Election models are everywhere. Nate Silver (on Substack!) has one. 538 has one. The Economist has one. Decision Desk HQ has one.
But, what, exactly are they? And do they work?
The simplest explanation is that election modeling is an attempt to combine both poll data and and other underlying data into one big prediction. Like economic indicators. Fundraising. Etc.
It’s a sort of hybrid between straight polls and the “keys” of American University professor Allan Lichtman. An attempt to say that neither total reliance on polling nor ignoring polling totally works.
For me, it’s another data point to consider when thinking about who will win the election. And, if I am being honest, it’s probably the election tool I trust the most — since it combines both polling and lots of other data points that have historically influenced elections.
Most of the election models right now are producing a similar result: A narrow win for Kamala Harris.
Decision Desk HQ released their revamped — post-Biden — model this morning and they have Harris with a 55% chance of winning and predict her with 276 electoral votes.
Here are the Decision Desk model odds of Harris and Donald Trump winning each swing state:
The Decision Desk prediction is similar to where 538’s model sees the race: A 59% chance of Harris winning. (Sidebar: There have been significant questions raised about the new 538 model. It’s worth reading about them here.)
The Economist model has the race more of a 50-50 proposition.
The broader point here — from the models at least — is that this is a very close race. Harris may have a slight edge right now but it is slight.
Much more breakdown of models and what they mean in today’s edition of The Morning.
This is ALWAYS a FREE post. But if you’d like to support my independent journalism — and a get a ton of perks along the way including a 30-minute, one-on-one Zoom with me — become a paid subscriber today! It’s $6 a month or $60 for the year.
Share this post