Metaresearch

Metaresearch, or meta-research, is the study of research methods, reporting, reproducibility, and the overall efficiency of the scientific enterprise. It aims to understand how studies are designed, analyzed, and published, and to identify practices that improve reliability and reduce waste. Metaresearch methods include systematic reviews of methodological quality, meta-epidemiological analyses that compare treatment effects across studies, simulations, re-analyses of existing data, and experiments that test how different research practices affect results.

Key topics are replication and reproducibility, statistical power and p-values, data and code sharing, preregistration and registered reports, publication bias and selective reporting, and research waste such as unnecessary redundancy. The field often analyzes large bodies of literature to estimate the prevalence of bias and the robustness of findings, using tools like p-curve or z-curve diagnostics and multi-study replications.

Applications include policy changes in journals and funders, such as adopting registered reports, encouraging data sharing, preregistration, and improved reporting standards. Notable organizations include the Center for Open Science and various consortia conducting replication projects. Early influential work includes John Ioannidis's critique of research reliability and high-profile replication projects in psychology and other fields.

Critics note that metaresearch findings can vary by discipline and methodological choices, and that implementing changes can be challenging. Nonetheless, metaresearch seeks to increase transparency, credibility, and efficiency in scientific research.