网络多样性视角下分裂断层对子群极化的影响研究

摘要

在网络多样性与子群相关研究基础上,将分裂断层理论引入技术创新网络,从身份、地理和知识三个多样性属性分析分裂断层在技术创新网络中的具体表现.利用2000年至2013年我国电子信息产业联合申请专利为样本构建组织间技术创新网络,运用多元分层回归方法实证检验分裂断层、成员异质性与位置嵌入性对子群结构的共同影响.研究结果表明:由多个不同成员构成的技术创新网络,节点属性聚合能产生潜在的分裂断层,但分裂断层对子群极化的直接影响并不显著;然而,在控制成员异质性或位置嵌入性的共同效应后,分裂断层的直接作用变得显著,并且当成员异质性程度适中时,分裂断层对子群极化的影响比成员异质性很低或很高时更强;此外,当位置嵌入离散性程度较高时,分裂断层对子群极化的作用显著提升.研究结论为理解技术创新网络结构变化的微观过程提供了新的理论视角,也对探索网络子群形成和发展的内在机理具有指导意义.%Based on previous studies of network diversity and subgroup,we extend divisive faultlines from groups of individuals to technological innovation network,analyze the concrete forms of divisive faultlines in technology innovation network from three diversity attributes:identity,geography and knowledge.Then we use joint application patents data in Chinese electronic information industry from 2000 to 2013 to build inter-organizational technology innovation network,and test the impact of divisive faultlines,actors' heterogeneity and positional embeddedness on subgroup polarization conjointly through multiple hierarchical regression method.Results show that potential divisive faultlines will appear when the node attributes aggregation in technology innovation network is composed of multiple different actors,but its direct impact on subgroup polarization is not significant;However,when the common effect of actors' heterogeneity or positional embeddedness is controlled,the direct effects become significant,and the divisive faultlines will have a bigger impact on subgroup polarization with moderately actors' heterogeneity than with weak or very strong actors' heterogeneity.In addition,the effect of divisive faultlines on subgroup polarization is increased significantly when positional embeddedness dispersion are at high levels,which demonstrates potential divisive faultlines will be triggered by some key mechanisms such as positional embeddedness dispersion.This research conclusion provides a new research perspective for understanding the micro process of technological innovation network structure change,and has guiding significance for exploring the inner mechanism of network subgroup formation and development too.