Network Sampling |
|
| Coordinator 1 | Dr Tom Emery (ODISSEI) |
Simple random sampling remains central to the design of most social surveys today. Simple random samples were developed using the simple tabular format of population registers and sampling frames. In recent years, greater computational capacity and more intensive linkage of administrative sources have led to a wider array of auxillary data and linkage being available within administrative data systems during sampling. At its most intensive, this has resulted in the creation of Population Scale Networks (Van der Laan et al, 2023). These population scale networks allow for network sampling techniques that have been developed in fields such as computer science to be applied in the area of social surveys (Zhang, 2021). Such techniques include random-walk methods and community detection techniques which allow for the selection of individuals to be determined by their surrounding network characteristics. These techniques allow samples to be drawn which allow for the statistical dependence between individuals to be accounted for whilst also treating them as independent sampling units (unlike in snowball sampling). These techniques are particularly useful for studying social dynamics and the potential transmission of behaviour or attitudes across populations at various scales. In this session, papers are invited which utilize population scale networks or similar approaches in the design of surveys. We encourage the submission of both theoretical and empirical work, inclusive of simulations or pilot studies and especially those papers addressing the practical challenges of fielding such surveys.