law.school.admissions: Law School Admission Council data

In fairml: Fair Models in Machine Learning

Law School Admission Council data

Description

Survey among students attending law school in the U.S. in 1991.

Usage

data(law.school.admissions)

Format

The data contains 20800 observations and the following variables:

• age, a continuous variable containing the student's age in years;

• decile1, a continuous variable containing the student's decile in the school given his grades in Year 1;

• decile3, a continuous variable containing the student's decile in the school given his grades in Year 3;

• fam_inc, a continuous variable containing student's family income bracket (from 1 to 5);

• lsat, a continuous variable containing the student's LSAT score;

• ugpa, a continuous variable containing the student's undergraduate GPA;

• gender, a factor with levels "female" and "male";

• race1, a factor with levels "asian", "black", "hisp", "other" and "white";

• cluster, a factor with levels "1", "2", "3", "4", "5" and "6" encoding the tiers of law school prestige;

• fulltime, a factor with levels "FALSE" and "TRUE", whether the student will work full-time or part-time;

• bar, a factor with levels "FALSE" and "TRUE", whether the student passed the bar exam on the first try.

Note

The data set has been pre-processed as in Komiyama et al. (2018), with the following exceptions:

• DOB_yr, the year of birth, has been dropped because it is (nearly) perfectly collinear with age, and thus it is redundant;

• decile1b has been dropped because it is (nearly) perfectly collinear with decile1, and thus it is redundant.

In that paper, ugpa is the response variable, age and race1 are the sensitive attributes and the remaining variables are used as predictors.

References

Sander RH (2004). "A Systemic Analysis of Affirmative Action in American Law Schools". Stanford Law Review, 57:367–483.

Examples

data(law.school.admissions)

# short-hand variable names.
ll = law.school.admissions
r = ll[, "ugpa"]
s = ll[, c("age", "race1")]
p = ll[, setdiff(names(ll), c("ugpa", "age", "race1"))]

m = nclm(response = r, sensitive = s, predictors = p, unfairness = 0.05)
summary(m)

m = frrm(response = r, sensitive = s, predictors = p, unfairness = 0.05)
summary(m)

fairml documentation built on May 31, 2023, 6:02 p.m.