I’m the skunk at the garden party when it comes to statements about humans’ ability to detect deception. There are people like Paul Ekman who claims he can detect high-stakes lies at very high levels of accuracy and can train others to do the same. There are many book authors who also pretend that they can teach you to tell when someone is lying, sometimes with ludicrous tips (e.g., is the person you suspect touching his or her nose?). They probably make much more money than I do from writing books on deception.
My belief is that people are barely better than chance at knowing when people are lying and when they are telling the truth. That is based on a review paper Charlie Bond and I published based on a statistical analysis (meta-analysis) of every study ever done -- there were hundreds of them -- on human deception detection success. We found that if you ask people directly whether someone is lying or telling the truth, they will, on the average, only be correct about 54% of the time, when they could get 50% correct just by guessing.
But today in the New York Times, in “The search for our inner lie detectors,” a reporter raised the question of whether we might know more at an unconscious level than is apparent from our explicit judgments of whether or not a person is lying. Since he quoted me in the article, I thought I’d tell you a bit more about my previous research on this topic. You can find a more complete discussion of my own research on this topic (and other people’s) in The Hows and Whys of Lies; here I’ll just share some highlights.
In most of my studies of deception-detection, my colleagues and I showed people video clips of people who were either lying or telling the truth. (I’ll call the people viewing the video clips judges of deception, in the layperson sense of the term. In some studies, the judges interacted face-to-face with the people who were lying or telling the truth, rather than just watching them on video, and sometimes they listened to audio clips or read transcripts.) We always ask the judges to tell us whether they think the person in the clip was lying or telling the truth. That’s a direct, explicit measure of deception-detection.
We also ask the judges other questions – indirect questions. Here are some examples of questions we asked them about their impressions of the persons who were lying or telling the truth:
We also asked the judges about their own feelings after each clip that they watched (or heard or read, or each interaction they had). For example, we asked them:
We did studies like these with adults and we also did a developmental deception-detection study with people ranging in age from 6th graders through college students. In almost every instance, we found that people could distinguish lies from truths more effectively by their answers to all of their indirect questions than they could by telling us directly whether they thought someone was lying or telling the truth.
For example, people are more likely to feel comfortable and confident in their judgments, and more likely to feel that they got enough information, and less likely to feel suspicious, when the person they were judging was telling the truth than when that person was lying. They also think the person they are judging seemed less comfortable and more ambivalent when that person was lying than when that person was telling the truth. All of those distinctions are sharper than the distinctions made by direct judgments of whether the person was deceptive, or just how deceptive the person seemed.
In sum, when people are asked directly whether someone is lying or telling the truth, their accuracy is just a little bit better than chance. But when you ask them a less direct question, that does not involve calling someone a liar, then they are a little better than they were before. They still aren’t great at separating the liars from the truth-tellers, but they are better than they were with the most explicit questions.
My favorite study on this topic was not my own, but a doctoral dissertation of a student in my lab, Eric Anderson. His study was based on that awkward couple thing, where one person in the couple points to someone else and asks the partner – Do you think that person is attractive?
Anderson brought people into the lab and showed them pictures of other people. For each picture, he asked them whether they thought the person was attractive. Behind a one-way mirror, their romantic partner was watching them. So was a total stranger, of the same sex as their romantic partner. (So if a woman was watching her boyfriend, then the other person watching her boyfriend was also a woman.) But, importantly, the people watching from behind the mirror could not see the pictures of the people being evaluated for their attractiveness. They were simply asked whether the person answering the questions (“Do you think that person is attractive?”) was lying or telling the truth. They were also asked some of the other indirect questions listed above.
When asked directly whether the person was lying or telling the truth, the romantic partners were barely better than chance, at 52%. The complete strangers weren’t all that great at deception-detection either, but at 58%, they were better than the romantic partners. So who knows better whether your partner thinks someone else is attractive – you or a stranger? If we go by direct questions about lying or telling the truth, the answer is that perfect strangers know better.
But what about those indirect measures? Both the romantic partners and the strangers felt like they got more of the information that they needed, and they also felt more comfortable, when the person they were judging was telling the truth than when they were lying. They felt more suspicious when the person was lying than when the person was telling the truth. That means that they all achieved a certain level of indirect deception detection. Again, their indirect deception detection was not that great but it was better than their direct deception detection (based on their answers to the explicit question of whether the person was lying).
Now here’s the final twist: Even though the romantic partners were worse than the strangers at detecting deception when asked directly whether the person was lying, they were better than the strangers at indirect deception detection. Their feelings of comfort and of having gotten enough information, and their judgments of suspiciousness, separated the lies from the truths more effectively than the strangers’ judgments did.
Here’s my final paragraph about this research, from The Hows and Whys of Lies:
What we do not yet understand, but would very much like to explore, is this disconnect between partners’ direct ratings of deceptiveness and their gut intuitions. Are the partners not aware that their feelings of confidence and suspiciousness and perceptions of other people’s suspiciousness are varying in ways that could be meaningful? Do they have any clue at all that there could be a link between these kinds of feelings and whether or not their partner is lying? And if they were clued in on this clue, would it even matter? Could they use that information effectively, or would their attempts to use it undermine the process whereby they form these meaningful impressions and intuitions? And finally, the more sinister question: If they could use this information to find out who their partners really did find attractive, would they really want to know this? Maybe they should just let sleeping frauds lie.
Note: If you are interested, you can find links to my other books about deception (as well as my singles books) here. Below are links to other blog posts about the psychology of lying and detecting lies. Next time, back to my real passion: Living Single!
I. First, Some Truths about Lies
II. Lying and Becoming a Liar
III. Figuring Out When You Are Getting Duped