Significance Editing in the Australian Survey of Average Weekly Earnings

Abstract

Editing survey returns is a sub­stantial component of the cost of running sample surveys. This paper presents a method for reducing the resources allo­cated to editing. The method is based on estimating the effect on the survey esti­mates of resolving the edit queries for each unit. This method has been introduced to the Australian Survey of Average Weekly Earnings. As a result the number of respon­dents contacted to resolve edit queries has been more than halved without affecting the quality of the final statistics. This has reduced the total respondent burden and also the overall costs of conducting the sur­vey. It has also enabled an earlier publica­tion of the survey results. This paper gives a description of the technique, and also pre­sents results from an evaluation study that was conducted prior to its implementation.