University of Waikato I1.09 (I Block, 1st floor)
Missing data is a common phenomenon in large scale surveys. This nonresponse can be of two types; unit and item nonresponse. The literature on missing data presents various available methods to deal with non-response. In this presentation I would like to acquaint you with a nearest neighbour imputation method. Our method is based on dissimilarity measures, is nonparametric and handles categorical and continuous covariates without requiring any transformations.
As the imputed values are not true values, estimating the variance of the parameter of interest using standard methods would underestimate the variance if no allowance were made for the extra uncertainty due to imputation. Hence, in this presentation some methods on estimating the variance of the parameter of interest are also discussed.