David Leonhardt writes today about the misguided campaign against using standardized tests like the SAT in college admissions. It's easy for lefties to hate on the SAT, but there was a time when lefties loved the SAT because it was a way of identifying smart kids who didn't necessarily grow up in great neighborhoods or go to great schools:
Administrators at Harvard, who pushed for the creation of the tests, saw them as a way to identify talented students from any background. The administrators believed that these students would go on to strengthen the country’s elite institutions, which were dominated by a narrow group of white Protestants, as Nicholas Lemann explained in “The Big Test,” his history of the SAT.
There may be some level of unfairness in SAT scores, since rich kids can afford tutoring and test prep and that sort of thing. But that turns out to be pretty modest. Other common metrics used for college admissions—high school GPA, recommendation letters, subject tests, strength of curriculum—are far more unfair. They strongly favor elite kids, which is the whole reason the SAT was widely adopted after World War II in the first place.
(It's worth pointing out that this was the era of the GI Bill and the expansion of college admittance broadly. Adoption of the SAT wasn't originally designed to identify more Black or Hispanic kids, it was designed to find poor and working class white kids who previously had neither the money nor the encouragement to get into college.)
And like it or not, tests like the SAT really do have a good track record of predicting success, especially at elite universities. This is from a recent paper that Leonhardt cites:
As the paper shows, SAT scores are also much more predictive of college grades than high school GPA. Ditto for success after college.
No university in America (as far as I know) relies solely on SAT scores for admittance. It's one piece of many. But it's a pretty good piece because it is predictive of success and it isn't very biased against poor kids or students of color. Tests don't always produce the results we'd like, but that's usually not because the tests themselves are bad. It's because they reflect a reality that's bad. That's where our focus should be, not on shooting the messenger.
We really should only require SATs for admission. Grades are a crock. The only reason grades are used is because it enables various forms of affirmative action. Unfortunately, it's also what leads to the dominance of Asians in admission. Asians get slightly higher test scores, but much higher grades.
We should also revert back to an earlier SAT. The current one is very coachable for the top 30% of students.
Make $280 per hour. Getting a job is not easy. In any case, you have access to a wealth of resources to help you with your work style. Become motivated to promote hundreds of jobs through job boards and vx20 career websites.
Take a look at this................................. https://careershome74.blogspot.com/
Yeah, this push boggled my mind from the very beginning. Like, every single other metric outside of testing are ones that are quite obviously biased in favor of wealthy kids. I hope this "no SAT" issue gets put to rest quickly.
I wonder if there is any correlation between pro- and anti- SAT opinions and how well that individual liked/felt that the SAT helped them...
Take a look at that chart. We're talking about first year performance. After that, the test's ability to predict performance drops off dramatically. You could get better correlations using zip codes and family income.
You write "take a look at that chart." Apparently you missed a good bit of the article.
You need to take a look at the whole Times article by Leonhardt. There is a comparison of correlations to attending an elite graduate school, and one comparing correlations to working at a prestigious firm. Those are not transient effects limited to the freshman year.
During the pandemic, when kids couldn't take the SAT/ACT, many colleges went to test-optional or test-blind. Many schools are "in review" on what their longer-term policy will be. I'd guess more will be considering or requiring test scores in future years.
As it is today, UCs and CSUs do not look at test scores (but in some cases will use scores for placement of incoming freshmen). UC test policy is probably the most contested issue in college admissions here. Don't know if any changes are in store. Elsewhere, private schools tend to be test-optional. Though Caltech, like some Ivies, do not consider scores at all. MIT, otoh, is requiring scores again.
Re test-optional schools: anytime I've seen school data, the percentage of accepted students sending test scores is invariably higher than the percentage of applicants sending test scores. That's not a surprise (nobody wants to send an optional low score), but it does show schools are probably still using test scores as one factor in admissions.
Another change: the "diversity essay," a mainstay on many college applications, seems to be undergoing some changes. More schools seem to be moving away from language implying race or gender (directing students to interpret "community" in their own way), or ditching the diversity essay altogether.
I don't see how colleges can continue to rely on essays of one kind or another, either. ChatGPT can spool out a great story about the challenges you overcame as a young Filipina girl growing up in LA.
Kevin, where have you been? Don't you know that in our modern parlance, any standardized tests lead to "stratification" and we can't have that. It goes against the principle of "equality of outcome", which is deemed a Law of the Universe, or something like that.
By the way, the new mayor of Chicago, when he was a teacher, deliberately did not conduct or encourage his students to take any of the "pre-" tests, which have a modest impact in creating a level playing field for those who can't afford tutors. Why? He opposed standardized testing so he made sure his students would prove him right by performing poorly.
i never understood the 'pre' tests anyway; you take the test and get a score but you have no idea what questions you missed or why, so not much value there as far as i can see
They have some value; for a lot of folks, they help ensure that when they take the actual, it's not the first time they are seeing that format/type of test.
Pre-SAT gets you scholarships.
See though, the problem with that graph -- of "elite" universities, which only includes "Ivy League +" -- is privileged grading curves.
Do this graph against the top 10 public universities instead.
Long, long ago, I was one of those white kids from a poor mill town whose SATs helped me get into an "elite" university. I think they still have a role to play, though it's much trickier since test prep courses are so prevalent.
What exactly is this chart showing? All the students are above average. 1250+ is top 25% of college students, and all the SAT scores on the chart are from successful students with GPAs between all Bs and all As.
The chart caption says "Elite Universities."
Among students at such universities, I expect the number with below-average SAT scores or mediocre high school GPAs is so small that it isn't worth including on the chart.
The SAT and the GPA are both terrible measures as they are single valued
In the UK and here in NZ the High School kids take subject exams - you can be great at Maths and Science and crap at English
Intelligence is simply NOT a single value
To be fair, the SAT isn't a single value either--it's two (one for verbal, one for math). They get added together as a shorthand, but most schools I think care about the breakdown between them in admissions.
As well as about GPA and the grades that go into it, as well as application essays etc.
In the UK and here in NZ the High School kids take subject exams - you can be great at Maths and Science and crap at English
Many American students do this as well. They're called "achievement tests," and they're extremely common among more academically gifted university applicants.
The SAT and the GPA are both terrible measures as they are single valued
This seems incorrect in the case of the SAT. As Kevin point out, the test has a very strong record—backed up by extensive research—in terms of its predictive value with respect to academic success at the undergraduate level.
The other issue is that SAT test prep typically raises scores by 100-150 points, the difference between the 25th and 75th percentile in college scores of admitted students, and on the low side between the SAT score of the 25th percentile and 'not admitted'.
A student self-motivated, or parentally motivated to do a month or three of test prep certainly is more likely to do better in college, so maybe the SAT does predict college success!
The impact of test prep depends on who you ask.
If you ask the test prep companies to do the study, it's 100-120 points (The Princeton Review, Khan Academy).
If you ask the College Board (which develops the SAT) to do the study, it's 33 points.
If you ask someone without an obvious and severe conflict of interest to do the study... you get an answer that doesn't show up on a quick Google search, I guess, since I couldn't find any with numbers attached. What source are you using?
Hmmmm....
So college level GPAs are the assumed best measuring stick.
But high school level GPAs are determined to not be good measuring sticks?
I sense a bit of a problem here....
You took a shortcut in your thinking that lost sight of the subject.
The challenge is to identify the best (or at least the better) *predictor* of college success. Nobody said high school grades are not a good measure of anything. What they measure is, of course, high school success . What was found is that they are not such a good *predictor* of college success.
No I didnt take a shortcut, but I do like that turn of phrase. In fact, youre statement is the shortcut that loses sight of the subject.
Youve (and others) assumed that college GPAs are the best measure of college success. But this is just an assumption and its an odd one to make when discounting the predictive value of HS GPAs.
This data appears to show that SAT/ACT scores are better predictors of college GPAs. That college GPAs are the best measurement or best guide for who deserves a spot in college is merely an assumption and the arguments against HS GPAs should be a warning that this assumption might not make sense.
Deciding what to measure and making values judgements based solely on what is easiest to measure is a huge problem in education and life in general. This article should make this obvious....but measurement bias is a powerful one.
Yes:
I remain gobsmacked that this isn't immediately obvious to everyone. TL;DR: When your metric becomes a target it ceases to be a measure.
From Kevin:
"It's easy for lefties to hate on the SAT, but there was a time when lefties loved the SAT because it was a way of identifying smart kids who didn't necessarily grow up in great neighborhoods or go to great schools"
I hesitate to read too much into anything, so sue me if I'm wrong, but providing a mechanism (no matter how imperfect) by which people from different social and/or economic backgrounds all get as much of a fair chance as possible sounds like the argument for equal opportunity.
But, unless you've been asleep for a few decades, it's clear that modern lefties have abandoned the concept of equal opportunity. Or, perhaps it may be more accurate to say that there's a cohort of the Left (perhaps even a radical cohort) that never accepted the equal opportunity premise in the first place, and now they find their influence (in education, Democratic politics, publishing, media, and elsewhere) either ascendant or at least greatly out of proportion to their own (relatively small) numbers.
Instead, the idea of a "systemic" problem has more or less become unquestionable dogma on the left. This perspective is typically found in reference to issues like "systemic racism." The SATs, of course, are a part of that corrupt "system," and as such they must go.
And the only way to circumvent an irredeemable system is to give up on the pipe dream of gradual progress and instead demand equal outcomes. Out with equality, in with equity.
Here is an example from the city of San Francisco: Lowell High School. I have no personal connection to the school, but it was a big story a few years ago. It is (or was) an apparently well-renowned High School that includes the likes of Supreme Court justices among its alumni.
But with the wave of radical Leftyism following the incredibly counter-productive "reckoning" during the summer of 2020, and using the pandemic as an initial pretense, the San Francisco school board first temporarily and then permanently decided to remove the admissions test to Lowell.
Is this an exact parallel to SAT tests? No, but there are similarities. In each case, a cohort on the Left wants to circumvent want it sees as a hopelessly corrupt system in order to further promote their idealized utopia.
And Lowell is in fact a good example since the SF school board was fairly explicit in its reasoning: https://go.boarddocs.com/ca/sfusd/Board.nsf/files/CF94CB0ADE93/%24file/Resolution%20-%20Lowell%20Interim%20Admissions%20Policy%20for%202023-2024%20SY.docx%20-%20Google%20Docs.pdf
"[O]n February 9, 2021, in response to concerns about ongoing, systemic
racism at Lowell High School, the Board of Education approved Resolution 212-2A1"
This was not only motivated by the belief in a "systemically racist" system, it was also at least in part prompted by the modern Left's rejection of the entire concept of merit.
Believe it or not the California Community College (CCC) system provides an online DEI glossary of terms, and that positively Orwellian repository of doublespeak define merit as follows:
"[M]erit is embedded in the ideology of Whiteness and upholds race-based structural inequality. Merit protects White privilege under the guise of standards (i.e., the use of standardized tests that are biased against racial
minorities)"
I will also just quickly add that all CCC teachers are now required to teach and promote these DEI concepts and risk losing their jobs if they fail to comply. In other words, asserting a belief in merit is now a fireable offense in the lefty circles that Kevin is discussing: (https://govt.westlaw.com/calregs/Browse/Home/California/CaliforniaCodeofRegulations?guid=I5F7BF9004C6911EC93A8000D3A7C4BC3&originationContext=documenttoc&transitionType=Default&contextData=(sc.Default)&bhcp=1)
But these lefty beliefs and practices bashing standards and merit are wildly unpopular. Three members of the SF schoolboard in question were recalled by San Francisco voters (hardly trump-style reactionaries), and I believe Lowell High School has reverted to its admissions test for future graduating classes.
So Kevin, to his credit, appears to be positioning himself against the ascendant, and profoundly unpopular, Lefty culture of our times. And I believe that we all need to do everything we can to isolate this faction from the Democratic party in order to give Biden his best shot at defeating trump.
The math scores on the SAT / GRE are critically important indicators for math heavy subjects - STEM ... In general, students need to be up to a reasonable level of math proficiency before they can pursue these subjects - and they are going to need a reasonable mathematical aptitude as well for some of the fields.
Please look at Appendix figures 3 and 4 (especially panels C and D) and you'll get a better idea of the issue with SAT/ACT scores. The authors report that the scores correlate with college achievement when segregated by high school privilege, but if you segregate them by student privilege (e.g. low income students or first generation students) the correlations are MUCH weaker.
I also have questions about the presentation of the paper (for example, what is a data point and where is the representation of variability in each data point?) but the main issue that educators have noted is the weakness of college admission tests in predicting success of these less privileged students.
The push against standardized tests is underlain by the Kendiist assumption that anything that shows a racial disparity is racist. What this misses is that lots of things influence other things. Racism moves through our social system like groundwater pollution, often becoming visible far from where it originated.
We see racial disparities in test scores, yes. But the tests are not racist - rather, they are accurately measuring the impact of racial disadvantage on kids' intellectual development.
Kevin wrote about this before in the context of IQ tests: https://www.motherjones.com/kevin-drum/2019/06/heres-why-the-black-white-iq-gap-is-almost-certainly-environmental/
The College Board has never claimed to predict anything beyond first year GPA. That's the only thing your graph shows and that only for high end students at elite universities. Color me unimpressed.