Catch me on the other site: @ + (same as my website). I’ve also got a few codes if anyone needs some.

RT @jenniferdoleac: I spent Friday with researchers and practitioners at a @JPAL_NA convening, talking about the value of causal evidence (….

Suggestions on best (e.g. good performance, not crazy expensive) VMs where I could run some heavy, one-off, computations in R? I have a total of like 8T of data and I need to combine geospatial data by state (largest state is like 1T). Azure? AWS? All suggestions welcome #rstats.

RT @andres_bafer: Two weeks left to submit to the Valle Nevado Workshop on Economics of Education. Great papers and a fantastic ski 🇨🇱⛷️!!….

RT @andres_bafer: 🚨📢 3rd Workshop on Economics of Education, Valle Nevado, 🇨🇱 (Aug 19-22). Confirmed speakers.🎓 JOHN FRIEDMAN (Brown) .🎓 S….

Ah! You can also check out the full paper here:. (And a shoutout to `tinytable` and `modelsummary` for making pretty tables 💕).

This app allows us to look at the recruitment process at diff stages, and to have objective measures of performance throughout a players' career. And in this case, it could be that early decisions (HS) could compound disparities later on as well (NFL). Anyways, this was fun!.8/8

But why is this important? Measuring discrimination in labor markets is not easy, especially because a lot of the times we lack data (on recruitment, performance, etc.). Sports labor markets have the advantage of having *a lot* of data. and who doesn't like that? 7/n

If some teams discriminate, how much does that cost them? I do some sims under a scenario where teams randomly choose a QB with a similar HS performance to the one I obs. Teams that are likely to discriminate would ⬆️ their prob of winning significantly if they didn't (6.5%) 6/n

Ok, but it could be that recruiters are trying to max the performance of _the team_. so they track non-white QBs to other positions (most likely, a wide receiver (WR)). However, it doesn't look like HS QBs are better WR than. WRs. In fact, it appears like they are worse 5/n

Ok, non-white QBs are less likely to be recruited. so what? Maybe they are worse?. Not according to the data 📊. They have diff play styles (e.g. pocket passers vs dual threats), but QBs of color still tend to outperform white QBs in relevant measures 4/n

There is a significant diff in recruitment (for both the transition to college and to the NFL), which is probably not surprising to anyone that has watched football. One interesting thing is that for HS recruitment, there results are moderately robust to hidden bias 🤓 3/n

To assess whether there's actual discrimination, I use a benchmark and an outcome test. The 1st one shows diff in recruitment cond on other vars. The 2nd one shows diff in performance by race. Gaebler & Goel(2024) show that under a monotonicity asspt, at least one test is right!

🚨Working Paper Alert🚨. "QB or Not QB? Measuring Discrimination in Football Labor Markets". TL;DR. I find that there is significant discrimination when recruiting QBs from HS to college (and from coll to the NFL!) based on their race. and this costs teams wins🏆 and money💰 1/n

Camino al @cepchile como en los viejos tiempos… Ni que fuera el 2011.

RT @StatModeling: A 10% swing in win probability corresponds (approximately) to a 0.4% swing in predicted vote.

RT @chuga: It's (almost) time to exorcise your money fears 👻 Copilot Money lands on iPad in just a few hours!. The before and after says it….

RT @2plus2make5: Please retweet: I am recruiting PhD students at Berkeley!. Please apply to @Berkeley_EECS or @UCJointCPH if you are intere….

RT @Netlify: this one's for the devs who collaborate with nontechnical folks

RT @OurWorldInData: People tend to think there are more immigrants in their country than there really are. In the United States, for examp….

RT @VincentAB: 📚🚨 I posted 11 new chapters of my upcoming book!. Model to Meaning: How to Interpret Statistical Results with #marginaleffec….