SCU, like other law schools, would like to have a high bar pass rate.
Suppose you are given the job of using statistics to achieve that goal.
You have available figures on all graduating students for the past ten
years, showing their entering LSAT's and GPA's, their courses,
grades while at law school, and whether they passed the bar on their
How might you use that information to improve the school's bar pass
rate? What problems might your solution encounter?
It is useful to know how good a job different law schools do of
training their students—both from the standpoint of the ABA and AALS,
which would like to know which law schools they should approve of, and
from the standpoint of students who would like to know which school to
go to. The performance of a law school's graduates, whether judged by
bar passage, employment, or salary, is not a very good measure, since
the students differ coming into the law school as well as coming out.
If the entering students at Harvard are much smarter, or much harder
working, than the entering students at Podunk Law, they are likely to
do better after graduation even if Harvard teaches them little or
How might you use statistics to deal with this problem—to evaluate law
schools by value added? Better yet, how could you do so in a way that
would help each student decide which law school is right for him?