面板数据是截面数据和时间序列数据的组合,本文设计的一定时间段内多个企业R&D能力评价的24个指标体系就是典型的多指标面板数据.运用传统因子分析方法解决多指标面板数据因子分析问题,其效果不能完全令人满意.本文尝试借助SPSS和Excel统计工具,用改进的因子分析方法以江苏省科技成果转化专项资金资助企业为例解决这一问题.在分析中首先根据因子载荷阵提取与企业R&D能力有关的五个公共因子,然后根据2000-2007年各年参与企业数目进行8次因子分析,最后根据改进公式得出40家企业的公共因子得分和综合总得分.
The evaluation system designed here to assess enterprise R&D abilities in certain period consists of 24 indexes, which are of typical multi-index panel data that is defined as the combination of section data and time serial data. Owing to the fact that traditional factor analysis approaches can't handle with multi-index panel data, it attempts to tackle them by improved factor analysis methods as well as SPSS and Excel with the example of the enterprises imbursed by the Special Funds for Transforming Scientific and Technological Fruits in Jiangsu Province. First, five common factors are drawn from the whole, which are concerned with the enterprises R&D abilities in input, output, employees performance, management, and implementation respectively. Then, it makes factor analysis eight times separately with enterprises in accordance with specific time spans. Lastly, it resorts to an improved formula to work out the forty enterprises' common factors scores and total scores.