A survey of publication bias within evolutionary ecology

摘要

Publication bias has been recognized as a problem in ecology and evolution that can undermine reviews of research results. Unfortunately, direct tests of publication bias are extremely rare. Here, we quantify a well-discussed but, to our knowledge, previously untested form of publication bias: the publication of results with and without estimates of effect size. We find that results published without effect sizes are a biased sample of those that are published. This further complicates the already difficult task of compiling quantitative literature reviews and meta-analytic studies.