NORC extensively employs a variety of validated methods for maximizing unit and item response rates (including many methods pioneered by NORC staff), complex surveys of public policy importance require the application of advanced imputation techniques to facilitate appropriate interpretation and use of data by researchers and policymakers. NORC imputation experts routinely employ varied imputation approaches including hot deck imputation, logical imputation, model-based imputation, and adaptations of these methods to meet the substantial methodological challenges of a particular survey project.
While every effort is made to increase response rates, inevitably some sample units will be missing from the completed cases. If the cases are missing at random, then the nonresponse may be said to be ignorable, contributing only to the sampling variance. However, the nonresponse is not ignorable and may represent a bias if the missing cases tend to be different from those that responded. NORC statisticians have published on the issue of ignorable and non-ignorable nonresponse and bring this expertise to bear on these problem sets.