Decision making under conflict: Decision time as a measure of conflict strength

Everyday experience suggests that decision making is often accompanied by conflict. For example, both farreaching decisions—for instance, whether to have chemotherapy or surgery as a treatment for cancer—and trifling decisions, of whether to go to a concert or an opera, can generate conflict. A conflict is resolved by making a choice between alternatives. Kurt Lewin was among the first to bring together the concepts of decision making and psychological conflict (Lewin, 1951, 1931/1964). Depending on the valences of the choice alternatives, conflict situations emerge that have been classified as approach–approach conflict situations, avoidance– avoidance conflict situations, or approach–avoidance conflict situations (Miller, 1944). In an approach– approach conflict situation, a decision is made between desirable alternatives; in an avoidance–avoidance conflict situation, the choice alternatives are undesirable; in an approach–avoidance conflict situation, the choice alternatives possess both desirable and undesirable features. Furthermore, the uncertainty about possible consequences of a decision may intensify conflict (see, e.g., Hogarth, 1975). Several approaches that invoke conflict situations in the analysis of decision making have been proposed (e.g., Busemeyer & Townsend, 1993; Coombs & Avrunin, 1988; Diederich, 1997; Hull, 1932; Janis & Mann, 1977; Lewin, 1931/1964; Mellers, Schwartz, Ho, & Ritov, 1997; Miller, 1944; Townsend & Busemeyer, 1989), but as Tversky and Shafir (1992) pointed out, there exists neither a standard definition of conflict nor a generally accepted procedure for measuring conflict. Assuming that decision making in conflict situations is more difficult than in noconflict situations, they suggested the deferring of a decision as an indicator of conflict. Further experimental findings indicate that it takes longer to reach a decision in avoidance–avoidance conflict situations than in approach– approach conflict situations (e.g., Berlyne, 1957; Böckenholt, Albert, Aschenbrenner, & Schmalhofer, 1991; Busemeyer, 1985; Dashiell, 1937; Hansen, 1972; Houston, Sherman, & Baker, 1991; Janis & Mann, 1977; Luce, Bettman, & Payne, 1997). The purpose of the present study was (1) to introduce a model that actually predicts a decision time pattern depending on conflict situation type and (2) to probe the assumption that decision time can be used as a measure of conflict strength in risky decision making. To test this idea, experimentally multiattribute choice alternatives were constructed. The advantage of multiattribute choice alternatives is that they may have desirable attributes, undesirable attributes, or both. Moreover, the level of desirability/undesirability may vary for each attribute. Finally, the variability of their values can be manipulated for each attribute, so as to test the assumption that uncertainty about the consequences may induce conflict. Conflict will be manipulated by varying the relative desirability of the choice alternatives. The resulting choice probability pattern will depend on the preference for particular attributes and will allow one to infer the type of conflict experienced by the participant. Thus, the choice probability pattern will be used to predict the ordered relations of decision time that presumably indicates the conflict strength. In the following, I will briefly present a dynamic stochastic decision model for binary multiattribute decision problems, called the multiattribute decision field theory (MDFT), and its specific predictions for decision making in conflict situations. Details—in particular, the mathematical derivations—are given in Diederich (1996, 1997). MDFT is a sequential comparison model that extends and generalizes the so-called decision field theory (DFT) of Busemeyer and Townsend (1992, 1993) from