Understanding covert and criminal behaviour from a social network perspective is gathering increasing currency. While the standard social network paradigm assumes that network data has been collected though eliciting ties from respondents in a predefined set of individuals, covert networks pose obviously challenges in several respects. Firstly, the individuals in the network might not be known a prior. Secondly, what constitutes a relevant set of individuals and ties might be ambiguous.
Lastly, and perhaps crucially, collecting data using standard approaches presents the researcher with numerous obstacles. In this talk the key target of inference is understanding patterns of interactions of covert actors and the role of foci and settings in covert activities. We assume that data are unstructured, in the sense that they have not been collected through standard social network instruments. We assume that we have some information of personal ties between individuals, affiliations of individuals to organisations, events and locations, as well as some information about how the organisational entities and events are themselves connected. Adopting a multilevel network perspective, we simultaneously allow for leveraging these different types of interactional information and the fact that data are partial in all domains. For the former issue, we can for example utilise any tendency for individuals belonging to the same foci to have a higher propensity to be relationally tied that people in different foci and that conflict between organisations may be mirrored in enmity between their constituent members. For the latter issue, recent methodological advances for sampled and partially observed networks allow us to use information for observed parts of data to reduce uncertainty about unknown quantities. We illustrate different aspects of this framework in the context of three covert network datasets. The first is a network of Indonesian terrorist based on ethnographic accounts of the terrorist group after its demise. The second example is a large network relating to organised crime of criminals active in a Canadian province. The third is based on data collected on former combatants in the Democratic Republic of Congo, where friendship and instrumental ties as well as armed group affiliation, have been collected through a hybrid link-tracing design in the field.
Johan Koskinen joined the Department of Social Statistics at the University of Manchester in 2011 having previously worked at the Universities of Stockholm, Melbourne and Oxford. Dr Koskinen has contributed extensively to methodological development in social network analysis to enable innovative applications by several disciplines. He is one of the co-authors of the RSiena statistical network analysis package for longitudinal network analysis and a contributor to the MPnet software package, one of the most commonly used statistical software packages for network analysis. His methodological contributions are often developed in collaboration over substantive research projects with applied researchers and he is active in disseminating best practices through frequent workshops. He has also co-written two books on social network research methods aimed at practitioners. One of them a book on exponential random graph models (Cambridge University Press) that was awarded the 2016 Harrison White Book Award by the American Sociological Association. His current research concentrates on extending current statistical methodology for modelling social interaction to social networks of multiple types of nodes using data collated and collected from different sources.