A checklist for choosing between R packages in ecology and evolution

Abstract

The open source and free programming language R is a phenomenal mechanism to address a multiplicity of challenges in ecology and evolution. It is also a complex ecosystem because of the diversity of solutions available to the analyst.Packages for R enhance and specialize the capacity to explore both niche data/experiments and more common needs. However, the paradox of choice or how we select between many seemingly similar options can be overwhelming and lead to different potential outcomes.There is extensive choice in ecology and evolution between packages for both fundamental statistics and for more specialized domain-level analyses.Here, we provide a checklist to inform these decisions based on the principles of resilience, need, and integration with scientific workflows for evidence.It is important to explore choices in any analytical coding environment-not just R-for solutions to challenges in ecology and evolution, and document this process because it advances reproducible science, promotes a deeper understand of the scientific evidence, and ensures that the outcomes are correct, representative, and robust.

Keywords: R programming language; checklist; guidelines; heuristic; open source; paradox of choice; reproducible science; statistical methods; tools.

Figure 1

The number of R packages for common statistical and ecological/evolutionary concepts. The estimates for each term were generated using the R package “packagefinder” to search CRAN directly from the R console. The minimum number of packages returned was 3, and the maximum was 2,876 effective July 2019