The 2024 Presidential Election is coming closer and closer, practically knocking on our door. And as the day comes nearer, it seems everyone from traditional news channels to youtubers to bloggers have made their predictions, including me. However, I have confidence that my prediction will be eerily accurate. This is because instead of basing it on arbitrary factors, I have made a model that is data backed with objective assumptions to predict the election result.
The methodology of my model is fairly simple, as I only used two factors: past results and current polling data. Furthermore, my model is restricted to the seven swing states (Pennsylvania, Wisconsin, Michigan, Nevada, Arizona, Georgia, and North Carolina) since what happens there is what matters.
Past results: For the past results, I only used the 2016 and 2020 Presidential elections, since the elections before 2016 had vastly different trends, and 2016 also marked Trump’s first run. I gave the 2020 Presidential election 3 times the weighting of the 2016 Presidential election because of recency. Based on only using past results , Harris would have a lead in Pennsylvania, Wisconsin, Michigan, and Nevada, giving her the election.
Current polling data: The second factor of my model is current polling data dated Oct 6th 2024. I have used both 538 and realclearpolitics polling aggregates as the data in my model . However, as we have seen in the last two presidential elections, the polls have often been off. In order to normalize this factor , I took the polling error from 2020 for both the polling data sets and multiplied it by ¼ to account for the improvement pollsters have made to correct their errors. I adjusted both current aggregates based on the errors and weighted them equally. The result of using the current polling data as a sole predictor gives Trump an edge in every state except Michigan and Nevada, handing him the election. Interestingly, without accounting for error Harris would have a slim lead in both Pennsylvania and Wisconsin, giving her the election.
Finally, I took both data sets (past results and current polling data) and weighted them equally. The results are shown in the table and map below:
| State | Margin (+ denotes Harris Victory, – denotes Trump victory) |
| Pennsylvania | +0.29% |
| Wisconsin | -0.24% |
| Michigan | +1.23% |
| Nevada | +1.52% |
| Arizona | -1.2% |
| Georgia | -1.2% |
| North Carolina | -1.6% |
Based on the model Trump will win a very close race, with the deciding state being Wisconsin. In fact, if we were to use 2020 election turnout, Trump would win the election by a mere 8000 votes.