Not all are good.
But it's always a good idea to question your assumptions. Statistics neither lie nor tell the truth. They are interpreted.
Words are the same way.
"Net."
Did I just refer to something you fish with? Or did I type the Russian word for "no"?
You can't tell, can you? They look exactly the same.
There's just one problem.
Je n'ecris pas chacun d'eux. J'ai ecris le mot francais pour "clean."
I didn't type either. I typed the French word for 'clean'. But you probably ruled out the "right" answer because of assumptions that you carried into the thread, modified by the false choice I put in front of you.
It's the same with stats, esp. in the era of data dredging and unpublished results. So some bilingual psych studies were very informative. They showed that to be bilingual was to have greater mental focus, IQ, etc. Except that only the studies that showed that were published; the other studies that showed no results weren't. The result? The published studies lost all statistical significance. There's no "there" there because the statistics didn't lie but the researchers did. The assumption was that the data were random. They weren't. They were chosen. So the statistical analysis didn't apply. (Actually it's more likely that the psychologists involved didn't lie; they were just over their heads when it came to the statistical methods they applied mechanically. They failed to question their assumptions.)
But seriously--the more "advocate" the research is likely to be, the more socially relevant or "important" in some economic or political or social way, the more likely it is to be non-reproducible or to be rejected if the results don't match expectations.
= new reply since forum marked as read