I've reworked my chart from last week to account for how red/blue a state is. This is the new version, and while I'd like you to go to my S*bstack and read the article that explains how I got here (unpaywalled, with jokes!), I'll give you the gist of it here.
๐งต1/11
Post
Conversation
This is the chart from last week -- I'm using "Bernie-endorsed" as a proxy for "leftist" and measuring statewide performance relative to Kamala '24. This is "competitive races" according to the NYT, plus Ed Markey because I wanted all five Bernie-endorsed candidates in there.
Leftists complained that the chart doesn't recognize that it may be harder to cause a big blue shift in an already-blue state than in a red one. If true, that *could* help the Bernie cohort. I wish that they'd made their point in a less-snotty way, but I decided to look into it.
Using methods described in the article, I compared each state's blue shift from '24 to '26 with partisan lean, and it seems that how red/blue a state is *does* moderately affect the likelihood of a big blue shift. Lots of variance, but a moderate correlation (Pearson r = -.55).
But does this help the leftist candidates? The line of best fit suggests that each point of partisan lean correlates with 1.4 points of shift, so I gave each candidate a bonus or penalty based on that. A D+10 state gets a candidate a 14 bonus points, R+5 = 7 point penalty, etc.
FWIW, I thought this was heavy-handed because it treats a broad trend like an iron law. But I wanted to give the "state lean explains away the trend" argument every benefit of the doubt. I re-ran the chart and got this:
Redemption for Ed Markey! Though, he maybe shouldn't be included; he's the only candidate not in a competitive race, and is one of only two already a Senator (Ossoff is the other). Non-leftist candidates in competitive races still fare 5.6 points better on avg than the leftists.
And when I did this for ALL Senate races, it became clear that my adjustment was far too strong. Because this chart is gibberish; the worst performers are all in deep-red states, the best all in blue ones. The "lean" effect is so huge that candidate quality is being washed out.
There's no way that 8 candidates under-performed Kamala on the best night for Dems in 50 years. John Reed is not George Washington reincarnated. The competitive races are clustered in the middle, which made me think that I shouldn't have mixed Markey into the "competitive" group.
So, I weakened the effect: This chart is simply the mid-point between my first, "unadjusted" chart and my second, "super-adjusted" chart. Markey's gone. The normies here (including Osborn, who isn't a Dem) fare 10 points on avg better than the leftists. 11 points without Osborn.
Here's the full article, not paywalled. I publish 3-4 articles a week and two podcast episodes. And I typically have a lot more jokes and a lot less social science.
Access high-speed internet on the go with Starlink. Equipped with a built-in Wi-Fi router, Starlink Mini is ideal for travel, camping, boating, remote work & more.
Order online in minutes.
There are five senate candidates listed on Bernie Sanders' "endorsements" page, so those are the ones I highlighted. So-chosen because I needed a proxy for "leftists" (if *I* had chosen who's a leftist, people would have quibbled with that).
But Talarico was endorsed by Our Revolution. Why are we not counting that?
There are five senate candidates listed on Bernie Sanders' "endorsements" page, so those are the ones I highlighted. So-chosen because I needed a proxy for "leftists" (if *I* had chosen who's a leftist, people would have quibbled with that).
This is really good stuff. I do feel like the corporate wing of the party was on an all out attack against the left wing, while the left wing did a much better job of accepting defeat when it came in the primary and supporting the winner. Not sure how that would be calculated.