A compensatory model for simultaneously setting cutting scores for selection-placement-mastery decisions (OMD-RR 94-17)

A method is proposed for optimizing cutting scores for a selection-placement-mastery problem simultaneously. A simultaneous approach has two advantages over separate optimization. First, test scores used in previous decisions can be used as "prior data" in later decisions, increasing the efficiency of the decisions. Then, more realistic utility structures can be defined using final success criteria in utility functions for earlier decisions. An important distinction is made between weak and strong decision rules. Weak rules are allowed to be a function of prior test scores. Conditions for optimal rules to be monotone are presented, and it is shown that optimal weak monotone rules are compensatory by nature. Results from an empirical example of instructional decision making illustrate the differences between simultaneous and separate approaches. Subjects were 71 medical students receiving interactive video or computer-aided instruction