接續駱明慶(2002)對1954-2000年台大學生組成的分析，本文使用2001-2014年台大學生的學籍資料，分析多元入學對於學生來源分布的影響。結果發現，透過對個別高中推薦學生名額的限制，「繁星計畫」和「繁星推薦」使得台大學生在高中畢業學校與學生居住縣市的集中度，以及學生居住地平均所得均下降，學生高中畢業學校的總數則明顯增加。雖然相較於「考試分發」，「個人申請」在高中畢業學校的集中度和居住地的平均所得均較高，但考量制度上先「個人申請」、後「考試分發」的時序安排，「個人申請」結果不滿意者才會再參加「考試分發」，我們無法得到「申請入學」對明星高中和高所得家庭學生較為有利的結論。此外，「個人申請」比例的逐年增加，以及「個人申請」相較於「考試分發」錄取的女性比例較低，可能導致了整體女性比例的下降，但「繁星計畫」和「繁星推薦」則有提高錄取學生女性比例的效果。

Continuing the analyses of Luoh (2002), this paper uses student records from 2001 to 2014 to investigate the effects of the multi-channel admission program on the composition of NTU students. First, by limiting the number of students from each senior high school recommended to NTU, the Stars Project reduces the concentration of students from elite high schools, diversifies the locality and reduces the average township income where students reside. The Stars Project also effectively increases the total number of senior high schools from which NTU students graduated. Although we find that ＂admission via application＂ has a higher proportion of students who graduated from elite high schools, and has higher average township income, due to the time sequence of ＂admission via application＂ and ＂admission via the joint college assigned-subject examination,＂ we still can not conclude that ＂admission via application＂ is more advantageous for students with better socioeconomic family backgrounds. In addition, the increasing proportion of ＂admission via application＂ in the multi-channel admission program, along with the fact that a smaller proportion of female student are admitted via application, may have contributed to the overall slight decrease in the proportion of females admitted to NTU. In contrast, the Stars Project increases the proportion of females admitted.