Sunday, February 03, 2008

Law School Accreditation and Bar Passage Rates

The Volokh Conspiracy Blog has a recent post on and discussion of ABA proposals to make law school accreditation depend on bar passage rates.

As anyone involved with law schools knows, bar passage rates are a hot topic closely connected with the issue of affirmative action. A recent article in the Stanford Law Review by Richard Sanders argued that affirmative action by law schools injures blacks; the student who would have done reasonably well in (say) a third tier law school with other students about as able as he was gets admitted to a first or second tier law school eager to increase its (racial) diversity, and ends up never passing the bar. I discussed the issue in a post some time back. After hearing about the article I asked a colleague likely to be unsympathetic with its conclusion if she knew of a good rebuttal. Having read the article she suggested, whose argument ultimately came down to "Sanders is a racist," I concluded that if that is the best rebuttal available, Sanders is probably right. If anyone knows of a better rebuttal, I would be interested in the cite.

The easiest way for a law school to raise its bar passage rates in order not to be at risk of losing its accreditation is by reducing its willingness to admit black students with low LSATs. Thus the ABA proposals, apparently made in response to pressure from the Department of Education, are likely to be seen as an attack on affirmative action. They are also seen as threatening traditionally black law schools. If Sanders is correct, the first may actually be a good thing; the second is not.

In my view, the fundamental problem with the ABA proposal is that it is measuring the wrong thing. There are two different mistakes one would like law schools to avoid:

1. Admitting students who won't benefit from admission and probably won't pass the bar, in order to raise the racial diversity of their student body.

2. Rejecting students who will benefit from admission but will still have a lower than average chance of passing the bar, given where they are starting from, in order to raise the school's bar passage rate.

The real test of the education provided by a law school should be some measure of value added rather than quality of graduates (as measured by bar passage rates). The fact that a school can get a student who starts at the 99th percentile of the LSAT to pass the bar isn't much evidence that it is worth going to. If the ABA is going to set standards along these lines, they should be based, not on average bar passage rates, but on the relation between bar passage rates and entering LSATs.

Doing that is pretty straightforward. Run a regression for each state with LSAT as the independent variable, bar passage probability as the dependent variable, to find, on average, the probability that a student who starts with a given LSAT will end up passing the bar. Use the results of that regression to calculate what the bar passage rate ought to be for a given class at a given law school if the school is providing an average education. If the actual rate is more than a fixed amount below that, the school doesn't get accredited. That way accreditation provides the information that matters to the student--how much more likely am I to pass the bar if I go to this school or that?

Of course, I also have reservations about other parts of the system--but at least, if they are going to do this part, they ought to do it right.

Accreditation requires a yes/no decision. The approach I have described also makes it possible to provide more detailed information to applicants trying to decide whether to go to law school and, if so, what school to go to. Each school could report bar passage rate for its graduates over the past ten years as a function of entering LSAT, both as an absolute number and relative to the average figure for the state.

What a student wants to know, after all, is not how good a job the school does on average but how good a job it is likely to do for him. If, as Sanders' article implies, a student with a less than stellar LSAT is likely to learn less at Harvard Law School than at some less prestigious school better adapted to teaching students like him, that is information the student ought to have.


15 comments:

Anonymous said...

Are LSAT scores really that closely correlated to bar passage? I have never attended law school, but my understanding is that the LSAT is a test of logic, while bar exams are tests of legal knowledge.

Mike Linksvayer said...

the student who would have done reasonably well in (say) a third tier law school with other students about as able as he was gets admitted to a first or second tier law school eager to increase its (racial) diversity, and ends up never passing the bar.

So passing the bar is much more strongly correlated with class rank than school quality?

Seth said...

Is there any strong reason to base accreditation on this statistic, rather than just publishing it and letting prospective students decide for themselves?

David Friedman said...

Mark asks about correlation between LSAT and bar passage. My understanding is that it's high; there may be figures in the article I linked to in my post.

With regard to Mike's question, I think the point is negative correlation, LSAT held constant, with school quality for people at the lower end of LSAT. Higher quality schools admit, on average, people with higher LSAT's, and you have to separate out the two factors so as not to attribute to the school the results of the characteristics of the incoming students.

With regard to Seth's point, I don't see a whole lot of point to accreditation, as I thought I at least suggested in the post. But if it's going to be done, it ought to be done right.

Raphfrk said...

What are the benefits of accreditation, is it just that the college can say that they are accreditated, or does it effectively close the college down ?

Anonymous said...

David, please help me.

What does this crucial sentence mean?

Having read the article she suggested, whose argument ultimately came down to "Sanders is a racist," I concluded that if that is the best rebuttal available,

Does it mean that name-calling is the only effectiver rebuttal smart people can think of?

Regards,
Bill Drissel

David Friedman said...

Bill asks about my comment on the rebuttal to the Sanders article.

I'm not sufficiently involved to try to search out everything anyone has written pro or con in the controversy. So I asked a colleague whose political views made her likely to disagree with the article if she knew of a good rebuttal. She pointed me at an article which she hadn't read but had heard was a good refutation.

That article, so far as I could tell, contained nothing close to a rebuttal, and its central argument was indeed name calling. I concluded that if there was a really good rebuttal my colleague would probably have heard of it, hence that there probably wasn't.

David Friedman said...

raphfrk asks about why accreditation matters. My understanding is that in many states, you have to have graduated from an accredited school to be allowed to take the bar. In some cases there are alternative qualifications that can be used instead, such as some sort of apprenticeship.

David Friedman said...

Bill asks about my comment on the rebuttal to the Sanders article.

I'm not sufficiently involved to try to search out everything anyone has written pro or con in the controversy. So I asked a colleague whose political views made her likely to disagree with the article if she knew of a good rebuttal. She pointed me at an article which she hadn't read but had heard was a good refutation.

That article, so far as I could tell, contained nothing close to a rebuttal, and its central argument was indeed name calling. I concluded that if there was a really good rebuttal my colleague would probably have heard of it, hence that there probably wasn't.

happyjuggler0 said...

David Friedman,

As near as I can figure, the whole issue of school accreditation comes down to an extra "layer of protection" for the end consumer, above and beyond licensing and getting hired by a firm or persuading clients to hire one in the first place. To that end:

What are your thoughts as a libertarian, and/or a professor of law/economics, regarding mandatory government licensing of lawyers to begin with? How about school accreditation, do we really even need this government interference in our affairs?

Secondly, if mandatory government licensing is for some reason is deemed a good idea (rightly or wrongly), why on earth do we let the ABA (or in the medical profession, the AMA...don't even get me started on more mundane professions like cutting hair) make up the rules? At best they should merely be unofficial advisors to our elected officials and the permanent bureaucracy in my minarchist opinion.

Finally, going back to the whole school area, is there an admissions ceiling of some sort on law schools or number of students? It is my understanding there is one for prospective doctors. Any thoughts on how to agitate effectively on changing this?

Sorry for the multiple questions in one post.

Mike Huben said...

David:

It occurred to me that bar exams might differ from state to state (and indeed they do), so I decided to check the wikipedia article. That's a relatively minor problem with your idea (which overall doesn't strike me as bad.)

What struck me as making your idea impractical was how students actually prepare for the exam. Generally, the law school doesn't prepare them for the exam: "most students in the United States attend a private bar exam review course which is provided by a third-party company and not their law school."

That raises all sorts of other questions on the correlation between the review companies and passage rates, and whether there is more discrimination (economic, most likely) in that private sector. After all, you may have completed law school on a need-based scholarship, but have trouble affording a first rate review course.

David Friedman said...

Mike writes:

Generally, the law school doesn't prepare them for the exam: "most students in the United States attend a private bar exam review course which is provided by a third-party company and not their law school."

That doesn't mean that the law school doesn't prepare them for the exam, just that in addition most take an additional review course.

I don't know if there are data on how the particular review course affects bar passage.

David Friedman said...

Happy Juggler asks my views on accreditation. I, of course, think it ought to be done privately and serve as a source of information to both prospective students and employers, not a legal restriction.

Given that it is done by government and is a legal restriction, it is almost certain to be controlled by the profession for familiar public choice reasons; I'm not sure if it makes much difference whether that is done explicitly, as in the case of the ABA, or implicitly.

Mike Linksvayer said...

I understand that you have to separate school and student quality. I still find the claim implicit in the text I quoted above interesting and surely verifiable.

Another way of stating the claim: a student with a given quality is more likely to pass the bar eventually when the student attends a lesser law school.

pops said...

The idea that schools don't prepare you for the bar is false. My girlfriend currently attends a first tier law school that requires all students to take a mock bar exam. It has no effect on their grade, but students who fail are not-so-subtly nudged into additional tutoring, which is usually offered by the school. The extent of economic discrimination is probably not overwhelming. First, the discrimination would already be controlled for, at least to some extent, by the study as Friedman designed it. Preparation for the LSAT is virtually the same as for the bar, students pay to take private courses. So I don't think the correlation wouldn't be too compromised.

And sure, the classes aren't cheap-- about $1000. But considering that students have received an education valued at $120,000, the additional $1K expense to gain the ability to utilize that education seems paltry, especially since (conjecture) the average earning of unlicensed attorneys are almost certainly much lower than those who pass the bar. For many years now student loan institutions have offered loans specifically for law school grads preparing to take the bar.

As a soon-to-be 1L, I'd agree with Dr. Friedman that the most important statistic to the majority of students is employment info... although there's obvious problems in collecting that info:

http://online.wsj.com/public/article/SB119040786780835602.html