Adaptive Scoring: Big Problem with the Digital SAT (dSAT)
Starting in 2024 the SAT will be fully online. This is the third part of my series on the new digital exam; you can read the first two parts:
Benefits of the New Digital SAT (dSAT)
Problems with the New Digital SAT (dSAT)
One of the biggest changes, and likely a significant challenge for test takers, is the new digital SAT’s adaptive scoring.
Old Scoring Methods
In 2016 the SAT changed to only count a student’s correct answers. Before this students earned points for right answers, but also lost a fraction of a point for wrong answers as a type of guessing penalty.
From 2016 through 2023, each SAT had a score conversion table that equated the number of correct problems with an score. For example, if you got 58 of the 58 math questions right, you had a perfect score of 800. 40 of 58 would give you a 600.
Different tests would have slightly different score conversions, but in general the same number correct (plus or minus a few questions) would yield the same results. This meant the scoring system was clear, consistent, and predictable.
What Is Adaptive Scoring?
Adaptive scoring is new for the digital SAT, but not new for standardized testing. Most of the graduate school exams have been digital and score adaptive for years.
Each part of the new SAT will have two scored modules. (Two modules of reading/writing and two modules of math.) If a student does well on the first module, they will be given a harder second module of that subject. Harder modules mean higher scores. If a student does poorly on the first module, they will receive an easier second module that comes with lower scores.
How Does It Complicate Test Prep?
The first complication is that two students could answer 30 answers correct on the digital SAT math and have two different scores.
Currently College Board has released very limited practice materials and only vague score conversion charts. Students no longer have a clear and consistent target number to work towards as they prepare.
In working on my new SAT prep course materials, I reviewed the College Board materials and scoring guidelines. 30 questions correct out of 40 total questions in math could score as low as a 500 and as high as a 700. In other words, that 30 questions right could place a student anywhere from the middle half of all test takers to the top 10%. That’s a big difference. And a hard one to wrap your head around if you are trying to set score goals as you practice.
How Will Adaptive Scoring Impact Test Day?
My biggest concern with adaptive scoring is its impact on a student’s test day plan and experience.
Because performance on the first module is essential to earn higher scores, students need to go into the test warmed up and ready to excel starting with the first question. This will be a challenge for many students who start out nervous and only to settle into a good testing rhythm after a number of problems.
Add to this the fact that many students will be less familiar with the College Board’s Bluebook testing app and its features. They may not utilize all the tools available to them to maximize results.
What Can Students Do?
The first step is appreciating that the new digital SAT is not scored in a traditional way. Understanding adaptive scoring principles helps.
Next, students should practice enough that they are familiar and comfortable with the test questions and digital platform so they can do their best on the first module of each question type.
All students should warm up their brains before the actual exam, just as an athlete would warm up before an important game. Adaptive scoring makes this step even more critical.
Finally, until we have more data, practice materials, and score conversion examples from actual exams, students need to remain flexible. We just don’t have the same clear and consistent score numbers we used to.
Conclusion
Once you understand what adaptive scoring is and how it is used on the new digital SAT, you can take steps to prepare. This is probably the one area of the new SAT that is going to be least predictable for the near future.