Prediction of the Winner in 2020 American Presidential Election (Pre-Recorded)

• In Event: Forecasting Elections

Abstract

A mathematical model was developed using engineering analysis based on the principles of probability and statistics. Polling data for the general election from well-known and credible sources (ie. Survey Monkey/Tableau, Survey USA, YouGov/CBS, ABC/Washington Post). The basic premise is that polls will capture the pulse of the people. The model used is a dynamic model, dependent on timing of the poll and political events surrounding the country at that time. Density function of the polling data is constructed, the validity of which is checked using the well known chi-square test. The mathematical analysis showed that the polling data followed a normal distribution. This was followed by the determination of the parameters of the normal distribution and finally the probability of the winning of each candidate (in the 2020 American Presidential election, it is Biden of Democratic party and Trump of Republican Party). The developed mathematical model predicts both the final popularity vote as well as the electoral college votes. The mathematical used predicted popularity vote for Biden and Trump as 49.22% and 45.58% and electoral college votes of 350 and 188 respectively which is pretty close and within the statistical margin compared to the actual American Presidential election results. Another important observation that can be made is that the “silent majority” that helped President Trump in 2016 didn’t play any significant role in 2020 American Presidential elections which would have skewed the results.
The details of the work are shown below.

Histogram creation methodology

Collect all sampling data and determine the range of the data
Using range, determine the width of each bin in which the data will be sorted.
Example: If your data ranges from 1-100, then 5 bins of width 20 can be determined. Bin 1 can have numbers 1-20, bin 2 can have 21-40 and so forth.
Next, a table of frequencies needs to be determined. Examine the data and note where a data point falls in the bins chosen. Make a tally of this count.
Use this tally to draw the histogram. The height of each bar for each bin corresponds to the frequency of the bin at the base of the bar.

Determining Mean and standard deviation of the data

To find the mean of the data the following expression was used:
µ=∑▒x_i /n
Where xi represents all the data points and n is the number of data points in the sample under consideration
To find the standard deviation:







DECLARATION OF WINNER

This information, in addition to declaring the winner, also assigns the electoral votes for the two candidates in the presidential elections. Thus, the information about the popular vote as well as the electoral votes is collected.
These two essential pieces of information, are judiciously combined to declare the winner in the 2020 presidential election.
SAMPLE OF POLLING DATA USED



THE POPULAR VOTE: SAMPLE OF STATISTICAL ANALYSIS OF STATEWIDE POLLING DATA



























FORECASTING the 2020 PRESIDENTIAL ELECTION
Winning Probability









FORECASTING the 2020 PRESIDENTIAL ELECTION
Popular and Electoral College







FINAL PREDICTION

Joe Biden will win the 2020 presidential election
The mathematical model predicts Biden will win by a wide margin (electoral and popular votes)
Popular Vote: Biden 49.22%; Trump 45.58%
Electoral Vote: Biden 350; Trump 188

Authors

• Chandrasekhar Putcha, California State University, Fullerton

• Brian Sloboda, UoP