|
Activity Number:
|
513
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #309812 |
|
Title:
|
Sampling Design and Inference for Social Network Data: Estimating Epidemic Potential with Contact Tracing Data
|
|
Author(s):
|
Krista Gile*+ and Mark S. Handcock
|
|
Companies:
|
University of Washington and University of Washington
|
|
Address:
|
Box 354320, Department of Statistics, Seattle, WA, 98195-4320,
|
|
Keywords:
|
survey ; social science ; random graph model ; epidemiology
|
|
Abstract:
|
Social network models are of increasing importance for applications from disease spread to social capital to national security. Numerous sampling strategies are used to gather network data for model fitting, including egocentric and link-tracing designs. In the case of reportable sexually transmitted diseases, network sampling is often conducted through "contact tracing," in which only infected individuals are sampled. This talk applies principles of statistically optimal design to sampling relational data. In particular, we study various contact tracing designs used to collect data to fit exponential family random graph models and estimate epidemic potential. We report advances in sampling and optimal design relevant to social science and epidemiological research.
|
