Professor Noshir Contractor
Northwestern University, USA

27th, Nov.
4:30pm ~ 5:30pm

Room 516, Redmond Barry Building, University of Melbourne.

Recent advances in digital technologies invite consideration of organizing as a process that is accomplished by global, flexible, adaptive, and ad hoc networks that can be created, maintained, dissolved, and reconstituted with remarkable alacrity. This presentation describes a multi-theoretical multilevel (MTML) model of the socio-technical motivations for creating, maintaining, dissolving, and reconstituting knowledge and social networks.  Using examples from his ongoing research on communities involved in a wide range of activities such as Communities of Practice, disaster response, public health and massively multiplayer online games (WoW - the World of Warcraft), Contractor will present the development of a contextual
multi-theoretical multilevel model to understand the emergence of networks in 21st century organizational forms.

Dr Mark Tranmer
School of Social Science, University of Manchester, UK

28th, Nov.
1:00pm~2:00pm

Room 516, Redmond Barry Building, University of Melbourne.

Over the last 15 years I have applied statistical models to investigate a range of substantive issues in social science. In many cases the analyses used geographical datasets and involved multilevel models; an excellent model framework in which to investigate individual and area effects and to recognise the fact that people are not independent  - the population has structure. Multilevel analysis thus allows the substantive assessment of effects or relationships at different levels of the population - i.e. the individuals and the (geographical) groups. Another way of grouping individuals is via their network(s), and geographical and network effects may both occur in a particular population.
Here I consider whether some of the issues considered and lessons learnt in geographical analysis are useful in the development and interpretation of social network analysis. I also make a few suggestions for studying networks in multilevel populations.

Dr Johan Koskinen
Department of Psychology, The University of Melbourne

Wednesday the 6th of Dec.
11:00am ~ 12:00pm

Room 213, Richard Berry Building, The University of Melbourne .

We consider here a probability model for the edge set of a graph that is commonly referred to as the exponential random graph model (ERGM), and its extension, the curved ERGM. Although some issues remains to be resolved when it comes to how to specify the ERGM, this class of models holds some promise when it comes to capturing network processes. Currently the favoured methods for statistical inference are Markov chain Monte Carlo (MCMC) Maximum likelihood estimate (MLE) and an MCMC implementation of the Robbins-Monroe algorithm, both of which rely on the properties of the method of moments for exponential family distributions. We propose instead to take a Bayesian approach that (i) yields clearly defined answers in terms of probabilities (the asymptotic properties of the MLE are not fully understood in the case of the ERGM); (ii) offers a rich picture of uncertainty (the MLEs and approx. s.e.'s do not adequately reflect the uncertainty stemming from the pronounced dependencies between observations); (iii) makes allowances for penalising "degenerate parts" of the parameter space using proper subjective prior distributions; (iv) provides us with a natural and probabilistic approach for handling missing data; (v) offers a principled and probabilistic procedure for performing model selection; (vi) provides us with posterior predictive distributions; etc.
How to implement a Bayesian inference scheme for the ERGM is, however, far from straightforward. It is clear that in all but trivial cases we have to rely on numerical methods. It is probably fair to say that as far as numerical methods go, MCMC is the gold standard. Thus far, however, efforts at designing an MCMC algorithm for the ERGM has been hampered by the fact that it is typically not possible to evaluate the normalising constant (the partition function) in the likelihood function. Although the (pure) MCMC does not require that we can evaluate the normalising constant in the posterior distribution it usually requires that we can evaluate the likelihood function. Recently an auxiliary variable MCMC (SISA; our acronym) was proposed that circumvented the need to evaluate the partition function. The key being to introduce an auxiliary variable defined on the same state space as data. However, while SISA performs sufficiently well in order for it to be useful for "simpler" models like the Ising model, it seems as if it runs into serious problems when applied to the ERGM. It is not only a question of whether the mixing is good or not, rather it is a question of whether it mixes at all. The reasons for this being so are easily understood when the SISA is understood in terms of the Simple Importance Sampler (SIS). We propose a solution (LISA) where the (single) auxiliary variable is replaced by an auxiliary variable defined on an extended state space. Whereas SISA may be seen as an algorithm that performs a one-sample point SIS in each iteration of the Metropolis-Hastings sampler, LISA performs a bridged (linked) importance sampling (LIS) estimation in each iteration, with the number of bridging distributions and sample points chosen to tune mixing. The extra number of calculations necessary to perform LISA as compared to the SISA is negligible. We illustrate LISA when applied to the analysis of the Ising on a 50x50 grid and a network for a New England law firm.

Dr Dale Hanson
Senior Lecturer, School of Medicine, James Cook University
Emergency Physician, Meckay Base Hospital

Room 718, 7th Floor, Redmond Barry Building , The University of Melbourne

The network idea has penetrated the Health Sector but essentially as a metaphor for a "new" way of working. The Ottawa Charter for Health Promotion emphasises the importance of “strengthening community action” as one of 5 key practice domains necessary to promote community health and safety. If indeed, collaborating with community agents is an important vehicle for promoting health then it is imperative to utilise research methodologies able to analyse how these social systems work. This research program utilised Social Network Analysis to describe and analyse Mackay Whitsunday Safe Communities (MWSC) – a community based safety promotion coalition.Arguably, the most critical decision in any study is defining the population under investigation. This study used a snowball methodology to allow respondents to identify people they believed had an important role in promoting safety in the Mackay Community. However, there are important advantages and disadvantages of snowballing methodologies.A questionnaire regarding the nature and quality of relationships was
distributed throughout the network and analysed using UCINET software.

Social Network Analysis proved a useful tool for documenting the growth of
social capital within a community safety promotion coalition. Two
distinct forms of social capital have been documented: firstly, the growth
of network cohesion and secondly, the critical role played by a small
number of key actors who perform an important brokerage function in the
network. MWSC is clearly on open network, in which relationships with external agents are critical to network function. A closed study of MWSC alone, would not have identified the critical brokering role undertaken by the networks leaders who have a key role mobilising resources on behalf of the network. The myth that it is possible to develop sustainable community safety promotion coalitions that can become totally self sufficient after a short period of development funding must therefore be questioned on theoretic and empirical grounds.

Professor John Padgett,
Department of political science, University of Chicago

Conference Room 1206,
Redmond Barry Building, The University of Melbourne

Dr. Malcolm Alexander
School of Arts, Median and Culture, Griffith University QLD

Seminar Room 1004,
Redmond Barry Building, The University of Melbourne

Dr. Carter Butts,
Department of Sociology and the Institute of Mathematical Behavioral Sciences,
University of California Irvine

Tuesday 15th Nov. 2005
12:00pm ~ 1:00pm

Seminar Room 1005,
Redmond Barry Building, The University of Melbourne

Dr. Johan Koskinen
Stockholm University

Seminar Room 1004,
Redmond Barry Building, The University of Melbourne