A Study On Simple Heuristics Strategy Of Feature Predicting When Categorization Is Uncertain

Category research is an important part of cognitive psychology .Knowledge about categ- ories plays an important role in people's cognitive activities. Especially reaches on feature pr- edicting when categorization is uncertain, have great significance.At the present, there are two theories for features predicting: single-category theory and rational model. Both theories agree that when categorization is uncertain, people will choose the most likely category to be a target category for a new object. The difference between the two theories is that single-category theory agrees people only pay attention to the feature properties in target category, while rational model believes that people are not only concerned with the feature properties of the target category, but also the feature properties of non-target categories. Feature predicting is a result of studying information about all categories.Both theories believe that feature predicting is the category-based induction. Thus, rese- archers differ with factors affect people's feature predicting. Murphy and Ross investigated the role of representativeness and diagnostic characteristics of the feature in 2005.Their resul- ts showed that representativeness impacted feature predicting, but diagnostic did not.However, Zhang Juan and Mo Lei obtained different results that the feature proportion of all the memb- ers in categories did not work in feature predicting.We believe that in Murphy and Ross's study 2005, the direction of objective representa- tiveness and category representativeness is the same, so the roles on feature predicting can not be separated. In this study, we us similar geometry materials to Murphy and Ross in 2005 .In our two experiments, fixed factors are objective representativeness and category representa- tiveness. Each of them has two levels: a balanced and an unbalanced level, combining into fo- ur conditions. So in our study, we investigate cases including direction consistent and discord. We use cognitive psychology and eye movement research method to explore the role and the mechanism of two kinds of representativeness .One is the feature proportion of the objective members in target category, called objective representativeness. The other is the feature prop- ortion of all members in target category, called category representativeness. Experiment One explores which representativeness people's reasoning based on under four conditions. In Eye movement experiment, we do further exploration on the role mechanism.The results show that: when categorization is uncertain, the target categories of informa- tion are concerned more during feature predicting. Reasoning process follow the simple heur- istic strategy. People will first choose the representativeness which can provide information f- or judgment. When the representativeness information is not sufficient to judge, people still d- on't consider non-target categories, a representative will be randomly selected as a reasoning foundation. This indicates that two feature predicting patterns exist: reasoning by objective re- presentativeness and by category representativeness.