because if you taught someone relativity and quantum mechanics without explaining the experimental data and the different ways that were tried of reconciling them that failed, they wouldn't believe you.

I like this, I'd never thought of it that way. Griffiths also justifies taking a somewhat historical approach by claiming it as a hack for the student's brain -- that our minds are built to process stories, to process narratives, and so by introducing each particle one at a time, through the events that led to its discovery, he can better fix the identity of that particle in the student's mind.