[arstin]

This is such a complicated statement to respond to! :) You could follow the thread of irony, of the kinds of abilities "analysis" is equivocal between, of what relevance evidence might have...what it might be applied against, what might even count as evidence, and, obviously, what results we even want!

Here's just one bigger picture thought to consider. For a reasonably motivated and bright person, it's pretty easy to teach yourself programming (I'd be willing to bet most people on this forum are self-taught). And, after you get some basics, pretty easy to teach yourself nice tidy applied math-y things like Bayesian reasoning. Likewise it's easy to teach yourself science. One reason why is that in every case you can self-correct: the program doesn't work, the calculation is wrong, the world says otherwise.

That simply isn't the case with the humanities. You need the guidance of an expert for a while.

I'll leave it at that for now, just noting that to the extent humanities help with "analysis" it's probably going to be especially beneficial with messy, open problems where even the criteria for success may be vague and shifting. Thinking critically about product design rather than improving an algorithm, to bring it into HN.