Researchers don't yet know how to explain these correlations -- and since the study was not a controlled trial, it's not clear whether marijuana or some other factor in marijuana users' lifestyles actually accounted for the beneficial effects.

CNN: Marijuana: The next diabetes drug?

The study itself contains a study designed to look for genes that might explain why some people respond and others don't.

FORBES: Race-Based Medicine, Genomics And You

The next is that this variation has little to do with things you might suppose would explain it: class sizes, hours of study per subject, and spending per pupil.

ECONOMIST: Education and the wealth of nations