由於政府研發資源有限,如何做最有效的運用與管理是一個重要的問題。本文以我國農業科技計畫為例,旨在比較資料包絡分析法(DEA)與多屬性決策(MADM)績效評估之研究,分別就方法論、資料處理與評估結果進行績效評估之比較。實證結果指出,以DEA模式分析我國農業科技計畫之績效值為0.924,其中以園產品最具效率,其次依序為畜產品、水產品、農產品及林產品:另以MADM針對五種產品研發效率進行分析,其排序為畜產品、園產品、水產品、農產品及林產品。由此可發現此兩種方法應用於績效評估皆為適合之方法。兩者主要差異可歸納為:1.DEA以數學規劃方式,透過模式運算,得到較為豐富的決策資訊;2.MADM屬非函數型之模型,其計算方法較DEA簡便易行;3.實證結果顯示兩種方法之結果相近,具有應用之價值。本文研究結果對於衡量農業科技計畫績效提供一套評估方法,同時對於農改單位提供決策的參考及提升研發資源分配之決策品質。
Due to increasingly strained governmental resources, effective management and efficient utilization of government funds for R&D projects has become a critical issue. This study is conducted to compare the data envelopment analysis (DEA) and multiple attribute decision making (MADM) of measurement and their effect on the performance evaluation of R&D projects of agriculture in Taiwan. Empirical results indicate that the average efficiency of DEA is 0.924. In addition, it is found that horticulture production are, on averages , the most efficient ones among sample R&D projects of agriculture, followed by livestock production, fishery production, crop production and forestry production. MADM efficiency rankings of the five types of livestock production, horticulture production, fishery production, crop production and forestry production. In comparing these two methods, the major differences are : 1. DEA shows the greater information for decision-making, 2. MADM is much easier when it come to data input and analysis process as compared to the DEA, 3. In analyzing the result, it is observed that they both had similar ranking for performance evaluation. The result of this study will provide useful information for agriculture administration agency a set of decision-making guidelines and rational resource allocation decisions.