The purpose of the present study was to develop a standardized questionnaire to measure proactive and reactive aggression in children. In Study 1, the questionnaire was administered to fourth- through sixth-grade children (n = 700) to investigate its factor structure and internal consistency. Moreover, in order to examine reproducibility, it was administered twice, at the interval of about three weeks, to 132 of these children. Statistical analyses of the data have yielded three factors: proactive-relational aggression, reactive-expressive aggression, and reactive-inexpressive aggression; and have confirmed that the questionnaire was reliable and internally consistent. In Study 2, 14 homeroom teachers of fourth- through sixth-grade nominated children who were most characteristic (n = 106) and least characteristic (n = 99) in each of the three factors. In addition, fifth- and sixth-grade children (n = 86) assessed their same-sex classmates according to the sentences that characterized each of the three factors. These children selected were also participants in Study 1. The results showed that the scale scores were consistent with the nominations and assessments, indicating the high construct validity of the questionnaire.