Richard Penny

Statistics New Zealand

Imputation variance estimation for the Accommodation Occupancy Survey

This is joint work with Raazesh Sainudiin, Department of Mathematics and Statistics, University of Canterbury

We formulate the problem of imputation variance estimation for the Accommodation Occupancy Survey (AOS) run on behalf of the Ministry of Tourism by Statistics New Zealand and develop a methodology to address this problem. We use Bayesian nonparametric methods and bootstrap techniques to provide consistent estimates of the imputation variance under the assumption of homogeneity within the predefined imputation cells.