1352.0.55.062 - Research Paper: Issues Associated with Producing a Longitudinal Dataset of Businesses (Methodology Advisory Committee), Jun 2004  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 06/07/2004  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All
  • About this Release

About this Release

Recently the ABS has begun to create a Business Longitudinal Dataset (BLD) which will contain business information to facilitate analysis at the microeconomic level on a range of policy issues based around growth and performance. The aim was to create this dataset by combining annual administrative data, such as Australian Taxation Office and Australian Customs Service data, with ABS collected survey data. That is, without the need for data collection. However, during discussion with potential users of this dataset, the desire for business characteristic data has emerged. Such data is not collected by any current ABS business surveys and so the need for a specific survey has arisen. the current vision is a survey vehicle used to collect the characteristics data, which is then matched to available administrative data and other ABS survey data to form the BLD. Thus the BLD design issue is in fact a longitudinal survey design issue. It is the desire to produce a dataset which will be suitable for many different analytical purposes. This has presented a design challenge for us. the ABS has some experience at designing longitudinal surveys which will produce datasets for a particular analytical purpose, but much less experience in designing to produce a dataset which will be suitable for multiple analytical purposes.

This paper discusses the major issues associated with creating a business longitudinal dataset for multiple analytical purposes, describes how these impact on the development of the BLD and states the methodology intended to be used to overcome these issues.