In this paper we argue that the literature underestimates the value of primaries because it focuses on overall average effects. We argue that primary elections are most needed in safe constituencies, where the advantaged party's candidate can usually win the general election 舒 even if she is low quality. If the main role of elections is to select good candidates, then advantaged party primaries in open seat races are particularly consequential. We provide evidence that these primaries are especially effective at selecting high quality types. This appears to be driven both by differences in the proportion of high quality candidates competing in the primaries and also by voter behavior.