How Can the Left Win in Working-Class America?

A recent study suggests that with effective messages and candidates, the left can make gains where it is currently weak.

Last November, Jacobin, YouGov, and the recently established Center for Working Class Politics published the results of an experimental study on working-class political views in several US states. The study, called Common Sense Solidarity: How a working-class coalition can be built and maintained, was the subject of much discussion and debate in the wake of its publication. Here, two of the report’s authors, Jared Abbott and Leanne Fan, speak with Socialist Forum about the report’s research methodology and main findings, responses to and criticisms of the report, and what they see as its implications for democratic socialist and progressive political strategy in the US. This transcript has been edited for length and clarity.

Socialist Forum: What was the motivation behind the study?

Jared Abbott: The people involved in the study were Bernie Sanders supporters who thought it made sense for us to try to get a handle on how progressives might be more successful in electoral campaigns going forward. The purpose of this was not just about helping DSA or socialists, it was about the progressive movement in general. So some of the study’s findings are quite relevant to people in DSA and some are less relevant, because DSA plays a distinct role in US politics compared to the broader progressive movement. That was the basic idea, and we came up with a plan for trying to figure out what types of progressive candidates might be most appealing to voters that progressives have the hardest time reaching. Those tend to be, as we all know, low-education working-class voters of all races, particularly whites but also Latinos and to a lesser extent African Americans and Asian Americans.

The strategy we used involves creating hypothetical candidates that vary across a bunch of different dimensions in terms of demographics and messaging. We went into the field to talk with working class voters in Nevada, Michigan, Wisconsin, Pennsylvania, and North Carolina to see which of the various types of hypothetical progressive candidates we came up with would do the best. The idea was to take that information to hopefully generate conversations about the political efficacy of different kinds of progressive campaigning.

Leanne Fan: The Bernie 2020 campaign was certainly the biggest inspiration. There was definitely a lot of accepted wisdom floating around among Bernie supporters, particularly the idea that a strongly progressive platform would draw out non-voters and that a bold, progressive platform is not a campaign liability, but is in fact more likely than an establishment platform to beat Trump. One of the things I was really interested in testing was whether Bernie’s style of messaging, which effectively articulates the “99% versus 1%” ideology, while using simple language, is resonant with voters.

Another inspiration was my experience helping to manage an progressive Democratic primary campaign for a Massachusetts state legislature seat. State House transparency rose to the top of the Massachusetts post-Bernie movement’s agenda as the tactic to “tenderize” the Democratic incumbents and unify a cohort of progressive challengers amid the Movement for Black Lives and the devastating economic impacts of the pandemic. The difficulty of explaining this issue to voters left me with an intuition about potential disconnects between some progressive issue priorities and what actually motivates people to vote.

In the most non-technical, jargon-free way possible, briefly run us through the basics of the methodology you used to do the study.

Jared Abbott: The survey is giving each survey-taker a choice between two hypothetical head-to-head candidate profiles. Instead of asking every respondent to choose between the same two candidates, we randomly generated tens of thousands of candidates based on nine candidate characteristics. The advantage of this kind of design is that you are able to isolate each candidate characteristic by respondent preference. We could then assess the relative popularity of candidate characteristics within the same category (what kind of candidate soundbite is most popular) amongst survey respondents, which we also further divided (by rural/urban, race, etc.).

How did you find the survey respondents?

Jared Abbott: We worked with YouGov, which is a well known polling firm, and we decided that we wanted to look primarily at the types of voters that Democrats in general and progressives in particular struggle with the most. According to the data, those are basically low-education voters who are generally working-class. Additionally, we wanted to talk to these kinds of voters in places with tight electoral margins, so we picked voters from five swing states mentioned earlier.

The last thing we wanted to be able to do was compare urban and metropolitan-area voters with rural and small town voters. We sampled more small town and rural voters than is ordinarily the case so we could draw some conclusions about different types of working-class voters in different geographic locations.

Many people seemed to have questions or criticisms about the definition of “working class” that you used for the study, but educational attainment wasn’t the only metric you used to define it. Tell us a bit about why you used the approach you did.

Jared Abbott: We don’t think or want to say that a lower level of educational attainment is the only way to define the working class. We make that very clear in the report.  But basically, we were interested in understanding how progressives and Democrats can reach the voters that are hardest for them to reach, and that tends to be lower education working-class voters.

We understand, of course, that many people with college educations, millions of them in fact, are also working class by Marxist or more sophisticated sociological or political science definitions. We’re not in any way trying to claim that those folks shouldn’t be considered part of the working class. But for practical reasons, we were most interested in the lower education voters, who are often some of the most difficult voters for progressives to reach, and that’s why we focused on them in this study.

Leanne Fan: What we ended up doing was trying to articulate more clearly what are the subgroups within the category of non-college educated voters. In the study, we also asked about the skill level of work, level of supervision at work, manual vs. mental responsibilities, self-identification, income, and education level of parents. The first three respondent characteristics provided an opportunity to build a richer “working class” variable, based on the work of Erik Olin Wright. Those respondents who do routine work, with less autonomy at work, and perform mostly manual tasks are those most commonly understood as working class, while those who have high degrees of autonomy, skills, and perform mostly mental tasks are understood as professional or middle class.

What were some of the payoffs from using these kinds of approaches?

Jared Abbott: The practical payoff is that we can speak with much more precision about how very specific characteristics of political candidates may be viewed by voters. If we were to look at all the progressive candidates that have actually run in real elections and then do a statistical analysis of how well they did, we wouldn’t necessarily learn as much. First of all, there aren’t enough of them to run a statistical analysis that would allow us to look at a lot of these characteristics, although we are building a database of progressive candidates to try to do something like that using real-world data.

Even then, there are so many other factors that could affect why a candidate did well in a given area and less well in a different area. Our approach allowed us to bracket that and do an experimental test that holds everything else constant to examine the impact of different demographic characteristics and messaging styles on working-class voters’ opinions.

The payoff of using the definition of “working class” we used is twofold. The first one is that it corresponds to the universe of voters and potential voters that we were interested in studying. Because of the territorially-based nature of our country’s political system, progressives and socialists need to get more support in red and purple states where we currently don’t run many candidates, and we wanted to understand how progressives can get more support in these areas. We don’t have to win over everybody, but we need to get more support here if we want to have any hope of being more successful in state legislatures and in Congress.

The second one is that, like I said before, if we included a broader universe of voters we simply wouldn’t have the same statistical power. That is to say, we wouldn’t have enough people in the survey to allow us to really drill down and make comparisons across a bunch of different dimensions.  In future studies, we will definitely look at different aspects of the working class, but these are the pragmatic reasons why we took this approach in this particular study.

What do you think are some of the downsides of the methodology you used?

Jared Abbott: The main downside of the methodology is that it’s not very realistic. We showed people written-out descriptions of hypothetical candidates. Now, this particular methodology has been tested by political scientists in different contexts to see how it stands up relative to real world conditions, and in general it does pretty well. That said, we did not expose respondents to real candidates in a real election. So it’s very important to keep in mind that this is one exploratory study that offers some unique and interesting insights, but it has to be verified against real world data. Ours is only one way of looking at these questions.

So that is an important limitation, which we acknowledge and don’t want to downplay. One downside of the class definition we used is that we were not able to compare higher-education working class voters to lower-education ones. That is a limitation we acknowledge and for practical reasons couldn’t be avoided, but we’re going to be looking at some of those differences in future studies.

The last thing I’d say is that we looked at different sorts of messaging and issue priorities. We basically had three different types of issue bundles we assigned to each candidate, and to create those we had to make choices about which of the million things we could have included. Because of that, there were certain things we didn’t test, everything from climate change to foreign policy. In terms of messaging, we used actual language from real candidates, but of course, if you subtly change the language that has an impact on results.

We are aware of the fact that the results are sensitive to the messaging we choose, as is the case with all surveys. That’s really important to acknowledge. Of course, we also think that the benefits largely outweigh the cost because we got some important insights we couldn’t get otherwise. But there are major limitations, like there is with any survey of this type.

Leanne Fan: Another limitation was that we were eager to get a lot of information, so we ended up packing a lot into the survey. So some of the sample sizes of each type of working class voter, for example, became really small. With small sample sizes, it was more challenging to come to really firm conclusions about candidate preference across different groups of respondents.

Now that we’ve addressed motivation and methodology, let’s talk about the findings. What are some of the main findings or takeaways from the study that you think are most relevant and interesting to progressives in general, and to DSA members in particular?

Jared Abbott: I think there are two or three headlines of particular relevance to DSA. The first is that socialists and progressives should try to compete not just in their traditional strongholds in places like New York and California and Massachusetts, but also in less hospitable territory in red and purple states. They can effectively use bold progressive messaging in those areas to attract working-class swing voters of different kinds in a way that frankly surprised me. The Bernie Sanders-style messaging we tested was very popular among some of the hardest to get working-class voters, like swing voters and independent voters who tend to be more conservative than Democrats. Even pretty bold racial justice messaging, things like “end systemic racism,” was pretty popular among working-class white respondents. It wasn’t a liability for a candidate to include things like that in their messaging style when it came to appealing to white working-class respondents.

There are many hopeful findings in the study from the perspective of how progressives and socialists can get beyond the ceiling that I fear they’re approaching in traditional strongholds. They should, of course, keep winning in those places, but they also need to go beyond that. The study offers some hope, I think, that we can do better and that we should be more bold and aggressive in trying to court working class voters that we’re currently not getting. DSA, frankly, doesn’t have much of a base among non-college educated working-class voters, but it could in the future.

The second thing I would say is that we need to have more working-class candidates. Ours is not the first survey to indicate this, but we found very decisively that candidates who were construction workers or teachers were much more popular than candidates with middle-class or upper-middle-class backgrounds. So it really is important for progressives and socialists to try to have candidates who came organically out of the working class.

We also have to start wrestling with the major strategic dilemmas we face right now. As Leanne said before, many of us thought that the Sanders campaign would dramatically pull in the non-voting electorate. That’s what Bernie said he was going to win on, and that didn’t happen. We also found that being associated with the Democratic Party did not hurt candidates in our study. I know a lot of people in DSA feel like working class people just hate Democrats because so many of them are awful neoliberals. But our study found that working-class people did not seem to care whether a candidate distanced themselves from the Democratic Party, what mattered most was the demographics and the message. If we want to figure out how to mobilize populist sentiment more effectively, we need to be more careful about our assumptions. Just because the Democratic Party is failing the working class in an objective sense doesn’t necessarily mean that an electoral approach grounded in distancing ourselves from the Democratic Party will be effective among working-class voters.

In some sense, our findings suggest that if socialists and progressives want to be more effective, they should probably pursue slightly more pragmatic approaches in certain geographical areas. On the other hand, the study suggests that there’s a much bigger picture outside of the electoral arena that we need to confront. I’m talking about the slow boring of hard boards over time to build a working-class constituency that could be open to the kinds of stuff we’re talking about. I think that we’re at a sort of strategic crossroads in terms of pragmatism versus pushing the envelope in different dimensions.

Leanne Fan: What we’re really asking is, how do we elect people who will pass the legislation that is overwhelmingly supported by Americans? The answer to that question is not just in the pockets of blue where a substantial number of DSA members can push a socialist challenger to beat an establishment Democrat. What can DSA chapters in electorally competitive districts be supported to push for candidates who could win us our political agenda?

Jared, you nodded toward this tradeoff between pushing the envelope and pragmatism. What are the areas where you think we can push the envelope, and where should we be more pragmatic? What exactly do you mean by pragmatic?

Jared Abbott: By “pragmatic,” I don’t mean trying to come to some sort of mushy centrist position that is supposedly more appealing to swing voters, or changing your political orientation to appeal to swing voters. We find you don’t really need to do that. Like I said before, we found that highlighting ending systemic racism is actually not much of a liability among working-class swing voters. It was relatively popular among basically everybody in the survey, although less so among white voters. The same goes for economic policy. We asked people about a job guarantee, and while it wasn’t super popular, maybe because people just aren’t familiar with what that even is, it was not unpopular. It always got net favorability among all candidates in the survey.

We basically found that you can go out there and say, “I’m for a job guarantee and ending systemic racism” and still be competitive among low-education working-class voters, which is a great thing. We also found that Medicare for All was generally quite popular. For me, pragmatism has more to do with the style of messaging, not the content or the issues themselves. It involves trying to avoid language that might not be palatable or just not very recognizable to voters who aren’t already progressive or socialist activists. That’s the only place in the survey where I think we find pragmatism, in this sense, being important.

The broader question concerns what the survey might imply about socialist strategy more broadly. Should socialists and progressives focus on trying to build as big a progressive coalition they possibly can by using all the tools they can to appeal to the broadest section of the working class? In that sense, our survey offers a lot of interesting insights that point to a more pragmatic approach. Alternatively, should socialists focus on building an explicitly socialist electoral base pushing the envelope of US politics? The answer depends on what you think is more important given the current political context.

Leanne Fan: Maybe it is as simple as saying progressives shouldn’t be afraid to run in competitive districts and to run on these issues. I think it could show you what has campaign liability relative to the other characteristics we chose.

Jared Abbott: I think a lot of this would sound roughly like the way Bernie Sanders speaks. Bernie is strongly committed to all of the different fights that we as socialists are committed to, but uses very clear, simple, and plain language and uses a very consistent message grounded in an “us versus them,” working class versus elites kind of framework. That’s what we’re talking about. It’s a matter of messaging style, as well as the relative emphasis that you put on different things in different contexts. I think we have to be aware that depending on the context, if you really want to reach a voter who might be a little bit more hesitant than you about abortion rights or the need to reform the police, you might want to lead with something that’s more related to jobs or economic justice, and then try bring in the other issues you care about. In that context, you’re trying to use a messaging style that disarms people and builds the strongest bridge you possibly can between what you’re saying and what they’re saying.

I don’t think this is any different than what many progressive organizers working in difficult contexts would do. I’m not sure who would want to quibble with the idea that messaging in this kind of way matters, and that’s basically all we’re trying to say.

Leanne Fan: In the experiences I’ve had with abolitionist candidates, they don’t knock on a door and lead with, “Hi, I’m your neighbor. I want to defund the police.” It’s more like, “I want to talk about community safety and what community safety actually means.” I don’t think it’s really controversial to say that how we craft our messaging matters and that we have to be attentive to the material needs of voters. What this report shows us is that in the narrow realm of electoral politics, we, as people who want to see universal programs in this country, can build our majoritarian coalition without compromising our values and vision of racial justice but we need to be vigilant about how it is accomplished.

Could we say then that for you it’s not a matter of whether we should be talking about racial justice, or abortion rights, or trans rights, or adopting more conservative positions on these questions for the sake of electoral gain, but a matter of how we talk about them?

Jared Abbott: Yes, 100%.

Educational attainment has an age gradient, which is to say that younger people, on average, have higher levels of educational attainment than older people. In that sense, might it be wiser from a long-term perspective to not worry about being pragmatic when it comes to older, particularly white voters in order to shore up the strongest possible base among younger voters?

Jared Abbott: It might be. The difficulty is that younger voters are by and large not the voters of today. This is one of those areas where progressive strategy in general might differ from DSA or socialist strategy in particular. Given the potentially existential threat to American democracy we’re currently facing, and give the very high political stakes of the moment we’re living in, any reasonable approach for progressives broadly speaking has to focus on making sure Donald Trump doesn’t become president again, and making inroads into state legislatures and congressional districts where we’re currently weak. There are just not enough young people we can mobilize to do this right now. There might be in the long term, but not now.

It’s a really difficult question, but I think that at the end of the day, the answer is that we have to do both. We have to mobilize historic numbers of young voters while also making meaningful inroads with low-education voters. But by doing both, this means that we really have to take the need to appeal to a broader section of the working class very seriously. I don’t mean just white, low-education voters. Biden’s support among Latino and African American voters went down quite a bit compared to Hilary Clinton. That’s a really serious problem because those are the linchpins of the Democratic and progressive base.

That’s the progressive picture in general, but I think DSA has a different kind of role. DSA is about trying to build a coherent socialist or social democratic movement in the long term that could be a genuine vehicle for working-class political power in the United States. It makes sense for DSA to focus on developing the consciousness of younger voters and building the constituency that hopefully can reshape politics over the medium to long term. But I don’t think that sort of logic should dominate progressive strategy in general given the extraordinarily high stakes we face right now.

It sounds like you’re arguing for a division of labor on the left, where some take on a certain approach in certain contexts and others take on a different approach in different contexts.

Jared Abbott: Exactly.

One of the things that comes through very clearly from the study is that working-class voters tend to respond very positively to working-class candidates. But as we all know, genuinely working-class candidates are few and far between in American politics. Why do you think that’s the case, and what can be done to increase the number of working-class people in politics?

Leanne Fan: Generally speaking, it is much harder for working-class people to feel like they have the social network to be able to run a successful campaign. In this respect, I would like to learn more about how unions have cultivated their members to run as candidates. That is hopefully a path forward here. There are a lot of structural barriers standing in the way of working-class people who want to run for office. Put simply, asking for money is much easier to do if you’re rich and you have a phone full of rich contacts.

What other issues or questions do you plan to address in future research?

Jared Abbott: We have a few things in the hopper. First, we want to try to understand the effects of Republican messaging relative to progressive messaging. That’s something that we weren’t able to get in this particular study. One of the projects that we’re working on next, which is also going to be a survey-based experiment, will involve showing working-class voters different sorts of Progressive messaging and then following up with a hypothetical Republican response. We want to see not only how well these sorts of progressive messages work in the abstract, but how well they work relative to conservative messaging.

Another thing we’re working on is understanding differences in the degree to which different parts of the working class care about different kinds of issues. This will also include more highly-educated working-class respondents. And we’re working on a large database of real-world progressive candidates. We’re collecting demographic and messaging data about all kinds of different candidates so we can look empirically at real candidates and real results to see how what we found in the first study might play out in the real world.

If any DSA members have ideas about things they think we should study, please let us know. We can’t guarantee that we can do everything, but we’d love to hear and maybe try to incorporate some of those ideas.

Leanne Fan: Broadly speaking, we want to understand better how certain issues can be most effectively incorporated into a successful election campaign. One example is understanding better how candidates can raise people’s expectations of what we deserve in this country, another example would be how to more effectively integrate demands for things like racial justice with universalistic economic demands. Another topic I’m interested in is understanding voter appetite for a redistributive agenda. I think the projects that we’re interested in will take up those sorts of themes.