Klaas Sijtsma, L. Andries van der Ark
We discuss reliability definitions from the perspectives of classical test theory, factor analysis, and generalizability theory. For each method, we discuss the rationale, the estimation of reliability, and the goodness of fit of the model that defines the reliability coefficient to the data. Similarities and differences in the three approaches are highlighted. Finally, we provide a computational example using generated data to illustrate the differences among the different reliability methods.