David Shor
@davidshor.bsky.social
7.7K followers 84 following 190 posts
Head of Data Science at Blue Rose Research, based in NYC, originally from Miami. I try to elect Democrats. Views are my own. he/him🌹
Posts Media Videos Starter Packs
davidshor.bsky.social
Both graphs are restricted non-white people
davidshor.bsky.social
Hard to know exactly - some combination of TikTok and Covid changing people’s news consumption to be way more online (discords, WhatsApp groups, etc).
davidshor.bsky.social
I think it’s probably related to the decline among immigrants we’re seeing everywhere too - there’s been a really fundamental change in the information ecosystem of people who don’t follow politics closely bsky.app/profile/davi...
davidshor.bsky.social
Non-white voters are rapidly trending toward right wing parties in Canada, the UK, and potentially other countries as well as the US.

www.washingtonpost.com/business/202...
davidshor.bsky.social
This Norway paper is pretty good for showing whats going on - when I looked through old Canada polling I saw a similar story too.

The real outlier is Germany, where its young women moving left rather than young men moving right.

bsky.app/profile/rube...
rubenmathisen.bsky.social
The pattern is the same whether we measure polarization through voting or left-right self-placement. Boys have shifted to the right and girls to the left.
davidshor.bsky.social
If you add all the left and right wing parties together the age trend is pretty clear as day
davidshor.bsky.social
If you look at the California high school mock election data (take with the biggest grain of salt you can imagine), you see that the cohort of kids who would have millennial parents seem to be way more Democratic
davidshor.bsky.social
I don’t think this is the main story, but I think some of it is just that our parents were boomers and their parents were gen x.

People under 25 are 7% less likely to say their parents were Democrats than 30 year olds.
davidshor.bsky.social
I think the Gelman model is empirically right but doesn't explain as large a fraction of cohort effects as people think.

Boomers were left wing in every country, Gen X was right wing in every country, millennials were left wing in every country, and now Zoomers are trending right everywhere too.
davidshor.bsky.social
In general the two biggest predictors of partisanship right now are socioeconomic status and political engagement.

Those two things are also the biggest predictors of answering a political survey. It's a huge issue.
davidshor.bsky.social
This is a good thread that someone on my team did that makes the case with public data bsky.app/profile/davi...
davidshor.bsky.social
1) It's improper to analyze party registration data without taking into account registration date.

Analyzing voter file data both cross-sectionally and within-person, it's clear that non-voters have become much more Republican in recent years relative to voters
davidshor.bsky.social
The "Harris did better than polls suggested among old voters and worse with young voters" comes through pretty clearly if you look at precinct data.

The basic issue is that the partisan gap between highly engaged and highly disengaged young voters went up *a lot* from 2020 to 2024
davidshor.bsky.social
It’s the voter demographic section here from the post election mega YouGov poll weighted to the election results. This is comparing Labour+LibDem+Green+SNP vs Tory + Brexit. en.m.wikipedia.org/wiki/2024_Un...
2024 United Kingdom general election - Wikipedia
en.m.wikipedia.org
davidshor.bsky.social
Shows up in polling data too bsky.app/profile/davi...
davidshor.bsky.social
I did an all left vs all right comparison and saw the same thing in the UK - I think the same is true in Canada from eyeballing but haven't checked too closely
Reposted by David Shor
Video of this year's Harvard Hutchins Center forum: What went wrong? What's next? Thoughtful and provocative conversation. Jonathan Capehart, Astead Herndon, Sherrilyn Ifill, David Shor, Laurence Tribe, Charlene Hunter-Gault www.youtube.com/watch?v=ld4k...
WATCH LIVE: The 2025 Hutchins Forum
YouTube video by PBS NewsHour
www.youtube.com
davidshor.bsky.social
Shows up in individual level data too!

bsky.app/profile/davi...
davidshor.bsky.social
I did an all left vs all right comparison and saw the same thing in the UK - I think the same is true in Canada from eyeballing but haven't checked too closely
davidshor.bsky.social
My tweet did not talk about polls at all.
davidshor.bsky.social
I did an all left vs all right comparison and saw the same thing in the UK - I think the same is true in Canada from eyeballing but haven't checked too closely
davidshor.bsky.social
Political science as a discipline has a lot to answer for: It has been clear this has been happening for the last decade but many prominent academics ignored it because of a combo of shoddy empirical work and it not fitting neatly into then trendy racial-conflict centric theories of politics
davidshor.bsky.social
Non-white voters are rapidly trending toward right wing parties in Canada, the UK, and potentially other countries as well as the US.

www.washingtonpost.com/business/202...
davidshor.bsky.social
One fun graph for pride: the single strongest demographic predictor I’ve found so far for predicting Cuomo vote share is what fraction of registered voters in the precinct identify as heterosexual
davidshor.bsky.social
The mayor race on Tuesday saw massive turnout differentials - turnout in @zohrankmamdani.bsky.social’sbstrongest precincts was up ~30% while turnout in Cuomo’s best districts were ~ flat.

These are big differences - essentially two different elections that happened to fall on the same day.
davidshor.bsky.social
This was the point of this tweet, right.
davidshor.bsky.social
Traditionally there's been a quant/qual split on ad length in Democratic politics, with [usually more moderate] quants pushing for shorter 15 second ads and [usually more progressive] qual folks pushing for 60 second ads.

Great to see @zohrankmamdani.bsky.social sided with the nerds!