就 SCI 和中国两家数据库(CSCD、CSTPCD)共同收录的中国科技期刊的 IF 值分年份做了多种回归分析。虽然 SCI与中国两家数据库用作信息源的期刊品种和总数都相差极大,但对它们统计得出的 IF 值所做的线性回归分析的相关系数仍多次达到统计上显著或极显著水平。从而可以预期,SCI 尚未收录的大量中国期刊中 IF 值较高者,在以 SCI 收录的期刊为信息源时其 IF 值也不会很低,这支持 SCI 对中国科技期刊的收录就 IF 值而言择优不明显的结论。 Several regression analyzes were conducted on the IF value of Chinese sci-tech journals, which are collected together by SCI and China’s two databases (CSCD and CSTPCD). Although SCI and China’s two databases have very different types and total number of journals used as sources of information, the correlation coefficients of linear regression analysis on their calculated IF values ​​still reached statistically significant or extremely significant levels many times . Thus, it is expected that the higher IFs in a large number of Chinese journals not yet included in SCI will not have very low IFs when they are used as sources of information in SCI journals, which supports the IF value of SCI journals in China Promise is not obvious conclusion.