Spring School: RSiena to model longitudinal social networks

26Dec10

It may seem odd to advertise a “Spring school” in the middle of this particularly snowy and chilly winter… but better get organised in advance!

The Centre for Business Networks Analysis at the University of Greenwich in London is organising a 5-day Spring School fin Analytical Software for social scientists (both PhD students and confirmed researchers) on April 4-8 and as part of that, I will give a one-day workshop on RSiena (5 April).

RSiena is an approach to the study of longitudinal social network data, that is, repeated observations of a directed graph on a given node set. The challenge with social networks is that ties between nodes (actors, whether they are individuals, firms, organisations, or other entities) may depend both on nodes’ own attributes (their qualities as well as forms of similarity between them) and on the previous existence of other ties (with endogenous phenomena such as reciprocity, transitivity, cycles, popularity). Longitudinal observations have an advantage over cross-section data, in that they include information collected at several points in time and therefore, allow estimating the probability of observing a tie between two nodes at date t, based on ties that were observed at date t - 1. At the same time, the sheer presence of several waves of observation adds a further layer of complexity. The statistical model to analyse such data has been developed by Tom A.B. Snijders (University of Oxford and University of Groeningen) and subsequently implemented, by Tom and his collaborators, in the RSiena software, a package in R.

My workshop will closely follow Tom’s own approach and teach the theoretical model as well as the practical use of the software. Focus will be on the bases of the statistical methodology, on examples of possible applications to the social and economic sciences, and on use of the software. The morning session will provide an intuitive understanding of the statistical approach and present the basic operation of the software. The afternoon session has more applied character and will allow participants to follow some steps of data preparation and analysis hands-on. It will also briefly introduce to some more advanced models for the simultaneous dynamics of networks and behaviour and other topics such as model specification and goodness of fit checking.

I will largely rely on Tom’s own teaching material, and I thank him in advance for making it available to me.

It will be helpful for participants to know how to run R and how to give basic R commands on their machine; but further knowledge of R is not required. Participants are advised to read in advance the tutorial given in Snijders, T.A.B., Steglich, C.E.G., and van de Bunt, G.G. (2010), Introduction to actor-based models for network dynamics, Social Networks, 32, 44-60, and to consult the RSiena website.

For more information and to book your place at the Spring School, please see here.

About these ads