Which way are key demographic groups leaning in the 2024 election?

We’re now just about a month away from Election Day, and both former President Donald Trump and Vice President Kamala Harris have been hard at work trying to persuade voters to pull the lever (or fill in the bubble, or touch the screen …) for them in November. Their efforts to persuade voters are all over the news, too: Whether it’s men, young voters, Latinos or Black voters, the campaigns are eager to discuss their outreach efforts to specific demographics that could be key to victory in November.

To see how well those outreach efforts are doing, we pulled together data from all the national polls in the last month to check how the electorate, and particularly these key demographic groups, have moved in response to last month’s campaigning. In particular, we checked in on whether the emerging Harris coalition has changed since our last look at patterns in demographic support back in August.

At that time, Harris had a 2.8-percentage-point lead in our national polling average, not too different from her 2.6-point lead on Friday, Oct. 4 at 10 a.m. Eastern. So it may be unsurprising that, among most demographic groups, there haven’t been many significant shifts since August.*

What does a racial realignment really look like?

In August, Harris was more or less polling even with President Joe Biden’s 2020 performance among white voters. But owing to an increase in support among white, college-educated voters, she’s slightly improved on that margin in the last month. She’s made small gains with Black and Hispanic voters too, but she continues to lag far behind Biden’s 2020 performance among these voters, a dynamic that has been fairly consistent throughout the 2024 campaign (we first noted Democratic erosion among these groups in spring of 2023).

To get a sense for how it might change the electoral map if Harris marginally improves upon Biden’s performance with white voters but slides significantly among Black and Hispanic voters, we turned to 538’s Swing-O-Matic. Keeping all other variables equal, we set the margins among white, Black and Hispanic voters equal to the average margin we currently see in crosstabs, and the result is an extremely close race.**

According to the Swing-O-Matic, if Harris improves upon Biden’s margin among white voters by only 2-3 points,*** she could still eke out a win even if she loses significant ground among Black and Hispanic voters. The results indicate a victory of exactly 270 electoral votes for Harris, with Trump earning 268 electoral votes. And the outcomes among the swing states are fairly predictable: The northern battlegrounds of Michigan, Wisconsin and Pennsylvania, which are whiter than the Sun Belt swing states, would stay in the Democratic camp, while all the southern battlegrounds would vote for Trump.

Of course, this exact scenario may not come to pass. Notably, we haven’t changed any other variables here: Turnout remains the same among all groups as it was in 2020, which is probably not going to happen. In addition, this scenario assumes uniform shifts nationwide, which is also unlikely: Hispanic voters in states like Florida, Arizona and Texas have shifted further to the right than those in states like California or Pennsylvania, for example.

Plus, it’s likely that the polls won’t be exactly on the mark. Every election cycle features some amount of polling error, which would push us away from this razor-close scenario in one direction or the other. Five of the seven key swing states in this scenario would be decided by a margin of less than 2.5 points, so even a historically small amount of polling error — 3 points in either direction — would probably flip at least two states (and possibly the election).

It’s also worth noting that, among pollsters who focus on particular subgroups, the decrease in support for Harris among nonwhite voters is less pronounced. For example, in a UnidosUS/BSP Research poll of Hispanic voters conducted in August, Harris earned 66 percent of the two-party vote among Latinos, higher than the 58 percent we see in national crosstabs. And in a Howard University survey of Black voters in battleground states, Harris earned 87 percent of their two-party support; she’s at just 82 percent in national crosstabs.

Nonwhite voters have traditionally been more difficult to reach in surveys, so it may be that polling of the full electorate is understating Harris’s support among these demographics. However, even taking these more generous estimates of Harris’s support among Black and Hispanic voters into account doesn’t change the picture much: Using the results from the UnidosUS and Howard University surveys in our Swing-o-Matic scenario flips only one state, Nevada, to Harris’s column.

The gender gap is narrowing — but not among young voters

A recurring theme of the 2024 election has been Trump’s attempts to make inroads with younger voters, particularly young men. And while, overall, Harris is coming fairly close to Biden’s 2020 performance among voters aged 18-29 in polling (in September, she averaged 60 percent of the two-party vote among this age group, compared with Biden’s 63 percent performance in 2020), other data suggests that young men and women may actually be drifting in different political directions.

As we noted in August, the Harris campaign had slowly been moving men into her camp, a trend that has continued into September: Her two-party vote share among men has grown about 2 points, while her polling among women has been fairly stable. While we don’t have enough crosstab data to look at young voters by gender in the polls overall (and the sample sizes for such a crosstab would likely be too small to be useful), we can use surveys specific to young people, such as the Harvard Youth Poll, to get some insight into how those voters are thinking about the election. And in the Harvard Youth Poll’s most recent survey from September, we see an enormous difference in the gender gap among young people versus in the electorate overall.

While the gender gap (the difference between the share of men and the share of women who say they will vote for Harris) in our average of national crosstabs has narrowed to about 9 points, the gender gap among young voters has widened significantly, from just 4 points in the spring Harvard Youth Poll between Trump and Biden to 17 points in the latest poll, even as Harris has improved on Biden’s polling among both young men and young women.

On the other hand, the gender gap among young voters in the Harvard Youth Poll isn’t too different from what we saw in the 2020 election. While the Pew Research Center’s validated voter survey from 2020 doesn’t include a breakdown of voters by age and gender, we can get a rough estimate of the gender gap from the 2020 exit polls. According to that data, 52 percent of young men said they voted for Biden, while 67 percent of young women said the same — a gender gap of 15 points, not too different from the 17-point gap in the Harvard Youth Poll. So while the gender gap among young voters has widened significantly since the spring, it may be less reflective of the GOP’s success in moving young men to their side and more about the young electorate returning to where it has been in the past.

* * *

Now, this analysis comes with a caveat: You probably shouldn’t make too much of a big deal over small differences in crosstabs between one month and the next. Crosstabs can be quite noisy, with small sample sizes and large margins of error, so a shift of a point or two is probably not indicative of significant changes in the electorate overall.

That being said, the overall picture here is generally not one of change, but of stability. Despite a month’s worth of campaigning, a debate that was watched by nearly 70 million people and hundreds of millions of dollars worth of paid advertising, the 2024 race does not appear to have shifted much at all — not even at the level of specific demographic groups. You can see this in our national polling average too: Since Aug. 30, Harris’s lead in the national polls has varied by less than a point. (Specifically, it has been between 2.4 and 3.3 percentage points every day.)

As with the last two elections, this race may come down to voters who decide who they’re supporting late in the game, a group that is historically less engaged with political news and may not be paying attention to the campaigns yet. It’s tempting to think that, if we stare at these polling numbers long enough, the truth of who will win the election will jump out at us. But the reality is, late movement in the race can be unpredictable, and the winner hasn’t been determined yet.

Footnotes

*Each candidate’s two-party vote share is calculated as the percent of respondents in each poll who say they plan to vote for that candidate 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. For the August averages, 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. For the September averages, we included all publicly available national polls that published crosstabs, had a final field date between Sept. 1 and Sept. 30 and were released by 10 a.m. Eastern on Oct. 4.

**For this scenario, the margins used are the topline margin among each demographic group, rather than the margin of the two-party vote.

***There is no definitive source on how different demographic groups voted in the 2020 election, since all votes are anonymous. According to the Pew Research Center validated voter study of the 2020 electorate, this change represents a shift of 2 points among white voters, while according to the 2020 Cooperative Election Study, this represents a shift of 3 points. Unless otherwise noted, Biden's performance in the 2020 election in this article is calculated as the percent of voters who voted for Biden divided by the percent who voted for either Biden or Trump, according to the Pew Research Center.