Return to Article Details The Performance of Multiple Imputation in Social Surveys with Missing Data from Planned Missingness and Item Nonresponse