特征捆绑的3 种多项式模型比较——位置不确定理论

正确捆绑的概率随着靶子与非靶子间距离的增加而增加，亦即错误的特征捆绑是由对特征位置知觉的错误造成的，这是位置不确定理论的一个重要观点．通过对特征捆绑的3种多项式模型对比，我们认为，位置不确定模型可以排除错觉性结合实验中的猜测成分，用数学参数精确地估计真正错觉性结合发生的概率，对错觉性结合的测量效果最佳．它不仅修正了传统的错觉性结合数据分析方法的不足，同时也是对特征整合理论的重要补充．

A SET OF MULTINOMIAL MODELS OF FEATURE BINDING--LOCATION UNCERTAINTY THEORY

A critical issue of location uncertainty theory was that the probability of correct binding would increase with the distance between items, i.e., feature binding errors occurred due to the incorrect perception of feature location. Compared among three multinomial models of feature binding, the authors thought that the model of location uncertainty permitted an accurate estimation of psychological parameters for the probability of the true illusory conjunction by eliminating the guessing in the multinomial models. The measuring effect for the illusory conjunction was the best among three models. It not only modified the shortage of traditional method of data analysis, but also was an important supplement of feature integration theory.