Collaborative Research: Substantive and procedural Rationality in Decisions under Uncertainty

该项目旨在生成和分析数据，以便在实验室中理解在不确定情况下的选择。与以前的工作不同，本文的目的不是揭示违反特定理论的行为，而是在丰富的选择环境中对不确定情况下的选择提供积极的解释，这种环境允许对个人行为模式的一般表征。为了刻画一个人在不确定情况下的决策，有必要对每个受试者在大范围的选择集上产生许多观察。为此开发了一种创新的图形界面，受试者在计算机屏幕上看到涉及两个(或三个)风险资产的投资组合选择问题的几何表示。受试者通过简单的“点击”来选择投资组合。这个直观和用户友好的界面允许每个受试者在广泛的选择场景下快速而有效地得出许多决定。这些数据的初步发现表明，受试者的行为是一致的，因为选择使一组一致的偏好最大化；强调这项测试不依赖于任何特定的不确定情况下的选择理论，而是对合理性的更基本的测试。下一项任务是描述选择行为中的(一致)模式。这是通过两种方式完成的。首先，作为一种描述性练习，研究受试者是否使用可识别的启发式(“经验法则”)来做出决定是有用的。对两资产实验产生的数据的检验表明，确实存在共同的启发式方法来指导许多受试者的选择。其次，为了将这种启发式选择方法与实质理性联系起来，估计了一个类型混合模型(TMM)。TMM通过假设个人试图不完美地最大化其潜在的(一致的)偏好来预测个人对启发式的选择。两资产模型的初步结果表明，TMM在描述受试者行为方面是有效的。该项目将采用类似的方法来研究对三种风险资产的选择，以及研究模糊情况下的选择，即在“真实”概率未知的情况下进行选择。该项目将开发必要的工具，包括方法论和分析性工具，以提供对不确定情况下决策的全面的个人层面的分析。更广泛的项目的两个明显特征是新的实验设计，以及将消费者需求理论的工具应用于实验室中的个人决策。广泛影响：不确定性下的决策进入经济决策的每个领域，因此对这一领域的根本贡献将具有广泛的影响。此外，已经开发并将通过该项目改进的实验平台和分析技术可以应用于多种类型的个人选择问题。目前正在开发的实验方法的更精细版本将允许受试者之间的互动。所有用于分析的实验和程序都将通过网络公开。

This project aims to generate and analyze data for understanding choice under uncertainty in the laboratory. In contrast to earlier work, the purpose is not to uncover violations of particular theories, but rather to provide a positive account of choice under uncertainty in a rich choice environment that allows for a general characterization of the patterns of individual behavior. To characterize an individual's decision-making under uncertainty, it is necessary to generate many observations per subject over a wide range of choice sets. An innovative graphical interface was developed for this purpose, where subjects see on a computer screen a geometrical representation of a portfolio choice problem involving two (or three) risky assets. Subjects choose portfolios through a simple 'point-and-click.' This intuitive and user-friendly interface allows for the quick and efficient elicitation of many decisions per subject under a wide range of choice scenarios.Preliminary findings with these data suggest that subjects' behavior is consistent in the sense that choices "maximize" a consistent set of preferences; it is emphasized that this test does not rely on any particular theory of choice under uncertainty, but is rather a more fundamental test of rationality. The next task is to characterize the (consistent) patterns in choice behavior. This is done in two ways. First, as a descriptive exercise, it is useful to study whether subjects use identifiable heuristics ("rules of thumb") to make their decisions. Inspection of the data generated by the two-asset experiment suggests that there are indeed common heuristics that govern many subjects' choices.Second, in order to connect this heuristic approach in making choices to substantive rationality, a type mixture model (TMM) is estimated. The TMM predicts an individual's choice of heuristics by assuming that the individual is attempting to imperfectly maximize his underlying (consistent) preferences. Preliminary results with the two-asset model suggest that the TMM is effective in describing subjects' behaviors. An analogous methodology will be applied to studying choice over three risky assets and to the study of choice under ambiguity, that is, choice when the "true" probabilities are not known.This project will develop the necessary tools, both methodological and analytical, for providing a comprehensive individual-level analysis of decision-making under uncertainty. Two distinctive features of the broader project are the new experimental design, and the application of the tools of the theory of consumer demand to individual decision-making in the laboratory.Broader Impact: Decisions under uncertainty enter every realm of economic decision-making, so fundamental contributions to this area will have broad-reaching implications. Further, the experimental platform and analytical techniques that have been developed and will be refined through this project can be applied to many types of individual choice problems. More elaborate versions of the experimental method currently under development will allow for interaction between subjects. All experiments and programs for analysis will be made publicly available through the web.