摘 要 本文首先以最後一年公布最低錄取原始總分的1996學年之第一類組為基準年樣 本，採用其最低錄取加權總分、各科加權、高低標準設定，以及學系的其他特性估 計最低錄取原始總分，估計與實際最低錄取原始總分間的誤差大小做為選擇模型的 依據。接著，以此關係式推估1997至2001學年未公布的最低錄取原始總分及各學年 的排名，並計算以1996學年真實排名為基準的1997至2001學年估計排名的變化。最 後，分析加重計分權數是否會影響排名變化。研究結果發現，1996學年已是前段排 名的學系其未來排名的變化與加重計分無顯著關係，而1996學年已是中後段排名的 學系，加重計分反而與排名惡化有顯著正相關。

Abstract The lowest scores (LS) of student admission for all college departments in the yearly Joint College Entrance Examination (JCEE) had been published until the year 1996. Since1996, only the lowest weighted scores (LWS) have been published. JCEE data of 1996 allow us to employ LWS to estimate LS, and to select a model which has sufficient R2 as well as the lowest difference between estimated LS and true LS. This model is then used to estimate the unpublished LS for each of the years between 1997 and 2001. The estimated LS is ranked each year to calculate each department’s ranking. Finally, we examine whether magnitude of weighting affects departments’ rankings. It is found that for the top 100 departments in 1996, magnitude of weighting has no impact on their rankings between 1997 and 2001. For the remaining departments, however, magnitude of weighting deteriorates their rankings between 1997 and 2001.