly and less accurately when lying in response to the routine questions than when telling the truth. In other words, it was easier to tell the truth than deny it, consistent with the research literature on the cognitive demands of lying. However, for the mission questions, the opposite was true. The participants’ response times and accuracy suggested they found it easier to answer in line with their false alibi, than tell the truth about what they’d actually done on the mission.

The next experiment looked at another cognitive sign of lying based on pointing movements. The mission and alibi were the same as before, but this time the participants answered questions on an iPad, tracing their finger towards onscreen Yes or No options which popped up randomly at different sides of the screen. Normally when people lie in this kind of situation, they hesitate before starting their response, they move more slowly, and instead of making a beeline for the lie response, their finger’s trajectory is indirect, curving towards the honest response. All these effects were seen in the current experiment when participants denied the truth of routine questions. But for the mission questions, the reverse was the case: it was the confessional true responses that were slower and curvier.

There are a number of competing theories for what’s behind these findings. An alibi might weaken the presence of the true information, or become such a strong mental model that now it’s the first point of call. But regardless, the new results don’t look good for what was considered a solid lie detection application – and existing attempts to apply this cognitive approach to real-life court cases (such as this investigation in 2008 that studied the speed of a woman’s answers alongside brain scan evidence) may need to be looked at with renewed skepticism.

Lying upside-down: Alibis reverse cognitive burdens of dishonesty

Alex Fradera (@alexfradera) is Staff Writer at BPS Research Digest

5 thoughts on “Cognitive approach to lie detection rendered useless by made-up alibi”

  1. In what order were the questions requiring answers on a compter screen asked? Were they answered in blocks or at random? Was the order of the questions reversed for half the participants randomly assigned i.e. counterbalanced? How old and computer- literate were the participants? Were there any differences in responses between age groups, gender etc.?

    As a participant in many such experiments in the lab, the more familiar I became with the task and the screen, the faster my responses became over time – the practice effect. In giving the alibi after the task, maybe the alibi was more salient and more easily retrievable than the task – recency effect?

    Research into false memories (memories that are believed to be accurate and true by the individual, even when the memory can be proven false), indicate that it is certainly possible for ‘false memories’ to mimic ‘true’ responses in their characteristics. I think it is accepted now, that memories (retained in the short and long term) become more embedded and easily retrieved following rehearsal / learning, which is presumably, what the participants had to do to make their alibi believable.

    Memories are constructed and reconstructed and subject to a battery of cognitive biases involved in the formation of beliefs about what really happened, rather than a ‘video recording’ account. This common mistake about memory renders traditional questioning and lie- detection techniques unreliable at best and where relied upon in a court of law, very dangerous ‘tools’ indeed.

    Whilst I welcome the challenge this research poses to accepted ‘wisdom’ about speed of responses and lie detection, I would want to have a great deal more research with external validity.

Comments are closed.