Union members have historically been more likely than nonunion members to vote for Democratic Party candidates. However, the differences between union and nonunion voters have changed over the past several presidential elections. Union member support for Democrats reached a recent high during the 2024 election after falling to a low point during Donald Trump’s first presidential election in 2016. In 2016, union members, compared with nonunion members, preferred the Democratic candidate by a 4.4 percentage point margin—a far smaller difference than in 2012, when union members preferred the Democratic Party candidate by a 13.7 percentage point margin.1 But since 2016, union member support for Democratic Party candidates has increased each election cycle, hitting an 8.2 percentage point margin in 2024.2
This trend of increasing union support for Democratic candidates has been driven largely by women: College-educated union women steadily increased support over the past two presidential elections. Working-class union women shifted sharply toward Democrats in 2024. As a result, union women, compared with nonunion women, preferred the Democratic candidate in 2024 by more than in 2012. Meanwhile, male union member support for Democratic Party candidates remains below 2012 levels, particularly for male union members without a college degree.3
![This twilight view from Arlington, Virginia looks east across the Potomac River toward the monuments on the National Mall in Washington, DC. Seen here are the Lincoln Memorial, Washington Monument, US Capitol Building and the Library of Congress.]()
This Center for American Progress Action Fund issue brief analyzes the union vote in 2024 compared with previous elections. Using results from the 2012–2024 Cooperative Election Study (CES)—which represents the largest academic poll of verified voters, allowing for detailed breakdowns of union voters’ preferences among demographic groups—the authors compared Democratic presidential candidates’ share of votes among union and nonunion voters, charting how the advantage that Democratic candidates have among union members has changed over time. In particular, the analysis yielded the following key findings:
- Union voters were 8.2 percentage points more likely to vote for Kamala Harris in 2024, compared with nonunion voters. This margin is nearly twice as large as the 2016 margin of 4.4 percentage points and slightly higher than the 7.7 percentage point margin Joe Biden won by in 2020.
- Higher support for Democrats among union members has been driven mainly by women.
- Working-class women union members supported Harris by a margin of 14.1 percentage points, compared with nonunion women voters—a slightly higher margin than among college-educated union women (13.7 percentage points).
- Male union members with a college degree supported Harris by a margin of 9.2 percentage points, compared with nonunion men with a college degree.
- Union men without a college degree were slightly more likely to support Trump than their nonunion counterparts (by 2.2 percentage points), after being slightly more likely to vote for Democrats in 2016 and 2020.
- Similar gender gaps occurred among Hispanic and white working-class union members. Hispanic and white working-class union women supported Harris by 11.4 and 12.4 percentage point margins, respectively, compared with similar nonunion voters. In comparison, Hispanic and white working-class union men favored Trump by 6.4 and 2.1 percentage point margins, respectively, compared with similar nonunion voters.
- Much of the decline in working-class male union support for Democrats in 2024 occurred among the youngest voters—those ages 18 to 29.
Ultimately, the votes of union members will shape future elections’ outcomes, as union members comprise 9.9 percent of workers nationwide and more than 11 percent in key battleground states such as Michigan and Pennsylvania; and recent presidential elections have been quite close.4 Politicians who can articulate an economic agenda that appeals to all union members—men and women, working-class and college-educated voters—will likely do well with this group.
Why union men and women vote differently
An analysis of the reasons behind the changing voting preferences of male and female union members is beyond the scope of this issue brief. Still, these changes suggest that union women may be more receptive to their union’s political messages than men, particularly noncollege men, because they are less “cross pressured” by their views on the candidates and other policy issues. In other words, women may face fewer issues or identity forces pushing them to support Republicans and thus are more persuaded by their union’s political messages. In contrast, men, particularly noncollege men, may face greater forces pushing them to support Republicans and thus may be less persuaded by their union’s political messages.
Though economic issues remain top of mind for many voters, a candidate’s attitudes toward unions and workers is just one of many factors a voter may consider when making their choice, and both men and women and college and noncollege voters differ in terms of the issues they rate as priorities.5 Voters are pressured by a range of priorities and facts about candidates and must decide how to balance these cross-pressures when voting; gender is one such factor, and research has found that cross-pressures can produce different voting outcomes by gender.6
Over the past several election cycles, Democratic candidates have expressed positions on policies that reflect high levels of support among female voters—such as abortion,7 an issue that women prioritized much more highly than men while also believing that Harris would better reflect their interests.8 In contrast, Trump has been perceived as stronger on issues that male voters, particularly men without a college degree, prioritized—such as guns,9 portraying strength, and paying attention to the concerns of men.10 Other candidate specifics, including Donald Trump’s personal appeal to men compared with women as well as the Democratic Party’s choice of female candidates in 2016 and 2024, may also affect the cross-pressures union men and women face. At the same time, older union men without college degrees may have more firmly developed political identities and thus be less influenced by these cross-pressures than younger union members.
Why union voters prefer Democrats
Support for Democratic candidates is generally higher among union voters,11 as academic literature shows that unions can inform voter preferences and increase voter turnout.12 Democratic candidates have typically been more supportive of pro-worker and, specifically, pro-union policies,13 although, in recent elections, some Republicans, including President Trump, have tried to court favor with union members. In turn, unions educate their members about candidate stances on economic issues and increase turnout for candidates who represent working-class interests.14
Empirical studies have confirmed that union members are more likely than nonunion members to vote and are typically more likely to favor Democrats, directly tying Democratic electoral performance to union strength.15 Not only do labor unions increase voter turnout, but counties with higher shares of union members have also been less likely to vote for Republican candidates in past presidential elections.16 One study found that “right to work” laws, which reduce union membership, shrunk county-level Democratic presidential vote shares in elections from 1980 to 2016 by 3.5 percentage points, compared with border counties with no right-to-work laws; these laws also lowered voter turnout by 2 percentage points.17
Additionally, union voters who support worker-oriented policies are more likely to vote for their union’s endorsed candidate,18 and union members prefer candidates they perceive as advocates for the working class.19 Finally, unions effectively elevate working-class voices into elected offices, as unions support members running for office, and states with more union voters favor more candidates from jobs with high unionization rates.20
Increasing support for Democrats among union members is driven by women
In 2024, union voters favored Kamala Harris over Donald Trump by a wide margin. Figure 1 shows results that use data from the 2024 Cooperative Election Study to compare union and nonunion vote shares for Trump and Harris. Harris won 59.2 percent of the union vote, compared with only 50.9 percent of the nonunion vote among voters who were employed full time. Although Harris narrowly won among all voters with full-time jobs, Trump’s margins among retired, part-time, and nonworking voters were enough to win more of the popular vote than Harris.21 This result is consistent with other exit polls such as the VoteCast survey conducted for The Associated Press and Fox News.22
Kamala Harris’ performance among union voters in 2024 was the strongest among any Democratic presidential candidate since 2016, part of a trend in rising support for Democrats among union members in the elections after Donald Trump’s first presidential run. Figure 2 shows the difference in vote shares for Democratic candidates between union and nonunion voters over the past four presidential elections, with Harris winning by an 8.2 percentage point margin among union voters compared with nonunion voters. This margin reflects Harris’ share of 59.2 percent of the union vote, compared with 50.9 percent of the nonunion vote, among voters who were employed full time. This was a slight improvement over Joe Biden’s 7.7 percentage point margin among union voters compared with nonunion voters and the largest margin since Barack Obama’s reelection in 2012. Exit polls confirm that Harris performed slightly better than Biden among union members.23
In recent years, midterm elections have also followed a general trend of more substantial support for Democratic candidates among union members. CES data on elections in the House of Representatives show that since 2018, Democratic support has increased among union voters compared with nonunion voters, reaching a 12.3 percentage point margin among union voters compared with nonunion voters for House Democratic candidates in 2022.24 Midterm elections tend to be influenced by lower turnout and voters’ reactivity to the current administration, so something similar to the bump among union voters in 2022 may well occur in 2026.25
Recent increases in union member support for Democratic presidential candidates have been driven primarily by union women boosting their support for Democrats. Women comprised 43 percent of union voters overall, 31.3 percent of working-class union voters, and 51.8 percent of college-educated union voters in 2024. Because women have made up similar proportions of the union vote since 2012, the changes in union support for Democrats result from improvements in Democratic support among union women who vote, not women making up a larger proportion of the union vote.
61%
Share of female working-class union members who voted for Harris in 2024, compared with 46.9 percent of working-class women not in a union
Figure 3 shows how Democrats’ electoral advantage among union women compared with nonunion women increased in 2024 for those with and without college degrees. Harris won working-class union women by a 14.1 percentage point margin compared with nonunion working-class women and won college-educated women by 13.7 percentage points compared with nonunion college-educated women—in both cases larger margins among these groups of union women than in 2012. These margins represent significant differences in overall voting outcomes: 61 percent of female working-class union members voted for Harris in 2024, compared with 46.9 percent of working-class women not in a union.
In 2024, college-educated union men supported Democrats significantly more than nonunion college-educated men, while working-class men voted similarly regardless of union membership. Men made up 57 percent of the total union vote in 2024 and comprised 68.7 percent of working-class union voters and 48.2 of college-educated union voters. As shown in Figure 4, college-educated male union voters maintained high support for Democrats, compared with nonunion college-educated men, in 2024 and other recent elections. In contrast, working-class union men were slightly more likely to vote for Trump in 2024 compared with nonunion working-class men, by 2.2 percentage points, after several elections of slight preference for Democratic candidates.
Male working-class union members began voting like their nonunion counterparts starting in the 2016 presidential election, although, in 2020, Biden made some gains with this group of voters, which held in the 2022 midterm election. As a result, it is possible the differences in voting outcomes between college-educated and working-class union men in 2024 may be driven by the unique appeal of Trump as a candidate among working-class male voters, though it is also possible that working-class men have developed a closer affinity for the Republican Party as a whole.
Union support for Democrats among working-class women has grown similarly among white and Hispanic voters. Working-class women in unions support Democrats by large margins, compared with nonunion women without four-year college degrees, especially in 2024. These margins are very similar among white and Hispanic working-class women, as shown in Figure 5. Among Black working-class women, differences between union and nonunion are much smaller, as Black working-class women voted for Harris by a wide margin of 90 percent—compared with only 10 percent for Trump. Yet the differences between union and nonunion Black working-class women were slightly higher in 2016 and 2020 compared with 2024.
In recent elections, both white and Hispanic working-class union men voted similarly to their nonunion counterparts in terms of their favorability for Democrats. White working-class men have voted very similarly to nonunion white working-class men since 2016, as shown in Figure 5. While the votes of Hispanic working-class men have varied more over time, Hispanic union and nonunion working-class voters also converged in 2024. Union Hispanic working-class men favored Trump in 2016, compared with nonunion Hispanic men; preferred Biden by an 11.9 percentage point margin in 2020; and returned to Trump in 2024, favoring him by 6 percentage points. Among Black working-class men, the differences between union and nonunion voters from year to year followed similar trends as those of white working-class men over time. However, there was a greater difference between union and nonunion voters.
Whether or not working-class union men vote differently from nonunion men varies considerably with age. Younger working-class union members, those ages 18 to 29, have been less likely than their nonunion counterparts to vote for Democrats in most elections since 2012 and moved sharply toward Trump in 2024. And male working-class union members ages 30 to 49 have voted similarly to their nonunion peers since 2016, with a slight preference for Republicans. In contrast, the oldest working-class male union members, those ages 50 to 64, favored Democrats in every election since 2016, as shown in Figure 7. These differences reveal that, even among working-class men, union membership plays a varying role in a worker’s voting choices depending on their generational cohort.
Conclusion
This analysis shows that union support for Democrats has grown significantly since 2016 due to increasing support among union women and varying levels of support among college-educated and working-class union men. Female union member support for Democrats in 2024 was above 2012 levels, while support for Democrats among noncollege union men has fallen sharply since 2012.
More research is needed to understand why voting patterns have shifted. Although Donald Trump has run as the Republican presidential candidate in every election since 2016, the Democratic candidates have varied in terms of their identity, campaign strategy, and record as elected officials. Further research may be able to determine what roles policy issues, the state of the economy, candidate identity, and messaging have played in influencing voting outcomes among union members.
Unions educate their members about economic policy and drive turnout for candidates who represent the interests of working people. Historically, this has increased union support for Democratic candidates. Although the advantage enjoyed by Democrats diminished in 2016 during Donald Trump’s first presidential run, in every presidential election since then, support for Democratic candidates has increased among union voters compared with nonunion voters—a trend driven primarily by women. Unions remain a critical means for improving the popularity of pro-worker candidates, but pro-worker candidates should consider how to address the growing divide among working-class and college-educated voters, even among union members.
Methodology
This analysis relies on the 2012–2024 samples of the Cooperative Election Study, formerly the Cooperative Congressional Election Study. The survey, conducted just before and after voting for elections, offers demographic and voting data on voters, cross-checked against records of verified voters to produce a representative sample of voters for each election. For 2024, the analysis uses a preliminary release of the dataset and only includes voters who reported voting in the 2024 election; the authors restricted the analysis to employed voters to ensure the union and nonunion populations are as comparable as possible.
Results are reported in this brief as the difference in the vote share won by the Democratic candidate between union and nonunion voters. The vote share won by the Democratic candidate is calculated including votes for third-party candidates. During midterm elections, results are reported as the difference in the proportion of total votes for a candidate for the House of Representatives compared with the proportion of votes for the Democratic candidate among nonunion voters. Hispanic workers are defined as workers who identified as being “of Spanish, Latino, or Hispanic origin or descent,” and to avoid double-counting respondents, white, Black, and “other or multiple race or ethnicity” voters only include non-Hispanic respondents. Working-class voters are all voters without four-year college degrees.