For the first time this cycle, there was just one debate night, and only 10 candidates made the cut — so now we’re trying to make sense of what happened when the front-runners shared the stage. In recent weeks, the polls have shown a top tier of three to five candidates, with former Vice President Joe Biden leading, Sens. Bernie Sanders and Elizabeth Warren tied for second, and Sen. Kamala Harris and Mayor Pete Buttigieg of South Bend, Indiana, a distant fourth and fifth — but did that change last night?
Some candidates from the lower-polling tiers had strong performances — former Rep. Beto O’Rourke delivered an impassioned speech on gun violence and Sen. Cory Booker spoke nearly as much as Biden, though Booker is only polling at 2.1 percent on average (based on 21 debate-qualifying polls). But as you can see from the FiveThirtyEight/Ipsos poll conducted using Ipsos’s KnowledgePanel, the overall picture hasn’t shifted much yet — although Warren does seem to have done the most to boost her campaign. Here’s what we’ve learned so far about what viewers made of the debate and the candidates’ performances:
Which candidates wowed the crowd?
First of all, how did viewers in our poll think the candidates did on Thursday night? To answer this, we compared debate-watchers’ ratings of the candidates’ performances to their pre-debate favorability scores1 to see if any well-liked candidates failed to impress or if anyone got high marks despite lower favorability. By this measure, O’Rourke and Warren were the biggest standouts, though Buttigieg and Booker also made a positive impression. But Biden and former Cabinet secretary Julián Castro — who memorably clashed — got the lowest scores relative to their pre-debate favorability.
Who gained (or lost) potential supporters?
Another way to assess who won last night’s debate is to see who convinced more voters to at least think about voting for them. Most candidates saw some change in the share of likely Democratic primary voters who were considering supporting them, though not all changes were positive. Warren, for example, saw the biggest increase in voters who were considering her — almost 4 percentage points, while Harris lost more than 2 percentage points of potential support. But for most candidates, the numbers stayed pretty much the same as they had been before the debate. Even for those whose debate performance stood out — like Biden and Castro, who got relatively poor grades, or O’Rourke, who got a strong rating — there was little change in how many likely primary voters said they were considering voting for them.
Which candidates appeal to the same voters?
With many voters in our poll still considering multiple candidates, we were also interested in examining which candidates share potential supporters. Perhaps unsurprisingly, the two candidates who are being considered by the most voters in our poll — Biden and Warren — also tended to draw a high proportion of other candidates’ supporters, too. Seventy percent of Buttigieg’s supporters are also considering Warren, for example, while 65 percent of O’Rourke’s supporters are also considering Biden. Although many respondents in our survey said they were considering Sanders, fewer of his supporters are considering supporting other candidates. In fact, Biden and Sanders had the most exclusive supporters — 24 percent of Biden’s supporters and 18 percent of Sanders’s supporters aren’t considering any of the other candidates who participated in the debate.
Who made a positive (or negative) impression?
You can also look at the change in candidates’ favorable and unfavorable ratings to understand who got people feeling more positively about them (or perhaps gained unwanted notoriety). So after Ipsos polled voters before and after the debate, we calculated the change in candidates’ net favorability (favorable rating minus unfavorable rating). O’Rourke may not have picked up many potential supporters, but he did improve his net favorability rating by more than 8 points with his debate performance. Castro, meanwhile, took the largest hit, dropping 6.7 points in net favorability, which could be related to his heated exchanges with Biden.
|candidate||before debate||after debate||change|
Who spoke the most?
Though respondents to the FiveThirtyEight/Ipsos poll thought Biden’s debate performance was less impressive than Warren’s, it wasn’t because he didn’t get a chance to talk. Of all the candidates on the stage last night, Biden had the highest word count, with over 3,000 words spoken. Booker and Warren, the next two most prolific speakers, were about 600 and 750 words behind, respectively.
Booker’s place as the second-most-prolific talker is even more impressive considering that he’s polling in the low single digits. (The polling average is based on 21 debate-qualifying polls released between June 28 and Aug. 28.) Castro also spoke more than anticipated given his polling average (1 percent), holding the floor for longer than both Buttigieg and Sanders. Sanders, in fact, had the second-highest polling average going into the debate, but was third from the bottom in words spoken, beating out only O’Rourke and businessman Andrew Yang.
Who mentioned Trump?
In addition to counting the words spoken by candidates, we also tracked the number of times each candidate mentioned President Trump by name:
Harris was the clear leader, mentioning Trump 11 times, though as a group, the candidates talked about Trump considerably less often than they did in either night of the second debate. And some of the candidates who spoke the most, such as Biden and Warren, seemed to avoid Trump, each mentioning the president only once. Sen. Amy Klobuchar, meanwhile, got through the whole debate without saying Trump’s name even once.
So did the single-night debate change the game? Thursday night’s debate drew about 14 million television viewers, which is more viewers than both nights of the second debate, but still slightly fewer than those who tuned into watch the first debate. And if our poll with Ipsos is indicative of voters’ reactions, then the needle didn’t move all that much. But for those of you who preferred the two-night approach, you might be in luck — the fourth debate, set for Oct. 15 and potentially Oct. 16, might be split across two nights, since at least 11 candidates have qualified so far. (The Democratic National Committee hasn’t yet confirmed what it will do, however.) Either way, we’ll be here live blogging, so stay tuned!
Do you want even more debate coverage?
Cool graphics from other sites
- The New York Times compared how long each candidate talked about each issue.
- Bloomberg charted the time devoted to each issue in the first three primary debates:
- Nigel Chiwaya at NBC News looked at who attacked who on the debate stage (you can see how much Biden was the focus of others’ barbs).
- And The Washington Post looked at who invoked Obama positively and how often.
And here’s more great post-debate analysis:
- Fact-checks from ABC News, as well as five key takeaways.
- Vox’s roundup of the winners and losers and the most substantive answers.
- Debate recaps from The Washington Post’s preeminent Dan Balz …
- … and the Associated Press’s wonderful Julie Pace.
- From The New York Times, a recap of what Trump was doing while the Democratic candidates were debating.
Finally, check out the rest of our debate coverage:
- Our “Who Won The Third Democratic Debate?” project, in which we partnered with pollster Ipsos to show how voters reacted to Thursday’s event.
- And our live blog of everything that went down Thursday night.
Additional contributions from Aaron Bycoffe.
Pre-debate favorability was calculated by assigning a 0 to 100 score to each respondent’s answer to the favorability question, where “very favorable” is equal to 100, “somewhat favorable” is equal to 75, “heard of, no opinion” is equal to 50, “somewhat unfavorable” is equal to 25 and “very unfavorable” is equal to zero. Scores were then averaged to create an overall favorability index for each candidate. Respondents who hadn’t heard of the candidate were not included.
Laura Bronner is FiveThirtyEight’s quantitative editor. @laurabronner
Ella Koeze is a visual journalist for FiveThirtyEight. @ellawinthrop