4.5 CHANGES IN SURVEY METHODOLOGY IN 2003–04
CHANGES IN METHODOLOGY IN 2003-04
There were a number of changes to the survey methodology introduced in 2003-04. Some of these were a consequence of the integration of the HES and SIH. The main changes which could impact on all data items were:
- previous SIH cycles had selected dwellings from those that had been respondents for eight months in the Monthly Population Survey (MPS), whereas in 2003-04 the SIH sample was drawn from dwellings not recently included in an ABS household survey (possible change in response bias)
- the sample size of the SIH was increased from 10,211 households (comprising 19,400 persons aged 15 and over) in 2002-03 to 11,361 households (comprising 22,315 persons aged 15 and over) in 2003-04 (lower sample error)
- interviewer use of a laptop computer - while the 1998-99 HES also used this collection method, the 2003-04 survey is the first SIH to use computer assisted interviewing (this may have improved data capture)
- the HES sample weights were calibrated to the same set of independent demographic benchmarks as the SIH and were calibrated to SIH estimates of the number of households by tenure type and SIH estimates of aggregate household income by state and territory and broad source of income
- editing and imputation procedures were changed - in particular because the SIH sample is no longer drawn from households who have participated in the MPS, responses given in the MPS are no longer available as a basis for imputation.
The changes in survey methodology relating to specific data items were:
- current income from own unincorporated business and investments was measured using respondents' estimates of expected income in the current financial year, whereas previously these data items were estimated based only on information about reported income for the previous financial year - this change had a significant impact on the coverage of such income streams in current income measures
- the collection of details about the assets and liabilities of the household may have improved the quality of reporting of associated income streams
- the instrument wording has been changed to explicitly ask that reported dividends include the value of imputation credits - previously this direction was only included in interviewer instructions
- expenditure on pay TV and internet service provider accounts were specifically collected in the household questionnaire on a 'last payments' basis, resulting in more reliable expenditure data for these items than was previously obtained when estimates were based on expenditure in the two week diary period
- expenditure data relating to the purchase of dwellings was collected with respect to the previous 3 years - previously this expenditure had been collected in the HES in relation to the previous 2 years
- information relating to some household loans was collected using a different methodology - for those loan accounts that have a redraw facility and have regular income (such as wages) deposited into them, respondents were not asked to provide a 'usual repayment' - instead they were asked to provide the amount that the principal outstanding usually decreases by in a 6 month period and this was used in conjunction with information collected on interest to derive a repayment amount
- changes to the derivation process used to estimate income tax liability in the HES to only use data items available in the SIH
- details of previous financial year income were collected form all persons - in previous SIHs this information was not collected from people who had only arrived in Australia in the current financial year
- details of hours worked were collected from all employed persons - in previous SIHs, this information was only available for employees
- unlike previous SIHs, data on repayments and principal outstanding on mortgages for other purposes (ie for purposes other than building, buying, altering or adding to the selected dwelling) excludes mortgages that were used for business or investment purposes.