The 538 website, noted for its analytic approach to popular topics such as sports and politics, takes on an issue very near and dear to the hearts of many folks here at BF. In this article, 538's Benjamin Morris suggests that just because the average 3-point shot is more efficient than the average long 2, that doesn't mean teams should leave long 2s out of their arsenal.
As Randy Wittman suggested when asked about this very topic, Morris writes that the defense often dictates what a good shot is, and an open long 2 is a better shot than a contested 3.
It'd be interesting to see even deeper analysis on how closely contested a shot has to be to start swaying the open-contested equation. I'm pretty sure the camera system installed at NBA arenas can deliver data on how close defenders are to each shot, so we should be able to figure out how open is open enough. Brad Beal takes a lot of long 2s off of picking action that allows him to get his shot off without being blocked, but with a defender close enough to bother him at least a little.