|Page tools: Print Page Print All RSS Search this Product|
11 While comparisons are made between 2001-02 survey results and earlier iterations of the Legal Practices Survey, the reader should bear in mind that the survey was not designed to support accurate estimates of change, and should exercise caution when comparing 2001-02 results to earlier surveys. The effect of sampling variability on historical comparisons is discussed in paragraph 23.
12 Investigations into data quality for the 1998-99 survey revealed that not all respondents were using as strict a definition of pro bono work as was intended. The questionnaire wording was improved for 2001-02 to make it clearer that free first consultations, cases involving contingency fee arrangements, and incentive discounts for regular clients should not be included. A new category of legal aid cases, at a reduced fee or without expectation of a fee, was also introduced to prevent some legal aid cases being reported in other categories of pro bono work. The 2001-02 survey estimates showed a lower level of pro bono work than in 1998-99, but this is thought to reflect the methodological changes made to the collection of the data as well as non-sampling error. For an explanation of non-sampling error refer to paragraph 24. The estimates of pro bono work should be used with caution.
RELIABILITY OF THE DATA
13 When interpreting the results of a survey it is important to take into account factors that may affect the reliability of estimates. Such factors can be classified as either sampling or non-sampling error.
14 The estimates in this publication are based on information obtained from a randomly selected stratified sample of legal services practices and organisations in the Australian business population. Consequently, the estimates in this publication are subject to sampling variability, that is, they may differ from the figures that would have been obtained if all units had been included in the survey. One measure of the likely difference is given by the standard error (SE), which indicates the extent to which an estimate might have varied by chance because only a sample of units was included.
15 There are about two chances in three that a sample estimate will differ by less than one SE from the figure that would have been obtained if all practices in Australia had been surveyed, and approximately 19 chances in 20 that the difference will be less than two SEs.
16 Sampling variability can also be measured by the relative standard error (RSE), which is obtained by expressing the SE as a percentage of the estimate to which it refers. The RSE is a useful measure in that it provides an immediate indication of the percentage errors likely to have occurred due to the effects of random sampling, and this avoids the need to refer also to the size of the estimate.
17 The following table contains estimates of RSEs for a selection of the statistics presented in this publication.
RELATIVE STANDARD ERRORS FOR TABLE 1.1 SUMMARY OF OPERATIONS
. . not applicable
na not available
np not available for publication but included in totals where applicable, unless otherwise indicated
Copyright ã Australian Bureau of Statistics, 2003
18 As an example of the above, the estimate of total income for legal practices in 2001-02 was $10,636.1m and the RSE was estimated to be 2%, giving a SE of approximately $212.7m. Therefore, there would be two chances in three that, if all units had been included in the survey, a figure in the range of $10,423.4m to $10,848.8m would have been obtained, and 19 chances in 20 (i.e. a confidence interval of 95%) that the figure would have been within the range of $10,210.7m to $11,061.5m.
19 The estimates presented in the main tables of this publication are recorded to a high degree of accuracy, to enable estimates to be accurately added together, or otherwise used to derive new statistics or indicators. This precise recording is not meant to imply a high degree of accuracy in the estimates. RSEs should be used as a guide to determine the appropriate degree of precision to use when working with the estimates to draw conclusions. For example, if the estimate of total income of $10,636.1m has a SE of $212.7m and there are two chances in three of the true value lying between $10,423.4m and $10,848.8m, then an appropriate precision to use when citing the estimate is $10,600m. Most of the estimates presented in this publication are only accurate to two significant figures (estimates in thousands should be rounded to the nearest hundred, estimates in hundreds should be rounded to the nearest ten, and so on).
20 The sampling variability for estimates at the state/territory level was higher than for Australian level aggregates. Within states/territories, the sampling variability, and therefore the RSEs of estimates for smaller states and territories were higher than for the largest states. Survey estimates for the smaller states and territories should therefore be viewed with more caution than those for other states. RSEs for Queensland, South Australia, Western Australia, Tasmania, Northern Territory and Australian Capital Territory were typically 2 to 5 times higher than the corresponding national estimate RSEs. RSEs for New South Wales and Victoria were typically 1.5 to 2 times greater than the corresponding national figure.
21 The sampling variability for estimates for solicitor practices at the smaller size category level was also higher than for all practice aggregates. Survey estimates for the size categories with fewer than 10 principals/partners should therefore be viewed with more caution than those for all practices. RSEs for estimates relating to practices with one principal/proprietor were typically 2 to 3 times higher than the corresponding all practice estimate RSEs. RSEs for the size categories 2 principals/partners, 3-5 principals/partners, and 6-9 principals/partners were typically 3 to 6 times greater than the corresponding all practice figure. RSEs for the 10 or more principals/partners category were typically fairly similar to the corresponding all practice estimate RSE, due to the large proportion of these practices selected in the survey sample.
22 The sampling variability for estimates for barrister practices by barrister type or years at bar were also higher than for all practice aggregates. RSEs for estimates by barrister type were typically 1.2 to 1.5 times greater than the corresponding all practice estimate RSE, and RSEs for estimates by years at bar were typically 2 to 5 times greater.
23 The sampling variability for estimates of movement or change that are obtained by comparing 2001-02 survey results with previous results are also subject to high levels of sampling variability. The standard error (SE) of the estimate of change is approximately 1.4 times the standard error of the 2001-02 estimate. For example, the survey estimated that total solicitor practice income increased from $6,191.5m in 1998-99 to $8,378.6m in 2001-02, an increase of $2,187.1m. The standard error of the 2001-02 estimate of total income is $167.6m. The standard error of the estimate of change was estimated to be 1.4 times $167.6m, or $234.6m. This can also be expressed as a relative standard error of 10.7% of the estimate of change in total income.
24 Errors other than those due to sampling may occur in any type of collection and are referred to as non-sampling error. For this survey, non-sampling error may result from such things as deficiencies in the register of businesses from which the sample was drawn, non-response, imperfections in reporting and/or errors made in compiling results. The extent to which non-sampling error affects the results of the survey is not precisely quantifiable, but its impacts can be broadly identified. One such example is in relation to pro bono work. Many legal practices, particularly large practices, did not keep records on the number of pro bono hours worked by solicitors/barristers, hence they provided an estimate of its value. Every effort was made to minimise non-sampling error by careful design and testing of the questionnaire, efficient operating procedures and systems and the use of appropriate methodology. Survey estimates subject to a high level of non-sampling error have been suppressed or provided with relevant cautions.
25 Estimates that have an estimated relative standard error between 10% and 25% are annotated with the symbol '^', estimates with an RSE between 25% and 50% are annotated '*', whilst estimates with an RSE greater than 50% are annotated '**'. Caution should be exercised when using these estimates.
26 Where figures have been rounded, discrepancies may occur between the sum of the components and the total. Similar discrepancies may occur between a proportion or ratio, and the ratio of the separate components.
27 Data contained in the tables in this publication relate to all operations of legal services practices and organisations in Australia during the year ended June 2002. Financial estimates included the activity of any legal practices or organisations that ceased or commenced operations during the year. Counts of practices included only those practices that were operating at 30 June 2002. Employment included only those persons working for a legal practice or organisation during the last pay period ending in June 2002.
28 ABS publications draw extensively on information provided freely by individuals, businesses, governments and other organisations. Their continued cooperation is very much appreciated; without it, the wide range of statistics published by the ABS would not be available. Information received by the ABS is treated in strict confidence as required by the Census and Statistics Act 1905.
DATA AVAILABLE ON REQUEST
29 Inquiries about these statistics and more detailed statistics than those presented in this publication should be made by telephoning the contact shown on the front page.
Back to Main Features
These documents will be presented in a new window.