Speaker
Martin Everett
Description
In a sunbelt presentation Everett and Borgatti proposed extensions to the EI index and Yule's Q to take account of overlapping group membership. Unfortunately their formulation did not give the required maximum score in the situation they predicted. In this paper we demonstrate why their suggested approach was incorrect and propose an alternative that does not have the same problem. In keeping with their original approach we propose two forms of the measure one based on time spent in activities and one based on skills or interest.
Topics | • Egonetworks and applications |
---|---|
Keywords | Homophily, overlapping groups, EI index, Yule's Q |
Primary authors
Martin Everett
Prof.
Steve Borgatti