期刊论文被引量与下载量关系研究

Research of Relationship Between Downloads and Citations of Journal Papers

[目的/意义]探讨论文被引量与下载量之间的关系,对论文影响力评价有重要的意义。[方法/过程]通过CNKI数据库,以《中文核心期刊要目总览》中理工农医类及经济、历史、法律、哲学类期刊2006年刊载的55 000多篇论文为基础,分析不同类论文的被引量分布特征,比较同被引量论文的下载量以及相近下载量论文的平均被引量,采用Spearman相关性方法计算不同类论文下载量、被引量的等级相关系数。[结果/结论]不同类论文的被引量分布具有一致趋势:随被引量的增多论文数比例较快地减小直至为0,相应的被引量分布可以由指数衰减函数近似函数表征。在绝对数量上,论文的下载量、被引量存在较大差别,二者之间的相关性不明显,与下载、被引的自身特征以及各类论文的文献使用特征有关。Spearman等级相关分析表明,论文层级上,下载量、被引量的等级相关性较强,某一篇论文在一定时期内的下载量可以依靠其排序位数预测其统计意义上的被引排序数,可以为论文评价提供参考。

[Purpose/significance]Discussing the relationship between citation and download is important in evaluating journal paper influence.[Method/process]Based on more than 55000 downloaded papers published in 2006 from journals of science engineering agriculture medicine economics history law philosophy on Guide to Core Journals of China(GCJC)included in CNKI the paper analyzed the citation distribution characteristics of different disciplines of papers compares the downloads of papers with the same citations and the average citations of papers with similar downloads and uses the Spearman Rank Correlation method to calculate the grade correlation coefficient of downloads and citations of different disciplines of papers.[Result/conclusion]The citation distributions of different disciplines of papers have a consistent trend:with the increase of citations the proportion of these papers decreases rapidly to 0 and the corresponding distribution of citations can be characterized by the exponential decay function approximation function.In absolute quantity there are great differences between paper downloads and citations and their correlation is not obvious they are related to their own characteristics of downloads and citations as well as the literature use characteristics of different disciplines of papers.The Spearman Rank Correlation analysis shows that at the paper rank level there is a strong correlation between downloads rank and citations rank;the downloads of some paper in a certain period can rely on its ranking number to predict the cited ordinal number in its statistical sense which can provide a reference for paper evaluation.