New paper out, w/Mike Ransom: "Do High School Sports Build or Reveal Character?" ftp.iza.org/dp11110.pdf /1 
In the paper we examine whether being an athlete in high school causes better educational, labor market, and health outcomes as an adult. /2
We find that almost all "effects" of sports are actually due to selection. i.e. correlation is not causation /3
We do find that, for men (but not women!), there is a causal effect on exercising regularly as an adult /4
The data we use to answer the question comes from three separate nationally representative longitudinal surveys: NLSY79, NELS:88, and Add Helath /5
The method we use is a generalization of instrumental variables: we measure how much the sports effect changes when additional controls are included in the model, and then use that measured change to predict how unobservable controls would influence the outcome /6
HS sports is a case where running an experiment would not be feasible (ethical issues), and there aren't really any viable quasi-experiments. So examining selection patterns is the best way to make a causal claim. /7
Our results show that claims about sports building leadership skills or other non-cognitive skills are overblown. Sports can and should still have a place in schools, but not for their skill-building. /8
The injury component of athletics is not often discussed, but Ray Fair has shown that injuries can be very costly, esp. for contact sports: /9 nytimes.com/2017/09/29/hea…
Finally, in an #econlife twist, this project was a dream fulfilled: I got to co-author with my Dad! Also, we end up citing independent work from another econ parent-child combo: Ray Fair and @ProfEmilyOster. #econlife is the best life. 10/10
you had asked about this paper back in January. This is the finished product! (or, at least, the initial-submission product)
