Bayesian Multiscale Multiple Imputation with Implications to Data Confidentiality November, 2008
Scott H. Holan, Daniell Toth, Marco A. R. Ferreira, Alan F. Karr · 2008
Many scientific, sociological and economic applications present data that are collected on multiple scales of resolution. One particular form of multiscale data arises when data are aggregated across di!erent scales both longitudinally and by economic sector. Frequently, such data sets experience missing observations in a manner that they can be accurately imputed using the method we propose known as Bayesian multiscale multiple imputation. This method borrows information both longitudinally and across di!erent levels of aggregation to produce accurate imputations of missing observations as well as estimates that respect the constraints imposed by the multiscale nature of the data. Our approach couples dynamic linear models with a novel imputation step based on singular normal distribution theory. Although our method is of independent interest, one important implication of such methodology is its potential e!ect on confidential databases protected by means of cell suppression. In order to demonstrate the proposed methodology and to assess the e!ectiveness of disclosure practices in longitudinal databases, we conduct a large scale empirical study using the U.S. Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW). During the course of our empirical investigation it is determined that several of the predicted cells are within 1% accuracy, thus causing potential concerns for data confidentiality.