Predicting the 2017 NCAA Tournament Using Schools’ ACT Scores
It’s March Madness time again! Last year, in an effort to find a hidden edge in my office pool, I asked the question: “What if Standardized Test Scores Predicted NCAA Tournament Winners?” and used the result to choose the winners from among the 16 teams remaining at the time.
This year the project has been expanded to include the entire 68-team field, using schools’ reported 25%/75% percentile ranges for the ACT only. At first glance, the predictions appear…well, mixed. On the one hand, traditional powers Duke, Michigan, and UCLA find themselves in the Final Four. On the other hand, tournament first-timer UC Davis is predicted to hand Kansas the first-ever 16-1 upset (you heard it here first!). Princeton is predicted to emerge from the West (by far the brainiest region, which, in addition to the Tigers, features Vanderbilt, Notre Dame, Bucknell, Maryland, and first-timer Northwestern, any one of whom would have done some damage in the other three regions) to win the tournament.
(For the record, Princeton has never finished better than 3rd place and hasn’t made it out of the second round since 1967. But hey, they’ve got really high ACT scores!)
A note about methodology: Of course, on their websites, schools don’t report single ACT “average” scores, but they do report 25% and 75% percentile ranges, which is what has been used for this exercise.
And why ACT scores only, you ask? The answer lies more with logistics than anything else. With the redesigned SAT only about a year old, the SAT information that is to be found on school’s websites is inconsistent at best: some schools are reporting current SAT percentiles, some are still reporting SAT scores on the old 2400-point scale, and still others are unclear about which version of the test their numbers are coming from (one school, Kansas, does not report SAT information at all).
The following charts show school 25% and 75% percentile ACT score ranges for the teams in each of the four tournament regions: