This new article by psychologist Satoshi Kanazawa concludes that more intelligent people are more likely to be politically liberal [HT: Ronald Bailey]. It has gotten a great deal of media attention, for example from CNN and Time. In reality, the article doesn’t actually prove any such thing. It has several significant methodological flaws.
I. Conflating Liberalism with Universalism.
Kanazawa uses a highly idisoyncratic definition of liberalism: “genuine concern for the welfare of genetically unrelated others and the willingness to contribute larger proportions of private resources for the welfare of such others.” This definition doesn’t distinguish liberalism from conservatism or libertarianism. It distinguishes universalism from particularism. For example, a libertarian who believes that free market policies best promote the welfare of “genetically unrelated others” and contributes a great deal of his money to charities promoting libertarian causes counts as a liberal under this definition. The same goes for a Religious Right conservative who believes that everyone will be better off under socially conservative policies and contributes lots of money to church charities. In fact, recent research by Arthur Brooks shows that conservatives and other opponents of government redistribution give more, on average, to charity than other members of the population.
When Kanazawa actually correlates measures of intelligence with views of particular issues, he finds that, controlling for various other variables, more intelligent General Social Survey (GSS) respondents are less likely to support government-mandated efforts to ” reduce the differences in income between people with high incomes and those with low incomes.” This is hardly consistent with claims that the more intelligent are more politically liberal in the conventional sense of the term.
II. Relying on Dubious Ideological Self-Identifications.
To be sure, Kanazawa also cites surveys showing that the more intelligent are more likely than others to describe themselves as “liberal” and less likely to call themselves “conservative.” However, decades of research show that large percentages of the population have a poor understanding of political ideology and have a poor grasp of the meaning of terms like “liberal” and “conservative” (at least as they are understood by pundits and scholars). This is part of the more general phenonemon of widespread rational political ignorance. Ideological self-identifications are notoriously unreliable as indicators of real political views. For example, more African-Americans describe themselves as “conservative” than “liberal,” even though this description fits neither their issue positions nor their voting patterns. In recent decades, the term “liberal” has acquired a negative connotation, so much so that many liberals have taken to calling themselves “progressives.” This makes it likely that some liberal survey respondents won’t identify with the term, especially among the less-educated and less politically knowledgeable.
The shortcomings of self-identification are exacerbated by the fact that many voters hold views that are hard to place on a one-dimensional liberal-conservative spectrum. Some 10-15% of Americans hold libertarian views, and there is also a significant proportion who are the opposite of libertarian (socially conservative, economically left-wing).
Finally, Kanazawa fails to control for the effects of political knowledge, which is probably correlated with intelligence, but also distinct from it. Previous research shows that increased political knowledge causes people to become more libertarian (economically conservative and socially liberal) than they would be otherwise, with the important exception of views on taxation. This is almost exactly what Kanazawa finds to be the impact of intelligence, including the taxation result.
III. Are Ideologies Endorsed by More Intelligent People More Likely to Be Correct?
I suspect that much of the public interest in Kanazawa’s study is driven by a perception that political views endorsed by more intelligent people are more likely to be true. This, however, is a dubious inference. Even intelligent people have incentives to be rationally ignorant about politics and to do a poor job of evaluating the information they do know. I do think that, other things equal, a political view is more likely to be correct if it is more likely to be endorsed by people with greater knowledge of the issue (controlling for other factors that may affect their answers). While knowledge and intelligence are likely to be correlated, they are not the same thing. Ultimately, the fact that a political ideology is more likely to be endorsed by more intelligent people is only a weak indicator of its validity.
Interestingly, Kanazawa himself does not claim that intelligent people are more likely to endorse liberalism because it is true. Instead, he argues that the result is due to the fact that liberalism is more at odds with our genetic instincts than conservatism is, and intelligent people are more likely to endorse “novel” ideas. I don’t agree with the “Burkean conservative” view that we should have a strong presumption in favor of following tradition. But the opposite presumption is also an error.
UPDATE: Kanazawa also argues that his data prove that more intelligent people are more likely to endorse atheism. If time permits, I will address this issue a separate post.
UPDATE #2: Here is a link to an ungated PDF of Kanazawa’s article [HT: James Joyner, who also has links to various other critiques of Kanazawa].