This thesis tests the hypothesis that increasing urbanization in the South is positively correlated with rising Republican voting in presidential and gubernatorial races. A measure of urbanization was derived by subjecting socio-economic data from three censuses for all southern counties to factor analysis. This measure was regressed against Republican percentages of presidential vote in 1952, 1960 and 1968, and against GOP percentages in governor's races closest to the census years. The coefficients of correlation were uniformly low, reaching as high as .50 only once in each case. It was concluded that urbanization accounts for little in explaining variation in Republican voting and that contradictory findings are the result of reliance on less powerful analytic techniques, misunderstanding of more powerful ones or inadequate operationalization of key concepts.