Reading some of the back-and-forth in this thread, it struck me that some of the discussion was about data, some was about models, some was about underlying reality, but none of the discussion was driven by statements that this or that pattern in data was “statistically significant.” Here’s the problem with “statistical significance” as I …