Objective. This study develops and tests a model of political regionalism that posits that if regions are politically exceptional, then individuals sharing the same profile but living in these different regions will have divergent presidential voting patterns (King, 1996).

Methods. Analyzing presidential voting behavior from 1952 to 2004, I use logistic regression techniques to test a regional model of homogeneity (southern exceptionalism) versus a unit model of homogeneity (South and Non-South are statistically similar).

Results. The findings show that the South's presidential voting patterns are exceptional in the 1950s and during the civil rights era but, starting in the Reagan era, southern exceptionalism waned. These findings also show that the South is converging with the non-South (northernization) relative to the influences of race, family income, union membership, in-migrants, and gender, and the non-South is converging with the South (southernization) relative to the influences of education, blue-collar workers, and age.

Conclusions. Both economic class and race variables contribute to the demise of regional exceptionalism; however, race plays a more persistent role. Given the process of “southernization” and the instability of the predictors of presidential voting for the South over time, I conclude that the study of the South as a region should continue until the process of change subsides and a new equilibrium is found.