There are many definitions of what a scientist is but that of a person that addresses questions using the scientific method is a good one. One of the tenets of science is that science autocorrects and that, through peer-review, only the most rigurously tested truths emerge to become part of the main body of science.
I finally managed to find the time to read this article from The Atlantic. The article describes how Dr. Ioannidis and his colleagues have found out that a lot of scientific papers describe research that is flawed in one respect or the other. Apparently, in medical sciences, as much as 90% of the published information that doctors rely on is flawed. That is indeed troubling and highlights, as the article points out that any resarch proving a hypothesis is a vastly more likely to be published than one that does not. Does this mean that medical resarch is based on the shaky grounds of thousands of papers with false results? It would seem so from the article:
Ioannidis was putting his contentions to the test not against run-of-the-mill research, or even merely well-accepted research, but against the absolute tip of the research pyramid. Of the 49 articles, 45 claimed to have uncovered effective interventions. Thirty-four of these claims had been retested, and 14 of these, or 41 percent, had been convincingly shown to be wrong or significantly exaggerated. If between a third and a half of the most acclaimed research in medicine was proving untrustworthy, the scope and impact of the problem were undeniable.
You can read this statement in a number of ways but what it tells me is that the most influencial research is usually not taken at face value but further examined and tested. Which is why we know that those high impact papers are not completely right, because our most important results are tested time and time again. Which is why we can trust them.
Even better, with the raise of journals like PLOS ONE it is increasingly easier for researchers to publish negative results. This is certainly an area where we could do with some improvement. With the right incentives we should know about flawed research and flaws in research a lot faster so that scientists do not spend time replicating dubious work.