Hot enough for you yet?
Several folks emailed me this Wall Street Journal Op-Ed about the declining use of the SAT for college admissions. Author Naomi Riley correctly points out the distinction between differential group impact and bias, and deftly summarizes the differing attitudes of the test prep giants Kaplan and the Princeton Review. Having dealt with representatives of both companies, I can attest to the veracity of her description.
Also sent along by several readers was this “criticism” of the SAT, which quickly devolves to the level of a conspiracy theory. I’m used to testing critics holding up the example of one college as “proof” that the SAT isn’t useful for predicting performance at any college. But claiming that a not-for-profit institution is, for the sake of profit, deliberately misrepresenting its content and deliberately choosing items to disadvantage certain students? In order to “spread out” the distribution of scores? That’s a new one for me. As I read it, his theory begins with the assumption that SAT scores would “clump” up and not be meaningful in the absence of biased items because everyone’s ability is so similar on the constructs taught in the classroom. Wow. I give him credit for managing to compress such a profound misunderstanding of item development, scaling, test assembly, testing companies, general psychometrics, and the state of modern education into such a short essay.
On a related note, many news outlets are reporting that the new SAT doesn’t predict college GPA any better than the old version. This particular article quotes the College Boards SVP of operations, Laurence Bunin, as noting that, despite the small rise in predictability, the addition of the writing component may have had a beneficial impact on high-school level writing classes. The College Board website also points out that the results show that the SAT is a better predictor than high school grades for minority college applicants. Interesting, then, the increasing number of colleges who have recently ditched the SAT in order to “diversify” their campuses.
Oh, and if you think you might want to take the SAT multiple times, but don’t want colleges to see your, um, “practice” takes? Starting in 2009-10, that’ll be no problem.
Plenty of methods have been suggested to help make kids more optimistic about their educational future and more motivated about taking tests. Here’s one of the more interesting ones.
Strange news from my old stomping grounds – the Chapel Hill-Carrboro (NC) school district is considering implementing one heck of a floor effect. Even if an assignment is not turned in or a test is not taken, 61 is the lowest grade that can be given? The district’s director of programs has a bizarre explanation for the plan, which is that giving really low scores won’t help failers get better. Technically, I guess that’s true, but forcing the lowest score to be a 61 doesn’t actually help someone who’s doing terrible work to get better either. It just raises their average, not their understanding. Once again, the messenger is being confused with the message.