Assuming 'e'

• Activity Number: 313
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Ashley Amaya*+ and Gary Euler
• Companies: NORC and National Center for Immunization and Respiratory Diseases
• Keywords: resolution ; response rate ; National Immunization Survey ; address-based sampling

Abstract

A response rate is defined as the total number of participating units divided by the total number of eligible units in the sample. However, researchers often do not know the total number of eligible units in the sample as many sampled units go unresolved. Researchers are left to estimate the number of eligible units in the sample. AAPOR response rate equations account for this by recommending that all unresolved sample units be multiplied by 'e,' the eligibility rate among unobserved units, but make no recommendation for how to calculate 'e.' Among multimode address-based samples, calculating 'e' is especially challenging - are non-responders vacant units; was the answering machine linked to the sampled address; can flags on the delivery sequence file help? We evaluate various methods of calculating 'e' to adjust for these unknowns and their ability to predict unobserved eligibility.