关系研究的新取向:社会网络分析  

A New Approach of Relation Research: Social Network Analysis

作　　者：徐伟[1] 陈光辉[1] 曾玉[1] 张文新[1] 

机构地区：[1]山东师范大学心理学院,济南250014

出　　处：《心理科学》2011年第2期499-504,共6页Journal of Psychological Science

基　　金：全国教育科学"十一五"规划教育部重点课题(DBA070073);国家自然科学基金项目(30970905);教育部人文社会科学重点研究基地重大招标项目(08JJDXLX270);"十一五"强化建设重点学科(发展与教育心理学)建设经费的资助

摘　　要：社会网络分析(Social Network Analysis,SNA)是用社会实体之间的关系来描述、解释和预测社会现象的一种研究取向。SNA提供了一种深入探究社会环境特征及其对个体心理发展影响的方法。本文基于SNA的发展历程,依次介绍了中心性分析、小团体分析、位置分析、QAP以及统计模型法。SNA在社会学研究中得到了较多应用,近年来在心理学研究中开始受到重视。Our contacts with other people can shape our view of the world,reinforce our identity,and the interactions provides us with all kinds of opportunities and resources to get things done.Traditional research mainly uses attribute data which reflect an individual＇s attitude,viewpoints and behaviors to study the relation.Social network analysis（SNA） is a new approach which focuses on the connections among social entities,especially the relational links and structures neglected in the traditional relation research.It argues that social context contains social ties and connections,not the total of different units simply.SNA uses relational data rather than attribute data to study the relation.Relational data are the contacts,ties,connections which relate one person to another and can＇t be reduced to the properties of the individual themselves.SNA provides the means for deriving a more complete view of a given social environment.This paper reviewed some functions of social network analysis in terms of psychology.Fundamental functions of social network analysis include centrality analysis which can reflect one＇s status and power,clique analysis which can assign individuals to subgroups,position analysis which can find the similar status or power＇s sets,and QAP which is the method of correlation and regression analysis to explain the relationships of different relational matrixes.Furthermore,SNA can also analyze the relationships between attribute data and relational data by the statistical models.Until now,the models mainly includes p1 model（log-linear model） which allows us to detect the dyadic directed relation,p2 model（random effects model） which allows us to study the attribute covariates of nominators and their targets between dyads,and p＊ model（exponential random graph model） which allows us to study network structure and the attribute covariates.UCINET and StOCNET are the two main kinds of software to execute these analyses.Social network analysis provides tools that can be used to answer

关 键 词：社会网络分析 关系数据 中心性分析 小团体分析 p2模型 p＊模型 

分 类 号：B842[哲学宗教—基础心理学]