Reis及Shaver(1988)將親密感視為自我表露與知覺同伴回應間的動態歷程。為驗證此理論，本研究於上、下學期分別以羅徹斯特互動量表記錄參與者日常生活中自發性的社會互動十天，之後並就其互動對象以Q-分類法測量人際親密感。研究結果顯示自我表露與知覺同伴回應在橫斷面上均與關係親密感具顯著相關，其獨立效果也均達顯著水準。此結果在上、下學期均表現出一致的趨勢，顯示相關組型具有穩定性。交叉分析顯示，上學期之各項指標均對下學期之變項具顯著預測力，但關係的親密感比社會互動具有較高的穩定性。上學期之社會互動對下學期之親密感的預測力優於上學期之親密感對於下學期之社會互動的預測力。而且相對於自我表露，知覺同伴回應也傾向於有較強的效果。此結果支持理論的假設，亦即自我表露與知覺同伴回應確為影響關係親密感的重要變項，而在社會互動中發展關係知覺時，知覺同伴回應可能在親密歷程中又扮演了更為核心的角色。結論中筆者也針對本研究對研究設計與分析方法的貢獻進行討論。

In the past three decades, the debate of which factor, self-disclosure or partner responsiveness, plays the central role in the process of intimacy is still on and the dynamics among these factors remain unknown. According to Reis & Shaver (1988), however, the intimacy process begins when one person expresses personally revealing feelings or information to another person. To become intimate, the discloser must also feel understood, validated, cared for by the listener. Thus, this theory suggests that the specific combination of self-disclosure and perceived partner responsiveness plays an essential role in the development of relationship intimacy. Although the model does provide an insightful perspective, little empirical research has yet been done to test this model. The current study is design to test this theoretical model, looking at the role of self-disclosure and partner responsiveness in the process of transforming daily social interactions to relationship intimacy. Furthermore, a longitudinal follow-up was conducted, in which both relationship intimacy and social interaction data were collected both in the Fall and Spring semesters. Participants recorded their daily social interactions using the Revised Rochester Social Interaction Record and their relationship intimacy with the interaction partners were measured by a Q-sort method. Because this study was a simulated within-subject design, hierarchical within-subject regression analyses were conducted, which ruled out individual differences and permitted us to focus exclusively on relationship processes. Two regression analyses were first conducted for the two cross-sectional phases separately. In both cases, the simple effects as well as the unique effects of self-disclosure and perceived partner responsiveness were all significant. The results supported the hypothesis that both self-disclosure and perceived partner responsiveness are functional in the process of intimacy. The consistent results from both semesters also indicate that this correlation pattern was reliable. Cross-lag analysis showed that all the variables in time 1 were predictive of all the variables in time 2, indicating that the causal directions might not be linear nor unidirectional. Nonetheless, there were other trends especially worth noting. First, the highest correlation was between two relationship intimacy indices, whereas the interaction indices were relatively less reliable compared to the relationship intimacy. Second, both cross-sectional and cross-lag analyses showed that perceived partner responsiveness had relatively stronger effect sizes than did self-disclosure. The theoretical implications of these results were discussed. In addition, the contributions of the current methodology to the study of personal relationships and relationship processes were also discussed.