Mismatch: The End of a Liberal Dream

February 23, 2016

The most disturbing thing about mismatch research (examining the contention that a student can be adversely affected attending a school where her level of preparation is substantially lower than that of her typical classmate ) is that it demonstrates a tense inequity: recipients of affirmative action at selective colleges are not as smart as non-recipients. That’s the blunt truth, and nobody likes to acknowledge it.

Smart means something specific and local, of course, in this case math and verbal aptitudes. Those are what the SAT and ACT tests measure, and they are what calculus, freshman composition, organic chemistry, and dozens of other first-year courses demand to greater and lesser degrees.

If two students enroll in a statistics course and #1 scored 150 points higher on the math SAT than #2, he is a whole lot smarter in that class. It is entirely possible that #2 exceeds #1 in other aptitudes, such as the skills that go into drawing and painting, but those won’t help in Statistics 202. Student #2 is inescapably cognitively disadvantaged. In order to compete with #1, #2 must work twice as hard, logging more hours of homework, stopping by the instructor’s office each week, and using the school’s math tutoring service. Sadly, that is unlikely to happen, and #2 shall soon enough shift out of STEM fields and head toward an easier (for him) major or drop out altogether.

This is the mismatch catastrophe of affirmative action, and the strongest current argument against it.

But progressives don’t believe it. They can’t, because if mismatch is real, then a crucial article of progressive faith will fall.

The article is this: people are products of circumstances, and if we alter the circumstances, we can improve them. That premise obviously applies to affirmative action. Yes, the progressive admits, Student #2 comes into college less prepared than #1, but that’s not because he is less intelligent. It’s because he came out of an environment that didn’t cultivate math aptitudes as well as #1’s environment did. Once we place #2 in the same environment as #1, aptitudes will equalize sufficiently for #2 to function competitively among his peers. That’s the progressive rationale.

If only it were true. But the fact is that aptitudes are not so fluid. It is true that recent research has demonstrated that cognitive gains can happen among adults, but in those studies, the gains were highly specific relative to a single task such as the ability to comprehend patterns in matrices. Furthermore, the subjects underwent specific training in completing it.

No cognitive psychologists believe that the ordinary life of a college student provides the kind of deep-intelligence training that will enable him to raise his SAT math score 100 points after a semester on campus. Even if we allow a near-total influence of environment on intelligence (that is, reducing the “heritability” factor to nothing), a change of environment cannot produce significant changes in aptitude fast enough to benefit Student #2 in the first year. By the time a person reaches age 19, intelligence has hardened too much to rise with a semester of higher education, no matter how much academic support and the company of high-achieving peers surrounds him. It takes longer than that, even with total and concentrated immersion.

The only way for affirmative action policies to overcome the mismatch problem is for colleges to create a wholly separate extracurricular habitat for recipients. This means extensive daily tutoring and other academic support. They won’t raise math aptitudes much, but they will enable students to complete the coursework and perform on exams at a higher level. Some of the campus protests made recently by African American students, including the Black Students at Emory (my home campus), add this component to the list of demands. In this aspect, the students are correct. They need more help, and universities that have admitted underprepared students through affirmative action are duty-bound to provide it.

But it’s a necessity that proponents don’t want to acknowledge. Progressives don’t like genetic or other biological explanations for group differences in intelligence. They smack of fate, and they (supposedly) dissuade us from working for progressive reform. But environment, too, is fate, for all practical purposes. That’s the sad truth, and for supporters of affirmative action to ignore it is to show them as ideologues, dogmatic and anti-science.

Discussion

gospace

February 26, 2016

Navy Nuclear Power School deals (or dealt) with test score mismatch. When I went through, the class was divided into sections, based on GCT/ARI (now ASVAB) scores and by rate. (Electronics Tech/Electricians Mate/Machinists Mate). The lowest sections were exempt from watchstanding, and everyone started out with 20 hours mandatory study time, that had to be logged in at the schoolhouse. That’s above and beyond the 40 hours spent in class. The top sections were encouraged, but not required to study, and didn’t have to log their study time in, but were encouraged to.

Sailors in the top sections could work their way into mandatory study by doing poorly on tests. Don’t know if those in the lower sections could work their way out. Probably, but it didn’t concern me.

Even with the mandatory study time and all the help available if asked for, the failure rate in the lower sections was higher then that in the upper sections.

A test score cannot tell you with 100% certainty whether a particular individual will succeed or fail in a particular academic endeavor. But the score will tell you that “X” percent with a score below a magic number will fail, while another “Y” percent with a score above another magic number will succeed. One thing the military is good at is sorting people into career fields where they can succeed, or at least, not fail horribly.

Reply
Charles W. Abbott

February 25, 2016

It seems to me that the SAT is measure of some mix of intelligence and preparation.

The level of preparation changes over time. My understanding of SAT is that it is a pretty good predictor (especially) of first year performance.

Prof. Bauerlein, I think this article doesn’t adequately develop some points you make elsewhere, such as in _The dumbest generation_.

For any verbal score, it tends to measure competence with “low frequency words” that come up in print or if you have verbose peers / parents who delight in wordplay.

I tend to suspect that the overall hypothesis of _Mismatch_ is confirmed by empirical evidence. But your argument as couched here is too extreme.

If SATs measure preparation as much as raw intelligence, the highly intelligent but underprepared can catch up with effort over time. It’s still a shock to the student to be “the dumbest guy in the room.”

Reply
Howard Schwartz

February 24, 2016

In the age of microaggression, no program to boost the performance of affirmative action admits can be successful. The issue is not so much about grades, but about what other students are, supposedly, thinking about you. If students of your type require special programs to bring you up to speed with regularly admitted students, this could be interpreted as meaning that you aren’t as bright as they are. Whether the other students think that way or not doesn’t matter; projection will do just as well. And then we get into the whole thing about pervasive racism, and they feel unsafe, and so on.
It doesn’t take much.

Consider that at Yale you have some of the most highly selected, and competitive, white and Asian students in the world. Imagine that two of them are having a dialogue in class, with each of them trying to show how smart they are. A black student, let us say, knows that he cannot follow the argument; it is over his head. He feels inferior, in that regard. But instead of leaving it at that, he withdraws from the recognition that the feeling of inferiority comes from him. He gets rid of it by attributing it to the students having the conversation. He is not inferior; they are racist. They are microaggressing against him by sending the message that he is inferior. And who is going to tell him that he is wrong?

Reply
bdavi52

February 24, 2016

But it’s so hard to leave a merry-go-round once you’ve jumped on! It goes faster and faster and you hold on tighter and tighter, so afraid you’ll fall off.

The explosion of Student Tantrums Against Systemic University Racism drives one absolutely critical question: What Racism? A question, of course, which has never been answered — cannot be answered — because there is no answer: no demonstration, no proof, no indication that any major/minor American University systematically, procedurally discriminates against Black students and/or Black faculty. None. There is only the PERCEPTION that such discrimination occurs.

“Do you doubt that I feel discriminated against!” we’re asked. No, of course, not. We’re certain you feel exactly the way you feel — what we doubt, very simply, is whether there is any evidential ground upon which this perception stands. And yes, at Missouri, for instance, they can point to 3 alleged racial incidents: 1 drunken insult (perpetrator expelled), 1 unvalidated racist epithet tossed from a passing truck by strangers, and 1 poopy swastika, scrawled in a bathroom. 3 incidents, 35K students, two years — yielding a student/day incident rate of .00001%. That ground seems pretty insubstantial — very difficult to justify this feeling of ‘pervasive racism’ given a .00001% incident rate.

Well then, we’re told, what about the vivid outcome imbalance problem — so quickly visible in student/faculty/administration populations. Blacks are significantly under-represented.

Yes, they are. BUT — given multiple generations of AAction….given multiple decades of specific, highly-targeted effort to increase URM counts every single time a job opening develops… how on earth can we continue to believe that outcome imbalance is related to some racist practice which remains absolutely invisible? Worse, how can we continue to search for the White Racist Villain in this melodrama when the answer to the imbalance question is, in fact, so painfully clear.

Black HS graduation rates (though steadily improving) remain @ 15 points below White rates. As reported by the ACT and the United Negro College Fund, “Sixty-two percent of African-American students who graduated from high school in 2014 and took the ACT met none of the organization’s four benchmarks that measure college readiness, which was twice the rate for all students.” Add to that the extraordinary disconnect between this very real lack of scholastic competence and the unrealistic attitudes also displayed by these very same college-bound Black students: 95% of all Black freshmen are “quite determined” to finish their college degree…89% say that “going to college is the most satisfying thing to do at this point in their life”? Is it really any wonder that the collision of these collegiate fantasies (sadly reinforced by AAction mismatched recruitment) against the immovable harshness of academic reality drives massive failure & non-graduation rates (20 points below White rates)….which drives, in turn the fear, anger, and frustration we see in all these protests?

And then, of course, we have the “Robust, & Replicable pro-Black Bias in Social Judgment” as reported by U.Virginia researchers, Ebersole and Nosek

The fault, it would seem, lies not in the stars, but in ourselves.

But to accept that judgment, to hold AAction responsible….to hold the creation of false standards responsible (which allow the equivalent of a 310 pt. SAT boost to AA students (Espenshade/Radford))…to hold, in fact, the non-performers responsible for their non-performance (which has significant implications which ripple all the way out to family structure & culture)….is to let go of the merry-go-round. And damn, it will really, really hurt to let go of the belief so many have clung to for so long.

Reply
Harry East

February 24, 2016

“intelligence has hardened too much to rise with a semester of higher education,”

Did you even read that back to yourself?

Let’s consider your example, though. We have the SAT score… you assume that this is a true measure of intelligence (at least with respect to maths)… but everyone knows that questions have to be raised about the ability of basically any assessment to capture “true intelligence”. It is also well known that, especially with things like the SAT, technique matters a lot. I am also sure that you’re aware that the students entering university from wealthier (and sub or urban) backgrounds have a lot more access to the sorts of resources which, from the perspective of true intelligence, artificially boost their SATs. That is, it is possible to do better on tests than most people because you are better at taking tests than most people.

Another problem from working from SAT scores (other than assuming that they are a good measure of true intelligence) is that you just looked at the difference. If Student #1 entered university with a Maths SAT of 800, and the other 650 (81.25%) we have two intelligent students (or, at least, ones good at taking SATs). Yet, this is the same as if we are talking 650 and 500 (62.5%).

The thing with mismatch is that it’s basically the idea that some people, if chucked in a stream that flows too fast, will drown, but if you chucked them in another stream, that starts out at their pace, they can succeed. (You can see here the necessary assumption that the difficulty of the work offered differs /enormously/ between, say, Harvard and the local state university.) Someone with a maths SAT of 650 is probably able to swim a whole lot better than someone with a SAT of 500… in other words, you can’t give us the relative scores because the absolute ones (to the extent that they are meaningful*) are what matters to mismatch. The question is whether or not someone with x entrance ability is able to keep their head above water.

This was the issue with Scalia’s remarks when Fisher was up at the Supreme Court most recently: he seemed to assume that everyone admitted under holistic review methods was at an absolute level where they would drown. That’s an extremely sketchy assumption.

Mismatch has validity as an idea insofar as it’s very difficult to learn anything if you feel overwhelmed the entire time. You cannot focus. You can’t provide distance between yourself and your situation. Old strategies for success (brought through from school) do not work for the overwhelmed because their heads are in and out of the river (esp. if they are also away from old support networks). To that extent, their emotional rather than academic wellbeing is what you would want to work with primarily. We might call mismatch “persistent academic culture shock”.

However, you have tried to present us with some data based analysis. Sadly, this data was presented in the context of a fundamentally flawed scenario. If indeed it is true that universities do not make their students more intelligent (or, at least, bring those who do slightly worse up), then perhaps someone who enters with 650 is better off in a stream that they handle with ease… but doesn’t the struggle matter? are they really better off with an A+ from wherever than a B from Harvard and being able to tell employers that, “When the going gets tough, I don’t quit and find something easier”?

*The SAT is also a single test capturing a specific moment in time. Compare and contrast education systems where university admission (from an academic perspective) is based entirely on a whole year evaluation of student performance (+ whatever specific subject requirements). The demands represented are radically different… coursework + exam, is a better representation of understanding than just exam or just coursework (assuming an exam is even meaningful in the discipline).

Reply
1. bdavi52
  
  February 26, 2016
  
  Intelligence is, perhaps, not the best descriptor for what drives the Mismatch Problem. Scholastic ability, maybe, would be more accurate. Demonstrated academic performance, perhaps. But in any case what Mismatch addresses is the gap between any student’s basic academic quality (not potential) and the required quality or stringency of the academic hurdles which must be cleared to be successful in college.
  
  Obviously, the larger this gap, the more difficult the collegiate experience becomes. The fascinating question, of course: how big is too big a gap?
  
  To your point, some Gap is not only to be expected (HS to College) but is absolutely required if real learning is to occur. If I’m running a 7 minute mile in HS and winning and College only asks me to run a 7 minute mile, well yes, I’ll win — but I won’t improve. I won’t learn anything. I won’t be pushed, asked or expected to step beyond what I’ve already mastered. And I leave College with my HS ‘quality’ unimproved.
  
  But, obviously, when we admit students who would otherwise not qualify by any of our normal measures…if we admit them only because they’re cosmetically diverse…then the gap between when they are currently capable of doing and what we will require them to do will be absolutely bigger and more significant than the normal HS/College gap we impose upon our other freshmen (who themselves may struggle, even given their lesser Gap).
  
  Right now SAT/ACT, standardized national testing, is the only reasonably objective, shared measure available. Clearly better than GPA (which, ideally, measures consistent performance over time….but which also is entirely a function of faculty/school quality)…it is still limited in what it shows. Great Test Takers do better than Poor Test Takers. Scores can be marginally boosted with special prep. But standardized test results (in combination with GPA) do allow us a glimpse into peer to peer quality…and they do, in fact, generally correspond to college performance in large populations.
  
  Given a large Gap (a gap significantly larger than that wrestled by their peers) it is severely unrealistic to expect that any remedial course or two or three will compensate for what is obviously absent. A hundred-point SAT Quant separation will not be corrected by a course in Kiddie Algebra. The severity of this gap will then impact the balance of the college career, to Bauerlein’s point, shifting major choices and even driving crash & burns (Black college graduation rates currently running 20 points under White rates). Like asking the 7 minute miler to compete on a D1 team which averages 4:15…he will not just fail, he will spectacularly fail.
  
  One thing to fail because you didn’t work hard enough…another thing to fail when you’ve been set-up to fail, when your qualifications for success are so far below the ‘normally/average’ qualifications that you become the 7 minute miler being lapped by every peer. That’s not just failure; that’s humiliation.
  
  Reply
  1. Harry East
    
    March 1, 2016
    
    Apologies, I am not very concise.
    
    Actually, academic potential does matter. Everyone can learn how to swim, right? But if you’re chucked in a pool where your assumed knowledge of swimming is greater than your own (mismatch), then it’s quite a bit more difficult to improve (e.g. you’re still floating and this pool started beyond treading water). Yet, if you are particularly good at learning, or dedicated, you may be able to get close enough to be able to get something from being in that pool.
    
    That being said, if you have a greater potential to learn, that you are mismatched is, in some sense, a bigger issue than someone whose potential to learn is lower. After all, the scale of the lost learning (caused by mismatch/persistent academic culture shock) is greater. Put both of these students in a pool/stream/class we’re they are not mismatched, and the one with the greater potential will learn more. At the same time, as we metaphorically demonstrated above, if you have greater potential to learn, you’re more able to cope with mismatch than the other person.
    
    But, of course, you’re completely right. A weight lifter doesn’t get stronger by sticking with the same weights (although their endurance may well improve). In fact, I was grappling with a manifestation of this issue the other day: was a course in one major too similar to my courses from another major (answer: no, it is, for one thing, more theoretical)?
    
    Let’s imagine that the intended jump up (Gap I believe was your terminology, I’ll use jump) of going from school to uni can be quantified. Imagine that 90% in school translates to 80% at uni, 70% to 60%: a ten point difference. The university assumes that everyone it admits has the potential to make up this gap (and it measures potential through entrance criteria).
    
    This potential consists of one’s capacity to learn, but it’s difficult to learn if you arrive at that first lecture having a basic idea of differentiation, but this course is% starting out with integration which you haven’t encountered before. Thus, your absolute learning — being on the same page as everyone else when you start — and your capacity to learn are what determines your ability to overcome the jump up. After a few weeks or whatever, the university imagines, that any given individual will now be able to get a uni 90% if they entered having been able to get a school 90%.
    
    The impact of mismatch on this is that we have someone whose potential is limited by starting at least a page behind. Maybe they have the capacity to learn that their own personal >10 percentage point jump up doesn’t matter. (That is, they start close enough behind that they can catch up.) Perhaps more commonly, they don’t. These are the people who experience persistent academic culture shock… and maybe it cannot be resolved.
    
    But insofar as I had a point beyond criticising the way Bauerlein framed mismatch, if your experience of “mismatch” means that when you graduate from Harvard with a B average rather than an A average, you’re quite likely better served than if you’d gone somewhere that expected a seven minute mile… where you could have had an A.
    
    This is something that I feel discussion of mismatch has ignored. That you could do better somewhere less intense when what you get is a B- is a non-meaningful case of mismatch. In fact, as I have tried to point out above, it could in fact be marketed in a better way than an A- from somewhere where you weren’t “mismatched”. As you rightly say, extensive major switching/dropping out is where mismatch can be concerning… but when you want to link mismatch to affirmative action you run into problems.
    
    As we have acknowledged (and you made explicitly clear), tests are generally indicative of scholastic aptitude rather than true intelligence/knowledge. They can be pulled in different directions by things like laziness or technique preparations or preparation techniques (e.g. exam doping, cramming). Exams are performance and performances are manipulatable. An across time grouping of tests/assessments… if comparable (as you say, GPA is flawed beyond being a crude average)… is a bit fairer but still not ideal. The ability to manipulate performance is not an innate human talent: it’s external. That, John, is able to access a SAT tutor and Jonny can’t has nothing to do with the individuals and everything to do with their circumstances.
    
    The idea behind AA is that you can influence the circumstances of John and Jonny’s children/grandchildren by what happens to John and Jonny. (This is well supported: consider Pisa and resilience, and social mobility.) The challenge mismatch presents is that it is possible to bring someone into an environment, a stream, where they can’t keep their head above water. I think you’d find very few people who would disagree with either of these notions. That we should try to increase social/educational mobility and that mismatch can happen are fundamentally reasonable propositions.
    
    Now, I’m hardly familiar with a range of specific AA policies, but it is specific AA policies that affect the applicability of mismatch. If we consider Fisher, as it was last at the US Supreme Court, the form of AA was holistic admission (possibly I have the exact terminology a little confused) for everyone who didn’t make the top 10% of their graduating class. Now, the specific challenge, in some sense, was that the graduating class policy… due to the state’s informal segregation of different ethnicities, rural/urban schools and rich/poor schools… was AA enough, and that the holistic stuff on top of that was not specific. The question we’re interested in here, though, is whether or not this specific policy (as it exists) leads to problematic mismatch.
    
    Scalia brought this problem of mismatch up in the oral argument stage. However, in what I am told is uncharacteristic of him, this was very inarticulately conveyed. It seemed as though his argument was that all Black Students would be mismatched. This, and less extreme, logic (whether intended or not) doesn’t bring anything to bear on the specific process. What matters is whether the University of Texas demonstrates mismatch.
    
    That’s not actually a particularly simple question to ask. After all, one has to classify all students based on their ethnicity and how they were admitted to the university. We’re probably interested in four year graduation rates, and retention too. So, we’d probably do a Poisson analysis of these counts, trying to find out if there was any relationship between AA admission, retention and graduation. It would also be important to chuck in where they go (e.g. tougher programmes like engineering, things with competitive entry to majors), gender and SAT/ACT scores (somehow we’d need that to be one variable: possibly a factor with levels like excellent, good etc.). I think where educated would also be important (at country/state levels Texas, California, overseas and, if possible, if the school was rural or urban).
    
    After having done such an analysis (and I’ve possibly ignored something that matters), if a relationship was found to exist between AA admission and retention/graduation, we could infer that mismatch may be happening as a result of AA admissions. If we did this as a regression analysis we could probably estimate the size of the effect too.
    
    If someone has done some sort of suitable analysis, I’d be interested. But this is what actually would show mismatch as a relevant part of discussion of AA/liberal dreams… at least as far as those apply to Fisher.