ABSTRACT
Developing selection and classification systems, or evaluating their effects, requires some identification of the goals that are to be served by the system. From this perspective, one major overall goal seems to be to make selection and classification decisions in such a way that aggregate individual performance is improved, maximized, or optimized. Making this goal operational requires the use of one or more specific measures of performance. By design or by default, the choice of specific criterion measures for performance assessment is guided by a model or theory of what performance is. Consequently, the objectives of this paper are to:
Provide a label and a brief description for each of a set of alternative ways one could think about, or model, job performance for purposes of selection/classification criterion measurement. The set of alternative models to be described is meant to include all possible alternatives. Sometimes a particular point of view is made explicit by the proponents for it, and sometimes it must be inferred from what they actually do. However, nothing major should have been left out. If something seems to be missing, it is because of either the author’s faulty knowledge or his faulty descriptions, for which there will most likely be penalties.
Outline the circumstances under which a particular model might be more appropriate than the others, particularly with regard to the distinctions among differing subgoals for selection and classification.
Outline the principal measurement implications of the different models, particularly for selection and classification research. Although this paper is not about the pros and cons of specific measurement methods, a major reason for having a model, or theory, of performance is to guide the search for, or the construction of, observable measures.
