|
Abstract:
|
The usefulness of economic estimates often depends on two characteristics: accuracy and timeliness. Editing methodologies have a direct impact on both of these attributes. For decades, statisticians have researched the implications of over-editing. It is widely agreed that the over-editing of survey data can deplete valuable resources, delay information release, and introduce bias. However, change in editing methodology has been slow and research on specific methodological issues sparse. Our primary objective is to increase the efficiency of the editing process by reducing costs and improving timeliness of data releases without diminishing data quality. In this paper, we investigate i) current survey specific editing practices, ii) the validity and impact of these practices, and iii) the feasibility of stopping the data editing process earlier. The remainder of the paper discusses moving toward an adaptive editing design by implementing changes to current editing methodologies and investigating the use of automation and Big Data techniques.
|
