But a careful look at the available data strongly indicates that Harris’s coalition looks very different from Obama’s and is still struggling to match the contours of Biden’s 2020 coalition.

This pattern is clear when you break down the Harris coalition by key demographics and compare her support with that of earlier Democrats. To do this, I used the Catalist data for the 2012-2020 presidential elections and New York Times-Siena College data on likely voters for this cycle. A cautionary note: We don’t know at this point who exactly will vote this November. Turnout will surely differ in some ways from four years ago. But using the current data on likely voters is the best way to see how the Harris coalition is evolving and how it differs from 2020 and earlier Democratic coalitions.

White working-class voters: In the June Times-Siena poll, Biden was losing White working-class voters by 32 points; in the new September poll, Harris is losing this group by slightly more, 36 points, worse than Biden’s 26-point loss among these voters in 2020. Although other polls have Harris doing somewhat better than the Times-Siena poll indicates among this demographic, the overall pattern suggests the Harris coalition has less White working-class support than Biden’s, renewing the general Democratic slippage among these voters in recent years. The Times-Siena data also indicate that the Harris coalition includes less non-White working-class support than Biden’s did in 2020.

Black voters: In the June Times-Siena poll, Biden was carrying Black voters by only 39 points over Trump. But in the first Times-Siena pollafter Biden dropped out and Harris was the presumptive nominee, her margin increased to 53 points and in their latest September poll the margin for Harris among Black people was 64 points. That’s a lot of progress. However, Biden’s margin in 2020 among these voters was 81 points. With six weeks until the election, Harris will need to work hard to match Clinton’s 86-point lead over Trump among Black voters in 2016 or Obama’s 93-point margin in 2012.

Hispanic voters: Latino voters, on the other hand, do not appear to have improved much for the Democrats since Harris got into the race. In the June Times-Siena poll, Biden was ahead of Trump by 14 points among Hispanics, which is actually a bit more than Harris’s lead among these voters (12 points) in September. This level of support for Harris is 11 points less than Biden’s 23-point Hispanic margin in 2020, which in turn was 16 points less than Clinton’s in 2016.

Young voters: Biden showed persistent weakness among young (18- to 29-year-old) voters before he dropped out. The Democratic margin among these voters averaged around nine points in pre-dropout Times-Siena polls, which has improved to an average of 17 points since Harris entered the race. However, that 17-point lead is still significantly less than Democrats have attained among these voters in the past three presidential elections when their advantage was a rock-steady 23 points in 2012, 22 points in 2016 and 23 points in 2020.

Seniors: Before Biden left the race, he enjoyed a small, three-point margin among voters 65 and over. Since Harris entered the race, however, Democratic performance among senior voters has deteriorated. The September poll has Harris running seven points behind among this group. That would be in line with recent Democratic performance: Obama lost these voters by six points in 2012 and Clinton lost them by eight points in 2016.

Women, men and the gender gap: Democrats of late tend to do much better among women than among men. This election is no exception. Before Biden dropped out, he was doing 17 points better among women than men (plus 5 among women, minus 12 among men compared with Trump in the June Times-Siena poll). Since then, the gender gap has widened substantially. In the September poll, the gap was 26 points — a 12-point advantage for Harris among women and a 14-point deficit among men.

If Harris’s lead among women is now similar to Biden’s lead in 2020 (13 points), her deficit among men is significantly worse than Biden’s in 2020 (minus six points among male voters). Other data see Harris’s deficit among men as less dramatic than the Times-Siena poll but are still consistent with a deterioration in male support relative to Biden in 2020. It therefore appears that the widening of the gender gap relative to 2020 might not be a favorable development for the Harris coalition, as it is mostly based on a decline in male support rather than an increase in overall female support. The latter is despite sharply increasing liberalism among the youngest women voters.

While Harris has improved on Biden’s margins among some demographics, the data suggest she is not yet replicating the coalition that won the White House for Democrats in 2020. She might improve her showing among some groups by Election Day — Black and younger voters are two possibilities — but for now, she is underperforming her party’s historical patterns with non-White and working-class and younger voters.