​​​Virginians have started casting ballots for governor. Recent polls suggest a close race. A Monmouth University poll attributed the tight margin in part to the division between urban and rural Virginia.

In this way, the Virginia race is a microcosm of recent trends. In 2020, urban voters supported Joe Biden 2-to-1, while rural areas backed Donald Trump by the same proportion. This urban-rural divide has expanded dramatically in the past decade, reaching its widest point in the last presidential election. What drives this deepening divide — and will we see it again in next year’s midterm elections? Our research suggests that the answer lies in Americans’ ideas about race.

How we did our research

Trying to understand what causes this political rift between rural and urban counties is a complex and nuanced undertaking. Pundits often offer vague explanations claiming that cultural differences fuel resentment. Earlier this year, James Carville said Democratic messaging doesn’t resonate with rural voters, coming across as “urban, coastal, [and] arrogant.”

But cultural differences are difficult to quantify. We may think we understand how rural and urban Americans approach life differently, but which particular factors really divide the two groups? We chose attitudes about three cultural markers to examine what distinguishes urban and rural American voters: faith, gun ownership and race.

We analyzed newly released data from the Cooperative Election Study (CES), the largest nationally representative academic survey of Americans interviewed before and after the 2020 election. CES data reveals that urban and rural Americans have significantly different attitudes about each of these cultural markers.

For example, while many Americans identify themselves as evangelical Christians, even more do in rural communities. About 48 percent of rural voters identify as evangelical Christians, while fewer than one-third of suburban and urban voters consider themselves “born-again” Christians, at 32 and 30 percent, respectively. Notably, White evangelicals identify as Republicans at a rate of about 75 percent. Rural voters are also significantly more likely than suburban or urban voters to say that they pray daily.

Rural and urban voters also differ on “gun culture.” In the CES, 61 percent of rural voters said there was at least one gun in their household. By contrast, only about 1 in 4 urban Americans own a gun. According to the Pew Research Center, gun ownership among Republicans (44 percent) is more than double the gun ownership rate among Democrats (20 percent). This partisan gap remains even after controlling for demographic factors such as gender, education and race.

Views on race relations and systemic racism also distinguish urban and rural America. Some social scientists measure racial attitudes using a scale called racism denial — that is, they ask respondents whether they agree or disagree that White people have certain advantages. Levels of racism denial have been shown to strongly predict vote choice in recent elections. A 2018 Pew Research Center report found that a majority of rural Americans deny that White people in the United States benefit from societal advantages that Black people lack. These differences between rural and urban opinions on racial advantages held even among non-White Americans. The 2020 CES also asked whether respondents agreed or disagreed that White people in the United States have certain advantages because of the color of their skin. Notably, more than 45 percent of rural Americans disagreed with this statement, while just 1 in 5 urban Americans denied that Whites have advantages.

Racism proves to be the most important predictor

So which of these factors, if any, is most closely linked to the 32-point gap in 2020 presidential vote choice between urban and rural voters? To answer this question, we calculated the urban-rural divide and used a regression model to control for each factor individually to see whether the urban-rural divide persisted after accounting for each explanation. Put simply, the smaller the remaining urban-rural divide after controlling for an individual factor, the more evidence we have that the factor explains that divide. The graph below summarizes our findings.

The first bar shows that rural Americans of all races were 32 points more likely to vote for Trump than urban voters. When we controlled for evangelicalism, the urban-rural divide dropped by just six percentage points. In other words, even after accounting for born-again Christianity’s prevalence in rural areas, a 26-point gap in vote choice between urban and rural Americans remains, suggesting that faith-related differences are relatively limited for understanding Republicans’ rural strength.

Accounting for gun ownership closes the urban-rural divide more than faith, reducing the gap to 21 points. However, most of the urban-rural divide remains unexplained even after accounting for higher rates of gun ownership in rural areas.

Racial attitudes among all Americans best explain the gap in vote choices between rural and urban areas. Controlling for racism denial, the gap in vote choice between rural and urban Americans drops to just eight percentage points. In other words, the different rates of racism denial among rural and urban Americans appears to explain about three-quarters of the urban-rural gap in voting for Trump.

To put our findings another way, if voters in urban and rural areas acknowledged White privilege at the same rate, the urban-rural voting divide would be relatively small, just eight points. That the divide is actually 32 points speaks to the powerful role that racism plays in fueling this gap. Perhaps it’s no surprise, then, that Virginia Republican candidate Glenn Youngkin attacked the teaching of critical race theory, making it a cornerstone of his campaign.

Our findings suggest that messaging isn’t the problem, as Carville asserted. Rather, rural Americans prefer Trump’s racially charged politics and denial that racism exists. Fueled by a core disagreement over racism in the United States, the urban-rural divide is likely to continue in 2021 and beyond.

Zachary L. Hertz (@zacharylhertz) is a master’s student studying quantitative methods and social analysis at the University of Chicago.

Lucas B. Pyle (@Lucasbenji17) graduated from Tufts University in May and majored in political science and religion.

Brian F. Schaffner (@b_schaffner) is the Newhouse Professor of Civic Studies at Tufts University and co-author of Hometown Inequality: Race, Class, and Representation in American Local Politics (Cambridge University Press, 2020).