Myneni Sahiti, Cobb Nathan K, Cohen Trevor
National Center for Cognitive Informatics and Decision Making in Healthcare University of Texas School of Biomedical Informatics at Houston, TX, USA.
Stud Health Technol Inform. 2013;192:807-11.
Unhealthy behaviors increase individual health risks and are a socioeconomic burden. Harnessing social influence is perceived as fundamental for interventions to influence health-related behaviors. However, the mechanisms through which social influence occurs are poorly understood. Online social networks provide the opportunity to understand these mechanisms as they digitally archive communication between members. In this paper, we present a methodology for content-based social network analysis, combining qualitative coding, automated text analysis, and formal network analysis such that network structure is determined by the content of messages exchanged between members. We apply this approach to characterize the communication between members of QuitNet, an online social network for smoking cessation. Results indicate that the method identifies meaningful theme-based social sub-networks. Modeling social network data using this method can provide us with theme-specific insights such as the identities of opinion leaders and sub-community clusters. Implications for design of targeted social interventions are discussed.
不健康行为会增加个人健康风险，且构成社会经济负担。利用社会影响力被视为影响健康相关行为干预措施的根本所在。然而，人们对社会影响产生的机制了解甚少。在线社交网络提供了理解这些机制的契机，因为它们以数字方式存档成员之间的交流。在本文中，我们提出一种基于内容的社交网络分析方法，该方法结合了定性编码、自动文本分析和形式网络分析，从而使网络结构由成员之间交换的消息内容决定。我们应用此方法来描述戒烟在线社交网络QuitNet成员之间的交流情况。结果表明，该方法能识别出有意义的基于主题的社会子网。使用此方法对社交网络数据进行建模，可以为我们提供特定主题的见解，例如意见领袖和子社区集群的身份。本文还讨论了对针对性社会干预设计的启示。
