Using modern methods for missing data analysis with the social relations model: A bridge to social network analysis

Social network analysis identifies social ties, and perceptual measures identify peer norms. The social relations model (SRM) can decompose interval-level perceptual measures among all dyads in a network into multiple person- and dyad-level components. This study demonstrates how to accommodate missing round-robin data using Bayesian data augmentation, including how to incorporate partially observed covariates as auxiliary correlates or as substantive predictors. We discuss how data augmentation opens the possibility to fit SRM to network ties (potentially without boundaries) rather than round-robin data. An illustrative application explores the relationship between sorority members’ self-reported body comparisons and perceptions of friends’ body talk.     

Bayesian analysis for partially observed network data,missing ties,attributes and actors

We consider partially observed network data as defined in Handcock and Gile (2010). More specifically we introduce an elaboration of the Bayesian data augmentation scheme of Koskinen et al. (2010) that uses the exchange algorithm (Caimo and Friel, 2011) for inference for the exponential random graph model (ERGM) where tie variables are partly observed. We illustrate the generating of posteriors and unobserved tie-variables with empirical network data where 74% of the tie variables are unobserved under the assumption that some standard assumptions hold true. One of these assumptions is that covariates are fixed and completely observed. A likely scenario is that also covariates might only be partially observed and we propose a further extension of the data augmentation algorithm for missing attributes. We provide an illustrative example of parameter inference with nearly 30% of dyads affected by missing attributes (e.g. homophily effects). The assumption that all actors are known is another assumption that is liable to be violated so that there are “covert actors”. We briefly discuss various aspects of this problem with reference to the Sageman (2004) data set on suspected terrorists. We conclude by identifying some areas in need of further research.     

Using Trellis software to enhance high-quality large-scale network data collection in the field

Trellis is a mobile platform created by the Human Nature Lab at the Yale Institute for Network Science to collect high-quality, location-aware, off-line/online, multi-lingual, multi-relational social network and behavior data in hard-to-reach communities. Respondents use Trellis to identify their social contacts by name and photograph, a procedure especially useful in low-literacy populations or in contexts where names may be similar or confusing. We use social network data collected from 1,969 adult respondents in two villages in Kenya to demonstrate Trellis’ ability to provide unprecedented metadata to monitor and report on the data collection process including artifactual variability based on surveyors, time of day, or location.     

Revisiting the accuracy problem in network analysis using a unique dataset

A series of papers published by Bernard and colleagues in the late 1970s and early 1980s, dubbed the “accuracy studies,” called into the question the validity of self-reported perceived communication in the study of networks, showing that such reports explain only about 20 % of the variance in directly observed communication. Questions remain about how well the kinds of organizations studied reflect typical formal organizations, the studies’ short observation periods, and manual observation methods. This study revisits the accuracy studies using a unique dataset comprising 144 weeks of network surveys and machine classification of 7,000 h of audio recordings to measure observable communication in a software engineering unit employing 54 people. Results show that correlations between perceived and observed communication over the weeks studied have a lower average than that reported in the accuracy studies but vary considerably from week to week. It also replicates results of earlier research showing that participants tend to overreport communication when they perceive a strong structural relationship to the alters they are rating. This study solidifies our knowledge about network self-reports using a stronger data foundation than prior research employed. Its results, along with the previous research, suggest that perceived communication is not so much a flawed measure of observable communication as it is a related, yet distinct phenomenon. This highlights the need for developments in theory and modeling that articulate the relationship between perceived and observed communication, and Network Reticulation Theory is suggested as a viable approach.