摘要
近几年,风险决策和“赌博范式”一直是经济学和心理学的研究主流,但奈特不确定性的选择问题没有引起学者的关注,尤其是旅游营销领域消费者潜在的认知过程。文章引入贝叶斯学习理论,还原旅游消费者学习过程,通过两阶段实验来探索感知价格和不确定质量信息对旅游消费者学习及其决策的影响机制。贝叶斯学习理论假设通过提供动态旅游广告信息的眼动追踪实验来观察消费者在奈特不确定性决策过程中的学习机制进而得到验证。结果表明,奈特不确定性下,旅游消费者在不同的信息状态空间St中基于不同感知价格水平,进行了不规则的贝叶斯学习;信念变化遵循一阶马尔可夫的学习规则;贝叶斯学习是一个双系统加工过程。总体而言,结果支持贝叶斯学习模型的信念整合原则。
In recent years,risk-related decision-making and the“gambling paradigm”have been prominent topics in economics and psychology research.However,decision-making under Knightian uncertainty has not attracted the attention of scholars,especially those studying consumers’cognitive processes in relation to marketing.The main statistical methods for dealing with uncertainty in decision-making can be divided into neoclassical methods(e.g.,expected utility theory),parameter correction methods(e.g.,prospect theory)and Bayesian methods,but the first two types focus on optimal decision-making under risk but not uncertainty,resulting in bias.Some recent studies have found that Bayesian learning is a reasonable means of overcoming Knightian uncertainty.However,there has been no empirical research on Knightian uncertainty,especially in relation to consumer decision-making.Given the lack of consensus in the literature on the learning rules used by consumers in real-world decision-making environments,these rules need to be evaluated with sufficient flexibility to accommodate Bayesian learning patterns.In this study,we propose a new method to evaluate the dynamic changes in beliefs that occur during consumers’Bayesian learning.We conducted an experiment and used data on the subjects’eye movements to confirm that the subjects used learning(or belief update)rules in their decision-making process.We refrained from complex comparative model fitting,instead using the classical method of hypothesis testing to investigate predictions regarding general differences in dependent variables and the effects of manipulating information relating to the price and quality of the product.The results of Bayesian learning are dependent on the cognitive process of belief updating,and the entire process of time and attention distribution is an important indicator of beliefs.In this study,we chose decision time,mean fixation duration,pupil dilation,distribution of attention,and direction of information search as the dependent variables.First,we investigated these measures from an overall perspective,aggregating them over the entire Bayesian learning process.Second,we analyzed changes in these variables during the decision-making period by dividing the learning process into several parts(i.e.,time bins).The results showed that the participants adopted a“certainty”strategy in state space St-1 when they perceived a high price or a low price,that is,they chose certain products by 70%.However,in state space St,when the perceived price was low,the probability of selecting uncertain products exceeded 60%,while when the perceived price was high,the consumers engaged in irregular learning,and the probability of choosing uncertain products was random.In state space St-1,the fixation proportion of the consumer on the product was separated by two-thirds of the decision time,while in state space St,they separated prematurely the fixation proportion of product by one-third of the decision time.Overall,the proportion of uncertain products in the late stage of decision-making showed a downward trend,while that of certain products showed an upward trend.Positive(alternative-based)and negative(dimension-based)search indicators exist in relation to information searching,indicating that consumers apply dual systems(reasoning and heuristics)to decision-making under Knightian uncertainty.In summary,when facing Knightian uncertainty,tourism consumers use irregular Bayesian learning based on different perceived price levels in different information state spaces,changes in consumers’beliefs follow the first-order Markov learning rule,and Bayesian learning is a dual-system process.
作者
宋红娟
SONG Hongjuan(School of Tourism,Hainan Tropical Ocean University,Sanya 572022,China)
出处
《旅游学刊》
CSSCI
北大核心
2020年第8期33-47,共15页
Tourism Tribune
基金
2019年度海南省教育厅高等学校科学研究项目“海南国际旅游岛文化展示与形象管理研究”(Hnky2019ZD-34)资助。
关键词
奈特不确定性
贝叶斯学习
旅游消费者
价格决策
Knightian uncertainty
Bayesian learning
tourism consumer
price decisions
