Biden on track for a knife-edge win against Trump
Key takeaways:
Joe Biden would beat Donald Trump by 287 electoral college votes to 251 if the presidential election were held today.
Biden would also win the popular vote by 49.6% to 46.9%.
However, Biden’s projected win is based on the narrowest of leads in key battleground states: in Arizona, Pennsylvania and Wisconsin he is winning by less than half of one percentage point. If Trump were to overtake Biden in Pennsylvania alone, he would win the election.
RFK Jr. increases Biden’s lead in key battleground states when included.
The dashboard exploring the data is also available here.
Stack MRP projections
This post summarises the results of Stack’s latest MRP (Multilevel Regression and Poststratification) modelling for the 2024 US Presidential election. Stack has used this technique in the past to accurately predict the result of elections in the UK (2019), US (2020 and 2022), and Spain (2023). This poll is among the most comprehensive of the race to date and follows Stack’s first MRP of the US presidential election, which was released in November last year.
We predict that, if the presidential election were held today, Joe Biden would win a second term in office by narrowly beating Donald Trump in the Electoral College. We project that Biden would win 287 Electoral College votes to Trump’s 251 (see the dashboard at the top of this page).
Although Biden would once again comfortably win the popular vote against Trump by 49.6% to 46.9%, his path to victory in the Electoral College currently hinges on razor-thin majorities in a handful of key battleground states. According to our poll, Biden would once again carry the states of Arizona and Wisconsin by a margin of 0.2%, while Trump would win back the state of Georgia by 0.7%. Our results also project that Biden would hold onto Pennsylvania by the thinnest of margins, less than 0.1%. However, if Trump were to overturn Biden’s lead in Pennsylvania, and all else were to stay the same, the Republican candidate would be the one heading for a second term in the White House: in what would be the narrowest of Electoral College wins, Trump would beat Biden by 270 votes to 268.
The results of our latest MRP also show that the inclusion of third party candidates would not change the overall picture, with Biden still winning the race by 287 electoral votes to Trump’s 251. However, offering voters the choice of third-party candidates does benefit Biden in key battleground races: in Arizona, Nevada and Wisconsin, his lead over Trump improves by one point when compared to the head-to-head. Biden’s margin also improves, albeit by a more modest amount, in the key swing states of Pennsylvania and Michigan.
When comparing this poll’s results to our last MRP of the US Presidential race, which we published in November last year, we see that Biden is making consistent gains across multiple demographic groups, especially older, white, rural, suburban and lower educated groups. At the same time, however, he is currently underperforming with younger voters and ethnic minority groups that have traditionally formed part of the Democratic base. Although Biden won 65% of 18-24 year olds according to exit polls from 2020, he would only win 53% of this demographic if the election were held tomorrow. Similarly, we see that Biden is performing worse with African Americans, Hispanics and Asians, with his vote share within these groups dropping by two to three points compared to 2020.
“The 2024 election cycle is truly on a knife edge. According to our in-depth research, President Biden is currently on course to cling on in the White House. But this is only by the narrowest of margins: only a percentage point in some states. The Midwest states that took Donald Trump to victory in 2016 are sticking with Biden for now, but it is effectively a toss-up in many of these key battlegrounds, there is everything to play for and every vote will count. Meanwhile the race for Congress is just as tight with the Republicans currently set to regain the Senate and retain control of the House. This is going to be a truly nail-biting contest.”
Methodological note
Finally, a brief note on methodology. Stack interviewed 7,793 respondents from across the US about their political opinions between 27th March and 13th April 2024. These surveys then powered our tried-and-tested MRP model. MRP modelling is a statistical technique that uses large-sample national polling data to create estimates of public opinion at lower geographic levels. This model effectively used census data to generate a geographically detailed view of the make-up of the US electorate. It then estimates the relationship between these demographic and geographic variables and voting behaviour to project vote intention results at the state level.
