Journal impact measures in bibliometric research
Journal impact measures in bibliometric research
复制标题
DOI：
10.1023/a:1014848323806
复制
发表时间：
2002-02-01
期刊：
影响因子：
3.9
通讯作者：
Moed, HF
中科院分区：
文献类型：
--
作者：
Glanzel, W;Moed, HF
尤金·加菲尔德提出的影响因子是以引文为基础的衡量科学期刊重要性和绩效的指标。它可能是在文献计量学本身以及科学界以外使用的最受欢迎的文献计量学产品。首先，简要回顾了三军情报局影响因素的背景和历史，以及它背后的基本思想。交叉引文矩阵用于可视化影响因子的构建和几个相关的期刊引文指标。讨论了影响因子的优点和缺陷。描述了不同作者为引入更复杂的期刊引用指标所做的几次尝试，以及许多旨在纠正影响因子的方法学局限性的指标不成功的原因。下一部分致力于分析基本的技术和方法论方面，在此背景下，研究了计算和使用期刊引用指标可能存在的偏差和扭曲的最重要来源。然后总结了应用语境的主要特征。最后一部分是关于期刊引文测量的统计可靠性问题。结果表明，与常见的误解相反，统计方法可以应用于离散的“偏态”分布，并且这些统计的统计可靠性可以作为比较分析中应用期刊影响度量的基础。最后，讨论了科学期刊引文影响的单一、多么复杂的衡量标准是否充分或不足的问题。
The Impact Factor introduced by Eugene Garfield is a fundamental citation-based measure for significance and performance of scientific journals. It is perhaps the most popular bibliometric product used in bibliometrics itself, as well as outside the scientific community. First, a concise review of the background and history of the ISI impact factor and the basic ideas underlying it are given. A cross-citation matrix is used to visualise the construction of the Impact Factor and several related journal citation measures. Both strengths and flaws of the impact factor are discussed. Several attempts made by different authors to introduce more sophisticated journal citation measures and the reasons why many indicators aiming at a correction of methodological limitations of the Impact Factor were not successful are described.The next section is devoted to the analysis of basic technical and methodological aspects, In this context, the most important sources of possible biases and distortions for calculation and use of journal citation measures are studied. Thereafter, main characteristics of application contexts are summarised.The last section is concerned with questions of statistical reliability of journal citation measures. It is shown that in contrast to a common misbelief statistical methods can be applied to discrete 'skewed' distributions, and that the statistical reliability of these statistics can be used as a basis for application of journal impact measures in comparative analyses. Finally, the question of sufficiency or insufficiency of a single, howsoever complex measure for characterising the citation impact of scientific journals is discussed.