Evidence-appraisal glossary
Multiplicity
Multiplicity is the problem that arises when a study runs many statistical tests at once. Each test carries its own chance of a false positive, so testing many outcomes, subgroups, or time points inflates the overall probability that at least one "significant" result is a fluke.
Also called: Multiple comparisons, Multiple testing, Multiplicity problem.
What it is
When a study asks one question with one test, the false-positive rate is whatever alpha was set (often 5%). But real trials often test several endpoints, compare multiple treatment arms, split patients into subgroups, or look at the data repeatedly over time. Every extra test is another roll of the dice, so the chance that some comparison crosses the significance threshold by luck alone climbs well above 5%. This inflation of the family-wise error rate is called multiplicity.
How to use it when reading a study
Count how many comparisons were actually run, not just the ones highlighted. Ask whether the primary endpoint was pre-specified and whether the authors applied a correction (for example Bonferroni, Holm, or a hierarchical testing plan) to protect the overall error rate. Treat a lone "significant" subgroup, secondary outcome, or interim look with skepticism if many were examined and none were adjusted. Findings from unplanned, unadjusted analyses are hypothesis-generating, not confirmatory.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.