chapter  9
24 Pages

Generalizability Theory

In several of the preceding chapters, we made substantial use of classical test theory. These chapters were based on the premise that the set of components in an instrument under consideration was prespecified, i.e., fixed throughout our pertinent discussion. Although one could make decisions as to which components to include in revised versions of instruments being investigated, the sets of possible components for them were considered as given beforehand. Within this framework, sometimes called statistical, sampling is typically carried out only from a population(s) being studied, rather than from a large pool of test components.