在本論文中，研究者認為影響受訪者跨族群交友的主要因素主要可從兩方面來看，分別為兩者連繫的同質性、連繫強度、接觸場合、社經地位差序變項，以及受訪者個人的世代、族群、教育程度與跨族群交友偏向程度。透過1997年三期三次「台灣地區社會變遷調查」社會網絡與社區組調查資料，本論文使用階層線性模型（Hierarchical Linear Models，HLM），從受訪者與他人的兩者關係所形成的連繫層次，以及受訪者個人層次來對台灣跨族群交往的現象作進一步的分析。研究結果發現，受訪者個人特質，例如受訪者教育程度、所屬族群，較兩者連繫特質，例如兩者連繫強度、認識場合，更強烈地影響了跨族群連繫形成的可能性，這樣的結果在過去由於方法上的限制，無法得到驗證，透過此研究，可對台灣族群相關網絡研究在建立跨族群連繫的可能因素上，提供了不一樣的看法與解釋。 / In this paper, I argued that what impacts the respondents constructing the cross-ethnicity ties can be treated from two perspectives. One is the tie-level, such as the homogeneity, tie strength, contact contexts, and socio-economical resources variables between two actors. The other is the individual-level, like the generations, ethnicities, levels of education, and the degrees of the preference for making cross-ethnic ties of the egos, that is, the respondents. I tested my hypotheses using the social network data in 1997, called Taiwan Social Change Survey (TSCS). Through building the hierarchical linear models, some advanced analysis could be done for understanding the cross-ethnic interaction phenomenon in Taiwan by considering the tie-level and the individual-level at the same time. The research results showed that when talking about the possibility of making cross-ethnic ties, personal characteristics, like levels of education, and ethnicities, are more important than tie characteristics, such as tie strength and meeting places of the two. However, this outcome cannot be examined in the past because of the methodological restrictions. Through this research, it provides different sights and explanations for discussing the possible reasons for constructing cross-ethnic ties in the social network researching area in Taiwan.