Tips on how to construct a polls-only objective Bayesian model that goes from a state polling result in probability of winning the state
With the presidential election approaching, a matter I, and I expect many others have, is does a candidate’s polling in a state translates to their probability of winning the state.
On this blog post, I need to explore the query using objective Bayesian inference ([3]) and election results from 2016 and 2020. The goal might be to construct an easy polls-only model that takes a candidate’s state polling lead and produces a posterior distribution for the probability of the candidate winning the state
where the posterior distribution measures our belief in how predictive polls are.
For the model, I’ll use logistic regression with a single unknown weight variable, w:
Taking the 2020 and 2016 elections as observations and using an acceptable prior, π, we will then produce a posterior distribution for the unknown weight
where
