Geek’s Guide to the Election 2024
The 2024 PEC Geek’s Guide to the Election is here. Let’s use this entry at the Fixing Bugs in Democracy site for an open thread. (No subscription required, it’s free.)
Liveblog below.
November 5, 8:52pm: Remember, reporting is uneven. Faster-reporting: Georgia, North Carolina, Michigan, Virginia, Florida, Ohio, and Colorado. Slower: Wisconsin, Pennsylvania, Nevada.
10:39pm: Generally speaking, Presidential and Senate races are coming in slightly more Republican-leaning than pre-election polls.
Trump will win Iowa by 12 points. Looks like that Selzer poll was a fairly big miss. Seems to be a win for stratification of polls thought.
Generally, I am impressed with the technical quality of the New York Times predictive models.
11:47pm: In 7 House races, the margins are running about 3 points more favorable to Republicans than pre-election polls. Two races in the Geek’s Guide (PA-08, IA-03) are going narrowly Republican at the moment. This is still a small sample, and it is not clear yet which party will end up in control. But it does argue against Hopium-level overperformance for Democrats.
November 6, 12:44am: Looking at House returns in 29 races that are mostly done counting, it appears that so far Democrats might gain 5 seats (NE-2, NY-4, NY-19, NY-22, VA-2) and Republicans might gain 2 seats (PA-7 and PA-8). Control of the House is still up in the air. There is a chance the House will be even more closely divided than it is now.
The Presidential race still depends on Pennsylvania, Michigan, Wisconsin, Arizona, and Nevada. Things are headed toward Trump, but it will take a day or two to play out.
Good night.
For an open thread, please go over to Fixing Bugs In Democracy.
I suppose it could be because results aren’t yet finalized, but looking at the national popular vote, it isn’t so much that there was a shift from Democrats to Republicans for 2020 to 2024, because the Republicans got about the same number of votes in both years. The difference between the two years is that, numerically, 1 of 7 Biden voters didn’t vote for Harris. Seemingly this has nothing to do with the country embracing Trump, at least not more than it did four years ago. Instead it seems to have much more to do with folks who normally would have voted for the Democrat instead abstaining / protesting. (FWIW, third-party numbers didn’t change much from 2020 to 2024.)
Do you have state by state, or county by county, or precinct by precinct data that could do this same analysis at a finer scale? For example, did Pennsylvanians really move towards Republicans, or is it better explained by Republicans getting their usual numbers but usual Democrats were instead abstaining or protesting?
What is the total number of precincts in PA? What is the average margin, Trump vs Harris, the number of votes per precinct, in 2024? The same question for Arizona, Georgia and Wisconsin? How does that compare to 2020? I have read estimates that it was 9 votes per precinct in 2020 in PA.