Adaptive Mixed-Mode Survey Designs Accounting for Mode Effects

• Activity Number: 437
• Type: Invited
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Johannis Schouten*+
• Companies: Statistics Netherlands
• Keywords: Nonresponse error ; Measurement error ; Responsive survey design ; Web surveys

Abstract

Adaptive and responsive survey designs have focused mostly on reducing nonresponse error in single mode surveys. However, to date, the most influential and most debated survey design feature is the survey mode. The quality - cost differential between modes is very strong and worldwide statistical offices have gradually begun to redesign their surveys to mixed-mode designs. Given the large differences in costs and given large anticipated differences in mode-specific survey error, i.e. the mode effect, the survey mode is the main design feature to be adapted to the respondent. Such designs need to extend their scope to measurement error and they need to minimize mode effects with respect to some benchmark. In this paper, a number of approaches is presented to account for mode effects in adaptive survey design. For surveys with only a few key variables, the optimization can be directed specifically at the minimization of the corresponding mode effects. A case study is presented linked to the Dutch Labour Force Survey.