The purpose of this study is to investigate the possibility of using the Baum test as an effective means of assessing elementary school students' aggression levels. The subjects were 323(165 male, 158 female) elementary school students from grades 4 to 6. Measurements were the aggression scale and the Baum test. This study adapted 15 interpretation criteria, consisting of 7 total indices and 8 detailed indices. Data analysis was performed using Cohen's kappa value, cross tabulation, and discriminant analysis. The results indicated that a total of 6 indices, 2 total indices and 4 detailed indices, showed statistical significance. In the case of total indices, continuative stroke and appropriate tree size were more prevalent in the drawings by low aggressive children. In terms of detailed indices, trunk(knots), trunk(shading), and branches(severed) were more commonly observed in the drawings of low aggressive children. But leaves was more prevalent in the drawings by high aggressive children. All other indices proven to show meaningful differences in the Baum test cross tabulation were performed a discriminant analysis. The results indicated an 89.2% discriminatory power for the variables, and the Baum test characteristics of stroke, tree size, trunk knots, trunk(shading), branches(severed), and leaves were statistically significant variables. In conclusion, the Baum test can be an effective tool for assessment of elementary school students' aggression.