chapter  2
26 Pages

: Generalizability theory: Concepts and principles

The principal aim of a generalizability study, or G study, is to evaluate the characteristics of a given measurement procedure and to estimate measurement precision. To this end, the different sources of measurement error in play must rst be identied, so that their relative importance as error contributors might be quantied. Relative importance is indicated by relative sizes of estimated variance components, typically estimated using classical ANOVA. It is these estimated components that are used to calculate measurement errors and G coefcients. It is these same components that are used in optimization studies, or “what if?” analyses, to predict how errors and coefcients might change in response to future changes in study design.