One cannot identify causation based upon these scores. Students self-identify majors, sometimes more than one, or sometimes none at all; others self-select into taking the LSAT altogether (opting for medical school, business school, or a lucrative career instead of law school). Therefore, it is emphatically not necessarily the case, based on this data, that these majors cause students to perform better or worse on the LSAT. It simply describes them.
The chart above identifies the mean highest LSAT score and mean undergraduate GPA based on self-identified major, for majors with at least 80 students taking the exam, among all law school applicants.
A forthcoming chart will have the same information, but solely for law school matriculants; that is, for people who actually matriculated to law school this year. Stay tuned!
UPDATE: I had some glitches in the first post, so I removed the matriculant data and will save it for another post. The matriculant data is now available here.