The Sorting Function: Evidence from Law School

Journal of Legal Education

Volumn 51, Issue 4, 2001, Pages 533-552

  Ippolito, Richard A ^a  

^a GEORGE MASON UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0035746025     PISSN: 00222208     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (25)

1. Market Signaling (Cambridge, Mass., 1974).
2. I designed the course around a microeconomics textbook, Robert Pindyck & Daniel Rubinfeld, Microeconomics, 4th ed. (Upper Saddle River, 1998). I taught the first semester without using mathematics as such, but I relied on extensive use of diagrams. In addition to the customary topics (indifference curves, cost curves, equilibrium concepts, surplus measures), I emphasized applications (price supports and ceilings, farm programs, taxes and subsidies) and those aspects of price theory that pertain to law topics, including property rights, the Coase Theorem, externalities, imperfect information, warranties, reputation, and Becker-Stigler deferred-wage contracts. See Gary S. Becker & George J. Stigler, Law Enforcement, Malfeasance, and Compensation of Enforcers, 3 J. Legal Stud. 1 (1974).
3. I designed the course around a microeconomics textbook, Robert Pindyck & Daniel Rubinfeld, Microeconomics, 4th ed. (Upper Saddle River, 1998). I taught the first semester without using mathematics as such, but I relied on extensive use of diagrams. In addition to the customary topics (indifference curves, cost curves, equilibrium concepts, surplus measures), I emphasized applications (price supports and ceilings, farm programs, taxes and subsidies) and those aspects of price theory that pertain to law topics, including property rights, the Coase Theorem, externalities, imperfect information, warranties, reputation, and Becker-Stigler deferred-wage contracts. See Gary S. Becker & George J. Stigler, Law Enforcement, Malfeasance, and Compensation of Enforcers, 3 J. Legal Stud. 1 (1974).
4. Of 233 students, 221 completed the survey form. In a few instances, there were some missing data in the forms completed. In the empirical work, I included dummy variables denoting missing information, though I do not report the coefficients on these variables. This procedure retains the other information in the survey but keeps these observations from influencing the coefficient on the appropriate variable.
5. I knew the students' LSAT score, seat placement in class, gender, and day/evening classification. The study was done with the permission of the law school on the condition that names, once data were collated and entered, were eliminated from the database, and that the individual data remain confidential.
6. The 160-point difference in SAT scores and the 10-point difference in LSAT score each represent two standard deviations for these scores respectively. See note 12.
7. Candidates with the highest subjective discount rates are unlikely to go to law school in the first place, but among those in the pool, presumably, there is some dispersion of discount rates over the relevant range. Classic references implicating discount rates in education are Gary S. Becker, Human Capital: A Theoretical and Empirical Analysis (New York, 1964) and Victor R. Fuchs, Time Preference and Health: An Exploratory Study, in Economic Aspects of Health, ed. Victor R. Fuchs (Chicago, 1982). See also Theodore Schultz, The Value of the Ability to Deal with Disequilibria, 13 J. Econ. Literature 827 (1975). I also show the relevance of discount rates for productivity across workers in Pension Plans and Employee Performance (Chicago, 1998).
8. Candidates with the highest subjective discount rates are unlikely to go to law school in the first place, but among those in the pool, presumably, there is some dispersion of discount rates over the relevant range. Classic references implicating discount rates in education are Gary S. Becker, Human Capital: A Theoretical and Empirical Analysis (New York, 1964) and Victor R. Fuchs, Time Preference and Health: An Exploratory Study, in Economic Aspects of Health, ed. Victor R. Fuchs (Chicago, 1982). See also Theodore Schultz, The Value of the Ability to Deal with Disequilibria, 13 J. Econ. Literature 827 (1975). I also show the relevance of discount rates for productivity across workers in Pension Plans and Employee Performance (Chicago, 1998).
9. Candidates with the highest subjective discount rates are unlikely to go to law school in the first place, but among those in the pool, presumably, there is some dispersion of discount rates over the relevant range. Classic references implicating discount rates in education are Gary S. Becker, Human Capital: A Theoretical and Empirical Analysis (New York, 1964) and Victor R. Fuchs, Time Preference and Health: An Exploratory Study, in Economic Aspects of Health, ed. Victor R. Fuchs (Chicago, 1982). See also Theodore Schultz, The Value of the Ability to Deal with Disequilibria, 13 J. Econ. Literature 827 (1975). I also show the relevance of discount rates for productivity across workers in Pension Plans and Employee Performance (Chicago, 1998).
10. For example, David Laband and Bernard Lentz show that part of the difference in salaries of lawyers is explained by differences in the hours worked. See Is There Sex Discrimination in the Legal Profession? Further Evidence on Tangible and Intangible Margins, 28 J. Hum. Resources 230 (1993).
11. A Study of Wages and Reliability, 39 J.L. & Econ. 149 (1996).
12. In this context Hasnas was looking for evidence that the student had read the assigned materials and had made an effort to disentangle the issues. Students who could not demonstrate that they had met this standard were given a mark of unsatisfactory and a reduction in grade point. Subsequently Hasnas offered the failed students the opportunity to recoup their lost points by agreeing to be called on several times during the remainder of the semester. If they were ready and prepared each time, they could erase their negative entry. Only three students earned the negative grade; all agreed to the unlimited exposure option to future questions; and all redeemed themselves during the process. The redemption possibility was not announced beforehand, and this was Hasnas's first year teaching at George Mason.
13. Cases and Materials on Torts (St. Paul, 1994).
14. Hasnas used Prosser, Wade and Schwartz's Torts: Cases and Materials, 10th ed. (Westbury, 2000).
15. Fall 1999 was the first time either Hasnas or Grady taught the course at George Mason. Entering students had to base their choice on the teachers' general reputation, the fact that Grady was a senior professor and dean, and information posted by Hasnas and Grady at their Web sites that provided fairly extensive information about their courses.
16. I ran the regression for day students only because Grady did not teach in the evening. Setting the dependent variable as "Grady," and using the same independent variables as in column 3, Table 1, the coefficient on SAT verbal is -.0017 (t-value=2.17). Thus, for each 100 points in SAT verbal score, the probability of choosing Grady fell by 17 percentage points. The mean of the dependent variable is .45. By contrast, the coefficient on LSAT score was .009, which is not statistically different from zero (t-value =.63).
17. While the difference in performance for either group is not statistically different from zero, the difference between these groups (.49) is statistically significant at the 95 percent level of confidence.
18. As an experiment, I started with the regression reported in column 2 of Table 1 and recorded the coefficient (and t-value) on the variable denoting whether the student was an economics major; these were 2.18 (t=3.92). I repeated the regression four times, each time adding one variable in the following order: LSAT score, SAT score, college math, and the character attribute variables (absent that day, row selection, and the Grady and Hasnas signals). The resulting coefficients on the variable denoting economics major were as follows: 1.49 (t=2.88), 1.28 (t=2.43), 1.13 (t=1.87) and .41 (t=1.40). It is well known, for example, that economics majors score relatively high on the LSAT. See Michael Nieswiadomy, LSAT Scores of Economics Majors, 29 J. Econ. Educ. 377 (1998).
19. Alternatively, perhaps some students become so embarrassed when speaking in public that the disutility of doing it is not sufficiently remunerated by a one-third bump in grade; or perhaps some students are extraordinarily risk averse and attach a large weight to the possibility of receiving the negative grade penalty.
20. For example, see Robert Wood et al., Pay Differences Among the Highly Paid: The Male-Female Earnings Gap in Lawyers' Salaries, 11 J. Lab. Econ. 417 (1993), which shows that even after holding constant childcare responsibilities, female lawyers earn less than their male colleagues.
21. The course covers game theory, finance, statistics, and decision-making under uncertainty. It requires students to make many calculations involving present values of tort claims, probability trees, and so forth.
22. This number is 5 fewer than the number of students who took his course (see Table 4); these five left the law school after the first semester.
23. That is, the coefficient on the absenteeism variable is -.06. Thus, three absences imply an effect equal to -.18. The coefficient on the sit-in-back variable is -.20. The total of these effects is -.38.
24. I used average SAT scores for the Princeton Guide to the Best 331 Colleges (2000 edition). For GMU law students who attended a college not on this list, I classified them with a "not rated" dummy variable. For those from rated colleges, I used their average SAT total score as a measure of the underlying quality of undergraduate institution.
25. It is worth noting that I reestimated the regressions in the first two columns of Table 3 to see if college average SAT scores mattered in the law school grades and the Economics I grades, and I found no significant effects. Stacy Dale and Alan Kreuger (Estimating the Payoff to Attending a More Selective College (Princeton, 1999)) show that one's own SAT performance is all that matters for later success, not SAT performance of the college of origin. My results do not necessarily conflict with this finding, because I do not use college mean SAT scores to measure quality, so much as to measure the relative importance of mathematics versus nonquantitative emphasis in the curriculum.