1/When #attrition looks like this in an endline survey in an RCT: Baseline disadvantage reduces endline participation, but only in control (here higher values of Kessler 6 = worse) Why? It seems like successful interventions can change who remains observable!

"Crime - boy I don't know." Except we kind of *do* know some things. Coming Fall/Winter 2026.

This is what I will be reading/focusing on: www.whitehouse.gov/wp-content/u...

Like this in particular:

So many academic papers written on what Fab 5 Freddy sums up in three sentences (from the incredible "Everybody's Fly")

New WP. #medsky #econsky

2/ There are A LOT of DID studies in clinical journals now. #Event #studies are increasingly popular. This is a good thing! But it has become easy to plug-and-play without much thinking. "Let's take policy X to look at outcome Y using DID!" With enough X and Y you have a career's worth of papers.

Basically, would you like a graph like that on the left (theory driven sample restrictions and controls) or on the right (data driven weighted controls seeking to "impose" pre-trends)? (Note - in paper, I will isolate main estimate in its own figure. These would be for robustness checks)

1/ For my #book project (which I just turned in!), I looked at how #health varied by #income in states where people were more or less likely to be #stuck (i.e., experience disadvantage in multiple domains). In #stuck places, everyone's health is worse off.

1/ New #paper by Chuck Manski, @johnmullahy.bsky.social, and me in the Journal of Business and Economic Statistics (@amstatnews.bsky.social).

3/ We believe we have a new #measure of well-being, one that combines multiple aspects of American life. We will be deploying a new survey in a few months to track stuckness + writing an academic paper to show the way forward. In the meantime, please read the words of our survey #respondents:

2/ Our report develops a measure of #stuckness and shows how it is strongly correlated with #health -- even after accounting for a wide range of well-being measures. It shows how stuckness crosses class-boundaries. It shows us the myriad ways people can get stuck.

Teaser: The % of working age Americans from our survey who strongly agree with the statement: "I currently feel stuck in life."

Un-American.

"James River Dream" - sketch using oil painting setting on an iPad app.

🚨@oppforhealthlab.bsky.social has a new website!!🚨 Fully redesigned, and full of helpful summaries of our work and associated resources. Our mission: "Advancing health by expanding opportunity and hope." Find out how: opportunityforhealth.org

The #income-#health gradient among working-age adults in the United States in areas with high vs. low economic opportunity @oppforhealthlab.bsky.social

"Holiday card" - study and final.

"On the Brooklyn Bridge" by my 7 year old and me.

Salience of sunk costs meets ChatGPT. (Paper in question: www.damianclarke.net/research/pap...)

Rather than write the first thing up and get into into a clinical journal, we decided to chase the oddities. We ended up with a rich interesting result (which many people seem to hate, but thats a story for later): arxiv.org/abs/2510.26857

My entire vibe for 2026 will be to show that #great things are still #possible when it comes to improving #population #health. My message to #researchers is to work on big substantive questions and deep science. There is always someone out there looking for #evidence to make things #better.

A discussant who knew the way to my heart. #ASSA2026

And sometimes the relative lack of data to buttress part of your argument means you have to turn to analagous situations to make your point. E.g. our paper is about pneumonia, but we needed to think about rheumatic fever a bit:

It's neat how you can leverage qualitative historical analysis Our ppr: ⬆️ early exposure to #antibiotics, ⬇️ pneumonia in infancy, ⬆️ long-run economic outcomes. Historical record: pharmacists shaped access to antibiotics Our data: long-run impacts ⬆️ with baseline shr of pharmacists per cap.

IT LIVES! First draft: 2011 This draft: www.dropbox.com/scl/fi/xh3sg... In between: 12 years of being under review (of which half are on us). +1 coauthor. For me personally: a few moves, a few jobs, a few kids.

Interesting that it goes up to 5 years but not more, given the separation of survival curves that we saw in an older paper (which we are working on updating because of a story I can tell you over a scotch.)

From my #book #project on feeling stuck in America. The share of Billboard top 100 songs with the term "stuck":

What happens to self-reported race/ethnicity when you change the way you ask it? There is some discordance. And it is consequential for estimating quantities of interest, like disease risks by population group. See our ppr: arxiv.org/pdf/2501.023... #econsky #medsky @johnmullahy.bsky.social

6/ By selectively removing exposed observations experiencing the outcome, the critique introduced a collider bias in the opposite direction. The more observations you remove, the worse it is. We illustrated this in a simulation (a 2x2 experience based on estimates from our paper):