Testing the Scenario Hypothesis: The Effect of Alternative Characterizations of Uncertainty on Decision Structuring

目前人们对可能有助于改善困难决策的决策支持有相当大的兴趣，例如规划者在考虑不确定的长期环境变化时所做的那些决策。以前的研究和实践表明，基于情景的规划过程可以通过设计额外的和适应性的选项来帮助决策者重组决策。这项研究通过比较决策支持工具的决策来检验这一情景假设，这些决策支持工具以两种方式之一提供预测：要么通过提供最佳估计概率，要么通过提供替代情景。该假设指出，情景版本将导致决策选项成功重组的可能性更高。参与者面临着管理假想渔业的挑战。他们以个人或团体的身份，为延伸到未来的一段时期制定了适当的捕捞限制政策。然后，参与者使用交互式渔业模式决策支持工具，帮助他们评估备选方案并设计新的备选方案，同时对可用备选方案及其潜在绩效施加合理的限制。与会者被告知，关键投入数据，如渔业的可持续产量，是科学分歧的主题，导致对未来的严重不确定性。渔业模型将表明，与其他决策选择相比，特定的适应性管理战略将使参与者满意？在对可持续收益和其他不确定性的广泛预期上，目标相互冲突。但最初给予参与者的选项不会包括这种适应性战略，他们也不会立即发现这种战略的存在是显而易见的。主要的实验操作是决策支持工具是根据最佳估计预测还是根据情景来组织信息。主要的结果衡量标准是确定成功的适应性战略的概率。帮助决策者设计新的、可能更成功的决策选项的能力可能是决策支持的一个重要方面。但许多关于决策制定和决策支持的研究都集中在一组固定的选项中进行选择，而不是设计新的选项，以减少困难的权衡。深度不确定性的存在？决策者可能对意外情况和概率不确定的情况？可能会使人们更难理解或商定替代备选方案的潜在后果，从而更难有效地设计新的备选方案。这些实验的结果揭示了复杂、高度不确定的政策决策所涉及的社会和心理过程。这项工作还有助于改进决策支持方法，并有助于就传达极不确定信息的最有效手段进行政策讨论。

There is considerable current interest in decision support that might help improve difficult decisions, such as those that planners make when considering uncertain, long-term environmental changes. Previous research and practice suggests that scenario-based planning processes can help decision makers to restructure decisions by designing additional and adaptive options. This research tests this scenario hypothesis by comparing decisions using decision-support tools that provide forecasts in one of two ways: either by presenting best-estimate probabilities or by presenting alternative scenarios. The hypothesis states that the scenario version will lead to higher probability of a successful restructuring of the decision options. Participants have the challenge of managing a hypothetical fishery. Acting as individuals or in groups, they set a suitable catch-limit policy for a period extending into the future. Participants then use an interactive fisheries model decision-support tool that helps them to evaluate options and to design new ones, while placing reasonable constraints on the options available and their potential performance. Participants are informed that key input data, such as the sustainable yield from the fishery, are the subject of scientific disagreement, resulting in deep uncertainty about the future. The fisheries model will show that a particular adaptive management strategy will, in contrast to other decision options, satisfy the participants? conflicting objectives over a wide range of expectations about sustainable yield and other uncertainties. But the options initially given the participants will not include this adaptive strategy nor will they find it immediately obvious that such a strategy exists. The primary experimental manipulation is whether the decision-support tool organizes information in terms of best-estimate forecasts or in terms of scenarios. The primary outcome measure is the probability of identifying the successful adaptive strategy. The ability to help decision makers design new and potentially more successful decision options could be a valued aspect of decision support. But much research on decision making and decision support focuses on choice among a fixed set of options, rather than on the design of new options that might reduce difficult tradeoffs. The presence of deep uncertainty ? where decision makers may be unsure about contingencies and probabilities ? can make it even more difficult to understand or agree on the potential consequences of alternative options, and thus to design new options effectively. The results of these experiments shed light on the social and psychological processes involved in complex, deeply-uncertain policy decisions. The work also informs improvements in decision support methodologies and contributes to ongoing policy discussions regarding the most effective means of conveying deeply uncertain information.