Skip to main content
Menu
The Third Debate Influenced Those Who Watched … But A Lot Of People Didn’t

The final numbers from our poll with Ipsos are in, and the big takeaway is that the third Democratic presidential debate didn’t seem to have a big effect on voters’ views of the primary. Ipsos interviewed the same group of likely Democratic primary voters twice — before and after the debate — and the poll found that the share of respondents who said they were considering voting for each candidate largely remained the same. Nor did the debate help voters eliminate some choices: The average number of candidates voters said they were considering (2.5) remained unchanged. But that doesn’t mean there weren’t important shifts in the race.

If the debate had a winner, it probably was Sen. Elizabeth Warren. Although the share of voters who were considering supporting her went up by only 2.4 percentage points, it was the biggest increase in the field. She was also the only one of the top three candidates to gain potential supporters; former Vice President Joe Biden and Sen. Bernie Sanders both lost ground. In addition, Warren received the highest marks for her debate performance from respondents in the FiveThirtyEight/Ipsos poll conducted using Ipsos’s KnowledgePanel.

The third debate may have also shaken up the second tier of candidates. For example, South Bend, Indiana, Mayor Pete Buttigieg and Sen. Kamala Harris now have roughly equal numbers of people thinking about voting for them. (Going into the debate, Harris had about a 6-point lead over Buttigieg on this question.) Former Rep. Beto O’Rourke, who has long been foundering in the polls, saw a 7-point uptick in his net favorability rating (favorable rating minus unfavorable rating), meaning he made a positive impression on voters, though it did not meaningfully increase the percentage of Democrats who were considering voting for him. Sens. Amy Klobuchar and Cory Booker also saw modest increases in both their net favorability rating and the share of respondents who were considering them. But former Secretary of Housing and Urban Development Julián Castro, who has struggled to break out, may have landed himself in jeopardy. Castro had the biggest drop in net favorability rating (it dropped 10 points) and received the lowest marks for his debate performance.

So what do we know about the voters driving these changes? Well, the biggest thing we know is that the 1,465 respondents who watched the debate reacted much more strongly than the 2,004 respondents who did not watch. This is certainly no surprise, but it suggests that the secondhand messages (like news coverage and social media reactions) that non-watchers may have been exposed to didn’t affect their opinions very much. For example, most of the movement in O’Rourke’s and Castro’s1 net favorability ratings was driven by debate watchers. In fact, opinions on O’Rourke barely moved among those who didn’t watch the debate.

Candidates made an impression on debate watchers

Change in net favorability for candidates in a FiveThirtyEight/Ipsos poll taken before and after the third Democratic primary debate, among debate watchers and non-watchers

debate Non-watchers debate Watchers
candidate Before after change before after change
O’Rourke +23.6 +24.1 +0.5 +26.1 +41.4 +15.3
Booker +22.6 +20.5 -2.1 +34.8 +42.2 +7.4
Klobuchar +5.6 +6.0 +0.4 +13.2 +20.0 +6.8
Buttigieg +27.3 +26.5 -0.8 +42.9 +49.1 +6.2
Warren +43.9 +45.9 +2.0 +56.9 +62.6 +5.7
Biden +44.1 +44.1 +0.0 +47.6 +51.1 +3.5
Sanders +44.7 +44.2 -0.5 +43.8 +44.4 +0.6
Harris +29.3 +29.3 +0.0 +36.8 +37.0 +0.2
Yang +10.8 +11.4 +0.7 +22.2 +18.8 -3.4
Castro +12.4 +9.7 -2.7 +32.6 +9.7 -23.0

From a survey of 3,473 likely Democratic primary voters who were surveyed between Sept. 5 and Sept. 11 and again from Sept. 12 to Sept. 16, of whom 1,465 watched at least some of the debate and 2,004 did not watch (four people did not answer the question).

Overall, debate watchers developed more positive feelings toward just over half the field, including O’Rourke; businessman Andrew Yang saw a small drop, though not nearly as large as Castro’s, and Sanders and Harris saw little change. However, among non-watchers there was far less movement. Their opinions changed by more than a point for only three candidates — Warren, Booker and Castro — with Warren the only one who improved her standing in their eyes. Castro had the biggest drop here too, but Booker also decreased in the estimation of non-watchers, which is a little odd, considering he had the second-biggest boost in his net favorability ratings among those who watched the debate.

Finally, let’s turn to the question of whom Democrats are considering voting for and how it differed between debate watchers and non-watchers. Interestingly, the two respondent groups moved in opposite directions for seven of the 10 candidates.

Debate watchers and non-watchers reacted differently

Change in the share of respondents considering voting for each candidate in a FiveThirtyEight/Ipsos poll taken before and after the third Democratic primary debate, among debate watchers and non-watchers

debate Non-watchers Debate watchers
candidate Before after change before after change
Buttigieg 20.3% 18.5% -1.8 25.4% 29.8% +4.4
Klobuchar 5.8 5.2 -0.6 7.8 11.1 +3.4
O’Rourke 15.0 13.1 -1.9 16.9 20.3 +3.4
Warren 41.4 42.3 +0.8 50.1 53.0 +2.9
Booker 11.6 11.5 -0.1 16.2 18.4 +2.3
Yang 7.9 7.8 -0.1 12.0 13.0 +1.0
Someone else 4.4 4.2 -0.2 2.9 2.8 -0.1
Castro 5.5 5.3 -0.2 11.4 9.0 -2.4
Harris 27.0 23.4 -3.5 30.0 27.7 -2.4
Biden 55.0 55.9 +1.0 58.3 55.6 -2.7
Sanders 41.7 42.4 +0.8 40.0 37.0 -2.9

From a survey of 3,473 likely Democratic primary voters who were surveyed between Sept. 5 and Sept. 11 and again from Sept. 12 to Sept. 16, of whom 1,465 watched at least some of the debate and 2,004 did not watch (four people did not answer the question).

For example, Buttigieg picked up ground among debate watchers but got less popular among non-watchers, and the same was true for Booker, Klobuchar, O’Rourke and Yang. Meanwhile, Biden and Sanders lost potential supporters among debate watchers but gained among non-watchers. This could be a result of noise in the data, but it could also reflect how, according to other polls, those two remain more popular among voters who are not highly informed or have yet to tune into the race.

Harris was notable in that she was the only candidate whose change in support was greater among non-watchers than among watchers — and both were negative, which helps explain why she lost the most potential voters of any candidate. Likewise, Warren was the only candidate who gained more potential voters among both watchers and non-watchers, which is another point in favor of her “winning” the debate, however small that victory might be.

Footnotes

  1. Notably, before the debate, those who went on to watch were much more likely to have a favorable impression of Castro than those who wound up not watching; it looks like he burned goodwill among people who were previously inclined to like him, which could be an especially bad sign for him.

Laura Bronner is FiveThirtyEight’s quantitative editor.

Nathaniel Rakich is FiveThirtyEight’s elections analyst.

Comments