Polls Since The Second Debate Show Kamala Harris Slipping

Polls since last week’s Democratic debate haven’t shown the sort of dramatic swings that we saw after Round 1 — but they do show some shifts. In particular, they show further downward movement for Kamala Harris, who had already lost much of her bounce following the first debate.

Candidates Attend Second 2020 Democratic Presidential Debates

Polls since last week’s Democratic debate haven’t shown the sort of dramatic swings that we saw after Round 1 — but they do show some shifts. In particular, they show further downward movement for Kamala Harris, who had already lost much of her bounce following the first debate.

So far, there have been five national polls conducted entirely after the debate that allow for a direct comparison to an earlier poll by the same pollster. These are the polls from Quinnipiac University, Ipsos, Morning Consult,1 YouGov and HarrisX.2 Because I’m feeling a little fancy — and because these polls are something of a mixed bag in terms of sample size and quality — I weighted the polls based on sample size and the pollster’s rating, as we do for our Trump approval tracker and in our election models. Here is the before-and-after comparison for each candidate between the post-debate polls and the most recent pre-debate polls from the same polling firms.

Harris is biggest decliner after second set of debates

Weighted average of five post-debate polls, with comparison to pre-debate polls by the same pollster

Weighted average of polls
Candidate Before Debate After Debate Change
Biden 30.2% 28.4% -1.9
Sanders 15.3 17.1 +1.8
Warren 13.0 14.6 +1.6
Harris 10.7 7.9 -2.8
Buttigieg 4.7 5.4 +0.7
O’Rourke 2.8 2.6 -0.2
Booker 1.7 2.5 +0.8
Yang 1.9 1.6 -0.3
Gabbard 0.8 1.3 +0.5
Castro 1.1 1.2 +0.1
Klobuchar 1.0 0.9 -0.1
Gillibrand 0.7 0.7 +0.0
de Blasio 0.6 0.5 -0.1
Ryan 0.7 0.5 -0.2
Williamson 0.5 0.5 +0.0
Bullock 0.3 0.4 +0.1
Delaney 0.6 0.4 -0.2
Bennet 0.5 0.3 -0.2
Hickenlooper 0.3 0.2 -0.1
Inslee 0.0 0.1 +0.1

Polls included in the weighted average are Quinnipiac (weight 1.40), Ipsos (1.49), Morning Consult (2.13), HarrisX (1.16) and YouGov (0.94). Only candidates who participated in the debates are listed in the table.

Harris was in the single digits in all five post-debate polls and was off by nearly 3 percentage points on average as compared with the pre-debate polls. Earlier this week, I discussed how Harris seems to be stuck in between Joe Biden, on the one hand, and Elizabeth Warren and Bernie Sanders on her other, more-left leaning hand. And that’s starting to show up in the numbers. If you look at the Quinnipiac poll, for instance, there’s no single group of Democrats — say, wealthy or young or black Democrats — among whom Harris is polling at any higher than 10 percent, whereas Biden, Warren and Sanders all have fairly distinctive bases.

Meanwhile, Biden has fallen by almost 2 points in the post-debate polls. It’s not clear that this is necessarily because of his debate performance; according to YouGov polling for HuffPost, there were slightly more Democrats who said the debate improved their views of Biden than those who said it worsened them. But the debates can serve to showcase other Democrats’ talents, a process from which the well-known Biden stands more to lose than to gain. Keep in mind, though, that Biden recovered almost entirely from what was a much bigger polling slump after the first debate.

Warren and Sanders have also gained a point or two since the debate. In Warren’s case, the movement is part of a slow-and-steady climb upward, although the debates do seem to at least partially account for that forward movement, as there were far more voters who said last week’s debate improved their impressions of Warren than those who said it worsened them.

For other candidates, polling movement can be hard to come by. Pete Buttigieg and Cory Booker both got positive marks from voters for their debate performances, but they gained only a percentage point or so in the topline numbers.

And other candidates that various media outlets (including us here at FiveThirtyEight!) claimed performed well in the debates, such as Steve Bullock, John Delaney and Marianne Williamson, didn’t see their numbers move at all, whether because voters actually disliked how they debated (as they clearly did in the case of Delaney) or just prefer other options.

After the debates, I issued a set of my periodically updating, not-to-be-taken-too-seriously nomination tiers, which reflect my subjective impression of the candidates’ relative likelihood of winning the nomination (i.e. there’s no statistical model behind these). But that was based on a lot of educated guessing of how voters would react to the debates. Now that we have actual data in hand, a few further tweaks are in order.

Nate’s not-to-be-taken-too-seriously presidential tiers

For the Democratic nomination, as revised on Aug. 8, 2019

Tier Sub-tier Candidates
1 a Biden
b Warren
c Harris ↓, Sanders ↑
2 Buttigieg, Booker
3 a O’Rourke ↑, Klobuchar, Castro
b Yang
4 a Inslee, Gillibrand, Gabbard
b Everyone else

Note: Steve Bullock was demoted into the “everyone else” tier.

I said last week that I was on the verge of moving Sanders back into Tier 1, and I’m going to go ahead and do that. I still have lots of concerns about his chances, most notably that the data suggests that he potentially has a high floor but a low ceiling. But I also have a lot of concerns about Harris, who I’ve demoted into Tier 1c — and who has fallen to a pretty clear fourth place in the polls.

In fact, other than Biden (who can win just by holding on to his current support, more or less) and Warren (who has the most upward momentum), none of the candidates has a particularly compelling theory of the case right now. Sanders’s is something like: Hope voters get tired of Biden; hope Warren stalls out, perhaps because she’s competing with too many other candidates for white, college-educated voters; win a couple of the early states with, say, 25 percent of the vote, and then hope the party rallies around you. There’s a lot of hoping in there — but every candidate faces uphill odds in such a large field, and at least Sanders has a distinctive brand and a decent-sized, fairly enthusiastic base.

I’m also slightly moving up Beto O’Rourke, which comes after a long series of downgrades for the Texan. Partly I just think I’d been overreacting before in demoting him below the likes of Amy Klobuchar and Julián Castro, both of whom O’Rourke polls slightly better than. But he’s also been forceful in responding to President Trump — and to the media — after the El Paso and Dayton shootings, and gun control is a high-priority issue for Democrats.

Finally, Bullock gets demoted into the “everyone else” part of Tier 4. He’s quite unlikely to make the third set of debates next month, and Biden seems to have monopolized the market for Democrats who are looking for a moderate white guy.


Footnotes

  1. For Morning Consult, I’m comparing their new poll to their full-week poll from July 22-28. Morning Consult also released some one-day samples before and after the debate, which I’m not using.

  2. Various outlets conduct polling with HarrisX. I’m using the most recent HarrisX poll, which was conducted in conjunction with ScottRassmusen.com, and comparing it to the last HarrisX polling for Ramussen that was conducted entirely before the debate.

Nate Silver is the founder and editor in chief of FiveThirtyEight.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *