科技评价中同行评议与指标体系关系的研究——以《泰晤士报》世界大学排名为例

 为了研究同行评议与指标体系评价结果之间的关系，本文基于《泰晤士报高等教育增刊》2007年世界大学排名数据，利用多元回归分析和Kappa一致性检验法进行了研究。研究发现，数据丰富程度影响同行评议与指标评价的一致性；在数据不足的情况下，辅以同行评议进行综合评价是较优的选择；可用回归拟合优度对数据丰富程度进行判断；同行评议也有其适用范围，权威的同行评议可以作为指标体系权重赋值以及指标选取的重要依据；在数据较为丰富的情况下可以向同行评议专家提供原始数据。

Study between peer review and multi-indicators evaluation in scientific and technology assessment

This paper analyzing peer review and multi-indicators evaluation based on multiple linear regression and kappa agreement test using the data of the 2007 times higher-QS world university rankings.The results show the data abundance affects the agreement of peer review and multi-indicators evaluation.Evaluation with peer review and multi-indicators together is a good choice while data lacking.Multiple linear regression is a good method to assess the degree of data supply and indicators choice.Excellence peer review is benchmark of indicator weight giving.Peer review can't be replaced by other methods in certain situations.Supplying original data to peer review experts is not always necessary.Multi-indicators evaluation is more stable and impersonal.