The Risk In-Basket

When experienced decision makers are presented with a choice between two alternatives, one risky and one relatively safe, which do they choose? Frequently the answer is neither. Consider, for instance, a top-level business executive faced with a threat by a union at a subsidiary company to go on strike unless a time-andmotion study is immediately discontinued. The two obvious actions of the executive are (1) ignore the threat and continue the study or (2) discontinue the study. The typical executive, however, is likely to collect more information (about the potential payoff from the study, about the union's position, etc.), to initiate negotiations with the union leaders, to delay taking any immediate actions, or to delegate the decision to the executives of the subsidiary. A typical experimental approach to studying decision making under uncertainty is to ask a subject to choose between two lotteries defined In this study over 400 top-level executives confronted four risky situations cast in an in-basket format. Each situation involved a choice between a risky and a sure alternative. Having the executives respond by writing memos to specify how the situation should be handled allowed an examination of a rich variety of actions for modifying risk. Responses in the form of probability equivalences and rating scales were also obtained. The results showed that the executives had a relatively strong tendency toward risk seeking, particularly when the payoffs were negative. All three response modes gave generally consistent information about risk propensity, but there were significant differences across the four situations. Overall, between one-fifth and one-half the time, the executives attempted to modify the risks especially through collecting new information. Delegating, delaying, and collecting information were positively associated with a desire to avoid risks. The Risk In-Basket was an effective way to assess risk propensity and to study risk modifications. * This research was supported by grants from the Technology Branch of the Canadian Department of Industry, Trade, and Commerce and from the Social Sciences and Humanities Research Council of Canada. We wish to acknowledge the major contributions of our colleague William T. Stanbury to the design, data collection, and coding. Alfred Kwong assisted on the design and John Bassler contributed to the data collection and coding. We are grateful to Vinay Kanetkar and Brad Cousins for their work on the analysis. The comments on an earlier draft by an anonymous referee were helpful.