NOTE: Thanks to the summer 2016 release of percentile data, we have published a new post on Easier 2016 SAT Percentiles.
With the first administration of the new format SAT rapidly approaching, we see lots of students starting to work through tests in the Official SAT Study Guide (2016 Edition). Scoring the new SAT turns out to be much easier these days, especially when you ignore the subscores and cross-test scores. Understanding what those scores mean, however, appears a lot tougher.
The problem, of course, lies in the very arbitrary nature of the 200-800 scale. Obviously, scores improve as they rise towards a perfect 800. But how much better is a 670 than a 620 or a 520? We can only understand the true achievement our test scores represent when we see their percentiles, the depiction of what percentage of the testing population we scored higher than.
Too bad the College Board hasn’t released percentiles data for the new SAT.
These new tests diverge sufficiently from previous versions of the SAT to require brand new assessments of scoring patterns across a cohort of test takers. In the past, the statisticians and psychometricians at the College Board could at least approximate percentiles for SAT scaled scores. This time around, however, the test includes question types that have never been widely administered on College Board tests, or at least in any way that counted. Thus, we have no percentiles data for the new SAT and likely will not have any until mid-summer at the earliest.
Fortunately, we can at least estimate percentiles rankings. The new SAT score is based on two section scores, namely Evidence-based Reading and Writing (EBWR) and Math (M). EBWR, in turn, comes from equal parts Reading and Writing.
The former version of the SAT, before we forget it completely, was based on three section scores: Critical Reading, Math, and Writing. We have plenty of valid percentile data for administrations of that version of the SAT. The problem is that, while old Math percentiles may be useful guides for what the new Math percentiles might be, the old Critical Reading doesn’t correlate to the current Evidence-based Reading and Writing. In fact, neither does the pre-2005 SAT Verbal section. Since EBRW scores are based on both Reading and Writing, we can use both sections of the previous SAT to estimate percentiles, as follows:
(CR percentiles + W percentiles/2) + M percentiles = EBRW percentiles + M percentiles
The table below provides CR, M, and W percentile data:
No, these percentiles probably won’t correspond exactly to the rankings the College Board finally comes up with after the spring administrations of the SAT. On the other hand, the current 200-800 scales will also likely change. Estimates of SAT percentiles may not be exact, but they can be very helpful in assessing relative ability and progress. Once you understand your percentiles, you can assess how much work you’ll have to commit to improving them!