John Rosenberg of Discriminations is aghast at this bizarre short piece published in the New York Times (Michael Berube, “Testing Handicaps,” Sept. 21, 2003, Section 6, page 18, fee subscription required). Apparently, Professor Berube is concerned not only with racial score gaps on the SAT, but every other sort of gap that's possible, and foresees "fairness" on the SAT as norming the test for each group in the same fashion that golf scores are handicapped:
I've heard well-meaning people suggest, for example, that the test be "race-normed": the average "black" score is 857 and the average "white" score is 1,063, so let's treat the black kid with an 1,100 the way we'd treat the white kid with a 1,300. This is the kind of suggestion that provokes immediate and justified derision from conservatives -- often in the form of sports analogies: if we're going to spot black students 200 points on the SAT, they say, let's give white wide receivers a 10-yard head start on black safeties and cornerbacks.
But there is a way to "norm" the SAT, not only for race but for sex, income, region and level of parental education. (Every one of these variables is critical. Rural students average 998, while suburban students average 1,066. Boys outscore girls by 43 points. Most important, children of parents who have graduate degrees outscore children with parents who didn't finish high school by a staggering 272 points.) And the best way to do it is by taking a page from a sport whose country-club associations belie its deep structural commitment to redistributionist justice: golf.
Golf is proud, and rightly so, of the fact that its handicap system allows hackers to play alongside champions. And if only the SAT were as well organized and as egalitarian as the U.S.G.A., every high-school student would be assigned a handicap. We already have all the numbers we need; all we need to do is to combine "region" and "parental education" with the race-gender-class triad, and we can issue remarkably precise handicaps -- more precise even than golf handicaps, since the SAT's permit neither mulligans nor "winter rules."
Please tell me this was intended to be satire. Surely, a professor of literature cannot be serious when he states that an exam, one intended to separate test takers by ability as clearly as possible, should be normed on so many variables that everyone essentially ends up with the same score. Surely, he cannot be serious when he suggests that the world of college admissions should take lessons from the world of sport, especially after he calls similar sports analogies justifiably ridiculous. Surely, he cannot be serious to suggest that the score gap between those who come from educated families (who have presumably benefited from both nature and nurture), and those who do not is a problem, an indicator that the test is unfair, rather than a validation of the fact that the test measures academic skills which do not exist, nor develop, in a vacuum.
It's hard to top John's reply to this:
Berube assumes that the SAT can be fair only if the distribution of scores of members of every known group is identical. As Kimberly Swygert would say (in fact, has said here about a similar argument), Horse Puckey! The SAT is not a test of moral worth. A test is fair if it is accurate, not if its scores are distribued equitably by race, gender, class, region, etc., etc.
Emphasis mine. It's astonishing to see article after article address the racial score gaps on the SAT while providing virtually no evidence about whether those score gaps are measuring something accurately. If we want to talk about social engineering, we can talk about the impact of score gaps, and the use of affirmative action in admissions, and the performance of students once admitted to college, but from a testing perspective there's no evidence of bias here, and no need to address any score gaps by tinkering with the test.
Professor Berube, having hit bottom, digs deeper:
Take a black girl from rural Alabama whose parents make under $10,000 and did not graduate from high school, and put her up against the wealthy white boy from Lake Success whose parents have Ph.D.'s. Before she sets pen to paper, she could be facing an 848-point SAT deficit. If we assign her only 80 percent of the parental-education gap (217.6 points), 60 percent of the income gap (155.4 points), 30 percent of the racial gap (61.8 points), 20 percent of the regional gap (13.6 points) and 10 percent of the gender gap (4.3 points), the 452.7 point handicap will help us gauge her true talents more accurately. Fair enough, no? The reason we'd have to scale the percentages is that the various categories overlap, but I don't see any problem there -- certainly it's easier than scoring a golf match on the Stableford system.
Never mind that this young black woman was obviously cheated by her school system, or was reared in a home that did not value education, or was herself responsible for failing to learn enough material to graduate from high school (unlike Berube, I do not assume that her skin color means that she cannot be held accountable for her own choices). What on earth has convinced Professor Berube that this woman's "accurate" score is 452.7 points higher than what she produces when she takes the SAT? What on earth convinces him that she will do well if "handicapped" into college?
Why does the "wealthy white boy" not deserve the higher score he gets? Because he's white? Because he's a boy? Because his parents are PhD's? Professor Berube has a PhD; does this mean he would hope that his children's SAT scores be normed downwards, despite the amount of effort they've expended in doing well in school?
Appalling. Just appalling.
The important thing is that golf -- a game, notably, in which success has long been tied to race, sex and income -- has much to teach the College Board. Thanks to golf's handicap system, I can someday fill out a foursome with Tiger, Annika and Michelle Wie, and the U.S.G.A. will give me about a stroke a hole. And that's the kind of diversity of which any campus, and any country club, can be proud.
Um, unless I've missed something entirely, Tiger didn't win those tournaments because of any allowances for his race, sex, or income. He won because he's damn good at the game. Professor Berube can be as proud of his handicap as he wants; at the end of the day, no amount of data manipulation will affect who gets to don that green Masters Tournament jacket, just as no amount of data manipulation should affect who gets into college, either.
Posted by kswygert at October 13, 2003 01:08 PM