個體必須具有足夠的社會技巧才能表現出適當的社會互動行為，以幫助個體獲得他人接納以及適應環境之能力。本研究旨在探討社會技巧促進系統（Social Skills Improvement System，簡稱SSIS）的團體介入課程，輔以同儕介入，能否提升國小低年級兒童的社會技巧表現。本研究採單一受試之跨受試實驗設計，以四名國小二年級具有社會技巧訓練需求之學生為對象，每名研究對象配對一名班級同儕。本研究所稱SSIS團體介入課程共有10單元，每單元模擬真實情境中可能發生之事件，對應10個目標社會互動行為與社會技巧步驟。介入前，研究者先針對四名同儕以任務腳本充分練習使其熟練介入策略及每單元的社會技巧步驟後，始與研究對象配對一組參與SSIS團體介入課程；介入中，研究者說明與示範正確的社會技巧，配對之同儕與研究對象共同演練社會技巧步驟，提供研究對象回饋如增強、讚美或提示，以及協助研究對象修正行為。本研究以社會技巧步驟達成率檢核研究對象各階段的社會技巧表現，視覺分析四名研究對象介入前後的表現，均有臨床顯著變化，說明SSIS團體介入課程輔以同儕介入能提升低年級學童社會技巧表現，兩者間具有功能關係，以及具有維持效果和高社會效度。研究發現研究對象與配對之同儕回到原班級中鮮少互動，故研究對象缺乏動機表現出習得的社會技巧，建議未來社會技巧課程能在班級自然情境及以班級發生的真實事件施行，期能強化維持效果。

Purpose: Individuals require sufficient social skills to demonstrate appropriate social behavior in order to gain acceptance from others and to adapt to their environments. Everyone has the need to learn social skills and be accepted by others in groups. In school settings, peers inherently constitute long-term interaction companions for students. The current study investigated the effectiveness of the Social Skills Improvement System (SSIS) with peer-mediated intervention in improving young pupils' social skills. Methods: A single-case concurrent multiple-baseline-across-participants design was used to examine the effectiveness of SSIS with peer-mediated intervention in improving the social skills of four elementary school students who required social skills training. First, the researcher taught the peers how to initiate conversation with target students by modeling appropriate social skills and providing feedback and prompts. Peers who mastered these tasks participated in the SSIS sessions with the target students. This study applied 10 SSIS units to 10 target social behaviors. During the intervention, the researcher instructed and modeled correct social skills. The paired peers practiced social skills with the target students in simulated events, the researcher provided prompts or praise and corrected the target students' behavior to enhance their performance. Results/Findings: The results demonstrated a functional relationship between the SSIS with peer-mediated intervention and social skills performance. The intervention had a significant effect in that it could improve the target students' social skills performance and maintenance effect. The teachers' questionnaire responses indicated that this research had high social validity, the SSIS was effective, and learning social skills was important to young pupils. Conclusions/Implications: The SSIS with peer-mediated intervention could stimulate target students to demonstrate appropriate social behavior. Concurrently, peers' social skills performance changed when they participated in the intervention. However, when the target students and paired peers returned to the classroom during the maintenance phase, they did not demonstrate the learned social skills or interact with one another. The reason for this is that the peer-mediated intervention could not be maintained in the class; hence, the target students had little motivation to demonstrate appropriate social behavior. Accordingly, implementing the SSIS in a natural classroom setting for the whole class is recommended. Teaching social skills through real events that occur in the classroom contributes to social skills maintenance.