The publication of research reports was an essential part of the Enlightenment. The first two scientific journals were the British Philosophical Transactions of the Royal Society and the French Journal des Sçavans, both of which started their activities in 1665. Today, 359 years later, the number of scientific journals exceeds 30,000.
The primary purpose of publishing a report in a scientific journal is to document that a specific observation has been made or an experiment performed. This communication of hypotheses, methodologies, and reviews of what is considered known or suspected supports the advancement of science and facilitates the exchange of ideas and collaboration. For individual researchers, publishing in reputable scientific journals provides recognition of their work. It contributes to their professional reputation and helps establish them as authorities in their field. In contrast, undocumented research is rapidly forgotten, and scarce resources have been wasted.
However, in empirical sciences, most, if not all, findings are uncertain due to sampling variation and because of unconditional reliance on possibly unfulfilled assumptions. Focusing just on the findings and neglecting their uncertainty is potentially misleading and makes the findings scientifically meaningless. Showing empirical support for one’s conclusions is as important as the conclusions themselves, which implies that statistical evaluation and presentation are critical. Many journals rely on statistical reviewers to check the methodological quality of submitted manuscripts and suggest improvements in terminology, analysis, results presentation, discussion and conclusions. Unfortunately, not all manuscripts can be improved. Ronald Fisher described this problem: “To consult the statistician after an experiment is finished is often merely to ask him to conduct a post-mortem examination. He can perhaps say what the experiment died of.”
