CONFIDENTIALITY AND CONSENT
It is important when obtaining administrative data for statistical use that the information provided remains confidential and private no matter how the data ends up being utilised. This means that no business or individual should be able to be identified from any of the statistical releases based on the administrative data.
The provision of accurate and timely administrative data is often reliant on public goodwill and cooperation. The reporting units (e.g. persons, businesses) about whom the data are being collected should be advised about the intended use of the data that they are providing.
Concerns about the protection of confidential information can cause a decrease in public confidence and a subsequent loss of accurate data due to non-response. For example, perceived misuse of EFTPOS data will not only be bad for the reputation and later effectiveness of the receiving agency in their survey activities, but also lose customers for the data custodian supplying the EFTPOS data. This in turn would reduce the data custodian's willingness to respond to any future data requests.
Problems in regards to confidentiality can occur during the hand over of administrative data if appropriate measure aren't taken with the security of the data transfer. It is important to ensure that the systems used in the transfer of data are compatible to avoid excess data handling. It is also important to make sure that all privacy and other applicable legislation have been investigated and adhered to prior to the agreement of data acquisition.
When using administrative data to link existing data sources along with complying with all relevant legislation it is worth while considering and resolving the governance of the use and access of the newly linked data source to ensure that privacy is maintained.
Some methods for minimising privacy concerns and protecting confidentiality when using administrative data include:
- informing respondents of how their data will be used;
- following privacy principles and legislation (both of the receiving agency and data custodian);
- be sensitive to cultural differences;
- honour guarantees made to respondents;
- restrict access to unit record data (i.e. data relating to individual units); and
- develop guidelines on access and release of the data to ensure that no business or individual can be identified from any statistical release.