Quote:
Quote by: ByaKya  MOST studies seem awfully biased? Back that up |
Well, I was afraid to use the word "most," but I must say that that's the impression. Once a few studies are unreliable, people are led to believe that the whole field is unreliable. I'm not saying that it is, though.
Then again, the contradicting conclusions of some scientific studies kindles this impression. Of course, there are some reasonable explanations for these, and scientific methods are often used, but often more bias than is necessary is included.
Here's an example of confounding . If we look at the increased number of drownings that occur at the beach during summer and the increased number of ice cream sales, then we see a correlation. Some unfortunate fellow might also see a causation. However, I think it's pretty clear that ice cream does not cause drowning. There was confounding variable, and the increased drownings were really caused by increased number of people swimming since there are more people at the beach. However, someone intent on forming a certain conclusion will ignore confounding variables that have no negligible effect and employ too much bias in making a conclusion. Their hypothesis test would be flawed. This is often what occurs in studies, such as those saying that chocolate is good for health. Of course some would naturally say that because they like chocolate; maybe it
is good for health, but their studies often don't prove it adequately.
Quote:
Quote by: ByaKya Probably screwed up?
They are a bit more educated they do this for a living, so im going to go with the "they probably didnt screw up" theory if you dont mind |
I didn't see this part of the message the first time I read it. Nonetheless, see above. I think that just because someone is a statistician, it doesn't mean he/she is good or is strictly interested in maintaining the integrity of the practice. Scientists aren't
always looking for the truth. Just look at that Korean guy and his pseudo-research on stem cells. They might have other agendas, unfortunately.