13th November 2012
In the olden days, scientists knew that money didn’t noticeably matter for incumbents: If incumbents won or lost it wasn’t because of differences in campaign spending. But in crude regressions, it looked like money mattered for challengers. Now, with fancier regressions, it appears money matters little for challengers as well, especially in well-studied, data-heavy House elections.
The old results were driven by omitted variables: Weak incumbents drew in stronger challengers, and hence more challenger cash. The challenger cash was a sign of the challenger’s relative strength, not much of a cause of her relative strength.