A nonresponse adjustment method: local regression imputation

Missing data occur because because of unit nonresponse or item nonresponse. Unit nonresponse is usually handled by weigthing adjustment, whereas item nonresponse is handled by a form of imputation. in this paper a regression imputation method is developed to deal with the item nonresponse problem in the estimation of the mean of a finite population. The aim of this procedure, denoted as local regression imputation method, is to make the estimates obtained by traditional regression imputation methods robust against violation of the basic assumption, that could be ignored during the explorative phase of the sampled data.