Lab member Christopher Marcum presented findings from the Improvisation in Emergency Response Project at the Annual Hazards Workshop in Bloomfield, Colorado. This project is an NSF funded research initiative charged with understanding the relationship between improvisational and conventional behavior in uncertain situations, especially during disaster response scenarios. Our case studies span two major U.S. domestic disasters—the 1995 bombing of the Alfred P. Murrah Federal Building in downtown Oklahoma City and the 2001 terrorist attacks on the World Trade Center in New York City, two types of response agencies—police and fire, and two types of data sources—informant reports and radio transcripts. This is an extensive comparative study of responses to anthropogenic disasters. The analyses summarized here come from a variety of research methods including simple cross-tabs to complex simulations and relational event modeling. The datasets used in this research are packaged as R data objects, complete with documentation, for future public distribution.
Several lab members presented research at this year’s annual Sunbelt Conference.
Zack Almquist presented “Predicting Regional Self-Identification from Spatial Network Models” which employs large-scale spatial network simulations in order to predict regional identification in the US.
Sean Fitzhugh’s work on network robustness in the context of the World Trade Center disaster response was presented by Butts. He finds that organizations with day-to-day involvement with disaster response are less robust to loss of personnel with coordinative roles than are organizations without such involvement with disaster response.
Lorien Jasny concluded her stay in Italy as a visiting fellow at Trento University as she worked on networks of immigrant advocacy groups with Professor Mario Diani. Her Sunbelt presentation looked at structure in propositional networks of political behavior and formulated appropriate baseline models.
Emma Spiro presented research on two projects. Her work extends traditional measures of brokerage by exploring brokerage behaviour and opportunity in dynamic networks. Her work also includes a new measure of dynamic brokerage as she she demonstrated in a case study of interorganizational collaboration. She also used data from a micro-blogging service to examine changes in users’ local networks following exogenous, hazard-related events. She utilized the longitudinal nature of the data to examine seasonality as well as endogenous and exogenous variation in the pace of change of local network dynamics.
Lab PI Butts will present with assistance from lab members Zack Almquist, Sean Fitzhugh, and Emma Spiro at Duke University’s 2010 Political Networks Conferenceworkshop on Network Analysis with statnet. The workshop will be held on Wednesday, May 19th and will serve as an introduction to network analysis with practical applications in individual, organizational, and international relations. Necessary data files and instructions can be found here.
Lab PI Butts will be speaking to the Stanford University MAPSS Colloqium this Thursday, May 13, on the use of relational event models to study social interaction. This talk will feature work from the NCASD Lab on statistical models for inferring behavior patterns from relational dynamics, and will highlight our recent findings regarding communication during the WTC disaster.
The most recent issue of the Calit2 Interface Magazine features an interview with Lab PI Butts on the societal implications of social media. Recent research by Butts and Cross on the dynamics of blog networks is also highlighted. This and previous copies of Interface Magazine can be found here.
Lab PI Butts will be the featured presenter at this year’s Applied Statistics Workshop of the Southern California chapter of the American Statistical Association. The workshop will be held at RAND’s Santa Monica, CA office on 4/30/2010, and will provide a practical introduction to the statistical analysis of social network data, as well as pointers on the use of the statnet suite of network analysis tools. Registration is required.
Lab member Zack Almquist will be speaking to the UCI Population, Society, and Inequality (PSI) colloquium on April 27 on the use of R to analyze spatial and demographic data from the US Census. This talk will feature an overview of his recently released R packages for Census data management, as well as examples of how these tools can be used for practical data analysis tasks. The talk will be held in SSPB 4206 on the UCI campus, at 12:00PM; all are welcome.
Lab PI Butts will be speaking to the Frontiers of Network Science Lecture Series at Penn State University, 4/2/10, on “Modeling Complex Interactions in a Disrupted Environment: Relational Events in the World Trade Center Response.” This talk will present work from the NCASD Lab on the modeling of dynamic event systems, with particular emphasis on emergency radio communications at the WTC. Also taking place that day at UMass Amherst (Department of Sociology) will be a talk by Lab Member Ryan Acton, titled “Logging On, Going Out: Group Size Dynamics in Online/Offline Spaces.” Acton’s talk is about his current work on the study of online behavior, and the formation of offline groups via online mechanisms. If you are in either venue, drop by to see our members in action!
Two new IMBS technical reports by Lab PI Butts are now available on the IMBS web site. The first, “A Note on Generalized Edges” (MBS 10-03) introduces basic terminology and formalisms for representing complex multi-party relationships, facilitating the treatment of phenomena such as joint task performance, mental models, and actor/bystander interactions that are difficult to model using traditional network analytic approaches. The second paper, “Bayesian Meta-Analysis of Social Network Data via Conditional Uniform Graph Quantiles” (MBS 10-04) provides a family of techniques for leveraging a well-known tool for analyzing individual networks — conditional uniform graph quantiles — to draw inferences regarding populations of networks from sampled data. These methods are especially applicable to meta-analytic applications, in which one seeks to pool information from multiple case studies. Both of these papers showcase our current research on methods for the analysis of network data, with an emphasis on tools and ideas that can be applied in practical settings.
Calit2 article highlights findings from a recent paper by Lab members Butts and Cross. Published in the Journal of Social Structure, this paper shows how blog networks during the 2004 election season evolved in response to both events on the campaign trail and to the cycles of everyday life. To learn more, see the original paper and theCalit2 feature.