News for hackers. Stuff that matters.
Re:Tech journalism is less diverse than tech (2020) by skulk on Friday November 19, 2021 @11:41AM (#29279096)
What do you think about the system on the subcontinent? I'm not sure how it's done in Bangladesh, but in India your higher education is basically decided by your score on a few objective measures which are orders of magnitude more difficult than the SAT.

It's actually mind-boggling how much harder IIT JEE math is compared to SAT math.

Re:Tech journalism is less diverse than tech (2020) by thenanyu on Friday November 19, 2021 @12:34PM (#29279853)
Chinese Gao Kao tests are similarly much harder than American tests; but at the end of the day I don’t think absolute difficulty matters too much as it’s a comparative test.

As long as there is differentiation it does it’s job. If the test taking population is very large the tests might need to be more difficult to make finer distinction

Re:Tech journalism is less diverse than tech (2020) by angelzen on Friday November 19, 2021 @12:42PM (#29279963)
Not surprisingly, the population of STEM graduate programs is vastly dominated by foreign students. "81 percent of full-time graduate students in electrical and petroleum engineering programs at U.S. universities are international students, and 79 percent in computer science are."

https://www.insidehighered.com/quicktakes/2017/10/11/foreign...

Re:Tech journalism is less diverse than tech (2020) by quantumwannabe on Friday November 19, 2021 @12:51PM (#29280094)
That is more because a grad degree is a unnecessary expense for an American who wants an industry job, so the only Americans who go for grad degrees are people who want to be in academia. A masters contains a lot of repeated information from a bachelors, and many masters programs from "reputable" schools are now diploma mills that prey on international students willingness to pay large amounts of money an American degree.
Re:Tech journalism is less diverse than tech (2020) by cameldrv on Friday November 19, 2021 @01:14PM (#29280378)
Advanced degrees also make getting certain visas easier. The relative value for spending money and two years on a Masters or six years of your life on a PhD is much higher if it carries with it a ticket to stay in the U.S.
Re:Tech journalism is less diverse than tech (2020) by filmgirlcw on Friday November 19, 2021 @01:35PM (#29280641)
I was going to say the same thing. Immigration into the US is extremely difficult if you aren't already famous/wealthy/well-connected, that if you want a visa, coming in through education is your best bet. And getting a job with a company that will sponsor your visa (to say nothing of one that will sponsor your potential green card) is already difficult, but being a student and having a graduate degree is a good incentive.
Re:Tech journalism is less diverse than tech (2020) by notaniitian on Friday November 19, 2021 @05:35PM (#29283503)
It's not meritocratic. Almost 70% seats are reserved in IITs.

https://en.m.wikipedia.org/wiki/Reservation_policy_in_Indian...

Relative ranking is something no one should adopt unless they have a population problem. Unlike SAT, your rank relative to your peers in the exam decides your fate. The problem with such system is it encourages people to participate in the race earlier. Parents are sending their kids earlier to JEE prep every year and increasing the difficulty. It will continue to go up unless it's humanely impossible.

Coaching industry built around it has bigger market value and revenues than government spends on IIT (the test that JEE is supposed to get you in).

Our most valuable startup is byjus, an edtech focused on helping students pass JEE. More and more foreign companies like Amazon are also opening JEE prep centres here.

https://academy.amazon.in/

Please don't adopt it.

Despite above, our educational outcomes in higher education are worse than most developed countries including US. (We are good at employment at tier 1 colleges though)

These students get burned by the time they get into college of their choice. There isn't much focus on actual academics and coursework unlike developed countries where universities get more difficult. Further, most students are encouraged to practice leetcode from year 1 instead of focusing on academics.

Everyone is slacking off after getting in so curving keeps their grade good despite a huge drop in effort.

Re:Tech journalism is less diverse than tech (2020) by OJFord on Friday November 19, 2021 @08:13PM (#29284660)
Recommend Netflix's Kota Factory that I happened to watch recently to anyone interested in this. (Not vouching for any accuracy, I'm British, (and it's a fictional drama series) but this comment reminded me of it since it does portray the sort of madness(?) described.)
Re:Tech journalism is less diverse than tech (2020) by commandlinefan on Friday November 19, 2021 @04:29PM (#29282786)
> What do you think about ... how much harder IIT JEE math is

Seems to me that there are only a couple of different ways somebody could feel about that. The first is to be opposed to the "the smarter you are, the more money you make/better you do in life" meritocratic egalitarianism that these tests support. The second is to support that ideal (whether you agree we're there or not), but insist that tests don't accurately measure intelligence.

Either way, while I can sort of see where the opponents are coming from, I don't see them offering many good alternatives. I don't think it's cosmically "fair" that being born to insanely rich parents leaves you better off in life than being born to unemployed drug addicts, but the only "solution" anybody's ever put forward is massive ongoing wealth redistribution which is demonstrably worse. Similarly, it's not necessarily fair that people who are born smart do better than people who are born stupid, but what do you do? Require as many stupid doctors as smart doctors? Demand that for every smart highway engineer, a stupid one be included as well?

I can also sympathize with people who insist that tests don't accurately measure intelligence - I've always done well on tests, so this system has worked out well for me, but I can see how somebody could be otherwise intelligent but unable to perform on tests. But again... what do you do? Just declare everybody equally intelligent and don't measure anything? Like it or not, tests do measure something that does correlate with that hard-to-quantify trait we call "intelligence" and while there are regrettable false positives as well as false negatives, it's the best we have.

So I wish our tests were harder so there'd be a more fine-grained stratification of actual ability, even if I ended up being closer to the middle than I am now - that would be ok, because a lot of other people would be closer to the middle too! A tests that accurately measured intelligence such that only the most intelligent people in the world could dominate it would necessarily be almost impossible to score perfectly on it, by definition. In the face of such a test, a 50% would be something to brag about and the top scorers would be off curing cancer or something for the benefit of all of us.

Re:Tech journalism is less diverse than tech (2020) by brendoelfrendo on Friday November 19, 2021 @05:30PM (#29283454)
I guess it depends on what you want out of the test; if you want something meritocratic, what we find is that testing ends up reinforcing existing social strata more often than not, because wealth often correlates to factors like quality of education, which in turn correlate to test scores. A stupid rich person will always be stupid, but a smart rich person will probably have access to better schools, private tutoring, and test preparation resources.

So there's the challenge that the anti-testing/testing reform crowd wants to solve: how do you call testing meritocratic when the equality of opportunity isn't there? And are we ok with that?

Re:Tech journalism is less diverse than tech (2020) by sokoloff on Friday November 19, 2021 @06:41PM (#29284053)
If we’re trying to measure something (say strength), should we measure that thing as directly as we can, or should we add a bunch of adjustments to make things more “fair”? Do we want to know how strong someone is, or how surprising it is that they’re as strong as they are given their starting point?

If someone has a home gym, a personal trainer, and time to workout 2x/day, 6 days a week, how much should we penalize their results? Zero? 25%? 50%?

I think that being strong as demonstrated on tests of strength is a valid (and the most valid I’m aware of) measure of “merit” in the context of strength. I think the same is true in other fields as well, including scholastics.

Re:Tech journalism is less diverse than tech (2020) by harryh on Saturday November 20, 2021 @01:07PM (#29290348)
"less than 14% of the variance in SAT scores can be explained by SES in two massive and representative samples"

https://freddiedeboer.substack.com/p/please-think-critically...

Re:Tech journalism is less diverse than tech (2020) by commandlinefan on Saturday November 20, 2021 @06:07PM (#29292619)
What does SES stand for? Your link doesn’t appear to define it and I can’t find it defined anywhere else.
Re:Tech journalism is less diverse than tech (2020) by harryh on Saturday November 20, 2021 @08:26PM (#29293361)
Re:Tech journalism is less diverse than tech (2020) by paisawalla on Friday November 19, 2021 @06:54PM (#29284176)
> if you want something meritocratic, what we find is that testing ends up reinforcing existing social strata more often than not

Compared to what? It's not a criticism to simply state that standardized testing fails to be perfectly meritocratic.

Re:Tech journalism is less diverse than tech (2020) by zeusk on Friday November 19, 2021 @12:32PM (#29279824)
The physics and chemistry aren't any easier. I'd say at-least Physics had the largest delta compared to SAT subject tests or even APs.

While the SAT was asking about basic kinetic motion in classical mechanics, JEE was asking about wave-particle duality and calculating uncertainty in motion of photon-electron collisions iirc.

Re:Tech journalism is less diverse than tech (2020) by rayiner on Saturday November 20, 2021 @12:06AM (#29285727)
That’s basically how law school admissions is done in the US. It’s 95% an index score combining LSAT and GPA. It works fine. I think people attribute the desperation and intensity of the Indian system to the tests, but it’s more the limited number opportunities compared to the population.