210 – Contributed Poster Presentations: Survey Research Methods Section
Addressing Outcome-Related Nonresponse Through a Doorstep Interview
Wendy Van de Kerckhove
Westat
Not missing-at-random (NMAR) nonresponse occurs when the reason for not responding to a survey directly relates to the outcome of interest. This source of bias cannot be addressed through standard weighting adjustments that rely on auxiliary variables. The Programme for the International Assessment of Adult Competencies (PIAAC) administers adult skills assessments (e.g., literacy, numeracy) in a country’s main language(s). It is subject to NMAR nonresponse because of persons that are unable to participate due to a language barrier, learning/mental disability, or reading/writing barrier. This paper provides background on the handling of such cases in data collection, weighting, and estimation for the first cycle of PIAAC. It then describes plans to further reduce the extent of potential bias due to NMAR nonresponse in the second cycle by introducing a brief multi-lingual questionnaire called the Doorstep Interview.