The Evil Assumption Underlying Ricci
By Sam Barr | July 2, 2009 at 8:39 amDrew Westen, writing for The New Republic, is the first commentator I have seen who has hit on a key assumption made by the five-man Ricci majority: the belief “that in a racially diverse city with a racially diverse fire department, it is perfectly plausible that the best candidates for leadership are all white.”
This might seem an unfair characterization of the conservatives’ position. How could a reasonable person believe such a thing, especially in an era when our president, the most important leader in the country, is a black man? Surely the majority acknowledged that New Haven’s written and oral exam was flawed, but argued that the city had to live with the test it chose, because it had created an expectation that the test results would be used.
Well, the majority did make that argument, although, as Justice Ginsburg noted, it’s a flawed one: “The legitimacy of an employee’s expectation depends on the legitimacy of the selection method…. Title VII surely does not compel the employer to hire or promote based on [a flawed test].” In the second paragraph of her opinion, she noted that nobody has “a vested right to promotion.”
Furthermore, the majority opinion does in fact rely on the pretension that New Haven’s exam was — I hesitate to say perfect — without major flaws. That assessment, unavoidably, rests on the assumption Westen identifies: that an unflawed test could produce such racially lopsided results, and therefore that white men must simply be better candidates for fire-department leadership positions, as a class, not as individuals, than black men.
Thus Justice Kennedy’s statement that New Haven “rejected the test results solely because the higher scoring candidates were white” (emphasis added). Justice Alito adds the insulting insinuation that the test results were thrown out to please some two-bit race-mongering black preacher. The possibility that the test results might be invalid, might be the result of a terrible method of evaluation, is not taken the least bit seriously.
Though the legal ruling may not do so, the majority’s assumptions and reasoning absolutely tear the heart out of the idea of disparate impact: no longer should government agencies or private businesses be suspicious of their evaluations if they produce results that defy logic, defy the assumption that blacks and whites are equal. Instead, they should just say to themselves, perhaps blacks and whites aren’t equal after all; maybe the white guys are just going to be better firefighters, or policemen, or middle managers, or whatever.
Killing such a pernicious line of reasoning is worth the price of upsetting the white firefighters’ “legitimate expectations.”
Filed under: Current Events and Issues | Tags: Race, Supreme Court | 5 Comments »
The tests were pre-approved as acceptable for use by an outside firm that the city has used many times for certifying racial acceptability. After the fact, the results did not reflect racial diversity. The test covers material that the city deemed necessary for advanced firefighters to know.
1. Unless the Court could find issue with either the test’s approval process or with its content, how can it have a problem with the outcomes?
2. If the test had produced all black and Latino promotion candidates, would you consider it progressive or unfairly biased?
3. How do you write questions about firefighting tactics that favor American-born English-speaking whites over American-born English-speaking blacks?
4. How is the expectation of fairness in promotions met for everyone if results can be thrown out without any problems being identified with the selection process?
1. The question is not exactly whether the Court “could find issue” with the test, but the related question of whether it should have found that New Haven had reasonable grounds for finding issue with the test. And I believe that the outcomes clearly speak to the inadequacy of the test. This is what I’m getting at in my above post: Either you think that it just couldn’t possibly be the case that a fair, valid process could produce such racially lopsided results (as in, you think it impossible that there isn’t one African-American in a city that is majority-minority who is qualified for promotion) or you think the opposite. And it seems to me that if you think it’s likely that a valid test would produce such wacky results, then you think it’s reasonable to assume that the white men were just better qualified, as a class, than the black men. Which, as I said, is a pernicious thought. The conservatives claim to be taking individuals seriously, not sacrificing an individual’s rights to the idea of group entitlements. But when you get right down to it, they rest on an assumption of one group’s outright statistical superiority. Because statistical, not individualized, superiority is all that could explain the results… unless there was something wrong with the test.
Now I’ve gone on and on, but I will answer the other questions sometime… later.
The test may well have been flawed, and I’ve read plenty that support that, but you can’t conclude so from the fact that white men did better in this case. The sample is much too small. The test could be entirely fair and white and black firefighters on average could be equally able, and you could still have a small group of candidates where no black candidates made the cut.
Sorry I let this fall by the wayside.
2. Unfairly biased. Westen actually makes this point well in the article I linked to. Imagine that New Haven had decided to test its candidates by having them play basketball or football, and then evaluated them not only according to physical ability but for leadership skills and team-playerness as well. Perhaps (given what we know about the NBA and NFL) the results would be disproportionately skewed towards the African-Americans. Surely physicality, leadership, and team-playerness are important qualities in firefighter captains, perhaps even more important than one’s studiousness, which, if one has it in abundance as Frank Ricci does, can certainly raise one’s scores on written tests. Nevertheless, I would object to such a minority-favoring result: I operate on the assumption that a fair and valid test will produce results roughly in line with the racial proportions of the test-takers.
3. This is very naive. It’s like asking, How do you write SAT questions that favor American-born whites over American-born blacks? For one thing, you can absolutely do so. For another, it’s not necessarily the questions themselves but the fact that it’s a written test: African-American culture is more oral than others. (I recognize that there was an oral component, but it was weighted only 40%, to the written exam’s 60%, which is a much different ratio than the one most other departments use, if they use such tests at all.) Your question also assumes away racial discrepancies in education, which are persistent and widespread. And it is the wrong question to ask about these tests: the right question is whether they could be expected to evaluate candidates for leadership in the department. Common sense and the evidence suggest it was not. Common sense: I did very well on the written portion of my driver’s test, because I studied hard and can memorize arcane and useless facts with ease (i.e., how many car lengths one needs to stay behind the car in front). But I am not a better driver than anyone else, necessarily, because of that fact. Because you can’t capture all of what makes a good driver in a written test. I suspect that being a firefighter captain is rather similar. Evidence: as Ginsburg noted, a survey showed that 2/3 of cities surveyed were using “assessment centers” to evaluate candidates for promotions. These put firefighters through actual scenarios and see how they react. And they produce racially proportional results. Which, as I have been saying all along, is what one ought to expect from a fair and valid test. Expecting otherwise is to assume the inherent superiority of one race over another. Other fire departments weight oral tests over written tests, look at on-the-job performance before narrowing the field of candidates, etc etc.
4. It can’t. Identifying problems with the selection process is absolutely essential. I have been trying to demonstrate why New Haven had ample grounds for suspecting something awry with its selection process. Then-current law said that New Haven largely had the benefit of the doubt, not least because conservatives have been weakening employment-discrimination law for decades in order to favor employers. But last week the Court abruptly reversed course and decided that employment-discrimination against white men (if you really think throwing out the results of a bad process is discrimination, when of course the city would subsequently develop a new, better test in which everybody, white and black, would be eligible to compete) is different in kind.
5. (re mholsen) I think that this is a valid question to raise. But the sample really wasn’t too small. A sample of 5 or 10 or so would be statistically unhelpful, but you don’t actually need much more than 20 or 30 sample units in order to expect that the test (or experiment, or whatever) will produce non-random results. In any case, the chance this being the cause of randomness would be slim, though non-zero. And the fact that other departments produce racially proportional results with other (better) processes bears this out. Finally, I don’t think the city could validly “conclude… from the fact that white men did better” that the test was faulty. They also had to supplement that with some assessment of how the test was faulty, and how they would develop a better test if the results were thrown out. On the facts, I believe they did this.
In this discussion there has been much talk about what the facts show, and, in fact, the disagreement on the Court was really over the facts than over the ideology. Well, at least the confusion over the facts didn’t help. Undoubtedly one’s view of the facts (whether one sees a city trying to rectify decades-old racial disproportions in its leadership by scrapping a terrible evaluation method, or a city kowtowing to pressure from noisy black preachers) is influenced by one’s ideology, but even still, the facts are sufficiently hazy that it was really a bad idea for the Court to take this case at all. Bad facts make bad law.
3) If you deny that things can be true regardless of race, class, and background, then maybe you’re right. Sorry, but if you speak English and understand the basics of American culture, then neither of these questions ought to prove problematic:
a) “Johnny and Beth went to the store. They spent $3.75 between both of them. If Johnny spent $2.30 on 6 apples and Beth spent the rest on 3 cupcakes, how much did each apple and each cupcake cost?”
b) “A two-floor household with three windows open has fire spreading up the staircase to the second floor…”
If there really are gross racial discrepancies, might that indicate inferior learning conditions or bad home situations, and not just racial bias? I’d rather wonder why a question about basic math is confounding many kids of any given race (hint: it probably has something to do with grossly inadequate education or bad parenting, for any number of reasons) than look for ways to “sensitize” the tests. One actually helps kids learn and succeed; the other makes them victims.
Whether or not written tests are advisable (I would recommend more practical testing, too, but that’s not the question at hand), going back through to find issues in hindsight is not appropriate. It undermines the expectation that studying to the test (I’m assuming reviews and criteria were discussed in advance, but not positive of that) and knowing the material that the city deems relevant will get you promoted.
I would not expect the results of any test to roughly reflect the racial makeup of its takers. I would expect immigrants to know the rules of citizenship better than native-born folks. I would expect suburban yuppies like me to know more about chemistry than inner city kids (though I’d be pleasantly surprised to find the opposite). I would expect people who had access to better education and better family conditions growing up to have better discipline and knowledge later on. I would expect an ex-con to know more about negotiation and self-protection than you or me.
It is wrong that this situation favors whites too much. That’s a testimony of the need to do more to give minority children better opportunities and more chances to succeed, and of the need to strengthen families and communities. It is not a sign that we need to manipulate results after the fact to make equality where, unfortunately, it is not earned by the individuals in question.
As for anti-discrimination laws being too tilted to favor employers, this is not a question of discrimination as such. It is a wrongful claim of discrimination on the part of an employer. Things ought to be based on merit, not identity. Generally, employers have greater incentives to favor merit over anything else in hiring, due largely to the profit motive. That is not always the case, especially when it comes to public employment, just as municipalities were far more willing to support Jim Crow laws than businesses. Businesses want the best worker in most (not all; that’s why the law exists) cases; government has any number of other considerations.