Document Type
Student Research Paper
Date
Summer 2021
Academic Department
Mathematical Sciences
Faculty Advisor(s)
Dr. Sakar Sigdel
Abstract
The 2020 US Presidential Election was unique in many ways, and held a number of surprising results. Although the 2020 presidential election is over, there is insight to be gained by analyzing the factors that may influence a voter’s choice and building a model that could predict future presidential elections. Using the Democracy Fund Voter Study Group’s Nationscape public opinion survey, we’ve constructed a model using multiple logistic regression with L2 regularization for predicting which candidate a given respondent will vote for, taking into account how various factors such as age, ethnicity, education level, and orientation influence voter decisions. Our model correctly predicts Biden winning the popular vote as well as the electoral vote, which is consistent with the observed 2020 presidential election results.
Recommended Citation
Hanley, Aidan, "Predicting the 2020 US Presidential Election" (2021). Summer Scholarship, Creative Arts and Research Projects (SCARP). 60.
https://jayscholar.etown.edu/scarp/60
Notes
Scholarship, Creative Arts, and Research Project (SCARP)