But getting back to the time frame issue. It seems to me that by allowing disparate-impact challenges, we’re defeating the point of testing. The point of a test is to ensure that the worthy [pass] and the unworthy [fail]. In the case of New Haven firefighter tests, to ensure that the eight most Lieutenant-worthy candidates for Lieutenant became Lieutenant and the seven most Captain-worthy candidates became Captain and everybody else didn’t get promoted. And it seems to me that if the test is going to do a good job or a bad job of measuring worthiness to [pass] is something that can be determined beforehand. And if someone alleges the test fails at its purpose before hand, there’s an opportunity to 1) unbiasedly and accurately analyze that allegation on the merits and 2) correct the test if necessary 3) with less motivation for people to allege faults that are noise obscuring the signal of valid allegations of fault. Whereas allowing these after-the-fact disparate impact claims based on a hypothetical disconnect between form and function means the analysis is post-hoc (something to discourage just on principle) in an environment where considerations other than the test’s objective measurement of objective [pass]-worthiness can affect both the analysis (e.g., consideration of the identity or quantity of who passed and who failed) and the allegations of flaws (e.g., “Me fail English? That’s unpossible! Ur test is teh suxorz!”). Limiting challenges to the test’s validity wouldn’t have eliminated incorrect allegations of flaws, but it would have reduced them and it would eliminate any invidious motives for them (leaving only allegations people who are just plain honestly wrong). So it seems to me a procedural bar on after-the-fact challenges to a test other than pre-grading challenges by takers to specific elements of the test (“Question 33 was flawed”, “The team we had to lead during the simulation was laypeople, not professional firefighters, and so was unrealistic because we wouldn’t have had to explain technical terms while giving instructions during a real system”) is a good thing. Hence, a procedural bar on disparate-impact claims… is a good thing, I guess.

And in the end, I read Ginsburg’s statement “they had no vested right to promotion” and want to tell her, “and the people who failed the test had even less.” Whatever her reasons, she’d give people who failed a do-over, and that doesn’t sit right with me. Do it once, do it right, or to hell with you – that’s a world worth working to create, but not the one her opinion would move our country towards.