The Intra-class Correlation Coefficient (ICC) for metric variables quantifies the absolute reliability or consistency among raters by estimating the proportion of total variance attributable to between-rater agreement within a specific groupings structure. Formally defined as a ratio derived from analysis of variance, it adheres to strict assumptions regarding normality, sphericity, and fixed versus random rater models to distinguish true signal from measurement error in continuous data domains. This metric operates strictly within the subfield of psychometrics and quantitative methodlogy, serving as a fundamental rule for establishing consensus validity prior to advancing reliability analysis techniques for categorical or nominal variables with multiple evaluators.
Prerequisite chain context: requires Cohen's Kappa for Two Raters in Statistics.