Multivariate Small-area Estimation for Mixed-type Response Variables With Item Nonresponse

Abstract Many surveys collect information on discrete characteristics and continuous variables, that is, mixed-type variables. Small-area statistics of interest include means or proportions the response variables as well their domain means, which are mean values at each level a different categorical variable. However, item nonresponse in survey data increases complexity small-area estimation. To address this issue, we propose multivariate mixed-effects model for subject to nonresponse. We apply method two structures where missing completely random by design. use empirical from separate studies: pet owners dataset National Resources Inventory. In these applications, our proposed leads improvements relative direct estimator predictor based univariate model.