Abstract

The divisive primary hypothesis predicts that when a primary election contest is hard-fought, or the eventual winner of the primary wins by a close margin, the party will do poorly in the general election. This relationship between primary and general election outcomes was first posited by Key (1953) and tested by Hacker (1965), and has been a venerable part of the elections literature ever since. However, despite the divisive primary hypothesis’ very compelling theoretical underpinnings, it has a tepid empirical track record. For every study which finds that divisive primaries harm party nominees, there’s another one which says they don’t.