New Stanford Law Review Article about black law students

[John Galt]

i disagree with the conclusion of the author, however I also believe JohnGalt's points are weak and almost preposterous.  If you're willing to go along with his argument that those in the bottom 10% of class ranking are likely passing the bar at the highest rate, then good luck

HA! Weak and almost preposterous? I would expect sharper critism than this inlight of all our LSAT training. Galt didn't make any conclusions; he simply pointed out the flaws of the argument.

Good job Galt! Here's to critical reading and analysis.

Obviously King can't read.

[Nemesis]

i disagree with the conclusion of the author, however I also believe JohnGalt's points are weak and almost preposterous.  If you're willing to go along with his argument that those in the bottom 10% of class ranking are likely passing the bar at the highest rate, then good luck

HA! Weak and almost preposterous? I would expect sharper critism than this inlight of all our LSAT training. Galt didn't make any conclusions; he simply pointed out the flaws of the argument.

Good job Galt! Here's to critical reading and analysis.

Obviously King can't read.

Kaplan - it'll get you everytime.

I KEED! I KEED! 

[King80s]

In fact, there is the alternative that the black students with higher grades were the ones who did poorly on the bar.

enough said.

by the way, the study clearly makes the correlation between class ranking and bar performance.  If you know anything about correlations, then what JG posted is hard to digest as a 'flaw'.  When there is a high probability established, you merely pointed out the exceedingly low probability scenario.  Are low probability scenarios 'critical thinking', apparently they are.  Again good luck

[King80s]

i disagree with the conclusion of the author, however I also believe JohnGalt's points are weak and almost preposterous.  If you're willing to go along with his argument that those in the bottom 10% of class ranking are likely passing the bar at the highest rate, then good luck

HA! Weak and almost preposterous? I would expect sharper critism than this inlight of all our LSAT training. Galt didn't make any conclusions; he simply pointed out the flaws of the argument.

Good job Galt! Here's to critical reading and analysis.

uhh.. not sure what lsat training you're referring to, but it's basic logic and statistics..  It isn't a flaw if someone doesn't establish a causal correlation between two correlations.  Causality is rarely even an issue in human behaviors.  The author never implied the two correlations were connected, but used them as multiple lines of support.  Hopefully you've had some statistical training outside of 'lsat study'... but it's rather amusing this is the basis of your argument / personal attack.

Since JG's argument is essentially a straw man, as are the other points, there's no need to seriously consider them.

[John Galt]

i disagree with the conclusion of the author, however I also believe JohnGalt's points are weak and almost preposterous.  If you're willing to go along with his argument that those in the bottom 10% of class ranking are likely passing the bar at the highest rate, then good luck

HA! Weak and almost preposterous? I would expect sharper critism than this inlight of all our LSAT training. Galt didn't make any conclusions; he simply pointed out the flaws of the argument.

Good job Galt! Here's to critical reading and analysis.

uhh.. not sure what lsat training you're referring to, but it's basic logic and statistics..  It isn't a flaw if someone doesn't establish a causal correlation between two correlations.  Causality is rarely even an issue in human behaviors.  The author never implied the two correlations were connected, but used them as multiple lines of support.  Hopefully you've had some statistical training outside of 'lsat study'... but it's rather amusing this is the basis of your argument / personal attack.

Since JG's argument is essentially a straw man, as are the other points, there's no need to seriously consider them.

King, you do understand that Sanders argues that there would be an 8 percent increase in the number of black lawyers graduating and he argues more of them will pass the bar. NONE of the evidence he produces even remotely supports such a claim. He just pulled that out of his arse. More likely, if black kids that go to Georgetown, or Cornell could only get into a lower ranked schools, they'd probably wouldn't apply to go to law school anyway. He assumes that just because the percentage of black students in law school has stayed constant for a number of years that it would stay constant even after affirmative action was abolished. One need only look to the UC system after proposition 209 and to schools in Texas after affirmative action was banned there to see that Sanders argument fails the most basic test of plausibility.

Next, if you notice in his study, he uses the ABA study of bar test takers broken down by race in a number of instances, but to prove this dramatic gap in passage rates, he uses only testakers in the California bar and then generalizes from there. One would have to ask, why did he not use the ABA data that he so heavily relies on throughout his analysis to help substantiate this most important point to his argument.

Next, he doesn't use class rankings to establish that a certain set of kids fail the bar. He uses two things: 1) general correlations between class rankings and bar passage rates; and 2) general correlations between LSAT score and class rank. The problem with this whole line of reasoning is that he never establishes that these correlations relate to the group he is studying. As has been proven many times, general correlations don't necessarily carry over to being consistent when applied to groups. In fact, specifically, there are many studies to show that the predictive value of the LSAT is diminished greatly when predicting the performance of African Americans. By assuming that these correlations apply to a specific group as well, he fails to eliminate the alternative I was talking about. I never argued that the alternative was true or plausible.

Sanders never establishes how failing the bar the first time hurts African Americans other than it may hurt their job prospects. This claim is dubious at best. First, most firms hire students before they take the bar and allow for the student to take it multiple times. A person's career is not in jeopardy after failing the bar once. Next, if all these african americans began to go to lower ranked schools, it is not clear how their job prospects would get any better. In fact, most of the data suggests that their job prospects would get even worse.

King, even the most die hard Sanders supporters must acknowledge the flaws presented. He presents a compelling analysis, but not a persuasive one. For a person so quick to boast about basic stats, it is a wonder you haven't criticized Sanders for his equivocation and shady mathematical analysis.

[Nemesis]

uhh.. not sure what lsat training you're referring to, but it's basic logic and statistics..  It isn't a flaw if someone doesn't establish a causal correlation between two correlations.  Causality is rarely even an issue in human behaviors.  The author never implied the two correlations were connected, but used them as multiple lines of support.  Hopefully you've had some statistical training outside of 'lsat study'... but it's rather amusing this is the basis of your argument / personal attack.

Since JG's argument is essentially a straw man, as are the other points, there's no need to seriously consider them.

No one is trying to personally attack you. The flaw is mistaking a correlation for a casual relationship. At best,  a high positive correlation simply suggests a casual relationship. And the answer is yes, I have had extensive training in statistics. The analysis doesn't even present statistically sound support for the conclusion attempted.

[thenextstep]

The other issue I have with Sanders analysis is failing to account for social class in his analysis.  Not because class is more important than race or anything of that nature (because that's not really true either), but because so many thousands of studies have found how important social class is in predicting academic attainment and other forms of status attainment.  We have mountains of proof that social class explains a great deal of performance on standardized tests, which when he is basing so much of his analysis on how LSAT score predicts law school GPA which predicts bar passage, he needs to account for class.  Maybe it doesn't change his findings, but it's a gaping hole right now.

Also Sanders does not adequately deal with the problem of individual effects versus school effects.  To do that would require a much more complicated statistical technique (hierarchical linear modeling), but I'm sure he could do it.  Analysis of K-12 schooling and performance using that technique has found strikingly different results than doing analysis the way Sanders does it (as in using his techniques one would find that which schools you attend at the K-12 level only explain 10% of achievement gaps, while using HLM that figure it at least 40% because you are more accurately measuring school-level effects).  So it is at least plausible that his analysis would be quite different that way. 

It sounds like the people on this board defending Sanders are assuming that those us critiquing him are doing so just because we don't like his findings.  I cannot speak for John Galt, but I don't think it is true for me.  While I do support affirmative action, I would never claim that it is perfect (few policies work the way that we might like them to).  If his claims about performance are true though (meaning they are reproducible from year to year and would stand up to using other statistical methods to measure the effects in a different way), I really would say that we need to really look at why that is.  Especially in light of his finding that the achievement gap on the bar and in school is MUCH lower in historically black law schools, that's really important and interesting.  His claims about what would happen without affirmative action are really just speculation because you cannot make such a sweeping change of an institution and expect everything else to stay the same.  As John Galt points to, look at the UC system.  I take issue with the analysis on a methods level as well as the logic of his argument being weak and not convincing.  But I will look at the critiques and see if any of them point out the flaws better than either Galt or I have.

[SCgrad]

I really don't see how this study, if you want to call it that, is helpful.  Even if some correlation does exist, the study can't make any reasonable claim as to what would happen if the rules were changed, there are too many factors, anything said is speculation.  The bottom line is if a student goes to a top school, does poorly and fails the bar, he or she only hurts himself or herself.  Also, there are a number of possible reasons for any student doing poorly at a school, inability is only one of them.

[King80s]

i disagree with the conclusion of the author, however I also believe JohnGalt's points are weak and almost preposterous.  If you're willing to go along with his argument that those in the bottom 10% of class ranking are likely passing the bar at the highest rate, then good luck

HA! Weak and almost preposterous? I would expect sharper critism than this inlight of all our LSAT training. Galt didn't make any conclusions; he simply pointed out the flaws of the argument.

Good job Galt! Here's to critical reading and analysis.

uhh.. not sure what lsat training you're referring to, but it's basic logic and statistics..  It isn't a flaw if someone doesn't establish a causal correlation between two correlations.  Causality is rarely even an issue in human behaviors.  The author never implied the two correlations were connected, but used them as multiple lines of support.  Hopefully you've had some statistical training outside of 'lsat study'... but it's rather amusing this is the basis of your argument / personal attack.

Since JG's argument is essentially a straw man, as are the other points, there's no need to seriously consider them.

King, you do understand that Sanders argues that there would be an 8 percent increase in the number of black lawyers graduating and he argues more of them will pass the bar. NONE of the evidence he produces even remotely supports such a claim. He just pulled that out of his arse. More likely, if black kids that go to Georgetown, or Cornell could only get into a lower ranked schools, they'd probably wouldn't apply to go to law school anyway. He assumes that just because the percentage of black students in law school has stayed constant for a number of years that it would stay constant even after affirmative action was abolished. One need only look to the UC system after proposition 209 and to schools in Texas after affirmative action was banned there to see that Sanders argument fails the most basic test of plausibility.

you realize that Sanders is explicitly acknowledging his findings are less relevant for those in higher ranked schools such as Georgetown and Cornell?    It's a reasonable measure since the confidence at the upper end of the bell curve is much less than the centroid mass.  I must be naive in thinking out of 130+ law schools, that are certainly good enough for non-black candidates, somehow those outside of the Georgetown / Cornell ranking would cause most of these AA recipients to not pursue law school?  (however, i'm willing to accept the argument for much lower ranked schools)  The point to emphasize is these numbers are geared toward group and not individual behavior, hence the issue of pointing out an error in believing the two correlations must be linked causaly.  Oh, and BTW, he uses index numbers and not soley the LSAT so your beef w/ the lsat, while true mischaracterizes his method.. see pg 36-39, also he notes his findings are less relevant at the top schools, which you're focusing on.

BTW: pg 60-77, the discussion of LS GPA, directly challenges the assertion that he's pulling this out of his arse


Next, if you notice in his study, he uses the ABA study of bar test takers broken down by race in a number of instances, but to prove this dramatic gap in passage rates, he uses only testakers in the California bar and then generalizes from there. One would have to ask, why did he not use the ABA data that he so heavily relies on throughout his analysis to help substantiate this most important point to his argument.

this is flat out wrong... see page 80


Next, he doesn't use class rankings to establish that a certain set of kids fail the bar. He uses two things: 1) general correlations between class rankings and bar passage rates; and 2) general correlations between LSAT score and class rank. The problem with this whole line of reasoning is that he never establishes that these correlations relate to the group he is studying. As has been proven many times, general correlations don't necessarily carry over to being consistent when applied to groups. In fact, specifically, there are many studies to show that the predictive value of the LSAT is diminished greatly when predicting the performance of African Americans. By assuming that these correlations apply to a specific group as well, he fails to eliminate the alternative I was talking about. I never argued that the alternative was true or plausible.

again, he's showing group behavior, and multiple correlations in an attempt to find independent variables.. He is not stating these variables are linked by necessity.  You might be the first to argue that while LS and GPA are positively correlated for most groups, that somehow they have no correlation or negative correlation in his study.. but again this is tangential to his argument.



Sanders never establishes how failing the bar the first time hurts African Americans other than it may hurt their job prospects. This claim is dubious at best. First, most firms hire students before they take the bar and allow for the student to take it multiple times. A person's career is not in jeopardy after failing the bar once. Next, if all these african americans began to go to lower ranked schools, it is not clear how their job prospects would get any better. In fact, most of the data suggests that their job prospects would get even worse.


Another straw man, and not necessary.  See pg 88.  If comparing a 45% graduation rate for said group compared to a 78% for the control, not graduating certainly hurts one's job prospects!





King, even the most die hard Sanders supporters must acknowledge the flaws presented. He presents a compelling analysis, but not a persuasive one. For a person so quick to boast about basic stats, it is a wonder you haven't criticized Sanders for his equivocation and shady mathematical analysis.


Yes, there are flaws, but not the ones you presented.  Your points are not 'necessary' ones .  Sander's analysis is not fair to any AA individual since it treats the 'group blindly' mostly concerned with those around the first and second moments, however, the same inequity also factors in granting the AA boost in which he is arguing.  His major findings are sound and indicate a significant factor, however his flaw is asserting his remedy is the ONLY remedy.

 

[navyseal69]

I really don't see how this study, if you want to call it that, is helpful.  Even if some correlation does exist, the study can't make any reasonable claim as to what would happen if the rules were changed, there are too many factors, anything said is speculation.  The bottom line is if a student goes to a top school, does poorly and fails the bar, he or she only hurts himself or herself.  Also, there are a number of possible reasons for any student doing poorly at a school, inability is only one of them.

To address each of your three points:
1) It is helpful in the sense that it provides some empirical evidence concerning the performance of African American individuals who are unnaturally elevated to higher ranked law schools through the process of AA. Meaning- How do they perform once they are there? Would you prefer we allow AA and just not check its academic effects?

2) It is dismissive of you to say that if an African American goes to a law school beyond his ability and fails, then it is his own problem. AA is undertaken by American society, in large part, to advance a generation of successful African Americans and create a diverse classroom. If these individuals are failing or performing in mediocrity at alarming rates, it affects the entire social program and diminishes the cause.

3) When you say there are a number of possible reasons for a law student performing poorly beyond just a measure of ability, what do you really mean? Of course there are students that face personal or health issues, but what issues are present to blacks and not whites that would account for such gargantuan gaps in performance and bar passage rates?