Estimation of proportions of white and non-white voters voting for a candidate with applications to possible violations ofsection 2 of the voting rights act

Abstract

Throughout the South, black citizens are challenging at-large methods of election on the grounds that they are discriminatory against minorities and are in violation of Section 2 of the Voting Rights Act of 1965 as amended in 1982. The United States Supreme Court has established tests that must be met to prove a violation. Arguing whether these tests have been satisfied involves estimation of the proportion of voters of each race voting for each candidate in a series of elections. Estimation is complicated by the fact that the analysis must be done on precinct totals rather than on the votes of individual voters. The Supreme Court has considered two methods of estimation, extreme case analysis and ecological regression. These methods are compared with three additional estimation methods, weighted least squares, two stage least squares and maximum likelihood. Extreme case estimators can have considerable bias. In the setting of elections in five political subdivisions of South Carolina, it is shown that th...