本研究目的以Sonstroem and Morgan（1989）的健身運動與自尊模式為基礎，提出健身運動與自尊假設模式，以99位安養護機構老年人（平均年齡81.56 ± 7.21 歲）為研究對象，驗證以下研究假設：一、身體自我效能對身體勝任感有直接影響效果。二、身體勝任感對身體接受與身體自尊有直接的影響效果。三、身體接受對身體自尊有直接的影響效果。四、身體自尊對整體自尊有直接的影響效果。五、身體自尊對身體勝任感與整體自尊具有中介的效果。六、身體接受對身體勝任感與身體自尊具有中介的效果。本研究所使用的研究工具有：身體自我效能量表、身體勝任感量表、身體接受量表、身體自尊量表、整體自尊量表，所收集的資料以徑路分析及多元階層迴歸分析進行資料處理。研究結果顯示：一、身體自我效能對身體勝任感的直接效果為0.67，身體勝任感對身體接受的直接效果為0.61，身體勝任感對身體自尊的直接效果為0.35，身體自尊對整體自尊的直接效果為0.45。二、身體自尊在身體勝任感與整體自尊兩者關係間扮演完全中介的角色；身體接受在身體勝任感與身體自尊兩者關係間扮演部分中介的角色。本研究結論是健身運動與自尊模式在本研究中獲得支持。

The purpose of the study was to establish the Exercise and Self-Esteem Model based on Sontroem and Morgan's (1989) Exercise and Self-Esteem Model. Ninety-nine elderly persons aged above 65 years (81.56 ± 7.21 yr) from the institutional care homes participated in this study. The hypotheses of this study were to evaluate whether: 1. Physical self-efficacy could influence physical competence directly. 2. Physical competence could influence physical acceptance and physical self-esteem directly. 3. Physical acceptance could influence physical self-esteem directly. 4. Physical self-esteem could influence global self-esteem directly. 5. Physical self-esteem could mediate the relationship between physical competence and global self-esteem. 6. Physical acceptance could mediate the relationship between physical competence and physical self-esteem. The research instruments used in this study were Physical Self-Efficacy Inventory, Physical Competence Inventory, Physical Acceptance Inventory, Physical Self-Esteem Inventory and Global Self-Esteem Inventory. The collected data were analyzed by path analysis and multiple hierarchical regression analysis. The major findings were as follows: 1. The effect of physical self-efficacy on physical competence was 0.67, the effect of physical competence on physical acceptance was 0.61, the effect of physical competence on physical self-esteem was 0.35, and the effect of physical self-esteem on global self-esteem was 0.45. 2. Physical self-esteem totally mediated the relationship between physical competence and global self-esteem, where physical acceptance partially mediated the relationship between physical competence and physical self-esteem. The conclusion was that the exercise and self-esteem model was supported in this study.