Political data @GSG. Former academic, campaign staff. UMich PhD (go blue!) he/him/his. politics, surveys, stats, soccer, D&D, dogs. Tweets are my views only
Joined April 2010
- Tweets18.2K
- Following4.8K
- Followers1.9K
- Likes70.9K
Alex Von Hagen-Jamar retweeted
Election results show how Americans voted, not why. In Poynter, NORC’s Jennifer Benz and Tom Rosenstiel explain how the Civic Health & Elections Project, a survey of 85,000+ registered voters in all 50 states, can help journalists fill that gap.
Read the full article: go.norc.org/4BpUexK
ALT A quote from an op-ed in Poynter that reads “A vote is not an opinion. It is a choice, often a complicated and begrudging one. The reasons behind it deserve more than a single crosstab. Doing better requires data that do not currently exist in usable form.” Underneath are two headshots of the authors, the first is identified as Jennifer Benz, Senior Vice President of Public Affairs & Media Research. The second is Tom Rosentiel, Senior Fellow of Public Affairs & Media Research. A source citation is at the bottom. The NORC at the University of Chicago logo is at the top.
Alex Von Hagen-Jamar retweeted
To add additional complication to this question:
1) undecideds turn out at lower rates than average (~1/3rd did not vote in some 2022 data we had, vs ~1/5th of full sample).
2) in reality, it's not just undecideds that move late. defectors can be an equally impt factor.
Alex Von Hagen-Jamar retweeted
LOTS of long term solutions need to be enacted here but in the short term, I think it's helpful for campaigns to know that ChatGPT users are more likely to be young men, and also skew Asian, Black + Hispanic.
Alex Von Hagen-Jamar retweeted
As my colleague @alexvhj astutely pointed out the history of polling errors in deep red (& deep blue) states also consistent with undecideds in these states breaking to partisanship. A question this year how it plays out with direr straits for POTUS party than seen in yrs.
Alex Von Hagen-Jamar retweeted
Very excited about some new research from Caucus AI today – we found nearly 1 in 5 ChatGPT answers about the midterms is citing an AI-generated voter guide. These sites are mostly anonymous, mostly registered in the last year, and they don’t rank highly in traditional search.
Alex Von Hagen-Jamar retweeted
New from Caucus AI: 18% of ChatGPT’s September answers about 2026 candidates cited AI-written voter guides from little-known, mostly anonymous sites, up from 7% in August.
Should this concern us? It definitely raises a lot of questions about reach and transparency in AI search.
Alex Von Hagen-Jamar retweeted
Here's where Latino polling stands in October: most estimates land above 2024. But there is disagreement over whether Dem support will fall in the 2020/2022 vs. 2016/2018 tiers. @EquisResearch pegs support around '20 right now (due largely to recall weight).
Alex Von Hagen-Jamar retweeted
Polling always survives this stuff bc it's unique
There is no other reliable way to
- ask an arbitrary question
- of giant, complicated population
- get a representative answer
- relatively fast, any time
Survey data isn't perfect. But, for a lot of purposes, it's the only game in town!
Alex Von Hagen-Jamar retweeted
There are way too many otherwise knowledgeable people on this site asking questions like "how does polling recover".
Pollsters didn't do anything wrong! A bunch of people did fraud and got caught by pollsters!
Alex Von Hagen-Jamar retweeted
Our investigation into allegations of data fabrication at the firm SoCal Strategies, and an alleged association with the admittedly fraudulent firm Wedgewood Polls, has resulted in the pollster wiping itself from the face of the earth.
Full story at 50+1: blog.fiftyplusone.news/p/mor…
NEWS: The website and social media accounts for pollster SoCal Strategies have been taken offline after FiftyPlusOne reached out to the firm Tuesday to verify data quality and potential ties to Wedgewood Polls, which admitted to fabricating data this week. blog.fiftyplusone.news/p/mor…
Alex Von Hagen-Jamar retweeted
Ilying is a poor strategy in politics because it's fragile. It forces you into dilemmas about whether to maintain the lie or respond to reality. Ultimately true things can be a huge pain in the ass in the moment, but in the long term are much more durable.
Alex Von Hagen-Jamar retweeted
Absolutely not. We don’t make spending or strategic decisions based on external polling.
🚨Wedgewood Polls released a survey on October 14 showing the Mississippi Senate race at Republican +4. The fabricated results prompted Senate Majority PAC, aligned with Chuck Schumer, to announce a $1.2 million television ad investment on October 20, described as its first major spending in the race after prior zero investment.
Wedgewood later stated that the poll data was fabricated. The admission followed an investigation into the firm’s methodology and data integrity.
Follow: @MorseReport
Alex Von Hagen-Jamar retweeted
Seeing some *discourse* about the feasibility of certain election results.
While there’s some merit to it, periodic reminder to folks that the range of possible outcomes is WAY larger than you may think.
For reference, in 2018, D+6.5 MI Senate and D+6.8 OH Senate.
Alex Von Hagen-Jamar retweeted
Kudos to FiftyPlusOne for catching this. It's going to be a big problem for Democrats in the 2028 primary -- huge incentives for bad actors to manufacture momentum or move betting markets.
We need to align on a way to prove that a poll talked to real voters.
We were investigating Wedgewood for faking their data and emailed them for comment this morning, ahead of publishing the findings tomorrow. They are trying to cover up their fraud with the "we were doing a social experiment" excuse.
Full story here: blog.fiftyplusone.news/p/fif…
Alex Von Hagen-Jamar retweeted
For a long time I’ve wanted to write an article showing the huge variation that you can get from LV models.
We do just that with our Decision Desk poll released today.
Alex Von Hagen-Jamar retweeted
2022 was weird: Democrats over performed—as polls indicated—in the most competitive states/seats. And—as polls indicated—got pummeled in bluer and redder places. I think pundits…somewhat sympathetically…tended to see one as “the story” and conclude the other was wrong
This is intuitive if you expect late deciders to tend towards underlying partisanship.
We're in the last 3 weeks when polls get graded.
I built a tool that breaks down polling error in the last 2 midterm cycles in the last 3 weeks.
Nationally, the polls were pretty good, but red state polls were too blue and blue state polls too red. x.patrickruffini.com/polling…
