[collabs]

something that came to my mind as I was reading the comments here -- the thing is that in the quest for professionalism, we have sidelined a lot of people who would be good at teaching in favor of people who are good at jumping hoops. there is a famous quote saying "when the measure becomes a target, it stops being a good measure"

[ryukoposting]

Goodhart's Law (that quote) is actually one of the motivations for moving away from ACT and SAT as college entry benchmarks. "Teaching to the test" is a rampant problem in the US.

UC is seeing flaws in departing from those benchmarks, though. The thing is, % of students getting admitted to college is itself a measure for schools and school districts. If GPA is how you get kids into college, well...

It's not a teacher problem, it's a district and state problem. As a teacher, if kids are failing your classes (which nowadays seems to be "getting anything less than an A") your school district blames you.

To me, it seems that Goodhart's Law is an inherent problem for education in the information era, no matter how you cut it. If there's one good thing that can be said about ACT and SAT, they're relatively difficult for schools to game. GPA inflation is trivial.

[antonyt]

I've never understood the "teaching to the test" argument against these tests. Take a look at some of the math SAT sample questions: https://satsuite.collegeboard.org/media/pdf/digital-sat-samp...

How would you "teach to the test" for these in a way that looks different from just teaching arithmetic, algebra, trigonometry, etc?

[ryukoposting]

With standardized testing, you lose a lot of detail about how the student got to their answer. Nuance is lost. But yeah I agree, with math it seems like that isn't such a big problem compared to, say, reading or science.

It does create some perverse incentives, to be sure. "Test mills" are an ongoing issue, especially in urban areas, in both public and charter schools. Basically admin guts all liberal arts programs, theater, music, history, etc, institutes some draconian discipline system, and kids just do practice tests over and over until they graduate from high school. Great standardized test scores, and virtually zero practical value to be had from the education the kids received. I know someone who got a 30 on the ACT and didn't learn that Africa was a continent and not a country until 9th grade.

[indoordin0saur]

I agree that Goodhart's law doesn't really apply to well designed tests.

[zahlman]

> ...actually one of the motivations for moving away from ACT and SAT as college entry benchmarks. "Teaching to the test" is a rampant problem in the US.... If there's one good thing that can be said about ACT and SAT, they're relatively difficult for schools to game.

You understand that you're contradicting yourself here, yes? The entire point is to have a difficult-to-game test. Teaching people to do well on the SAT looks an awful lot like actually getting them to understand the things that the SAT is intended to ensure they understand (plus a little bit of generic test-taking skill that would apply equally well to any test in the same format). And if you don't have that, you only have things that are worse.

[matwood]

For almost all math at the HS level, teaching to the test is exactly what you want.