Please read my sibling comment. The appropriate response to this study is:
The people who designed the study used absolute garbage for inputs. They should be heavily mentored or fired for ever proposing it, much less for releasing misleading information to a gullible public.
This has nothing to do with preconceived notions. This has everything to doing credible statistical work.
While I would not say smoking/chewing tobacco has any health benefits, the statistics they bandy about for smoking are skewed [i]heavily[/i] to weight the desired results, too. Sorry to be the unpopular guy and say this... bring on the pitchforks & someone get a rope! When a death certificate has space allocated to 'Death caused by tobacco: Probably' a sensible person should know where the numbers will trend regardless of the data received. Again, not defending tobacco, just saying that when you lie and manipulate data to make a valid point more valid(?), the results are tainted. The billions spent on synthesizing tobacco is also suspect, but yes, inhaling vegetable matter can't be good for anything.
IE: A man smoked in Vietnam for 1.5 year tour. Never smokes again for 42 years. Man develops neuroendocrine cancer and dies after stage 4. Tumors found in lungs(and everywhere else), death certificate's official cause of death? 'Probably' tobacco. One more stat for the witch hunt, one less for the Agent Orange cover-up/compensation. Yay statistics!
are you saying that no one should ever do observational studies with self reported data? do you think that this is the first study like this? do you think that scientists do not understand the limitations of these techniques? do you honestly think that NO conclusions can be drawn from observational studies?
how are you supposed to get funding for (expensive) studies with high statistical power without first conducting the studies with low statistical power?
One day of self-reported data extrapolated over 18 years? Knowing what we know about psychology of self-reporting and the changing of diets for people over time, you think that this study means anything?
Is it any surprise that it shows very odd, contradicting results based upon age groups?
do you honestly think that NO conclusions can be drawn from observational studies?
Well, I didn't say that - but if the study referenced is the best they could do then they might as well give up.
Haven't you noticed the meta analysis of many studies recently that show that most of the original studies were flat-out wrong? Do you think maybe it might be because of weak studies like the one in this article?
The people who designed the study used absolute garbage for inputs. They should be heavily mentored or fired for ever proposing it, much less for releasing misleading information to a gullible public.
This has nothing to do with preconceived notions. This has everything to doing credible statistical work.