A Harvard study finds exercise is associated with absence of obesity in white women, not in black women. This sounds like a result I would have wanted to suppress, if it were my study. See a Reuters article for details.
I'll bet I know why they included it though - in order to have significant findings. If they rolled all participants into one batch they probably didn't have much in the way of findings, so they segmented the subjects by race and, viola, they had something to report. In the bad old days, when I needed to get things published, I was guilty of this kind of data mining.
Later - Another thing concerns me about this finding - implied causality, that is, that exercise prevents obesity. How about the alternate interpretation that obesity prevents exercise? That makes sense to me.
Doing exercise while obese is like doing exercise wearing a backpack weighing 50-100 lbs., much more tiring and less exhilarating. The alternate conclusion is of less utility to the folks who want us all to exercise.
