Standardized test curves: The process of equating
People frequently ask me which test date or dates will have the best curve and will therefore give test-takers the greatest scoring advantage. In school, students often lament a well-prepared peer who “ruined the curve” of a particularly challenging classroom test, and most people assume that the same logic also applies to standardized tests. While standardized test scores are indeed adjusted, the process differs from traditional “curving.” Unlike school exams that a teacher might curve after s/he administers the test and evaluates student performance and score distribution, standardized tests are curved (a process called “equating”)before the administration of that test.
Standardized tests are so named because schools can compare student results across years, providing a consistent scale for student evaluation. For each test, a student receives a raw score (total number of correct answers) and a scaled score (raw score adjusted to a predetermined scale). While the raw to scaled score conversion can vary slightly with each test, the overall distribution of scores will be the same. Standardized tests are engineered and scaled such that the score distribution fits a bell curve: most students will score close to the average (the middle part of the curve) and very few will do extremely well or extremely poorly. (You do not, in fact, get 200 points on the SAT just for putting your name on the test, but if you get even a single question right anywhere in the test, you are likely to earn at least that many points.) Equating takes into account the difficulty of the questions and anticipates student performance to insure that, regardless of difficulty, test scores will maintain the same distribution.
The way in which a test is equated can still affect your score. For example, the June 2018 SAT received widespread complaints (google “June 2018 SAT curve” if you like) as high scoring students felt their math scores were lower than they should have been. On this particular test, if a student missed two math questions, s/he received a score of 750. On previous tests, two missed math questions would have resulted in a 790 or 780. What’s the difference? The difficulty level of the test. The math on the June 2018 SAT was too easy—it did not contain enough hard elements and thus the equating of this test was extreme. As a result, many students saw their accuracy increase on this test while their scaled score actually decreased. Arguably College Board should have better engineered the June 2018, but given that the test was easier than normal, it stands to reason that each error would have a greater impact.
What does all of this mean for you and your student? First, you always want a hard test! The more difficult the material, the better the curve, so each mistake becomes less significant, particularly at the top of the scale. Second, plan to take multiple official tests to make the most of these slight differences. At the top of the scoring scales, the difference between a 33 and a 34 or a 1550 and a 1570 can be a single question. Third, start your prep and testing early so that you have room in your schedule to add an additional test if needed.
Understanding the equating process is just one more step toward reaching your goal score. By taking multiple official tests and understanding that the degree of difficulty will vary between tests, you can leverage your own content strengths and test-taking skills to achieve your personal best.