Experimental Research

Abstract

The theoretical work in Chapters 2 through 7 has been undertaken to characterize a class of multi-stage decision situations, that despite their importance have received almost no experimental treatment, and to propose and develop simple response models for decision behavior in such situations. As such, our main purpose has been to stimulate new theoretical and experimental research rather than summarize findings from already published experiments. The three topics treated in the present chapter have been selected to suggest various areas of research in which the formulations or models presented in the preceding chapters may be found useful. Section 8.1 describes a computer-controlled experiment specifically designed to test the models developed in Chapters 2 through 5 in a simulation of quality control. We describe this experiment in detail both because of its substantive results concerning the predictive power of the various response models and the use it makes of the SOC plot introduced in Chapter 6. In the psychophysical experiment described in Section 8.2 both time and the nature of the observations are completely different than in Section 8.1. Yet, this experiment may be modeled as a TDC problem, and its major findings may be accounted for by asymptotic results derived from model FPB. The similarity between detection of change and simple repeated-trial learning situations is exploited in Section 8.3 to provide a criterion method for sequentially evaluating subjects’ performance in learning experiments.