Interesting result on the effect of presenting a "No opinion" option on this Washington Post-Ipsos poll www.washingtonpost.com/politics/202...

Dr. Sunghee Lee (@sungheelee.bsky.social‬) featured in this NYT article examining the consequences of deep NIH cuts on health disparities research. Her work on dementia in underrepresented populations was defunded despite its critical value. Read more: myumi.ch/W6dZb #HealthEquity #SocialResearch

I got asked to take uncertainty intervals off a chart because the lower tail showed potentially a 0% lift and like nah dawg I will not be doing that. You can take uncertainty out of the chart but you can't take it out of the system --- if you wanna delude yourself, that's between you and god

🚨 All data nerds 🤓: ANES 2024 full release is now available!

🚨 For all data nerds 🤓 out there: incoming announcement for a new data release!

THREAD: There are a lot of unanswered questions 12 hours after President Trump — who, according to the Constitution, has no final authority over the census — released a social media post calling for unprecedented changes to how the government produces data that forms the foundation of U.S. democracy

Even if you dislike the results, don't kill the polls!

I *really* hope ACS never gets to Trump's attention over the next few years...

AAPOR Statement on the Removal of BLS Commissioner McEntarfer aapor.org/statements/a...

Past vote weight-ers, rejoice!!! 🎉 www.pewresearch.org/decoded/2025...

Who in their right mind would poll Mamdani in a Democratic primary for president?! Oh, right, of course! 🤭

Trump is threatening Brazil with 50% tariffs for no reason whatsoever -- It's Brazil that has a trade deficit with the US...

This has to be a joke, right?

This week's E is for Epi newsletter: what the NYTimes' hit piece on Zohran Mamdani can tell us about the importance of good survey design. Read it here: open.substack.com/pub/epiellie...

Since I reposted this yesterday, I feel obliged to also post this important correction to the charts:

Turns out it's much harder to poll when a large part of the electorate is new and/or young voters

@badsurveyq.bsky.social‬ is in the house! Make sure to tag them for all of your favorite (or least favorite) bad survey questions!

This is brilliant! And for photo/image sharing, bring back the fax machine! Also, the survey research in me thinks: "Yay, make RDD great again!" 🥳

I love how journalists and writers are getting an independent space to freely express themselves, but this Substack model is quickly becoming unsustainable for consumers of this type of media (even more so than the streaming model). I hope we start to see some bundling options at some point...

This interview is a great lesson about liberalism and it brings hope that there is an alternative path in politics through grace and empathy.

Reminder that if these polls are using online opt-in panels, results for young people can be misleading. www.pewresearch.org/short-reads/...

One of my son's toys is the survivorship bias aircraft 😍🤓

Strength In Numbers partnered with Verasight to ask a sample of 1,000 Americans to describe the first few months of the Trump presidency in their own words. Republicans used words like "good" and "great." Democrats wrote "terrible" and "treason." More: www.gelliottmorris.com/p/what-voter...

With that said, here would be my advice (that nobody asked for) to pollsters running pre-election polls in the US: if you are not already, try to invest more in election modelling, especially using the voter file. +

We are excited to team up with @um-psc.bsky.social to create an International Hub for the important survey and popul. research that is done in the social sciences. Bill Axinn, founding Director, brings many years of experience to the position. isr.umich.edu/news-events/... @um-src.bsky.social

The issue of evaluating pollsters based on their, so-called, accuracy in pre-election polls is that this is turning much more in a matter of election modelling than survey methods. If you have very good models, you can get pretty decent estimates even with a crappy methodology. 🧵 +

According to the June The Economist/YouGov Poll, 30% of Americans think immigrants shouldn't have the same right to express political views online as U.S. citizens. About half of Republicans/Trump voters shares that view. So much for freedom of speech! d3nkl3psvxxpe9.cloudfront.net/documents/ec...

Not exactly related, but this reminded me of this excellent "experiment" the @upshot.nytimes.com ran back in 2016: www.nytimes.com/interactive/...

(Bayesians at least update priors based on data in mechanically consistent manner, polymarket is pure vibes)

This is once again to show that betting markets are not oracles or even forecasters per se, they just react to data. Not necessarily a bad thing, that's essentially what Bayesians do, they update priors based on data. But that's all it is.

Very, very concerning

Fine, as a sampling statistician I'll go ahead and say it: river sampling, especially recruited online, is not a serious sampling methodology and there is absolutely no reason to take its results at face value. Especially when the pollster is very shady about its methods.

Having the acting director of the US Census Bureau saying that they "might not prioritize enhancements and those sorts of things" due to loss of personnel is quite concerning, especially in an era that the field has to constantly change and adapt to deal with societal and technological challenges.

Just ran a Monte Carlo simulation to check the performance of weighting by recalled vote on correlation and regression coefficients, under *non-ignorable nonresponse*, and results seem quite promising, despite the non-ignorability!

This footnote on the NYT analysis that has been ciculing around tells exactly the pain Connecticut has been causing on those of us using national county-level data 😅😫😭

We desperately need to bring @BadSurveyQ to this plataform...

This is completely infuriating. People work extremely hard and changed their entire lives to move to a whole new country, away from their family and friends, to pursue the highest level of education just to have everything they worked for suddenly taken away from them for no good reason at all...

Cutting high quality surveys conducted by the US Census Bureau on important topics such as alcohol use and internet access? I don't even have words to describe how incredibly frustrated and furious I am with all this nonsense...

I understand the necessity of some RV polling from a sample frame perspective, but I likewise find the use of LVs a little baffling -- it adds a whole bunch of flux/uncertainty and sort of undercuts the idea that what people think matters, regardless of whether they vote in specific elections.

If you want to have a honest conversation about polling error and transparency this thread and article are a must read! Highly recommended for anyone interested in polling methodology/transparency. It's for pieces like this that it's really worth subscribing to the newsletter!

“Samples are not given. They must be selected, assigned or captured.” - Leslie Kish Source: projecteuclid.org/journals/sta...

Half of the United States population lives in the largest 144 counties.

Yeah, go figure, non-ignorable nonresponse is a really hard problem! There is not a single good solution out of it. There are some approaches, such as using pattern-mixture or Heckman selection model, but they rely on *very* strong assumptions and on having strong correlates to vote choice. +

#AAPOR25 is on the books! Now, time to enjoy some of St. Louis has to offer! 🍖 Awesome to finally meet in person some of you folks here! And always great to see again my people! Already looking forward to seeing everyone next year in the City of Angels! 😎

📈 We're nearing the home stretch at #AAPOR2025, but we have a full afternoon of sessions ahead of us. Gather your strength and follow along! To see a full schedule of SRC researchers' presentations, visit: myumi.ch/8qqye

📈 @rnishimura & Paul Burton, w/Matthew DeBell, used data from the @electionstudies.bsky.social to discuss the principles of sample stratification design and its role in objectives often seen in national United States election surveys. They will discuss their findings at #AAPOR25 myumi.ch/bVEx6

Important update on Wikipedia entry for the Likert scale!

Come over and say "Hi!" at #AAPOR Poster Session III at 1PM. I'll be showing the approach we used to stratify the Primary Sampling Units of the Fresh Face-to-Face 2024 American National Election Studies. #AAPOR25 aapor.confex.com/aapor/2025/m...

It's an early morning #AAPOR session, but never too early to watch a @bradytwest.bsky.social talk! Today he's discussing important considerations when looking for reference probability samples to be used for inferences on non-probability samples -- joint work with @randridge.bsky.social #AAPOR25

From Wikipedia entry on the Likert scale 😅