Best take I’ve seen on social media all week. I am a big fan of candy corn.

And while Weintraub is at the bottom of the graph, the Republicans at the top by comparison are more extreme on the scale.

www.nytimes.com/2024/06/10/u...

Insert caveat here about ecological inference! 😀

R-squared for the models.

What's up w/Washington County? It was in the top 4 until 1980. This is partly a measurement effect, but also it seems that the share of Francos also dropped, from about 17 percent in 1960 (where foreign born was the numerator) to ~10 percent in 2021 (where any French or Canadian ancestry was).

Comparing percents across different metrics is not ideal, so I'm trying ranks within year.

In more recent years, the ACS is great for this (at the town-level!). Before 1990, we need to rely on metrics that tend to focus on country of origin or language used at home.

The paper is "to be drafted." We've been working on getting the data in shape. The proj started as a focus on the Francos in ME politics. We needed a lot of town-level political data, and so I've spent the last year wrangling lots of files from the SecState that they dug out of the archives.

Thanks for the help @shennabellows.bsky.social.

Big thanks to Maine Secretary of State Shenna Bellows (and her office) for helping us get access to the town-level data over the decades. They scanned yearly pdf documents, and we moved them into Excel.

And because Bowdoin is in Brunswick, here is the change there:

Lots of Maine voters are "unenrolled" but this was inconsistently reported over the years. Today, about 30 percent of registered Maine voters are unenrolled with a party. This is down from high 30s in early 2000s.