社会与工程心理学研究室知识库

在长达上百万年的进化过程中，人类凭借智慧成功地完成觅食、选择配偶、保护后代等“生存任务”，最终登顶食物链顶端，正是因为我们经过漫长的自然选择历程，拥有有别于动物的独特技能——做出好的决策。在历史发展过程中，有三类决策对人类的生存和繁衍至关重要：一是风险决策，二是跨期决策，三是空间决策。人类不断地做出类似上述适应性决策，社会才得以不断向前发展，因此探索对人类生存和繁衍至关重要的风险决策、跨期决策和空间决策的基本机制问题具有重要的理论意义和现实意义。

风险决策、跨期决策、空间决策是典型的人类决策类型。这三类决策的基本加工过程是否不同或者相同？决策过程是否统一地遵循补偿性策略或非补偿性策略？学界对此一直有不一致看法。尽管强调三类决策存在系统差别以及强调补偿性加工策略的观点长期占据主导地位，但近年来人们越来越意识到三类决策加工过程的相似性以及非补偿决策策略或许更符合真实决策。然而，探究该问题具有较大的挑战性，因此学界迟迟未能给出令人信服的回答。在大量文献综述和前期实验基础上，本论文认为三类决策可能统一地遵循非补偿性的决策策略。基于此，提出本研究拟解决的科学问题：“风险决策、跨期决策和空间决策的加工过程是否一致？三类决策是否会统一遵循非补偿性、基于维度的决策策略？”。

为回答这一关键科学问题，本博士论文的研究一通过构建三类决策所共享的行为效应和探索行为效应背后的解释机制来回答加工过程是否一致以及一致的决策策略问题。研究1a分别在跨期和空间决策中发现了类似风险决策中的次比例效应，非补偿性理论中的维度差判断可以同时解释三类决策中出现的“次比例”效应；研究1b通过操纵实验参数进一步探索三个决策领域是否会以同样的方式出现“次比例的反例”，结果发现通过操作参数中的维度差可以改变次比例选择模式，从而出现次比例反例，非补偿性理论中的维度差判断同样可以预测行为选择结果。研究一的结果证据暗示了三类决策共享行为效应，并一致地遵循非补偿性策略进行决策。

研究二利用眼动追踪技术对三类决策的过程一致性和一致遵循的策略提供过程证据。结果发现：三类决策在反应时、注视点平均注视时长、长注视点比例和信息加工深度等眼动过程指标上不存在显著差异，且贝叶斯因子中高强度的支持该假设，说明三类决策在决策过程上具有一致性，同时三类决策的SM值结果表明风险、跨期和空间决策统一遵循非补偿性的决策策略；其次，通过引入选择/拒绝反应模式和基线选择任务深入探究三类决策是否遵循补偿性的加权求和过程，发现基线选择任务有别于自由选择任务，同时，基线选择任务不受反应模式的影响，而自由选择任务受到反应模式的影响，这说明个体的自由选择任务有别于包含审慎计算过程的基线选择任务，由此为证明“人们的自由选择是遵循非补偿性、基于维度的决策策略”提供了更多的证据。

最早对风险、跨期和空间决策进行的研究主要基于单结果的实验参数，然而面对更复杂的决策选项，还尚缺乏深入研究。单结果到双结果并非简单的增加参数，反映在决策研究中是质的飞跃，尤其对于非补偿性理论而言，无法在众多的维度中进行表征。基于此，研究三将触角延伸到双结果决策领域中，探索双结果决策的决策过程是否依旧存在一致性，其一致性策略是否沿袭非补偿性决策策略。结果发现，三类双结果决策在决策过程中具有共享的决策机制，一致的遵循非补偿性、基于维度的决策策略，同时双结果中基线选择任务同样有别于自由选择任务，值得注意的是双结果实验参数中更能体现出选择和拒绝反应模式的差异。

本研究的创新之处在于首次通过构建“次比例”行为效应将风险、跨期和空间三类决策纳入同一体系，并通过眼动追踪技术、引入反应模式和双分离逻辑等多种方法同时获得了关于三种领域（风险、跨期、空间）决策的过程性证据，同时在数据分析方法上将传统的假设检验与贝叶斯因子分析相结合，冀在决策任务过程分析和贝叶斯因子分析技术趋于成熟的时机下，为研究结果提供汇聚性证据，适时探索回答风险、跨期和空间决策是否具有共同机制这一科学问题。

In the million years of human evolution, humans have the wisdom to successfully complete “survival tasks” such as foraging, mating, protecting offspring, and finally reaching the top of the food chain. That phenomenon happens precisely because of the long process of natural selection. Our unique skill as human beings, that is, making good decisions, can essentially distinguish us from annimals. In the historical development process, three types of decisions are deemed critical to human survival and reproduction: risky, intertemporal, and spatial choices. Humans continue to make adaptive decisions similar to the above, thereby enabling human society to continue moving forward. Therefore, it is necessary to explore the mechanism of risk, intertemporal and spatial decision making that are vital to human survival and reproduction as well as its internal mechanisms. The issue of consistency has important theoretical and practical significance.

Risky, intertemporal and spatial choices are three typical decision type. Are the processes of these three decision type different or the same? Does the decision process uniformly follow compensatory or non-compensatory strategies? The long-standing debates among the theoretical camps of compensatory, option-based and non-compensatory, dimension-based decision making have not yet been unified. Although the viewpoint of emphasizing the systematic differences among the three decision domains, and emphasizing compensatory processing strategies have long been dominant, people have increasingly realized consensus of the three decision types and the shared non-compensatory decision-making strategies. However, it is challenging to systematically explore this question, a convincing answer has not been given.

Based on a large number of literature reviews and previous experiments, this paper argues that the three types of decisions may uniformly follow the non-compensatory decision-making strategy. Based on this proposition, the following scientific question is raised in this research: "Are the processes of risk, intertemporal and spatial decision making consistent? Will the three types of decisions follow a non-compensatory decision strategy?"

To answer this key scientific question, study 1 aims to answer the question of whether the processing and decision making strategy are consistent by constructing the behavioral effects shared by the three kinds of decisions and exploring the mechanism behind the behavioral effects. Study 1a found that the subproportionality effect in intertemporal and spatial choices, similar to subprobortionality in risk choice, was found. Meanwhile, the intra-dimensional difference in the non-compensatory theory can simultaneously explain the “subproportionality” effect in three decision domains. Study 1b manipulated the experimental parameters to further explore the existence of counter-example subproportionality. Results show that the manipulated intra-dimensional difference can lead to the counter-example subproportionality choice pattern, and the intra-dimensional difference judgment of the non-compensatory theory can predict the choice pattern. Results of study 1 suggest that the three types of decision domain share similar behavior effects and consistently follow non-compensatory strategies to make the final decisions.

In Study 2, the eye-tracking technique is introduced to further investigate the process consistency and strategies of the three types of decision making. The results showed that three kinds of decision at the reaction time, the mean fixation time, depth of information processing, among others, have no remarkable differences. Additionally, the Bayesian factor shows moderate evidence supporting the finding that the three choice domains remain consistent in the process index. Meanwhile, SM value results showed that the risk, intertemporal and spatial choices uniformly follow the non-compensatory strategies. Second, the selection-rejection response mode and baseline choice tasks are introduced further to explore whether the three kinds of decision domains follow the compensatory process. Results found that the baseline choice task is different from the free choice tasks. Moreover, the baseline choice task is not affected by the response mode, whereas the free choice task under the influence of response mode shows that free choice task is different from the baseline choice task.

The earliest studies on risk, intertemporal, and spatial decisions are based on one-outcome experimental parameters. However, in face of other complex decision options, the current studies remain lacking. The change from one-outcome to double-outcome is not a simple parameter increase but a qualitative leap in decision research, especially for non-compensatory theory, which cannot be represented in many dimensions. Based on this finding, study 3 extends its reach to the field of double-outcome decision making to explore whether the decision-making process of double-outcome decision remains consistent and whether the consistency strategy inherits non-compensatory decision-making strategy. The results show that the three types of double-outcome decisions have a shared decision-making mechanism in the decision- making process and consistently follow the non-compensation decision-making strategy. Meanwhile, the baseline choice task in the double-outcome experiment is also different from the free choice task. Notably, the double-outcome experiment parameters can better reflect the difference between the selection and rejection response modes.

The innovation of this research lies in obtaining procedural evidence for decision making through various methods such as eye-tracking technique, response mode. Simultaneously, it combines traditional null hypothesis testing with Bayesian factor analysis in data analysis to analyze the process of decision-making. At the time when the Bayesian factor analysis technology is becoming mature, it provides convergent evidence for the research results and timely attempt to answer the scientific question of whether risky, intertemporal and spatial choices share a common mechanism.