A Recursive Method for Improving Network Scale-Up Estimation for Hidden Populations

• In Event: 296 - Section on Alcohol, Drugs and Tobacco Roundtables (one-hour). Table 01. Networks and Substance Use

Abstract

Abstract
Researchers are often interested in studying populations that are difficult to reach through traditional survey methods. A collection of techniques to access these populations have sprung up to provide both convenience and representative data for interested researchers. However, many of these methods are considerably more expensive and difficult to implement than a study which uses a mailed survey to random addresses. The network scale-up method provides a middle ground for researchers who wish to estimate the size of a hidden population, but lack the resources to conduct a more specialized hidden population study. Through this method it is possible to generate population estimates of a wide variety of hidden populations using traditional survey tools such as telephone or mail surveys. However, it can be difficult to ascertain how accurate the provided estimates are and identify problems with the estimation process. We suggest a way to examine network scale-up estimators in a way that allows researchers to gauge how accurate the estimation process and if specific scale-up predictors are performing particularly poorly. The method works through a recursive process of back estimation to identify poor predictors and remove them individually before rerunning the data checks. This paper discusses the network scale-up method, our recursive data check process, and discussed the influences of our process upon predictors and the final population estimates.
Keywords: Network Scale-Up Method, Hidden Populations, Social Networks

Authors

• Patrick Habecker, University of Nebraska-Lincoln

• Kirk Dombrowski, University of Nebraska-Lincoln