Perspectives on Spatial Analysis in the Social Sciences
University of Washington, Seattle, WA
19-23 June 2000
Workshop Coordinator:
Michael D. Ward
Professor of Political Science & Member of CSISS
University of Washington
Co-sponsor with CSISS and host institution:
The Center for Statistics and the Social Sciences
University of Washington
www.csss.washington.edu
This workshop is intended to expand awareness and knowledge of spatial methodology in the social sciences. It is intended to service primarily younger scholars in a broad range of fields in social science.
The workshop will focus on two main threads of spatially informed analysis: Markov Chain Monte Carlo analysis and Network analysis, each joined with on-going applied work in the social sciences.
The first two days of the workshop will be devoted to Markov Chain Monte Carlo analysis, led by Julian Besag, who is in large part responsible for helping to introduce this breakthrough into the statistical sciences. MCMC methods were initially developed in physics to solve a spatial problem that is very similar to one faced by many social scientists today: the dependence of observations on one another. Thus, there is a very close and precise relationship between MCMC methods broadly cast and spatial statistical methods as applied.
Calendar:
Monday, June 19: Julian Besag, Department of Statistics
9:00-10:15 Overview and introduction to statistical inference via simulation
10:30-12:00 Intoduction to Markov chains & Markov random fields
12:00-3:00 Lunch
3:00 - 4:15 Monte Carlo and Markov chain Monte Carlo; Gibbs sampler, etc.
Tuesday, June 20: Julian Besag, Department of Statistics
9:00-10:15 Bayesian & frequentist applications of MCMC I
10:30-12:00 Bayesian & frequentist applications of MCMC II
12:00-3:00 Lunch
3:00 - 4:15 More advanced topics in MCMC; challenges for the future. Applications will include topics in spatial statistics and network models
4:30-6:30 BBQ for workshop and local participants
Martina Morris will lead sessions on the third and fourth days on the use of spatial methods to analyze network data, such as used by demographers and others to model the transmission of HIV. There is a lot of progress being made in using random graph models for networks building on Besag’s work as well as the specific network application started by the work of Ove Frank (and first reported in the 1986 article on the spatial network among a small subset of monks). In this context “space” is given a less Cartesian definition and a more socially defined spatial process is employed (some details can be found at the Web site: Kentucky.psych.uiuc.edu/pstar/). Information on Martina’s recent demography workshop in Thailand is available on-line at www.pop.psu.edu/events/iussp/chiangmai.html.
Wednesday, June 21: Martina Morris, Departments of Statistics and Sociology
9:00-10:15 Network data collection
10:30-12:00 Statistical modles for local networks
12:00-3:00 Lunch
3:00- 4:15 Statistical models for complete networks
Thursday, June 22: Laura Koehly, Department of Psychology, University of Iowa and Phillipa Pattison, Department of Psychology, University of Melbourne
9:00-10:15 Statistical estimation
10:30-12:00 Examples and applications
12:00-3:00 Lunch
3:00- 4:15 Modeling spatial and temporal "neighborhoods"
The morning seminar on day five of the workshop will bring a variety of local scholars together for brief presentations of their scholarship in a spatial context. This session will highlight ongoing, unfinished research projects currently underway at CSSS.
Friday, June 23:
9:00-10:00 Suzanne Davies, Dept of Geography, "Employment Dynamics and the Occurrence of Timing of Residential Mobility.
10:00-11:00 Katherine Stovel, Dept of Sociology, "Spatial and Occupational Mobility: The impact of expansionon careers
at Lloyds Bank
11:15-12:15 TBA
12:15- 12:30 Adjournment
[TOP]
|
