Workshop: Method Selection and Adaptation

Method Selection and Adaptation
Presenter:  Ulrik Brandes

Which centrality measure should I use? Why are there no interesting regular equivalences? Is modularity clustering appropriate for my data? Or should I use this fancy new machine learning approach that everyone keeps recommending? And how do I defend my choices when reviewers disagree?

This workshop will introduce the pivotal notion of network position, i.e., how an actor relates to everyone else in the network, as a means to inform method selection by theory and context. The approach reveals tacit assumptions in commonly used methods, and facilitates adaptation without requiring a degree in mathematics.

Reframing network analysis in positional terms is a different way of thinking about social structure aiming to reduce the gap between methods and theory. As plenty examples will show, it applies to social structures and personal networks alike, and supports mixed-method approaches.