IA Understanding the Effects of Choice Set Complexity in Stated Preference Methods

经济学家采用陈述偏好法来衡量消费者从各种环境计划中获得的使用和非使用价值。最近，经济学家和联邦机构已经开始探索通过从条件估值方法转向所谓的联合方法来推广陈述偏好方法的可行性。条件估值法只给消费者提供一个价格变化的选择，而联合估值法给消费者提供一个包含多个选择的选择集，每个选择都沿着一些属性变化。联合分析的支持者认为，这样可以更完整地描述项目属性所产生的效益，而且与传统的说明偏好方法相比，这样做可能具有更高的统计效率和更低的财务成本。然而，一些著名的经济学家表示担心，联合分析中使用的高度复杂的选择集可能会压倒消费者的认知能力，从而导致不可靠的使用和非使用价值的估计。这场争论仍在继续，因为很少有关于选择集复杂性对经济效益估计的影响的研究。因此，EPA和NOAA等公共机构仍然不确定联合方法与其他非市场评估技术的相对优点。这种不确定性阻碍了司法、监管和公共投资决策中获取和使用有关经济利益的信息。 我们进入这场辩论的核心问题进行了一些严格和系统的研究：如何选择集的复杂性影响的方式，消费者揭示他的环境设施的价值？为了回答这个问题，我们在一个精心设计的陈述偏好实验中改变了选择集的复杂性，然后从实验的几个维度来评估选择结果。首先，选择不一致（即，犯错误的倾向）随复杂性而变化？第二，消费者的信息处理策略是否会随着选择集的复杂程度而变化，从而显著改变他们的选择（从而改变经济学家对收益的估计）？第三，消费者的个人特征如何与选择集的复杂性相互作用，从而影响选择一致性和信息处理策略？第四，消费者对选择集复杂性的主观评估与我们作为研究人员可以客观量化的复杂性度量之间有什么关系？ 探讨这些实质性问题将使我们能够确定目前实践的局限性，并扩大对既定偏好方法的研究领域。我们分析的中心是一个选择的行为模型，它整合了心理学、经济学和市场营销领域的几种现有文献。从这个模型中，我们开发了一套假设，我们可以进行统计测试，以提供上述问题的答案。通过测试这些假设，我们将开发方法来识别和控制选择集的复杂性，在规定的偏好应用程序。我们相信这些方法将成为该领域未来研究的标准工具。在本提案中，我们只为新创建的数据集的分析师寻求资金。作为最近完成的哈佛大学研究项目的一部分，首席研究员设计并实施了联合实验，这些实验由经济学和营销研究领域备受尊敬的专家进行了同行评审。然而，作出了一项战略决定，将这种原始研究支助用于获得尽可能高质量的数据，并在初步分析中只回答必要的政策问题。我们现在已经用尽了这一原始资金，并正在寻求额外的支持，以解决重要的方法问题，可以使用这些创新的声明偏好调查实验进行检查。

Economists employ stated preference methods to measure the use and non-use value that consumers derive from a variety of environmental programs. Recently, economists and federal agencies have begun to explore the feasibility of generalizing stated preference methods by moving from the contingent valuation method to so-called conjoint methods. The contingent valuation method presents consumers with only one alternative, for which the price varies, while conjoint methods present consumers with a choice set that contains several alternatives that each vary along a number of attributes. Proponents of conjoint analysis argue that this allows a more complete characterization of the benefits derived from the project's attributes, and may do so with greater statistical efficiency and lower financial costs than traditional stated preference methods. However, several prominent economists have expressed concern that the highly complex choice sets used in conjoint analysis may overwhelm the cognitive capacities of consumers and thus lead to unreliable estimates of use and non-use value. This debate continues because there is very little research about the effects of choice set complexity on estimates of economic benefits. As a result, public agencies such as EPA and NOAA are still uncertain as to the relative merits of conjoint methods versus other non-market valuation techniques. This uncertainty impedes the acquisition and use of information on economic benefits in judicial, regulatory and public investment decisions. We step into this debate with some rigorous and systematic research on the core issue: How does the complexity of a choice set affect the way in which a consumer reveals his value for environmental amenities? To answer this question, we vary the complexity of the choice set in an elaborate stated preference experiment and then evaluate choice outcomes with regard to several dimensions of the experiment. First, how does choice inconsistency (i.e., the propensity to make mistakes) vary with complexity? Second, do consumers' information processing strategies vary with the level of choice set complexity in a way that significantly changes their choices (and thus economists'estimates of benefits)? Third, how do the personal characteristics of consumers interact with choice set complexity to affect choice consistency and information processing strategies? Fourth, how are consumers' subjective assessments of choice set complexity related to the measures of complexity that we, as researchers, can objectively quantify? Answering these substantive questions will enable us to identify the limits of current practice and to extend the frontiers of research on stated preference methods. At the center of our analysis is a behavioral model of choice that integrates several strands of existing literature in the fields of psychology, economics and marketing. From this model, we develop a set of hypotheses that we can test statistically in order to provide answers to the above questions. By testing these hypotheses, we will develop methods to identify and control for choice set complexity in stated preference applications. We believe these methods will become standard tools for future research in this area. In this proposal, we seek funding only for the analysts of a newly created data set. As part of a recently completed Harvard University research project, the principal investigator designed and implemented conjoint experiments that were peer-reviewed by highly respected experts in the fields of economics and marketing research. However, a strategic decision was made to invest this original research support in obtaining the highest possible quality of data and to answer only the requisite policy questions in the initial analyses. We have now exhausted this original funding and are seeking additional support to address the important methodological issues that can be examined using these innovative stated preference survey experiments.