The goal for this poll was NOT to measure the overall popularity of each candidate, but to see if the debate changed people’s minds about who they were planning to vote for. Please see the methodology section below for more information.
We first asked 18,267 Americans between 10pm EST and 3am EST this question: “Voters can be very passionate about their political beliefs. Which of these best describes your view?”
Answers
People who answered that they are either a moderate Democrat, a liberal Democrat or an Independent (in total about 34%) moved on to the next question of the survey.
We asked those 6,211 people this question: “Did you watch the Democratic Debate tonight?”
The 1,369 people that responded yes moved on to be asked this question: “BEFORE tonight’s debate, who would you have voted for?”
The 'I’m not sure' answer was always shown last, the candidates were rotated so that there was no position bias.
We then asked those same people this final question: “AFTER you saw tonight’s debate, who will you vote for?”
They were presented with the same answer set as the previous question. Once again, the candidate name answers were rotated in position, to ensure that people had to proactively pick their candidate, not just select the first one.
Since the goal of the poll was not to predict how many votes each candidate would get (although the percentage distribution seems to align roughly with many other polls) we did not control for which states the respondents were from. We also did not control for age or gender but did attempt to get roughly a 25% distribution of women under 50, women over 50, men under 50 and men over 50.
We set out to poll as many people as quickly as possible to see how the debate impacted the opinions and backing of each candidate so the numbers should be adjusted using preexisting state polls if used to predict the actual impact on election outcomes.
The key metric we were looking for was by what percent did each candidate improve or decline in their standing following the debate. The results are listed below in order of most to least popular according to how supporters were leaning before the debate.
21.3% of people indicated that they would have voted for Biden before the debate. 19.9% of people indicated that they would still vote for Biden after the debate.
15.3% of people indicated that they would have voted for Bloomberg before the debate. 15.5% of people indicated that they would still vote for Bloomberg after the debate.
14.6% of people indicated that they would have voted for Sanders before the debate. 13.6% of people indicated that they would still vote for Sanders after the debate.
7.4% of people indicated that they would have voted for Warren before the debate. 6.9% of people indicated that they would still vote for Warren after the debate.
5.3% of people indicated that they would have voted for Buttigieg before the debate. 4.8% of people indicated that they would still vote for Buttigieg after the debate.
4.1% of people indicated that they would have voted for Steyer before the debate. 5.0% of people indicated that they would still vote for Steyer after the debate.
3.5% of people indicated that they would have voted for Klobuchar before the debate. 3.7% of people indicated that they would still vote for Klobuchar after the debate.
28.5% of people indicated that they were not sure who they would vote for before the debate. 30.6% of people indicated that they did not know who they were going to vote for after the debate.
Raw data available upon request.