2) The curve is wrong
Hypothetically, if you distributed the scores of all students sitting for the SAT on a curve, with or without accommodation, it should approximate the normal curve (a.k.a. the “bell-curve”). When the College Board plotted the 2005 results of students taking the test with accommodations, the results yielded not a bell-curve but rather a bi-modal distribution (meaning the distribution was top and bottom heavy with a disproportionate number of low scoring and high scoring students rather than a tendency toward the mean). This greatly alarmed the College Board that the population of students receiving accommodation did not mirror the rest of the population.
Except that the curve isn't wrong at all.
A bimodal distribution is *exactly* what you would expect of this population. The high scoring students are the ones who know the material and just need some extra time to demonstrate that, and the low scoring ones are the ones who don't know the material or have some other issue that extra time isn't helping.
But the issue is... where were the kids whose scores were in the middle, then? THAT is what was concerning. Not that some kids were high scoring, which as you say is entirely expected. But that there were TWO bell curves-- one VERY high, and one very low. Well, as you note, very low could mean that the accommodations aren't appropriate, or that the students simply don't know what is being measured.
But why is everyone else who tested with accommodations scoring over 700 on both reading and math all of a sudden?
That was the nature of the concern.
That
some of that cohort would is completely understandable. But there should still be more of them scoring toward the mean than the tails-- assuming that the test norms the same way with and without accommodations, which is the entire premise of the asterisk going away.
What the College Board feared that it meant instead-- and this is a perfectly valid assumption based on the data available both in College Board's own stats and also in the California study of THEIR cohort... is that as many as 20% of those kids are using accommodations that they don't really need, for a diagnosis that they probably don't really have.
A disproportionate number of those high-scorers with accommodations come from private schools, are white, and of high (often very high) SES. They score at something like the upper quartile without accommodations. But WITH extra time, they can elevate their scores by another 10-15 points.
That's nothing like the students who NEED the accommodations, whose scores often rise by 20-30 points, lifting them from the lowest three quartiles into the middle of the distribution.
But it's what (at least potentially) was creating a bimodal distribution rather than a bell curve. You had "one size DOESN'T fit all.. huh" on the one end, and "great kids that just needed a chance" on the rest of the distribution, plus a bump from "going to take every advantage that I can possibly get my hands on" at the top.
Not all kids that need and deserve accommodations are academic superstars. At least, they shouldn't have to be to be "worth" it. (Back to my conversation with the snotty counselor, actually...)
