Data Protection Impact Assessment and Quality Assurance Statement for HL1, HL2, HL3 and PREVENT1 data collections

Quality assurance of Scottish homelessness data

March 2019, reviewed February 2023

This document contains information on the administrative data sources used by homelessness statistics and analysis team in the Scottish Government. It provides background to the data collections along with details about the steps the Scottish Government Homelessness Statistics and Analysis team takes to quality assure the statistics it produces.

Background to the statistics

The homelessness statistics and analysis team publishes statistics relating to homelessness in Scotland twice each year. These are an annual publication produced in summer, and a 6-monthly update produced in January/February. In addition we produce an annual publication covering activities relating to Housing Options services provided by local authorities when households seek assistance for housing related issues.

The data we use to produce these publications comes from four separate data collections.

• HL1 - Data on homelessness applications are collected on the HL1 return, a continuous case level electronic data capture system. This case level data allows analysis of applications and breakdowns by combinations of characteristics. Cases can also be grouped to household level to identify instances of repeat homelessness.
• HL2 - Data on households in temporary accommodation at the end of each quarter is collected on the HL2 form which is completed by local authorities and records summary snapshot details.
• HL3 - Data on households in temporary accommodation is collected in the HL3 data collection at case level. This includes information on the type of temporary accommodation offered and taken up by households as well as the dates households entered and exited temporary accommodation. It also collects information about placements that were unsuitable or in breach of the Homeless Persons (Unsuitable Accommodation) (Scotland) Order 2017. Records about temporary accommodation also link to homelessness applications as recorded in the HL1 data collection.
• PREVENT1 -This data collection collects data relating to the implementation of Housing Options work by local authorities. It includes information about the reasons for a Housing Options approach, the type of prevention activity offered, information about each individual within the household and information about the outcome of the case. It also includes information about whether the Housing Options approach links to a homelessness application that has been recorded in the HL1 data collection.

The data source and supplier

In the case of all four data collections the data suppliers are the 32 local authorities in Scotland. Their data returns are based on data extracts from their own case management systems. These systems are used by local authorities to administer their duties towards homelessness.

Risk/profile matrix

In order to determine the types of assurance and documentation required for Scottish homelessness statistics we utilise the Quality Assurance (QA) matrix outlined in the UK Statistics Authority Administrative Data Quality Assurance Toolkit. This matrix then informs the level of assurance required for each of the four main practice areas relating to quality assurance of administrative data, as outlined in the UK Statistics Authority Administrative Data Quality Assurance Toolkit. The four areas are:

• Operational context and data collection
• Communication with data supply partners
• QA Principles, standards, and checks by data suppliers
• Producers’ QA investigations and documentation

The public interest profile of homelessness statistics in Scotland can be considered to be high, given the political sensitivity of homelessness, the levels of media coverage it receives, and the important public health issues.

The data quality concern can be considered low, given the clear agreements with data providers about what data will be provided, and the appreciation for the context in which data are collected. The operational nature of the underlying data requires that the data be of sufficient quality for operational use. In addition a large number of data validation checks are carried out by the Scottish Government, and we also compare data with previous years, raising queries with local authority staff as required.

Information for how we meet the required level of assurance for each of the four areas of practice is provided on the next page.

Practice area 1: Operational context and data collection

Data is collected by local authorities in the course of the work they carry out supporting households who are homeless or threatened with homelessness (HL1), providing households with temporary accommodation (HL2 and HL3) or carrying out Housing Options work (PREVENT1).

Information is collected from households at the application stage in each case, recorded on the operational systems of the local authority and updated as the case progresses. Data is then submitted by local authorities to the Scottish Government on a quarterly basis in accordance with the data specification for each collection. These data specifications are available at the Scottish Government’s Homelessness Statistics website.

In the case of HL1, HL2 and HL3 data is submitted to the Scottish Government in csv format via government secure email to the Homelessness Statistics inbox at homelessness_statistics_inbox@gov.scot. PREVENT1 data is submitted in XML format via the Scottish Government’s online data collection system procxed.

Because homelessness applications, temporary accommodation placements and housing options activities can take place over extended periods of time data submitted at the end of each quarter may still be ‘live’, and so may be updated in future data collections. As a result statistics based on this data is based on a snapshot at a particular date and is often revised in later publications.

Practice area 2: Communication with data supply partners

Scottish Government statisticians work closely with data suppliers in local authorities in the course of data collection. Clear data specifications for each data collection have been developed by the Scottish Government and made available to local authorities on the Scottish government website. A data protection impact assessment has also been carried out and is also available online.

Data collection takes place on a quarterly basis and communication between the Scottish Government with regards to quality concerns is ongoing within each collection period. This includes the production by the Scottish Government of quarterly reports summarising collected data and providing comparisons with previous data for quality assurance purposes. Where data quality issues are found local authorities are asked to correct errors and resubmit data. Where issues have been identified by local authorities they will instruct us to correct the data they previously submitted or else resubmit correct data.

In addition, the Scottish Government also engages with data providers to consult them on the implementation of any changes and to keep them informed of the progress with these. Scottish Government also offer to discuss changes in detail and work alongside data providers to ensure the requirements are understood and correctly implemented.

The Scottish Government also coordinates the homelessness statistics user group (HSUG). This is a group of homelessness statistics users and data providers which provides the opportunity to discuss issues relating to homelessness statistics. Before the pandemic, in-person meetings were held twice a each year. Since then, regular written updates with offers to meet virtually have been provided by Scottish Government. The frequency and format of meetings going forward is currently under review, but it is anticipated there will be at least one annual meeting with the opportunity to attend in person.

Practice area 3: QA Principles, standards, and checks by data suppliers

As per the data specification, data suppliers have a number of validations built in to their data collection system.

To support the quality assurance work of data suppliers the Scottish Government homelessness statistics and analysis team produce quarterly reports for each local authority after collection and data validation (outlined in the next section) are concluded. These reports compare submitted data with previous years, enabling any unusual variations to be identified and investigated. Where any errors are identified local authorities will resubmit corrected data.

Practice area 4: Producers’ QA investigations and documentation

For each data collection a large number of automated data validation checks are carried out. These ensure that data submissions comply with the data specification provided by the Scottish Government. In the case of HL1 for example more than 170 separate tests are applied to the data.

In the case of HL1, HL2 and HL3 these validation checks are carried out using a bespoke IT system developed by the Scottish Government statisticians based on the SAS statistical software program. Where any validation errors are discovered an automated report is returned to the local authority, prompting the local authority to correct the issues and resubmit the data.

Data validation for PREVENT1 is carried out using the procxed data collection platform developed by the Scottish Government. In this case details of validation errors are made immediately available to the data supplier and the data will not be accepted by the Scottish Government until all errors are resolved.

After validation is concluded  Scottish Government statisticians carry out further quality checks of the data. We may then approach local authorities with further queries based on these.

We check also for under-reporting of cases in the HL3 by comparing details provided across the HL1 and HL2 datasets.

Limitations

There are some limitations with Scottish homelessness data. In the case of all four collections the data captures information about those households who engage with local authorities, so is likely to under represent the true scale of homelessness in Scotland. For example some rough sleepers or ‘hidden homeless’ (for example those staying with friends) will not be captured in the data.

Because Scottish Government policy has been for Housing Options to be developed locally, there is no single definition of Housing Options operating across all local authorities. As a result there is some inconsistency in how Housing Options activities are recorded within the PREVENT1 data collection, making comparisons between local authorities difficult.