Scientific reproducibility

Scientific reproducibility is the ability to obtain consistent results when experiments are repeated under specified conditions. It depends on clear methods, reliable measurements, and transparent reporting.

Repeatability: The closeness of results from repeated measurements under the same conditions, operator, and equipment. It describes consistency when the original setup remains unchanged.
Experimental design: The planned arrangement of treatments, controls, measurements, and analysis in an experiment. Design choices determine which conditions a repeat must preserve.
Replication crisis: A widespread concern that findings in several research fields do not consistently recur in follow-up studies. It brought reproducibility problems into public and scientific debate.
Registered report: A journal article format in which a study's question and methods are peer-reviewed before results are collected. It reduces incentives to publish only results that support a hypothesis.
Many Labs 2: A large collaborative project that tested the replicability of 28 findings across laboratories and cultures. Its results illustrate how context and study design shape replication outcomes.
Reproducibility: The closeness of results obtained under changed conditions, such as different laboratories, operators, or equipment. It names the broader cross-condition standard often distinguished from repeatability.
Standard operating procedure: A documented set of instructions for consistently carrying out a routine process. Procedures reduce variation in how experimental steps are performed.
Publication bias: The tendency for studies with positive or striking results to be published more often than null results. Selective publication hides failed repetitions and distorts the evidence base.
Data sharing: The practice of making research data available for access and reuse by others. Shared data allow independent checks of published analyses.
Registered Replication Report: A format for coordinated replications in which protocols receive peer review before data collection. It tests whether rigorous advance planning changes the reliability of replication evidence.

Linked from 15 pages