CMIST Afternoon Seminar: The two-level social relations model for binary data with application to Twitter following among NBA basketball players
|Dates:||14 June 2016|
|Times:||16:00 - 16:00|
|What is it:||Seminar|
|Organiser:||School of Social Sciences|
|Who is it for:||University staff, Adults, Alumni, Current University students, General public|
|Speaker:||Dr George Leckie|
The social relations model (SRM) is a statistical model for analysing continuous measurements on social interactions among individuals within groups. The SRM has proved particularly popular in social psychology, for example in round-robin studies of interpersonal perception, but it has an as yet untapped potential to be used to study the increasing amount of dyadic, relational and network data appearing more generally in the social and behavioural sciences. In this talk, we propose extensions to SRM to handle: (1) binary measured social interactions; (2) social interactions across as well as within groups; (3) covariates to explain, not just describe variation. These extensions greatly increase the range of problems to which the SRM can be applied. We motivate and illustrate these extensions with an application to Twitter following among NBA basketball players. We model team and player variability in Twitter following and reciprocity allowing for substantial differences between players on the same and rival teams. We then relate these patterns to player and team characteristics as well as to specific characteristics of the team and player dyads which they form. We formulate all models and extensions as multilevel models and fit them using Markov chain Monte Carlo methods as implemented in the WinBUGS software.
Tea/coffee from 3.45. All welcome. No booking needed.
Dr George Leckie
Role: Senior Lecturer
Organisation: University of Bristol
Travel and Contact Information
Humanities Bridgeford Street