Comparing Estimate Bias and Variances Using Calibrated Weights with a Different Number of Constraints in a Sample with Oversamples in Specific Groups

• Activity Number: 153
• Type: Contributed
• Date/Time: Monday, July 30, 2007 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Nuria Diaz-Tena*+ and Rob Cole and Partha Dass
• Companies: TNS-Global and TNS-Global and TNS-Global
• Keywords: weight adjustment ; weight calibration ; marginal targets ; web surveys ; noncoverage ; oversample

Abstract

TNS conducts a web-based customer satisfaction for each of the major wireless service providers in 20 Designated Marketing Areas (DMAs). The survey measures both overall satisfaction as well as performance on a large number of attributes. The client needs to be able to look at the data in a variety of ways: the entire wireless market; by specific wireless provider; and particularly by specific wireless provider within individual DMAs. TNS computes weights to compensate for noncoverage, nonresponse and oversamples. To compensate for response bias, we decided to calibrate age, gender and income within DMA and provider. Previously, the calibration was done only by age, gender, income and provider within a DMA. Increasing the number of constraints to calibrate within area and provider improved the effective sample sizes of the DMAs by providers.