A "Significant" Shift in Statistical Interpretation in Ecology and Evolution: A Helpful Guide for Authors and Reviewers

"Statistical significance", wherein the deviation of some measure of a sample population is greater than an expected threshold (usually p > 0.05) is a concept and practice strongly embedded in the life sciences. However, the use of "statistical significance" has been deprecated by the American Statistical Association (ASA), and that news has been slow to spread across disciplines. The result of this misuse is that many authors are wrongfully claiming "no effect" of a variable when p > 0.05 or confidence intervals (CIs) overlap zero. To ensure scientific rigor in Ecology and Evolution, we offer this brief guide to authors and reviewers and strongly encourage authors to use the ASA statement on p-values. The ASA suggests that authors report effect sizes and attendant error in ways that let users judge support for the hypotheses represented in a model, avoiding the binary "true or false" of p = 0.05 or overlapping CIs. Furthermore, we encourage authors to acquaint themselves with numerous other common mistakes in statistical interpretation highlighted by the ASA and subsequent authors. By incorporating these principles into statistical reporting and results, information and patterns that might otherwise be overlooked remain available for readers to evaluate. We recognize that changes in reporting practices will not occur overnight, so our aim is to help authors adopt ASA principles: to that end we describe potential problems and offer guidelines for authors and reviewers to follow, to avoid incorrect statistical interpretations and spurious conclusions.