<p>I don’t think College Board has any way of adjusting their scoring for different testing populations. They can only make adjustments for differing raw score bell curves. Let’s assume that students who take the SAT Subject Tests (a very high achieving test population) often take the May SAT I so that they can take the Subject Tests in June. Let’s assume that as a group, they score very well in terms of their raw scores. College Board will subsequently determine that the test was easy and impose an artificially harsh curve. The curve would be artificially harsh, not because the average question was easier, but rather because the average test-taker was more capable. Can anyone explain what is wrong with my analysis?</p>