The Truth Wears Off

[Damian]

Many rigorously proved scientific results start shrinking in later studies. What went wrong?
Long but fascinating read about the Decline Effect in scientific research.

http://www.newyorker.com/reporting/2010/12/13/101213fa_fact_lehrer

As an SEO I would like to know more about how medicial scientists conclude something is good or bad for humans for example, from testing just a few hundred or thousand people when there are so many unknown variables for every test person that could skew the results. If I would understand better I think I could learn some reverse engineering methods I could use on Google's Algo.. essentially the same problem IMHO.

[JamesR]

from testing just a few hundred or thousand people when there are so many unknown variables for every test person that could skew the results.

Excellent point.  I think alot of the realm of science is done this way - conclusions are found in the realm of probability rather than certainty.