Abstract: Previous literature on co-authorship network mainly characterized the research network structure in some scientific fields by means of social network analysis, but rarely investigated the factors which influenced the positions of stars in co-authorship network. Based on the co-authorship network in the field of innovation network from 1990 to 2012, we investigated the effects of academic paper quality on the network positions of stars in co-authorship network. Research findings indicate that the network positions of stars are influenced not only by the quantity of academic papers, but also by the quality of academic papers. Specifically, the net citation frequency can be used to predict the size of an author’s ego-network; collaborating with geographically distant authors negatively affects the efficiency with which they can reach others in the co-authorship network; the star authors with highly specialized knowledge are more likely to take the role of mediator, i.e., specialist would be the ideal knowledge broker in a co-authorship network.

摘要： 以往关于合著网络的文献多数采用社会网络分析方法，来刻画一些学科领域科研合作网络的结构特征，但很少有研究考察合著网络中明星作者网络位置的影响因素。本文以1990~2012年“创新网络”领域研究中的合著网络为对象，考察了学术论文的质量特征对明星作者网络位置的影响。研究结果表明，明星作者的网络位置不仅受到合著论文数量的影响，也受到学术论文质量的影响。具体而言，净被引次数可以很好地预测作者自我网络规模的大小；与地理位置较远的作者合作对明星作者在合著网络中接近其他作者的效率存在显著负向影响；高度的知识专业化对明星作者发挥中介作用具有显著正向影响，即学科专家更适合作为“知识中介”。

关键词: 合著网络, 学术论文质量, 网络位置, 社会网络分析, Co-authorship network, Quality of academic papers, Network position, Social network analysis