Recently, President Trump visited Montana to campaign against U.S. Sen. Jon Tester — hoping to help Tester’s Republican challenger Matt Rosendale this November.
Many pundits are treating the upcoming midterms as a referendum on Trump. But, far away from D.C., in places like Montana, candidates are sparring over something else: Who really has an authentic connection to the state? And, as our research reveals, connection to place matters a lot to voters.
Tester is one of a handful of “red state Democrats” battling for reelection just two years after Trump defeated Hillary Clinton in his state — by a resounding 21 points. Tester’s strategy been simple: Convince voters that he is the most authentically Montanan candidate in the race, or perhaps more accurately, persuade voters that he is the least non-Montanan.
Tester is not alone in trying to demonstrate an authentic connection to place. To date, both candidates have focused their campaigns almost entirely on showing that their opponent isn’t connected to Montana — while proving that they are the true Montanan.
The Tester campaign has hammered on the fact that Rosendale is a native of Maryland — and would rather develop than protect Montana’s public lands. Rosendale’s GOP allies, meanwhile, have pointed voters’ attention to the fact that Tester bought a home in the D.C. area and — they claim — has embraced “the swamp” and lost touch with Montana.
But is there any evidence that these sorts of appeals work? Our research, forthcoming at Political Research Quarterly, suggests that yes, this can be a potentially effective strategy, but that it may work differently in rural and in urban places.
Here’s how we did our research
To study the effects of geographically-based appeals, we conducted an experimental study. The study featured a convenience sample of about 1,000 respondents from across the U.S. and was conducted online via Amazon’s Mechanical Turk.
We divided respondents into three groups, each of which saw one of three simple political advertisements designed to mimic a standard campaign mailer. One of these ads used urban imagery, featuring the skyline of the largest city in the respondent’s state. Another used rural imagery featuring the largest federal or state park in the respondent’s state. The third ad featured a control image with no geographical imagery.
The ads looked like this:
Respondents could identify themselves as living in an area that was very rural, somewhat rural, more rural than urban, more urban than rural, somewhat urban, or very urban. Each ad was seen by roughly one-third of our respondents, regardless of where they live on the urban-rural continuum.
We then asked respondents a series of questions about the candidate that included an overall favorability rating as well as ratings of the candidate on specific dimensions, including the candidate’s ability to understand and know how to help people in the respondent’s state.
Yes, the imagery mattered – but differently to rural and urban voters
Respondents who reported living in a very urban area were statistically more likely to approve of the candidate portrayed as urban than the candidate portrayed as rural, by 13 percent. When compared to the control candidate with geographically neutral background, urbanites rated the rural candidate as roughly the same.
In addition, urbanites of all stripes — including those identifying as “somewhat” and “very” urban — rated the urban candidate as being more likely to understand and know how to help people in their state than the rural candidate. These findings suggest that candidates hoping to represent urban voters may want to show that they identify with the city.
Rural respondents, meanwhile, were very hostile toward the urban candidate. Indeed, very rural respondents were more than 20 percent less likely to approve of the urban candidate than either the rural or the control candidates. Considering how simple and subtle our experiment was, these results are especially striking.
Rural respondents also rated the urban candidate as less likely to understand people like themselves than the control candidate — but rated the rural and control candidates as roughly the same.
Taken together, these results suggest that Tester, Rosendale and other such candidates are right to emphasize that their opponent isn’t from “here,” or whatever rural voters might consider to be “here.” In other words, while candidates trying to appeal to rural voters may sometimes get a boost from signalling a shared “sense of place,” what’s even more important is painting their opponents as being out of step with rural values — that is, as “urbanizing” the opponent.
Place-based connections in the 2018 midterms
As the midterms draw closer, expect to see more ads in which candidates stress that they belong to the places they hope to represent, while their opponents do not. For both groups, the city is a mobilizing emotional force. Urbanites want to know the candidate identifies with their city — while rural voters love to hate candidates from the city.
What might that mean in practice? Red-state Senate Democrats defending seats in primarily rural states that Trump won two years ago — such as Montana, North Dakota, and West Virginia — might try to paint their opponents as being out of touch outsiders. In states with more urban voters — for instance, Missouri and Indiana — candidates might take care to signal a connection with and understanding of those cities’ residents.
Back in Montana, the senate election is shaping up to be another close one. That means the race may well be decided by which anti-Montanan caricature voters dislike more: “Maryland Matt” Rosendale or “swamp rat” Tester.
Nicholas F. Jacobs is a PhD candidate in the Woodrow Wilson Department of Politics at the University of Virginia.