在社會階層與社會流動的文獻上，有兩個基本假設主導著婚姻配對的討論。第一 種假設是同質性地位通婚，主張人類社會裡最普遍也最盛行的配對方式是社會地位相近者彼 此聯婚。第二種假設是「男高女低」的配對方式，認為女性傾向於嫁給社會地位比自己高的 男性。 本文使用一些最近的全臺問卷調查資料，利用 log-multivariate models 的統計分 析方法，初步探討臺灣的婚姻配對模式，分析結果可以支持「同質性地位通婚」的假設。不 論是就賦予地位（如族�暺P階級背景）來說，或是依成就地位（以教育取得為例）而論，我 們都觀察到臺灣民眾傾向於和自己社會位置相近者結婚。這種傾向若展現在族�撉瘧搣吨W就 是族�擗滷B的現象，若流露在出身背景的考量上就是「門當戶對」的選擇。不過，同類相聚 也不是人類配對現象唯一的法則。當異質性地位通婚發生時，我們發現「男高女低」的原則 可以充分解釋選擇性的教育配對，但不能完全說明階級之間婚姻流動的現象。另一方面，我 們發現一旦控制住人口年輪在屬性分配上的差異後，「門當戶對」在配對選擇上的重要性相 當穩定，並沒有隨著時間而改變；反之，族�擗滷B傾向則有減弱的趨勢。

　　　　　"Like marries like" is a long standing hypothesis in sociological inquiries into mate selection. In practice, the prevalence of homogamy with respect to a variety of social attributes, such as race, ethnicity, religion, and socioeconomic achievements, has been observed in several countries. This paper explores trends and patterns of assortative mating in postwar Taiwan, with particular attention paid to the tendency toward status homogamy. Marriage choice is a multidimensional phenomenon. A distinction is made between ascriptive and achievement-oriented characteristics in marriage selection. Ascriptive status homogamy is measured by the similarity of spouses with respect to their social origins, such as ethnicity and father's class position. The achieved dimension of status homogamy is evaluated by the similarity of spouses' educational attainments. All of these aspects of status homogamy are empirically examined by applying log-multivariate models to cohort data on intermarriage collected from recent islandwide social surveys. The empirical results indicate a significant tendency for people in Taiwan to marry homogamously. Besides, assortative mating patterns has remained stable regardless of class background or educational attainment. On the other hand, ethnic endogamy has decreased across cohorts examined.