| 其他摘要 | Human beings live in groups. When faced with important decisions, they often seek input from other members of their group in order to make informed choices. While group decision making can improve the quality of decisions, it can also lead to phenomena such as groupthink, group polarization, and social loafing, which can weaken the efficacy of decision making and lead the decision to the wrong direction. In today's society, where cooperation has become increasingly important, it is essential for individuals to understand how to interact within a group to arrive at effective and reasonable decisions. The key to addressing this issue lies in conducting in-depth research on group decision-making process, developing and refining techniques that enhance the quality of group decisions, and understanding the nature of the negative phenomena that emerge during group decision making.
Experiments are a common method of investigating group decision making. However, group experiments often require strict controls and are difficult to conduct on a large scale, making it challenging to collect sufficient data. Furthermore, there are many confounding factors in group interactions, making it difficult to establish clear causal relationships between experimental variables and decision outcomes. Computational modeling and simulation can complement group experiments. They can quickly and easily adjust the levels of various factors that affect group decision making, and perform quantitative analysis of the influencing factors and their interactions. In addition, the results of computational modeling and simulation can help researchers propose reasonable research hypotheses and determine experimental conditions for testing these hypotheses, thereby enabling more targeted experimental research and a deeper understanding of the process and mechanisms of group decision making.
The research in this dissertation is based on the framework of a group signal detection model. It incorporates multiple rounds of decision making and multi-task settings to construct a group decision-making model that is better aligned with real-life scenarios. Subsequently, computational modeling and simulation research were conducted to address both the Delphi group decision method and the social loafing phenomenon. Online and offline experimental platforms were then used to perform behavioral experiments and verify the simulation results.
Study 1 focuses on the Delphi group decision method and analyzes the key factors that affect decision quality in group decision scenarios that involve multiple rounds of information exchange. Additionally, it explores parameters that are conducive to improving the effectiveness of the Delphi method. The modeling research shows that group size, stopping rules, and the noise level of members' information integration can all affect decision efficiency. The results of the experimental research indicate that the highest accuracy is achieved when a larger group is assembled (i.e., a 19-person group) and when more stringent stopping rules are applied (i.e., the decision-making process stops only when all members make the same choice).
Study 2 focuses on the social loafing phenomenon. Through modeling and experimentation, we construct a theory of attention allocation in group decision making, and explore methods to reduce social loafing, and even use social loafing to enhance group members' overall wellbeing. When group members need to allocate attention resources to group tasks and individual tasks, modeling results show that the optimal strategy for members is influenced by the behavior and abilities of other members, specifically: (1) when only one member in the group can loaf, that member should loaf; (2) when other members in the group are expected to loaf, that member should also loaf; (3) when every member in the group is capable of loafing, if their decision abifity is low, all members should not loaf; otherwise, there should be some degree of loafing. Subsequent experimental results show that social loafing is widely present in group decision making tasks, and the loafing strategy of group members was somewhat consistent with the optimal strategy.
Overall, this research proposes techniques to optimize the Delphi method of group decision making and a group attention allocation theory that can explain social loafing. It provides both technological and theoretical insights to leverage the advantages and mitigate the disadvantages of group decision making, and opens up new paths to improve the effectiveness of group decision making. The main innovation of this study lies in the integration of computational simulations and behavioral experiments, as well as the application of cutting-edge techniques to explore the mechanisms and methods for improving the quality of group decision making. In terms of research methodology, this study extends the existing framework of group signal detection models and develops an online group experiment platform, which can improve the efficiency of group research. The exploration and discoveries of this research have filled the long-standing knowledge gaps in the field of group decision making. Overall, this research has provided innovative theoretical, methodological, and practical findings, led to new research questions, and made valuable contributions to the field. |
修改评论