Quantifying HIV Transmission Risk in Sex/Drug Networks

Source: NICHD, 1R01HD041877
Active: 3/15/02-2/28/07
Investigator: Martina Morris, PI

                   Mark Handcock (University of Washington)


We have little statistical theory to guide the sampling of data from networks.  Network sampling involves two units:  nodes and links.  While this can be thought of as a multi-level sampling design, the two levels are not nested in the traditional manner.  The project we propose here seeks to make systematic use of current network data to examine the information loss under alternative sampling strategies, and to develop the statistical theory for network sampling.  The specific aims for this project are:


1.0  Conduct and empirical examination of the impact of sampling design on network ascertainment in the Colorado Springs study.

1.1  Construct subsamples from the Colorado Springs data set based on three different "pure" sampling strategies:  cross-sectional, longitudinal and link-tracing.

1.2  Conduct descriptive analyses of network properties with each subsample.

2.0  Develop the statistical theory and methods for network estimation based on partial network sampling designs.

2.1  Establish the conditions under which the sampling mechanism is ignorable.

2.2  Develop random graph network models for the sampling mechanisms employed in the Colorado Springs and Bushwick data sets.

2.3  Assess the impact of different sampling mechanisms using simulation methods.

2.4  Assess the impact of different out-of-design sampling on the estimation of network parameters.


3.0 Develop recommendations for practical and efficient sampling designs for monitoring the population dynamics of HIV transmission.