The Effect Of Cognitive Reflection Capacity And Task Complexity On Bayesian Reasoning

Previous studies of Bayesian reasoning have consistently shown that people perform poorly on classic Bayesian reasoning.In order to improve people's performance on Bayesian reasoning,the researchers reduced the difficulty of Bayesian reasoning from the complexity of inferential materials,such as problem situation,information representation,structure complexity and so on.However,cognitive reflective ability is an important subjective factor of Bayesian reasoning.People with high cognitive reflective ability are more likely to suppress false intuitions,their cognitive resources are abundant,their reasoning grades are more stable.So,the effects of cognitive reflective ability and task complexity are discussed in this study.Important decisions in daily life are based on a variety of information cues,not single information cues,and information presentation is diversified and visualized.Decisions are made through tedious calculations.Therefore,this study takes into account the characteristics of daily life information,considers task complexity from three aspects:information presentation,structural complexity and mental steps required to get the correct answer.On the basis of the findings above,this study aimed to discusses the effect of cognitive reflective ability and task complexity on the Bayesian reasoning.In order to test this hypothesis,in experiment 1,a 2x3 between subjects experiment was in designed,the experiment aim to discusses the effect of cognitive reflective ability and information presentation in the Bayesian reasoning.The independents variables were cognitive reflection capacity("high cognitive reflection capacity" and "low cognitive reflection capacity")and the presentation format of the statistical information("the natural frequency format"?"the probability format"and "the probability format with tree diagram presentation").The dependent variables was the accuracy of the estimated result of the Bayesian reasoning that the absolute value of difference between posterior probability and normal value.The statistical methods that we used in the study were the analysis of variance and t test of independence.In experiment 2,a 2×2 between subjects experiment was in designed,the purpose of this experiment is to explore the effect of thinking styles and structure complexity on the Bayesian reasoning.Participants completed a Bayesian reasoning task under the natural frequency format in a 2(the rational and experiential thinking styles:high group vs low group)x2(the structure complexity:complexity vs classic").The dependent variables was the same as the experiment 1.In experiment 3,a 2x2 between subjects experiment was in designed,the purpose of this experiment is to explore the effect of the thinking styles and the number of computational steps required to get the right answer on the Bayesian reasoning.Participants completed a Bayesian reasoning task under the natural frequency format in a 2(the rational and experiential thinking styles:high group vs low group)x2(the number of computational steps required to get the right answer:a computational step vs four computational steps").The dependent variables was the same as the experiment 1.The results were as followed:(1)Task complexity of three aspects:information characterization,structure complexity,and mental steps required to get the correct answer have an impact on Bayesian reasoning,the inference accuracy of participants in the complex Bayesian reasoning task is lower.(2)The cognitive reflective ability has different role in different Bayesian reasoning:The facilitation effect of natural frequency does exist among individuals with different cognitive reflection capacity;When the structure of the Bayesian reasoning in natural frequency presentation become complex,higher group of the rational thinking styles than lower group inference accuracy is higher,and the role of experiential thinking styles just the opposite;whether under the condition of a mental step or four mental steps,the rational and experiential thinking styles has no effect on the accuracy of natural frequency Bayesian inference.(3)The findings support the nested sets account(NSA)of the Bayesian reasoning.