Electoral College Landslide Predicted due to Shy Voter and Mail-In Ballot Effects: Trump 354, Biden 184

In the 2016 election many pollsters failed to take into account the state-by-state nature of the electoral college. For 2020, most pollsters are failing to take into account the “shy voter” effect and the rejected mail-in ballot effect.
The following analysis takes voter rolls as of 9/01/2020 as supplied by Real Clear Politics and adjusts for the “Shy Voter” phenomenon and the fact that mail in voting will be a much larger percentage of ballots cast this year.

Shy Voter Effect
https://www.cloudresearch.com/resources/blog/election-2020-poll-respondent-honesty/

Nate Silver’s 538 (mainstream poll analytics) site admits that only one of their pollsters considers the “shy voter” to be a factor. The majority of mainstream pollsters discount the shy voter effect because they assume that the shy Biden voters will cancel out the shy Trump voters. While this may be true in aggregate nationally, it fails to take into account the fact that the effect may vary by locality based on societal norms. For example, in a Republican leaning state, the Biden voter may perceive a social stigma with revealing his/her vote. The analysis assumes that there is a greater stigma and perceived social and economic repercussions associated with admitting voting for Trump than for Biden in a competitive or Democrat leaning state. This appears to be validated by the CloudResearch study that indicates that 11.7% of Republicans, 10.5% of Independents, and 5.4% of Democrats are reluctant to reveal their true voting intentions. For this analysis, the shy voter average is R=12%, I=11%, D=5%, and is weighted depending on the Real Clear Politics classification of a given state’s political leaning: a Democrat is more likely to be “shy” in a Republican state and vice versa.

Rejection of Mail In Ballots Effect https://fivethirtyeight.com/features/why-rejected-ballots-could-be-a-big-problem-in-2020/

Due to COVID, mail-in voting will be a much greater proportion of the votes cast. A Pew research study in June indicated that 83% of Democrats versus 44% of Republicans are considering mailing in their votes. Data on the rejection of mail in ballots due to voter error is sparse, but in Philadelphia rejections due to the “naked ballot” issue was 6% in the primaries. For this analysis, it is assumed that Rejection Rate = 5%, D=40%, R=20%. Given that Democrats are more likely to vote by mail, a greater percentage of their ballots could be impacted.

Turnout
Turnout is assumed to be meaningfully higher than 2016, approximately 143.5 million ballots cast.

Electoral College Forecast
Democrat = 184
Republican = 354
Probability of Trump victory under this scenario 85%