That said, I never liked this case on the basis of fairness. The dyslexia doesn't move me. It's one thing to argue over criteria--say, should race play a role in college admissions? It's another thing to argue that after we've agreed upon a criteria, we should scrap the results because we don't like how they look. This is just me talking. Some legal scholars, and likely some of my guest bloggers, may feel different.To conclude that New Haven acted unconstitutionally is to assume that the Constitution's 14th Amendment mandated a policy of strict colorblindness by state and local governments. Maybe it should have. But the historical evidence that it did is weak. Certainly the conservatives on the Supreme Court have not tried to argue that it did: originalist analysis has been notably absent from their opinions in affirmative-action cases.
Judicial restraint has also been absent. That virtue is best understood as a finger on the scales, tipping judges in close cases against invalidating the actions of Congress or state or local governments. To invalidate laws without a strong argument that the Constitution requires doing so is precisely what conservatives usually mean by "judicial activism."
In their outrage over the Ricci verdict, I suspect, conservatives have gotten carried away by their laudable fervor against race-conscious policies. But we on the right, of all people, should know that not every wrong has a judicial solution. Conservatives are moved, as well, by their empathy for the Frank Riccis of the world. When President Obama has talked about empathy on the bench, conservatives have responded that, given free rein, it can lead judicial reasoning astray. On race, unfortunately, we are illustrating our own point.
« My abbreviation was young when I caught the cases | Main | Wimmen's Work And The Midlife Crisis » SCOTUS Rules For Ricci29 Jun 2009 10:48 am
We knew this was coming. I've read up on this case quite a bit, and the fascinating thing is that it looked like the law was on the city's side. From Ramesh Ponnuru, last week:
Comments (69)Post a comment |
Today's Headlines From The Atlantic |
The Beautiful Struggle: A Father, Two Sons, and an Unlikely Road to Manhood 
Copyright © 2010 by The Atlantic Monthly Group. All rights reserved.
Coates
I am far from a legal scholar but if you get a minute you definitely want to read Justice Ginsberg's dissent. In fact she threw her weight around so much on this one that Not only did the conservatives write the decision, they also had all these other "concurrances" including one written by Alito which did nothing more than attack Ginsberg's dissent really.
Like I said I don't know the law but I can read english and in Ginsberg's dissent she actually makes a strong case that disparate impact IS something that is supposed to be in the critera of title IX decisions and not just disparate treatment.
Hell read it for yourself. I think its insightful even for us laymen.
http://www.supremecourtus.gov/opinions/08pdf/07-1428.pdf
Yeah, this seemed like just the case for some kind of conservative judicial activism (gasp!). To put it simply, if the law was on the city’s side in this (which a lot of people who know more about the law than I do, seem to believe is the case), then the law just doesn’t seem fair.
Of course, a lot is going to be made of this with regard to Sotomayor, but I tend to think she probably comes off good on this. It would look bad if the decision weren’t close, or if there wasn’t a solid legal argument backing her decision, but that doesn’t seem to be the case. My assumption is that there will be a little bit of grandstanding on the case from Republicans, but that they should be easily shut up or made to look foolish (assuming she can make what everybody seems to believe is the sound legal argument in favor of the ruling that’s just been overturned).
Oh and one more thing.
I thought the most compelling part of Ginsberg's defense is the effect that the ruling will have on businesses that try to self police themselves to be title IX compliant. As she said, should these businesses actually be proactive and without the government having to interfere (which is what conservative judicial advocates have pressed for for some time)try to make sure they are not using discriminatory practices, this decision would open them up to costly lawsuits with a low threshold of proof for the plaintiffs. Its really a very fascinating read.
Yeah, I'm gonna read it later today. Again, everything I've read has said that the lower courts were right on the law. I also understand that courts are supposed to be ruling on law, not what seems to a layman (like me) to be fair.
The law's not always an easy dichotomy, though, which is why being a judge requires the ability to interpret the law. It sounds like SCOTUS decided to overturn precedent in this case, which means that everyone beneath followed the law as they understood it and as previous courts had interpreted it, and this court decided to toss all that out. Some people would call that judicial activism. I call it the prerogative of the highest court in the land.
Situations change and the way we interpret laws changes based on that. The difference between me and so-called originalists is that I acknowledge that from the start, while originalists twist themselves in circles trying to pretend it ain't so. And you'll notice, not much originalism pops up in the majority decision. Wonder why that is?
Ginsberg's correct as far as she goes, New Haven was just trying to avoid a Title IX complaint which they probably would have lost, but I think most people sense that something unfair happened here and that there has to something balancing the other side of the equation. I need to read her decision in depth, but I'm not really sure where she thinks that line should be drawn.
And, to be honest, I don't find the argument that nobody really got hurt because nobody got promoted very compelling.
As for Ponnuru's point, yeah, you know, it's fun to mock the other side's hypocrisy, but that's really irrelevant to right and wrong. These are constitutional issues IMHO, and while I don't like the current Court very much, I do think this is an area of law where it's appropriate for the Court to step in.
... I think most people sense that something unfair happened here and that there has to something balancing the other side of the equation.
Interesting. That sounds like a pretty good argument for having judges that use... what's the word... oh yeah, empathy. ;-)
Well, you know, I think it *would* be a good argument for empathy, except that I don't think the SC has traditionally (or currently) been suffering from a lack of empathy for white men.
This goes along with my other point. I agree that this case illustrates the contradictions and/or hypocrisy on the right. And that's important to point out. But once we pass that by, we're still left with some difficult issues to work out.
Well Ponnuru isn't really "mocking the other side's hypocrisy." To the contrary, he's asking for some sort of intellectual consistency. From my perspective, I'm not mocking the other side either. My feelings on the case are documented.
During my earlier reading about this, my impression was that the argument is that when you pick criteria ahead of time, as they did, which then shows itself to potentially have bias built in that you didn't notice earlier, it's better to go back and make sure that there wasn't any bias.
Maybe I bring my scientist-filter into everything, but it reminds me of validating scientific results. If you run an experiment and get consistent results that weren't what your model predicted, it's integral to the scientific process to comb through your experiment (and have others do the same) to make sure there weren't outside factors influencing the experiment. You want to make sure you have the most correct data possible before drawing conclusions. So in this case, if the city comes up with a different test that gets the same results on the same group of people, then great - no worries about bias, and it really was merit-based. And if not, they've discovered a flaw in their system that was invalidating their results.
“So in this case, if the city comes up with a different test that gets the same results on the same group of people, then great - no worries about bias, and it really was merit-based. And if not, they've discovered a flaw in their system that was invalidating their results.”
Yeah, I don’t think so. Separate findings in a second test don’t necessarily prove flaws in the original test any more than similar findings would prove no flaws existed. You’re talking about a small group of people taking two different tests. In order to make a scientific argument in favor of bias, you’d have to have a larger pool of people taking the same test(s), ideally, over a period of years. By forcing a retest, all you do is try to avoid the problem, which is, in fact, exactly what New Haven was trying to do. They were trying to avoid being sued and labeled as racist, so they changed the rules in the middle of the game. Not to be fair, mind you, but (ironically) to avoid the appearance of being unfair.
Yeah, I was using that as an analogy, not an exact correlation to the situation. I am well aware of the problems of a small sample size, and that a second test could definitely have problems as well.
I can't speak for New Haven's intentions, whether they really wanted to be fair or just wanted to avoid the appearance of unfairness. I tend to give the benefit of the doubt on questions like that.
"I can't speak for New Haven's intentions, whether they really wanted to be fair or just wanted to avoid the appearance of unfairness."
This is actually established:
"The city said it acted because it might have been vulnerable to claims that the exam had a "disparate impact" on minorities in violation of the Civil Rights Act of 1964, and sought to avoid lawsuit from minorities who were turned down following the test."
http://www.cbsnews.com/stories/2009/06/29/national/main5121634.shtml
Furthermore, this was the second test! The whole point of paying outside consultants to develop a test was due to accusations that the previous promotion method was unfair for the same reasons, that not enough minorities were promoted. So KT should have no concerns.
While you're correct that a second test wouldn't necessarily prove or disprove the idea that the first test was biased, it would help reduce error from statistical fluctuation. It's not useless - just not as useful as we'd like.
This whole case to me seems like an exercise in the perils of small-number statistics and the dangers of setting criteria which won't allow adjustment post-results. If certain outcomes are unacceptable (as in this case), then don't write yourself into a corner with inflexible rules. Ironically, I'm sure that the whole point of developing that kind of criteria was so that they couldn't be accused of prejudice - when no one trusts the management to make decisions in good faith (and let's face it, the history of racism makes it plenty reasonable to not trust them), you're left with trying to design decision criteria based on an "unbiased" merit-based measurement.
"This whole case to me seems like an exercise in the perils of small-number statistics and the dangers of setting criteria which won't allow adjustment post-results."
Maybe. I don't know. I think I'd agree with you if the case were, fundamentally about the fairness of the test. It wasn't.
In fact, the case was never about whether or not the results of the test proved bias. Afterall, the guys who passed the test were the ones bringing the suit. And the city didn't--to my knowledge--make the argument that it was a biased test.
The city threw out the results not because they believed the results proved bias but because they were afraid that accepting the results of the tests as they were would leave the city liable in discrimination suits brought by the African Americans that didn't pass.
In the end, this case wasn't about the test. It was about whether the city should be allowed to change the structure of the objective process by which it grants promotions simply because the results of the process yielded an unforeseen result.
Re: BreakerBaker
Honestly, I don't know what the correct decision is from a legal standpoint. It seems to me that the city designed poor criteria from the start. You're right that that isn't what the case was about - or rather that isn't what the legal argument is about. My point relates more to what I think the parties should have been fundamentally concerned with rather than what they were concerned with. And that applies to everyone - the city and both sets of firefighters. In the end, I feel as there was far too much focus on whether bias was present than on what is the best way to determine promotions and how to best measure those aspects.
While it probably wouldn't have been the better legal argument, I personally would have had a lot more respect for the city if it had said either, "We're tossing the test because we don't think that it measured the qualities that were relevant well enough and is constraining our ability to make the best promotions possible" or "We stand by the test and while the racially disparate results are unfortunate, we believe that doing well on it is an essential part of making a good officer. We don't believe the test is fundamentally racially biased." Standing behind "we didn't want to get sued" just makes them look like they were more concerned with appearances than with fairness.
hrf:
Like I meant to say, I think you and I generally agree. And I think just all commenters (regardless of where we think the court should have come down on this) agree that this whole thing could have been avoided. That the city played this about as poorly as they could have. First they created very restrictive criteria for promotion, then they threw out the criteria when they realized how restrictive it really was.
They designed a system that would tie their hands. And then acted flabbergasted when they couldn't move.
Standing behind "we didn't want to get sued" just makes them look like they were more concerned with appearances than with fairness.
"We didn't want to get sued" is probably more about not getting sued than it is about appearances per se. As it turns out, they got sued anyway, and maybe there's a lesson there, but if you're some kind of government agency (or anyone at all, really) with constrained resources, you've got to take seriously the possibility that you can avoid spending millions of dollars and disrupting your own operations for years.
It doesn't seem to me that there is any way that all of the participants in this can go away from it feeling that they have been treated "fairly". In that sense, the city can't win. Somebody gets pissed off no matter what, all they can do is try to minimize the damage from the things they already think they know about.
"We didn't want to get sued" is probably more about not getting sued than it is about appearances per se.
I don't doubt that they tossed the test because they didn't want to get sued. And I'm not unsympathetic to them not wanting to get sued. But I think that motivation, frankly, is a terrible primary one. Some self-reflection on whether the initial test and criteria were reasonable to begin with would be nice. Otherwise, it's all about self-preservation.
In that sense, the city can't win.
True. It's a Catch-22, but largely one of their own making. Once they issued that test with those criteria, they were set up for failure.
"Once they issued that test with those criteria, they were set up for failure."
Only if they weren't prepared to stand behind the tests regardless of what the results were. In the end, the main failure was wanting a particular result.
KT,
The problem is simple, and it's two-fold:
1) On pretty much every rigorous standardized test, there are significant differences in average performance between different groups (generally, Asian Americans perform the best, and African Americans perform the worst).
2) According to the four-fifths rule that the EEOC uses, those significant differences in average performance -- which apparently aren't a legal issue when they show up on every MCAT, SAT, LSAT, ASVAB, etc. -- are implicit evidence of bias on a job-related test such as the firefighter test.
The SCOTUS ruling wasn't broad enough to challenge the four-fifths rule itself, but that's at the heart of the problem here. Going by the logic of the fourth fifths rule, I wonder if someone will challenge the validity of those other standardized tests I mentioned above. It seems logically inconsistent to acknowledge average performance differences by ethnic groups on military and academic standardized tests and then act like it's a scandal when similar performance differences show up on a firefighter test. Logically, there are really only two possibilities: either average differences in performance on job-related tests such as the firefighter test are to be expected, given similar performance differences on every other standardized test; or, the disparate average performance by ethnic groups on job-related test is implicit evidence of bias and the same is true of the disparate average performance on every other standardized test.
"1) On pretty much every rigorous standardized test, there are significant differences in average performance between different groups (generally, Asian Americans perform the best, and African Americans perform the worst)."
Seriously, will you lay off already! You make this same argument (and usually as the very first point in your larger argument) *every* single time a discussion about minorities (read black people) and merit comes up. Really, your larger point in this comment is essentially: "black people score lower on every single type of standardized test and therefore really are intellectually inferior. It's just that no one wants to admit that because it's not P.C."
This is coming from a black woman who spent her entire K-12 education in GT / Honors courses, currently has a PhD, and regularly scores at the high end of any standardized test: Your comments on minorities (black people) and merit are a one-way ticket to Stereotype Threat. Give it a rest already.
Socgrad:
You have failed to refute his point, or even address it. All you are doing is asking him to stop raising an argument based on some fairly solid facts.
Socgrad,
I haven't made any gratuitously invidious comments here, so I'm not sure why you feel the need to claim that my "larger point" is to put down black people. Give that a rest.
The reason I brought up the performance differences on other standardized tests is because they are directly relevant to this case. On academic standardized tests (such as the ones you regularly score at the high end of) or on the military's vocational test, no one presumes bias when there are differences in average performance by race or ethnic group, and yet they do presume bias on non-military, job-related standardized tests. Do you think this is logically consistent? If so, why?
As for your point about stereotype threat, let's assume that you are correct and the reason why Latinos and African Americans (read: Latinos and African Americans, not "black people"), on average*, tend to under-perform other groups on standardized tests is because of that. What, then, would be your proposed solution? To scrap all rigorous objective tests entirely? Then you are left with subjective measures of merit in hiring, which opens the door to nepotism, cronyism, and racism in making hiring decisions.
As I've suggested elsewhere, the best balance between hiring the best-qualified workers and achieving the political goal of diversity in municipal hiring may simply be quotas: decide how many candidates you want from each ethnic group and then pick the best candidates from each group. That would be better than lowering standards for everyone. If you have a better solution, I'd love to hear it.
*To make explicit what I assumed was implicit by the phrase "on average": this doesn't mean that there aren't individuals in each group that score very well or very poorly on these tests; there are.
In recent studies, I came across a factoid about standardized testing. A fundamental part of test design that all designers must adhere to - reliability and validity. Reliability refers to whether the test is culturally neutral; validity asks whether the test actually measures what it says it measures. If either of these things are in question, then the credibility of both the test and the administrators are in question. It's a difficult science, and it's controversial, as with the 'SAT is culturally biased' accusations of the 80's and 90's. I believe most administrators would want to toe the line (to avoid situations like Ricci), but it's hard. Not in the "oh so much work" sense, but even just where to start. When confronted with the data from the results, the city was right to question both reliability and validity. They should have gone more conclusively after the test, though, and tried to follow up with something peer-reviewed as reliable and valid.
Our culture loves tests. When you question the purpose, the validity, or the results of tests, you immediately get painted as a whiner. I think the controversy stems from our almost religious adulation of these powerful sheets of paper that sort out the good from the bad. Attacking tests attacks the culture's notion of righteousness.
That said, I would love to hear the pedagogical justification for a written test to judge the worth of a firefighter. Seems more like a visual-spatial and kinesthetic kinda job. I wonder if they took points off for spelling in the essay section.
That said, I would love to hear the pedagogical justification for a written test to judge the worth of a firefighter.
There couldn't possibly be an intellectual component to firefighting -- it's all just brutes hauling hoses around and driving cool trucks.
Is that really my argument? Wow, that's not what I thought I meant. Read up on multiple intelligence theory, I don't have a prejudice on which intelligences are legit and not legit. I just think the best firefighters are the ones who are best equipped to deal with the phenomena in front of them. A multiple-choice linguistic-based test is not going to be the best choice, in this case.
Im a big fan of multiple intelligence theory. Thanks for the plug.
http://en.wikipedia.org/wiki/Multiple_intelligence_theory#Gardner.27s_categories_of_intelligence
Actually if you go and read the decision you will find that in point of fact it has been argued for many years that written tests are the wrong way to go when it comes to standards for promotions with firemen. It is what it is.
"That said, I would love to hear the pedagogical justification for a written test to judge the worth of a firefighter."
First, the test wasn't to be a firefighter, but to be a firefighter officer, i.e., a supervisor and leader of firefighters, someone who needs to make quick, correct, life-or-death decisions in emergency situations. And second, presumably, the pedagogical justification is that there is a lot of technical information a firefighter officer is required to know, about different structures, types of fires (e.g., chemical, electrical, etc.), different ways to put out fires, etc.
I'd just note that the tests weren't for the position of firefighter per se, they were lieutenant and captain exams. At that level, there are definitely things you want your officers to know that can easily be tested on written exams (current laws, procedures, common building structures, some basic knowledge of flammable vs. non-flammable materials, etc.)
On the flip side, I think most reasonable people realize that being a good fire captain probably involves lots of things that aren't easily tested on a written exam. Personnel management and public relations come to mind immediately, and those are also skills where having some diversity among the officer corps contributes significantly to a well-functioning department.
To your point about the justification for why a written test is even needed or weighted so much for a firefighter Ginsberg absolutely went all in on that one. I don't know if she was just repeating what had been presented or if she did her own research but she pretty much pointed out the folly of using written tests as a barometer for promotion to a job that relys moreso on stuff that wouldn't be on a written test. She also went through New Haven's troubling history of discrimination with firefighters. Very interesting stuff.
By the way, a question I submitted over at scotusblog is this...
What does this mean for minorities who HAVE been discriminated against because of the use of a racially biased test or criteria now who may try to bring a lawsuit which may end up in front of the Supreme Court? It would seem that the majority decision makes the burden of proof in that situation incredibly high now.
My cousin is a firefighter, and took firefighting officer tests (not in New Haven, but he says they are all similar). The test is basically fire science combined with some behavioral stuff.
1.) "Given fire conditions X, how long until a building of type Y collapses?"
2.) "What special hazards are often present in a Z building?"
3.) "At time X, in a building of type Y and size Z, how many people can you expect to be present?"
4.) "In a fire of type Z an a building of type X, where will people typically flee to?"
Firefighting requires visual-spatial and kinesthetic skills. It also require you to be know how a fire works, particularly at the officer level (the guy giving orders). Do you have 7 minutes to get everyone out or 9? If you've already evacuated 15 people, do you expect there are any more inside? Should the grunts carry a 25lb hazmat kit, or is that just dead weight?
You've improperly defined the statistics terms you've used. Reliability and validity are actually content neutral and have nothing to do with race. In fact, the terms are used in statistical tests that are entirely divorced from human subjects.
Reliability refers to the ability of a test to produce consistent results from consistent inputs. One form of this is test-retest reliability: Does the test produce the same results when it is given to the same subjects multiple times?
Validity refers to the ability of a test to accurately measure the variable it claims it's measuring each time it tries to measure that variable. Obviously this can't happen unless the test is reliable. Thus, reliability is a necessary part of validity.
The classic simile involves darts. A dart thrower who hits the same part of the board -- let's say the bottom right edge -- every single time is completely reliable. A dart thrower who completely misses the board every time is completely reliable. But only a dart thrower who consistently hits his target -- let's say the bull's eye -- is, in statistical terms, valid.
Count me as somebody who likes tests. I don't love them. I like them. Should written tests be the only metric by which we gauge a person's aptitude? Whenever possible, we should probably make the habit of viewing objective test results alongside other metrics (for the purposes of this test, I would think interviews with the applicant, his supervisors, and peers, as well as a thorough review of past performance evalutations would be appropriate).
But, yeah, I think it's foolish to throw out a test simply because we question whether or not the ability to perform well on the test would be materially relevant to the job. Personally, I do think it's relevant. But this case was never about whether or not the city should use a test. It was about whether or not and why the city should be able to cast out a test (and all scheduled promotions) based solely on the unforeseen results of the test.
When confronted with the data from the results, the city was right to question both reliability and validity. They should have gone more conclusively after the test, though, and tried to follow up with something peer-reviewed as reliable and valid.
To me, this is the heart of the issue. The city threw out the test not because they thought it was biased, but because they were afraid to get sued. That's not only their practical justification, but their legal justification. That's weak. For me, the case doesn't have any kind of moral heft unless you establish whether the test is actually biased or not. The state didn't even bother to try to find out. They just screamed OH NOES and invalidated the results. In what way is that useful to anyone? I believe the SCOTUS issued new rules today about when it's approproate to take action because of a fear of lawsuits and when it's not. Perhaps that will be useful.
I know the case is tied in to much larger issues, but the more I read about it, the narrow and esoteric the thing seems. I can't get worked up about it one way or the other because all the decisions have been so narrow and the deep moral/philosophical questions seem quite far from what folks are mulling over on the bench. That said, Ginsburg's dissent is a tour-du-force. Taking it as a policy statement, I could not agree more.
Bitten on the ass by my own snark. And it's funny how one gets the most response when one frames an argument poorly.
I don't know the nature of this test in particular, I'm just an advocate for more authentic assessment in all cases. I don't disagree that science, math, and vocabulary play a role, it's just that I was assuming that it was over-weighted, as it is in most educational assessment. You test what you can test, and it's easier to frame a collapsing building as an algebra problem than as a complex, multi-layered judgment call. I'm not invalidating the algebra, more saying that the emphasis on this one way of knowing can come at the expense of other ways of knowing.
Anyway, I agree that a fair assessment would include a written component in addition to several types of performance assessment, virtualization and modeling tests, and administrative skills assessment. This is closer to what I meant, though it lacks the snark I love so well.
BreakerBaker:
But this case was never about whether or not the city should use a test. It was about whether or not and why the city should be able to cast out a test (and all scheduled promotions) based solely on the unforeseen results of the test.
good point, I bow
Roberts finally sides with citizens instead of government. So at least we now know that he values the rights of the oppressed white majority above the authority of government.
It's always a shock reading two viewpoints on one issue that are so diametrically opposed. The opinion and dissent are from two different worlds - one purely legal and the dissent by Ginsberg that takes an account of history as well. (I skimmed the opinions though, so I may be wrong on that note.)
I thing the history of discrimination in the country is important - but part of me keeps wanting to know why no black firefighter scored high enough. Part of me is just skeptical on legislating fairness, or even using a test to insure fairness.
"I thing the history of discrimination in the country is important - but part of me keeps wanting to know why no black firefighter scored high enough."
From the summary that TNC linked to a couple of days ago, I think that one of the primary reasons for the racial gap was because the way they did the scoring meant that the test was highly susceptible to the problems of small-number statistics. The key problem with the test to me was the following: "Written questions would count for 60 percent of the total. And they were ranked by exact score, rather than rounded into categories spanning, say, five-point differences, like letter grades. And under the city's charter, only the top three scorers for any given position could be promoted into it." This means, as I understand it, that the same set of people are likely to repeatedly fill up the top three slots - a well-qualified black firefighter might come in fourth, frequently. It doesn't mean that the black firefighters necessarily fell so far behind their white and Hispanic co-workers on the test. And, in fact, "Nine black and six Hispanic candidates had passed the tests" - they just didn't manage to fill out any of the top three slots for the particular positions in question.
All in all, my sense is that the test was just poorly designed. This isn't to say that it was necessarily racially biased in the traditional sense, just that the level of granularity imposed by the decision criteria meant that the sort of result which occurred (in which no black firefighter was eligible for promotion) is not especially unlikely. Leaving aside the question of whether there is an underlying racial gap in ability and also whether the test even measures the most pertinent skills, the small size of the sample and the particular ranking used means that getting a result which looks biased may just be an outlier measurement in the base distribution. It doesn't prove that the black firefighters were less qualified - even in the sense of this particular test - and it doesn't even prove that the test itself had built-in racial biases.
"Part of me is just skeptical on legislating fairness, or even using a test to insure fairness."
For me, that's what it really comes down to. If it's important to have a racially diverse set of officers (and I think that it probably is), then that should explicitly become part of decision process. Americans are wedded to the idea that it's possible to rank people on "merit" - and that it's possible to numerically measure said merit. I think that the reality is a lot murkier. I'm reminded of the Supreme Court case on affirmative action and college admissions a few years back where UMichigan's method (assigning a number to each applicant and adding a bonus for being a minority) was determined to be unconstitutional while Harvard's wasn't (a more qualitative, subjective adjustment based on race, among other things). Harvard, every year, gets enough applicants that if it chose, it could compose a freshman class where everyone scored perfectly on the SATs, or, less controversially, one where everyone was valedictorian of their high school. While this might be "fair" from a strict numbers point of view, it's not clear that this is the best way of achieving Harvard's objective - wanting a freshman class with a broadly diverse set of interests and backgrounds. In the end, "fair" is always going to be a subjective quantity and at some level, we have to just trust people to act in good faith.
Basically, I think that New Haven probably approached the problem wrong from beginning to end. I get the sense that they were more interested in avoiding accusations of racism than in determining to what extent having a diverse officer corps was important.
For me, that's what it really comes down to. If it's important to have a racially diverse set of officers (and I think that it probably is), then that should explicitly become part of decision process. Americans are wedded to the idea that it's possible to rank people on "merit" - and that it's possible to numerically measure said merit. I think that the reality is a lot murkier.
Ths is most definitely the heart of the issue, but Americans are so wedded to the ideal of "colorblindness" as an attainable and desirable goal (I would disagree on both points) that the argument that "merit," might include something like racial/ethnic/cultural background or gender or etc. is anathema.
So we get this really weird discourse in which there is a dichotomy between "merit" which is an objective quantitative measure (as if) and "preferences" which is subjective and unfair. In reality, "merit" differs significantly depending on what people decide the optimal outcome is and "preference" is something that you give anyone whom you chose. The question as it relates to justice is: did everybody get a fair shot?
But answering that question pushes us back into the territory of the dichotomy that you've pointed out. What counts as meritorious? How can/ought it be measured?
Our reliance on tests to measure aptitude (rather than discrete and concrete knowledge acquisition) is, I think, more superstition than science. The only thing SATs and ACTs have been shown to measure acurately is success in the 1st year of college. And yet, we've let this tool dominate the college admissions process. Why? Because it makes us feel better to have a numeric decision rule.
As a person who grades papers, I totally sympathize, but as all my fellow graders know, pretending there's no subjective element is total rubbish.
Hell, as someone who used to grade problem sets, I sympathize. I used to divide problems into points - so many points for getting this part right, so many for a different part, etc. But then you run into the kid who approaches the problem from such a vastly different direction that the point divisions don't necessarily work well. Or once I found that a kid had gotten almost the entire problem correct - but when talking with him later, he hadn't understood a bit of what he'd written down and had basically stumbled on most of the correct method by sheer accident. Multiple choice tests might have removed my own subjectivity, but wouldn't have made the test itself any fairer in the sense of measuring ability. Perhaps, in the end, that's the real message. For all of our best efforts, there's a certain amount of subjectivity in the world and we'll never be able to entirely design it way. Nor, perhaps, should we.
Colorblindness is absolutely an attainable goal. The manner in which you can accomplish it is very simple: withhold from the decision maker any knowledge of color.
As for defining "merit", you simply pick your objective. In this case, the objective is "minimize # of fire related deaths". Then "merit" is "best able to minimize the number of fire-related deaths/injuries".
You may not be able to perfectly measure everything, but "# of fire related deaths" is an objective quantitative measure.
And yes, you will get a different answer if you define merit differently, e.g. "best able to sell insurance." So what?
I generally agree with a lot of what you're saying, but the implication that racial/ethnic/cultural background or gender should be viewed as meritous seems to me to be a pretty dubious one.
While you're right that merit is not something that can plausibly be gauged solely through objective means, I'm at a loss to determine why it is more unseemly to determine who is and is not worthy of promotion based on the results of a multi-tier process that is gender and color blind than it is to multi-tier process if the results can be trumped to serve gender or racial quotas.
Mind you, I am not talking about hiring or college admittance procedures.
NinjaZombie, you're right. I misspoke. Colorblindness, in the literal sense of "having no knowledge of the color of applicants" is attainable, but it's not desirable because the "colorblindness" in this sense ignores the social reality that there ain't no such thing in our lived experiences, the experiences that create the social conditions for our life chances (See Ginsburgs dissent for a more eloquent exegesis of the reasons why than I could possibly muster).
In terms of "minimize # of fire related death" as an objective standard, I think you illustrate the point. First, minimize the number of fire related deaths is a retrospective or purely theoretical question. Applicants can not be objectively shown to minimize the number of fire related deaths prospectively. What one can measure is their capacity to minimize the # of fire related deaths. The question then is: what are the best measured of such a capacity? Does this written test attempt to measure whether candidates have the firefighting acumen to adapt to fast-paced, high risk situations or does it measure whether they have a basic understanding of the physics of fire? Does it attempt to do both? The truth is the ability to minimize the # of fire related death is probably constituted by both these things. A writen test can likely only show mastery of the later. If it's a good test (which is in dispute up the thread) then it might be a necessary component of figuring out whether the applicant has the desired capacity, but it would not be sufficient.
In the end, 'm not really arguing about whether the test is good or not. I don't know and apparently, the state never even attempted to find out, but I am making a larger argument about how we've talked ourselves into thinking that everything can be made objective and the further fiction that only those things which are objective are fair. I would argue that's simply not the case. Fairness is an essential criterion, but it's not always easily determined especially when dealing with social issues.
Regarding colorblindness, it seems quite desirable here. Fire is colorblind, so an optimal decision procedure must also be.
Indeed, it may not be the best possible measurement. But it is the best presently available measurement. All your objections are valid. All measurements are noisy. One may be able to construct better firefighting tests.
However, until you actually have that better test, you have no reason to discard this one. It's imperfect, but it's better than nothing. If you play a poker variant where you see 2 of your opponent's cards, you don't ignore that info just because you can't see the other 3 cards. You accept your uncertainty and make the best guess possible. You don't ignore the info you have, and make a guess based on irrelevant criteria.
Regarding objectivity: objectivity of a test simply means that the scoring of it is unrelated to the subjective opinions and perceptions of the grader. Any multiple choice test, for example, meets this criteria. A test need not be accurate to be objective.
As for fairness and objectivity, I definitely disagree with you here. If an exam is not objective, then the outcome will change depending on the subjective opinions of the grader. So is a subjective exam fair regardless of who's subjective opinions influence it? Or are only a certain set of people capable of giving a subjective fair exam? If so, which people?
Real issue now is, will the black firefighters sue the city and say the test was discriminatory? If they do, the SC will be in the trick bag. How could they not take the case. New Haven will probably have to settle the suit to avoid complete confusion and chaos. I don't know if this is is a bad decision or not by the Court but it is a trip that the Court took the case considering the issue.
If they do sue, the city will end up in the very place it was attempting to avoid - in court because it administered a test that might have been biased. Heads I win, tails you lose. In that sense, I have a lot of sympathy for New Haven. It seems to me they were trying to do the "right" thing.
I wonder if this will put more pressure on the test-makers. I can imagine a city requiring the testing company to assume liability for suits stemming from biased results.
Here is a theory for the legally educated. I call it the cash register theory. When I was a teenager, I had a job at the mall, my first boss warned me about the cash register. "Just be honest about your drawer." "Sometimes it will be over, sometimes under. No one will get suspicious unless it is always perfect." When asked why, he explained that it would be clear that I was manipulating the money if it was always right.
I would suggest the same applies here.
And for a layer of irony, I am a white fire captain promoted from a test that was challenged by other white candidates because too many minorities were promoted. A federal jury found in their favor regarding disparate impact. We're still waiting for the judge to rule on remedies.
That sounds interesting to me. Assuming it wouldn't be giving away too much personal info, is there a case description or case record you could link to?
@ The Ninja Zombie
First, if you're going to use block quotes on my comment, use my comment as I wrote it, don't cut and paste the comment to reconfigure it to your liking.
Secondly, stereotype threat is a real issue in standardized testing for minorities (and white women). It not that "someone, somewhere may suffer from stereotype threat" because of comments like Dave's. What I'm saying is that minorities (and white women) *do* experience stereotype threat when they are routinely confonted with comments like Dave's and the attitude of the inherent inferiority of minorities and women that underlies those comments.
Third, I didn't "fail to refute" Dave's point. I'm specifically not addressing merits (lack thereof) of the point because it has already been addressed by other commenters, *every* single time he brings it up (including in this thread). I'm not going to keep beating my head against the wall of his disbelief.
Fourth, I didn't say please.
Socgrad,
See my question above about stereotype threat. Assume that is responsible for the average performance differences between different groups on all rigorous standardized tests. What then? What is your solution? I suggested a solution above. I'm interested in hearing yours.
Dave,
Well, a first step would be education about the concept, so that people would not think it was "ridiculous and offensive" as Ninja put it.
And, okay, I know it's just Wikipedia, here's what they have to say about tested solutions. They only speak about African Americans here, but I'm sure the same could be applied to girls, and let's say, engineering tests, for instance. I have no idea if these solutions really would work, or what percentage of scores it would truly affect. But, it seems fair to me to at least explore stereotype threat as a possible factor in test scores:
If the results of this intervention are actually replicable, then maybe they'll do something similar before firefighter exams.
Socgrad, I didn't cut and paste your comment. I completely
rewrote it, preserving the main point but rewording, with the primary goal of mocking your post.
As for stereotype threat, I googled it, and it sounds ridiculous and offensive to me. Are minorities/white women so fragile that hearing viewpoints they disagree with/dislike causes them to perform poorly? I'm not saying it's wrong (I don't know enough about it), but I do find it surprising.
Incidentally, I skimmed this thread. I did not observe anyone refuting Dave's point. In this thread, he only has one reply (yours) which explicitly does not even attempt to address his point.
I skimmed the previous Ricci thread, but saw no refutation there either. Could you point out a comment I missed? Also curious, which of Dave's facts were refuted? Was it the point about racial differences in most test scores, or is it the point about the EEOC's 4/5th rule?
The fact that you had to google stereotype threat because you knew nothing about it beforehand demonstrates that you need to do a *lot* more reading and research on racial disparities in testing (and race relations in America) before mocking my comment.
As for the mocking, if you act like a snotty two year old, you'll get treated like one, with me essentially ignoring you.
Notice I did reply to Dave, he actually had substantive things to say in his reply.
@ DaveinHackensack:
I was wrong to assume that you were making invidious commments about black people in your post. You may in fact be that rare soul who can look at the results of multiple standardized merit tests that rank average black scores as the lowest of all racial groups and *not* conclude that black people as a group are intellectually inferior.
I didn't say, even imply, that I think the sole reason for the racial gap in standardized tests is due to stereotype threat. What I said is that stereotype threat is a real phenomenon, it does affect minorities' and women's scores on standardized tests, and that being regularly told by people that your group (on average) always scores the lowest on standardized tests is one of the things that causes steretype threat.
I do understand the difference between low average scores for a group coexisting with high scores for some individuals within that group (graduate training remember). I do also understand that *never* in our country's history have society, media, or politicians made that distinction. Accepting the *truth* of consistently lower average scores for black people on standardized merit / ability tests (especially when the race-neutrality of those tests is still in question) *will* translate into a mainstream acceptance of the idea of *all* black people being inherently inferior. Every black person, regardless of their actual ability will be assumed to be less capable, less competent, and less qualified than the white person next to them, regardless of that white persons' actual ability. Sound familiar? Aren't we as a society still trying to crawl out from under this particular rock?
Lastly, your idea of quotas as a way of balancing the desire for diversity with the need for "the best qualified candidate" assumes that you won't ever encounter a situation where the objectively "best qualified candidate" is a minority rather than a white person. In your proposal, the best qualified candidate from the black pool is the best qualified *black* candidate, not the best qualified candidate. The white candidate will always be considered the "best qualified candidate". The other candidates will be "the best from X minority group" with the implicit assumption that these minority groups are not (and can't be) as qualified as the white group.
Screw that, direct competition, with serious legal enforcement of anti-discrimination laws and a real committment to ending economic disparities in K-12 education is the way to ensure diversity *and* merit in employment.
Great post. You are much more qualified to take on DaveinHackensack than I was in the last post about this subject!
Socgrad,
I'm going to use block quotes to quote parts of what you wrote and respond to them below, but I will be quoting your words as is.
Fair enough. But what causes the gaps is less relevant to this case than that the gaps have proved to be fairly intractable over the years. That means that, unless you have a silver bullet to change that (perhaps that intervention Jennifer D. mentions above is such a silver bullet, but I'm skeptical), then you have to deal with a world in which there will be similar gaps on job-related tests. That means, firstly, not assuming that any such gaps are implicitly the result of biased test-writing.
I assumed you understood initially (I figured I was clear enough in using words such as "average" and "tend"), but then you did very well on standardized tests, were in gifted & talented programs growing up, and had a Ph.D., and were a black woman. There are two reasons I could think of why you might have mentioned that: as an attempted refutation of my point about the average group differences on standardized tests, or as an appeal to authority (i.e., I'm smart; therefore, what I say should be given more weight). I explicated the point about individual variation within groups on the chance you might have mentioned your credentials as an attempt at refutation.
Please. Society, the media, politicians, and voters make that distinction all the time today. It's 2009. That's how no one on Merrill Lynch's board questioned Stan O'Neal's intellect when they made him CEO of the company*; same with Ken Chenault at American Express, Don Parsons at Time Warner, Ron Williams at Aetna, etc. That's how the voters in a mostly white state elected Deval Patrick governor, and how the voters in a mostly white country elected Barack Obama president; etc. Similarly, who in the media questions Neil Degrasse Tyson's intellect? If anything, the media goes out of its way to portray African Americans, in particular, in a positive light. Think of how often black nuclear families appear in advertisements, when such families comprise a minority of African American families in real life. Think of how often African Americans are portrayed as judges, scientists, etc. in fictional portrayals. Think of how infrequently blacks are portrayed as criminals in a show such as Law & Order.
No it doesn't. It assumes that when you are dealing with groups (e.g., a group of firefighters taking a test), it's unlikely that blacks will be represented among the top scorers in proportion to their share of the population. Although it didn't happen in New Haven, it's certainly not impossible for the highest scorer on a similar test to be black.
That's true, and that's an inevitable downside of quotas.
You assume that a white candidate will always score the highest on the test. Why? Asian-Americans routinely outscore whites on standardized tests.
We already have more than serious legal enforcement of anti-discrimination laws -- we have the presumption of guilt of any employer when non-Asian minority candidates under perform on job-related tests. And as for ending economic disparities in K-12 education, look up the history of the Abbott Decision in my home state of NJ. We spend more per student in poor, minority districts in New Jersey than the state average, and we've been doing this for two decades. This hasn't done much to ameliorate racial gaps in test scores.
*No, he is not the CEO now.
Regarding the issue of being "color blind" in hiring decisions, whenever I hear this kind of talk in a discussion about Affirmative Action, it rankles. The fire department sure wasn't color blind 30, 40 years ago. (Hell, 20 years ago.)
How do we think we got to the point in our history now where blacks can even consider applying for a supervisory job at the fire department? People died, and people fought their asses off and sued, that's how.
This stuff doesn't magically happen: one day in 1974 in San Francisco there are only 4 black firefighters (that's a true number) because of blatant racism, and then the next day (or rather, ten years later after intense litigation) they declare "okay, now we're color blind" and from then on the department magically transforms to a more fair representation, i.e., not 90% white and male.
The issue is murky, yes, and the solutions to arrive at fairness are not always obvious, but I am always at the short memory of Americans when it comes to Affirmative Action.
By the way, this is coming from someone who thought that the Ricci guys who scored highest should get to keep their promotions in this case. I just think they need to research what happened and design a better test for next time.
That should be: always amazed by the short memory of Americans.
Indeed. The Fire and Police Depts were once dominated by White ethnics like the Irish (famously) many decades ago, when patronage systems were the order of the day in major urban cities. They understood it as looking out for their own in a world where everyone else is doing the same.
There's still an advantage to be had which is rooted in those old ways; when you look at families that have generations of cops or firefighters, naturally you have to factor in a pipeline in which fathers use their personal connections from decades of service to keep those jobs in the family.
As with legacy admissions in universities, until people are willing to mount wholesale attacks on that process, they can miss me with their weak complaints about wanting a "colorblind" process.
this is coming from someone who thought that the Ricci guys who scored highest should get to keep their promotions in this case. I just think they need to research what happened and design a better test for next time.
That's why judges get the big money :-) I mean, that's easy to say, but legally the test results are either discriminatory or they aren't. You can't really rule "ok this time, but just this once" (except for Bush v. Gore of course). Either the disparate impact test is valid or it isn't.
In practice, I agree with you. New Haven seems to be acting in good faith, this stuff is tough, they're going to make mistakes, they should just try to do better next time. But I think everyone's attempt to put this case into a little corner and treat it as an exception isn't right. The disparate impact test directly leads to situations like this; people should be honest about what they really think disparate impact means and what you can deduce from it (not just in a legal sense, but in general).
always amazed by the short memory of Americans.
I hope you're from the US. Otherwise, posting a huge generalization like this, especially after the wide variety of viewpoints you see in these comments, is pretty rude.
Oh yes, thanks. I forgot the wonderful word "some" in front of Americans.
I need to read the opinions again, more carefully, but suspect that will confirm my thoughts that the roots of the problem lie in the political process of 40+ years ago.
Yes, disparate impact is part of the law, but so is disparate treatment. More importantly -- and almost certainly in order to make Title IX more palatable politically -- the law is interpreted to protect both minorities and whites.
Hence the problem which, despite comments about "clear precedent," has not previously been addressed to my knowledge (at least not based on the district court opinion, which I did read thoroughly): what do you do in a situation when either action you take potentially violates the law, only with respect to a different group of employees? When there is disparate impact but before there's a showing of an acceptable alternative that meets business requirements and has less disparate impact?
Deny promotions? Disparate treatment of the white firefighters, and it was disingenuous of the lower court to pretent it wasn't, just because no one got promoted. You set up a process and criteria; if that had resulted in all minority firefighters being promoted, it would have been accepted; but when it resulted in all white firefighers being promoted, it was not. The white firefighters were treated differently than minority firefighters in the same situation would have been, and we all know it.
Accept the results and promote the white firefighters? Disparate impact of the minority firefighters. That's not necessarily a violation of the law, of course, UNLESS the minority firefighters can show an acceptable alternative that meets business requirements and also would have had less disparate impact. But if there is an acceptable alternative, yes, disparate impact is also a violation of the law.
How do you balance the rights of ALL parties? How do you protect the rights of one side without violating the rights of the other side? Perhaps it would work if the city, rather than just throwing out the results, made a strong, good-faith effort to identify an acceptable alternative. (Of course, it would have been best to do that BEFORE using the test.) If and only if the city can make a reasonably clear case for an acceptable alternative, THEN toss out the results of the first test -- but don't do it just because you think you might lose a lawsuit. The POSSIBILITY of losing a lawsuit means that there MIGHT be a better alternative. Should that be enough to justify disparate treatment of the white firefighters? Or should the city have to demonstrate a greater probability of losing -- or even that there clearly was a better alternative? Or to put it another way, does the city have an obligation to defend its test, until there's more proof of a better alternative, rather than just concede the disparate impact issue? Can you engage in disparate treatment of one group without more support for the disparate impact claim?
(If they had done that at the time, and come up with the better alternative, and THEN threw out the results of the test, they would have been on much stronger ground. There may well be better alternatives, but it surely looks like the city didn't really investigate until long after they threw out the results and denied promotions.)
I think that's where the court is, right now; trying to find an acceptable solution that doesn't overly disadvantage either side, and therefore fits within the political compromise of discrimination law. At least in theory, putting the city to the trouble of a good-faith attempt to identify a better alternative before it accepts the disparate impact to the white firefighters.
Did they get it right in this particular case? Maybe, maybe not. Is there a better solution to the conflict between the protections granted to both sides? Perhaps.
Maybe the law should say that intentional discrimination against the white/young/male/native-born/JudeoChristian/etc. is acceptable to avoid disparate impact of minority/old/female/foreign-born/Muslim/etc., even before there is evidence of a better alternative test with less disparate impact. That is, maybe disparate impact by itself should be enough and maybe when there is a conflict between the two sides, you should put a thumb on the scale in favor of the historically disadvantaged.
There are certainly reasonable arguments to be made for such a reworking of the law. But that's not the compromise Congress -- or the American people -- agreed to.
Stereotype threat does not explain the test score gap. To the extent that stereotype threat exists, it exacerbates the existing gap.
As with legacy admissions in universities, until people are willing to mount wholesale attacks on that process, they can miss me with their weak complaints about wanting a "colorblind" process.
Working and middle class whites and Asians are by far the most disadvantaged by the current college admissions process. Both groups would cheerfully accept an end to legacy admits if it accompanied an end to affirmative action. So be careful what you offer.
More importantly -- and almost certainly in order to make Title IX more palatable politically -- the law is interpreted to protect both minorities and whites.
I agree, but one key point: the law wasn't written to protect both minorities and whites. At the time these laws were written, everyone assumed that a generation of boosting would end all disparities--that is, equal opportunity would lead to equal results. It was assumed at the time that the only protection needed would be overt discrimination against blacks (Asians and Hispanics weren't in the original equation).
Alas, the equal results never happened, and the skew that employers and college admissions exercised in order to achieve increasing levels of diversity became so egregious that whites finally complained and demanded that the laws be applied equally. Definitely a case of unintended consequences, and one that routinely sets most affirmative action fans to squawking.
I think Scalia's opinion was the interesting one--at what point will the Supreme Court be forced to face the constitutional issues involved when disparate impact clashes with equal protection?
Equal results "never" happened? Whites "finally" complained? Let's say whites just started complaining today. That would have given AA initiatives about 30-35 years to work, following generation, after generation, after generation of not just discrimination, but Jim Crow and slavery. But, in fact, some whites have been fighting against these policies since the very beginning. In the San Francisco firefighter affirmative action case I mentioned above, the case took ten years to resolve, meaning changes did not really begin to take affect until the late 1980s.
How long do you think it should take to arrive at some fairness considering the history of this country? Does 20 or 30 years seem like long enough? It does not to me.
I'm impressed by both the fairness and civility of much of the above discussion. My own view is that given the practical constraints involved, the Court made the right decision. Its impact is to ensure that problems about testing should be thought through before the test is designed rather than after the results are in, which makes sense to most people, particularly because most of us respond with a visceral sense of unfairness when those who passed a test are denied the promised benefits.
I'll add one further point that has probably not yet been made (although I may have missed it). As has been pointed out, the test included many items critical to split-second, life and death decisions that a fire captain must make to ensure the safety of building occupants and firefighters. These included the way a fire spreads, the appropriate means to contain it, and many other principles that must become second nature during an urgent situation. It is therefore important to realize that these principles were part of the study materials provided to the firefighters in advance of testing. Under those circumstances, failure to reproduce them under test conditions might not have been an unreasonable means of concluding that the same individuals might have failed to employ those principles during an actual fire, with potentially disastrous consequences.