The strange links between intelligence and prejudice

Human judgement often becomes less accurate when we train it on ourselves. Self appraisals commonly flatter our strengths and minimise our weaknesses. The average man overstates his height by 1.2cm and the average woman understates her weight by 1.4kg.

Judgements of our bodily dimensions may be prone to distortion but they are constrained by the brute facts of physical reality. A short person cannot claim to be tall without losing credibility.

However, when we judge our psychological characteristics we are not constrained in the same way. We may be remarkably inaccurate in our self assessments, as if we were observing our mental capacities in a fun-house mirror.

Self-assessed intelligence

These judgement biases have been studied in assessments of general cognitive ability or intelligence. Intelligence can be assessed formally using psychometric tests but it can also be informally estimated. Researchers have examined whether people’s estimates of their intelligence accurately reflect their psychometric intelligence.

Two striking findings have emerged from this research. First, people tend to hold inflated impressions of their own intelligence: most people think they are above average.

This is an example of the “better-than-average” effect, a widespread illusion of personal superiority. The illusion has been documented in people’s appraisals of their personality, health, work performance, relationship satisfaction and driving skill. People also tend to believe they are above average in their immunity to judgement biases.

A second key finding is that people’s self-assessed intelligence is poorly calibrated. There is only a weak relationship between self-assessed and psychometric intelligence.

Suppose we collected a sample of 100 people and selected one person whose self-assessed intelligence was in the top 50. There is only a roughly 60% chance they would be in the top 50 on psychometric intelligence, not much better than a coin toss.

If we combine the inflation and poor calibration of self-assessed intelligence, we arrive at a situation like the one shown below. Let’s take 100 people from the general population and divide them evenly into those who are above (blue) and below (red) average on psychometric intelligence. Let’s also divide them into those people (let’s say 80, a conservative estimate) who estimate their intelligence to be above average (dark) and those who estimate it to be below average (light).

The table makes a few sobering points. Only a slender majority of people (58%) accurately estimate where they sit relative to others. A large minority of people (36%) incorrectly estimate they are above average, dwarfing the group (6%) who underestimate their intelligence.

Most people who have above average intelligence correctly estimate they are above average. However, most people who have below average intelligence mistakenly make the same estimate.

This pattern exemplifies the “Dunning-Kruger effect”. That cognitive bias involves a tendency for people with relatively low ability to overestimate their ability, in part because they lack the capacity to recognise their lack of competence.

Intelligence, self-assessed intelligence and prejudice

Research on self-assessed intelligence shows the people who think they are above average are not the same as those who are above average. This discrepancy reveals itself powerfully in an article published this month by a team of Belgian psychologists.

The researchers examined a sample of Belgian adults from the general community. The sample completed a psychometric intelligence test and estimated their intelligence on a scale from 0 (least intelligent Belgian) to 100 (most intelligent Belgian). The average estimate was 67: roughly 85% of the sample believed themselves to be above average. The two ways of assessing intelligence were very weakly related.

The study also employed a measure of subtle racism, included because greater psychometric intelligence is associated with lesser prejudice. The researchers explored whether psychometric and self-assessed intelligence had the same or different links to racism.

Remarkably the two ways of assessing intelligence had opposite associations with subtle racism. As expected, higher psychometric intelligence was associated with lower racism, largely because more intelligent people thought about social groups in less crudely categorical ways. However, higher self-assessed intelligence was associated with higher levels of racism.

The explanation for this finding is that people who estimate their cognitive ability to be higher than others tend to perceive the social world vertically in terms of superiority and inferiority. Such people are high in “social dominance orientation”, an anti-egalitarian ideology linked to prejudice.

Similar findings have been found in studies of narcissism. Narcissistic people believe they are superior, have inflated estimates of their intelligence, and they also tend to hold more prejudiced attitudes.

In essence, the Belgian study shows that being intelligent undermines group prejudice, but believing one is superior to others in intelligence reflects and promotes it. When people make assessments of their intelligence they are estimating a cognitive strength, but perhaps also revealing an attitudinal weakness.