Skip to main content


Showing posts from August, 2018

An Interactive Retention Visualization

As I've written before, I think graduation rates are mostly an input, rather than an output.  The quality of the freshman class (as measured by a single blunt variable, average test scores) predicts with pretty high certainty where your graduation rate will end up. 

(Note: Remember, the reason test optional admissions practices work is that test scores and GPA are strongly correlated.  If you didn't have a high school transcript, you could use test scores by themselves, but they would not be as good; sort of like using a screwdriver as a chisel.  And the reason why mean test scores work in this instance is essentially the same reason your stock portfolio should have 25 stocks in it to reduce non-systematic risk.)

Further, choosing students with high standardized test scores means you're likely to have taken very few risks in the admissions process, as high scores signal wealth, more accumulated educational opportunity, and college-educated parents. That essentially guarant…