This forum is about wrong numbers in science, politics and the media. It respects good science and good English.
After reading SWN, Epi, Junkscience Judo, and mixed in my experience from the Nuclear industry, computing and engineering, I have come to the opinion that ANY epidemiological study outside the bounds of epidemic level problems (you know the kind where suddenly a lot of people get sick with similar symptoms from an unknown source) are pretty much worthless.
I watched a woman last week fret over Fire Retardant in children's sleepware because a study showed a 1.8 RR for autism in children who were exposed to Flame Retardant. She thought that the 80% increase meant that it was almost a 1 to 1 relationship. That if you were exposed to the retardant then you would get autism. I haven't been able to find the study to find out what the actual numbers were, but a 1 to 1 correspondence she describes would result in a relative risk of infinite (I perhaps exaggerate, but it would be greater than the RR of lung cancer and smoking of 24).
I have explained this subject to this individual before. She is an intelligent Master's Degreed (in Speech Therapy) person. The scare factor got her though.
I am not overly worried about fire retardant in clothing. Back in the days of open flame lamps and candles in every room, this was much more important. Now it is not nearly as useful as strapping a seatbelt onto your lap.
There are probably some very ethical Epidemiologists out there. The ones that get into the media. The studies that make news are all examples of epidemiology gone wrong. Epidemiologists have attained Lawyer status. They now seem to do more harm than good.
I should acknowledge that they (epidemiologists) are providing me entertainment. I get to be morally outraged at them and write in this corner of a the internet about them.
Entertainment is important.
Sandy S. has good insight on this sort of crap(US slang for the defecationary residue from the male of the bovine species) at this site:
Epidemiologists are fun like watching snakes and just as dangerous.
"Epidemiologists are fun like watching snakes and just as dangerous. - Gary K."
Not all of them. Some of the ones you see nowdays touting RRs of 1.05 are, but when I was in research, I knew a couple of REAL epidemiologists who understood the fraility of their studies.
One had me find a better random number generator. His analysis showed that the generator built into Fortran on our PDP11 wasn't truly random, (software generators often aren't,) and he wanted no bias in the way patients were assigned to different treatment schemes.
He also once said that anyone with a mission to cure disease X shouldn't rely on statistics. Emotional bias makes it easy to see a cure in the numbers where there really is nothing there.
But nowdays, anyone with Excel can be an expert. (People like John obviously excepted.)
There are probably thousands of Epis out there who are honest. They are doing very mundane things like analyzing cat food (although apparently that has taken a twist for excitement this month). The pet food scare here in the US was a prime example of when those Epi's come in handy. Problem is, those things don't happen all that often.