Many socio-geographical or political patterns also lead to vote distributions that resemble election fraud, for instance, producing an association between vote share and turnout. Differences in turnout and in party identification are often due to differences in the political behaviour of social groups. For instance, in the United States, whites tend to be characterized by higher turnout, and stronger support for Republicans. Hence, a positive correlation of the two is not necessarily evidence of error, ballot box stuffing, or other election malpractice (Grofman et al., 1999).