高校分类科研绩效评价的模型建立和实证研究

An Empirical Study on the Modeling of Classified Evaluation of University R&D Efficiency

对2000年至2010年间某地不同类别高校科研综合投入与产出的面板数据采用Pearson简单相关分析和Logistics回归拟合,分析得出投入与产出的正相关性强的关键因子。同时,采用极大似然估计对数据分析并建立相应的科研绩效的高校分类模型,该模型较好地验证了2000年至2010年间不同类别高校的科研绩效实际;研究得出了不同类别高校科研绩效差异预测的点估计和95%Wald可信区间估计范围情况预测概率与观测值之间的关系,定量地解释了不同类别高校科研绩效的差异与发展趋势。模型对于合理评价科研绩效以及对于科研活动进行预测和调控,具有重要的现实意义。

A Pearson Simple Correlation Analysis and a Logistics Regression Fitting have been employed to analyze the panel data about the overall inputs and outputs of R＆D of various types of universities in a certain area from 2000 to 2010. The strong positive correlation of the key factors between inputs and outputs was obtained. Meanwhile, the Maxi-mum Likelihood Estimation was adopted to analyze the data for the establishment of a classified model of university R＆D efficiency of various types,which appropriately verifies the actual efficiency of R＆D from 2000 to 2010. The point esti-mation for the prediction of difference in efficiency of R＆D between universities was concluded in the paper and the re-lation between observation and prediction probability of 95% Wald confidence interval estimation was also determined. The difference and development tendency of the university R＆D efficiency of various types was explained quantification-ally. This modeling has practical significance for the rational evaluation of efficiency of R＆D as well as the prediction and regulation of scientific research activity.