Law School Discussion

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Infinity

Pages: 1 2 [3] 4
The problem with your argument is that it is not appropriate to bring in risk-seeking/risk-neutral/risk-averse to this discussion.  It's easy when we are talking about dollars.  A dollar is a dollar is a dollar.  But the hypothetical "assuming-all-else-equal" when applied to applicants is silly.  All else is inherently going to differ from applicant to applicant.  I mean I see what you are going for in your argument.  But if Patrocles was raised in a Somalian hut and Achilles has never known poverty, and there is no indication that there will be a huge difference in ability, Patrocles will be the better addition to the class.

I respectfully disagree with the unstated premise of your counterexample: that the URM will have a clearly distinguishable background from the non-URM.  I agree "assuming-all-else-equal" is rarely the case, but my point is not that the candidates have equal backgrounds but indistingusiable or incommensurable backgrounds (remember, we're talking here about the average candidates, the 60 or 70 or 80% of the class that are not auto-rejects or auto-admits, sure, some of them may stand out with background stories, but i sincerely think that most average applicants, are, well, average).  I picture a kind of bell-curve in terms of the quality of an applicant's background, with the vast majority, irrespective of race, being normal.

The one thing that we can absolutely NOT say with ANY degree of plausibility, is that the individual white applicant with a 170 LSAT score is -- other things (i.e. GPA and such) being equal -- more meritorious than the individual black applicant with an LSAT score of 164. No way.

Okay, does that mean we can say that the black applicant is more meritorious?  Are they equally so?

I agree with "....."'s intended point.  Also, I don't think the scoreband helps your point because, again, marginally, the person with the higher LSAT score will have a slightly higher score band, even if it is 11 points wide and there is lots of overlap. 

The URM score distribution will be shaped the exact same way after the adjustment as it was before the adjustment.  Is there a wider range in URMs as non-URMs?

True, on average, the distribution would be the same.  But, again, when looking at specific individuals and making specific admissions decisions on a case-by-case basis, that cannot be known.

And you're right about the risk-seeking\risk-neutral distinction.  I am assuming that law-schools are risk-averse and would prefer the certainty of a sure bet.  This is certainly debateable, and may be where my argument goes ashore.

First, just to nitpick, you wouldn't have to be "excessively risk-seeking" to take the first choice.  Risk-neutral is sufficient.  Risk-seeking would mean that you would take the first choice even if the expected value were less than $100.

I don't really get your argument here.  It seems to me that you are saying that the URMs would have a wider range of scores than the non-URMs.  Where does this assumption come from?

My assumption comes from the argument that "Stereotype threat" is an average.  So adjusting for "stereotype threat" for URMs would lead to a wider range of scores because "stereotype threat" does not reduce scores by a constant.

None of these ways by which URMs may be admitted, though, is what we commonly understand by race-based affirmative action. What I am proposing is that we should add one more consideration for URMs -- (iv) that knowledge that the LSAT is probably not telling the truth about their true academic ability, and that, when in doubt, the other factors in their application file (adjusted GPA, recommendations, scholarships and prizes) should be seen as more accurate indicators of their scholarly merit.

This sounds like a fairly conservative position to me, and one that I think utterly fair and meritocratic. Will it get it right every single time? I doubt it. Is it fairer and more meritocratic? Yes. Is it a better justification than "diversity"? I think so. Does it address structural racism up and down the educational, cultural and professional system? No.  Glib? Maybe.


C00per06 -- I'm thinking that this response to Southside may also fit the question that you posed. Holler if it doesn't.

Red, I agree with almost everything you say, except the quoted section above.  My question here (which I just iterated to Southside in the above post) is that I do not see how it will be possible, very often, to distinguish between two candidates, one URM, one non-URM, with comparable backgrounds without using LSATs.  You mentioned 3 other avenues of admissions for URMs, all of which I agree with, through all of which I would imagine URMs are not UR (hence, in my PLS scenario, URMs are being admitted in great numbers among what I called the auto-admits).  However, at this stage, the hoi polloi of the admissions class, i can't understand how an examination of one person's background will really distinguish them from another person by too much.  It is not as if, at this stage, the non-URMs are all dunces with good stores and terrible backgrounds and the URMs have sub-par scores with fantastic backgrouds.  What i think is more a more plausible situation is two candidates with comparable backgrounds who are differentiated only by LSAT score.  Although on average these candidates might have the same LSAT score when adjusted for "stereotype threat," individually, that is not certain, and the lack of certainty leads me to think that, on the basis of merit or fairness, the law school would have to prefer the candidates more likely to have the higher merit indicated by the higher LSAT score.

There are lots of weird logical steps in your post, so I'll just try to pick out a couple of the bigger ones that I think sufficiently damage your larger point.

First, let's take your unsupported assertion that there is no way to determine which candidates may have been affected by stereotype threat and which haven't. I don't see why this is the case. Adcomms see lots of applicants every year, and they can start to discern patterns about what sort of students tend to be successful. Obviously, they can never be sure of what any applicant will do, so the whole process is a series of educated guesses. Given the existence of a persistent stereotype threat, adcomms can and should examine other relevant factors of minority applicants more closely to look for indications that their LSAT scores are not indicative of their prospects. This educated guesswork won't be perfect, but it will be a lot more accurate than relying on a measure that is known to be biased.

Second, you make a simple mathematical error in your example of the 20 grouped applicants. You state that each individual non-URM is more likely to have an acceptable score than each individual URM, even though the scores on average will be the same when accounting for stereotype threat. This is clearly not true. Even if you had no way of predicting who was most affected by stereotype threat, you would be just as well off picking randomly as you would be picking from only the non-URM group. If you do have some holistic methods of ascertaining who has been particularly affected by stereotype threat, which I have argued that you do, then you are much better off taking that information into account.

In reality, admissions don't work much like your hypothetical, but even if they did, your points wouldn't follow.

I confess I don't know how admissions really works (nor, I doubt, do most posters here), but I thought the story I told about it was relatively convincing, and your counterpoints don't make me think otherwise.

To take your points in reverse order: I did not make a mathematical error.  If you have 2 invesment options, and the first one pays out $100 every time, and the second one pays out $200 33% of the time, $100 33% of the time and $0 33% of the time, although the two bets average the same payoff (same "expected value," as they say), if you desire $100, you will certainly take the first bet (unless you are excessively risk-seeking), because the likelihood of getting $100 is 100% in the first bet but only 66% in the second bet.  So, perhaps, you could make an argument to me that T14 law schools want the best of the best and should be willing to be more risk-seeking to get an Achilles even though it might also mean getting a Patrocles, but, and this is just a matter of considered opinion, I would think that law schools would generally prefer to take an Ajax, a known quantity that will perform admirably but without much flair.  This metaphor probably isn't helping me out since I would take Achilles and Patrocles over 2 Ajaxes, but in the decisions I was talking about, there weren't any potential-Achilles around, since all the exceptional candidates had already been selected, and the logic, I think, holds.  Particularly, I think, law schools will be more risk-averse when filling out the majority of their class, so perhaps they take a few flyers on potential-Achilles, but not the vast majority, which would end up excluding URMs on the basis of merit because of lower LSAT scores.

On your first point, I dont think it answers the dilemma i posed.  I have already said that law schools should look closely at all candidates' files, and not more closely at some than others.  Regardless, my point is that the vast majority of candidates' backgrounds are largely indistinguishable or incommensurable.  And certainly law school admissions committees might detect patterns of backgrounds that correlate with success in school, but that is irrelevant: i would imagine such backgrounds are generally spread out over candidates from all races (remember, we're talking about average applicants here, not exceptional ones).  And if that is the case, my dilemma returns: how do you distinguish between two students with similar backgrounds indicative of success, but one has a higher LSAT score.  In these cases, the ones with higher LSAT scores are, because of my point above, marginally more likely to have the level of merit the school desires.

Whew. Lots of fun stuff to chew on here:

That's mostly my point, that the evaluation (and justification) of an idea is not its theoretical soundness but its practical feasibility (see, for instance, communism and other variants of Marxism and utopianism, the Star Wars anti-missile program, neoconservatism, and, the best metaphor of all in my opinion: the Tower of Babel).  By contrast, non-theoretically sound but practically beneficial ideas are often very useful or good: Newton's mechanics, penicillin, sleep, etc.).

Another one that made me chuckle. You could unpack the absurdities in this paragraph for days; it almost reads like a quote from a Monty Python episode. My favorite point is where you label sleep as an "idea."

Well if I can't convince, I'm happy that I entertain.  Anyways, I think you were reading my line a touch too literally, though you're right that I probably shouldn't have used the word "idea."  The point is, we use and do lots of things without knowing why or without being able to explain why they work (or explaining wrongly), whereas lots of things we can theoretically understand and explain how they would work, don't actually work in practice because that nice tidy world of theory is often removed from life.  This is obviously not a novel point, but I was making it to say that we don't necessarily need a theoretical justification of AA if it provides concrete societal benefits, but this point is really neither here nor there, so I'll shut up about it.

The frame
1. I am targeting my attention on the T14 schools, and engaging the debate strictly on the basis of scholastic merit1. This is because scholastic merit is the criterion against which the people who argue against affirmative action frame the debate. One can imagine other qualities that may be relevant: “leadership”, etc., but these complicate the issue from the outset, when, for the sake of ease of discussion, I would rather that they complicated the issue at the end of the discussion.

2. I am defining “affirmative action” as any admissions process that does not systematically discriminate against equally-qualified black, hispanic and native american applicants;

1"Acquired skills relevant to the study of law". I figure that we all more or less agree on what this means -- ability to follow and anticipate a line of inquiry; to read properly; to see shades of meaning; and to think critically, independently, and imaginatively.

6By the way, I don't see what is so fuzzy about looking at applications holistically. On the contrary, it strikes me that that is the way to achieve a greater measure of accuracy in judging an applicants' scholastic merit and potential. And no, they don't look at everyone holistically -- they assume that the 3.8/175 is bright and motivated and admit him/her automatically (what I have read suggests that approximately half of each class at schools such as Penn, CLS, Harvard is admitted automatically in this way), when we all know that that is not necessarily so: you can show 3.8/175 and be as dumb as a brick.

In any case, I am absolutely not arguing for a system that blindly adds 6-4-2 points to every black-hispanic-native american applicant’s index number.

I'm going to respond as carefully as I can, Red, since your well-considered post warrants it.  You asked:

"Do you think that I have not sufficiently problematized the assumption that the LSAT is a straight race-neutral measure of reasoning proficiency?"

I take no issue with your problematization of AA.  I am not an expert in social science or on "stereotype threat."  In my postings on this thread, I have assumed as true that "stereoytpe threat" is a real phenomenon that accounts for the gap in LSAT scores between different races.  I leave it to others more skilled in these areasto critique your argument on whether "stereotype threat" is real or accounts for the entire gap.  Thus, in all my postings on this thread, i take it as fact that the LSATs objectivity as a standardized test is undermined by"stereotype threat."

In my previous long post (a few pages back) I agreed that "stereotype threat" justifies AA in theory, as you suggest, as a way to neutralize test discrepencies and give an equal shake to all law school applicants.  So I accept your definition of AA as "any admissions process that does not systematically discriminate against equally-qualified black, hispanic and native american applicants."

I also concur with your definition of "merit" as "Acquired skills relevant to the study of law," from which it follows that race is not a factor in merit (at least not ideally).

My challenge to your justification of AA arises when I ask how do we implement AA in a meritocratic manner.  I understand that you are not suggesting "a system that blindly adds 6-4-2 points to every black-hispanic-native american applicant’s index number."  But how else can AA operate in practice?  And if AA cannot operate meritocratically in practice, then i do not think AA is justified by merit.  Another way of stating this point is to ask: what would an "admissions process that does not systematically discriminate against equally-qualified black, hispanic and native american applicants."  Because this is the definition of AA, if we cannot explain what such an admissions process is in practice, then we have not justified it since all we have justified in theory is an ambiguous concept.  The normal way of answering my question is: more "holistic examinations," or greater "contextualization," answers that i find vague.

The reason why I find "holistic examination" fuzzy is that it is generally employed as an argument, disguised in high-falutin academic language, against using numbers (LSAT\GPA) in admissions.  A holistic examination should include all aspects of a candidates application, including their numerical indicators.  Perhaps in certain circumstances for particular candidates, admissions committees give less weight to these numbers (thus, people may submit addendums explaining why they underacheived).  And that is proper, but those are limited circumstances determined on a case-by-case basis.

I do not see how a reasonable case-by-case determination can be made for applicants on the basis of race.  Although I'm uncertain about this, I imagine that it is tough for admissions committees to decide between most applicants' backgrounds, either because they are mostly similar (which would make sense for most students directly out of college, or just a year or two out of school) or because they are incommensurable (how do you compare the background of a person who overcame a brain tumor with one who worked and supported a child throughout college?  Is the person who overcame the death of both parents at a young age less than, equal to, or superior to someone who knows 4 languages?  What about if the person knows 8 languages? Or started their own business?).  In the few instances where a background is exceptional, we see those applicants being accepted to schools despite sub-par numbers, though not too low (no 140s or <2.0s at T14s).

I imagine this scenario:  Princeton Law School, the newest entry to the T14, subscribes to your policy of entirely meritocratic admissions based on holistic examinations.  It makes a few auto-admits and a few auto-rejects based on holistic examinations (thus kids with both good numbers and great backgrounds are admitted, and kids with bad numbers and unimpressive backgrounds are rejected).  That leaves 70 or 80% of the applicants.  URMs were admitted at a rate among the auto-admits that, if continued for the entire applicant pool, would result in URMs becoming Ms.  However, now that all the candidates with exceptional backgrounds are out of the way, URMs are at a disadvantage in comparison to non-URMs.  All these leftover candidates are largely indistinguishable: either they have average numbers and backgrounds, or there are some with good numbers and average backgrounds or good backgrounds and average numbers, but, at the margins, non-URMs always have slightly better numbers, particularly LSAT scores, with the result that, upon ranking all students on the basis of a holistic examination, URMs are disproportionately low on the admit list.  What does PLS do?  If it admits students purely on the basis of its list, URMs will continue to be URMs.

However, the admissions committee learns about "stereotype threat," which it believes accounts for the numbers gap at the margins.  But how does it apply this knowledge?  It does not want to blindly add points to every person subject to "stereotype threat."  It wants to know for sure who is unduly affected by "stereotype threat" and to what extent to bump them up the list fairly.  What indicators of "stereotype threat" are there?  How can PLS distinguish between candidates who performed as well as they could and those who actually suffered from "stereotype threat?"  I think there are none.  Let's make this situation more concrete:

While ranking, PLS has grouped together 10 non-URMs and 10 URMs with comparable backgrounds and GPAs.  However, the URMs all have LSATs 6/4/2 points lower (relative to their race), which, if accurate, are below the school's acceptable levels given an unexceptional background.  How should PLS rank them?  On average, these candidates are all equal, and, perhaps, if PLS could admit all of them, it should.  But lets say that it cannot: they are down to their last 10 spots, or these are the 20 candidates from Florida and PLS only wants 10 admits from Florida; now the individual rankings are crucial.  Despite being equal on average, it does not follow that PLS should consider them all equal when ranking them individually, because some of the URMs might have not been affected by the "stereotype threat" and others might have been affected by it by more than 6/4/2 points.  Nor does it follow, though, that greater "contextualization" or a more holistic approach will miraculously reveal which of these URMs suffered from the "stereotype threat" and to what extent.  Since it is meritocratic, PLS would, unless it is risk-seeking (which is highly doubtful since it is a law school admissions committee), rank the 10 non-URMs higher because each, individually, is more likely to have an acceptable LSAT score than each individual URM, even though, on average, they all have the same score (and obviously LSATs are not sure things, they are only indicators, but if the only difference between two candidates is an LSAT score, the only meritocratic course is to rank higher the student with the higher LSAT).  Because it is impossible to specifically determine the affect of "stereotype threat" on any of these individual candidates, it is more meritocratic to rank 10 170s higher than 10 165s, even if some of those 165s might really be 170s or 175s if the "stereotype threat" is properly accounted for in the individual cases.  Unless PLS boosts all of the URMs equally (and blindly) it seems impossible to have a fair or meritocratic method that would not rank the 10 non-URMs highest.

Thus it seems that PLS can either have a meritocratic admissions system or an AA admissions system, but not both.

Red said "e) (I think persuasively) answers the following common questions:
--“why should the sons and daughters of  rich black folk benefit from affirmative action when the chidren of Appalachian coalminers don’t?”

To me the key word in the quote was "should"--it doesn't necessarily say why they do, though at other points she has argued that AA is in part a response to flaws of the LSAT- Still assuming one takes stereotype threat as explained by Red as a reason why AA is jusitificed based on merit of the applicants, then there is still room for other threads that focus on the historical and political realities behind it,  unless of course we assume the lawschools use AA policies purely for reasons of merit.

This was, at least in part, the point Bass made.  I do kind of wish the "should" and "how" could be separated, even though C00per6 disagrees.  C00per6's point, I think, was that without a proper "how-to", there should be no corrective action even with evidence of a disparate impact.  (I find that strange, but whatever.)

That's mostly my point, that the evaluation (and justification) of an idea is not its theoretical soundness but its practical feasibility (see, for instance, communism and other variants of Marxism and utopianism, the Star Wars anti-missile program, neoconservatism, and, the best metaphor of all in my opinion: the Tower of Babel).  By contrast, non-theoretically sound but practically beneficial ideas are often very useful or good: Newton's mechanics, penicillin, sleep, etc.).

Also, I do believe that AA is a good policy right now because of its benefits (an empirical justification), with or without a theoretical justification.

I see.  Perhaps there should be another thread that can provide a forum for AA practice, under the assumption that it is theoretical justified on the grounds described here.

That might be a good idea---I've posted my question here because i think it pertains to Red's quest for a meritocratic justification of AA, but if others choose to try to help me answer it here or elsewhere, i'll listen and respond if i can. 

Pages: 1 2 [3] 4