抄録
The purpose of this study was to examine the usefulness of vector representations, diagrams constructed of geometric vectors, in statistics training. In 2 experiments, the effect of using the vector representations was compared with the conventional, algebraic method. In the preliminary experiment, 4 graduate students solved a problem about the regression analysis in both of the two ways. Only one student was able to solve the problem by using algebraic method. All the students succeeded in solving the problem with the vector representations. In the second experiment, 15 undergraduate students learned a text material on the regression analysis and solved three problems about this analysis. The vector representations sometimes helped the students learn the text material and solve the problems. However, these representations were not always useful. Three properties which made the vector representations useful were found: reducing the number of solution steps, easy and concrete manipulations, and a small number of operators. It is argued that students can easily learn statistics only when these properties help them.
