14 Comments

Poker players are an example of RH's class of "private professionals whose job is to detect lies," and I'm glad RF provided an example. I don't know what you mean by "criminologists," but I think that would fall under the category of public professionals, who do not seem to do very well, as RH said above, their livelihoods don't depend the quality of their work. I read in some secondary source, probably Gladwell, that although the median police aren't better than the median civilian, the proportion of really good police is higher. Perhaps Navarro is of this class. Or maybe that's not true, it's just easier to find good police because they get a reputation.

Re: criminals.Yes, the papers say they do better, but do they use studies designed without this objection in mind? I doubt it. Studies with half the people lying are probably cheaper than with 10%. An easy enough thing to check.

Expand full comment

Joe Navarro, formerly of the FBI, went up against a psychic, a poker player (Annie Duke) and a few other people in a televised experiment to see who was best at detecting lies. Navarro and Annie were the best. If the question is "is there any group that is actually good at detecting lies?", my money is on the criminologists and poker players. After all, their livelihoods depend on it very directly.

Expand full comment

Do private professionals whose job is to detect lies do any better?

Most of these papers seems to imply that private professionals perform just as badly as anyone else, but have more confidence in themselves.

It's the feedback that seems key.

They may be just as bad as the general population, except that they have defaulted to believing people lie rather than tell the truth.

The papers' claim that they are stictly better, not differently defaulted. And relationships in the criminal world are much more dependent on trust (with no legal system to enforce contracts, trust and fear are the main tools to do so). So it's not just a question of ferretting out the liars but of figuring out the truth tellers.

shouldn't it be possible to have hardened criminals watch the investor relations presentations of corporate officers

The best idea here would be to get people who were involved, as a group, in long-term financial fraud (like the Enron crew). They would probably be superb at this.

Expand full comment

If this is true, shouldn't it be possible to have hardened criminals watch the investor relations presentations of corporate officers, and point out which ones were suspect? Would it be possible to beat the stock market this way?

Actually, one would probably prefer to call upon a disciple of Ekman then a hardened criminal per se.

Expand full comment

TGGP,

The experience is valuable if combined with feedback on accuracy. The parole boards don't get rewarded or punished on the basis of recidivism rates among those they release.

Expand full comment

Maybe this is the answer to Robin's previous post. Use criminals to define the objective truth, then compare popular opinion to that.

Also, if you believe that politics is mostly a sham, as I do, then this would predict that voter turnout is low among criminals.

Expand full comment

"you'd expect that those whose job involves ferreting out liars, such as police officers or immigration judges, would be better at it than the rest of us. But this study claims that Swedish judges on the Migration Board (MB) are about as good at recognising the signs of lying as students"

"unless you have [...] a great experience in being lied to, the best is most definitely not to trust your gut."

The two quotes seem to conflict. Wouldn't people whose jobs involve ferreting out liars have a lot of experience with being lied to?

Expand full comment

Perhaps government civil servants do not have much relevant expertize in their tasks because voters have a hard time monitoring them. Do private professionals whose job is to detect lies do any better?

Expand full comment

I question, though, if prisoners (and people with paranoia, for example) are *accurate* judges of deception, specifically, if in being able to better detect liars their success is from simply suspecting more people are lying. And even if they do successfully catch all of the liars, how often do they wrongly claim non-liars? They may be just as bad as the general population, except that they have defaulted to believing people lie rather than tell the truth.

Expand full comment

As I suggested on my blog, maybe we should make sitting on the migration board a punishment for repeat offenders?

Good idea! But my favourite punishment for repeat offenders is to get them to fill in pages and pages of surveys. A specific population, of course, but within that population a hundred percent response rate and no self-selection...

Expand full comment

But there's also an evolutionary pressure pushing us towards being good liars. And our lie detecting abilities would have evolved socially, i.e. we'd be much better at detecting lies in people we know than in strangers.

The article "Vrij, A. (2000) Detecting Lies and Deceit: The Psychology of Lying and its Implications for Professional Practice," (which I can't access, grrrr) apparently shows:

accuracy rates fall below 60 per cent (Kraut 1980; Vrij 2000). This performance level is hardly impressive, considering the fact that an accuracy rate of 50 per cent is expected by chance alone.That's very poor.

So yes, our guts do provide some indication, but we normally overestimate its reliability - a dangerous thing. We also (in my experience) tend to look at lying differently from other issues (such how reliable is the evidence, is it biased, what does it actually say, etc...)

Take "he may be lying", "he may have made a mistake" or "he may be biased". People I know tend to be reasonably bayesian with the last two, but treat "he may be lying" as an either/or, not an estimate in between.

Expand full comment

As I suggested on my blog, maybe we should make sitting on the migration board a punishment for repeat offenders?

Expand full comment

Oh, finally. Our guts just have to be a tiny bit better than random to be valid as evidence to consider.

Expand full comment

"So, unless you have a criminal record or a great experience in being lied to, the best is most definitely not to trust your gut."

That seems like a wrongful, or at least invalid, conclusion.

It's certainly possible that most liars people are likely to encounter in fact will be nervous, and that this makes signs of nervousness a good indicator that something is fishy.

Possibly you should be careful about going with your gut when trusting or distrusting someone entails risk, but that's not really the same thing as your, I assume lighthearted, conclusion.

All it says is that we're overconfident about our ability to catch lies, we still have to weigh expected outcomes to know what side we should err on.

On a side note I find it a little bit difficult to believe that people are really bad at catching lies, since it seems we have a fair bit of brain tissue dedicated to detecting cheating in social situations. Evolutionary it would seem quite useful to be able to catch any objective clues towards deception. If there are any such clues I would be surprised if our ancestors weren't partly selected for their ability to detect those clues.

Also there's a huge difference between a random liar and an expert liar. Could it be that these studies mostly consider expert liars, or people lieing in situations where there are no pressure?

Expand full comment