PATTERNS AND SOURCES OF TICKET SPLITTING IN SUBPRESIDENTIAL VOTING

The primary source of divided government in the United States is voters who split their ballots between the parties. Yet there has been little comprehensive examination of either patterns or sources of ticket splitting in recent years. Instead, divergent lines of research have emerged, emphasizing such things as voter partisanship, incumbency, and a ''new'' (young, well-educated, even partisan) kind of ticket splitter, and their focus has been too often restricted to the atypical president-Congress pair. We seek to unify these research traditions in a comprehensive model of split-ticket voting and to test this model across the partisan ballot in a typical election setting-here, the contests for five Ohio state-wide offices in 1990. The model incorporates partisan strength, candidate visibility, and the individual characteristics that distinguish the ''new ticket splitters ''. The results support our partisan strength and candidate visibility explanations but provide little support for the emergence of a new type of ticket splitter.