2020 Presidential Election Analysis

Kayla Manning, Harvard College

View the Project on GitHub kayla-manning/test

Welcome to my election analysis blog! I will post updates every Saturday. Feel free to reach out to me (kaylamanning@college.harvard.edu) if you have any comments or questions.

Final Prediction

This forecast uses polling numbers as of 3 PM EST on 11/1/2020.

Post-Election Model Reflection

How did my model compare to election results? In this post, I analyze my blog in the context of the election results, hypothesize reasons for its shortcomings, and brainstorm ways to improve it in future iterations.

Testing Narratives: Would Trump have won in a COVID-free world?

In an alternate universe without COVID-19, would Donald Trump have won the 2020 election? Inspired by previous research on the impact of the 1918 Spanish influenza outbreak on elections, this analysis takes a look at the association between state- and county-level COVID outbreaks and Trump’s 2020 vote share.


The following posts analyze various components of presidential elections and provide updated predictions for the 2020 race. Each post ultimately informed the modeling decisions in my final prediction:

1. Introduction

As the 2020 election approaches, this first blog post looks to past elections to explore electoral swing in battleground states between the 2012 and 2016 presidential elections.

2. Economy

Economic conditions predict elections fairly well, but how does that translate to 2020 in the midst of an economic crisis brought on by a global pandemic? Also, does the relationship between vote share and economic conditions vary between incumbent candidates and new candidates from the incumbent party? This post explores three different models that investigate these questions.

3. Polling

This installment incorporates poll numbers in updated predictions, explores the accuracy of polls across sources, and compares two approaches taken by professional forecasters.

4. Incumbency

More often than not, incumbent presidents win their re-election bids. Incumbent candidates enjoy a large set of structural advantages that collectively form the incumbency advantage. Can Trump’s status as an incumbent carry him to a victory in 2020?

5. Advertising

This analysis examines political advertising on various media platforms and takes a look at the relative advertising strategies of candidates in battleground states.

6. Voter Turnout

How do experts anticipate COVID-19 affecting voter turnout? This post considers the possible effects of COVID-19 on turnout, explores how different forecasters account for the resulting variability, and looks at models that simulate fluctuations in turnout.

7. The Impact of COVID-19

After a brief exploration of the relationship between COVID-19 and polls, an updated model provides state-by-state predictions.