The Gaming Den

Frank, I think your dissection of the New Haven SCOTUS decision was a little backhanded.

The decision was a lot more nuanced than you gave it credit for. I felt the dissent had the stronger point, but the fear of litigation they're talking about is a peculiar function of how the CRA is written and enforced, not some blanket statement that applies to all law forever.

Explanation of the District Court's Reasoning in the original case.

Basically, you're right that the fear of litigation is a function of the way the EEOC interprets and enforces the law, rather than the law itself. Now, having not read the full decision, I would like to know if SCOTUS actually struck down the EEOC regulations or merely set up a true "damned if you do, damned if you don't" situation for employers facing similar situations.

The decision was extremely limited in scope. Their ruling didn't have many repercussions on the CRA and basically only said that if you're going to come up with a test that is ostensibly racially neutral then you have to abide by that test. I don't usually side with the conservative side of the bench but here, if you accepted the test was indeed racially neutral, they had the best argument.

The dissent brings up that they felt the test wasn't neutral, and I agree. However, considering that the test was designed to be neutral I think the majority decision side just accepted that and moved on.

Well, see, if New Haven went ahead and gave the promotions, they've be open to a lawsuit. Instead, they made a new test, looking at the errors of the last one...

The Supreme Court said that being fair wasn't necessary. They shouldn't have revoked the old test results when it was (and it was) found to be racist. Therefore, in the majority position, New Haven had no business acting to save itself from a lawsuit. They said that New Haven should've either changed the next test and waited to be sued, instead of preventing the action which would lead to the lawsuit.

In other words, it's okay to be racist, because it only matters if those disadvantaged are able to sue.

-Crissa

But disparate impact is measured before and after the test, by the test itself and its results.

The test was designed with best and demonstrable intentions to be racially neutral. The court's ruling is that that the results didn't bear that out is not grounds to throw out the test, and Title VII upholds that. The only real question is how much can you bring to bear on any test to determine its neutrality - the majority said that the work was enough, the dissent said otherwise.

The Court's decision is a decent way to resolve the paradox that the case confronted: how do you resolve an ostensibly neutral test that turns out not to be? Throwing out a truly neutral test simply based on the results violates the CRA, but to maintain a bad test in the face of disparate impact means you've violated the CRA. What about a test such as this, created by consultants with minority input throughout? It's a grey area and disparate impact is not measured solely by results.

If the test was implemented without concern for bias then there was ample reason to retest, but given the scenario that's not the situation. Now, I happen to agree with the minority because I think the test should have been re-examined, but I accept that's only one way to resolve the paradox.

Yeah, but it's been enough time since to know why the test gave the results it did.

But the Majority decided to ignore that, and just make a decision which is capricious - that the results shouldn't be thrown out if it's found to be in violation of title VII. Basically, you're go to go until you get caught.

-Crissa

Crissa wrote:Yeah, but it's been enough time since to know why the test gave the results it did.

I haven't seen the reasons listed anywhere, could you explain?

Crissa wrote: Instead, they made a new test, looking at the errors of the last one...

I thought they just threw out the test and hand-picked the promotions. Or was this new test developed between then and the lawsuit?

mean-liar: Lawsuits really aren't their only problem. They also have to worry about EEOC complaints. After all, the EEOC's guidance on Title VII says that a test where minorities pass less than 80% as often as whites is presumed discriminatory unless proven otherwise. Unless SCOTUS also struck down that regulatory language, you have the irony of a conservative court ruling that businesses must spend money when they could save it (by investigating whether the test is discriminatory instead of just throwing it out and finding other ways to determine promotion).

Absentminded_Wizard wrote:After all, the EEOC's guidance on Title VII says that a test where minorities pass less than 80% as often as whites is presumed discriminatory unless proven otherwise.

I have never heard this before and am shocked at this level of stupidity... it's...quite possibly one of the most fucking stupid things I have ever heard. Take the following example:

When I taught electrical engineering at university, students nominally in the same grade level (junior year) from the US were mingled with a class of students from Singapore. There were about 15 US students (all white I believe though there might have been a few Indians as well) and about the same number of students from Singapore. In lab work, the Singapore students absolutely outperformed the US students by a wide margin - easily 80%. Per the EEOC, the lab assignments were racially biased. WTF? Does anyone here seriously believe that the assignments we gave were biased against the white kids? Really?

What is even worse about this is that my university's EE department used T-Score as the basis for all grades, so all grades were relative to how the other students performed, not on some objective scale like 90+ = A or whatever (I once got a 100% on a test and my average went down in the class under a T-Score). This means that in addition to the discrimination on the assignment, the white people were further hurt by the success of their Singapore counterparts, as their GPA's went down based upon the racially biased tests.

*sigh* the idiocy abounds in America, in electrical engineering as well as in the government. Now there is a surprise...

LL, you have no concept of actual statistics and managing outside interference with your data.

In the case of the firefighters, all of them went to the same schools, were taught the same firefighting skills at the same firehouses, paid by the same wages, speaking the same language.

In your case, you have two pools of students which were taught by different schools getting different results.

While your result should result in them looking and seeing if the singapore students had a benefit - it should also look at outside factors. It might have been that the TA was helping one set of students better than the other set. Or maybe not.

But you're just to damned stupid to figure out the difference.

-Crissa

Oh, here's one online:

[url=http://www.stlbeacon.org/beacon_columnists/new_haven_s_test_was_flawed_-_why_is_it_wrong_to_recognize_that_ wrote:Saint Louis Beacon[/url]]We need to understand a few basic qualities of a test, which we often take for granted. Tests need to be valid and reliable. Furthermore, in the world of employment, tests are most helpful when they have a cut-off and/or can rank candidates in an order to aid with decision-making.

Reliability simply means that a person will score similarly over time. Validity refers to a test measuring what it is supposed to measure. For example, if we want to test individuals' ability to read a map, we would ask about their knowledge of directions and other related constructs rather than asking about their mastery of sports or politics. That seems obvious, right? Well, it is not so simple. Psychologists spend entire careers constructing and perfecting tests so that they are valid indicators of what they intend to measure. All of these test properties are collectively referred to as psychometrics.

The test in question lacked validity, a cut-off and the ability to rank order. Some of the flaws with the New Haven exam are cited in the dissenting opinion . With the test, a difference of several points was not statistically significant enough to be used in promotion decisions. It also did not differentiate between qualified and unqualified candidates. Previous promotional exams have been invalidated because of these shortcomings.

Basically, several of the items were not statistically valid. They weren't yes-no questions, but instead performance evaluations.

New Haven revoked the test after realizing this, and delayed promotions until they fixed the errors on the test, requiring those who passed to take the new test and be graded again. This resulted in many white guys who passed the prior test to not pass the new one with verified results.

That's not discrimination.

Crissa wrote:But you're just to damned stupid to figure out the difference.

Perhaps we may, at some future date, develop some sort of spray to fumigate your filth from the planet. Your very existence consumes oxygen I could be putting to better use. I have had crabs that I cared more about than your opinion. In short, If you can't express yourself politely, kindly shut your suckhole until I tell you I have need of it, mmmk?

Of course they went to different schools, of course the comparison isn't the same. That is exactly my point. My case shows how results can be perfectly fair and yet fail to pass the EEOC test for discrimination. The point is that the assumption of the EEOC is that any test that generates those results is discriminatory unless shown to be otherwise, which places the burden on the organization conducting the tests to dredge up all of the data and prove itself "innocent" when tests generate something other than a bell curve that is agnostic across racial lines (read as: every failure has a free lawsuit waiting if they don't like the outcome of a test under this EEOC stupidity). That sort of reasoning leads to madness.

Now, in the New Haven case, of which I have not heretofore mentioned, the test may very well have been invalid; I have never claimed otherwise and your useless points about the New Haven case are duly forgotten. My only point is that the a priori assumption of discrimination is ridiculous; it is especially grievous when it places the burden on the testor to justify itself.

Yeah, like why don't I just ask the NBA to you know, go off and spend a fuck-ton of cash showing me that its under-representation of whites is *not* discriminatory.

Koumei wrote:I don't know what's funnier: the hypocrisy of that sentence or the concept of people ever being polite on The Gaming Den.

Yeah good point. Just been a long few weeks - too much work and not enough gaming I guess.

Koumei wrote:I even managed to get a post edited by a mod!

Koumei wrote:

Lich-Loved wrote:In short, If you can't express yourself politely, kindly shut your suckhole until I tell you I have need of it, mmmk?

I don't know what's funnier: the hypocrisy of that sentence or the concept of people ever being polite on The Gaming Den.

I'd take the open hostility of TGD over false smiles and secret plots of, well, most of the message boards in the 'net any day of the week.

Fuck all of you?