確率的強化事態におけるシロネズミの選択行動

自己訂正法とガイダンス訂正法の効果

抄録

In Exp. I, a spatial 70:30 probability learning task was given in a Skinner box to each of the three groups of rats by the use of a free correction, a guidance, and a noncorrection method, respectively. Percent choices of the majority side was the highest for the noncorrection, intermediate for the guidance, and the lowest for the free correction group. Two possible interpretations, “nonreinforcement hypothesis” and “probability estimation hypothesis”, were tested in Exp. II, concerning the obtained difference in performance between the free correction and the guidance method. The results suggested that the difference would be due to the larger number of unreinforced responses made on error trials on the majority side than on the minority one only for the free correction group.