Key data quality information: Specialist Homelessness Services Collection, 2020–21
The AIHW plays a role in developing and maintaining national metadata standards. This work contributes to improving the quality and consistency of national health and welfare statistics. The AIHW works closely with governments and non-government organisations to achieve greater adherence to these standards in administrative data collections to promote national consistency and comparability of data and reporting.
One of the main functions of the AIHW is to work with the states and territories to improve the quality of administrative data and to compile, analyse and disseminate national data sets based on data from each jurisdiction.
Data Quality Statements are developed for each data set and made available on the AIHW Metadata Online Registry (METeOR).The 2020–21 Specialist Homelessness Services Collection Data Quality Statement is available from METeOR.
Breaks in time series
Clients subject to care and protection orders: Improvements made in 2015–16 to the method used to identify clients subject to care and protection orders mean that data from 2011–12 to 2014–15 are not comparable with data from 2015–16 onwards.
Source of income—DVA pension or payment: In 2017–18, the response options for source of income were updated and the three response options relating to payments or pensions from the Department of Veterans Affairs (disability pension—DVA, service pension—DVA and war widow(ers) pension—DVA) were replaced with a single response option of 'DVA pension or payment'. As the single 'DVA pension or payment response' option can include more payment types than the three options previously available, data on the 3 DVA pension or payments from 2011–12 to 2016–17 are not comparable with data on 'DVA pension or payment' from 2017–18 onwards.
The addition of ‘Other’ to the Sex item: On 1 July 2019 the inclusion of a Sex = Other category in both the Client and Unassisted persons components of the SHSC was introduced. This change had minimal impact on the collection with 0.8% of clients identifying as Sex = Other in 2019–20, in 2020–21 this was 0.4%. However, this change has meant that there has been a break in time series for some clients as their SLK (client identifier) has changed to reflect their change in recorded sex.
This change has also been implemented differently across states and territories with much higher numbers of clients identifying as Sex = Other in some states and in the unassisted collection. The Sex = Other option may have also been applied by some agencies to young children where their sex was not obvious or for those who did not want to state their sex.
Due to these issues, as well as the confidentiality concerns that accompany small numbers of clients, those clients identifying as Sex = Other have been included in the ‘Female’ category when reporting the data for 2020–21. This merging of Sex = Other clients with Female clients has not caused any impact on the validity of Female clients for 2020–21 as the number of Sex = Other clients is so small.
As more data is collected, and the data quality of this item improves, AIHW will review how this data is presented.
Data issues that require caution when making comparisons
Disability: Data for clients with disability who require assistance may not be comparable across age groups due to differences in the interpretation of the disability questions; this issue relates mainly to young children.
Presenting unit type: Data for presenting unit type may not be comparable across age groups due to differences in interpretation of presenting units and how they are recorded. This issue mainly concerns young children and presenting unit type ‘lone person’.
Housing crisis, financial difficulties and housing affordability: Improvements made during 2014–15 resulted in changes to the way agencies were required to report ‘main reason’ and ‘reasons for seeking assistance’. In addition, wording providing a specific example of housing crisis was removed from the section relating to reason for seeking assistance. Caution should be used when making comparisons over time as the reporting of these items may be inconsistent between agencies. These changes in agency reporting were evident in the data from all states and territories.
Children presenting alone: South Australia has a comparatively high number of children reported as presenting alone. This may be due to a difference in how presenting units are recorded in South Australia's client management system. Caution should be used when comparing data for children presenting alone in South Australia with other states and territories.
Case management: Some aspects of case management are recorded differently in South Australia’s client management system. Caution should be used when comparing data on case management for South Australia with other states and territories.
Services and Assistance: Assertive outreach: In 2017–18, there was a clarification made to the response option used to record clients who needed, or were provided, or referred assertive outreach services. The option was changed to specify that this service was directly targeted at rough sleepers. Due to this change, caution should be taken when comparing the number of clients receiving assertive outreach services before and after 2017–18.
COVID-19: From 26 March 2020, SHS agency workers were instructed to add ‘COVID-19’ in the free text section of the ‘Other reason for seeking assistance’ item if the client and/or the agency were affected directly or indirectly by the crisis. This change has been implemented differently across jurisdictions and therefore caution should be taken if making any comparisons between jurisdictions.
Funding for COVID related social support was also implemented differently across jurisdictions and SHSC data will only show services delivered where funding was provided to SHS agencies.
Jurisdictional changes: More detailed information on the responses by state and territory government to the COVID pandemic is available in the Specialist Homelessness Services: monthly data report.
Over the years there have also been other changes made by jurisdictions to specialist homelessness service delivery models and policies. If making jurisdictional comparisons over time please read the SHSC Data Quality Statement to ensure all these changes are considered in the analysis.
Improvements to data items
Mandatory data items: Changes made in 2014–15 resulted in a substantial improvement in data quality for mandatory data items and in particular resulted in a decline in the number of non-response or missing values for these data items. Care should be used when comparing results from 2011–12 to 2013–14 with results from 2014–15 onwards.
Housing situation: Following improvement in the derivation for housing situation used in the SHSC in 2016–17, clients with a tenure status of ‘life tenure scheme’ are now counted under the housing situation category ‘private or other housing (renter, rent-free or owner)’ if their dwelling status was ‘housing/townhouse/flat’. This change has very little impact on housing situation percentages and hence does not constitute a break in time series.
Age: In 2017–18, age and age-related variables were derived using a more robust calculation method. Caution should be used when comparing results with publications from December 2018 onwards that include 2017–18 data with previously issued publications.
Family and domestic violence: In July 2019 the following changes were made to improve the collection of information regarding family and domestic violence (FDV) in the SHSC:
- The inclusion of an ‘FDV agency’ option in the ‘Formal sources of referral’ item to capture clients referred from non-SHS funded FDV services.
- A change to the ‘Services and assistance’ received item to separately identify clients who received services for ‘FDV victims’ and/or services for ‘FDV perpetrators’.
- The inclusion of an FDV services category to the ‘Type of service requested’ in the unassisted persons component of the SHSC.
These changes have not resulted in any break in time series as the addition of the new FDV referral item has only increased the number of FDV clients by 0.5%. The change to collect both victim and perpetrator services can also be combined to reflect the general FDV services item used prior to 1 July 2019.
The data distinction between victim and perpetrator services has not been widely reported in 2020–21. This is due to the change being implemented differently between agencies causing data quality issues, as well as the small numbers of perpetrator services provided resulting in confidentiality issues.
Culturally and Linguistically Diverse (CALD) Clients: On 1 July 2019, a question was added to record a client’s main language spoken at home. A follow up question was also added for those who did not speak English as a main language at home. This second question collects information on how proficient the client feels they are at speaking English. These questions have been added to strengthen the data collected regarding clients from CALD backgrounds.
Clients with disability: On 1 July 2019 a National Disability Insurance Scheme (NDIS) indicator was included to gain an understanding of the number of clients who were receiving an agreed NDIS package of support at the time of presentation to an SHS agency. This indicator was introduced to strengthen information regarding clients with a disability.
Imputation and weighting
Data from 2011–12 to 2016–17 are weighted to account for variable rates of agency response and SLK validity. However, due to improvements in agency response and SLK validity rates, data for 2017–18 onwards are not weighted. Unweighted data for 2017–18 onwards are directly comparable with weighted data for 2011–12 to 2016–17. The removal of weighting does not constitute a break in time series.
The annual SHS report and accompanying products presenting financial year data for 2011–12 to 2016–17 are weighted. However, other AIHW publications that analyse the pathways of individual clients over time periods more than a single financial year and publications using SHS data linked with data from other collections, do not use weighted data.