Participation rates, socioeconomic class biases, and congressional elections: A crossvalidation

Theory: Recently, Leighley and Nagler (1992) presented evidence indicating that class biases in voter turnout remained relatively stable across presidential elections from 1960 through 1988. Nevertheless, few investigations test for class biases (or the stability of class biases) in midterm congressional elections. Given that presidential elections are such high-intensity affairs (Campbell 1993) some lower-class citizens likely participate in presidential elections as a result of the sheer salience of the campaign but may be less likely to be mobilized in the lower intensity midterm elections. Consequently, we might expect to find evidence of an increase in class bias in the shrinking congressional electorate rather than in the shrinking presidential electorate. Hypothesis: Class biases in voter turnout have not substantially increased since the 1960s. Methods: A demographic model of voter turnout is used in order to isolate the marginal impact of income on voter turnout across election years using the American National Election Studies 1958-94 and the Current Population Surveys 1974-94. Results: While class biases in any specific election are great (higher socioeconomic groups comprise the majority of voters) such biases appear to have remained relatively stable over time. The declining rates of turnout since the early 1960s have occurred among all segments of society, not just among the lower classes.