The latest Common Data Set (pdf) shows that the 25th-75th percentile spread on the SAT (math + reading) is 1400–1570 for the class of 2021. This is a big increase from the class of 2020, which was (pdf) 1330–1540.

Note that there are two ways to report the 25th/75th percentile spread on math + verbal. First, in some years/schools, you are given this number. For example, for the class of 2020, Williams just tells us that this is 1340–1523. Second, you can calculate it yourself by adding the 25th (75th) percentile of math to the 25th (75th) percentile of reading. For the class of 2020, this gives us 1330–1540. These two methods should be fairly similar. The difference, obviously, will depend on the correlations between scores across students. I use the second method for both years since Williams does not (why?) give us the “true” range for the class of 2021.

A commentator writes:

Something is not right with the SAT numbers. Scores don’t change as dramatically as these. Williams, like a few other schools, seems to be using the SAT concordance tables to conflate old SAT and new SAT scores to arrive at artificially high numbers.

An easy way to see why I am skeptical of thee number is to look at class of 2021 data for Stanford and Princeton and compare to Williams. Here are the links.

https://admission.princeton.edu/how-apply/admission-statistics

http://admission.stanford.edu/apply/selection/profile.html

Princeton’s middle 50 is 1380-1540
Stanford is 1390-1540

If you compare Williams’ middle 50 to Stanford and Princeton’s you can see something is amiss. Otherwise Williams is suddenly more selective than Stanford and Princeton.

And it is not that the new SAT is producing higher SAT scores. Most of the more competitive schools which have current data appear to have new SAT scores which are lower than the old SAT. The best comparison I’ve been able to make is to compare the 2015-16 (pre new SAT) scores. That is pretty close to where the new SAT scores come in.

I find this argument fairly compelling. But, at the same time, Director of Institutional Research Courtney Wade (and her staff) are smart and careful. Did they make a mistake or did Stanford/Princeton?

Facebooktwitter
Print  •  Email