Although I did not entirely ignore the subject. I should have paid more attention to people’s estimates of their own intelligence. Self-estimates are error prone, and may have negative consequences in real life, as well as making discussions about intelligence remarkable error-prone.
Adrian Furnham did several papers on this topic, and Sophie Von Stumm made a particularly good contribution in 2013.
Von Stumm looked at self-estimated intelligence, asking people to rate themselves in general (Distal Self Estimated) and then regarding a particular mental task they had just completed (Proximal Self Estimated) and compared it with actual test results on that intelligence test.
Von Stumm said:
Two meta-analyses reported that IQ scores correlated at about .30 with SEI, suggesting that people have some insight into the rank-order of their IQ score compared with others. At the same time, however, people generally overestimate their IQ with SEI scores being typically 1 Standard Deviation (SD; i.e., 115) above people’s actual IQ test scores. The above-average-effect is thought to result from people’s need to maintain optimistic self-judgements to lead a productive, happy life in a sometimes uncaring world.
The above-average-effect tends to be greater for people of low ability, who may be either unable (i.e., lack of insight) or unwilling (i.e.,embarrassment) to acknowledge their mental deficits. For example, college students with exam scores in the 10th percentile estimated their scores to be in the 60th percentile. By comparison, students who scored in the 90th percentile also estimated their scores to be in the 90th percentile.
Men tend to report higher SEI than women–a phenomenon which is known as the ‘male hubris – female humility’ bias. Also, children rate their fathers as more intelligent than their mothers, and parents think their sons are more intelligent than their daughters with (112 IQ points compared to 105 IQ points. That said, gender differences in SEI distortions are largely independent of actual intelligence or personality differences between men and women but they seem to be related to gender stereotypes.
Here are the results in a graph:
The graph shows (green line) the average IQ of each quartile, based on intelligence testing. The other lines show for men (squares) and women (circles) their self estimated intelligence in general (distal) and their estimate of how well they did on the intelligence tests in particular. People in the lower intelligence quartile strongly over-estimate their abilities. People in the upper quartile somewhat under-estimate their abilities.
Dunning-Kruger rules, and the extent of the problem is clear to see.
Here is von Stumm’s discussion about the reasons behind the discrepancy:
Dunning et al. (2003) suggested that top performers’ underestimation has a different source than the overestimation of poor performers: top performers know how well they perform in absolute terms, such as their raw test scores, but they overestimate how well other people are doing on the same test (Fussell & Krauss, 1992). Conversely, poor performers overestimate their intelligence because they ‘lack the skills to produce correct answers, [and]they are also cursed with an inability to know when their answers,or anyone else’s,are right or wrong’ (Dunning et al., 2003, p. 85). Overall, people tend to have imperfect information about their own intelligence and even worse information about others’ intelligence (Moore & Healy, 2008).
These finding have wide relevance. Half the population over-estimate their intelligence because they can’t work out how badly they have done, and the other half under-estimate their intelligence because they over-estimate the other half’s abilities.
As a consequence, half the population will have a sneaking suspicion that intelligence tests are a cheat, the other half a conviction that intelligence is so widespread that testing not necessary. Overall, everyone will have the mistaken impression that they are pretty much alike in intelligence, so actual differences in mental power will be compressed and minimized in public estimation.
Now a new paper has come out about romantic couple’s estimates of their intelligence. This is interesting in itself, and the results also throw light on popular views about intelligence, and perhaps also why manny people belittle intelligence testing.
People tend to overestimate their romantic partner’s intelligence even more than their own.
Gilles E. Gignac, Marcin Zajenkowski. Intelligence 73 (2019) 41–51.
One feature stands out:
Dwelling on this a moment, one thing becomes clear: many people are immensely deluded. They think themselves two standard deviations brighter than they really are.
In fact, the scores on the Raven’s Matrices were corrected for two decades of Flynn Effect. Without the correction, the scores would still be 1.5 standard deviations too high. Lake Wobegon on steroids.
Back to the main point: people seem to be over-estimating their intelligence by 30 IQ points and their partner’s IQ by 38 points in the case of women doing the judgments, and 36 points in the case of men doing the judgments. People are deluded about their abilities, and deluded about their partners’ abilities. Delusion plus 7 points. This is dreadful, but also highly illuminating. No wonder so many people hate actual intelligence tests.
We have seen that the estimates of intelligence are wildly exaggerated. It is small comfort that they correlate at roughly r= 0.3 with actual intelligence, which suggests limited insight. However, people value intelligence in prospective romantic partners, but for that to have real-life consequences they have to be able to judge intelligence accurately. They seem able to do so only to a limited extent, in that couples correlate 0.3 – 0.4 for measured intelligence, perhaps based more on verbal abilities, which are easier to judge.
Subjects were 218 heterosexual Polish couples, average age 28, and having spent 6 years together, time enough surely to note intellectual abilities or the lack of them. They had been selected by trained pollsters. Couples rated their own and their partner’s intelligence on a 25 point scale. Everyone did the Advanced Progressive Matrices test, and the 1994 norms were used. I assume none of them separated when they were debriefed about the results. (Actually, I would be interested in hearing the estimates given by divorced couples after the decree absolute).
Women were no better than men at predicting their partner’s results. Intelligence differences did not relate to relationship satisfaction. Although intelligence rates second or third in the list of desirable partner characteristics, couples do not show evidence of active assortment by intelligence. Perhaps they are matching up by prettiness (which was not measured in this study, but which has an impact in real life).
Why are these estimated so wildly exaggerated? The usual over-estimation is about 8 points, but that figure comes from a small sample of undergraduates, so is probably unrepresentative. If we stick with these results, leaving aside possible problems with self-estimation measures, the effect may be due to narcissism, which leads to poor decisions and more accidents. The finding is consistent with the observation that a majority of people think that they are better than average at most things. (It would follow from this that the average person does not understand averages. However, depending on the distributions, they may be using the median, not the average). One reason the self-estimates were so high is that they were made before intelligence testing, which generally reduces self-estimates by 5 points, because delusions are somewhat deflated. However, realism wears off after a week or so.
This is a good paper, well conducted, and modestly written up, with a clear view of what could be added in further research. In my view the couples aspect is interesting because intelligence turns out not to be much of a selection criterion. However, that finding is put in the shade by the sheer magnitude of the high regard people have for their intellectual powers. No-one wants to be average, even when they are.
I don’t do policy, but it seems prudent not to take people at their own high self-regard, but to test them carefully before letting them get close to heavy equipment.