Tag with skepticism that merits more thought. Can this all be grounded on the stereotype threat or is that merely a transmission mechanism? Does grounding this on stereotype threat shield what lies beneath. Could deeper analysis perhaps lead to a more targeted intervention that could address these issues. A quick analogy before I do more thinking:
Black male life expectency is much shorter than White male life expectency in America. In fact, as Sen points out, Black American males live shorter than Chinese and Keralan males (White American males live longer than both). It would seem silly to me to argue that the best way to address these health shortcomings would be to cite a number of studies on vitamin deficiency to explain the difference. That's sort of what this feels like. I don't agree with the principles, I just think that you've giving too much centrality to what is essentially a symptom and not the problem.
Stereotype threat is a symptom of
societal stereotypes, and not a symptom of URM insecurities as such. The remedy, therefore, and in the ideal, would be a society without negative stereotypes. That would eliminate the problem for this group - i.e. the highly-motivated, high-achieving URM applicant who has earned a UGPA that is sufficient to earn him/her consideration for a T14/T50 school.
Short of that, one could eliminate tests that do more than is acceptable to highlight these societal faults, and to invert these faults to make them appear to be instead the fault of the people who are negatively stereotyped.
Failing both of these, it seems to me justifiable and fair to remedy this effect via affirmative action.
My point is that there are underlying societal problems, that the current heavy reliance on the LSAT severely amplifies those problems via a well-documented process, and that schools should either not rely on the LSAT or, in the alternative, should use affirmative action as a remedy.
To do otherwise would be grossly unfair, in my view.