On election night, we're used to poring over maps as votes are reported. In general elections, this can get kinda boring, since the places where Democrats or Republicans do well are often very similar from year to year. Democrats dominated in big cities and college towns? Yawn. Republicans are racking up margins in exurban and rural areas? Wake me up when we get to the unexpected part.
As it turns out, primaries — at least in New Hampshire — have some geographic consistency, too. My University of Pennsylvania colleague Gall Sigler and I estimated just how tightly correlated the results of different Republican primaries in New Hampshire have been across the state's 320 precincts since 2008. We found some evidence that different GOP candidates aligned with the establishment tended to find support in the same communities, while Trumpier candidates also had similar bases of support.
For example, in 2010, the GOP primary for U.S. Senate pitted establishment-backed then-state Attorney General Kelly Ayotte against more conservative political outsider Ovid LaMontagne. Ayotte eked out a victory by under 2,000 votes.
Ayotte's support from establishment-aligned voters was positively correlated with that of then-Sen. John McCain, the winner of New Hampshire's 2008 presidential primary, and Sen. Marco Rubio's support in 2016. That said, both correlations were fairly modest. McCain, Ayotte and Rubio had overlapping, but by no means identical, bases of geographic support. In part, that's because they faced different fields of opposing candidates.
By contrast, Trump's 2016 support in New Hampshire was negatively correlated with McCain's and Ayotte's support. In other words,Trump won with disproportionate support from New Hampshire precincts that had backed Ayotte's and McCain's opponents.
These patterns have held up in more recent elections. In 2022, establishment-aligned former state Senate President Chuck Morse narrowly lost the primary for U.S. Senate to more right-wing retired Brigadier General Don Bolduc, who had publicly doubted the legitimacy of Biden's 2020 win. And Bolduc's performance was positively correlated with Trump's performance in 2016.
All this makes sense: When primaries pit the same factions within each party against each other year after year, they're going to produce similar maps. And in New Hampshire, those differences broadly track with levels of educational attainment: The places where Trumpy candidates tend to do well have smaller college-educated populations, and the places where establishment Republicans do well are more college-educated. I expect to see that divide reappear tonight.
—Dan Hopkins, 538 contributor