Abstract

A stage model was developed as a conceptual framework to explore limitations in the kindergartners' causal attribution. It was assumed in this model that a prevailing cause of a particular behavior is inferred through a series of stages: Interpretation of relevant information, comparison of actors with different causes, and selection of a particular actor. Predictions generated from this model were tested in the Smith paradigm (1975), using kindergartners and primary school children. Results obtained, in particular, patterns of interaction among variables indicated that perception of reward and justification of order influence the interpretation stage whereas type of a person who suggests an actor to do a certain thing affects the comparison stage. A decison made as to who is really a good actor (child) also seems to affect the latter stage. Furthermore, more than 70% of the kindergartners employed in the present study showed the discounting effect. Contrary to the results of the previous studies, only less than 5% of the kindergartners revealed the assimilation effect. These findings were obtained in three experiments with different stimulus stories as well as child subjects differing in socioeconomic backgrounds.