Applicant success belies SAT's value
Wake Forest freshmen are more high-caliber, diverse without test.
By Jane Stancill
for the Charlotte Observer
Across North Carolina and the nation, high school seniors are sweating their college applications and fretting about one number: their SAT score.
But not students aiming for Wake Forest University, which no longer requires students to submit the standardized test score. Wake Forest was the first highly ranked research university to announce the move away from the SAT in 2008.
Since then, the university in Winston-Salem has become more racially and socio-economically diverse. Pell Grant recipients almost doubled. Students of color increased from 18 percent to nearly 23 percent.
Along the way, the university also noticed an uptick in the number of students with an exemplary high school track record, which, research shows, is the best predictor of college success. The percentage of Wake Forest first-year students who graduated in the top 10 percent of their high school classes grew from 65 percent in 2008 to 83 percent last fall.
Thank you for this, Terry. We should all buck the SAT requirement. It's terrible.
ReplyDeleteWhat about the LSAT requirement for wannabe law students? Hell, there are plenty of ABA-accredited law schools that will gladly accept students with terrible exam scores, weak-ass UGPAs, and pathetic majors, i.e. "Political Science," History, English, etc.
ReplyDeleteOh my God. Nardo is a law student. And he shits on a self-chosen set of "pathetic majors." God, man, you are in the wrong place.
ReplyDeleteHistory, man. Now there's a bogus major. Who the fuck cares what happened a long time ago? It's not like America ever lost a war or anything. Jeez.
ReplyDeleteAt the risk of actually getting an answer...
ReplyDelete... what sort of undergraduate major do you think is apporpriate for law school?
Yes AdMonk ... this is how to advocate for the dismantling of the SAT machine -- using empirical evidence of its lack of validity to convince the powers that be to stop relying upon it.
ReplyDeleteHowever, cheating on the exam?
Sorry, still wrong.
This is not empirical evidence of a lack of validity. This is a direct and obvious consequence of dropping a test that taps heavily into intelligence, which is strongly correlated with race. You get more students lower in intelligence and SES, who are on the average going to be more racial and ethnic minorities, as well as qualify for more Pell grants. Not surprising, and not empirical evidence. They also saw an uptick in highly qualified _applicants_ based on high school experience - not surprising, considering all the press surrounding this change. Wake Forest got on more students' radars. That has nothing to do with how well the test works: rather, it's an advertising effect (they are better able to attract customers).
ReplyDeleteThe link between SAT scores and college grades is well established. The SAT is valid, and there is a mountain of evidence supporting that. Bringing in students with higher SAT scores results in higher grades, end of story.
The real issue is whether or not that is fair, since there are other factors that predict student performance nearly as well. Why use something that shows racial differences when you could use something else (or several something elses) and not see those effects? And perhaps even more critically, don't the students with low intelligence and high motivation need us the most?
Um, what the hell are you talking about intelligence correlates with race, and "lower in intelligence" people are "on average, going to be racial and ethnic minorities"? This is not the 1920s. No educator should be walking around with that kind of crap in his head.
ReplyDeleteEven the guy who invented the IQ test, Alfred Binet, ended up saying that IQ ITSELF did not correlate with intelligence, as there were kinds it didn't measure. He also stressed that IQ was not genetic but greatly influenced by background. And he was horrified by Americans' use of his test, especially by their use of it to "prove" racial superiority, as he invented it to advocate what we would now call differentiated instruction, so all children could be educated. If even IQ correlates with socioeconomic status, then the SAT is hardly a guarantor of intelligence. Or of racial superiority.
Christ on a creampuff. Maybe I dreamed this racist garbage and it will be gone when I wake up.
I do not believe that Ricardus is an educator. I believe that Ricardus is a troll of a particularly vile and racist kind. I request that his comment be removed. You're welcome to delete mine too once it's no longer relevant, RGM.
ReplyDeleteAnd I think it is important to keep the odd racist comment around, at least for a bit, so that we may all recall that the big "R" is still out there--maybe not THRIVING, but there and as malignant as ever.
ReplyDeleteI found Ricardus's comment perhaps poorly worded, but not overtly racist. I don't really know any statistics on the issue, but in my mind it stands to reason that IQ scores for racial minorities would be lower for the simple reason that even in this day and age, racial minorities are more likely to be in the lower class and have less educational opportunities. So minorities aren't less intelligent based on any innate quality, but rather as a result of their social and economic condition (which itself is the result of historical disadvantages, continuing racism, and bad policy).
ReplyDeletePlease correct me if this is just plain wrong! This is how I have understood the race and IQ issue, but I have never done any significant research on the topic.
@Ricardus: Gee, for someone who is the past has been concerned with research versus anecdotes, if you really are an academic at all I dare you to provide some actual evidence for your sweeping generalizations.
ReplyDeleteCurmudgeon (can I call you Mudge?), if Ricardus had said that IQ scores correlate with racial/ethnic minority status, that would not necessarily be racist, for the very reasons you cite. However, he did not. He said that INTELLIGENCE correlates with race/ethnicity, and went on to say that students lower in intelligence AND SES are more likely to be racial minorities. Finally, he said that SAT tests "show" racial difference, which they do not. You cannot line up a group of SAT scores and figure out which scorers belong to which race.
ReplyDeleteThe comment could be just poorly worded, but as we say in composition classes, there is no good thinking without good writing. If you sound racist because of poor word choice or garbled syntax, then people are going to assume you are racist. So I say leave the comment up and let him correct himself. Or not, and then we know.
@Reg W.
ReplyDeleteHis name is NANDO, and he is a graduate of the Drake School of Law. He runs a blog, www.thirdtierreality.blogspot.com on the scam that is law school.
I, too, would be interested in hearing what Nando thinks are non-pathetic majors for those planning to go to law school. The highest scorers on the LSAT are math/physics majors, with economics and philosophy/theology majors tied for second place. History comes in much lower, but I think it's an excellent major for prospective lawyers.
ReplyDeleteI'd give economics majors the edge as far as being the most likely to figure out whether getting a law degree is going to be worth whatever debt they incur. Philosophy majors - assuming today's philosophy majors actually do the reading for their courses - would have the edge when it comes to getting through hundreds of pages of cases per week. (I always wonder how students who hate to read anything longer than a tweet think they're going to make it through even the first semester of law school.)
The two lowest ranked majors? Prelaw and criminal justice. (The latter is not what people who want to practice criminal law study. It's what people who want to be prison guards and POs study.)
Strel, not sure how cool I am with providing extra info on commenters here even if said commenters have not made it too hard to find more about them...
ReplyDeleteThat being said, Nando doesn't seem to have very good advice. I'd rather not be discipline specific, but I work closely with this group of people, and the more successful law school students are not only those who make it to the top 5 or 10 schools. Farming schools exist in every single region, (schools to farm out future lawyers) and those regional firms are not going to travel all the way to New York to hire.
I think Aware and Scared is really worked up on cheating. Still. As though it isn't rampant in our education system. And that continues to confuse me, considering the content of this website.
AdMonk, Aware and Scared isn't "really worked up on cheating" precisely because of the awareness of it being rampant in the academy AND the content of this website.
ReplyDeleteWhat perplexes Aware and Scared are members of the academy and this community openly advocating (if not outright engaging in) the practice particularly to gain access to the academy.
Given the content of this website, that would seem to be hypocrisy writ large.
We're products of the same system as our cheating students. I'm not advocating cheating (as I clearly stated in the last thread by referring to cheating as "shooting one in one's foot" and using the word "stupid" to describe the decision to cheat about 4 times in a single comment). I'm just saying it isn't shocking, it's hardly the only time it would happen, etc.
ReplyDeletePeople keep talking about this automatic honesty thing at 30, or how impossible it is for them to have ever broken the law. That's bull shit. Everyone has broken the rules at some point. It's not good, but the system being circumvented via SAT cheating is terrible as well, so stop getting your panties in a bunch, everyone. Cheating happens. *gasp*
Oh, and just to clarify, becuase you seem to think that my engaging with the realities of life means that I support it.
ReplyDeleteCheating is bad.
I can't believe I actually have to write that in order to ward off accusations that I am pro-cheating.
For fucks sake.
@Frog: Mudge is fine! Better than CiT I guess (just as Frog is better than FaT!). I guess I gave Ricardus the benefit of the doubt, since it is only a blog comment and infelicities slip in. But we'll see if he corrects himself.
ReplyDeleteActually, I'm (in part) an intelligence researcher. I teach primarily doctoral courses, with an undergrad course thrown in every year or two.
ReplyDeleteRemember that group differences in intelligence/test scores/etc does not mean that any particular student of a particular race will have higher/lower lower intelligence scores (this is the ecological fallacy), nor does it mean that every person of a particular race has higher or lower intelligence than anyone else. They are called group differences for a reason - they are differences in means of groups, not individuals. If this concept is confusing to you, I suggest reading up on the definition of variance, hypothesis testing, and what ANOVA entails.
Ignoring the facts about group differences does not make the realities go away (which is a common approach by educators without any background in social science research, and only a gut feeling that "there _shouldn't_ be differences so anyone that says otherwise must be racist" - this is just ignorance).
By the way, when I refer to intelligence, I am more technically referring the general psychometric factor between cognitive loaded tests (common called "g") - what you get when you give a battery of cognitively loaded tests (verbal, non-verbal, language-agnostic, etc - the SAT is one example, but so is one flipping 3D shapes in your mind and drawing the result) and extract the common variance between them. I've given a variety of such tests and conducted the analyses myself many times - the general cognitive factor always appears, and there are always mean differences by race. As much as I wish it weren't true, it is. It is a far more healthy approach to admit this, investigate the cause, and try to correct it at a systems level (e.g. interventions for schools, states, nations, etc). Sticking your head in the sand solves nothing.
And as for all the various personal attacks toward me, I consider spreading misinformation about the SAT based on personal opinions in order to add to the hysteria and paranoia surrounding testing a much worse offense than speaking candidly about current research. If you don't like what I'm saying, start a research program and help us fix the real problems underlying these realities.
As a side note, IQ is an archaic measure of intelligence and is not used in current research. It does tap into intelligence somewhat, but it misses many components of intelligence (criterion deficiency) and also taps into many related but distinct constructs like memory (criterion contamination). As soon as you hear someone quote "IQ" at you means you are talking to someone ignorant of at least the last several decades of research.
ReplyDelete@Academic Monkey
ReplyDeleteHe doesn't teach, he is a graduate with a terminal degree, so he is free of the spy games we play here. I don't know what his real name is, nor do I care - I mentioned those facts that I do know because of Reg's rolled-eyes "he's a law student" comment.
@Ricardus: I am now genuinely curious. What "other factors" predict "student performance nearly as well" as the SAT, how well does the SAT do it, and what experimental controls assure that any of this is valid and reliable?
ReplyDelete^^ The SAT is a complete POS of a test. I'm proof of that when I spent 3 months studying how to game it, and my score went up 270 points.
ReplyDelete@StockStalker - One person is not proof of anything. Most of the studying effects come from relief of text anxiety and increase familiarity with the text format (a sort of re-test effect). Then there are cheaters, of course. But that's just a fact of life - somewhere around 75% of college students admit to cheating at some point in college.
ReplyDelete@Froderick - It depends a bit on what you're trying to predict. For example, you don't necessarily use the same predictors to determine first-year student performance versus four-year student performance. It also changes based on how you define student performance: are you interested in final grades, on overall GPA, on improvement from pre-college assessments, on social interactions, on community engagement, on knowledge gained, on skills gained, on satisfaction with college, on retention, on entry into STEM majors, etc. The heart of the problem is that you have to come up with an "index" to make a single selection decision based on all of this information - so each college/university/etc must decide what combination of student qualities they find most important in order to create that index. If you weight SAT too strongly, you'll end up with high grades, but you lose diversity. But if you ignore it completely, grades will drop unless you have other cognitive measures in your battery somewhere (high school grades are a reasonable proxy, but you end up with too much between-school variance, such that mid-quality students at high quality schools are disadvantaged in relation to high-quality students at low-quality schools - whether you think that's ethical or not is a personal decision). There is no magic bullet, which is why we're all still arguing about it.
Here is an excellent recent summary that will answer all your questions:
Sinha, R., Oswald, F., Imus, A. & Schmitt, N. (2010). Criterion-focused approach to reducing adverse impact in college admissions. Applied Measurement in Education, 24, 137-161.
Abstract: The current study examines how using a multidimensional battery of predictors (high-school grade point average (GPA), SAT/ACT, and biodata), and weighting the predictors based on the different values institutions place on various student performance dimensions (college GPA, organizational citizenship behaviors (OCBs), and behaviorally anchored rating scale (BARS)), can increase the proportion of some ethnic subgroups often disadvantaged by the use of only traditional measures such as the SAT/ACT. The sample consisted of 836 students from 10 universities across the United States. Results show that meaningfully different proportions of groups would be admitted to universities when the predictors included noncognitive measures and the weights for the various components in the battery were based on performance dimensions other than first-year GPA. These dimensions should reflect institutional values.
@Ricardus - Clarify this for me: are you saying that there are differences in TEST RESULTS between members of different racial and socioeconomic groups? Or are you saying that test scores indicate that there are INHERENT (possibly biological) DIFFERENCES between people of different racial groups, or between the rich and poor?
ReplyDeleteBecause the first acknowledges the social reality that race and SES change how people interact with social institutions, and the other is scientific racism of the sort popularized by J.F. Blumenbach. (Worth noting that even he realized he was wrong by the end of his life...too bad so-called scholars today insist on perpetuating this.)
Ricardus - have you read Gould's _The Mismeasure of Man_? Identifying g with intelligence is at best ignorant and at worst disingenuous. For bonus points, note that racists love to do it!
ReplyDeletePerhaps this is a field/domain difference. In social sciences research, "group differences" only implies "we measured something and found mean differences between the groups." If you accept that that measurement accurately reflects the underlying unmeasurable construct (e.g. SAT scores in part represent intelligence), then you can extend that logic to the construct.
ReplyDeleteI am not sure why everyone misinterprets this when discussing intelligence - I suspect heightened cultural sensitivity. If I say "there are gender differences in height," no one automatically assumes I am saying "all men are taller than all women" or "if you sort people by height, you can tell which ones are men and women." It's obvious: on average, men are taller. But talking about intelligence in the same way suddenly makes you a racist. There are also gender differences in spatial reasoning (a component of g) but people seem less sensitive about this.
I am not sure what you mean by "inherent." If you are asking if I think race and intelligence are causally related (i.e. that race or SES causes intelligence), no, I do not believe that, as there is little compelling evidence for that point. Rather, I believe that the general factor of intelligence ("g") exists, that this is the common factor among all tests tapping cognitive ability, and that there are differences on this construct among races. I do not hypothesize as to the reason.
And just to return to my original point: because these differences exist (regardless of why), if you were to drop SAT from a college admissions battery (with no other changes), you'd end up with a more diverse student body at the expense of grades. This is not evidence that the SAT is invalid; it is in fact the opposite.
@Dave - Yes. Gould has been thoroughly discredited in scientific circles for misrepresenting research, misunderstanding factor analysis (the basis of g theory), and promoting a political agenda at the expense of good science. It is certainly a convincing read though, if you are into that sort of thing - very effective rhetoric.
ReplyDeleteRicardus -
ReplyDelete> Rather, I believe that the general factor of intelligence ("g") exists
"g" is a common factor among IQ and related tests. Identifying "IQ test score" with "intelligence" is a common fallacy that is easily disproved by examining the hidden assumptions in IQ tests.
Crossed your post, Ricardus...
ReplyDeleteCould you provide some evidence that IQ test score and "intelligence" are correlated?
Could you provide a citation for your claim that Gould is discredited? _The Bell Curve_ doesn't count :)
@Dave - As I mentioned above, IQ test scores and intelligence are not the same thing. IQ measures both more than intelligence and less than intelligence, just like any cognitive measure. Rather, one should look to the common variance among cognitive test scores to see evidence of intelligence.
ReplyDeleteThere were a number of strong criticisms of Gould even around the time Mismeasure was written. The Bell Curve describes some of them, but it is similarly politically motivated. I suggest reading Korb's review of Mismeasure in the journal Cognition. It was in the mid-1990s somewhere.
@Ricardus
ReplyDeleteLooks like Korb's review has exactly one ISI-indexed citation since 1994. Not exactly evidence that Gould is "thoroughly discredited".
From my reading, Korb's only evidence that IQ tests measure intelligence is their correlation with "individual economic performance". This might be the wrong blog to advance that argument!
"Common variance among test scores" doesn't imply that the tests are actually measuring the thing you want them to measure. This is basically Gould's argument: the various flavours of IQ tests all measure more or less the same thing. Unfortunately there is little reason to call that thing "intelligence".
I think part of the confusion here is the word "intelligence." It's not one thing, but many--mathematical ability, artistic ability, musical ability, the ability to learn second or third or fourth languages, and on and on. What's loosely called "emotional intelligence--which includes perseverance and keeping at a difficult task in spite of obstacles or the lack of initial success is another kind of intelligence.
ReplyDeleteWhat's the Turing Test? Put a human being and a computer in separate rooms and talk to them by means of a keyboard. If you can't tell which one is which, then you have to concede that the computer is intelligent.
So here's a simple syllogism:
Intelligence is the ability to talk.
All human beings can talk.
Therefore,
I think part of the confusion here is the word "intelligence." It's not one thing, but many--mathematical ability, artistic ability, musical ability, the ability to learn second or third or fourth languages, and on and on. What's loosely called "emotional intelligence--which includes perseverance and keeping at a difficult task in spite of obstacles or the lack of initial success is another kind of intelligence.
ReplyDeleteWhat's the Turing Test? Put a human being and a computer in separate rooms and talk to them by means of a keyboard. If you can't tell which one is which, then you have to concede that the computer is intelligent.
So here's a simple syllogism:
Intelligence is the ability to talk.
All human beings can talk.
Therefore,