20 ms·

Test scores are an excellent way to distinguish moderate students from strong ones. In my experience, when you're distinguishing top applicants from each other,

Test scores are an excellent way to distinguish moderate students from strong ones. In my experience, when you're distinguishing top applicants from each other, much of the difference will come down to the intensity of test preparation. Of course, for these Ivy League schools, the overwhelming majority of applicants fall into this category. The desire to outperform drives a huge investment in test prep, and selects for students who have put all their time into that activity -- rather than actually being curious about the world.

ghaff 3y ago
Per a link re:MIT in these comments, at that point, the best "solution" (such as it is) is to say they have a good enough SAT and move on to other metrics that may not be as accurate but are perhaps more illuminating. Does it really matter if someone has a Math SAT of 750 vs. 800 or whatever the numbers are these days?
- jhanschoo 3y ago
  Indeed, at some point the metric becomes very subject to stochastic uncertainty. IQ tests administered by psychologists admit that there is a cutoff beyond which they fail, and so presumably the same situation exists with standardized tests.
kelipso 3y ago
> rather than actually being curious about the world Great phrase that allows anyone to make any kind of shit up to evaluate someone.
hintymad 3y ago
We can't just make the tests hard enough so preparations will have diminishing returns? When I look at the exams in elite universities like Tokyo University, JEEs, and etc, I'm not sure how preparation can help the students in different talent pools. An extreme example would be those selective STEM summer camps. For instance, students in SUMaC will pick a subject in number theory, abstract algebra, or algebraic topology to study for three weeks. I don't think everyone can even understand the topics no matter how hard they try.
Terretta 3y ago
It'd be incredible for "test prep" to have to mean "learn things" to result in better scores.

ghaff 3y ago

Per a link re:MIT in these comments, at that point, the best "solution" (such as it is) is to say they have a good enough SAT and move on to other metrics that may not be as accurate but are perhaps more illuminating. Does it really matter if someone has a Math SAT of 750 vs. 800 or whatever the numbers are these days?

jhanschoo 3y ago
Indeed, at some point the metric becomes very subject to stochastic uncertainty. IQ tests administered by psychologists admit that there is a cutoff beyond which they fail, and so presumably the same situation exists with standardized tests.

kelipso 3y ago

> rather than actually being curious about the world Great phrase that allows anyone to make any kind of shit up to evaluate someone.

hintymad 3y ago

We can't just make the tests hard enough so preparations will have diminishing returns? When I look at the exams in elite universities like Tokyo University, JEEs, and etc, I'm not sure how preparation can help the students in different talent pools. An extreme example would be those selective STEM summer camps. For instance, students in SUMaC will pick a subject in number theory, abstract algebra, or algebraic topology to study for three weeks. I don't think everyone can even understand the topics no matter how hard they try.

Terretta 3y ago

It'd be incredible for "test prep" to have to mean "learn things" to result in better scores.