Uses and misuses of meta‐analysis in plant ecology

The number of published meta‐analyses in plant ecology has increased greatly over the last two decades. Meta‐analysis has made a significant contribution to the field, allowing review of evidence for various ecological hypotheses and theories, estimation of effects of major environmental drivers (climate change, habitat fragmentation, invasive species, air pollution), assessment of management and conservation strategies, and comparison of effects across different temporal and spatial scales, taxa and ecosystems, as well as research gap identification. We identified 322 meta‐analyses published in the field of plant ecology between 1996 and 2013 in 95 different journals and assessed their methodological and reporting quality according to standard criteria. Despite significant recent developments in the methodology of meta‐analysis, the quality of published meta‐analyses was uneven and showed little improvement over time. We found many cases of imprecise and inaccurate usage of the term ‘meta‐analysis’ in plant ecology, particularly confusion between meta‐analysis and vote counting and incorrect application of statistical techniques designed for primary studies to meta‐analytical data, without recognition of the violation of statistical assumptions of the analyses. Methodological issues for meta‐analyses in plant ecology include incomplete reporting of search strategy used to retrieve primary studies, failure to test for possible publication bias and to conduct sensitivity analysis to test the robustness of the results, as well as lack of availability of the data set used for the analyses. The use of meta‐analysis is particularly common in community ecology, ecophysiology and ecosystem ecology, but meta‐analyses in ecophysiology are more likely not to meet standard quality criteria than papers in other subdisciplines. Fewer meta‐analyses have been conducted in plant population ecology. Synthesis. Over the past two decades, plant ecologists have embraced meta‐analysis as a statistical tool to combine results across studies, and much has been learned as a result. However, as the popularity and usage of meta‐analysis in the field of plant ecology has grown, establishment of quality standards, as has been done in other disciplines, becomes increasingly important. In order to improve the quality of future meta‐analyses in plant ecology, we suggest adoption of a checklist of quality criteria for meta‐analysis for use by research synthesists, peer reviewers and journal editors.