Suspicions Confirmed

If you're a conservative, you're not going to like this post. If you're a liberal, you probably knew all along that those pointy hats that Ku Klux Klan members wear were actually dunce caps.

photo by greenyourdecor

A new study by Gordon Hodson and Michael Busseri found that intelligence is negatively related to both conservatism and prejudice. That is, the higher people's I. Q., the lower their conservatism and their prejudice. These relationships were obtained in three separate samples. In two nationally representative British samples (almost 16,000 people), intelligence was measured when the participants were 10 or 11, while conservatism and prejudice toward racial minorities was measured in their early thirties. The potentially contaminating effects of both social class and education were statistically removed prior to analyzing the data. In a smaller American college sample (254 people), all three measures were taken at the same time, and prejudice measure was their dislike of gay people.

The relationships they obtained were the same in all three samples and are strong enough to take seriously. Of course, that doesn't mean that all racists are stupid or that all stupid people are racists. I did some back-of-the-envelope calculations based on the average correlations they report for the two British samples. (I will use the word “correlation” to describe the relationships that were found, although the data analysis actually used standardized regression coefficients, a slightly better estimate of the relationships.) If a British child was below the median in I. Q. at age 10 or 11, there was a 57% chance he or she would be above the median in ethnic prejudice at age 30-33. Alternatively, if a child was above the median in I. Q., the chance of being above average in prejudice drops to 43%.

The relationship of I. Q. to conservatism is actually stronger. A child below the median in I. Q. has a 63% chance of being above the median in conservatism. Finally, the strongest relationship is that between conservatism and prejudice. A person who is above the median in conservatism has a 74% chance of being above the median in prejudice. This is nothing new. The strong relationship between conservatism and prejudice has been well-known to social scientists since the 1950s. For example, recent surveys of "tea party" members suggest that they are more prejudiced than the average American. The mass media usually politely refrain from mentioning this relationship.

The authors claim that the relationship between intelligence and prejudice is mediated by conservatism. The term implies causality. In plain language, here's what it means. Intelligence comes first. (In this case, this is literally true, since intelligence was measured about 20 years before conservatism and prejudice.) Intelligence first has the effect of increasing conservatism. Then, as a result of their conservatism, these less intelligent people go on to become more prejudiced. Had they not first become conservatives, they would not have become as prejudiced as they did. The implicit assumption is that “intelligence causes conservatism, which in turn causes prejudice.”

The generally accepted way of testing a mediational hypothesis is this. First, compute the correlation between intelligence and prejudice. Then recompute the correlation between intelligence, partialling out, or eliminating, the effect of their common relationship to conservatism. This correlation should be significantly lower than the initial correlation between intelligence and prejudice. This is the case in all three samples, supporting the mediational hypothesis.

The catch here is that these data, strong as they are, do not conclusively support the everyday meaning of the term “mediates.” These are correlational data, and correlation does not mean causation. (The rising of the sun does not cause us to get out of bed, even though the two are highly correlated.  See the video below.) There are alternative interpretations. Since conservatism and prejudice were measured at the same time, it is possible that prejudice mediates the relationship between intelligence and conservatism. Given the pattern of correlations observed here, if the authors were to calculate the relationship between intelligence and conservatism removing the effect of prejudice, they would get the same result.

The idea that prejudice comes before conservatism is not implausible. It is possible that children have experiences, either in real life or through the mass media, that lead them to form negative attitudes toward minorities long before they decide that women should not work or that prisoners should receive longer sentences (two items from their conservatism scale). One way of looking at conservatism is that it is an ideology that provides a socially acceptable justification for a person's existing prejudices.

It is also possible that there are one or more unknown variables that are responsible for the relationship between intelligence, on the one hand, and both conservatism and prejudice, on the other. Here's my favorite “fourth variable” explanation. Maybe less intelligent people don't really enjoy thinking (because they're not very good at it). A psychologist might say that they are low in need for cognition, or that they are cognitive misers.

Conservatism is a much simpler ideology than liberalism. Its explanations of behavior usually focus on a single variable, usually a personal characteristic of the individual in question, i.e., “She is poor because she is lazy.” A liberal's explanation of poverty involves several variables, most of them external to the person, that are related to one another in complex ways over a long period of time. Conservative theories are more readily communicated on television not because they are more correct but because they are easier for a slightly inattentive audience to understand. Maybe it is the fact that people low in intelligence would rather not be bothered to think that causes them to accept both the simplistic ideology of conservatism and overbroad negative generalizations about minorities.

That's my theory, anyway. To the best of my knowledge, it remains to be tested.