Republican victories in Virginia and New Jersey last week are an ominous reminder of what is at stake for Democrats. Glenn Youngkin’s victory over Terry McAuliffe in the Virginia gubernatorial race is a clear indication that Joe Biden’s victory, particularly in such hotly contested purple states, was in large part a reflection of the fact that Donald Trump underperformed compared to more generic, establishment Republicans. Perhaps the most staggering difference was among non-college educated voters—who Youngkin won by a significant margin.

No doubt, making sense of these results in the coming weeks will hark back to the battle between moderates and progressives that followed the 2020 election. Moderates, especially those from “purple” districts like Pennsylvania Representative Connor Lamb, who narrowly held on to his seat, blamed progressive members for filling the Democratic Party agenda with “unpopular” and “unrealistic” issues like Medicare for All and defunding the police. Progressives like Alexandria Ocasio-Cortez fired back that members of the Progressive Caucus had no trouble winning their elections and blamed moderate defeats on ineffective digital marketing strategies and other tactical errors.

But the reality is that both sides are talking past each other. By lumping together issues like Medicare for All (a relatively popular proposal) with defunding the police (a wildly unpopular one) into one “progressive agenda,” moderates are failing to recognize the very different levels of support enjoyed by different planks in the progressive agenda—particularly among the working class.

Incoming New York mayor Eric Adams, for example, won both a primary and a general election campaign on an anti-crime platform in arguably the most progressive city in the country. On the other hand, the left-wing rebuttal is significantly weakened by the fact that members of the Congressional Progressive Caucus, on the whole, represent very safe Democratic districts. This point is driven home by the election results for Democratic Socialists of America–endorsed races in the recent election. But there simply aren’t enough safe Democratic districts to build a viable progressive congressional coalition.

On top of that, only five members of the Congressional Progressive Caucus represent competitive working-class districts, including majority-white and nonwhite districts. This is indicative of a broader problem across the party: Democrats have been slowly hemorrhaging working-class voters, the backbone of the party since the New Deal. The shift is most pronounced among non-college-educated voters, who are leaving the Democrats for the Republican Party.

Biden gained a little ground with non-college-educated whites in 2020, but Trump and Republicans still won that group by 26 points nationally. These trends are most acute among white voters but in recent years have become highly visible among working-class voters of color, too. In 2020, for example, Trump and down-ballot Republicans both made gains among non-college-educated Black and Latino voters. Similarly, in last week’s election in Virginia, Latino voters’ support for Democratic candidates was lower than expected.

Changes in national partisan trends—and particularly the erosion of the working-class base of the Democratic Party—are critical questions that must be reckoned with if the Democratic Party is to stand a fighting chance in 2022, 2024 and beyond. At their core, these debates reflect key presumptions about the kinds of policies and platforms that will appeal to new tranches of voters that Democrats so desperately need to attract.

But how can Democrats attract new working-class voters, particularly those in competitive districts? Do working-class voters actually support the kind of egalitarian economic agenda championed by candidates like Bernie Sanders? The answers to these questions are critical ones not only for Democrats electoral futures but also for building a strong progressive coalition in Congress.

A recent study by Jacobin magazine, YouGov, and the newly formed Center for Working-Class Politics, “Commonsense Solidarity,” takes a novel approach to exploring how progressives might attract more working-class voters. The report, which I coauthored, looks at the policy and candidate preferences of 2,000 working-class voters in five key swing states: Nevada, Michigan, Wisconsin, Pennsylvania, and North Carolina. This is a much larger sample of this demographic than appears in most polls and gave us the ability to focus on working-class voters in greater depth.

The challenge with these types of surveys and experiments is that it is difficult to get them to really simulate realistic voter decisions. To account for this, we asked working-class voters to respond to thousands of head-to-head electoral matchups between candidates with a range of different characteristics. By asking voters to simulate their behavior at the ballot box—selecting one candidate, and one set of attributes, messages, and priorities, over another—we were better able to isolate the factors that attract or repel working-class voters.

In addition to providing candidates’ full agendas, we also highlighted their “day-one priorities,” as a way of signaling the most important issues for a particular candidate. We also looked at what kinds of candidates were most popular, across different demographics and occupational backgrounds. Social scientists have long debated how best to study and measure social class—often opting for singular measures like education level or job type. But these singular measures give short shrift to class positions, which are inherently complex and multifaceted. Our study provides a new, multifaceted way to look at these.

Because we wanted to test the appeal of progressive candidates among working-class voters, we trained our focus on the subset of voters reasonably within reach of Democratic campaigns: that is, Democrats and Democratic-leaners, independents, and Republican-leaning voters. We excluded strongly partisan Republicans, leaving a sample that comprises about 70 percent of the working-class electorate. We call this very large group “potentially Democratic working-class voters.”

Our primary and most important finding is that there is strong support among working-class voters, including blue collar workers, for candidates who promise to deliver on universal bread-and-butter programs like Medicare for All or a federal jobs guarantee. And they prefer these candidates over those championing more moderate alternatives like “empowering small business” and “increasing access to affordable health care.”

We also investigated how voters responded to different types of rhetoric by crafting candidate soundbites drawn from different real-world candidates. In short, we were most interested in comparing how progressive messaging from candidates like Sanders and Alexandria Ocasio-Cortez fared against more moderate messaging from candidates like Joe Biden and Kamala Harris. But, within these ideological categories, we also wanted to compare how activist-inspired appeals fared against more universal, plainspoken language. Again, we found that universal, plainspoken rhetoric resonated more with working-class voters than the activist-inspired messaging.

While our respondents preferred candidates with a central focus on universal, bread-and-butter issues, we found little evidence that racial resentment was driving these preferences. In fact, potentially Democratic working-class voters were strong supporters of candidates who promised to end “systemic racism,” favoring them over rivals with a more general commitment to “equal rights for all.” To underscore this point, Black female candidates were far and away the most popular candidates among our sample (including among white working-class respondents). These results complicate the popular theory that the primary driver of white working-class electoral behavior in recent elections is a latent or resurgent racism.

The structure of our survey also allowed us to generate some interesting information about the popularity of particular kinds of candidates among the working class. Overall, working-class voters preferred working-class candidates over other types, particularly over lawyers and CEOs (who can blame them?). But we also got some interesting insight into some Democratic presumptions. For example, a mainstream moderate military veteran—the kind of Democratic candidate often celebrated by party leaders and the press—received the support of just 51 percent of survey participants. However, a progressive populist teacher, on the other hand, earned over 65 percent.

Perhaps the most important takeaway from the “Commonsense Solidarity” study is that working-class Americans who don’t vote are not automatic progressives, as politicians like Sanders often claim. We found little evidence to support the idea that these nonvoters are simply sitting on the sidelines and waiting for the right candidate or the right message to come along. Instead, we found that working-class nonvoters look a lot like other working-class voters. We found very little evidence that the lack of visible, progressive candidates is the obstacle.

This is an essential message for progressives to take to heart and one that will likely offer little comfort to those reeling from the recent election results. Expanding the Democratic base and building a stronger progressive majority in Congress is possible—but it will require a massive organizing effort that may not bear fruit for years, possibly decades, to come.