[ Question ]

How good is a human's gut judgement at guessing someone's IQ?

byhabryka 6mo25th Feb 201915 comments


In her latest post, Sarah Constantin writes:

Whatever ability IQ tests and math tests measure, I believe that lacking that ability doesn’t have any effect on one’s ability to make a good social impression or even to “seem smart” in conversation.

I remember that I made the opposite prediction a few months ago, while I was studying psychometrics more actively. In particular, I had the sense that because of the positive manifold (i.e. almost all positive mental attributes quite strongly correlate), it seems like judging someone's IQ should be relatively easy, and also quite valuable in assessing a large number of other positive qualities, which should make it quite evolutionarily advantageous to be able to assess it.

Two concrete experiments I've thought about:

1. Take a group of people whose IQ you now, then just take pictures of their faces (or short videos of them saying something), and then have a group of participants rate them on their intelligence (probably in order). See how strong they correlate.

2. Throw a bunch of participants into a group, have them talk to each other, then afterwards ask them about the relative judgement of the intelligence of the other members of the group, see how predictive they are.

These seem like very straightforward experiments, that I can imagine someone having already run, and where I would be pretty interested in the results. If only so that I can be better calibrated on how much to trust my gut judgement of someone's intelligence.

Does anyone know of any similar experiments that have been run?

New Answer
Ask Related Question
New Comment

4 Answers

Googling around phrases like 'perception of intelligence' seems to be a keyword for a relevant literature. On a very cursory skim (i.e. no more than what you see here) it seems to suggest "people can estimate intelligence of strangers better than chance (but with plenty of room for error and bias), even with limited exposure". E.g.:

Perceived Intelligence Is Associated with Measured Intelligence in Men but Not Women (Note in this study the assessment was done purely on looking at a photograph of someone's face)

Accurate Intelligence Assessments in Social Interactions: Mediators and Gender Effects (Abstract starts with: "Research indicates that people can assess a stranger's measured intelligence more accurately than expected by chance, based on minimal information involving appearance and behavior.")

Thin Slices of Behavior as Cues of Personality and Intelligence. (Short 1-2min slices of behaviour in a variety of contexts leads to assessments by strangers that positively correlate with administered test scores for IQ and big 5)

The amount of effort people invest in seeming smarter than they are, for example in situations like hiring or sales, suggests to me that people's gut judgments are easy enough to fool to be worth it.

I expect that what people are actually doing is using a few heuristics to make the judgment. I would also expect that courtesy of things like Dunning-Kruger, people towards the bottom will be as bad at estimating IQ as they are competence at any particular thing.

If we just stick with the intuition provided by fact that some people are really terrible at guessing this sort of thing and some people are not, I further expect that what people use as a heuristic changes with ability level. In example: people at the bottom might base it on whether someone agrees with them about stuff; those in the middle might focus heavily on whether someone has a smart-person job like doctor or lawyer; people towards the top might look more at the way someone does things, like some sign that they chose to think carefully about it or apply some technique known to them.

Faception is a startup which claims to have developed machine learning algorithms that can classify IQ and other things (including "academic researcher", "professional poker player", and "terrorist").

If it is the case that AI can actually do this well, then I'd take that as evidence that humans might be capable of it too.

(From a quick look it's unclear how successful they are. They seemed to be focused on military/security applications, i.e. detecting terrorists.)

Research on how humans attribute intelligence to each other suggests that most people attribute higher intelligence to efficient agents, i.e., agents whose actions most resemble what we believe is the best course of action in an uncertain situation. A minority attributes intelligence to agents based on the outcome of their actions.

People attribute intelligence to efficiency in proportion to their own ability to plan. Less competent planners attribute intelligence to outcome. To elaborate, a competent planner would evaluate the behavior of an agent and understand whether it acted rationally, regardless of its success. On the other hand, a bad planner cannot infer the reasoning of an agent; since he himself cannot foresee the best course of action, he would not understand why an agent made a particular decision, and would likely attribute high intelligence to an irrational but lucky agent since its behavior would result in a good outcome.

My interpretation of these results is that intelligent people are better at evaluating the intelligence of others. I believe that given enough time (and thus enough information to evaluate) an intelligent person can adequately estimate a peer's intelligence. It could be argued that "seeming" smart requires a sufficient degree of awareness and intelligence, although of a different kind; perhaps IQ tests do not have a wide enough scope to accurately measure different intelligence in its different forms.


Kryven, Marta. "Attributed Intelligence." (2018).