Monday, November 10, 2014

Autism Diaries: An unexpected wildcard on the SAT math test

J has just taken the SATs—hopefully for the last time. To anyone in the know, his scores—with the several hundred point gap between verbal and math sections--cry out “autism” (or, possibly, “gifted math student from a non-English speaking country who only had a few years of English instruction”). Given that J’s autism-related communication difficulties have him reading at a 7th grade level, there’s only so much his verbal scores might budge upwards. But an 800 in math is theoretically possible, if only J would avoid the sorts of careless mistakes—there are always 2-3 of these--that he invariably makes on practice tests.

The wildcard here are the 10 “student produced response questions.” Here, instead of selecting among 5 multiple choice options, the student enters a numerical answer by filling in bubbles in a number grid. Everything else being equal, these SPR questions are much easier to get wrong than the multiple choice questions: they allow, after all, a much broader range of possible answers. But they’re also easier to get wrong for stupid reasons in particular. If you misread a multiple choice question, your misreading often becomes obvious to you when you look at the 5 choices and see that none of them fits. If you misread a student produced response question, there’s only the number grid to clue you in, and it will do so only if your response is, literally, off the charts.

Mathematically speaking, the SPR questions strike me as generally easier than many of the multiple choice questions. Despite this, they’re the only sort of question J gets wrong on practice tests. He misreads the question, and then has no clue that he’s misread it.

You can sort of see why SPR questions are theoretically appealing. Multiple choice questions always get a bad rap, and SPRs, by comparison, look refreshingly open-ended. But this open-endedness goes only so far—and remains a far cry from the open-endedness one finds in college entrance exams in many other countries (for example, Finland). The end point of America’s SPR questions is still a single, right-or-wrong response, with no opportunity to show your work (not to be confused with “explaining your answer”) and get partial credit. In a way, therefore, SPR questions combine the worst of both worlds: they don’t rule out stupid mistakes, as multiple choice questions do, and they don’t allow partial credit, as truly open-ended responses do.

So I can only hope that J managed somehow to show an unprecedented level of vigilance vis a vis the 10 SPR questions that confronted him this past weekend.

No comments: