Re:Mounting criticism of NIH's relationship with the alcohol...
by jcranmer on Tuesday April 03, 2018 @05:08PM
(#16749252)
You don't need ethical flags to be cynical about the state of general "X makes me healthy" research.
Most studies, even the squeaky clean ones (in terms of possible conflicts of interest), tend to suffer from persistent methodological flaws. P-hacking and small N are depressingly common. Actually establishing rigor in controlling only the independent variable is difficult (e.g., alcohol-is-good-for-you tends to be control poorly for the preexisting conditions of teetotalers versus non-teetotalers). There's also the "standard" biases of publication bias and the like.
Most studies, even the squeaky clean ones (in terms of possible conflicts of interest), tend to suffer from persistent methodological flaws. P-hacking and small N are depressingly common. Actually establishing rigor in controlling only the independent variable is difficult (e.g., alcohol-is-good-for-you tends to be control poorly for the preexisting conditions of teetotalers versus non-teetotalers). There's also the "standard" biases of publication bias and the like.