Comparing models of voter turnout using American National Election Studies data

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Blair T. Johnson


Voter turnout, National Election Studies, Presidential elections

Subject Categories

American Politics | Social Psychology


In the US presidential elections of the 1950s and 1960s the percentage of voter turnout for US citizens of voting age was typically between the high 50s to low 60s. By 1988 voter turnout hovered around 50%, where it has remained for most subsequent presidential elections. Concern has arisen that the American electorate has become progressively detached from the electoral process (W. E. Miller, 1992). To better understand voter behavior, substantial research has been conducted by scholars in a variety of social science disciplines. Still, the proximal determinants of voter turnout have eluded researchers studying this vexing problem.

Particularly promising as models of voter turnout are the behavioral models extant within social psychology (e.g., Fishbein & Ajzen, 1975) because those models may allow for a more directed search for the proximal causes of voter behaviors. This dissertation examines the affect of several such variants that emerged from regression, path and neural network analyses of the NES database (i.e., the 1976, 1980, 1984, and 1988 presidential election samples). Several of these social psychological models that were tested were found to be more parsimonious and very comparable in predictive power with previous atheoretical voter turnout models reported in the literature (e.g., Freedman & Goldstein, 1996).

Additionally, the models that were the focus of this dissertation offer a social psychological perspective that helps explicate the individual level psychological constructs which appear to be associated with, if not determinative of voter turnout. Those psychological constructs are voter intentions and prior-related behaviors (e.g., Bentler & Speckart, 1979). It is hoped that future research will further develop an understanding of the antecedent causal factors, which trigger changes in the proximal determinants of voting. Further it is hoped that the Presidential Voter Turnout Model ( PVTM ) will spur research in social psychology as well as in the voter turnout field.