Graduation rates are always a hot topic in higher education, but often for the wrong reason. To demonstrate, I offer my parents. Here is a portrait of Agnes and Mark, married May 4, 1946. One night while I was talking to my brother, he asked, "Do you think mom was the way she was because dad was the way he was, or do you think dad was the way he was because mom was the way she was?" To which I replied, "yes." My point, of course, is that in complex relationships, it's always difficult--impossible, actually--to detangle cause and effect. And, despite the Student Affairs perspective that graduation rates are a treatment effect, I maintain that they are actually a selection effect. As I've written about before , it's pretty easy to predict a college's six-year graduation rate if you know one data point: The mean SAT score of the incoming class. That's because the SAT rolls a lot of predictive factors into one index number. These include acade