One aspect of the work that really impressed me
>was how painstakingly Chris had documented his experimental methodology,
>and I remarked on that to him. His response was, "Well, what else would
>you expect me to do. It's no good if someone else can't check my results."
> Even when it's published in a refereed journal, a paper can contain errors
>or even be dead wrong. But careful, properly documented experimental
>methodology makes it possible for others to catch your mistakes, and serves
>as an incentive to researchers to be responsible in their methodology and
>their reasoning.
Another related issue is the deliberate failure of papers to give sufficient
information to allow repeating the results. This is done in order to throw
off your tail in areas where research is particularly hot (or where your
data is particularly shoddy). An "accidental" error in a decimal place in
the m&m section of a paper from a competing lab cost me a couple of months
work during my graduate studies. My professor "knew" it was no typo, and
related other such horror stories, which are anecdotal, but very familiar in
the community. Such dishonesty is probably rare, but certainly not unknown.
I haven't forgotten that lesson. This only emphasizes more the critical
importance of m&m sections to scientific papers.
Art
http://chadwicka.swac.edu