<p>@siserune Thank you for sharing that paper. I was very surprised that only 15% of people with an SAT math score of 800 who retook got an 800. This prompted me to look at my SAT score reports online and I saw similar statistics (it was 16% instead of 15%) but there is one caveat not reported in the article. The statistic seems to only be for students who took the test the first time as juniors who then retook it as seniors. Although this is damaging for the hypothesis the statistic would probably not include exceptional 2400s if they exist so I wouldn’t say that conclusively falsifies the hypothesis. Similarly, although the sample size was very small, the evidence regarding high scores on the AMC 12 and SAT 800 was also damaging to the hypothesis. </p>
<p>I’m curious to what extent this is unique to math as the possibility for dumb mistakes is surely greater on math tests and the curve is harsher near the top. If people who got 800s on critical reading or writing go look at their SAT score report and go under your score details they should be able to find what percentage of students with an 800 on reading or writing when they first took the SAT as juniors got another 800 when they retook as seniors.</p>