Methodology

First off, there are three more schools in this study than in Brian Leiter's 2003 study. My study came about because I was first interested in comparing some other law schools that weren't in the 2003 study. This piqued my interest to use Leiter's methods to duplicate his results for all the schools in his study. The list of schools in my study is not meant to include every good law school, and just as Leiter admitted that there might be other schools besides the 22 he studied that performed just as well on the criteria used, there might be other schools besides the 25 I included that also perform well.

I used the Martindale-Hubbell database for my search.

I took extra care to sort out other schools with names that shared a word with a school in this study such as Chicago-Kent or Michigan State . I believe my work in this regard was quite accurate. The name variants I searched for can be seen in the name variants list. I don't think I missed any, but e-mail me if I did.

McGuireWoods appeared to be the only firm where entering their name in the search engine on Martindale-Hubbell produced results from other firms that had to be weeded out. After adding LLP to the end of the name, the extraneous results disappeared, which is why it's the only firm with LLP after its name in the data.

For the O'Melveny & Myers data I went to www.omm.com and copied the results from the search for each school into an excel file in order to easily and accurately count the number of lawyers from each school.

Numbers for firms Leiter uses as examples (such as Cravath, Swain & Moore) will differ from those in my table because, as Leiter himself says, those firms underreport attorneys to Martindale-Hubbell so the numbers that come from their website will differ from the numbers that come from Martindale-Hubbell. It's important to note that for all the law firms searched on Martindale-Hubbell, it might be that the number of attorneys from each firm is underreported compared to what one would find searching the firm's website if that site allows a search by law school.

Some of the firms in the Martindale-Hubbell database (such as Mayer Brown) list some of the names of their lawyers twice. Because I have no reason to believe that these double listings favor any particular law school(s), I have included these extra listings in my results because that was much faster than sorting through them. Because of the extra listings the total number of lawyers for all the law schools might be inflated, although they still might not be inflated because, as Brian Leiter points out, the database may give fewer attorneys than could be found through the law firms' own websites.

Following Leiter, I combined the LL.M. and J.D. graduates to calculate the per capita standings. The numbers of J.D. graduates came from www.lsac.org and the numbers of LL.M. graduates came from each school's website. The LL.M. numbers often took some searching to find and might not be up to date, so feel free to send me the newest figures for your school. I e-mailed the two schools (Emory and UCLA) that I do not have any LL.M. numbers for and until I hear back from them I cannot calculate their per capita score or their overall normalized score.

Following Leiter, for the normalized score I weighted the per capita score as 40%, the number of firms with at least one graduate as 40%, and the number of firms with at least 5 graduates as 20%. For each variable, the top score was taken as 100% and all the scores below it were expressed as a percent of the top score. It should be obvious that using different weights should give different results, and I may eventually add that kind of analysis.

For my opinions about the methodology and results, please read the discussion after viewing the results.

Back to main page - Data - Results - Discussion