Evaluation of the Aggies Automated Edit and Imputation System

Abstract

Data editing plays an important role in the survey process. The National Agricultural Statistics Service currently uses, in addition to some manual editing, an interactive micro-level edit system or a batch micro-level edit system, and an interactive macro-level edit system to edit reported data. Advantages of using these two edit systems are that: 1) the most complex edits can be incorporated and 2) the impact of editing at aggregate levels can be readily evaluated. There are, however, disadvantages with the use of the two edit systems: 1) a considerable amount of time and resources may be expended and 2) editing may not always be performed in a consistent manner. This paper evaluates a generalized automated edit and imputation system developed by the author called the Agricultural Generalized Imputation and Edit System (AGGIES). The AGGIES is appealing for the following reasons: 1) editing and imputation are fully automated, 2) the system provides consistency in the edit and imputation process, and 3) the system can be easily applied to any number of surveys, thus conserving resources to the development and maintenance of a single system. Comparisons between the AGGIES and the current edit and imputation procedures are made for expanded totals and the number and magnitude of variable changes. The data used for these comparisons are obtained from the Quarterly Hog Survey. The results reveal that the expanded totals obtained from using the AGGIES are similar to those obtained from the current edit and imputation procedures. Further testing on more applications is recommended.