Good news: Found the data I need. Bad news: It's text in a space-delimited wide format. BARF.

"Classified information is available upon reasonable request."

Khoa@khoavuumn.bsky.social · last yr.

If Hegseth had put classified info on Github, no one would have been able to access it.

If you are an author on the Many-Economists project but have not received any emails from me in the past month it means that we have outdated contact info for you. Please contact me immediately.

Reviewers: "We want to see X, Y, Z, M, N, P, Q and W additional analyses and robustness tests. Also, fit it under the page limit."

FWIW, I saw people (appropriately) fired for accidentally leaving small amounts of, realistically, "barely sensitive" classified information on military bases in Afghanistan. And nobody was surprised--they screwed up and everybody knew it.

This is such a cool project--I'm thrilled that I got to be a part of it. The big takeaway for me is that we need to do *less* research with *more* people on the team and try to do real code review and parallel analysis of our data.

Nick Huntington-Klein@nickchk.com · last yr.

After a long wait, the working paper for the Many-Economists Project: The Sources of Researcher Variation in Economics. We had 146 teams perform the same research three times, each time with less freedom. What source of freedom leads to different choices and results? papers.ssrn.com/sol3/papers....

INBOX ZERO. Don't any of you jerks email me. Oh, and happy new year.

Things got busy again and I didn't have time to do the puzzles the last few days, but I'm back! #AdventOfCode #RStats

Kids craft kits from like KiwiCo should be required to include no fewer than 10 pictures of the thing made by a child the target age advised by a parent who’s never seen it before. Don’t show me the perfect product made by an adult given 5 attempts…

Coming from a not-computer-sciency background, the Advent of Code is a fun challenge. I don't ever have time to complete the whole thing, but I'm keeping up with 2024 so far. And since most of my work is in #RStats so is my Advent of Code :-) #AdventOfCode adventofcode.com/2024/day/2

If you're planning Spring classes and Technical Debt would be relevant, my case on the Southwest Airlines meltdown is now available from Harvard Business Publishing. hbsp.harvard.edu/product/W365...

Reminder that the probably easiest way to import the citation for an #Rstats package to @zotero.bsky.social is probably to 1. run citation("package_name") in R 2. copy the BibTeX entry 3. go to #Zotero File > Import from clipboard 💡

Distracting myself by fiddling with something neat...using Expected Parrot's edsl library in-stream in an R data analysis to pass data to ChatGPT (in this case for classification) and then merging the answers back into R (w/ the reticulate package in RStudio). @johnjhorton.bsky.social

I spent a while trying to decide whether to hold class on Wed and decided yes. 1) University is open. 2) Students paid for that day. 3) I am paid to teach that day. I uphold my agreements. But. 1) I might forget to take attendance. 2) Heyyyy a pre-rec video of my lecture is already on the LMS.

I'll be talking about my paper "Covid-19 is (Probably) Not an Exogenous Shock or Valid Instrument" at Oslo Met's Pandemics and Society Seminar today via Zoom at 1600 CET/1100 US Eastern/1000 US Central. Minimal math, I'll make my argument with vibes. uni.oslomet.no/pansoc/2024/... #EconSky

Okay, how on earth did we miss the opportunity to call the field of studying healthcare IT "infirmation systems"? @ajh.bsky.social

Looking for examples of bad IVs for a seminar discussion. Just found a paper looking to assess the impact of socioeconomic factors on Covid-19. The paper uses WINDSPEED and AVERAGE RAINFALL 3 and 4 weeks ago to instrument for the NUMBER OF DOCTORS. It is WILD out there.

I get that a small number of high-stakes assessments has problems, and so we don't want grades based on that. But I'm getting tons of pushback now on grades being based on a large number of little, more formative exercises (from students who just won't do them). Can't win.

Chicago, you’re perfect.

The amount of time I spent coming up with a fake company and scenario for my midterm was totally appropriate, yes a very good and reasonable use of time. Pretty dope (AI generated) logo though.

Pls re-skeet for reach: CALL FOR PAPERS for a special issue of EER in honor of Nora Szech editors: S. Huck, N. Schweizer, M. Serra-Garcia #econsky 📉📈 www.sciencedirect.com/journal/euro...

Don't make me tap the sign... papers.ssrn.com/sol3/papers....

I am thrilled to announce that Bluesky is starting to feel like old science Twitter.

Pretty pumped to have seen a very convincing, very exogenous natural experiment. Now I just need to wait ~2 years to write it up so that I have sufficient data on the Post-Treatment period.

Pretty jazzed that I (along with lots of other contributors) got an email about this reproducibility study being accepted at Management Science…a great holiday present! osf.io/mydzv/

The other site has become convinced that I am a horse aficionado which is a) a totally fine thing to be but b) not me. The quantity of horse-themed ads in my feed is entertaining. Also, their algorithm is so borked at this point.

Pretty jazzed about this graph, both for the story it tells and the aesthetics.

Finally got my PhD completion gift to myself. Pretty jazzed about it.

Higher standards for code review and reproducibility would be good for social sciences but if we want that to be the standard then we need the resources to do it and the incentives need to be aligned. 1-3 author pubs and lots of them can't be the standard for tenure.