2020 USA Presidential Election Forecast
For an overview of these mathematically-based prediction models, see: 2020 Presidential Election True Dollar Journal Forecast. Donald Trump Wins. Biden Runs Against the Wrong Opponent and Loses.
For data tables that run the models, see the sections below in this work.
Models Designs
Each model runs off a data set. One data set comes from Real Clear Politics. The other data set comes from Axios / SurveyMonkey. The RCP data set is a calculated average derived from telephone-based surveys. The A/SM data set is a provided rolling average derived from online surveys.
For any state, the 75k models apply polling percentages to the total of likely votes from counties of with populations of 75,001 or greater and for the 50k models, those models apply the same polling results to the total of likely votes from counties with populations of 50,001 or greater. All models assume that "rural" voters (counties with populations of less than 75,000 or 50,000 respectively) will vote in the same percentages as they did in 2016.
The model designs arise from insider knowledge on how public opinion polling firms operate and thus from where almost all completed questionnaires will be derived.
Update: 1-Nov-2020
The models now are run with a synthetic data set which uses the RCP data when there are sufficient numbers of surveys dating from September 1, 2020. For states where no telephone surveys have been done by polling firms and for states with a small number of telephone surveys since September 1, 2020, data from the SurveyMonkey online surveys gets blended into the data for such states.
Gab It
