外省與本省族群間在平均教育成就上有相當顯著的差異。既有文獻，一方面對於外省族群身分之於教育取得是否有其獨立影響並沒有一致的答案，另一方面，大多假設族群效應為一定值，較少考慮到此影響力在不同群體身上是否有所差異。本文合併臺灣社會變遷基本調查多期資料來探究族群身分對於上大學機會是否有獨立影響力，分析其長期趨勢，並檢測省籍、性別、父親教育程度之間的交互作用。本文的主要發現有二：首先，對出生於1940、1950年代的年長世代來說，外省族群中上大學機會的兩性差距要比本省族群來得小。再者，在控制父母親教育程度、城鄉差距、父親職業地位之後，省籍效應同時依父親教育程度不同而有所變化；對高教育家庭背景的子女來說，並沒有統計證據顯示，省籍對於任何出生世代之上大學機會有影響；但在低教育家庭背景的子女中，則存在相當顯著的省籍效應--外省族群身份對一個人上大學的機會確有影響，對於出生於1940、1950年代的年長世代來說尤為顯著，在進一步控制公部門效應與手足數之後仍有影響；不過，此效應已隨時間推移而減弱，甚或消失。整體來說，族群身份對於教育成就的影響，與其所鑲嵌之社會脈絡息息相關，不僅依時代變遷，也因性別階級之不同而有所差異。

There has been a significant gap in educational attainment between mainlanders and other ethnic groups in Taiwan, especially in the early postwar years. Existing studies did not provide consistent answers about whether ethnicity has an effect on educational attainment, net of parents' education and socioeconomic background. Focusing on the probability of obtaining a bachelor degree, this study found that the mainlander effect varies not only with birth cohort also with gender and class. The findings include: first, mainlander effect was significantly stronger among females than among males in the early years due to greater gender egalitarianism within the mainlander group. Second, whereas no ethnicity effect was observed among those from highly-educated families, the probability of attending college for mainlander offspring from less-educated families was significantly higher than for other ethnic groups, even after further controlling for the "public sector" effect and the number of siblings. Ethnic difference in gender egalitarianism accounts for the early-year net effect to some degree. However, there are still some features left unexplained. Possible explanations to be explored include: language capital, social capital formed within mainlander villages, and new immigrants' atypical motivations.