推理判断中偏差反应的加工机制：冲突探查失败，还是抑制失败？

摘要： 通过控制自变量一致性（冲突、非冲突）与问题类型（基础比率问题、合取问题），对推理判断中偏差反应的加工机制：冲突探查失败，还是抑制失败，进行验证。反应时、反应自信度的结果支持抑制失败说，被试对冲突的探查是成功的，偏差反应的产生是没能抑制占主导的直觉。而冲突探查大小（Conflict detection size）的结果表明冲突探查存在个体差异，有些被试对两种加工间的冲突探查过程完美无暇(Flawless detection)，有些被试的冲突探查过程松散马虎(Lax detection)，而且受到问题类型的影响。
Abstract: Humans are always biased in their reasoning, judging and decision making. A critical question to answer these bias responses is that the biases arise from the participants’ failure to detect the conflict between heuristic processing and analytical processing, or from their failure to inhibit the dominant heuristic intuition? There are two different views on the efficiency of the detection process. Evans and Kahneman claim that the conflict detection between intuitive response and standard logical rule is typically quite lax, causing failure to engage in analytical processing and modify the default intuitive response. But other scholars, including Sloman and Epstein, claim that the monitoring is flawless because the two processes parallel competitive process. People know and perceive the conflict between the two responses; however they can't successfully suppress the dominant intuitive response between the two reactions. The purpose of this study is to verify whether the bias response is from the failure of the conflict detection or the failure of suppression. Participants solved a total of 20 base rate problems and conjunction Fallacy problems on the computer, half of which was basic ratio problems and the other half was conjunction principle problems. Two types of problems were presented: (1) conflict problems, containing stereotypes that conflicting the large base-rate group or stereotypical description; (2) non-conflict problems, containing stereotypes that matched the large base-rate group or stereotypical description. After making a judgment, participants were asked to grade confidence in their response from a scale of 1 to 5. SPSS19.0 is used to analyze the following data such as participants’ accuracy; response confidence; response latency and conflict detect size between the conflict and non-conflict problems. The results show that participants’ accuracy in solving conflict problems is significantly lower than that in solving conflict problems. Most of the participants made a heuristic response to the conflict condition and neglected the basic ratio and the probability of conjunction principle. If there is no conflict detection between intuitive response and standard logical rule, there should be no difference between conflict problems and non-conflict problems in