Though Republican candidates almost everywhere registered large margins among white voters without a college degree, Democrats ran much more competitively among the roughly half of that group who are not evangelical Christians, according to previously unpublished results from the 2018 exit poll conducted by Edison Research for the National Election Pool, a consortium of media organizations including CNN.
2016 Presidential election
2020 Presidential election
Belief, religion and spirituality
Continents and regions
Education systems and institutions
Elections (by type)
Elections and campaigns
Employment and income status
Females (demographic group)
Government and public administration
Government organizations - US
Immigration, citizenship and displacement
International relations and national security
Labor and employment
Low income persons
Midwestern United States
Political Figures - US
Population and demographics
Social and economic status
Southeastern United States
US Democratic Party
US Federal elections
US House elections
US House of Representatives
US political parties
US Presidential elections
US Republican Party
US Senate elections
Voters and voting
Workers and professionals
Democrats, the analysis found, ran particularly well this year among white working-class women who are not evangelicals, a group that also displayed substantial disenchantment in the exit poll with Trump's performance. Those women could be a key constituency for Democrats in 2020 in pivotal Rust Belt states such as Michigan, Wisconsin and Pennsylvania, where relatively fewer blue-collar whites are also evangelical Christians.
Nationwide, nearly three-fifths of blue-collar white women who are not evangelicals voted Democratic in last month's House races, while an equal number said they disapproved of Trump's performance in office, the analysis of exit poll results found. That was well over double the Democratic share of the vote among non-college white women who are evangelical Christians. And while Republicans last month still carried a majority among working-class white men who are not evangelicals, Democrats attracted about twice as much support from them as they did among the equivalent men who are evangelicals.
"It's another overlay to the conclusion that there are some parts of the white non-college population that are open to Democrats and can be moved a few points in your direction," says Ruy Teixeira, a long-time Democratic analyst of voting trends who now serves as a senior fellow at the liberal Center for American Progress.
Though these distinctions sound like fodder for a cocktail hour argument at a political science faculty lounge, they actually inform a backstage debate simmering among Democratic strategists about 2020. This debate has clear implications for the message the party develops over the next two years and the kind of nominee it chooses against Trump in the next presidential election.
At issue is how much emphasis the party should place on trying to recapture white working-class voters as opposed to maximizing turnout among its new base of minority, millennial and college-educated white voters, especially women and those in urban areas.
While some Democrats have come to view white working-class voters as largely a lost cause for the party in the Trump era, other party strategists, including some affiliated with organized labor, have privately argued that the large number of staunchly conservative evangelical Christians in the group has overstated Democratic weakness among them.
Strategists in this camp argue it would be a mistake for the party to downplay outreach to white working-class voters who are not evangelicals, especially the women in that group.
This bubbling private debate has given rise to a new and improbable acronym that some Democrats see as a potentially pivotal group for 2020: WNCNEW, as in white non-college, non-evangelical women.
"WNCNEW is the group Democrats should care about," one Democratic strategist insisted in an email. Those women represented about one in nine voters nationally this year and an even larger share in key Rust Belt battlegrounds.
In practical terms, the party and its presidential nominee in 2020 inevitably will try to turn out its new base and to regain blue-collar voters of both genders from Trump: the decision facing the party is not either/or. But Democrats will face a genuine choice of emphasis.
Two camps of Democrats
Many of the party's potential 2020 contenders appear better suited to energizing its new base than recapturing working-class whites: Sens. Kamala Harris, Elizabeth Warren, Cory Booker and Texas Senate candidate Beto O'Rourke might all fit into that category. By contrast, former Vice President Joe Biden, Ohio Sen. Sherrod Brown and centrists such as former Virginia Gov. Terry McAuliffe and outgoing Colorado Gov. John Hickenlooper might be better positioned to reassure working-class white voters than to mobilize the base.
Similarly, the choice on how the party positions on racially tinged issues, such as immigration and police reform, will also likely be influenced by this debate. If Democrats believe they can recapture meaningful numbers of blue-collar whites from Trump they may hesitate about alienating them with vanguard liberal positions on social issues — such as abolishing ICE — in the hope of energizing younger and non-white voters.
Different ways to assess white voters
The detailed exit poll results provided to CNN show clear openings for Democrats with some groups of white working-class voters. But they also indicate that Democrats still face significant headwinds with most of those blue-collar whites and that white evangelical Christians look increasingly monolithic in their support for Republicans. That's a big challenge for Democrats in many southern states, where half or more of working-class whites are also evangelical Christians.
To illuminate these subtle but critical distinctions, Edison Research analyzed results from the 2018 exit poll in its national sample and in seven states where voters were asked whether they are evangelical Christians. (To identify evangelicals, the exit polls rely on a self-reported answer of "yes" to the question, "Would you describe yourself as a born-again or evangelical Christian?")
In the states, Edison diced the results twice. First it separated white voters into those with and without college degrees. Then it further divided each of those groups into those who are evangelical Christians and those who are not.
For the national results, which have a much larger sample of voters to assess, Edison at CNN's request divided the groups a third time, in this case along lines of gender.
To start, this analysis underscores how many white working-class voters are also evangelical Christians. Nationwide, the exit poll found that evangelical Christians this year comprised fully 45% of all white voters without a college degree, a substantial portion of the total electorate. By contrast, evangelicals represented only one-fourth of college-educated white voters. (In 2016, the exit polls found that evangelicals constituted slightly larger shares of each group.)
In all of the southern states where the question was asked this year, evangelicals represented a majority of working-class white voters, including fully two-thirds in Georgia and Tennessee. Evangelicals were also a majority of white working class voters in West Virginia and Indiana and exactly half in Missouri.
That's a big hill for Democrats in those states, because the exit poll results show that they face virtually monolithic opposition from all segments of the evangelical community.
Evangelical support for the GOP candidate
At least 77% of white evangelicals without a college degree voted against the Democratic Senate candidates in Florida, Missouri and Tennessee, while 72% opposed defeated Democratic incumbent Joe Donnelly in Indiana, the exit polls found. In the Georgia governor's race, a breathtaking 89% of non-college white evangelicals voted for Republican Brian Kemp over African-American Democrat Stacey Abrams; 84% of those voters picked Ted Cruz over O'Rourke in Texas. Only Democratic Sen. Joe Manchin in West Virginia ran competitively, losing those voters by a narrow 52% to 46% margin.
In House races, the exit polls found that exactly three-fourths of white evangelicals without a college degree voted Republican, while only about one-fifth supported Democrats. Democrats lost the men in this group by 57 percentage points and the women by a still daunting 49 points.
The picture wasn't appreciably better among white evangelicals with a college degree. Nationally, Republicans again won almost exactly three-fourths of them in the House races, with relatively small differences between the men and the women. Results on that group were not available in as many states, because they comprised a smaller share of the total vote, but Republicans carried about 70% of them in the Tennessee, Indiana and Missouri Senate races; in Georgia, 83% of college-educated white evangelicals voted for Kemp.
The collapse of any meaningful distinctions among evangelicals reflects both their hardening loyalty to the GOP and the contraction of their overall numbers, says Robert P. Jones, chief executive officer of the Public Religion Research Institute, a non partisan group that studies religion, values and politics.
In PRRI surveys, he notes, evangelical Christians have declined from about 21% of the total population in 2008 to 15% this year. That erosion, Jones says, has been "asymmetrical," with younger and better-educated members becoming the most likely to leave the faith. That's left behind a group that is older and more uniformly conservative.
"As this group is shrinking and aging it is just becoming more and more homogenous," says Jones, author of the 2017 book "The End of White Christian America." "When you have that kind of attrition, and it's coming all from the low [younger] end, and the low end is more likely to be college educated and likely to be more liberal on a whole range of cultural issues and less anti-immigrant ... you start losing differences."
A different story among non-evangelical white voters
But if the 2018 election results highlight the solidifying uniformity of white evangelicals — male and female, college-educated or not — the findings simultaneously illuminate pressing reasons for Republican concern about white voters who are not evangelicals, both those with and without college degrees.
Although Democrats this year posted their best recent showing among white voters holding at least a four-year college degree, the GOP's continuing strength among such well-educated white evangelicals obscured the full extent of the party's decline.
In the national House exit poll, Democrats carried fully two-thirds of college-educated whites that are not evangelical Christians. That included not only a head-turning 71% of college educated white women who are not evangelicals but also 59% of the equivalent men. The shares that said they disapproved of Trump's performance were even higher in both groups: 74% of the women and 63% of the men.
Democrats carried about three-fifths of these non-evangelical white-collar whites in the Florida Senate race and about two-thirds of them in Indiana, Tennessee and Missouri. Even in Texas, O'Rourke carried 61% of them, while Abrams won 55% of them in the Georgia governor's race. Those geographically dispersed results testify to the breadth of the recoil from the Trump-era GOP among these well-educated white voters that Republicans in an earlier generation considered part of their base.
In a process that Trump has accelerated, Republicans in this century have instead come to see whites without a college degree as the foundation of their support. But the detailed results from Edison show some cracks in that base among the working-class whites that are not evangelicals.
In the national House exit poll, Democrats actually carried a slim 52% to 46% majority among non-college whites who are not evangelicals, though with a significant gender gap. While Democrats won the blue-collar women who are not evangelicals by 16 percentage points, Republicans won the equivalent men by 9 points. And while 56% of those non-evangelical blue-collar men said they approved of Trump's performance in office, 58% of the equivalent women said they disapproved. Those are the so-called WNCNEW that some Democrats see as an essential target for 2020.
Religion is not the only explanation
In the state races, the pattern was mixed. The 2018 results found that Democrats lost working class whites that are not evangelicals by big margins in three Southern states — the Texas and Florida Senate contests and the Georgia governor's race. But Democrats ran close to even among them in the Senate races in Tennessee, Indiana and Missouri (where Claire McCaskill carried a slim majority of them). Manchin won these voters comfortably.
As those state results suggest, religion alone doesn't explain the modern Democratic difficulties with the working-class white voters who anchored their coalition in the decades after World War II. Democrats this year captured only about two-fifths or less of white voters without a college degree in Senate contests even in several states where white evangelicals, according to PRRI data, are a very small share of the total population — including Nevada, Arizona, New Jersey, Michigan and Pennsylvania.
PRRI surveys also show that among non-evangelical whites, those without degrees are consistently more conservative than those with degrees on polarizing cultural and racial issues. Results from PRRI's 2018 American Values Survey, for instance, found that among whites who are not evangelicals almost exactly twice as many of those without college degrees as those with advanced education said that discrimination against whites is as great a problem as discrimination against minorities. That placed those blue-collar whites much closer to the position of white evangelicals than to that of their fellow non-evangelicals who have college degrees.
Likewise, the working-class whites who are not evangelicals came down closer to the evangelicals on such other questions as whether to build a border wall or to ban immigration from Muslim-majority nations; and if the country will benefit as minorities become a majority of the population. On most questions, non-college white women who are not evangelicals were about as conservative as the men — and both groups usually landed well to the right of college-educated whites who are not evangelicals.
Overall, says Jones, on cultural and racial issues, the distance between working class whites who are evangelicals and those who are not amounts only to "a difference in degree but not in kind."
All of this suggests continuing limits on the potential inroads for Democrats among working-class white voters, evangelical or not, particularly with Trump appealing to so directly to their cultural priorities and resentments. But these results also point to genuine openings for some Democratic improvement in 2020 compared to 2016, at least in northern states where fewer of the blue-collar whites are evangelical Christians.
Although changes in survey methodology may partly explain the difference, the 2018 exit polls showed that among both working-class white men and women who are not evangelicals, Democratic House candidates won a measurably higher share of the vote than Hillary Clinton did in the 2016 presidential race. In the heavily blue-collar Rust Belt states that tipped the 2016 election to Trump — particularly Michigan, Wisconsin and Pennsylvania — even small improvements might be enough to tilt the result the other way. As Teixeira says, when trying to reverse Trump's narrow margin of victory in the decisive states, "it all counts."