I've been slow to post a third entry this week because some of the comments on my first entry made me think more carefully about the table I posted Monday (entitled "White Michigan Grades at Private Firms with 100+ Lawyers) and the conclusions I drew from it. I went back to the data and offer below a refined analysis.
There are two significant weaknesses in Monday's table. First, I included all Michigan alumni in creating the deciles, but of course many alumni did not return surveys. These non-respondents are part of the "base" on which the rates are calculated. So the author of the first comment, who noted that the proportion of Michigan grades joining firms seemed way to low, had a good point. (Among respondents, about 37% started at firms with 100+ lawyers and 47% started at firms with 50+ lawyers. The numbers would probably be higher today because firms are bigger.) Second, the time periods covered by the first two columns and the third column are significantly different; the first column looks at graduates from 1977 to 1990, while the third (15-year column) looks at grads from 1967 to 1985. Even though I noted the data were not literal attrition rates, many readers took them to be that.
After some thought about these limitations, I concluded that a better way to look at the association between GPA and success at big firms was to look directly at survival rates for big firm associates. The Michigan alumni surveys ask respondents how long they stayed at their first job after graduation, and how many jobs they've had. It's thus possible to see how many grads who started at big firms stayed at the same firm for the next fifteen years (and, in nearly all cases, became a partner at the firm).
The table below summarizes such an analysis. Those tracked here were respondents to the 15-year alumni survey from 1987 to 2000 who reported that their first job after law school was at a firm with 50 or more attorneys; these are divided into grade deciles, and tracked by whether they are still at the same firm. (I used a 50-attorney threshold because many of these attorneys started their careers in the 1970s, when a 50-attorney firm was considered "big".)
|(a) Grade Decile||(b) White respondents to 15-year survey who started at firms with 50+ lawyers||(c) Those in (b) who were still with original firm 15 years later||(d) Those in (b) no longer at original firm||% Still at Original Firm (c/b)|
To illustrate how to read this table, look at the first row of data. Of 1192 Michigan 15-year alumni who provided information on their first job and who started their career at a big firm, 25 (2%) had grades that put them in the lowest decile of all Michigan students. Of those 25, only 1 was at the same firm 15 years later - a 4% "survival" rate (reported in the last column, as (c/b).
The last column thus gives us an idea of how grades affect tenure among those at big firms. The relationship is quite strong -- the correlation of grade decile and the "survival" rate is more than .90.
With this added context, let me make some observations responsive to the posted comments and other emails I've received:
1) I don't contend that high law school grades themselves have a direct causal effect on success in big firms. I contend that high grades measure various things that are related to later success. Those things might conceivably include actual learning in various law classes; more likely, they include facility with legal analysis, cognitive skills and writing ability, and also such things as ambition, drive, lack of a social life, etc.
2) Obviously another factor that affects "survival rates" to partnership are choices about career, family, and lifestyle. Those choices are certainly very different for women and men, and they might vary across racial lines, too. But in The Racial Paradox, I showed that during the early years of associate experience (the years for which the After the JD project supplies good data), white men and white women had essentially identical experiences, while those of blacks and Hispanics diverged dramatically. So for women, attrition due to personal choices or to "glass ceilings" presumably sets in later.
3) Several of those posting comments suggested that minority patterns of attrition must be far higher than those explained by these grade variations. The Michigan data can't really answer this question, because the samples of nonwhites in big firms are very small (a total of 43 nonwhites in the 15-year alumni dataset reported starting at big firms). The numbers we do have for UMLS nonwhites are roughly consistent with those for whites:
|(a) Grade decile range||(b) Nonwhite respondents to 15-year survey starting at firms with 50+ lawyers||Proportion of (b) at same firm 15 years later||Comparable rate for whites|
|7 or higher||4||25%||25%|
Again, the sample sizes are far too low to draw conclusions from this second table; the close match of percentages is largely happenstance. And of course the dynamics facing nonwhites are very different. As I observed before, the nonwhites with low grades face two extra obstacles: first, they are likely to be at more elite and challenging big firms than are the whites with low grades; second, they are likely to be "stereotyped" as second-rate even when they are not. On the other hand, minorities may benefit from some affirmative action at the promotion and partnership level -- though clearly, if this happens, it is much less than the use of racial preferences in the hiring of associates.
As this second table suggests, nonwhite hires at big firms have far lower GPAs than white hires; nearly two-thirds of the nonwhites are in the lowest two deciles, compared to 5% of the whites.
The story told in these short blog entries is a brief summary of a more complicated story laid out in The Racial Paradox. I've tried here to grapple with the responses of those skeptical of the data. My hope (perhaps a forlorn one) is that critics will go beyond mere skepticism to try and develop an alternative explanation consistent with all of the available data.