基於行為評量表在臨床的需求、評量表多元評量來源的必要性、評量表發展的本 土化，本研究發展一套由教師評、同儕評與學生自評多元社會評量表，量表包括適應行為與 不適應行為兩部分，各分七個分量表，全計十四個分量表，適應行為部分包括合群、溝通技 巧、主動、尊重\互惠、衝突處理、自我效能、與學習行為等，不適應行為包括攻擊、違規 、過動\衝動、退縮\膽怯、焦慮、人際問題、與學習適應問題等。 各版本各分量表的內部一致性α值介於.761∼.944之間，間隔四週的重測信度介於.484 ∼.945，評量者間一致性在教師評版為V.456 ∼ .810，惟，同儕評版較不理想，教師評與 同儕評量表的信度較優於學生自評。在三版之內容的效度均受肯定，惟，效標效度與內部同 質方面，他評式量表(教師與同儕)較學生自評式理想。最後對有關量表的運用與工具的研究 提出建議。

　　　　　The study was to construct a social behavior rating scale (SBRS) with multiple agents, including self report, peer rating, and teacher rating. It was devided into two major parts as positive and negative behaviors, which consisted of seven subscales respectively. They are group-related skills、communication skills、activation、reciprocation、conflict-solving、self-efficacy、and learning efficacy in positive scales， and aggression、delinquent-behavior、hyperactivity- impassivity、withdrawal-timidity、anxiety、interpersonal-maladjustment、and learning-problem in the negative scales. The major finding showed as following: 1.The cronbach α coefficient of subscales ranged from .761 to .944. 2.The 4-week test-retest reliability ranged from .484 to .945. 3.The inter-teacher consistence of the SBRS ranged from .456 to .810; however, the inter-peer consistence of the SBRS were not as satisfied as the teacher form. 4.The subscale scores were reasonably correlated with each other. 5.Three forms of SBRS were found to be able to differentiate different adjustment groups. 6.The sensitivity to intervention of the teacher from of SBRS was proven.