数网环境下论文引证文献数的多变量动态分析

Multivariate Dynamic Analysis of Citation Frequency in the Digital Network Environment

在数字化网络传播环境下,学术期刊论文被引情况的研究有了新的数据来源。影响因子的计算面临新的挑战,许多学者以单学科领域、单因素或传统因素、静态等方式展开了研究。文章以某综合性科学技术类期刊论文的数据为最小记录单位,研究传统变量和网络出版的特征变量对引证文献数的影响,针对变量的类型不同分别采用方差分析与回归分析方法,得到多元变量与引证文献数的关系。结果表明,对论文的引证文献数影响最为显著的变量是学科、下载数、同被引文献数。下载数与同被引文献数在评价论文的学术影响力时有着良好的稳定性与合理性。此外,引证文献数与发表年限有着密切的关系,且在5～6年进入稳态期,具有显著的动态特征。

In the digital network environment, there are new data sources for the research on the citation of academic journal papers, and the calculation of influence factors have been facing new challenges. Many scholars have paid attention to single subject field, single factor or traditional factor, static state, etc. Selecting the data of comprehensive scientific and technical journal papers as the minimum unit of records, and adopting both methods of variance analysis and regression analysis, this paper tries to make clear the influence of traditional variables and online publishing characteristic variables on citations, especially the relationship between multiple variables and citation frequency. It turns out that disciplines, the quantity of downloads, and the quantity of co-cited references are the most significant variables influencing citation frequency, and the latter two variables show high stability and rationality in evaluating the academic influence of papers. In addition, the citation frequency is closely related to how many years a paper has been published, and it shows an obviously dynamic feature though it might enter a steady state after 5-6 years.