Multiple imputation for longitudinal network data

Missing data on network ties are a fundamental problem for network analysis. The biases induced by missing edge data are widely acknowledged. In this paper, we present a new method with two variants to handle missing data due to actor non-response in the framework of stochastic actor-oriented models (SAOMs). The proposed method imputes missing tie variables in the first wave either by using a Bayesian exponential random graph model (BERGMs) or a stationary SAOM and imputes missing tie variables in later waves utilizing a SAOM. The proposed method is compared to the standard SAOM missing data treatment as well as recently proposed methods. The multiple imputation procedure provided more reliable point estimates than the default treatment. The results have relevant implications for the analysis of network dynamics under missing data.

[1]  Martina Morris,et al.  Fit, Simulate and Diagnose Exponential-Family Models forNetworks , 2015 .

[2]  Alberto Caimo,et al.  Bergm: Bayesian Exponential Random Graphs in R , 2012, 1201.2770.

[3]  Christoph Stadtfeld,et al.  Interactions, Actors, and Time: Dynamic Network Actor Models for Relational Events , 2017 .

[4]  C. Steglich,et al.  Applying SIENA: An illustrative analysis of the co-evolution of adolescents’ friendship networks, taste in music, and alcohol consumption , 2006 .

[5]  Michael Schweinberger,et al.  MAXIMUM LIKELIHOOD ESTIMATION FOR SOCIAL NETWORK DYNAMICS. , 2010, The annals of applied statistics.

[6]  Garry Robins,et al.  Bayesian analysis for partially observed network data, missing ties, attributes and actors , 2013, Soc. Networks.

[7]  Mark Huisman,et al.  Imputation of Missing Network Data , 2018, Encyclopedia of Social Network Analysis and Mining. 2nd Ed..

[8]  D. Rubin.  INFERENCE AND MISSING DATA , 1975 .

[9]  Tom A B Snijders,et al.  Representing Micro–Macro Linkages by Actor-based Dynamic Network Models , 2015, Sociological methods & research.

[10]  P. West,et al.  Drifting Smoke Rings: Social Network Analysis and Markov Processes in a Longitudinal Study of Friendship Groups and Risk-taking 1 , 2003 .

[11]  Alberto Caimo,et al.  Bayesian model selection for exponential random graph models , 2012, Soc. Networks.

[12]  L. Michell,et al.  Girls, pecking order and smoking. , 1997, Social science & medicine.

[13]  T. Snijders,et al.  Estimation of Stochastic actor-oriented models for the evolution of networks by generalized method of moments , 2015 .

[14]  T. Snijders.  The statistical evaluation of social network dynamics , 2001 .

[15]  Alberto Caimo,et al.  Bayesian inference for exponential random graph models , 2010, Soc. Networks.

[16]  Carter T. Butts,et al.  4. A Relational Event Framework for Social Action , 2008 .

[17]  Tom A. B. Snijders,et al.  Stochastic Actor-Oriented Models for Network Dynamics , 2017 .

[18]  Stephen P. Borgatti,et al.  Ethical and Strategic Issues in Organizational Social Network Analysis , 2003 .

[19]  Per Block,et al.  Forms of Dependence: Comparing SAOMs and ERGMs From Basic Principles , 2019 .

[20]  Tom A. B. Snijders,et al.  Introduction to stochastic actor-based models for network dynamics , 2010, Soc. Networks.

[21]  Harold D. Green,et al.  Analytic strategies for longitudinal networks with missing data , 2017, Soc. Networks.

[22]  Mark Huisman,et al.  Treatment of non-response in longitudinal network studies , 2008, Soc. Networks.

[23]  Stef van Buuren,et al.  Flexible Imputation of Missing Data , 2012 .

[24]  C. Steglich,et al.  DYNAMIC NETWORKS AND BEHAVIOR: SEPARATING SELECTION FROM INFLUENCE: separating selection from influence , 2010 .

[25]  Carter T. Butts,et al.  Research note: The consequences of different methods for handling missing network data in stochastic actor based models , 2015, Soc. Networks.

[26]  T. Snijders.  Models for longitudinal network datain , 2005 .

[27]  Garry Robins,et al.  Analysing exponential random graph (p-star) models with missing data using Bayesian data augmentation , 2010 .

[28]  Alberto Caimo,et al.  Simultaneous modeling of initial conditions and time heterogeneity in dynamic networks: An application to Foreign Direct Investments , 2015, Network Science.

[29]  James Moody,et al.  Network sampling coverage II: The effect of non-random missing data on network measurement , 2017, Soc. Networks.

[30]  Garry Robins,et al.  Exponential Random Graph Models for Social Networks: Index , 2012 .

[31]  James Hollway,et al.  Dynamic Network Actor Models: Investigating Coordination Ties through Time , 2017 .

[32]  T. Snijders.  Stochastic actor-oriented models for network change , 1996 .

[33]  Roger A. Sugden,et al.  Multiple Imputation for Nonresponse in Surveys , 1988 .

[34]  T. Snijders,et al.  Bayesian inference for dynamic social network data , 2007 .