|Page tools: Print Page Print All|
27.4 Estimates of the percentage change in average earnings (from both the previous quarter and from the corresponding quarter of the previous year) are published for each series. Estimates from the survey are cross-classified by sex, State/Territory, industry and sector.
27.5 Estimates are published on the following bases: original; seasonally adjusted; and trend (see Chapter 16 for further explanation of original, seasonally adjusted and trend estimates). Seasonally adjusted and trend estimates are available by Australia and State/Territory for each of the three main series listed above.
27.6 Data collected in the survey are compiled according to concepts and definitions outlined in Chapter 4 (employees) and Chapter 11 (earnings).
27.7 The scope of the survey is restricted to employing businesses (MUSs). In addition, the standard scope exclusions for ABS labour-related business surveys (outlined in Chapter 23) apply to this survey.
27.8 Details of numbers of employees (full-time adults and other employees), total gross weekly earnings (for full-time adults and other employees), and weekly overtime earnings of full-time adults are obtained on a quarterly basis from selected businesses, using a mail-out/mail-back collection methodology. Data for some Commonwealth and Territory Government organisations are collected electronically on a fortnightly basis from a centralised pay system. A small number of large private businesses also provide data electronically.
27.9 The survey reference period is the last pay period ending on or before the third Friday in the middle month of the quarter. Businesses are asked to estimate one week's proportion of total gross earnings and full-time adult overtime earnings if their payroll is not weekly.
27.10 Businesses which do not mail back their completed questionnaire within a reasonable period of time after the survey reference period are followed up by mail and then phone if necessary. Final response rates for the Survey of Average Weekly Earnings over the 1998-99 financial year averaged 98%.
27.11 The selection unit for the survey is the Management Unit/State (MUS). The collection and reporting units used in the survey usually correspond to the selection unit. However, where the MUS is unable to provide information required for the survey, it may be split into a number of 'reporting units'. For further information on statistical units used in ABS business surveys refer to Chapter 23.
27.12 A probability sample of MUSs (employing businesses) is drawn from the ABS Business Register using the process outlined in Chapter 23. Variables used to stratify the survey frame each quarter are:
26.13 Strata on the survey frame that are completely enumerated include those containing MUSs with benchmark employment greater than a set cutoff (this cutoff will vary for different States/Territories) and strata with a very small number of MUSs. Strata which are completely enumerated because they contain a low number of MUSs may become sampled strata if the number of MUSs in those strata increases sufficiently.
27.14 In addition to constraints outlined in Chapter 23, sample selection is constrained by the need to ensure that there is minimum overlap with other labour-related business surveys and with non-labour related business surveys.
SAMPLE SIZE AND ALLOCATION
27.15 Approximately 4,700 MUSs are selected in the sample to yield a live sample of approximately 4,000 MUSs.
27.16 The sample is allocated optimally across sampled strata using a technique designed to minimise the variance of average weekly earnings estimates at both the national and State/Territory level.
27.17 The sample is updated each quarter to reflect changes in the ABS Business Register. Approximately 8% of the sample selected from the non-completely enumerated strata is replaced each quarter. This process is called sample rotation (see Chapter 23 for further explanation).
27.18 Sample rotation is implemented for the majority of sampled strata, but is not implemented where the population of a stratum is so small that units rotating out of the sample would be rotated back in after only a short interval.
27.19 Estimates of total weekly earnings and number of employee jobs are computed for various combinations of State/Territory, sector and industry. Average weekly earnings measures are the quotient of the respective earnings and employment measures. Ratio estimation is used in all sampled strata, except in cases where there is a large proportion of units with zero benchmark employment, in which case number-raised estimation is used.
27.20 In both completely enumerated and sampled strata an automatic imputation procedure is used for units not responding, by applying imputed growth rates to the most recently reported employment and earnings data for these units, provided that data have been reported in either of the two previous quarters. This is referred to as Beta imputation. Otherwise, the Live Respondent Mean method is used to impute for missing data items.
27.21 Significance editing was introduced in September 1992. This technique means that editing is only performed on those survey values which will significantly impact on the survey estimate if left unaltered.
27.22 Survey outliers are handled by using the 'surprise outlier' technique.
27.23 New Business Provisions were introduced in May 1999. Adjustments are made to survey estimates each quarter to allow for births and deaths of businesses that have occurred up to the end of the survey reference period but which are not reflected on the survey frame.
27.24 For further information on estimation methods used in ABS Business Surveys, refer to Chapter 23.
TIME SERIES ESTIMATES
27.25 Both seasonally adjusted and trend estimates are produced for key series from this survey.
27.26 Seasonally adjusted estimates were introduced from September quarter 1983. The seasonal factors are reviewed annually to take account of each additional year’s original data. The review takes place in time for the results to be incorporated in each February issue of Average Weekly Earnings, Australia (Cat. no. 6302.0).
27.27 Trend estimates were introduced from August quarter 1993.
RELIABILITY OF THE ESTIMATES
27.28 Estimates from the survey are subject to both sampling and non-sampling error. The relative standard errors of survey estimates are published in Average Weekly Earnings, Australia (Cat. no. 6302.0).
27.29 The 'jack-knife' approach is used to calculate estimates of variance for this survey. For further information on the jack-knife technique for estimating variances, or on sampling and non-sampling errors, refer to Chapter 16.
DATA COMPARABILITY OVER TIME
27.30 In order to provide a high degree of consistency and comparability over time, changes to survey methods, survey concepts, data item definitions, frequency of collection, and methods of time series analysis, are made as infrequently as possible. Significant changes have included:
27.31 For further details contact the Assistant Director, Labour Employer Surveys Section, on Perth (08) 9360 5245.