Wacky Scores on the June 2018 SAT

SCORES FOR THE JUNE 2018 SAT HAVE BEEN RELEASED AND HIGH-SCORING STUDENTS ARE NOT HAPPY. THE CURVE ON THE MATH SECTION WAS REALLY UNFORGIVING TO HIGH-SCORING STUDENTS. THE TEST IS A REMINDER THAT STUDENTS SHOULD BE READY TO RETAKE THE SAT AND ACT AND THAT NO ONE ACTUALLY WANTS AN “EASY” SAT.

Another case of “Just when you thought it was safe…” after the dust had settled on the College Board’s newest version of the test, trouble has reared its ugly head.

Case in point: the June SAT Math scores for high-scoring students were computed using a score curve that was rather tough on those students. Miss x number of questions, and your score takes a relative nosedive. There is quite a disparity between this and previous tests’ Math sections:

SAT Math Curve (courtesy of Jed Applerouth)

Testing services explain how careful they are in equating the tests so that college admissions officers can rely on continuity, consistency, and equality of questions from test to test. College Board puts it this way via the Educational Testing Service (ETS) website:

Equating makes it possible to report scaled scores that are comparable across different forms of the test.

College Board really tries to ensure that there are not large differences between the SAT tests. ACT engages in the same quality assurance. Although from the get-go, the latest SAT Math sections exposed serious problems in the questions, June score reporting proceeded without anyone raising serious issues. Until now.

As soon as the June 2018 SAT scores were delivered to students, there was an outcry about the Math curve. Students with fewer wrong answers on this test than on previous SAT tests received lower Math scores!

Here’s where equating comes in: Because the June Math was easier, the curve was tougher than usual. Statistical equating is performed before the test administration. College Board knew the math was easier, meaning more students would get more questions correct, so the scale would be adjusted accordingly. In the end, the impact of a few missed questions had a larger than usual impact. OK, equating is supposed to take care of that. But.

Apparently, equating used for the June 2018 SAT seems to point to the fact that the College Board released a test whose scores make it too easy to distinguish among high scorers who received a score of 650 (86th percentile) or higher. For colleges that treat a 650, a 700, a 750, and an 800 as true indicator of differences in math ability, that’s a problem.

Almost 4,000 students have signed an online petition attacking the College Board for its scoring practices.

Students dismayed by their June 2018 SAT results have a few additional opportunities to retest. We trust that College Board and ETS have already taken note of the outcry, and will do everything within their power to mitigate this relatively strange June test administration. Is trust the same as hope?

P.S. There was a similar anomaly with the Reading/Writing score because, again, there were too many easier questions, and four test questions were actually dropped from the scoring because there were problems with them.