VOTER TURNOUT IN PRESIDENTIAL ELECTIONS, 1964-1976

This dissertation is concerned with recent changes in levels of voter turnout for presidential elections. Turnout decreased at each election from 1960 to 1976 despite a variety of changes taking place in American society which might have been expected to lead to increased turnout. The present research attempts to shed light on this paradox by means of a longitudinal analysis of a variety of surveys of the American electorate. Data collected by the Bureau of the Census were used in an investigation of changes in turnout across various demographically defined subgroups between 1964 and 1976. The Census surveys provide information about approximately 100,000 persons. Data collected by the University of Michigan's Survey Research Center (SRC) were used in an analysis of the attitudinal correlates of the turnout decline. The SRC samples ranged in size from 1500 to 2000 respondents. A comparison of the Census and SRC surveys revealed that although turnout levels were generally higher in the SRC series than in the Census series, the two series provided similar estimates of the relationships of turnout to such demographic characteristics as region, color, and sex. The traditional sex differential in turnout was found to have narrowed steadily from 1964 to 1976 so that by 1976, female turnout equalled that of males among non-Southern whites and Southern blacks and exceeded that of males among non-Southern blacks. Reversals of the traditional sex differential were particularly pronounced within younger, more educated segments of the population. Further analysis revealed that the sex differential in turnout was related to sex differences in patterns of labor force participation. Differential change by color and region between 1964 and 1968 was also observed; the turnout of non-Southern blacks decreased sharply over this four year period while that of Southern blacks increased markedly. No further change in the color differential was observed in either region from 1968 to 1976, although this finding was later discovered to be the result of increases in the color differential among younger, more educated persons and counterbalancing decreases in this differential among older, less educated persons. Also observed over this 12 year period was a decrease in the traditional regional turnout differential. From 1964 to 1968 and again from 1972 to 1976, Southern and non-Southern turnout rates converged. Approximately 27% of the turnout decline observed between 1968 and 1972 could be attributed to the lowering of the voting age. Among whites, the greatest decreases in turnout over the entire 12 year interval were observed among persons between the ages of 45 and 54. Among blacks, the greatest decreases were observed among the 25 to 34 year old age group while actual increases were observed at the oldest end of the age continuum. Patterns of change by education paralleled to a certain extent the age-related patterns, with the greatest turnout decreases from 1964 to 1976 being observed among whites with between 9 and 11 years of education and among black high school graduates. The attitudinal correlates of turnout examined in this research included measures of party identification, political interest, political efficacy, and political trust. It was found that the aggregate turnout decline from 1964 to 1972 could be statistically "explained" by decreasing party identification and increasing political cynicism. Further analysis revealed that increased cynicism accounted for the sharp decrease in turnout observed among non-Southern blacks between 1964 and 1968 and suppressed the increase in turnout observed among Southern blacks over this same period. The other attitudinal items, although cross-sectionally related to turnout in theoretically predicted ways, failed to explain away the turnout decline