Evaluation of Methods for Second-Stage Variance Estimation for the Drug Abuse Warning Network Survey

• Activity Number: 527
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Rong Cai*+ and Kathy Poneleit
• Companies: Substance Abuse and Mental Health Services Administration and Substance Abuse and Mental Health Services Administration
• Keywords: second stage variance estimation ; Taylor series ; replication methods ; DAWN

Abstract

The Drug Abuse Warning Network (DAWN) was designed as a single stage probability sample of hospitals with a 100 percent second stage sample of the full daily census of all drug-related emergency department (ED) visits. Recent introduction of sub-sampling of ED visits requires accounting for second stage variance. We discuss three variance estimation methods: a Taylor series expansion of the first and the second stages, Taylor series expansion in the first stage and half sample replication in the second stage and Taylor series expansion in the first stage and a multiple replication method in the second stage. DAWN 2008 data will be used for evaluating these methods. Recommendations will consider both theoretical and practical issues for implementing these second stage variance estimation methods.