Multilevel models for social networks: Hierarchical Bayesian approaches to exponential random graph modeling
详细信息 查看全文
- 作者：Andrew J. Slaughtera ; andrew.j.slaughter.civ@mail.mil" class="auth_mail" title="E-mail the corresponding author ; Laura M. Koehlyb ; 1 ; koehlyl@mail.nih.gov" class="auth_mail" title="E-mail the corresponding author
- 关键词：Multilevel networks ; Hierarchical models ; Exponential random graph models ; Autoregressive models
- 刊名：Social Networks
- 出版年：2016
- 出版时间：January 2016
- 年：2016
- 卷：44
- 期：Complete
- 页码：334-345
- 全文大小：795 K
文摘
- •
We develop fully Bayesian ERGMs for samples of networks and networks of networks.
- •
We demonstrate the effects of shrinkage and partial pooling for ERGM estimates.
- •
A fully Bayesian approach provides better fit than post-hoc meta-analytic methods.
- •
Results from post-hoc meta-analytic methods may be unreliable.