In his Stanford article, Sander breaks the nation's 180 or so law schools into six tiers, based roughly on their selectivity. Then he holds constant other factors that might affect a student's chances of passing the bar: undergraduate grades, LSAT scores, race, gender—and law-school grades. Oops. According to Ho, holding constant law-school grades to assess the effect a law-school tier has on passing the bar is like holding constant lung health to assess the effect smoking has on death. That throws off the results because lung health is one of the pathways by which smoking causes death.