启发式概率估计在反馈条件下的偏向研究

【作者】 黄凌云；

【导师】 刘电芝；

【作者基本信息】 西南师范大学 ， 发展与教育心理学， 2003， 硕士

【摘要】 启发式概率估计是当前国外自然推理研究的新兴的热点问题之一。非确定条件下的启发式推理是自然推理中最重要的部分，由于启发式具有经验性和意向性的特点，因此有关启发式的研究多限于理论假设，而主观概率的估计作为自然推理中命题判断的一种变形，与启发式一起形成了启发式概率估计这一新的研究亮点。启发式概率估计的研究范式能够将启发式理论假设与概率估计的可量化特点结合起来，可以更好地以实验的方法探索启发式推理的内部机制。尤为重要的是，启发式概率估计的研究突破了传统推理研究以形式逻辑为材料和评判标准的不足，研究重心转向现实思维活动的或然性本身。 研究反馈对启发式概率估计的作用，是通过揭示反馈如何影响启发式的各个环节，从而准确把握启发式产生偏向的心理机制。本研究从概率问题呈现方式的次序、概率问题的提问类型两个方面组成了几个主要的反馈条件，用两个因素分析实验来测量这些反馈条件下的偏向变化状况：实验一是验证问题的呈现次序与反馈这两个因素对概率估计的偏向影响；实验二是验证提问类型和反馈这两个因素对概率估计的偏向影响。实验得到了以下几个结果： 一、反馈能够改善启发式概率估计的内部过程。对于不同的概率问题呈现形式，反馈调节的效果也不同，频率式呈现方式条件下的估计效果好于细节样例式。 二、频率式和细节样例式属于两种不同的概率问题呈现形式，在启发式概率估计中，它们分别引发了两种不同形式的思维方式，即分布式和单一式思维。本研究进一步揭示这两种思维方式不存在交互作用，即使在反馈条件下也不能增强这种影响作用。 三、在提问类型这样涉及概率问题内容性的变化时，启发式概率估计更容易受启发式的易提取性抽象水平、易提取性的难度和思维特点等的影响，在多样的启发式过程中，如果非指导性反馈不能针对不同的影响因素提供有效的概率估计线索，那么它就不能发挥纠偏的作用。 四、反馈对于减少概率估计偏向的作用特点表明，反馈在纠偏方面具有条件性。这些条件中不仅包括了反馈本身的属性；就启发式各个环节而言，反馈作用只限于启发式的某个环节，或者是促进了某个环节的改善，从而在一定程度上能够起到纠偏的作用，但是条件的变化有时也会限制反馈的这种有限的促进作用。更多还原

【Abstract】 Heuristic probability estimate is one of the popular issue of natural reasoning in foreign psychological research currently. Heuristic probability estimate in uncertain condition is the important part of natural reasoning. Because of the traits of empiric and intentionality, most research on heuristic is limited to the form of theoretical hypothesis. As a transform of natural reasoning, subjective probability estimate, as long as heuristic become the new study point of heuristic probability estimate. Thus, the study paradigm of heuristic probability estimate combine heuristic theoretical hypothesis with probability estimate, so to further explore the mechanism of heuristic probability estimate especially, heuristic probability estimate break the shortcoming of traditional reasoning study which has formal logic as material and evaluation criterion, and the study emphasis has turned to probability itself in pragmatic thinking activity.Study on effect of feedback on probability estimate can well reveal how the feedback interfere with each link of heuristic, thus precisely grasp the psychological mechanism of heuristic bias. We designed several feedback conditions from the presentation order and questioning type of probability task and used two factor analysis experiments to measure the change of bias in such feedback condition: Experiment 1 was to verity the effect of order in form of presentation and feedback on probability estimate; Experiment 2 was to verify the effect of questioning type and feedback on probability estimate. The results are as follows:1. Feedback can promote the inner process of heuristic probability estimate. To the different probability question presentation style, the effect of feedback’s regulation is different. The estimate effect in frequency presentation style is better than in special cases presentation style.2. Since frequency and special cases belong to the two presentation styles of probability, they arose two kinds of thinking respectively. That is distribution thinking and single thinking which can not effect each other in heuristic probability estimate, even though in a condition of feedback.3. While the content of probability question is varied, heuristic probability estimate is familiar to be changed by some factors, such asabstract level of availability, difficulty of availability and characteristic of thinking. If non-instructional feedback can not provide the available probability estimating cues for these factors in various heuristic, it will lead a limited effect in correcting the bias.4. Feedback’s decreasing the bias of probability estimate is conditional. The conditions include the property of feedback itself. To heuristic, the feedback’s effect is limited only to some link, correcting the bias to some degree by improving some link. Otherwise, the change of conditions can also sometimes limit such positive effect of feedback.更多还原