Assessing Transparency and Reproducibility in Invasion Science

Policymakers and practitioners overseeing invasive species management depend on reliable research for guidance. Transparency and reproducibility are core features of reliable research, and prerequisites for successful study replication, but are evidently lacking in many science disciplines. Whether this shortfall characterizes invasion science remains unknown. We evaluated a sample of invasion science studies for their adherence to practices that enhance transparency and reproducibility, such as making data and code available, and explicitly considering statistical power. Our evaluation focused on published studies concerning two plant species invasive to riparian ecosystems in British Columbia, Canada (Elaeagnus angustifolia and Phalaris arundinacea), as these contributed to a broader systematic mapping initiative. Our systematic literature search yielded 746 studies, of which 45 met our predefined inclusion criteria (relevance rate = 6%). We assessed each study against a 14-item checklist (motivated by the Transparency and Openness Promotions guidelines) and a corresponding scoring rubric. On average, studies achieved a score of 26%, with no studies addressing statistical power, pre-registering their plans, and few making data or code publicly available. There is a clear need and opportunity for improving the transparency and reproducibility of invasion science research. We refer researchers to resources aimed at improving research practices, and discuss two practices that are especially important in the context of applied invasion science: power analysis and sharing data and code. We echo recent calls for educational and research institutions to expand access to training in open science, and urge policymakers and practitioners to consider transparency and reproducibility when seeking guidance from invasion science research.