Skip to content

How did Trump win the election? New interactive tool takes deep dive into voter demographics

The tool developed by Northeastern University Distinguished Professor David Lazer allows users to select for demographics including gender, race, income, education, age and voting preference in 2020 and 2024.

Silhouette of voters lined up in Arizona.
A tool developed by Northeastern University Distinguished Professor David Lazer, lets users compare the votes of demographic groups between the 2020 and 2024 elections. AP Photo/Matt York

Donald Trump will be the 47th president of the United States — the Associated Press called the race before 6 a.m. Wednesday — and experts are already trying to figure out where the election was won or lost.

A new interactive tool developed by Northeastern University Distinguished Professor David Lazer offers help, showcasing demographic shifts among the electorate from 2020 to 2024. 

“It’s allowing the user to summarize the vote breakdown for people within very specific sub demographics,” Lazer says. 

Trump secured victories in the key swing states of North Carolina, Georgia, Wisconsin, Pennsylvania and Michigan. He also won the national popular vote. The new tool allows the user to select for gender, race, income, education, age, community type (urban, rural or suburban), and voting preference in 2020 and 2024.

“Because our survey was quite large, we have some subcategories which are small percentages of the population, but are still big enough to make comparisons,” Lazer says. “You can really zoom in on some very specific categories.”

Portrait of David Lazer.
David Lazer, Northeastern distinguished professor of political science and computer science, developed the tool with results from the CHIP50 nationwide survey. Photo by Matthew Modoono/Northeastern University

The interactive tool was developed with data from the Civic Health and Institutions Project (CHIP50), a 50-state survey effort that polls Americans on opinions and behaviors.  

About 40,000 people indicated their vote preference in surveys from 2020. Another 40,000 people who report they are eligible to vote and either plan to do so or — in some cases, have voted early — responded with their choice for president over the recent two-month survey period. 

The interactive tool shows some interesting shifts in the electorate.

For instance, younger African Americans with a high school education or less shifted 20 points toward Donald Trump between 2020 and 2024, according to the tool, with the Republican nominee tripling his vote share among the demographic.

Older African Americans with a high school education or less, meanwhile, did not shift at all — with 95% going for Joe Biden and 95% for Kamala Harris.

“There’s a huge generational shift among African Americans,” Lazer says. “That’s the kind of thing you can do with this tool.”

Hispanic males with a high school education or less and who make less than $100,000 also shifted toward Trump. Biden captured 64% of that demographic; Harris gets 46%.

Illuminating these shifts will be crucial to understanding the outcome, Lazer says.

“On average, there’s been very little shift in the electorate,” Lazer says. “If you look at slices of the electorate, there have been very large shifts — younger people have shifted quite a lot, older people have shifted more modestly.”

So, the ultimate question: can this tool predict the election? 

“You know, the answer is no,” Lazer says. “First of all it’s a national tool, not a state-level tool and the election will be determined by the Electoral College.”

However, that being said, there is a lot of information that can be extrapolated.

“We did actually just post our penultimate estimates for the presidential race,” Lazer says. “What our survey is telling us is the most likely outcome is that Harris will win.”

But he adds some very important caveats. 

First of all, the tool uses data collected from a two-month period ending Oct. 26, thus, it may have missed any late movement in the polls. 

Lazer says the survey also may lean Democratic, comparing it to the polling website 538, which shows more Trump support than his tool does. 

Ultimately, the tool may be better at explaining than predicting results. 

“For people who really want to dig deep into what happened (in the election), I think it’s gonna get down to how do different sets of people behave? How did they vote?,” Lazer says. “This tool would be very powerful for that purpose.”