摘要：

【目的】 针对《中国学术期刊影响因子年报》中学术期刊影响力指数(以下简称“《年报》CI”)计算方法和俞立平等提出的CI修正方法(以下简称“YCI”)存在的不足,提出一种新的CI修正方法(以下简称“NCI”)。针对图情类期刊的复合数据和他引数据分别进行实证研究,达到简化CI计算过程、改善数据分布规律和得到更有效结果的目的。【方法】 根据复合(或他引)总被引频次(Total Cites,TC)和影响因子(Impact Factor,IF)数据的实际分布规律,提出对IF和TC分别采取开平方根和3次方根预处理,求得图情类期刊的NCI值,并与YCI值、《年报》CI值及其排序结果、CSSCI来源期刊以及排序结果进行对比分析。【结果】 经预处理后的数据既近似服从正态分布规律,又具有较高的区分度,还可以消除个别期刊因TC很大而IF很小对CI的不利影响,并可舍去调整系数K,计算过程更精简。【结论】 实证研究结果表明,NCI能比YCI从IF和TC中挖掘出更多有效信息,比《年报》CI更好地平衡IF和TC的作用,并且NCI值及其排序结果更合理,区分度和适用性也更好。

关键词: 学术期刊影响力新修正指数(NCI), 影响因子, 总被引频次, 正态分布规律, 排序, 实证分析

Abstract:

[Purposes] Based on the shortcomings of the clout index (namely CI or original CI) proposed in Annual Report for Chinese Academic Journals Impact Factors(namely AR) and the revision CI (namely YCI) proposed by Yu Liping, a new modified CI (namely NCI) is established in this paper. Based on the positive research of the compound data and the cites' data without journal self-cites of the journals of the library and information science (namely JLI), the CI calculation process is simplified, the NCI distribution is improved, and the NCI value and the rank of the JLI are more reasonable. [Methods] According to the actual distribution of the total cites (TC) and impact factor (IF) of the JLI journals, the extraction of square root for the IF and the extraction of cubic root for the TC (namely pre-processed) were taken in the NCI. The NCI, the YCI, and the original CI as well as their ranks and CSSCI with their ranks were compared and analyzed. [Findings] The distribution of the pre-processed data is obviously improved and nearly obeys to normal distribution with suitable discrimination of the NCI. The NCI overcomes the shortcomings of the overlarge TC with too large number of published articles and very small IF, and neglects the adjustment coefficient K. The calculation process of the NCI is simplified. [Conclusions] The positive research results of the JLI journals show that the NCI can dig up more information from the TC and the IF than YCI, and better balance the IF and TC than the CI. The NCI and its ranks possess more reasonability, and better discrimination and applicability than the YCI and the original CI.

Key words: New modified clout index of academic journal (NCI), Impact factor (IF), Total cites (TC), Normal distribution, Rank, Positive analysis