Replication

Replication is the repetition of a study or experiment to assess whether its findings recur. It tests the reproducibility of results under specified conditions.

Direct replication: A repeated study that closely follows the original study’s procedures and measures. It tests whether results recur when the original methods are followed as closely as practical.
Reproducibility: The ability to obtain consistent results using the same data, code, and analytical procedures. It is often distinguished from replication, which collects new data to test a finding.
Many Labs 2: A large collaborative project that tested the replicability of 28 psychological effects across laboratories and settings. Its coordinated design illustrates how replication can test robustness across diverse samples and contexts.
Publication bias: The selective publication of studies according to their results or statistical significance. When failed replications remain unpublished, the published record overstates how often findings recur.
Conceptual replication: A test of the same hypothesis using different methods or measures. It asks whether the underlying claim holds beyond the original operational choices.
Statistical power: The probability that a statistical test detects an effect of a specified size when that effect exists. Low power makes genuine effects harder to detect in both original and repeated studies.
Reproducibility Project: Psychology: A 2015 project that attempted to replicate 100 studies published in three psychology journals. Its results became a prominent empirical assessment of replication in psychology.
P-hacking: The use of analytical choices or repeated testing to obtain statistically significant results. Replication efforts exposed how flexible analyses can produce findings that do not recur.
Registered replication report: A format in which a replication protocol is peer reviewed before data collection and results are published regardless of outcome. It reduces publication bias by committing to report the replication before its outcome is known.
Effect size: A quantitative measure of the magnitude of a phenomenon or difference. Comparing effect sizes reveals whether a repeated result matches the original beyond mere significance.

Linked from 20 pages