A data set specification (DSS) is a group of data elements that specify information about one particular subject. A DSS also defines the conditions under which data on this subject are collected, whether they are mandatory or conditional, the verification rules to be used, the scope of the collection, and sometimes the sequence in which data elements are included.

For example, Community Housing DSS 2018– is a data set specification about the provision of community housing across Australia. The 62 data elements in this DSS include things such as the recipient’s address, their disability status, the main language spoken at home, the number of bedrooms and occupants in the dwelling, and details about the service provider organisation. The scope of this DSS is rental units being managed by a community housing organisation, and excludes tenancies managed under public housing and other programs. The data are collected by surveys of community housing providers and jurisdictional housing authorities, and provided to the AIHW once a year.

A National Minimum Data Set (NMDS) is a type of data set specification. An NMDS is the minimum set of data elements which must be collected and reported nationally. They are mandatory for all Australian states and territories to follow.

Within the health sector there are also two sub-types of data set specifications:

• National Best Endeavours Data Set (NBEDS). An NBEDS is a metadata set of health data which Australian jurisdictions agree they will make their best effort to collect.
• National Best Practice Data Set (NBPDS). An NBPDS is similar to an NBEDS, and is a metadata set for health data that is recommended for collection by agencies and organisations.

Clusters are a special kind of data set specification that can link together data elements within another data set specification, but they can’t be used on their own. They are like a ‘sub-DSS’ module, which can be ‘plugged in’ to data set specifications.