Yeah, I agree. The study is really large (5,000 kids!), so there’s bound to be conflicting variables.