本文旨在探討我國經濟部「科技研究發展專案計畫」(簡稱科技專案)之執行效率,並探討影響執行效率之因素。本研究以經濟部科技專案執行年報民國88至93年之專案為研究對象,首先,以分析層級程序法(Analytic HierarchyProcess, AHP)計算投入與產出變數之權重,進一步運用資料包絡分析法(Data Envelopment Analysis, DEA)進行科技專案執行效率之衡量與差異分析;最後,應用Tobit迴歸分析探討影響科技專案執行效率之因素。實證結果發現:執行效率平均為0.533,有分配不均的情形,大多數DMU屬於明顯無效率單位。效率差異分析方面,「機械與航太」領域之專案執行效率相對其他產業較佳,不同年度專案比較則以「93」年較佳;Tobit迴歸分析顯示「研發人力素質」、「研發經費密度」、「專利應用率」、「成果移轉」等四項對科技專案執行效率有顯著的正向影響;敏感度分析顯示利用DEA/AHP求得科技專案之執行效率值頗具穩固性。
This study explores the executive efficiency on "Technology Development Program (TDP)", and the factors influencing it. This research aims at investigations notified in the annual reports about the TDP from 1999 to 2004. First of all, Analytic Hierarchy Process (AHP) is applied to calculate the weighted proportion of input and output variables. Data Envelopment Analysis (DEA) is further used to analyze the efficiency and divergence on TDP. Finally, the application of Tobit regression probes into factors that influence the executive efficiency on TDP. The result is that the executive efficiency on the investigations related to "machinery and aerospace" is relatively better than the other industries, while the one in "2004" is better in comparison with the other annual investigations. Besides, Tobit regression analysis shows that "Human Resource", "Research Density", "Patent Application", and "Result Transfer" have significant positive effect on the executive efficiency. In addition, sensitivity analysis of DEA/AHP indicates the executive efficiency is robust.