Skip to content
Stories

NULab Faculty David Lazer Uses CHIP50 to Test Biden’s Poll Numbers

this is the correct image with higher resolution

Reposted from Northeastern Global News

By Cody Mello-Klein

President Joe Biden’s performance during the first presidential debate has been widely panned, with critics and even those within his own party calling for him to step aside. But the question remained: Would Biden’s shaky performance against former President Donald Trump result in a noticeable dip in the polls for the president?

According to a new report from the Northeastern University-led data project CHIP50, the answer is no.

Led by David Lazer, university distinguished professor of political science and computer science at Northeastern, the report indicates that the debate had little if any impact on people’s voting preference. Lazer hopes the report helps illustrate the dangers of making a mountain out of a molehill when it comes to the media interpreting data.

“Even the New York Times, which is usually better about this, talked about a very tiny shift that was totally insignificant statistically like it was evidence that it was a shift toward Trump after the debate,” Lazer says. “My hope is that reporters look at this and say, ‘Maybe we need to be careful in overinterpreting noise as actual signal.’”

Seeing the dominant narrative coming out of the debate, Lazer and the team at CHIP50 decided to test the hypothesis that Biden had lost ground in public opinion after the debate. Notably, Lazer says, they didn’t survey two different cross-sectional groups of people before and after the debate like most polls. Instead, the team was able to survey the same group of respondents from a survey conducted before the debate.

Lazer says using the same group of people helps make the results more precise, which is important in polling that inherently has a margin of error.

What the report finds is that Biden held on to 94% of the people who said they would support him before the debate. For Trump, 86% of people who said they would support him before the debate said they would do so after the debate. 

“What we see is that there is some churn –– maybe 10 percent or so of people change what they answer –– but that the net result is not a movement away from Biden,” Lazer says. “If anything, it seems that Biden is holding on to his people somewhat better than Trump.”

In fact, generally, the shift in support was actually more in Biden’s favor, although Lazer notes it’s not statistically significant and well within the margin of error.

There were very minor shifts between the candidates: 1% of people who said they would vote for Biden before the debate, switched to preferring Trump, while 3% switched from Trump to Biden. Similarly, respondents who said they were unsure who they would vote for before the debate were slightly more likely to switch to preferring Biden after the debate.

“[It all] points in the same direction, which is that it seems unlikely, based on our data, that things shifted toward Trump after the debate,” Lazer says. “If anything, our point estimate is a little more toward Biden, but I wouldn’t make a lot of noise about that.”

There were more significant, but still minor, shifts toward people preferring other third-party candidates after the debate. About 4% of Biden’s supporters and 6% of Trump’s supporters said they would prefer the “Other” category on the survey post-debate. Meanwhile, 6% of those who preferred another candidate before the debate shifted to both Biden and Trump after the debate, resulting in a small net impact.

The relatively stable trend of poll numbers in the 2024 election is indicative to Lazer of how polarizing politics have become where very little will sway voters from their preferred party.

“Trump was convicted of a set of felonies,” Lazer says. “The impact it had on surveys was zero. Biden had a debate where most people said it proved he was too old. Survey respondents said, ‘Yeah, I saw that. He’s too old. I’m still voting for him.’ The numbers just aren’t moving.”

Lazer says “the overconfidence we have in seemingly precise statistics” isn’t anything new during an election cycle, but he cautions that the 2024 election is different.

“In this case it’s actually materially affecting decision-making,” Lazer says. “If you make the narrative that things are cratering for Biden, it may affect whether Biden stays in the race. I don’t know if he should stay in the race or not, but decisions should be made on accurate information.”

More Stories

Decorative NULab logo.

Digital Integration Teaching Initiative Spring 2025 Call for Partnerships

09.30.2024

NULab Grants Fall 2024 CFP

09.17.2024

NULab Fall 2024 Calendar of Events

10.04.24
News