• Summary
• Downloads
• Explanatory Notes
• Related Information
• Past Releases

DATA QUALITY

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 error or non-sampling error, as defined below.

Estimates in this publication are based on information collected from a sample of registered motor vehicles, rather than a full enumeration, and are therefore subject to sampling error. They may differ from the figures that would have been produced if the information had been obtained for all registered motor vehicles.

Other inaccuracies may occur because of imperfections in reporting by vehicle owners or in processing by the ABS. These are collectively referred to as non-sampling error and may occur in any statistical collection whether it is based on full enumeration or a sample. Concerns about non-sampling error in previous surveys, most notably recall bias, led to the introduction of a new collection methodology for the 1998 SMVU. While the new collection methodology has resulted in an overall improvement in SMVU estimates, some data quality issues remained with the 1998 survey and these are described below.


SAMPLING ERROR

The sampling error associated with any estimate can be calculated from the sample results. One measure of sampling error is given by the standard error, which indicates the extent to which an estimate might have varied by chance because only a sample of vehicles was included. There are about two chances in three that a sample estimate will differ by less than one standard error from the figure that would have been obtained if all vehicles had been included, and about 19 chances in 20 that the difference will be less than two standard errors.

Another measure of sampling variability is the relative standard error (RSE) which is obtained by expressing the standard error 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 error likely to have occurred due to sampling.

The SMVU was designed primarily to minimize relative standard errors for estimates of total kilometres travelled at the State/Territory level for the main vehicle types.

Detailed relative standard errors for all variables can be made available on request.

NON-SAMPLING ERROR

The ABS was concerned about the extent of non-sampling error in the SMVU prior to 1998. The previous collection methodology relied on owners of selected vehicles recalling details of vehicle use over the 12 months prior to receiving the survey questionnaire. This dependence on a "recall methodology" for collecting motor vehicle use information from respondents led to increased non-sampling error where vehicle owners did not keep detailed records. The degree of record keeping by freight vehicle and bus owners was better, but even for these vehicles, not all of the information sought was readily available. This led to two main types of observable deficiencies with the quality of data entered on survey questionnaires; firstly, where respondents failed to provide a figure for particular questions, even after follow-up contact from the ABS; and secondly, where respondents' recall was inaccurate. A particular concern was the extent to which respondents were only able to provide rounded figures for the questions about distance travelled. Studies showed that inaccurate recall by respondents resulted in biased estimates of total distance travelled.

In the light of these concerns about data quality, the ABS undertook an extensive review during 1995 and 1996 of the methods employed to obtain vehicle use data, including tests of different methodological options. Following this review, the SMVU was redesigned from the previous triennial survey to an ongoing quarterly survey designed to produce annual estimates. The new quarterly survey was introduced from 1 August 1997.

The ABS is confident that the new "pre-advice" methodology has improved the quality of its estimates of motor vehicle use. The new methodology in which vehicle owners receive early advice about their inclusion in the survey encouraged a higher degree of record keeping about the use of the vehicle during the survey period, either within owners' systems or by using the worksheet provided. This reduced the reporting errors arising from inaccurate recollection of use identified as a deficiency in the previous collection methodology. In addition, the reporting of odometer readings taken at the start and end of the survey periods provided more reliable estimates of total distance travelled without the recall bias inherent in the previous methodology.


OTHER DATA QUALITY ISSUES

In spite of the overall improvement in the collection methodology, it was necessary to make some adjustments to the estimation process for the 1998 survey to account for vehicles not in the population from which the survey samples were selected. This occurred either because they were registered just prior to the commencement of the survey, or because of processing difficulties with information received from some of the motor vehicle registration authorities.

As 1998 survey selections were taken from vehicles registered at 31 October 1996, i.e. nine months before the beginning of the 1998 survey reference year, adjustments were made to account for the use of new motor vehicles registered after that date and up to 31 July 1998, as well as the re-registration during this time of other vehicles not registered at 31 October 1996. At the Australian level, the adjustment for vehicles being re-registered after 31 October 1996 accounted for approximately 1% of total distance travelled for all vehicles. For rigid and articulated trucks the adjustment was slightly higher at 2%. However the impact of the adjustment for new motor vehicles was much more significant and is detailed by type of vehicle in the following table.

The following table shows the effect of the adjustment for new motor vehicles registered after 31 October 1996 as a percentage of their contribution to total distance travelled for each type of vehicle.

	%
Passenger vehicles	11
Motor cycles	16
Light commercial vehicles	13
Rigid trucks	9
Articulated trucks	15
Non-freight carrying trucks	11
Buses	10
Total	11
(a) New motor vehicle registrations.