4 ms·
I went to Stuy and have always believed that requiring people to score high on an exam that tests verbal and mathematical reasoning skills is a good thing. Now
by snikolov 15y ago
I went to Stuy and have always believed that requiring people to score high on an exam that tests verbal and mathematical reasoning skills is a good thing.
Now, a score on a single exam is not a particularly great way to admit people for many reasons, but if we accept that it is an objective test of mathematical and verbal reasoning skills, why not help minorities score high enough (or let people know that there is an exam in the first place)?
In college admissions (at least in theory), affirmative action takes effect only after all the candidates have passed some minimum bar in terms of quantifiable things like SATs, grades, etc. Admission to Stuy is based only on a test so one can apply the same principle: Look at people holistically only after they've passed the minimum bar.
I think we should help minorities make the cutoff and then take a closer look at those applicants who scored high enough (though perhaps not using a hard threshold). Of course this is easier said than done, but I think ignoring such a test completely and relying on GPAs or interviews or other things alone would not be a good idea. Most people haven't "done" many things in middle school the way people do things in high school that distinguish them when they apply to college. Please correct me if I'm wrong on this. (And I actually think one of the goals is to make sure that kids have a lot more opportunities and encouragement to distinguish themselves by the time they finish middle school)
Now, perhaps the current cutoff for getting into Stuy is analogous to getting above a 2300 on the SAT, and elite institutions seem to agree that a 2300 is no more predictive of success than a 2200 or a 2100. So it would be interesting to measure how scores on the admission exam correlate to various metrics of success in college and later in life in order to determine a similar threshold above which the score doesn't matter anymore.
- rdl 15y agoI'm curious how it should work if you measure two separate scores (math, verbal) -- would you go for the people who have the highest combined score, the highest score in either, or the highest score in either but meeting a minimum standard in the other?
- snikolov 15y agoThat's a good question. I'm not sure how I would design it. The current test uses a combined score but there's a subtlety. The curve is such that you're better off getting 350 on math and 250 on verbal (or vice versa) than 300 on each. People have suggested that this leads to accepting a many people who are "pointy" rather than "well-rounded." There are people who take differential equations freshman year as well as people who will end up publishing books by the time they graduate.