ABSTRACT
In analyzing physical systems or examining variable relationships, the researcher is aided by the construction of one or more competing models. Underlying each model may be a particular theory or set of theories. The engineer may have several types of diffusion or compartmental models in mind, whereas the social scientist may consider contrasting correlation models in connection with path analysis. The selection of a best model from the collection of constructed models is often a choice among competing theories and is based on empirical results obtained from sample data. We refer to the theory of selecting among rival models based on sample information as model discrimination.
