Since this past June, high school students have erupted in an online protest demanding for the College Board to rescore the results of their June SAT® exam. Many students feel that the “curve” is too brutal and are confused as to how and why their SAT® scores are not going up despite getting fewer question wrong than in previous tests. Students have taken to social media platforms such as reddit and twitter to voice their complaints, one student posting, “The curve is insanity. I improved by getting 13 more questions correct, yet my score only increased by 20 points. This is outrageous!”
In terms of the Reading and Writing sections of the June 2018 exam, for example, students who were aiming for a 640 or 650 may have ended up with scores in the 500s. One reason the curve was stepper than usually probably stems from the 2 omitted questions from the Reading section and 2 omitted questions from the Writing section—a total of 4 questions retracted by the College Board after the exam date. When the College Board omits questions, there are fewer items to which they can compare students against each other, leading to steepening an already very steep curve.
So when we talk about this curve, what exactly does this mean?
Turns out, the SAT’s® curve isn’t actually a curve—it’s a scale. While a “curve” is compiled after a test is taken, a “scale” is determined by data sourced without actual test takers on exam day. The College Board creates this scale through a process called “equating,” where they select a group of anywhere between 1000 and 5000 students to take an unreleased exam. Then they compile the data from this test and cross reference it to real SAT® tests that these students took. By doing this, the College Board is able to map out all the SAT® questions and determine the difficulty of each individual question. The easier the questions are overall on a test, the more points each answer is worth. If you miss questions on an easier test, you’ll get more points docked off because the College Board expects many people to get it right.
So the irony is that the June SAT® was actually too easy. To accommodate for this incongruity, the College Board had to scale the June SAT® much more precipitously than any other revised SAT® exam. Typically, a raw score between 47-51 will give you a 700 on the Math section of the revised SAT®. However, the June SAT® required a raw score of 54 in order to get that same 700 score— 3 points outside the regular range. This is actually similar to the typical scale on the old SAT® math section, but for this exam, is indeed atypical.
For many students, the explanation of “equating” given by College Board isn’t satisfactory, causing students to start a petition and a hashtag to “#rescoreJuneSAT.” And they may have a point, this test was not “fair” for “all test takers.” Any testers who tend to make careless errors or run out of time (and thus leave questions blank) are particularly hurt by “easy” tests. On the flip side, those who struggle with content only may perform better on “easy” tests.
What’s most fair? Tests that are as similar as possible. In that sense, the College Board may still have some work to do.