Mr Campbell asked the Treasurer, upon notice, on 8 October 1996:

(1) How many persons (a) refused to complete the 1991 Census form and (b) were successfully prosecuted for their refusal.

(2) How many persons were successfully prosecuted for refusing to complete Australian Bureau of Statistics (ABS) surveys in each year since 1991.

(3) Why are the Census and ABS surveys compulsory.

(4) What surveys has the ABS conducted since 1991 which were not compulsory.

(5) What reliability can be placed on non-compulsory surveys if compulsion is necessary to ensure an unbiased sample.

(6) If non-compulsory surveys give valid results why (a) are not all ABS surveys optional and (b) does the ABS threaten persons who consider that the surveys invade their privacy.

(7) What databases are accessed by the ABS in order to get the names and addresses of persons to be surveyed or who refuse to or neglect to fill in forms.

(8) What is the minimum compliance rate which gives meaningful data.

Mr Costello —The answer to the honourable member's question is as follows:

(1)(a) 5,234

(1)(b) 91

(2)—

(1) excludes Population Censuses

(3) The word `compulsory' is not used in Australian statistics legislation, and the term `compulsory survey' has no accepted meaning.

The Census and Statistics Act 1905 provides the Statistician with two powers for the collection of data. `The Statistician . . . . may, either orally or in writing, request (emphasis added) a person' to fill up a form (Section 10(3)) or to answer a question (Section 11(1)). No offence is committed if a person refuses such a request. The Act also provides that `the Statistician may, by notice in writing . . . .. direct (emphasis added) the person' to fill up a form (Section 10(4)) or to answer a specified question (Section 11(2)). Failure to comply with such a direction without reasonable excuse is, in respect of each day, an offence subject to a fine not exceeding $100 (Section 14(1)).

In all its statistical collections the Australian Bureau of Statistics (ABS) first requests businesses and householders to provide the data it seeks ie. the ABS seeks willing cooperation. While the power to direct persons to respond is generally available, by the collections being tabled before Parliament under Section 6(3) of the Australian Bureau of Statistics Act 1975, these powers are rarely needed or used.

In these circumstances, it is not possible to define ABS surveys as compulsory or not compulsory.

(4) It is not possible to define ABS surveys as `compulsory' or `not compulsory'—see response to Question (3).

However, it is possible to say that five surveys since 1991 were conducted on an overtly voluntary basis as follows:

Survey of Embassies and Consulates 1992-93

Study of Workload Placed on Small Business by Government Paperwork (1992)

Population Survey Monitor (quarterly from August 1993 to May 1996)

Queensland Day Survey (1995)

Women's Safety Survey (1996)

In addition, some household collections contained some questions which householders were advised were voluntary e.g. the question on religion in the 1996 Population Census.

(5) It is not possible to define ABS surveys as `compulsory' or `not compulsory'—see response to Question (3). However, in all its surveys the ABS undertakes extensive testing to ensure the collection methodology used will provide reliable data of sufficient quality to meet the needs of the key users.

(6)(a) It is not possible to define ABS surveys as `compulsory', `not compulsory' or `optional'—see response to Question (3).

ABS and overseas experience (eg see ABS Annual Report 1987-88 p19-21) shows that surveys conducted on an overtly voluntary basis typically produce data of lower quality. ABS seeks to provide high quality data in all its collections. Experience shows that this is best obtained by seeking the willing cooperation of selected households/businesses but with the power to direct a response being available should it be needed. On occasions, particularly if the content of the survey is sensitive (e.g. the Women's Safety Survey) a survey may be conducted without the power to direct a response being available, providing field testing has shown that data of sufficient quality can be collected.

(b) ABS does not threaten persons who consider surveys invade their privacy.

(7) External databases of names and addresses of persons are rarely used to select samples for ABS household surveys. A recent example was the use of a Commonwealth Employment Service list in the sample selection for a survey of employment and unemployment patterns.

The names and addresses of those who do not respond to a household survey or the Population Census are obtained from the ABS contact with the household in question.

(8) There is no single compliance rate which can be universally applied to all surveys to give `meaningful data'. The ABS aims to produce data of sufficient quality to meet the needs of the users.