APPENDIX SURVEY CHANGES
CHANGE FROM CENSUS TO SAMPLE SURVEY
The Survey of R&D, Businesses had shown a steady increase in the number of businesses each year undertaking research and experimental development. The cost of collecting data from the additional units each year, along with the additional burden on providers, is unsustainable. As part of a review of the suite of R&D surveys, the decision was made to change the Survey of R&D, Businesses from a reduced scope coverage census to a sampled survey. This means that rather than contact all businesses who undertake R&D with intramural expenditure of $100,000 or more, data was collected from a sample of these businesses. The 2011-12 survey was based on a random sample of approximately 4,000 businesses which was stratified by industry and expenditure on R&D. Businesses with expected expenditure below $100,000 have been excluded from the sample. Due to the nature of R&D activity, some industries have been completely enumerated, that is, all businesses have been included for these industries.
The change of the collection from a census to a sample will have some impact on the estimates.
The move to a survey introduces sampling error. Sampling error results from differences in estimates from those that would be produced if a census had been undertaken. For this publication, sampling variability is measured by the relative standard error (RSE). Please see the Sampling Error section of the Technical Note for more information.
Estimates by industry (ANZSIC 2006)
The level of detail that will be available for industry, as collected using the Australian and New Zealand Standard Industrial Classification (ANZSIC), 2006 will be affected. This may particularly be the case for special data requests. Information below the subdivision level (2 digit), and in some cases, division level, may not be available. This may be caused by limited information being collected at these levels making the estimates unreliable, or application of ABS confidentiality provisions to the data.
Estimates by classification of R&D (ANZSRC 2008)
The level of detail collected using the Australian and New Zealand Standard Research Classification (ANZSRC), 2008 for fields of research and socio-economic objective will also be affected. Outputs from 2011-12 are available at the division level (2 digit) where in the past they have been available at the group level (4 digit).