School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff, United Kingdom CF10 3AT.
J Exp Psychol Gen. 2010 Feb;139(1):138-61. doi: 10.1037/a0018210.
We propose that biases in attitude and stereotype formation might arise as a result of learned differences in the extent to which social groups have previously been predictive of behavioral or physical properties. Experiments 1 and 2 demonstrate that differences in the experienced predictiveness of groups with respect to evaluatively neutral information influence the extent to which participants later form attitudes and stereotypes about those groups. In contrast, Experiment 3 shows no influence of predictiveness when using a procedure designed to emphasize the use of higher level reasoning processes, a finding consistent with the idea that the root of the predictiveness bias is not in reasoning. Experiments 4 and 5 demonstrate that the predictiveness bias in formation of group beliefs does not depend on participants making global evaluations of groups. These results are discussed in relation to the associative mechanisms proposed by Mackintosh (1975) to explain similar phenomena in animal conditioning and associative learning.
我们提出，态度和刻板印象形成的偏见可能是由于社会群体在多大程度上先前能够预测行为或身体特征的学习差异而产生的。实验 1 和 2 表明，群体在经验上对评价中性信息的可预测性差异影响了参与者后来对这些群体形成态度和刻板印象的程度。相比之下，实验 3 表明，当使用旨在强调使用更高层次推理过程的程序时，可预测性没有影响，这一发现与这样一种观点一致，即可预测性偏差的根源不在于推理。实验 4 和 5 表明，群体信念形成中的可预测性偏差不依赖于参与者对群体进行全面评估。这些结果与麦金托什（1975）提出的联想机制有关，这些机制解释了动物条件反射和联想学习中类似的现象。
