Hacker News Reader

Comment by ckemere | Hacker News Reader

While the article claims grades are less predictive than test scores, UTs policy is actually intended to have a similar locally-normalized effect, leveraging the wide variability of high school quality in TX…

View on HN

like many things, there are some pretty significant unintended consequences though. In my district in Texas, parents would send their kids to the worst high school in the district because it was dramatically easier to be top 4% and get auto-admit to UT that way

Granted UT Austin is really well ranked, that seems like a terrible move for the kids. Like, great, you got in to this really well regarded school, but your preparation to get there may well be considerably worse than nothing at all.

At the other end, the last time I attended a graduation at Highland Park TX, 4% was about the number of National Merit Scholars in the class.

They typically go out-of-state, but the competition for the next 4% is insane.

Goodhart’s lawing the Asian nerds into a high school with cholos and gangbangers

there ought to be a sitcom about this

Dartmouth wants to both achieve SES diversity and gets kids who are likely to succeed. If selecting the top of each class in the state gives them the diversity but doesn't give them the success rates, it's a less effective strategy. Their data seems to say that accepting medium-high SAT scores from poorer kids gets them both the diversity and the higher predicted success rates.

Reply on news.ycombinator.com