There seems to be a little confusion here. This thread is about published studies, not science.
Science is simply an evidence-based approach to solving real world problems. The closest emulator of the real world is *gasp* controlled trials, not anecdote. That's why we live in the 21st century and can live to be 80 years old. Because a logical mode of thought was adopted that allowed the advent of modern medicine via controlled trials.
Studies, on the other hand, can be quite dubious in nature. One must always look at the big picture. I've pointed this out quite a few times with, say, Indus Biotech. People want to get published, and to get published, you generally want positive results. Any company with this in mind has a stats guy that can properly manipulate (not illegally of course) the statistical data to display significance. It happens all the time, and it's why you should look at meta-analyses, structure/function relationships, and *gasp* yes anecdote (case reports)!