基于DEA方法的国家RD项目多重目标绩效评价比较.doc

Comparative evaluation of performance of national RD programs with heterogeneous objectives: A DEA approach Hakyeon Lee a, Yongtae Park a,*, Hoogon Choi Abstract The strategic importance of performance evaluation of national RD programs is highlighted as the resource allocation draws more attention in RD policy agenda. Due to the heterogeneity of national RD programs? objectives, however, it is intractably difficult to relatively evaluate multiple programs and, consequently, few studies have been conducted on the performance comparison of the RD programs. This study measures and compares the performance of national RD programs using data envelopment analysis (DEA). Since DEA allows each DMU to choose the optimal weights of inputs and outputs which maximize its efficiency, it can mirror RD programs? unique characteristics by assigning relatively high weights to the variables in which each program has strength. Every project in every RD program is evaluated together based on the DEA model for comparison of efficiency among different systems. Kruskal?Wallis test with a post hoc Mann?Whitney U test is then run to compare performance of RD programs. Two alternative approaches to incorporating the importance of variables, the AR model and output integration, are also introduced. The results are expected to provide policy implications for effectively formulating and implementing national RD programs. 1. Introduction As RD has been considered as a driving force for national competitive advantage, many countries have been raising RD investments through various national RD programs (Lee et al., 1996). Since RD investment is one of the most decisive elements in promoting scientific and technological progress (Wang and Huang, 2007), the effective use of the limited RD resources can be regarded as a prerequisite for benefiting from formulation and implementation of national RD programs. Thus, performance evaluations of RD programs need to be made so that the limited resources