Generalizability Theory: Estimation

Abstract Generalizability theory is a conceptual and statistical framework for the analysis and construction of measurement instruments. Among the most important concepts of generalizability discussed in the first section of this entry are the universe of admissible observations, universe and observed scores, random and fixed facets, crossed and nested designs, variance components, generalizability studies, decision studies, and generalizability coefficients. A discussion of the results of a generalizability study and a decision study of a crossed one‐facet random effects design and of a two‐facet crossed one‐facet random effects design is presented in the two following sections. A number of other designs that illustrate the versatility of generalizability theory are presented in the final section.