Originally Posted by
blackbishop351
Thanks to a fascinating course a couple years ago (that was actually listed in the Philosophy department), I'll point out that data can be "statistically significant" even though they're drawn from a small sample size. Lack of a realistic control group can even be accounted for.
On the other hand, it seems to me that statistical significance must be a term that was invented specifically for what I call pseudo-science - sociology, psychology, marketing, whatever. It accounts for a lot of subjective meaning as well - for example, studies have found that smoking increases the user's chance of contracting lung cancer by 0.2%, which is then called "significant" simply because of the subjective gravity of the disease.
Anyway. Enough of the rant. I personally try not to believe ANY reported statistics whenever possible...at least without a LOT of corroboration.