David Phillips
1.4K posts
David Phillips
@PhillipsEcon
Mostly on impact evaluation and poverty in the US. Caution: economist on Zoom may be taller than he appears.
University of Notre Dame เข้าร่วม Ocak 2016
543 กำลังติดตาม825 ผู้ติดตาม
David Phillips รีทวีตแล้ว
Feels like there’s a study to be done here on self report bias and measurement error in height #econtwitter
Joe Pompliano@JoePompliano
We’re going to see a lot of height changes this year because, with the rollout of ABS, MLB is now measuring each player's height down to the millimeter. The process is super strict — no shoes, no hats, knees exposed, back against the wall — and to account for potential shrinkage throughout the day, MLB is even requiring all its teams to take measurements between 10 am and 12 pm local time on their appointed day.
English
David Phillips รีทวีตแล้ว
Through a rigorous evaluation with Bridges to Success, LEO researchers @PhillipsEcon and William Evans found adult mentorship increases the likelihood of employment. Their paper is forthcoming in @JPubEcon. To learn more about Bridges to Success, click the link in our bio.
English
David Phillips รีทวีตแล้ว
Great for Georgetown. Ominous sign for the World Bank's research department amid lots of other rumors.
Georgetown Center for Economic Research@GCER_Georgetown
🎉 Exciting news! @dmckenzie001 has accepted the offer to join @GCER_Georgetown as the new Edmond D. Villani Chair Professor of Economics! We can't wait to welcome him to the Hilltop! 🙌 sites.google.com/site/decrgdmck…
English
@andy_garin @causalinf This is a better way of saying what I was trying to say. 😂
English
@causalinf @PhillipsEcon I’m mainly saying that if one does not buy into matching strategies, it’s hard to imagine a case where one would believe causal estimates from DD with covariates.
It’s hard to think of cases where relaxing from CIA to conditional PT makes for convincing identification
English
David Phillips รีทวีตแล้ว
Results from my Claude Code audit of six Callaway and Sant'Anna packages (two in python, two in R, two in Stata). Same specification, same dataset, same covariates, same estimator, almost never do they agree.
open.substack.com/pub/causalinf/…
English
@causalinf If it has been true in the past. If there's a long pre-period showing difference in levels but no difference in trends, then I think the traditional assumption (without Xs) is reasonable. If it is not true, I am skeptical that conditioning on Xs is fixing anything.
English
@PhillipsEcon We discuss this in section 4.2 of our JEL, “diff in diff: a practitioners guide”. When groups differ wildly on X, you’d need 1st diff mean y(0) identical to them anyway. We rarely would say simple comparison in means identify causal effects so why here?
English
@paulnovosad @alz_zyd_ 100%. For the same reason, totally worth going back and reading the assigned reading from high school as an adult. HS books are genuinely great for adults.
(And econ is the same. Gen-ed economics would be better consumed by 30 year-olds than 18 year-olds in most cases.)
English
@alz_zyd_ My theory is English curricula are based on how adults think about kids.
Had any teen ever read Catcher in the Rye or A Separate Peace and felt like it really hit them where they were at?
These books resonate with teachers and educators and are useless for kids.
English
Nobody reads these books because they suck. You don't need to try to get smart kids to read/watch Les Miserables or Crime and Punishment or Catch-22 or Dune because they're actually good books
Boze Herrington, Library Owl 😴🧙♀️@SketchesbyBoze
I hate that schools everywhere are closing their humanities departments and shuttering their physical libraries. How are students going to fall in love with Chaucer, Shakespeare, Austen if they never encounter them? We are rapidly entering a new dark age.
English
@causalinf (b) requires some strong institutional justification. Plus, as you show, there's a lot of additional layers of assumptions needed in any actual implementation.
English
@causalinf I don't think I follow...two possible assumptions: (a) parallel trends hold unconditionally or (b) parallel trends hold conditional on X. If someone claims (a) false but (b) true, my immediate reaction is skepticism.
English
David Phillips รีทวีตแล้ว
Our CEO, Jennifer Loving, will take the stage at the TED 2026 Conference in Vancouver next month. We're excited for her to share how Right at Home is leading innovative efforts to prevent homelessness, demonstrating how communities across the United States can ensure more people remain safely and stably housed.
This work is made possible in part by @TheAudaciousPrj, whose partnership in the Right at Home initiative is helping drive thoughtful, people-centered solutions to stop homelessness before it starts. conferences.ted.com/ted2026/speake…
English
David Phillips รีทวีตแล้ว
Preventing homelessness is more effective and far less costly than responding after crisis.
A new initiative from @RightatHomeUSA, guided by @LEOatND, will expand what works to keep families housed in 10 pilot locations nationwide: go.nd.edu/4bcc4a
English
David Phillips รีทวีตแล้ว
Right at Home is built on a simple, evidence-backed idea: When people receive timely support before losing their housing, more people remain housed.
Learn more about how #HomelessnessPrevention works: rightathomeusa.org
English
David Phillips รีทวีตแล้ว
Today we launch Right at Home, a national initiative to stop homelessness before it starts. Led by Destination: Home, this work is expanding on the country's first proven homelessness prevention initiative so that more communities can keep more people safely housed.
Learn more: bit.ly/RAHLaunch
@RightatHomeUSA #RightatHomeUSA #StopHomelessnessBeforeItStarts #HomelessnessPrevention
English
David Phillips รีทวีตแล้ว
I review a lot of papers at JDE and JPubE which share the characteristics of these AI papers.
Thorough data analysis of some comparative statics, as much robustness as you want, but not very strong on contextual understanding and using a design that doesn’t actually answer the question you think it does. I recommend reject on these, but it all gets published *somewhere* and often pretty decently.
English
David Phillips รีทวีตแล้ว
Nobody will ever beat the 1992 Olympic Torch lighting
English
David Phillips รีทวีตแล้ว
the Notre Damest thing I can think of. amazing.
Notre Dame Photos@PhotosND
Last night, more than 1,600 people gathered for Mass at our student-built ice chapel. ❄️ #LoveThee
English
David Phillips รีทวีตแล้ว
Incredible and inevitable—
Police chief says ICE agents violated his own officers' civil rights while off-duty.
A female officer was boxed in by ICE agents who demanded her papers, pulled their guns on her, and knocked her phone down as she tried to film.
English
Honestly, I stopped following IU a long while ago, but that was a fun one to watch on behalf of the long line of diehard IU fans from the Phillips family.
English