It takes a brave publication these days to claim an authoritative prediction for November’s US election.
The Economist has combined polling and demographic and economic data into a model developed with Columbia University political scientists Andrew Gelman and Merlin Hedemann.
And it’s currently predicting a landslide victory for Joe Biden.
The model is currently predicting 333 electoral college votes for Biden to Donald Trump’s 205.
As the site concedes, Hillary Clinton consistently outpolled Trump in 2016, before losing the electoral college and winning the popular vote by a much smaller margin than predicted.
The Economist thus argues that what the model produces of value is the “estimate of uncertainty around that prediction … if the model’s best guess is that Mr Biden will win 51.6327% of the vote, then there is probably a decent chance that he ends up between 51% and 52%, a very good chance he finishes between 49% and 54%, and virtually no chance that he secures less than 46% or more than 57%.”
Given polling’s awful record — across several countries — recently, this is a project well worth keeping an eye on.
Crikey is committed to hosting lively discussions. Help us keep the conversation useful, interesting and welcoming. We aim to publish comments quickly in the interest of promoting robust conversation, but we’re a small team and we deploy filters to protect against legal risk. Occasionally your comment may be held up while we review, but we’re working as fast as we can to keep the conversation rolling.
The Crikey comment section is members-only content. Please subscribe to leave a comment.
The Crikey comment section is members-only content. Please login to leave a comment.