The following started out as a comment to his post, but sort of kept going long past comment length. I won’t clutter his page with this missive, but perhaps the following may be at least a partial response to his commentary...
I’ll take up that challenge, Jack. Sort of.
That is, I would suggest a couple of reasons to be skeptical of this proposed legislation. They require a little more abstract reasoning and a little more knowledge of the way admissions operations work than the average politician is likely to be able to muster. I hasten to add that I am not defending the rationale that diversity is such an inherent good that it overrides all other considerations, but (for example) discussions of The Death of the Last Black Man in the Whole Entire World or Fires in the Mirror in my Advanced Play Analysis class have always benefited enormously from the presence of students with different demographic profiles.
But to the matter at hand… First off, it’s difficult to assert with any authority that discrimination is actually taking place, since no one can agree on what elements of a student’s record ought to be weighted more heavily: grades? class rank? standardized test scores? extra- and co-curricular activities? recommendations? interviews with admissions personnel and/or alumni? Should legacies count? Should the ability to pay tuition without institution-funded financial aid? Should the inability to do so count? All of these elements are part of the admissions evaluation process at universities, especially those at the elite level. (At my university, a student with a certain class rank and/or board scores is automatically offered admission. But we’d ultimately end up accepting probably 95% of the students rejected by Harvard or Yale.)
Grades are “objective”; except of course, they aren’t. Students take classes with different degrees of difficulty. Someone who gets a 90 in calculus and a 75 in phys ed is a better student than one who gets a 95 in phys ed and an 80 in algebra, but will have a lower average. Moreover, it almost goes without saying that the #20 student in a class of 400 at School X might be better or worse than the #1 student in a class of 40 at School Y. Using sports as an example: I happen to live in the small city where one of the top half dozen soccer players in US history went to high school. He was, of course, all-conference, etc. But an all-conference player in that league today might be lucky to get a scholarship to a Division II university, let alone even hope to become a star for the national team. So how do we compare, if all we know is “all conference”?
Of course, pols of both parties seem enamored of standardized testing as the ultimate arbiter of student (or teacher) success. The closest thing to an objective means of comparing students is to give them exactly the same test under exactly the same circumstances. But even such an endeavor is problematic, even apart from such considerations as recognizing that Student X is fighting off a cold, or student Y’s mother just got diagnosed with cancer.
Someone, a mortal, designed that test, and the questions are almost by definition going to favor someone who knows precisely those concepts, terms, or formulas over someone who may have mastered different, but equally important, material. I remember a question on one of those tests that asked me to figure out the Earned Run Average of a baseball pitcher. The question provided the formula, but because I knew baseball and statistics, I could glance at the multiple choice responses and determine the right answer in perhaps two seconds. Someone fully as adept at math but who didn’t know baseball would have spent a lot more time, and perhaps been rushed at the end of the exam, omitting questions or not having time to really think them through.
So it’s perfectly possible that a student taking a standardized test would score better or worse, perhaps even appreciably so, were they to take the exam a week earlier or later. But this isn’t even the biggest problem. Those who argue for significant weight for SAT, ACT, GRE, LSAT, MCAT (etc.) scores do so based on a fanciful belief that performance on standardized tests measures skill or intelligence. It does not. In fact, they measure test-taking ability. To provide just two examples from my own personal experience as a standardized test-taker:
I was pretty good at math when I was in high school, so in my junior year I was among the 15 or 20 students who took a standardized test under the auspices of the American Mathematical Association, or some organization like that. The test was difficult; no one could be expected to finish it in the appointed time. And it was scored like Jeopardy: harder questions were worth more points, and you literally lost points if you answered a question and got it wrong. I got a negative score! Still, I was asked back the following year, after several months of taking a class with the worst math teacher I ever had, and a couple of months with the next-worst. I adopted a different strategy… and got the highest score in the history of the school. Was I as bad as my junior year score? Nope. Was I the best math student in however long my high school had been administering that test? Not even close.
Skip ahead a few years. I took the GRE some five or six years after the second of the two freshman-level math courses (Calculus I and Intro to Finite Math) I took in college. It was 8:00 a.m. on a Saturday. I’d had about four hours of sleep, and I was more than a little hungover. Perfect conditions, right? I got a perfect score in math. Is it plausible that I deserved such a score? I suppose so. The test I took, unlike the subject exam in math, didn’t really test anything past algebra and a little introductory probability. But no rational person would think that my math skills exceeded my verbal skills (my verbal score was good, but not perfect).
But, as they say on the late-night infomercials, wait, there’s more. There are studies that show there is little correlation between board scores at one level and performance at the next. More to the point in the current discussion: I remember reading about (as opposed to reading) studies which concluded that the average African-American student with a certain board score would outperform the average white student with that same score. I forget the exact figures, and I’m not sure I could find them now, but the following provides the general idea: on average, a black student with a 1080 SAT score would get better grades in college than a white student with an 1100. I don’t recall that Asian-American students were included in the findings, but I would suspect that the same principles would apply. So should we be comparing the scores themselves, or their predictive value?
Of course, multiple studies have shown that these tests tend to privilege affluent white males: precisely the stereotype of the GOP pol. Were I of cynical disposition (perish the thought!), I might be tempted to suspect that the Republicans are using Asian-American rights as a front for their attempts to codify white preference. Wait, Republicans being as virtue-signaling and contemptuous of actual reality as the Dems? Surely not!
Be it noted: for all my distaste for standardized tests to measure, well, anything, I still use them in considering whether a particular student should be admitted into our program or receive some of our sparse scholarship money. This is different, however, from believing that a higher score on the SAT or ACT makes for an inherently better student. Taken as part of an overall assessment, a good score will do a student more good than a bad score will, but by itself it’s merely a statistic. Another sports analogy: the best shooter in the NBA will never have the highest field goal percentage, because he’s shooting a lot from 3-point range, where 40% is excellent, but there are centers and power forwards shooting well over 50% because they never take a shot from more than 5 feet from the basket.
So if the question is whether discrimination is a good idea, the answer is of course not. But the phrasing of the amendment opens up a lot of questions without answering any of them, especially about what, exactly, constitutes discrimination. Catch phrases and gotcha politics on either side of the aisle aren’t going to solve any problems.
An admissions policy which privileges class rank over SAT scores would probably benefit African-American students over Asian-Americans, but it is not inherently discriminatory, any more than a strategy that reverses those priorities would be. And certainly a procedure which attempts to predict success at the collegiate level rather than judging success at the high school level doesn’t seem to me to be problematic, although it could be called discriminatory, I suppose.
I don’t want some federal judge to decide how my alma maters, my employer, or indeed any other college or university runs its admissions process, absent clear and convincing evidence of actual discrimination as opposed to simply that which could be construed that way.
I’m not suggesting that we ought to break existing laws to advance a cause. There are no doubt cases in which the disparity in qualifications between accepted applicants of Group X and rejected applicants of Group Y is so profound that racial discrimination (against Asian-Americans, perhaps against whites) is the only plausible explanation. In such cases, the legal system does indeed need to be employed.
What I am suggesting, however, is that the Cruz amendment smacks more of politics—“look, we got the Dems to vote against equal rights!”—than of ethics. It seems too broad a statement to ensure just administration without further clarification. In something like collegiate admissions, one person’s discrimination is simply another person’s (ethically legitimate and comprehensible) difference in priorities.
As a career educator at the college level, I’d probably have voted against the Cruz amendment, too… although I’d be fully aware I was walking into a trap.
No comments:
Post a Comment