As an author, editor and methodological reviewer, I experience from time to time that only one methodological approach is considered correct: For example, reviewers say that confounders should only be selected on the basis of significant group differences (!), or statistical procedures are considered incorrect per se – probably out of unfamiliarity. In statistics and epidemiology, however, there is not just one right way to “answer” a question.

Many research questions can be investigated in very different and creative ways. Conversely, this does not mean that all research approaches are valid: Flawed study designs and analyses lead to wrong conclusions. This makes it all the more important to plan studies thoroughly and to openly discuss the study protocol in an interdisciplinary team.