我們使用2007－14年間入學之台大學生的學籍資料、入學考試成績和在學成績，探討不同管道學生入學後的學業表現差異，並藉此評估多元入學政策在台大實施的成效。我們發現，以個人申請管道入學學生的學業成績雖然顯著優於透過指考入學的學生，但學測成績的差異即可解釋兩者入學後學業表現的大部分差異；控制學測成績後，申請生和指考生的學業成績並沒有顯著差異。而在轉系與退學的面向，申請、繁星生無論在轉系、成績因素退學和主動退學的比例，皆明顯低於指考生。在繁星計畫／繁星推薦方面，繁星生平均的學業表現優於申請生和指考生。這不只是來自明星高中繁星生的貢獻，非明星高中繁星生的學業表現平均而言也顯著高於指考生。然而如果只看來自2011－15年內無人以申請或指考進入台大的高中的繁星生，其學業成績則顯著低於指考生。此一差異也可以用學測成績的差異來解釋；控制學測成績後，這些繁星生和指考生的學業成績並沒有顯著差異。

At the turn of this century, National Taiwan University (NTU) started to admit undergraduate students via different channels. Top ranked students in high school are considered by the Star Program. Students interested in the Application Program are first screened by their GSAT (General Scholastic Ability Test) scores, and those who pass this screening will be further evaluated by the departments. Students not admitted by these two programs take a national joint entrance examination (Exam) and will be matched with departments of various universities based on their scores. This study compares college performance of students who enter NTU via different channels. We find that students admitted by the Application Program perform better than those admitted by the Exam Program. However, this performance difference could be explained by the difference in their GSAT scores. Star students, on average, perform better than other students. Once we focus on high schools that have no graduates entering NTU via the Application or Exam Program, we find their Star students have lower college scores than other students. These Star students, usually from backward places, though lagging behind their classmates in GSAT scores, exhibit good upward mobility in their class ranks.