TECHNICAL NOTE CHANGES TO ESTIMATES RELATING TO NOT-FOR-PROFIT ORGANISATIONS RESULTING FROM IMPROVEMENTS IN ESTIMATION METHODOLOGY
CHANGES TO ESTIMATES RELATING TO NOT-FOR-PROFIT ORGANISATIONS RESULTING FROM IMPROVEMENTS IN ESTIMATION METHODOLOGY
1 Subsequent to the initial release, an error in the application of the estimation methodology has been identified, resulting in errors in the originally published estimates of not-for-profit organisations in Australia. This re-release presents updated estimates based on the application of the improved estimation methodology.
2 The survey of not-for-profit organisations in Australia was designed to be fully integrated with the Economic Activity Survey, the results of which are presented in Australian Industry (8155.0). This included using identical estimation methodology for producing estimates for both publications. This strategy was adopted to reduce the statistical reporting load placed on providers, and to produce estimates which were comparable between both surveys.
3 Generalised regression estimation methodology was used to calculate statistics for both publications. Under this methodology, sampled units are separated into groups with similar business characteristics, known as calibration classes. For each calibration class, population benchmarks of auxiliary data items are calculated. Survey weights of sampled organisations are then adjusted in such a way, that survey estimates of the auxiliary data items match the population benchmarks. These weights are then applied to all data items collected in the survey.
4 As the estimation methodology used in producing estimates for Not-for-profit Organisations, Australia, was identical to that used in producing estimates for Australian Industry, the calibration classes used were also the same. For this reason, the calibration classes included a mixture of both for-profit businesses and not-for-profit organisations.
5 Although the choice of calibration classes was suitable for use in producing the estimates presented in Australian Industry, they were found to be too broad for use in calculating the finer level estimates presented in Not-for-profit Organisations, Australia. In resolving this issue, new calibration classes were used, which accounted for the profit status and the market status of sampled organisations. This allowed the use of different generalised regression models, which resulted in improved estimates of not-for-profit organisations in Australia. This re-release presents updated estimates using these finer calibration classes.
6 The effects of the correction, as compared with the previously published estimates of not-for-profit organisations in Australia, differ by ICNPO group, and by data item. The table below shows the percentage change between the previously published estimates, and those presented in this re-release, for a number of key data items.
Changes to estimates relating to not-for-profit organisations resulting from improvements in estimation methodology(a)
Organisations at end June
|Culture and recreation |
|Education and research |
|Social services |
|Environment, development, housing, employment, law, philanthropic, international |
|Business and professional associations, unions |
|Other activities |
|- nil or rounded to zero (including null cells) |
|(a) For details of category descriptions, refer to the Explanatory Notes. |
|(b) At end of June. |
Further information about these changes can be obtained by contacting Julie Cole on (03) 9615 7562.