Researchers for the Graduate School of Political Management have developed a model to predict the outcomes of each contest during the 2020 Democratic primaries.

Members of the Public Echoes of Rhetoric in America Project, who work to study how the public responds to political messages, released two projections for the Iowa caucus and New Hampshire primary earlier this month. Researchers associated with the project said the model is the first of its kind based on its formula for making projections.

“This groundbreaking prediction model that incorporates the conversational echoes on Twitter offers researchers a promising new approach towards understanding the dynamics of presidential nominations,” Lara Brown, the director of the school, said in a release.

The project’s Iowa caucus model projected former Vice President Joe Biden to win the caucus with 24.1 percent of the vote. Former South Bend, Ind. Mayor Pete Buttigieg and Vermont Sen. Bernie Sanders are currently leading in that race, which remains too close to call.

The New Hampshire primary model predicted Sanders to take the contest. He has been declared the winner.

The model’s formula is based on the number of Twitter mentions of each candidate, the sum of money each candidate has on hand and each candidate’s endorsements, the release states. The model is the first to take into account the number of times a candidate’s name is mentioned on Twitter, according to the release.

The same team of researchers started a “Weekly Tweeterboard” last semester to track the number of Twitter mentions for 2020 presidential candidates and other politicians.

“Twitter mentions were particularly able to provide improved predictions not captured by the fundamental factors,” Meagan O’Neill, a research scientist for the model, said in the release.