In the Shadow of Silence: Modelling Missing Data in the Dark Networks of Crime and Terrorists

TL;DR

This paper investigates how to model and account for missing data in covert networks like crime and terrorist organizations using exponential random graph models, highlighting implications for data collection and analysis.

Contribution

It introduces a flexible framework for modeling non-random missingness in covert networks, improving inference accuracy and understanding of hidden network structures.

Findings

Model specification impacts inference accuracy.

Different missingness mechanisms significantly affect network analysis.

Framework helps identify potential biases in covert network data.

Abstract

The clandestine nature of covert networks makes reliable data difficult to obtain and leads to concerns with missing data. We explore the use of network models to represent missingness mechanisms. Exponential random graph models provide a flexible way of parameterising departures from conventional missingness assumptions and data management practices. We demonstrate the effects of model specification, true network structure, and different not-at-random missingness mechanisms across six empirical covert networks. Our framework for modelling realistic missingness mechanisms investigates potential inferential pitfalls, evaluates decisions in collecting data, and offers the opportunity to incorporate non-random missingness into the estimation of network generating mechanisms.