Activity Number:
|
21
- Non-Probability Sampling and Estimation: Fit-for-Purpose Designs
|
Type:
|
Topic Contributed
|
Date/Time:
|
Sunday, July 30, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Survey Research Methods Section
|
Abstract #323367
|
View Presentation
|
Title:
|
Redirected Inbound Call Sampling - an Example of Fit for Purpose Non-Probability Sample Design
|
Author(s):
|
Burton Levine* and Karol Krotki and Georgiy Bobashev
|
Companies:
|
RTI International and RTI and RTI International
|
Keywords:
|
nonprobability ;
Redirected ;
Inbound
|
Abstract:
|
This paper evaluates the external validity of estimates from a new non-probability sampling method with calibrated survey weights. Individual callers that misdial into a bank otherwise non-active toll-free numbers are given a survey using an Interactive Voice Response system. These callers comprise a nonprobability sample of the United States population. The questionnaire, designed to evaluate bias, mimics questions from the American Community Survey (ACS) and the National Household Interview Survey (NHIS). Sample weights are created by calibrating to the demographic distributions estimated from the ACS. The distributions of the respondents' unweighted demographic characteristics are surprisingly close to the population, a result that is consistent with other non-probability samples fielded using this methodology. Estimates of public health outcomes from the non-probability survey are compared to the ACS and the NHIS. We calculate bias in the survey estimates assuming the ACS and the NHIS estimates reflect the true population values. We conclude that the low cost and speed coupled with the minimal bias of this data collection technique may make this technique a viable option in som
|
Authors who are presenting talks have a * after their name.