What we do

We investigate the structure and dynamics of social networks, with a joint focus on method and application. We work on the development of cutting-edge statistical models for social network analysis, notably exponential random graph models (ERGM). These network models enable us to understand critical issues of importance to organisations and communities, such as innovation, trust and culture. We key areas of academic research are ERGM, networks and organisations, multilevel networks, and network dynamics.

Our major current research interests include:

Exponential Random graph models (ERGM)

ERGMs are prominent statistical models for social network structures. They take into account important endogenous structural effects such as network closure, degree centralisation, and reciprocity; and actor-relation (attribute) effects such as homophilly, and sender and receiver effects.

We also develop autologistic actor attribute models (ALAAMs), which can predict actor attributes from network structure and so model effects such as contagion and social influence. The theoretical implications of ERGMs are also an element of this work.

Publications

• Lusher, D., Koskinen, J., & Robins, G. (Eds.). (2013). Exponential Random Graph Models for Social Networks: Theory, Methods and Applications. New York: Cambridge University Press.

• Gallagher, H. C., & Robins, G. (2015). Network statistical models for language learning contexts: Exponential random graph models and willingness to communicate. Language Learning, 65 (4): 929-962.

Multilevel networks

A two-level network involves nodes at two different levels with different types of ties within and between the levels. This data structure can be used to represent many social systems that involve both hierarchy/level and networks, particularly organisations but also novel applications such as social-ecological systems. We have developed ERGMs for multilevel networks.

Publications

• Wang, P., Robins, G., Pattison, P., & Lazega, E. (2013). Exponential random graph modules for multilevel networks. Social Networks, 35(1), 96-115

• Brailly J., Favre G., Chatellet J., Lazega E., 2015, Embeddedness as a Multilevel Problem. A Case Study in Economic Sociology. Social Networks

• Brennecke, J. & Rank, O.N. (forthcoming). The interplay between formal project memberships and informal advice seeking in knowledge-intensive firms: A multilevel network approach. Social Networks, dx.doi.org/10.1016/j.socnet.2015.02.00

Big data networks

We are investigating inference for big data using parallel estimation of multiple snowball samples, in conjunction with colleagues from the University Svizzeria Italia (University of Lugano, Switzerland) and Northwestern University.

Publications

• Pattison, P., Robins, G., Snijders, T. & Wang, P. (2013). Conditional estimatation of exponential random graph models from snowball and other sampling designs. Journal of Mathematical Psychology, 57, 284-296

• Stivala, A., Koskinen, J., Rolls, D., Wang, P., & Robins, G. (2016). Snowball sampling for estimating exponential random graph models for large networks. Social Networks, 47, 167-188.

Networked innovation

Innovation no longer belongs to stand-alone corporate or government research and development (R&D) laboratories. It is the property of networks, where innovation occurs at the interstices of organisations, large and small, public and private, and the individuals nested within. These networks operate at intra- and inter-organisational, regional, national and international levels. Our research partners include the Commonwealth Scientific and Industrial Research Organisation (CSIRO), the Boeing Company, AusBiotech and the Australian Football League.

Publications

• Brennecke, J., Rank, O. N. (forthcoming): The interplay between formal project memberships and informal advice seeking in knowledge-intensive firms: A multilevel network approach. Social Networks,dx.doi.org/10.1016/j.socnet.2015.02.004

• Lomi, A., Lusher, D., Pattison, P., Robins, G. L. (2014). The focused organization of advice relations: A case study of boundary-crossing ties in a multi-unit business group. Organization Science, 25(2).

• Lusher, D., Robins, G, Pattison, P., Lomi, A. (2012). "Trust Me": Social Mechanisms for Expressed and Perceived Trust in an Organization. Social Networks, 34, 410-424

• Gilding, M. (2008). 'The tyranny of distance': biotechnology networks and clusters in the Antipodes. Research Policy, Vol. 37, no. 6-7 (Jul 2008), pp. 1132-1144.

Longitudinal and Dynamic Network Analysis 

For network data that have been collected at several time-points we may investigate the ways in which networks change. We develop and apply a range of different methods for longitudinal analysis. An example is the Longitudinal ERGM that can be estimated both using Method of Moments in LPNet or Bayesian inference. In addition we apply and develop extensions to Tom Snijders’ stochastic actor-oriented models (SAOM) implemented in RSiena. SAOM allow for simultaneous study of changes in behaviour and in social network ties.

Publications

Application of Bayesian SAOM to data messy data

• Bright, D , Koskinen, J., Malm, A. (2018). Illicit network dynamics: The formation and evolution of a drug trafficking network, Journal of Quantitative Criminology

A two-mode extension of SAOM

• Koskinen, J. & Edling, C. (2012). Modelling the evolution of a bipartite network—Peer referral in interlocking directorates. Social Networks, Vol. 34 (3), 309–322.

Application of an actor-oriented approach to housing moves

• Koskinen, J., Mueller, T., Grund, T. (2017). A dynamic discrete-choice model for movement flows, Pp: 107-117 in Perna, C., Pratesi, M. & Ruiz-Gazen, A. (eds.), Studies in Theoretical and Applied Statistics. Springer

SAOM for Ego-network data (Chapter 7)

• Crossley N., Bellotti, E., Edwards, G., Everett, M., Koskinen, J., & Tranmer, M. (2015). Social Network Analysis for Ego-Nets. SAGE, London

Application of LPNet

• Igarashi, T.: Longitudinal changes in face-to-face and text message-mediated friendship networks. In Lusher, D., Koskinen, J.H., Robins, G.E. (eds.) Exponential Random Graph Models for Social Networks: Theory, Methods and Applications, pp. 248–259. Cambridge University Press, New York (2013).

Bayesian Longitudinal ERGM

• Koskinen J., Caimo, A., & Lomi, A. (2015). Simultaneous modeling of initial conditions and time heterogeneity in dynamic networks: An application to Foreign Direct Investments. Network Science, 3(1): 58-77.

• Koskinen, J., and Lomi, A. (2013). The Local Structure of Globalization: The Network Dynamics of Foreign Direct Investments in the International Electricity Industry. Journal of Statistical Physics. Vol. 151, (3), 523-548.

The difference between continuous-time models and discrete-time models

• Block, P., Koskinen, Stadtfeld, C. J., Hollway, J., Steglich, C. (2018). Change we can believe in: Comparing Longitudinal Network Models on Consistency, Interpretability and Predictive Power, Social Networks, 52: 180-191.

SAOM workshops at ASNAC Adelaide 2020

Koskinen will be giving two training workshops at ASNAC, Wed 25 November 2020 (https://www.ansna.org.au/preconference-workshops).

Workshop 2: 11:30pm-1pm AWST (2:30pm-4:00pm AEST)

Hands-on introduction to SAOMs for newbies.

Workshop 5: 1pm-2:30pm AWST (4:00pm-5.30pm AEST)

A review of advanced use of SAOM.

Photo credit: aeruginosa via Foter.com / CC BY