Replication in field ecology: Identifying challenges and proposing solutions

1. Field ecology has been included in a 'replication crisis' that extends across many scientific disciplines. However, the underlying concepts of replication, reproducibility and replicability are not always clearly distinguished, and complicate the identification of best practices. Furthermore, conducting experiments under the high variability of natural field conditions reduces the capacity for replication relative to other biological disciplines working under controlled conditions. Field ecologists are therefore facing a significant challenge in assessing the replicability of their research with implications for overall confidence in study outcomes. 2. Through a review of the literature, we discuss several related aspects of experimental design that can enhance confidence in scientific outcomes. Specifically, we describe sample replication (repeat sample), within-study replication (repeat experiment) and between-study replication (repeat study) and how each can be used within field ecology. Since perfect between-study replication (i.e. direct replication) is generally not possible in field ecology, we suggest more explicit use of conceptual replication would enhance confidence in scientific outcomes. However, such changes require cultural shifts in practice among all participants in the scientific enterprise. 3. We suggest several tangible steps could be taken to improve confidence in ecological research: (a) increase the use of within-study replication before publication, (b) increase replicability for aspects that we can control (e.g. pre-register experiments, open data, publish code), (c) divest from novelty as the primary criterion for publication in leading ecological journals and invest in experimental design, (d) be sceptical of contradictory findings from studies testing similar research questions and (e) create a publishing environment that encourages more conceptual replication studies. 4. We believe adopting these practices will increase the confidence in results for field ecology. There are critical obstacles that could prevent some scientists from increasing within-study or between-study replication, including short-term funding mechanisms and the prospect of fewer publications. We suggest strategies to mitigate negative impacts to researchers, such as leading journals piloting new article categories and explicit mention of experimentally linked studies. We acknowledge that adopting greater replication in field ecology will require significant changes to cultural practices, but there are clear benefits for improving our confidence in science.