With just weeks remaining until 2024’s elections, “PS: Political Science & Politics” has published its election models edited by Wake Forest University faculty. 

The “Special Issue on Forecasting the 2024 US Elections” includes 12 different models forecasting the results of the 2024 United States presidential election, each predicting the winner of both the popular vote and electoral college. Professors Lina Benabdallah, Justin Esarey, Peter Siavelis, and Betina Cutaia Wilkinson of Wake Forest’s Department of Politics and International makeup “PS’s” editorial team. 

Of the forecasts included in the pre-published article, eight anticipate former President Donald Trump securing an electoral college victory and returning to the White House. 

Models in PS 

Like many US election models, those published in “PS” generally forecast their results based on two major metrics: prevailing economic conditions and an incumbent candidate’s popularity. Each prediction then varies in how heavily they weigh certain factors. 

Some models like East Carolina University political science professor Dr. Brad Lockerbie’s, lean on consumer financial forecasts. Dr. Charles Tien of Hunter College and Dr. Michael S. Lewis-Beck of the University of Iowa, meanwhile, consider incumbent approval rating with economic growth in their projection. Notably, the projections in the pre-published issue of “PS” minimally consider “the effect of the campaign on voter behavior.” 

Dr. Betina Cutaia Wilkinson is an associate professor and associate chair of the Politics and International Affairs Department at Wake Forest. As a public opinion scholar and co-editor of “PS: Political Science & Politics,” she evaluates the research and methods of election forecasts submitted to the journal. 

“For us, our main role is to maintain the status [and] rigor of the journal in terms of its ability to publish [research] that has strong theoretical or empirical contributions,” Wilkinson said. 

She explains that it’s important for political scientists to constantly scrutinize not only the results of these forecasts but also the methods that deliver them. There’s never a way to be certain, but forecasters are always readapting and considering new metrics in making predictions. 

“We have had, as political scientists, to be very careful to really do our due diligence in terms of really figuring out ‘What are the appropriate measures to put into these forecasting models?’” Wilkinson said.

Modeling in 2024 

Wilkinson highlights one nuanced metric found in this election cycle’s special edition: the strength of local political parties. Now a component in Philippe Mongrain and others’ prediction in State-Level Forecasts for the 2024 U.S. Presidential Election, the measure has been incorporated since former Secretary of State Hillary Clinton’s narrow defeat in key swing states in the 2016 presidential election. 

“Therefore, forecasting models have to focus on the states much more to be able to accurately predict what will happen,” Wilkinson said. 

The upset sent political scientists back to the drawing board but ultimately provided clarity in their model-making. After Clinton was predicted to win the race based on nationwide numbers, projections like Mongrain’s have readapted to forecast future elections using state-by-state metrics to better predict electoral college counts. 

The sophistication of models like Mongrain’s are especially valuable in this election cycle where several swing states are identified to make or break the election for either candidate. This “PS” Special Issue identifies Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania and Wisconsin as these key battleground states. 

But painting a picture with such fine detail can also influence the outcome of an election. “Horse race” coverage of any election, Wilkinson says, can both mobilize reactive voters wanting to decide a tight race — or even dissuade voters whose candidate is polling ahead from turning out to vote. 

The path to 270 

For any shot at the magic number — 270 electoral votes — among swing states, Harris has to win Pennsylvania, Wisconsin and Michigan. Any other battleground states — including Nevada, Arizona and Georgia — are just icing on the cake at that point. She gets to 270 — though not by much. 

Trump’s job is slightly easier. He can afford to lose both Wisconsin and Michigan, win Pennsylvania and comfortably walk through the 270 door. That calculus changes if Georgia goes blue again this cycle, but for this example let’s assume the state polling is right and Trump is leading there — if only marginally. The map just described is clearly the campaign’s current east coast strategy; flipping Pennsylvania red and crucially, keeping North Carolina red.

The Tar Heel state throws a wrench into all of these hypotheticals. 

If North Carolina flips, Harris can lose Pennsylvania while only needing to win one other swing state to win the presidency. Or she can lose both Wisconsin and Michigan, win Pennsylvania AND win both Nevada and Arizona and clinch 270 that way. Point is, she can run the board with North Carolina under her belt. Trump will have to play catchup in every other swing state if that happens. 

Forecasted results 

When the special issue’s individual models are averaged, a repeat result of the 21st century emerges. Vice President Harris earns 50.3% of the popular vote, although her forecasted 246 electoral votes aren’t enough to win her the White House compared to Trump’s 292. The state-level models, meanwhile, project Trump to take 312 electoral votes in a sweep of all seven swing states. 

The combined projection is not unreasonable. A Trump electoral college victory despite a marginal mandate for Harris would reflect a recent trend in presidential elections where the winning candidate fails to secure the popular vote — the third such occurrence since the turn of the century. 

James Watson contributed to this story.