人类决策是否遵循补偿性模型所预期的加权求和计算过程? 多属性决策、风险决策、跨期决策的心理机制探索
文献类型：学位论文
| 作者 | 苏寅
|
| 学位类别 | 博士 |
| 答辩日期 | 2013-05 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 北京 |
| 导师 | 李纾 |
| 关键词 | 多属性决策 风险决策 跨期决策 补偿性模型 加权求和 |
| 其他题名 | Is human decision-making based on a weighting-and-adding process？Explorations of the mechanisms of multiple-attribute, risky, and intertemporal choices |
| 学位专业 | 心理学 |
| 中文摘要 | 多属性决策、风险决策和跨期决策是人们生活中经常遇到的，与人类生存息息相关的三种决策。这三种决策的主流理论均认为人们是按照一种补偿性的规则对选项中的信息进行整合，即通过加权求和的方式为每个选项计算一个总分，选择总分最大的选项。然而，越来越多的研究对补偿性模型提出了质疑，并提出了一系列基于启发式决策过程的非补偿性模型。本文通过三个研究，分别对多属性决策、风险决策和跨期决策的补偿性模型进行检验，旨在解决人类决策是否遵循加权求和计算过程这一由来已久的争论。 研究一通过问卷调查和眼动追踪，分别从决策结果和决策过程的角度，证实了选项的共享特征对多属性决策的影响。而共享特征效应违背了多属性决策领域中补偿性模型的独立性假设。因此，人们的多属性决策可能并非遵循补偿性模型预期的加权求和的计算过程。研究二包含两个子研究。研究2.1中，我们通过问卷调查，发现了风险决策中的“隐藏零效应”。这一现象违背了补偿性模型的不变性原理。研究2.2采用一种全新的基于比较的实验范式，通过分析人们决策时的眼动轨迹证实了人们风险决策时的信息搜索与加工过程与补偿性模型预期的加权求和计算过程存在显著差异。因此，人们的风险决策可能并非遵循加权求和的计算过程。研究三包含两个子研究。研究3.1中，我们通过问卷调查，发现了跨期决策中的“伪即刻效应”。这一现象违背了补偿性模型的不变性原理。研究3.2同样采用基于比较的实验范式，通过分析人们决策时的眼动轨迹证实了人们跨期决策时的信息搜索与加工过程与补偿性模型预期的贴现求和计算过程存在显著差异。因此，人们的跨期决策可能并非遵循加权求和的计算过程。 基于上述三个研究，本文认为人们在多属性决策，风险决策及跨期决策时，可能并非遵循补偿性模型所预期的加权求和的计算过程，而更倾向于遵循某种非补偿性模型所预期的启发式决策过程。因此，一味地修正基于加权求和决策过程的补偿性模型对多属性决策、风险决策和跨期决策理论的发展可能是徒劳的。未来的研究应尝试对现有非补偿性模型进行修正或发展新的基于启发式决策过程且对选择结果有更高预测力的非补偿性模型。 |
| 英文摘要 | Multiple-attribute choices, risky choices, and intertemporal choices are all very common andimportant in our daily life. Mainstream theories in these three fields consistently suggest that, to achieve a decision, people have to weight and sum the information from all dimensions in a compensatory way, and choice the option with the highest overall value. However, evidence has been accumulated to pose challenges to the compensatory models. Three studies were performed, aiming to address the long-standing issue about whether multiple-attribute, risky, and intertemporal choices are based on a weighting-and-adding process. In study 1, weexaminedthe mechanism ofmultiple-attribute choice. We found that the common features of alternatives which were assumed to be cancelled during decision-making could actually influenceboth the decision process and outcome. This finding violated the independence assumption of compensatory models. We therefore suggested that multiple-attribute choices mightnot be based on a weighting-and-adding process.In study 2, we examined the mechanism of risky choice. We found a “hidden-zero effect”in risky choice, which violated the invariance principle of compensatory models. We also found that a significant difference between the process of information processing in a risky choice task and that in a baseline task in which people have to follow a weighting-and-adding process in making a decision. We therefore suggested that risky choices might not bebased on a weighting-and-adding process.In study 3, weexamined the mechanism of intertemporal choice.We found a “psudo-immediacy effect” in intertemporal choice, which violated the invariance principle of compensatory models. We also found that a significant difference between the process of information processing in a intertemporal choice task and that in a baseline task in which people have to follow the weighting-and-adding process to make a decision. Wetherefore suggested that intertemporal choices mightnot bebased on a weighting and adding process. Taken together, we suggested that human decision-making might not be based on a weighting-and-adding process that was predicted by the compensatory models, but be more likely based on a heuristic process that was predicted by the non-compensatory models. We implied the need to improve current theories or develop new theories that postulate the use of a heuristic process. Keywords: multiple-attribute choice; risky choice; intertemporal choice; compensatory models; a weighting-and-adding process |
| 学科主题 | 应用心理学 |
| 语种 | 中文 |
| 源URL | [http://ir.psych.ac.cn/handle/311026/19657]  |
| 专题 | 心理研究所_社会与工程心理学研究室 |
| 作者单位 | 中国科学院心理研究所 |
| 推荐引用方式 GB/T 7714 | 苏寅. 人类决策是否遵循补偿性模型所预期的加权求和计算过程? 多属性决策、风险决策、跨期决策的心理机制探索[D]. 北京. 中国科学院研究生院. 2013. |
入库方式： OAI收割
来源：心理研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。