I’ll do my best to make this an evergreen post: One I can just point to when we read yet another opinion article by the same types of people insisting that removing the SATs is going to harm first-generation and/or low-income and/or students of color and/or anyone who attends an under-resourced high school.
The latest is here.
The arguments are all bullshit, as I’ve written before; long version here and short version here. And while I would normally just tweet about this in the best tradition of Akil Bello’s #HateRead hashtag, I’m afraid threaded tweets aren’t the most long-lasting record of anyone’s thoughts. And this time, since they mentioned me in the essay, I thought a more formal response might be in order. So, a blog post.
As I mentioned in the linked post above, I’m weary of this argument. Well, of arguing. The SAT and ACT are dying, but like Freddy Krueger, they and their defenders seem to pop up on schedule in a pattern that seems just too well-sequenced to be a coincidence. See this from Richard Atkinson who tried to kill the SAT in the UC System (if you don’t want to view the whole thing, forward to about 2:28).
Here is the hate read you came for. And it starts in the headline.
The thing is, no one I know of has ever talked about banning admissions tests. But if you want to fire up those who have something invested in an outcome, talk about banning it (see also “cancel culture”). It’s frankly a cheap rhetorical device, but not totally unexpected. Even if every college in the country went test-free, the tests would not be “banned.”
Second, this paragraph.
Where to begin? For the sake of argument, let’s assume that the University of Phoenix (the archetype, apparently, for those not lucky enough to be “highly rejective” in another interesting rhetorical twist) and all the others from the mass of unwashed non-elites. (Don’t read that whole thing unless you really want to. Here’s the line you want from it):
Note the follow-up, rarely seen, double logical fallacy: Assuming that test scores are the same thing as academic ability is both a) the fallacy of begging the question, in which someone offers an unproven supposition as fact in support of an argument, and b) the fallacy of equivocation, in which two things are made equal when in fact they are not.
Finally, calling the National Merit Scholarship Program a “well-designed test,” much less “efficient, practical, and reliable,” is laughable.
Let’s move on. We can perhaps–possibly–overlook the use of the term “17-year-old minority girl,” to get to the heart of the problem here:
Notice how they’ve done it again: Working in the supposition of grade inflation, which Matt Barnum and James Murphy pretty much destroyed. The topic has become so ingrained that all the testing fan boys need to do is to make a casual mention of it. Mostly, though, notice that “standards of scholarship are not held in high esteem.” Passive voice, of course, to avoid any possibility of accusations of racism. And again, no one is talking about banning the tests. Literally no one.
Here comes the data. Sort of.
I never said the SAT is not predictive. In fact, it is, so I hope the authors will have the courage to issue a correction. I have said on numerous occasions that if you only had an SAT, you could make reasonably good admissions decisions. That’s because–by the College Board’s own admission–the SAT and GPA are consistent two-thirds of the time. (I’ll come back to this in a moment.) And no one–not even the testing agencies–say that the tests are better than GPA.
OK, you’re going to cite the UC Faculty Senate study that said tests are better than GPA. It is literally the only time I’ve ever heard anyone say that research supports this. Again, even the testing agencies say GPA is better. And Saul Geiser, who knows more about testing and its predictive validity in the UC System responded to the study, saying they’d left out some important variables.
My point–and the point of almost everyone not funded by the College Board who has studied this–is that tests add very little to the predictive equation once you look at everything else in the file. In other words, they’re redundant and not predictive. If you don’t believe me, perhaps the newly enlightened Dean of Undergraduate Admissions at Yale, Jeremiah Quinlan, might convince you.
Do the SATs predict graduation? Of course they do, to a point. Because if you’re wealthy, your score is higher. If your parents are college graduates, your score is higher. If your parents are not a part of an under-represented ethnic or racial group, your scores are higher. If your high school grades are higher, your test scores are higher. Tests by themselves? Meh.
This is from a document Purdue Enrollment Management submitted to its University Senate when discussing test optional. Even as a single factor at Purdue, SAT scores are really weak in predicting graduation. If you added the other factors, they might even turn up negative. (Note: This is a bad chart with the y-axis truncated to emphasize small differences. Still, note that the second and third bands have higher grad rates than the fourth).
What would a defense of the SAT be without #4 on Boeckenstedt’s list of Tropes?
This, again, is a lie. They are not fairer. They are not better. No one thinks this.
This? Nicholas Lemann and Joseph Soares would like a word about the history and the purpose of the SATs. The tests may have been designed to find “diamonds in the rough” but it was certain those diamonds would be of “superior genetic stock” who were somehow not wealthy despite their genes. Carl Brigham was a eugenicist, and the SAT wasn’t reviewed to ensure it had no racial bias until 1978.
Moneyball? Let’s look at analytics in selection. Here’s a piece I wrote about the NFL Combine scores and how they measure things that don’t really matter. And while it might be helpful for a lineman to be fast in a 40-yard dash, and it may be a sign of athleticism, it doesn’t mean you can block or tackle. Just like choosing the right answer from five on a test may be a skill you’d rather have than not have, but it’s not clear if that’s valuable in philosophy or history class. And using analytics can cause you to overlook a lot of talent.
Beyond that, the list of the top baseball players of all time includes mostly those who played long before there were analytics used to select them. Granted, this list is skewed to players who have played a long time prior to analytics, so in ten years we might see something different. But I’d hate to think of my youth without Willie Mays or Hank Aaron or Mickey Mantle.
Again, look at what Jeremiah Quinlan said. And remember that until the early- to mid-60s’, the period of greatest increases in educational attainment in the US, college admission in the US (at least outside of the Northeast) was, in fact, driven by high school GPA.
Finally, as James Murphy (again) has tweeted, grade inflation in high school has been a “concern” since the early 1970’s. It’s designed to distract you from the topic at hand, and to make wealthy white people fear that students from “weaker” high schools might be judged on the same things as their children.
Two closing points:
I don’t know the authors, but I’m flattered that they included my thoughts in their attack. I Googled them, and here’s what I found:
David H. Robinson is a professor at UT Austin. He’s written a few things in conjunction with College Board staff, that may or may not have been sponsored by College Board. If so, I presume he’d disclose that. But he seems a little snobby and appears to rush to judgment in his approach, not unlike someone who might write in support of the College Board. He’s liked tests since at least 2014, when this was published. (And I’ll admit to the logical fallacy of argumentum ad hominem here. I’m sort of over worrying about this stuff.)
I was unable to find much about Robert Bligh online that would speak to his knowledge about testing. Here is a summary I found of him in a book he wrote with Dr. Robinson.
And Howard Wainer was principle research scientist at ETS. So, some of the survival of the SAT goes to his ego, in all probability.
Inside Higher Ed needs to do a better job of disclosing when opinion pieces are written by people with some sort of financial or other interest to the topic at hand. It’s not the first time it’s happened: They previously published a piece supporting the SAT by two people who run a conference sponsored by College Board and ACT. The relationship does not mean the conclusions are wrong; it just should be disclosed. I’ve written to them about this before, apparently to no avail.
Remember the visualization on Discrepant Scores? I said I’d come back to it. If you look at that visualization of College Board data, you find that tests benefit men. They benefit white people. They benefit the children of people with graduate degrees. They benefit the children of high income people. So, look at who always writes these editorials. And ask what interest they have in the status quo.
A famous research study from 1988 might help you out. There are numerous reasons why people prefer the status quo. And many why they prefer to hold onto the tools they’re accustomed to, even if it (literally) means they might die.
Love tests if you like. Keep them at your institution if you do real research that says they’re important. I don’t care. I do care if you use specious arguments to support an opinion not founded in any real research.
Whew. That’s a lot. If you’re here, thanks for reading. I’d love to hear your comments, below.