When MIT, which was the first among highly-selective schools to unapologetically reinstate the testing requirement two years ago, was held up here as an example, it was allowed that while yes, it mi-ight be right for MIT (if that!), it surely cannot be right for all the other schools whether they exceed it in minimal required rigor (Caltech) or not (presumably everyone else).
4 Likes
Many large universities with less selective overall admissions have honors programs and scholarship programs that are much more selective.
Large universities with a large number of applicants may also benefit from getting more data points that can be plugged into a non-holistic formula.
1 Like
I think itās still tremendously hard for some.
Not only STEM. What about reading comprehension (the reading subsection on the ACT for example). Imagine being an English major and struggling with Shakespeare
While I agree in sentiment, whether the ACT especially can differentiate between a bad reader and a ānot fastā (but detailed) reader may come into play since its such a time sensitive test.
While the same can be said about STEM, Iāve always been of the stance that math facts should be memorized, not because thatās the best way to learn math, but rather because itās the most practical way to be highly conversant in algebraic math. None of the math in the tests is āhardā but wielding them effectively at that speed is a useful skill in the mathy stem majors.
But I donāt know. Iām really basing it on the fact that schools like MIT, Yale, Dartmouth, Brown, and others have mentioned in some form the correlation with longevity or success in the stem majors. They didnāt seem to make similar statements re: ebrw to my knowledge.
2 Likes
Thatās fair. Although Iāve always been of the understanding that speed of recall of stored information is an attribute of intelligence, or in this case aptitude (not sure of the difference). Imagine watching Jeopardy and instead of signal devices everyone was given as much time as they needed to come up with an answer. Maybe Ken Jennings isnāt a deep thinker, but his speed is a sight to behold!
MIT is primarily a STEM college. When did the others say anything about math/STEM and not EBRW/reading?
The UC study found SAT score was most predictive of within course GPA in social sciences (SAT + HSGPA explained 19% of variation in college grades) and least predictive in engineering (SAT + HSGPA explained 11% of variation in college grades). Overall the 2 SAT subsections were roughly equally predictive, but reading was slightly more predictive than math for humanities and social sciences; while math was significantly more predictive than reading for natural sciences and engineering.
3 Likes
SAT and ACT are not aptitude tests, they are achievement tests.
There are many students who may have high aptitude but suffer in underperforming K-12 schools and donāt do well on ACT/SAT tests because they havenāt been taught/learned what their peers haveā¦simply because of the schools they attend. To be crystal clear, I am not saying all humans have the same level of native intelligence. I expect we can all agree that the US K-12 education complex leaves much to be desired for many students, and does in fact leave children behind.
3 Likes
Any chance there are kids out there in underperforming schools who still knock standardized tests out of the park? Is this an impossibility? Because I think these are the kids that the newly restored test required schools are saying they are looking for.
It would be interesting to see what groups historically have had the highest and lowest SAT scores as that might shed light on the winners and losers moving forward.
At least for Harvard, the data from the Fair Admissions case and prior studies shows really interesting discrepancies between the admit rates among races and religions relative to those groupsā SAT scores.
If high SAT scores really lead to success (as suggested correct by the preponderance of Asians among Phi Beta Kappa alpha and Rhodes winners), other readjustments are likely coming as to the composition of the class. Hopefully those readjustments will not be considered as driven by bias if they simply follow the SAT distribution. Only fair right?
Of course there are students like that, and nothing I said suggested otherwise.
It depends what you mean by āgroupsā. For example, the answer would differ somewhat between HS graduates vs admits to selective college X.
Previous studies have found the following groups tend to be overrepresented among test optional admits compared to test submitter admits. That is scores are more likely to be a relative weak point in application.
- Women
- Lower SES
- First Gen
- URMs
- More Rural States (midwest, Maine, W Virginia, ā¦)
- Public HS
My feeling is that some high-aptitude kids in underperforming schools still manage to learn more than their peers from other sources (home, the internet, the library, etc). These kids would be better positioned to do well on tests, compared to their peers.
Others, who may also be naturally high-aptitude kids, might not have the same opportunities for educational experiences outside of school, or might not be able to take advantage of such educational sources because of other factors.
I was a low-income kid in a school where almost none of my peers went on to college (maybe one other kid in my HS classā¦). But our city had a really nice public library and it was peaceful there and a good place to hang out after school (when no adults were home at my house). So it just happened that I spent a lot of time there reading books and pursuing my nerdy interests with a few other nerdy friends. Something like this is going to affect how well a student does on tests (and in college), I think.
1 Like
My point is that if SAT = Success in school and thereafter, shouldnāt the average SAT score and admission rate be the same among races and across religions (after normalizing for SAT scores)?
Unless one race or religion is just ābetter peopleā or more likely to be better at sports or more āauthenticā generally. Well actually, I guess in that 70 pct of athletic recruits are caucasian, that group must be best at sports. Especially given athletesā much lower grades and SAY scores.
From Admission Beat podcast (Dartmouth, Lee Coffin) that I transcribed above:
(emphasis mine)
LC: Lee Coffin (Dartmouth)
JQ: Jeremiah Quilan (Yale)
JQ:
So at Yale, we were looking into this question before the pandemic just to understand how important standardized testing was in predicting how well a student would do at Yale, and it turns out actually that the SAT or the ACT is the single best predictor of a studentās academic performance at Yale. Um, and particularly the math SAT, in persistence in some of our science majors. Um, this is a bit counter to the national research, which suggests that GPA is a bit more predictive than standardized testing. But at an institution at Yale [sic], um, we find that the standardized testing is the single biggest predictor.
LC:
Yeah, I would just add to you that weāre studying the same thing and thatās the emerging storyline here [Dartmouth] as well.
I can see if I can dig up Quinlanās and Coffinās individual statements on the matter on their respective podcasts as well, if that helps.
1 Like
āBirds of a feather flock togetherā
I think itās likely that inherently high aptitude kids would congregate at places like libraries, etc. Not always, but more likely.
Michael Jordan figured out he was good at basketball, and you could find him frequentingā¦basketball courts.
We are talking about relatively weak correlations. Rather than SAT = ā¦ Itās more SAT is correlated with ā¦ such that SAT in isolation explains x% of variance, where x is generally a small minority.
shouldnāt the average SAT score and admission rate be the same among races and across religions (after normalizing for SAT scores)?
Different races/religions have different SAT score averages among the HS graduate population and have different rates of applying to Ivy+ colleges. There are many contributing factors to why these differences exist. If colleges can legally favor particular races/religions, they may intentionally give a boost to students from a particular group, leading to lower average admission qualifications (many qualifications average lower, not just SAT).
2 Likes
I didnāt listen to that podcast. The analysis linked on the website looks at first year GPA (across all subjects) and combined SAT score, rather than just math/science/STEM.
Strange qualifier. Seems only at Yale this is the case, not the wider realm of reality.
I appreciate that JQ was fitting claims to evidence.
Getting out in front of my skis I suppose. Toucheā