That was my immediate reaction too, but then I started reading it.
Their study is rather interesting, even if you aren't persuaded. (It's much wider ranging than the ABA encapsulation suggests, and contains a lot of good notes and references if nothing else.) Among other things, it reviews findings on the role of social economic status, family legal/professional background and religious background in selecting law schools and feeds all of that into a discussion of shifts in the industry. (At parts, it reads like the academic world decided to churn out a TLS flame.)
I'm going to drop some big chunks of it into the thread here, to spare people who just want the good bits from trawling through the PDF.
Here's a bit from their discussion of big law changes:
Forty years later, the process – and the elite firms themselves – looked very different.
Elite firms had grown enormously (by a factor of ten or more), and thus had to do hiring on a
much larger scale than before. The sharp increase in legal demand that accompanied this
growth meant that much big firm business was coming from completely new sources, where
social connections were less relevant. The rise of elite Jewish firms in the 1960s and 1970s had
brought greater pressure to bear upon older firms to emphasize intellectual virtuosity over
social connections. And the increased geographic spread of the elite firms meant that big firm practice
was far less concentrated in New York than had once been the case. All of these
factors led firms to both broaden the range of law schools from which they hired, and to place
greater emphasis on the grades of applicants. It is well known among those involved in law
firm hiring – though not discussed much in public – that the great majority of big firms have
adopted sliding scales that set rule‐of‐thumb grade thresholds at each school the firm selects
from. Thus, an elite Chicago law firm might consider the pool of possible hires to include the
top half of the class at the University of Chicago, the top third of the class at Northwestern, the
top fifth at the University of Illinois, the top tenth at Loyola Law School, and so on.
One can see the shift in patterns even over the shorter period embraced by the two
“Chicago Lawyer” studies by Heinz, Laumann and their collaborators. Here, we have created
two categories of private law firms called “big” firms and “huge” firms based on their size
relative to other firms at the time of each survey. In 1975, “big” firms had 60 or more attorneys
(putting them at the 75th percentile of respondents age 40 and under in a firm setting), and
“huge” firms had 150 or more attorneys (the 90th percentile of respondents age 40 and under in
a firm setting). In 1994, “big” firms at the 75th percentile had 220 or more attorneys and “huge”
firms at the 90th percentile had 500 or more. The changes over time are striking: In 1975, 92%
of the younger lawyers at huge firms had attended a prestigious law school, and by 1994 that
figure dropped to 52%.
OK....here's the wind-up:
The table suggests that the relative supply of top talent at the most elite schools (i.e.,
students with very high credentials) is greater than the relative supply of big firm partners from
those schools. If anything, moderate‐credential students who go to lower‐ranked schools seem
to be landing a disproportionate share of partnerhips. As we shall see, there is much evidence
that doing very well at a non‐elite school confers particularly high value on law students. Thus,
while it is true that Columbia students as a group are more likely to be recruited by elite firms
than Fordham students, it is no longer obvious – by any means – that a given student will be
more likely to eventually land a big firm job if she attends Columbia than if she attends
Fordham. A critical part of the calculus, and one consistently overlooked by law school
applicants, is whether she performs well in law school.
and the pitch:
whenever it is possible to link law school grade data to either short‐ or long‐term
outcomes among lawyers, the evidence is indisputable that grades matter. The After the JD
study reveals a sharp tradeoff between law school grades and tier. For all but the most elite
schools, the salary premiums for achieving high grades more than make up for the salary
depreciation associated with attending a lower‐ranked law school.
The AJD dataset covers lawyers at the beginnings of their careers and can say little
about the long‐term importance of law school performance, but the impact of law school
performance has a strong and long‐lasting reach. Consider, for example, Table 8, which is based
upon the University of Michigan’s surveys of law school alumni. The surveys gather data from
thousands of alumni five and fifteen years after graduation, and links alumni responses to
actual data on the alumni’s years of attendance, credentials, and law school GPA....
[The data show] two important phenomena. First, it is obvious that getting higher
grades at Michigan was associated with a higher chance of working at a large law firm. While it
is conceivable that some self‐selection is at work, it is much more likely that most of this
association is due to the law firms’ use of grades as a screening device in hiring (the 10th decile
alumni were less likely to start at big firms in large part because many of them first did judicial
clerkships). Second, GPA is strongly predictive of one’s survival at the big firm. Alumni with
high grades were several times more likely to still be at the same firm they started at – i.e., to
make partnership at that firm – than alumni with low grades. One should keep in mind that the
relationship must be even stronger than this chart suggests, because those with low GPAs in
this group were those selected by the big firms despite their low GPAs; that is, they were seen
as more promising lawyers by the firms despite their low grades, and thus presumably had a
higher chance of survival than would typical Michigan students with low grades. The
implication of this chart is that the big firms – which supposedly obsess over hiring associates
with high grades – were actually placing too little weight on grades for this pool of Michigan
Put differently, these patterns among the Michigan alumni provide very strong evidence
that law school grades are much more than a “credential”; they powerfully predict long‐term
success. Indeed, part of what makes this particular example probative is that, since our
measure of success is whether big firm associates stayed with and were promoted at their
original firm, we know that their grades as credentials had no relevance to a successful
outcome; it is highly unlikely that the firm promotion committees ever again looked at the
transcripts of these associates, once they had been hired out of law school.
Here's where they go with it, examining the interrelation of these factors and getting at the idea that transfers do worse (finally.)
To help disentangle the interaction between grades and school eliteness, we used data
from the Bar Passage Study (BPS), the longitudinal study of students starting law school in 1991
that LSAC conducted in the 1990s. In the questionnaire administered to participants as they
began law school, each person was asked a series of questions about their application process,
including whether they were admitted to their “first choice” law school and, if they were,
whether they enrolled in that school. About one respondent in eight was admitted to their first
choice school but did not attend it, often for geographic or financial reasons. These students
provide an excellent group for avoiding the selection effects that are so omnipresent in work on
school eliteness; since the students who go to their second‐ or third‐choice school were
admitted to their first‐choice (and generally more elite) school, they probably have the same
(unobserved) strengths as their peers who also got into, but actually attended, their first‐choice
schools. By using OLS regression to predict standardized law school grades in the BPS, we obtain
consistent estimates that attending one’s first‐choice school is associated with a drop in law
school GPA of about one‐fifth of a standard deviation. We also find that first‐choice students
attend schools that are, on average, about one‐third of a “tier” more elite than their secondchoice
peers who are otherwise similar in their credentials. Our preliminary inference from
these findings is that moving up a tier in the BPS hierarchy is associated with a GPA drop of
about three‐fifths of a standard deviation (about two tenths of a point in a typical law school
4.0 scale). The detriment to GPA was even greater for law students in the top two law school
tiers in the BPS (roughly equivalent to schools ranked 1‐50). For these students, attending a
first‐choice school resulted in a highly statistically significant loss of .39 standard deviations in
law school grades, suggesting that trading up an entire tier could result in a half point drop in
As suggested above by bk1
, whether or not you accept this prediction is key.
What we really want to know, however, is how these factors
directly compare to one another in shaping career outcomes....
The [OLS regression] results are rich and telling. In 1975, law school eliteness is
associated with higher incomes, though the relationship is not quite statistically significant for
even the top tier of law schools. The importance of self‐reported class rank is roughly
comparable to law school eliteness in its importance; even back then the scale of the
coefficients on class rank variables was greater than the tier variables. One key measure of
social eliteness – having a father in the legal profession – has a significant and large impact
upon income in 1975.
By 1994 the picture has changed substantially. Self‐reported class rank is the dominant
explanatory variable, with a 27‐percentile difference between those reporting that they are at
top versus the bottom of their class. The coefficients on law school eliteness are
strengthened as well, but less so. The range of predictive power associated with class rank is
more than double that of law school tier. Having a parent who is a lawyer no longer has any
impact on earnings, and neither do the other measures of social status.
Of course, as we noted earlier, self‐reported class rank is about the worst possible
measure of law school grades. The predictive power of law school grades is certainly even
higher than the 1994 regression suggests, given the large amount of noise in this database’s
measure of grades. Moreover, since we are here considering long‐term career outcomes – not
short‐term recruitment – the effect of law school performance is not simply a credentialing
effect of high grades leading to attractive job offers. Something about doing well in law school
is strongly associated with lasting career success, and proves to have more efficacy than law
school eliteness. The eliteness of one’s law school is, compared with grades, a relatively weak
explanatory factor in the 1994 equation. And while the grade coefficients are biased in a way
that understates their actual influence, the law school coefficients are almost surely overstated.
The Chicago Lawyers equations do not include measures of pre‐law credentials, such as LSAT
scores and undergraduate grades. Since these factors do predict income for broad crosssections
of lawyers, and since the tight hierarchy of law school admissions makes law school
eliteness a close proxy for student credentials, an unknown but probably large part of what
seems to be explained by school eliteness is actually just a measure of pre‐law credentials.
On the dataset:
“After the JD” (“AJD”) is an ongoing project sponsored by leading legal institutions to
create a panel database on the early careers of lawyers. The AJD surveyed in 2002‐03 a national
sample of several thousand lawyers who were in the 2nd to the 4th year of their legal careers.42
The sample is broadly representative of the national lawyer population, though minorities are
oversampled and lawyers in major legal markets (e.g., New York, Washington) are also
overrepresented. It includes a very rich set of background information that was aimed at
providing a detailed picture of the “social eliteness” law graduates brought to the job market.
The Law School Admissions Council (LSAC) also provided data on the relative LSAT scores and
undergraduate grades of all participants. Respondents also provided categorical information on
their grades and class rank. This is a particularly rich dataset, then, for examining the very first
stages of contemporary legal careers.
See the PDF for more in depth discussion of the models used to analyze the data, which adds factors (school rank, geography, grades, SES, religious background, etc) and observe changes in correlation in outcomes. Some of the notes there on particular markets etc might be interesting to some readers.
Here's the conclusion:
The first‐wave AJD shows no predictive power for a wide range of SES factors in
explaining early career income. But we have also seen two important qualifications.
First, having a religious affiliation is associated with higher income in both the AJD and Chicago data.
Since Catholics, Protestants, and Jews all appear to have some edge over nonbelievers and the
unaffiliated, these preliminary results suggest that it is the process of being part of an organized
religion itself that has a causal influence, rather than any sort of religious affiliation preference among employers. This is an interesting issue for further exploration. Second and more importantly, SES and social eliteness still play a critical role in shaping the aspirations and credentials of young people before they enter law school. The level of social privilege associated with legal education has changed remarkably little over the past half‐century. What has changed – and apparently largely disappeared – is any ongoing role of that privilege in shaping careers after law school.
The quantitative evidence also suggests that the importance of law school eliteness is
exaggerated in most discussions about legal markets. Law firms which once hired exclusively
from a narrow set of elite firms now hire associates from dozens of different law schools. The
AJD regressions show that the earnings premium of elite graduates diminishes markedly once
proper controls are included for cognitive skills and regional cost‐of‐living differences. And the
Chicago Lawyer regressions indicate that the earnings boost of an elite degree is smaller in a
cross‐section of all lawyers than it is in an analysis of new lawyers, suggesting that one’s elite
degree is a depreciating asset as one’s career evolves.
Law school performance, in contrast, is an extraordinarily powerful predictor of career
outcomes, and one that has clearly become more important over recent decades. Even when
the only measure available is a highly imperfect indicator of law school grades, that measure
dominates all others in predicting early career incomes, cross‐sectional career incomes, or
survival to partnership in big firms.
They then go on to philosophize a little about What It All Means.
That seems to be the meat of it, w/o reproducing the table and graphs. There's also an interesting footnote discussing why AJD self-reported grades appear to be much more predictive and reliable than self-reported class rank.