Nonreproducibility

Nonreproducibility is the inability to reproduce findings of a study when others attempt to repeat the work. In common usage, the term is connected to two related ideas: reproducibility, where others can obtain the same results using the original data and analysis, and replicability, where independent researchers obtain the same conclusions using new data or methods. Because terminology varies by field, nonreproducibility is often described as part of broader concerns about irreproducible research.

A range of factors can produce nonreproducible results. Insufficient reporting of experimental procedures, data, and code makes exact repetition impossible. Small sample sizes and low statistical power can yield unstable estimates. Analytical flexibility, p-hacking, and hypotheses formed after results are known (HARKing) can inflate false positives. Publication bias toward positive results, errors, and differences in materials or conditions across laboratories also contribute. Data quality issues and inadequate data management can further hinder replication.

Mitigation efforts include preregistration of study protocols, sharing of raw data and analysis code, and adherence to reporting guidelines. Registered reports, larger samples, and standardized methods improve reliability. Independent replication and meta-analyses help assess robustness across studies. The study of nonreproducibility has informed reforms in many disciplines and underpins ongoing movements toward transparency and open science.