In the last couple of posts, I ranked state bar exams by difficulty using aggregate data from ABA-approved law schools on bar passage rates, LSATs, and undergraduate GPAs. The first version used data from 2010-2011 and controlled for LSAT and GPA. The second version used additional data (2008-2011) and controlled only for LSAT. These rankings are approximate and imperfect, especially for smaller states, but provide better interstate comparability than passage rates alone for large states.
In the first post, I commented on changes in Michigan's bar exam in 2012 that resulted in much lower bar passage rates from previous years. Because 2012 data was not included in either ranking, the much greater difficulty of the new Michigan exam was not reflected in the rankings. In this post, I show how the changes in Michigan's bar exam affected passage rates at schools across the LSAT spectrum.
The chart blow shows how bar passage rates at ABA-approved schools varied with LSAT medians in California, New York, and Michigan, using a lowess regression (don't ask). New York was chosen for comparison because it had approximately the same bar passage rate as Michigan over the 2008-2011 period but ranked as a more difficult exam. The solid red line represents Michigan's bar exam from 2008-2011, and the red dots represent the 2012 Michigan bar exam results by school.
The new (2012) Michigan bar exam results show an uncanny similarity to California's bar exam results. Whereas in the 2008-2011 Michigan results schools at the lowest end of the LSAT spectrum had almost the same passage rate as those at the highest end of the spectrum, the passage rate is now lower across the spectrum, but especially at the lower end of the LSAT scale. The results for each school in Michigan's 2012 exam seem to cluster around the California line.
A few caveats are in order. First, these are aggregate data, so it is possible that they are distortetd by selection effects. Second, I am not arguing that California is "the best" or that Michigan's changes to its bar exam are desirable. In fact, I have deep reservations about the ability of the bar exam to assess the qualities necessary to serve clients effectively as a lawyer. I am especially hesitant to suggest that a bar exam should be made more difficult vis-a-vis students who have already incurred significant debt to attend law school. This post is simply descriptive, but perhaps I will have time (later) to say more about what would be normatively desirable.
If you are wondering how Michigan and New York could have about the same passage rate (for 2008-2011) when the New York line is below the Michigan line through almost the whole range, the reason is that the New York LSATs are concentrated near the high end, while the Michigan LSATs are concentrated near the low end. The same reason explains why the 2012 Michigan results cluster around the California line, yet California had a higher bar passage rate. You can see this in the fourth column of the table on this page.
I leave it to readers to assess whether the change was "raising the bar" or "moving the goalposts."