3 ms·

I think standardized tests correlate very highly (perhaps highest of any measure) to performance. I regularly see schools in the US with a 10% math proficiency

by bko 3y ago

I think standardized tests correlate very highly (perhaps highest of any measure) to performance. I regularly see schools in the US with a 10% math proficiency and 40% literacy proficiency with a 99% graduation rate. What's up with that? Standardized tests are the most fair and simplest way to evaluate students. Plus gaming the test is what exactly? Hiring a tutor and spending hours studying, much of it alone? And it's not a shock that success on studying and doing well on exams correlates to college where you're evaluated on how well you study and do on exams It reminds me of that Key and Peele sketch where they're planning a long con bank heist but by the end it's obvious they're just working a job and saving for retirement https://youtu.be/jgYYOUC10aM?si=uiceINWHRRf3brZ5 https://youtu.be/jgYYOUC10aM?si=uiceINWHRRf3brZ5

ghaff 3y ago
You probably want to do some test prep but you can do that with a test prep book at your local library. Years ago, in a grad course of some sort, we looked at some papers of predicting performance in school. Basically, even at the grad level, GMATs were far and away the best predictor of academic performance, followed by undergrad grades. Things like letters of recommendation were pretty worthless.
arp242 3y ago
The argument is that some people have access to private tutors who "teach the test" and others don't, not "some people study harder than others". And there is absolutely truth to that. I had to work after school, to make ends meet. Ain't got no time for extra studying. That doesn't mean we should do away with test scores altogether, but we do need to realize the limitations. These scores come with error bars and uncertainty, and it matters how you use them. "Everyone with a score higher than the requirement ends up in a lottery" tends to be reasonably fair, but "we take the top-200 scores" tends to be biased to the well-off (sometimes very biased).
- vladgur 3y ago
  We should advocate for social programs that allow high school kids “make ends meet” without impacting everyone’s academic acceptance criteria.
- bko 3y ago
  Harvard isn't some lottery that just pays out. It's a school. If you have to work after school and didn't have time to study for an exam, maybe you should go to a less competitive school. If you're not prepared enough, getting into a very challenging school isn't good. Its like I don't know how to read because I was taking care of my dying mother but now I want to get into a PhD program for literature. The PhD program wouldnt be doing me any favors by letting me in. Instead i would be frustrated, waste a lot of time and money and drop out eventually I agree that there is likely no difference in performance between a 1500 and 1600 SAT so a lottery is reasonable. Although I will say it's pretty strongly correlated even at the extremes. But why should a school solely focus on maximizing gpa? It should minimize failure but choosing between a student who will get a 3.5 or a 3.75 is pointless https://zeescorrelationstudy.weebly.com/ https://zeescorrelationstudy.weebly.com/
  - Dylan16807 3y ago
    > If you have to work after school and didn't have time to study for an exam, maybe you should go to a less competitive school. If you're not prepared enough, getting into a very challenging school isn't good. Getting extra tutoring, focused on improving your test results, does not make you more prepared for college. It's not like your analogy.
- gottorf 3y ago
  > some people have access to private tutors who "teach the test" and others don't, not "some people study harder than others". And there is absolutely truth to that. There is truth to the fact that some students have better access to study the test than others, but in my recollection, it really doesn't help those kids very much. I'm sorry I don't have any sources off hand, but I remember it being something like that all the extra prep will add maybe 10-30 points to the SAT. To put another way: a dumb rich kid with private tutors will still score lower than a bright poor kid who takes the test with no prep.
- yongjik 3y ago
  I agree with your concern, but the problem is that Harvard will have to choose somehow, and most other ways of choosing are even more biased toward rich parents: some high school kids are writing conference papers and starting a venture, just imagine what kind of social background would make these kids even consider such an opportunity. I don't know if there's any good answer.
gnicholas 3y ago
> I think standardized tests correlate very highly (perhaps highest of any measure) to performance This depends on what kind of performance you're trying to measure. There will be some people who test very highly but are much less inclined to persevere and follow-through. Grades can be a better measure of this (though with rampant grade inflation, their utility is undercut).
MrSkelter 3y ago
This is a tautological argument. Test scores correlate with performance as measured via tests. Many of the best computer programmers I know have zero formal eduction in the field. They would fail the tests beloved to Silicon Valley and yet have proven their worth time after time.
seanmcdirmid 3y ago
> Plus gaming the test is what exactly? Hiring a tutor and spending hours studying, much of it alone? Only if you aren't in Asia. Cram schools are a thing in India, China, Japan, Korea... Of course, willingness and perseverance to go an extreme in work ethic has to be related indicator of future success. But that's your competition if you are just hiring a tutor and spending some hours studying alone.
naijaboiler 3y ago
The only thing standardized tests measure very well in aggregate are parental income and educations level

by bko 3y ago

ghaff 3y ago

You probably want to do some test prep but you can do that with a test prep book at your local library. Years ago, in a grad course of some sort, we looked at some papers of predicting performance in school. Basically, even at the grad level, GMATs were far and away the best predictor of academic performance, followed by undergrad grades. Things like letters of recommendation were pretty worthless.

arp242 3y ago

The argument is that some people have access to private tutors who "teach the test" and others don't, not "some people study harder than others". And there is absolutely truth to that. I had to work after school, to make ends meet. Ain't got no time for extra studying. That doesn't mean we should do away with test scores altogether, but we do need to realize the limitations. These scores come with error bars and uncertainty, and it matters how you use them. "Everyone with a score higher than the requirement ends up in a lottery" tends to be reasonably fair, but "we take the top-200 scores" tends to be biased to the well-off (sometimes very biased).

vladgur 3y ago
We should advocate for social programs that allow high school kids “make ends meet” without impacting everyone’s academic acceptance criteria.
bko 3y ago
Harvard isn't some lottery that just pays out. It's a school. If you have to work after school and didn't have time to study for an exam, maybe you should go to a less competitive school. If you're not prepared enough, getting into a very challenging school isn't good. Its like I don't know how to read because I was taking care of my dying mother but now I want to get into a PhD program for literature. The PhD program wouldnt be doing me any favors by letting me in. Instead i would be frustrated, waste a lot of time and money and drop out eventually I agree that there is likely no difference in performance between a 1500 and 1600 SAT so a lottery is reasonable. Although I will say it's pretty strongly correlated even at the extremes. But why should a school solely focus on maximizing gpa? It should minimize failure but choosing between a student who will get a 3.5 or a 3.75 is pointless https://zeescorrelationstudy.weebly.com/ https://zeescorrelationstudy.weebly.com/
- Dylan16807 3y ago
  > If you have to work after school and didn't have time to study for an exam, maybe you should go to a less competitive school. If you're not prepared enough, getting into a very challenging school isn't good. Getting extra tutoring, focused on improving your test results, does not make you more prepared for college. It's not like your analogy.
gottorf 3y ago
> some people have access to private tutors who "teach the test" and others don't, not "some people study harder than others". And there is absolutely truth to that. There is truth to the fact that some students have better access to study the test than others, but in my recollection, it really doesn't help those kids very much. I'm sorry I don't have any sources off hand, but I remember it being something like that all the extra prep will add maybe 10-30 points to the SAT. To put another way: a dumb rich kid with private tutors will still score lower than a bright poor kid who takes the test with no prep.
yongjik 3y ago
I agree with your concern, but the problem is that Harvard will have to choose somehow, and most other ways of choosing are even more biased toward rich parents: some high school kids are writing conference papers and starting a venture, just imagine what kind of social background would make these kids even consider such an opportunity. I don't know if there's any good answer.

gnicholas 3y ago

> I think standardized tests correlate very highly (perhaps highest of any measure) to performance This depends on what kind of performance you're trying to measure. There will be some people who test very highly but are much less inclined to persevere and follow-through. Grades can be a better measure of this (though with rampant grade inflation, their utility is undercut).

MrSkelter 3y ago

This is a tautological argument. Test scores correlate with performance as measured via tests. Many of the best computer programmers I know have zero formal eduction in the field. They would fail the tests beloved to Silicon Valley and yet have proven their worth time after time.

seanmcdirmid 3y ago

> Plus gaming the test is what exactly? Hiring a tutor and spending hours studying, much of it alone? Only if you aren't in Asia. Cram schools are a thing in India, China, Japan, Korea... Of course, willingness and perseverance to go an extreme in work ethic has to be related indicator of future success. But that's your competition if you are just hiring a tutor and spending some hours studying alone.

naijaboiler 3y ago

The only thing standardized tests measure very well in aggregate are parental income and educations level