Politics: Campbell Sharman and Janette Stuart, ‘Patterns of state voting in national referendums’, XVI: 2, November 1981, pp. 261–270

Abstract

The relationship between cognitive ability and stated political preferences in the United States is examined with data from the General Social Survey, which includes a brief vocabulary test (Wordsum) as a measure of verbal ability. Since the 1970s, liberal and conservative self-identification became increasingly identified with the Democratic and Republican parties, respectively. Liberal self-identification has increasingly been related to higher Wordsum scores since the 1970s, but liberal-conservative differences rarely exceed the equivalent of 3 IQ points. Among Whites, those identifying themselves as “moderate” or “independent” have lower average Wordsum scores than those with stated ideological or political party preferences, contrary to the hypothesis that higher intelligence is related to less extreme political positions. The relationship between Wordsum and Democratic Party affiliation has moved from negative to neutral since the 1970s. In presidential elections, the most consistent finding is that voters scored substantially higher than non-voters. Those voting for the Democratic candidate had higher average scores than those voting for his Republican opponent since 2000. In regression models that control for demographics, higher Wordsum scores are associated with liberal self-identification but not with political party preferences. In conclusion, higher vocabulary scores are associated with a greater likelihood that people place themselves on the ideological and political spectrum and that they vote in presidential elections, but have only small relationships with liberal-versus-conservative self-identification.