New empirical paper from David Brockman et al:

A prominent literature argues that moderate candidates perform better in general elections, but a competing literature that emphasizes voters’ partisan loyalties contests these predictions. The 2020 Democratic Presidential primary represented an opportunity to speak to these debates due to relatively high voter information about multiple moderate and extreme candidates running in the same election. We present results from both an existing (n = 62,275) and an original (n = 12,725) national survey that asked how respondents would choose in a general election between one of the Democratic candidates and Republican Donald Trump. Our evidence is consistent with moderates having an electoral advantage: more moderate Democratic candidates receive more support against Trump than do more extreme Democratic candidates. Republican partisans contribute to Democratic moderates’ advantage: ≈2% select Trump against a more extreme Democrat but not against a more moderate Democrat. One of the more extreme candidates, Bernie Sanders, ostensibly deviates from this pattern, but only under unlikely assumptions about turnout. Our point estimates are larger than Biden’s margin of victory in pivotal states in the 2020 general election. Our findings lend further support to the hypothesis that moderate candidates are electorally advantaged.