How Does Education Level Impact Attitudes Among Voters of Color?

HG EDU Voters v1

After the 2016 election, all pundits wanted to talk about were white non-college voters. Democratic support had cratered with this group between Obama’s 2012 reelection and Hillary Clinton’s loss in 2016, and everyone wanted to know why.

But four years later, the 2020 election has illustrated that the story of Democrats’ shifting coalition is more complicated than that old narrative. Democratic support held constant with white non-college voters between 2016 and 2020, but their support with non-white non-college voters dipped from 81% to 75%. Seventy-five percent support is still high, but given the importance of Black, Hispanic, and Asian voters to the Democratic coalition, a 6-point drop-off has serious implications for Democrats both at the presidential level but also in down-ballot races with sizable non-white constituencies.

Since the November election, pundits and data-crunchers have debated exactly what this drop-off means, how significant it really is, and why it happened. One potential explanation is that there are growing cleavages among Black, Latino, and Asian voters by education. But it’s difficult to prove. Frustratingly, most polls do not have sufficient sample sizes to draw conclusions about non-white voters by education level. Many do not even have sample sizes big enough to break out non-white voters by their actual race or ethnicity, grouping them instead as “voters of color.” While the phrase “X group is not a monolith” is in vogue, most publicly-available data sources do treat Black, Hispanic, and Asian voters as such. Amy Walter highlighted this problem in a recent Cook Political Report article, recognizing that we have enough data to know that Democrats lost support with non-white non-college voters, but not enough data to have a sense of why or how Democrats might win them back.

To contend with this vacuum, Third Way is launching a new project to examine and understand educational divides among Black, Hispanic, and Asian voters on the most pressing political debates of our time. The data we do have reveals opinion divides on several key issues, but it also shows some areas of consensus among voters of color of every education level. For this analysis, we relied on the Cooperative Election Study (CES), a survey of 66,000 US adults that is conducted each year. The CES asks voters’ opinions on a range of important issues from abortion, guns, and policing to health care and social welfare policy. We have transferred this data into a more easily decipherable format and now are partnering with experts to sift through it all and uncover what these educational divides might mean for Democrats as they try to reverse negative trends with non-white non-college voters.

Looking at Black, Hispanic, Asian, and white voters each by education level offers a unique glimpse into the issues that divide American voters. Attitudes differ by race on many topics, but on some, education divides just as much. On abortion restrictions after twenty weeks, for example, Black and white non-college voters are seven percentage points apart, but Black non-college and college voters are separated by eleven. On most issues, non-college voters are more conservative than college graduates, but they are more progressive in a few key areas. And on a few issues, Americans are divided neither by race nor education but have seemingly reached consensus.

This report lays out the differences between Black, Hispanic, Asian, and white college and non-college voters on key issues. It highlights existing patterns, as well as areas where certain voter groups deviate from those patterns. And it emphasizes points of agreement across all groups: some just where one might expect, and some that may come as a surprise.  


The CES data provides a unique look at public opinion on social issues like abortion, policing, and immigration, as well as on economic issues like health care and social welfare policies. The overarching lesson from this analysis is that on many social issues, Black, Hispanic, and Asian non-college voters express views that are somewhat more conservative than their college-educated counterparts and fall somewhere between white college and white non-college voters. But on economic issues, results are a mixed bag; non-college voters of color are more conservative on some issues, fairly aligned with college voters of color on some issues, and more progressive on a select few.

Taken on their own, voters’ responses to some individual questions are surprising and counterintuitive. In some cases, majorities of groups agree with two or more conflicting statements. High shares of all groups, for example, agree with the statement that we should “always allow abortion as a matter of choice,” including a majority of white non-college voters. At the same time, large shares of all groups also agree with the statement that we should “permit abortion only in cases of rape, incest, or if the woman’s life is in danger.”

Rather than drawing a confident policy prescription from either of these two data points independently, we can conclude from these answers that voters are conflicted in their views. Digging into this data, we urge readers not to draw dramatic conclusions from one independent data point but rather to look at groups of questions on a topic holistically to see where voters land, and how the demographic groups might differ in their views.

Respondents in issue polls also tend to respond “yes” by default when asked binary “support” or “oppose” questions when they do not have a confident opinion. In this analysis, if an issue position has about 60% support, we will not refer to that as “overwhelming support” for the policy. Areas that we identify as consensus areas have over 70% support.

This study looks at responses to sets of questions on guns, policing, immigration, abortion, health care, and climate change. It also looks at responses to individual questions on raising the minimum wage, equal pay for equal work, and food stamps. Rather than analyzing the data by individual issue, we categorize the data by issues on which non-college voters of color are more conservative than college graduates, issues on which college and non-college voters of color are aligned, and issues on which non-college voters of color are more progressive. We also highlight issues with particularly high levels of support across all groups, where American voters agree with one another across both race and education lines.   

On many issues, education gaps are significant. This presents several challenges for Democrats. The fact that non-college voters of color are more conservative on a range of issues indicates that Democrats will struggle to win back working-class voters of all races without carefully navigating the policy and politics of those debates in a way that might cause tension with activists on social media. And voters’ tendency to agree with seemingly conflicting statements when asked questions in different ways is a stark reminder that many voters are not staunchly ideological and can be easily swayed away from progressive viewpoints when presented those ideas in different ways. Nonetheless, there are several areas where there are smaller education gaps, where non-college voters of color are more progressive, or where all voters are aligned. These are important to note if Democrats hope to win back more working-class voters in 2022, 2024, and beyond.

Party Identification Among Voters of Color by Education

The CES survey asks voters both their ideological alignment and their party identification, and for the most part, college graduates identify as more liberal and more Democratic than those without a college degree. This holds true across racial groups, though the gap is wider among white voters than others.

Meanwhile, non-college Black, Asian, and white voters are more likely to identify as conservative than their college-educated counterparts, though relatively small shares of non-college Black and Asian voters identify as conservative.

Hispanic voters stand out as one exception to this rule; college-educated Hispanics are slightly more likely than non-college Hispanics to identify as conservative. This means that Hispanic college graduates are both more likely to identify as liberal and more likely to identify as conservative compared to non-college Hispanics. A large share of non-college Hispanics identify as moderate, and a significant share (16%) do not identify ideologically at all. This suggests that college-educated Hispanics are quite divided in their opinions, with significant shares identifying as either liberal or conservative, and that non-college Hispanics are much less ideological and may not align precisely with either political party or movement.

When it comes to party identification, college graduates of all races and ethnicities are more likely to identify as Democrats. Majorities of Black, Hispanic, Asian, and white college voters all identify as Democrats. Notably, only 52% of non-college Hispanics identify as Democrats, and 49% of non-college Asian voters. Seven in ten (71%) of Black non-college voters identify as Democrats.

Educational Divides Among Voters of Color

It is no secret that college graduates are more progressive on a range of social issues compared to non-college graduates. Due in part to that fact, Democrats have seen great success in recent years with college-educated voters, including many white college graduates who have voted for Republicans in the past and hold more fiscally moderate views. The data bear that out; with a few individual exceptions, non-college voters of every racial group are more conservative on gun control, immigration, and abortion than their college-educated counterparts. Among white voters, education gaps are wide, often above 20 percentage points. Among non-white voters, gaps are narrower but nonetheless significant.

On guns, most voters support policies that would strengthen gun laws and oppose proposals to loosen them. But margins of support and opposition vary by both race and education. When asked if they support or oppose banning assault rifles, college grads of every racial group support a ban by wide margins. Support for banning assault rifles is about ten points lower for Black, Hispanic, Asian non-college voters than for their counterparts with a college degree. Still, majorities of all groups support such a ban.

Image Alt Text When it comes to immigration, both Black and white voters are divided by education level, although opinion gaps are much wider among white voters. When asked if they support increasing the number of border patrols on the US-Mexican border, 45% of Black college-educated voters are in support, compared to 49% of Black non-college voters. For white voters, the education gap on immigration is a chasm; 68% of white non-college voters support increasing the number of border patrols compared to 49% of white college voters.  

Interestingly, Hispanic college graduates express slightly more conservative views on immigration than Hispanic non-college voters. Over half (52%) of Hispanic college graduates support increasing the number of southern border patrols, compared to 49% of Hispanic non-college voters. When asked if they support building a border wall, among college graduates, Hispanic voters had the highest support at 32%, with white college graduates following about a point behind. Three-in-ten (29%) Hispanic non-college voters support building a border wall. Hispanic college graduates’ slightly higher rate of support for strict immigration laws could potentially relate to their distance from the immigration process itself. Working-class Hispanic voters may have closer ties to people who recently immigrated to the United States, and their policy views may reflect that closer proximity. Either way, Hispanic college graduates’ somewhat more conservative stance on immigration suggests that candidates and policymakers should take a nuanced approach to the immigration issue when communicating with Hispanic voters, especially those who have been in this country for generations.

When it comes to abortion and a woman’s right to choose, voters are highly divided by education and less so by race. When asked if abortion should be prohibited after twenty weeks, majorities of Black, Hispanic, and Asian non-college voters agree. Support for abortion prohibition after twenty weeks is about ten points lower among their college-educated counterparts. Notably, Asian and white college graduates are furthest to the left on this question, while non-college voters and Black and Hispanic college graduates are more supportive of abortion restrictions.

Image Alt Text When asked about abortion using different language, education gaps hold, but voters across all racial and educational groups are more supportive of the right to choose.

Areas of Agreement Among Voters of Color Across Education Levels

To test for differences on attitudes toward work and work requirements, we looked at the results of a question asking if voters support requiring able-bodied adults ages 18 to 49 who do not have dependents to have a job to receive food stamps. While there is a 17-point education gap among white voters on this question, the education gap among non-white voters is quite small.

Exceptions to the Rule: Areas Where Non-College Voters Are More Progressive

On a few select issues, non-college voters of color express more progressive views than their college graduate peers. Unsurprisingly, these issues are economic issues that would disproportionately benefit working-class people, including raising the minimum wage and lowering the eligibility age for Medicare.

Majorities of all voter groups support raising the minimum wage to $15/hour. Education gaps on this question are small, but Black, Hispanic, and Asian non-college voters are slightly more likely to favor raising the minimum wage than their college-educated counterparts. Only among white voters are college graduates slightly more liberal on this issue, though the gap is quite small and within the margin of error (1.5 percentage points).

Lowering the eligibility age of Medicare from 65 to 50 also resonates with non-college voters of color slightly more than with college graduates. The gap among Black and Hispanic voters is significant; 77% of Black non-college voters support lowering the eligibility age compared to 70% of Black college graduates. Among Hispanic voters, 73% of Hispanic non-college voters favor lowering the eligibility age, compared to 64% of Hispanic college graduates. Asian and white college graduates and non-college degree holders are fairly aligned on this issue.

One possible explanation for these gaps is that Black and Hispanic non-college voters might be older relative to Black and Hispanic college-educated voters and hence have more to gain from expanded Medicare eligibility.

Areas of Consensus Across Race and Education

Some Democratic proposals have overwhelming support across voters of every race and education level.

Lowering the cost of prescription drugs has remained popular in polling across racial and educational groups. Voters have also come a long way on supporting workplace gender equity. Over 90% of all voter groups support paying men and women equally for equal work.

And notably, Democrats have a real opportunity to deliver on police reform. While policing was a controversial flashpoint in the 2020 election, these data show that targeted police reform is an area of overwhelming consensus. Over 60% of all voter groups across race and education level support eliminating mandatory minimums for non-violent drug offenders, and over 85% of all groups support requiring police officers to wear body cameras. More than 80% support banning the use of chokeholds by police, and three-quarters support creating a national registry of police officers who have been investigated or disciplined for misconduct. Even ending qualified immunity gets consensus support, with over 74% of all groups in support. By contrast, proposals around decreasing or increasing the size of police forces are divisive across racial groups and educational attainment.


As the longtime party of the working class, the Democratic Party strives to represent working people of all races. But representing those interests requires listening and paying close attention to what these voters really want, rather than to the loudest voices in politics and on social media that claim to speak for them. The incredibly comprehensive CES dataset provides a useful first look into differences in perspective among voters of color by education level on some of the most important political issues of the day. It presents challenges, namely that non-college voters are not uniformly progressive on either social or economic issues. But it presents opportunity areas as well, as large swaths of American voters agree on a range of major progressive policy priorities. If Democrats want to woo back working-class voters of all races and stem losses with Black, Latino, and Asian non-college voters, there is ample opportunity for them to do so, provided Democrats are aware of the range of opinions among these voters and choose their priorities with those views in mind.


Get updates whenever new content is added. We’ll never share your email with anyone.