摘要
基于DEA对我国对高校投入产出绩效评价主要是利用经典DEA模型对高校相关科研、教学等活动以院系或学校进行计算和分析。然而,经典DEA模型采用Pareto-Koopmans的序关系,这种序关系下,每一个投入,产出的指标都被视做同等重要,不可相互补偿。其次,经典DEA模型还存在若评价单元有限的条件下,选取过多评价指标会出现区分度下降的问题。而在我国现实科研活动中,高校绩效评估体系已演变为一个复杂的、多目标的绩效评估体系,简单套用经典DEA模型容易导致评价结果偏差问题。本文针对我国高校覆盖自然科学、人文社会科学多个研究领域,不同学科有其自身学科特点,简单将高校各项数据加和进行整体科技投入绩效的分析存在忽略学科之间差异性和不可比性等问题,建构了多层次分类RD-DEA模型,并以教育部直属高校为例展开实证分析。评价结果显示,RD-DEA模型在允许高校突出显示自然科学和人文社会科学特点基础上,能够较好地体现评价中价值导向,同时还能兼顾现实高校科技评价多层次指标特点基础上依旧具有较好地辨识度,评价结果相对更加客观。
Performance evaluation based on DEA of higher education institutions in China mainly calculates and analyses the teaching and research efficiencies of universities or departments with classical DEA models. However classical DEA models are based on Pareto- Koopmans order which assumes that every input and output is equally important and cannot compensate mutually. Secondly,the discrimination power of classical DEA models will decrease if there are too many inputs and outputs indicators for limited number of DMUs. University performance evaluation system has evolved into a complex,multi- objective performance evaluation system. Simply applying classical DEA models will lead to deviation problem of results. In this paper,new RD- DEA model is developed,which can present different characteristics of nature science and social science and value judgment will be well adopted. An evaluation of 40 higher education institutions' panel data is presented as a case study. Evaluation results demonstrate RD- DEA model can reflect the value orientation and better and highlight the strength of different subjects. RD- DEA model has strong discrimination power and the evaluation results are more objective.
出处
《科研管理》
CSSCI
北大核心
2013年第S1期221-229,289,共10页
Science Research Management
关键词
DEA
多层次分类
高校科技评价
DEA
multi-level classification
research evaluation of higher education institutions
