Methods of Adjusting for Misclassification in Respondent-Driven Sampling

• Activity Number: 408
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Isabelle Beaudry*+ and Krista J. Gile
• Companies: University of Massachusetts, Amherst and University of Massachusetts, Amherst
• Keywords: Respondent-Driven Sampling ; Misclassification (Measurement Error) ; Network Inference ; Misclassification Simulation Extrapolation (SIMEX MC) ; Social Networks

Abstract

Respondent-driven sampling (RDS) is a sampling method designed to study hard-to-reach populations. Beginning with a convenience sample, each participant receives a small number of coupons, which they distribute to their contacts who become eligible. RDS studies ask participants to report on the number of contacts they share with the studied population. Also, a set of characteristics is observed for each participant. The accuracy of these attributes is not considered in current prevalence estimators. However, ignoring misclassification may lead to biased estimates. In this study, we first assess the sensitivity of estimators to varying levels of misclassification across simulated social networks of varying features. The main contribution of this study is to propose a statistically principled approach to correct for the bias introduced by misclassification. Also, we apply the Misclassification Simulation Extrapolation method (SIMEX MC) to correct for this misclassification. Finally, we compare the results of the two methods and discuss the circumstances in which they may be suitable.