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，認爲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（如族羣與階級背景）來說，或是依成就地位（以教育取得爲例）而論，我們都觀察到臺灣民衆傾向於和自己社會位置相近者結婚。這種傾向若展現在族羣的屬性上就是族羣內婚的現象，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，並没有隨著時間而改變；反之，族羣內婚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 pepole in Taiwan to marry ho­mogarnously. 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.