1. documentcloud.adobe.com/link/track?uri… This striking new paper by Uma Jayakumar and @Scott_E_Page provides some context for the university cheating scandal. Short version: water polo doesn't provide rich students with a leg-up because: cheating coaches, but instead because: water polo.

2. Their findings are straightforward. First high socioeconomic status kids go to different schools than lower socioeconomic status kids. Second these schools not only differ in their educational offerings, but in their extracurriculars. And this can make a big, big difference.

3. "New Trier, an elite public school... offers 18 varsity sports –including fencing and bass fishing –along with more than 150 clubs. ...In contrast, students at Romulus High School in Michigan choose from among only 9 sports and fewer than 20 extracurricular activities."

4. This means that students at elite schools have much higher odds of finding something that they are good at. Furthermore, activity choices are probably highly correlated. Athletes at non-elite schools have to compete with many more people in their own school and others

5. all vying to be the best at the same sport. Want to be a top 100 skier? Your odds are low - but they are far higher than the odds of becoming a top 100 basketball player. Finally, if you are high SES your parents can invest much more money and time in helping you to excel.

6. These different effects - more choice, less competition and more resources - all point in the same direction - a far greater advantage for people at some schools than others, independent of both variation in academic ability and variation in athletic ability across schools.

7. Building on a simple model and toy data, they suggest that "students from high socioeconomic status families and communities may be between five and twenty times more likely to exhibit the exceptional talents desired by elite colleges and universities than lower SES students."

8. These results should obviously not be taken as definitive or anything like definitive. Instead, they point attention towards a set of largely unconsidered biases in the system as it exists, and ways in which these biases can plausibly have a substantial multiplying impact.

9. Nonetheless, it would be startling if these biases did _not_ have a substantial impact given the plausibility of the mechanisms. And they are also plausibly related to differences in future career trajectory, for reasons discussed in Lauren Rivera book washingtonpost.com/news/monkey-ca…

10. [One of the more depressing aspects of the discussion of that book was that some reviews took an indictment of the system as a how-to manual]. Finis.