Estimating Technology Adoption and Aggregate Volumes from U.S. Payments Surveys in the Presence of Complex Item Nonresponse

• Activity Number: 255
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): May X. Liu*+ and Geoffrey R. Gerdes
• Companies: Federal Reserve Board and Federal Reserve Board
• Keywords: item nonresponse ; imputation ; ratio estimator ; multiple imputation

Abstract

In recent years, the Federal Reserve has surveyed banks on the number and value of check payments, and to track the adoption of new electronic check clearing methods that replace traditional paper-based methods. The data requested from each respondent consists of a hierarchy of nested totals and subtotals. The ability of banks to report data to this voluntary survey tends to vary, however, which leads to a complex pattern of item nonresponse. In light of this, independent estimation of aggregates by ratio estimator using only the reported data created violations of adding-up and other logical constraints among survey items. To overcome this, we investigated various item imputation methods. But these methods tended to overestimate the proportions of banks in the population that had adopted the new clearing technologies. Accordingly, we adjusted the methods to account for that bias.