I have serious doubts about the statistical controls you can do in the presence of very obvious and large confounders. Especially when the thesis doesn't seem t

I have serious doubts about the statistical controls you can do in the presence of very obvious and large confounders. Especially when the thesis doesn't seem to generalize particularly well (as they admit). Is there something cab drivers do, which is radically different from the numerous other driving occupations? Personally I find these very broad studies extremely uninteresting. Without making any accusations, having large,very broad datasets, make it very easy to find something, when there is just noise. What makes me despise it even more are the headlines it generates.