Paired data arise when an experimental or observational study design deliberately samples two units as a dyad to form a basis of comparison. Use this tag for any question about paired data or their analysis.

Paired data arise when an experimental or observational study design deliberately samples two units as a dyad to form a basis of comparison. Typically these units are highly correlated and thus warrant the use of statistical methods for correlated data. Examples of paired designs include twin studies, case-control studies with 1:1 matching of cases to controls, split-plot experiments, and pre-post tests. Examples of conditional likelihood based statistics include a paired t-test, conditional logistic regression, and McNemar's test.

history | excerpt history