睡美人与王子文献的识别方法研究

Defining and Identifying the Sleeping Beauty and Her Princes in Science

[目的/意义]研究睡美人与王子文献的识别方法。分析唤醒机制,为未来在学术交流体系中发现"王子"作者,发掘、唤醒低被引和零被引文献的潜在价值提供理论依据。[方法/过程]采用被引速率指标和睡美人指数两种客观指标识别1970-2005年临床医学四大名刊上发表的睡美人文献;基于以下4个原则寻找唤醒睡美人的王子文献:①发表于被引突增的附近年份;②本身被引次数较高;③与睡美人文献的同被引次数高;④在年度被引次数曲线上,王子文献对睡美人文献的"牵引或拉动"作用非常显著,即至少在睡美人文献引用突增的附近年份,王子文献的年度被引次数应高于睡美人文献。[结果/结论]由于考虑了全部引文窗的引文曲线,被引速率指标能够识别出那些被引生命周期长、至今仍持续不断高频被引的论文;睡美人指数能够快速识别出睡美人文献,但却无法反映年度被引次数达到峰值之后的引文曲线;将被引速率+发表最初5年年均被引次数两个指标结合起来能够更好地识别睡美人文献。分析发现,综述、指南、著作等"共识型"的文献对于引发那些提出了新思想但尚未被认可的睡美人文献的被引突增起到了关键作用。建议事后识别睡美人文献可采用客观指标与主观界定相结合的方法,事前预测睡美人文献要注意追踪其是否被"共识性"文献推荐和引用,学术评价要特别关注被引速率低的论文。

[ Purpose/significance] This paper proposes methodologies for defining and identifying the sleeping beau- ty and her princes in science. We try to provide some insights to the early prediction the potential for scarcely cited papers which may become sleeping beauties in the future. [ Method/process ] Two parameter-free indices, i. e. Citation Speed （CS） and Beauty Coefficient （B） were used to identify sleeping beauties （SB） published between 1970 and 2005 in NEJM, Lancet, BMJ and JAMA. In order to discover the papers about princes who wake up SB, we proposed that the prince candidates should （1） be published near the years when the SB began to attract a lot of citations ; （2） be highly cited papers themselves; and （3） receive substantial amount of co-citations with SB. In addition, the ＂pulling effect＂ of princes on SB should be strong enough, i.e. , the annual citation of princes is higher than that of SB, and at least on certain points in the citation curve where the annual citations of SB began to increase significantly. [ Result/conclusionl We find that be- cause CS takes the citation curve of the whole citation window into account, by CS one could identify the continuingly highly cited papers with a ＂very long＂ life cycle. Bis proved to be a good indicator to identify a SB, but it fails to cover the citation curve after the paper receives its maximum annual citation. Thus, we hold that the combination of CS and C5 （5 -year average annual citations after publication） may be a good choice to define and identify SB in the large-scale biblio- graphic dataset. Within the 10 SBs identified in our dataset, we find that reviews, guidelines, and original books may play a critical role in waking up the SB. Finally, we recommend that （1） the parameter-free metrics in combination with param- eter-based metrics could help to identify SBs ; （2） the fact that a never cited or poorly cited paper is cited by the consensus -building literature （ for instance, a paper would be regarded as a SB with a lower value of CS, because reviews, guidelines, and books） could be used as an indicator to predict to what extent ; and （3） in bibliometric research assessment, one should give special attention to papers such papers accumulate their citations slowly and thus show a longer citation impact.