国家重点实验室投入产出效率评价备受学术界和实践界关注。采用基于信息熵的理想解排序法（TOPSIS）和数据包络分析（DEA）-CCR综合评价模型评估国家重点实验室1990-2015年宏观效率变化情况。研究结果表明：两种模型分析的结论存在明显差异，前者参照理想解的临近距离得到总体呈上升趋势，反映出随着投入的增加，实验室的产出值也快速增加；后者参照投入冗余效率得到年份排序总体呈下降趋势，说明实验室的科研产出相对于增加的投入而言，仍具有很大的提升空间。比较分析后提出国家重点实验室宏观效率变化的4个阶段，以期为合理评价国立科研机构效率提供新的视角。

The efficiency evaluation on the input and output of state key laboratories has attracted much attention from academics and administration. This paper evaluates the efficiency changes of state key laboratories from 1990 to 2015 based on the information-entropy TOPSIS and DEA-CCR model comparatively. The results show that: by TOPSIS, there is an increase trend in general from the length of distance from the ideal value, meaning that the outputs has risen up quite fast compared with the added inputs; on the opposite, by DEA, the trend is on the downside, which means that there should have been much better outputs because of the rising amount of inputs. On the basis of the above models, four phases of efficiency changes of state key laboratories are proposed, in order to provide a new perspective to evaluate the efficiency of government-funded institutes.

F204

国家社会科学基金青年项目“国家实验室资源协同供给模式研究”(17CGL001)；国家留学基金委2017年国家建设高水平公派研究生项目(201706160073)