Overcoming the pitfalls of categorizing continuous variables in ecology and evolutionary biology

<ol> <li><span>Many metrics in biological research – from body size to life history timing to environmental metrics – are measured continuously (e.g., body size in grams) but analyzed as categories (e.g., large versus small). The pitfalls of categorization are well-recognized in statistics, but many scientists in the fields of ecology, evolution, and behavior may not be aware of this literature. These fields lack a review of common examples and feasible solutions to avoid the hazards of categorizing continuous data. </span></li> <li><span>Our goal was to summarize current practices of categorizing continuous predictors in ecology and evolutionary biology and provide guidance for overcoming those pitfalls. We conducted a mini-review of 72 recent publications in six popular journals to quantify the prevalence of categorization. We then summarized commonly categorized metrics and simulated a dataset to demonstrate the drawbacks of categorization using common metrics and realistic examples from ecology and evolutionary biology. </span></li> <li><span>We show that categorizing continuous variables is common (31% of publications reviewed), especially in the animal behavior field, and underscore that predictor variables – including abiotic, morphological, physiological, behavioral, and demographic metrics – can and should be collected and analyzed continuously. Our analysis of the simulated field dataset demonstrates how categorizing continuous variables can lower statistical power and change interpretation, especially when arbitrary breakpoints are used. Finally, we provide recommendations on how to keep variables continuous throughout the entire scientific process. </span></li> <li><span>Together, these pieces comprise an actionable guide to increasing statistical power and facilitating large synthesis studies by simply leaving continuous variables alone. Overcoming the pitfalls of categorizing continuous variables will allow ecologists and evolutionary biologists to continue making trustworthy conclusions about natural processes, along with predictions about their responses to climate change and other environmental contexts. We hope that this manuscript and its associated code will provide a useful lab practical for students and teachers to develop programming skills including data simulation, plotting, and model comparisons, as well as research skills including reporting and interpretation. </span></li> </ol>