Hacker News Reader

Comment by scarmig | Hacker News Reader

In theory, other measures of student achievement: grades, extracurricular activities, personal statements, letters of recommendation. GPA by itself isn't terrible for this, though pairing it with test scores is an improvement over both it by itself and test scores by themselves for prediction of academic success in college.

In practice: things that satisfy ideological goals and maximize fundraising outlook for the institution.

View on HN

According to a professor I know, the average quality/preparedness of students has dropped with the removal of the SAT where he teaches. A high SAT score was never going to get you accepted to a good school, but a low SAT score used to get you rejected. Removing that filter makes room for less objective measures (I know the SAT isn't particularly objective either) to get more weight.

To be clear, I strongly oppose removal of SATs as a metric used to evaluate candidates.

That said, there's a confounder here: schools which remove SATs as an admissions metric are schools that are looking for different things than academic preparedness/quality. It's likely that they're using methods and rubrics to deemphasize the predictive quality of GPAs as well.

GPA is fairly objective (though less so than the SAT) and predictive (more so than the SAT) as a standard. Accounting for the courses taken and the high schools the courses are taught at, you could select a class purely based on grades that is highly qualified and prepared (and incidentally would also have high SAT scores; GPA and SAT are correlated). But a university removing the SAT from admissions is also likely to be trying deemphasize the component of GPA that's predictive of quality as well.

To add one thing here: the professor in question is at a school in the top 20 worldwide. Almost every applicant had a basically perfect GPA from high school, but some of them had SAT scores that were 70th percentile or below. The SAT seems to "scale up" to that market much better than GPA as a predictor.

Right; GPA is a coarse representation of performance on coursework. A 4.0 at Stuy with coursework on real analysis and organic chemistry means something very different than a 4.0 at a struggling inner city public school where the hardest math class is Algebra 2.

My point is more that universities can (and do!) create a different representation of coursework performance that accounts for rigor. But a university that eliminates the SAT is also likely intentionally making that representation less predictive of undergraduate performance to allow weighing of things more than preparedness/quality as indicated by grades.

What makes SATs so important is that you get a very limited preparedness signal from kids with 4.0 at the crappy school; combining test scores with GPA allows for a selective school to get a much more meaningful signal for quality/preparedness.

My motivation here is to push back against the scores alone are enough idea, though it's an understandable reaction to the people who by all appearances think quality/academic preparedness should be a secondary concern in admissions.

According to the recent "Task Force on Standardized Testing" from the UCs (https://senate.universityofcalifornia.edu/_files/committees/... page 23) SATs are better predictors than GPAs for most demographics, although both together is unsurprisingly better than either alone.

This seems natural to me, as GPA seems inherently less objective (different schools, different teachers).

I think it's fairly strange that universities are getting rid of using them. One explanation is that it helps protect them from lawsuits like those against Harvard, as there is less incriminating evidence. I fail to see how it actually helps students.

> extracurricular activities, personal statements, letters of recommendation

These are just selection for middle-classness.

I would say upper to middle-upper class.

depending on which extracurricular activities, always some activity that most of people can't effort it, it is a indicator for distinguishing rich.

Reply on news.ycombinator.com