Harris is starting to rebuild the Biden coalition

It has now been just over four weeks since President Joe Biden dropped out of the presidential race and Vice President Kamala Harris became the de facto nominee of the Democratic Party. Over that time, Harris and her campaign have taken steps to reintroduce her to the American people, reimagining Biden's campaign to match her different personality and appeal to new voters.

That has led to a lot of speculation about what the "Harris coalition" could look like in November — and how it could differ from Biden's. To find out, we aggregated crosstabs for several key demographic groups from polls conducted since Biden dropped out of the race. Now, this data isn't perfect — as subsets of the larger sample, crosstabs are subject to larger margins of error, and this is also just one month's worth of data in what will end up being a three-and-a-half-month general-election campaign. But it is enough to give us an early sense of how the Harris coalition is developing. Relative to Biden's performance in 2020, she's made gains with some groups of voters, but there are still plenty of demographic groups with whom she's underperforming.*

That said, a look under the hood reveals that her campaign has been making inroads among some groups that will be critical for the general election.

Another historic gender gap?

Since 1996, a majority of women have voted for the Democratic candidate for president, while men have only voted for the Democratic candidate in one year (2008). In addition, there has typically been a sizable difference in what percentage of the vote the Democratic candidate receives from men versus from women. The widest such gender gap in recent history occurred during the 2016 election, in which 58 percent of the two-party vote share among women went to former Secretary of State Hillary Clinton but only 44 percent of the two-party vote share among men did — a gap of 14 percentage points, according to the Pew Research Center's validated voter surveys. In 2020, according to Pew, the gender gap fell back to 7 points, much more in line with the historical norm.

In 2022, there was some suggestion that the gender gap was widening in the wake of the Supreme Court's decision in Dobbs v. Jackson Women's Health Organization, but a significantly larger gap didn't materialize. Despite concerns about abortion being top of mind for many voters in that election, there was ultimately only a 7-point gap between the percentage of women and men voting for Democrats.

Until Harris took over, this year's gender gap was shaping up to be fairly typical: The average gender gap in polls conducted and released between the debate and Biden's withdrawal was 9 points, and polls generally showed Biden performing about 3-5 points worse among both men and women than he did in 2020.**

Since Harris took over, though, this appears to have changed. On average, Harris is performing about the same with women as Biden did in 2020; in polls conducted since the handover, Harris averages less than a point better than Biden's 2020 two-party vote share among women. But among men, she's performing an average of 4 points worse than Biden's 2020 performance.

However, the campaign seems to have made gains with men since its launch: In polls conducted the first week of Harris's candidacy, she averaged 43 percent of the male two-party vote, but in the last week, that average has grown to 46 percent. On the other hand, her polling among women has hardly budged over the course of the month.

Harris has made gains among Black voters and urban voters

One of the biggest questions about the 2024 election to date has been whether nonwhite voters (particularly young Black voters) might be more likely to vote for former President Donald Trump than they were to vote for Republicans in previous elections. In an average of polls conducted and released between the debate and Biden's withdrawal, Biden was earning only 77 percent of the Black two-party vote — significantly worse than the 92 percent he won among Black voters in 2020. Among Hispanic voters, the drop was less significant, but still meaningful: Biden was polling at 54 percent of the Hispanic vote, compared with the 61 percent of Hispanics who voted for him in 2020.

Harris started off around Biden's pre-debate level of Black support but has made inroads since. In the first week of her campaign, she garnered, on average, 78 percent of the Black two-party vote share in polls; by the third week, that had grown to 84 percent. This growth, however, may be stalling out; in the last week she earned 82 percent of the two-party vote among Black voters. And although Harris's vote share among Black voters has significantly improved upon Biden's 2024 polls, she still lags far behind his 2020 performance.

On the other hand, Harris hasn't had as much success persuading Hispanic voters. On average, she is winning 56 percent of the Hispanic two-party vote share, not too different from Biden's polls before he dropped out, but still behind his 2020 margins, and she hasn't meaningfully increased support among that demographic yet.***

Harris has also made gains among urban voters. Harris's two-party vote share with this group has grown from 58 percent in the first week of her campaign to 63 percent in the last week. This may be related to the rise in her Black support: Black voters make up a larger share of the urban population than the suburban or rural population.

Harris is doing better than Biden among rural and white non-college-educated voters

The only two demographic groups we examined that Harris is meaningfully outperforming Biden's 2020 two-party vote share with are rural voters and white voters without a college degree. On average, she is earning 37 percent of white voters without a college degree; Biden won just 34 percent of this group's two-party vote share in 2020. Similarly, among rural voters, she's improved on Biden's 2020 performance by about 5 points (from 34 percent for Biden in 2020 to 39 percent for Harris this year). Because there is significant overlap between rural voters and white non-college-educated voters, these two improvements may also be related.

Despite speculation in advance of the 2020 election that Biden would perform better than Clinton did in 2016 in rural counties, that improvement never materialized. According to the Pew Research Center, Clinton earned 37 percent of the two-party vote share among rural voters in 2016, while 34 percent voted for Biden in 2020. That number is a marked decline from former President Barack Obama's performance in 2012, when he earned 39 percent of the vote in rural counties, according to the news site the Daily Yonder.

So far, Harris's polling among rural voters more closely resembles Obama than Clinton or Biden. The Democratic Party may have reached a nadir among rural Americans in 2016 and 2020, and if Harris can continue to outperform, she has the opportunity to bring some of those voters back into the Democratic fold. This may also explain why Harris has looked so much stronger in polling of Wisconsin than Pennsylvania in spite of the fact that Pennsylvania voted to the left of Wisconsin in both 2016 and 2020; Wisconsin has a larger rural population than Pennsylvania.

You might want to take this with a grain of salt, though: In the past few elections, whiter and more rural states have been more difficult to survey, so there may be some reason to be a bit skeptical of this data. After polling missed the mark in the 2020 election, researchers at the American Association for Public Opinion Research noted that "polling overstated Biden's support in whiter, more rural, and less densely populated states." Of course, pollsters will have adjusted their methods and sampling strategies to try to avoid repeating this mistake in 2024, but whether that will be successful is yet to be determined.

Young voters have coalesced around Harris

Harris has also made significant inroads with young voters, a demographic that Biden had been struggling with in recent months. In polls conducted and released between the debate and his exit, Biden was averaging 56 percent of the two-party vote share among voters aged 18 to 29. This significantly lagged his performance in 2020, when he earned 63 percent of the two-party vote among young voters, and had been one of the polling trends giving Democrats the most heartburn before his exit.

They'll be glad to know, then, that Harris appears to be turning things around with this demographic. Since taking over the campaign, she has increased Democratic vote intention among voters aged 18 to 29 to an average of 61 percent, and she's continued to persuade more young voters to her side. The first week of the campaign, she was polling at 57 percent among these voters, on average, but by the third week, that had grown to 65 percent. We see a similar trend here as we do with Black voters: Their enthusiasm may be subsiding a bit, as Harris's vote share among this group fell to 61 percent in the last week of polling. However, she remains in a position to match or possibly exceed Biden's 2020 performance with young voters.

On the other hand, Harris's support has barely budged among voters aged 65 and above. She more or less matches Biden's 2020 performance with this group, and the percentage of older voters who say they intend to vote for Harris has not meaningfully changed since her campaign began.

We did look at other age groups, but we didn't have enough data to be confident in comparing them to the Pew validated voter research. (For example, Pew reports results among voters aged 30-49, but only three pollsters in the past month have included a crosstab of that age slice.) A preliminary analysis suggests that Harris is performing about as well with voters aged 30-49 and 50-64 as Biden did in 2020, but no better.

It's also worth noting that voters aged 18-29 in 2024 are different from those who were 18-29 in 2020. Voters who were 26-29 in the last election have aged out of this cohort now, and they've been replaced by a whole new group (voters currently aged 18-21) who have never voted before, and we may see different patterns emerge as more Gen Z**** voters age into the electorate.

* * *

Now, we probably shouldn't make too much of these differences. Overall, the Democratic coalition in 2024 looks pretty similar to the Democratic coalition in the last few elections. And Harris hasn't totally reversed some of the trends that we observed in polls of Biden, such as Democrats' softer support among Black and Hispanic voters than we've seen in previous elections. It's also quite early: Voters have only had four weeks to consider a brand new candidate, and crosstabs can be noisy due to higher margins of error than topline polling samples. But the data so far suggests that Harris is making gains with some key voter groups that Democrats have relied upon in the past and improving upon Biden in some meaningful, and sometimes surprising, ways — and that could make all the difference in November.

Footnotes

*Biden's performance in the 2020 election is calculated as the percent of voters who voted for Biden divided by the percent who voted for either Biden or former President Donald Trump, according to the Pew Research Center validated voter study of the 2020 electorate. Harris's numbers are calculated as the percent of respondents in each poll who say they plan to vote for Harris divided by the percent who say they plan to vote for either Harris or Trump. If a poll asked more than one horse-race question, we used the question with the fewest candidates (for these purposes, we treated a generic "other candidate" option like a candidate). If a poll released results among more than one population or sampling frame, we used the narrowest population; that is to say, we prefer likely voters to registered voters and registered voters to all adults. We included all publicly available national polls that published crosstabs and were conducted between July 21 and Aug. 17 and released by 9 a.m. Eastern on Aug. 20.

**We used the same methodology as described above for calculating Biden's performance in polls before he dropped out.

***Some pollsters include a "Hispanic" demographic group in their crosstabs, and others include a "Latino" group. We included both categories in the Hispanic/Latino averages.

****Defined as people born between 1997 and 2012.