It seems like we part ways at the last point in that article, or at least, what you're saying here would seem to be at odds with it. I'm inclined to agree with

It seems like we part ways at the last point in that article, or at least, what you're saying here would seem to be at odds with it. I'm inclined to agree with it, myself. It's true that these are messy problem domains, and that does make it harder to conduct and draw firm conclusions from scientific research. That inhibits progress, but it doesn't render the whole enterprise a lost cause.