績效評比中多項投入與產出指標權重的決定方法

Abstract

我們在對多個獨立組織間（例如多家醫院間、多個學校間、多家銀行間）做相對績效評比的時候，往往需要訂定評比的指標，包括投入指標和產出指標，而這指標往往又是多個的。為了要讓每一個組織獲得一個最終的績效值以進行各組織間的評比或排名，往往必須要在各個指標間給定彼此之間的權重，如此才能獲得一個最終的績效值（總加權產出/總加權投入）。

各指標彼此間權重的給定方法往往牽涉到不同的層面，而且常常會涉及到人為主觀的認定，造成有權力的人決定一切，不同的權重值對評定結果往往差異甚大，如此下來往往造成評估的結果有失客觀且造成評定結果效率不佳組織的不滿與反彈，因此如何決定指標的權重就非常重要且值得研究了。

針對權重由人為主觀給定的現象，有學者提出資料包絡法（Data Envelopment Analysis 簡稱DEA）來加以克服；而資料包絡法也的確克服了這個問題，但是資料包絡法雖然克服人為給定權重的現象，可是在實際的應用上卻仍然有一些缺點存在，而使得資料包絡法的應用受到了限制與批評。

本研究提出一種各項指標權重給定的客觀方法，排除人為因素的參與，也克服資料包絡法的缺點。讓組織間績效評比的工作能夠更客觀更為各個接受評比的組織單位所接受，進而可以將所有接受評量的組織經營績效加以排名、分析。

關鍵詞：績效評比、權重、資料包絡法

When we make performance rating among multiple independent organizations（for example multiple hospitals、schools and banks）,we often need to decide the input and output indexes, and they are usually multiple. However, in order to achieve this, we also need to give weight to each index. The decision making of each weight often involves different sides, and often decided by the man who has power. However different weights get the different result, and cause the performance rating unfair. So how to decide the weights become very important and is the goal of this research. In the light of weights decided by human factor, some scholars propose Data Envelopment Analysis（DEA）to overcome it. Although DEA indeed overcomes the problems, it still has few limits and criticize. This study presents a objective method to decide the weights, excludes human’s factor and overcomes the defects of DEA. The research results show the developed method is feasible and workable.

Key words：performance rating, weight, Data Envelopment Analysis（DEA）