Presidential Election Predictions
This prediction takes polls into account. However, as the past few elections have demonstrated, polls are only a piece of the puzzle are imperfect. This model therefore takes other factors into account such as:
The trend of each state in the past four national elections (2016-2022) as compared to the nation as a whole
Historic polling Error (national and by state)
This gist of polling error - since 1980, it is generally seen that the incumbent party tends to outperform the national polling average. Adding weight to past elections individually with 1) a Democrat in White House; 2) an open contest; and 3) a party seeking a second term in office, the national polling miss is 1.1%. This also gives the most weight to recent elections, considering the unpredictability that Donald Trump has added to polling in general.
For States, the 2022 midterm elections saw some notable polling misses in either partisan direction - those are applied as well, but ultimately shift results by less than a percentage point in almost all cases.
Given all of the above, this model predicts that Kamala Harris will win the popular vote by about 2.8%, somewhat less than the 4.5% margin of victory Joe Biden achieved in 2020. Harris is projected to win every state Biden won in 2020 except for Georgia, which will also give her an electoral college victory.
Biden & Harris - 51.3% (306 Electoral Votes)
Trump & Pence - 46.9% (232 Electoral Votes)
Close States (Less than 10% difference)
Senate Election Predictions
The same assumptions apply to the U.S. Senate elections, with the following additional built into the model.
How the incumbent performed relative to other candidates of their same party, in the same state, in the previous election (i.e., 2018). If the opposing candidate also ran in a statewide election in the past, their relative performance is also included, albeit it is weighed less heavily.
Given all of the above, this model predicts that the Republicans will gain the open Senate seat in West Virginia and narrowly defeat the incumbent in Montana, giving them a slim 51-49 margin in the Senate. This would give the Republicans control of the chamber; however, Republican Senator Lisa Murkowski has hinted at leaving the GOP in the past and registering as an independent, which could come into play in this scenario.
House Election Predictions
The same assumptions apply to the U.S. model, with the following additional inputs built into the model.
Generic Ballot (def. how Democrats in aggregate perform against Republicans in aggregate) - this takes into account the average polling of this measure. However, special elections in aggregate tend to be about as predictive as the generic ballot itself in past election cycles. The national environment is determined as such
(Polled generic ballot + democratic performance in special elections compared to 2020 presidential results in that same district)/2
Given this, the political environment is predicted to yield a 2.2% Democratic advantage.
To predict 2024 results, this is weighed against the national environment in the last house election, which had an ultimate generic ballot performance of a 2.7% Republican advantage
Given all of the above, this model predicts that the Democrats will gain a net total of seven seats in the House of Representatives, giving them control of the Chamber. Hakeem Jeffries would then become the African American to be Speaker of the House.
Close Races in Democratic held seats (Less than 10% difference or seat changes)