An Estimation Method for Matrix Survey Sampling

• Activity Number: 85
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Takis Merkouris*+
• Companies: Athens University of Economics and Business
• Keywords: Split questionnaire ; generalized regression estimator ; calibration ; composite estimator ; best linear unbiased estimator

Abstract

Matrix sampling, sometimes referred to as a split-questionnaire, is a sampling design that involves dividing a questionnaire into subsets of questions, possibly overlapping, and then administering these subsets to different subsamples of an initial sample. This design reduces the data collection costs and addresses concerns related to response burden and data quality, but also reduces the number of sample units that are asked each question. For matrix sampling with overlapping subsets of questions, we propose an estimation method that exploits correlations among variables surveyed in the various subsamples in order to improve the precision of the survey estimates. The proposed method uses a suitable calibration scheme, which is equivalent to a generalized regression procedure based on the principle of best linear unbiased estimation, and is computationally very convenient.