SAT (and ACT) scores are highly predictive of academic achievement at Dartmouth [pdf]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nostrademons on Monday February 05, 2024 @12:25PM (#39263780)

The point of the report is that SAT scores are less correlated with socioeconomic status than other metrics used by admissions like GPA, high school reputation, extracurriculars, essay content, etc. They're still correlated. One of the interesting findings of social science research is that a lot of different metrics like IQ (itself a correlation between different subtest performances), GPA, parental income, child's income, parental wealth, child's wealth, college acceptance chance, eventual education level, ability to delay gratification, general health, etc are all correlated - that's the idea behind the g-factor, a measure of general intelligence that highly correlates with wealth & income. But this research shows that among those correlates, SAT scores are actually less correlated with socioeconomic status than the other metrics that admissions committees can use.

This is also why the SAT was adopted by Harvard in 1934. Prior to that, most students attending Harvard came from private New England boarding schools that could only be afforded by the upper class. They saw an opportunity to diversify the student body by testing the general public and admitting those that scored highest. As an added benefit, the SAT was much easier to grade than previous essay-based admissions standards, and so it reduced cost for the University.

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nyrikki on Monday February 05, 2024 @12:53PM (#39264200)

A Harvard, Princeton, and Yale alumni survey found three-quarters live in zip codes that rank in the top 20 percent by income and education. Half live in the top 5 percent of zip codes.

So their efforts didn't work.

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nostrademons on Monday February 05, 2024 @04:04PM (#39267150)

What would the counterfactual be? The population of elite New England boarding schools, at least in 1934 when the decision was made, was pretty solidly top-1%. (It too has diversified recently because of efforts by those top prep schools to accept diverse applicants on scholarships, but a major applicant criteria they use, particularly for underprivileged applicants, is...get this...the SSAT!). If they're accepting 25% of students from outside the top 20%, that is way better than the situation 100 years ago.

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by sfink on Monday February 05, 2024 @01:30PM (#39264854)

Only if your only criterion for success is diversity. I don't think the intention has ever been to recruit evenly from all zip codes or whatever. My understanding is that the idea is to have a substantial portion come from a less advantaged population. (Heck, these universities still need to have a lot of wealthy kids so their families give them lots of money. A dozen self-made millionaires aren't worth much compared to a handful of trust fund billionaires.)

If I flip your numbers around, it actually sounds pretty good: a full quarter of the alumni come from the bottom 80% of zip codes!

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nyrikki on Monday February 05, 2024 @01:35PM (#39264949)

The priority is a large educated workforce while not tossing away potential talent because they were born to the wrong parents.

But note that teams built from cognitively diverse members are far more productive while homogeneous teams dramatically underperform.

https://hbr.org/2017/03/teams-solve-problems-faster-when-the...

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nick7376182 on Monday February 05, 2024 @01:03PM (#39264376)

Is that for alumni, not incoming students?

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by nyrikki on Monday February 05, 2024 @01:39PM (#39265038)

Currently, the Ivy League institutions are estimated to admit 10% to 15% of each entering class using legacy admissions.

Add that to the zip code correlation that is the strongest indicator of SAT scores and it does matter.

Students without parents that are college graduates are also at a disadvantage no matter what their potential is.

It is one of the barriers to social mobility.

[ Parent | on HN ]

Re:SAT (and ACT) scores are highly predictive of academic ac... by laverya on Monday February 05, 2024 @04:48PM (#39267711)

> Students without parents that are college graduates are also at a disadvantage no matter what their potential is.

It is possible both for SAT scores to predict a person's potential to contribute in cognitively demanding fields AND for there to be a strong correlation between SAT score and parental income/zip code/etc. In fact, this is true even of a theoretically perfect measure that can somehow peer into the future of every prospective student were they to be admitted.

Is there some reason you expect that the children of people who did not go to college will have the same level of potential as those born to college graduates? Would this hold true of the children of professors, or Nobel winners? Is there any place for 'nature' in your worldview, or is it purely 'nurture'?

[ Parent | on HN ]