The First Minister has stated the intention to end youth homelessness within a decade in Wales. To achieve this our collective understanding of both the causes and impacts of youth homelessness must improve.
Schools provide an environment where children, young people and their families can be supported and homelessness could potentially be prevented. So a primary area of interest lies in the relationship between education and homelessness. However, the information that exists on links between education and homelessness is based on small, qualitative studies and convenience sample surveys which do not quantify the impacts of homelessness on education.
This report presents findings of an exploratory study that had two aims:
- to test a methodology for identifying children and young people living in homeless households, linking together statutory homelessness and General Practice registration data
- to begin to explore the educational achievement of children and young people living in homeless households
This analysis relates to children and young people in homeless households, rather than children/young people who have made a homeless application themselves.
2. Preliminary findings
This analysis reflects children/young people living in households who were assessed by the local authority and were legally defined as being homeless. It should be noted that this definition includes households in a number of different homeless situations, and therefore there may be differences in education outcomes within the homeless definition used that this analysis may mask. Only the first education outcome post-homelessness is reported on.
In total, data relating to assessments and exams for 971 children and young people following the application for assistance could be identified through linking together the homelessness and education attainment data.
Table 1 provides an overview of the proportion of children achieving the expected outcomes for the core subject areas at each level. For foundation phase to Key Stage 4, the expected outcomes are known as the foundation phase indicator and core subject indicator. To provide context to the education outcomes of the homeless group, Table 1 includes the national outcomes for foundation phase 1 to Key Stage 4, split by Free School Meal status (FSM). National data are only available for single years, and the homeless cohort relates to the whole period of assessment (2012 to 2016). Therefore, direct comparisons are not possible and the data are shown for illustrative purposes only. Additional column includes outcomes for homeless group for the combined period 2012 to 2016.
A number of other studies on the effects of homelessness on aspects of education, all from the United States, have used eligibility for FSM as a proxy for poverty. This was then used in creating comparison groups with homeless young people. The FSM eiligibility status in the Welsh education data is taken to be an adequate, rather than a perfect measure of socio-economic disadvantage. Furthermore, it is the only proxy for socio-economic disadvantage at an individual level available within the education data set.
As illustrated in Table 1, the proportion of the homeless cohort who achieved the core subject requirements at each of the different levels appears to be more similar to the outcomes of FSM eligible students than those not eligible for FSM. The finding that FSM and homeless students may have similar levels of attainment could reflect the links between homelessness and poverty, and/or that students who were homeless were also therefore FSM eligible. The extent to which the homeless cohort were FSM eligible is something that will be explored in future analysis.
|Year of teacher assessment/exam||Homeless cohort|
|2012||2013||2014||2015||2016||2012 to 2016|
|Foundation phase indicator|
|Eligible for FSM||66||69||72||75||76||-|
|Not eligible for FSM||85||87||89||90||90||-|
|Key Stage 2 CSI|
|Eligible for FSM||67||70||72||75||77||-|
|Not eligible for FSM||87||88||90||91||91||-|
|Key Stage 3 CSI|
|Eligible for FSM||48||54||61||66||69||-|
|Not eligible for FSM||78||82||86||88||90||-|
|Key Stage 4 CSI|
|Eligible for FSM||22||23||27||29||33||-|
|Not eligible for FSM||55||55||59||61||64||-|
Source: Education achievement data from national teacher assessment and PLASC, Welsh Government
 Brumley, B. et al. (2015) The unique relations between early homelessness and educational well-being: An empirical test of the Continuum of Risk Hypothesis. Children and Youth Services Review. 48:31-37 Canfield, J.P. et al. (2016) Using a Person-Centre Approach to Examine the Impact of Homelessness on School Absences. 33:199-205; Deck, S.M. (2017) School outcomes for homeless children: difference among sheltered, doubled-up, and poor, housed children. Journal of Children and Poverty. 23(1): 57-77
 Taylor, C. (2018) The Reliability of Free School Meal Eligibility as a Measure of Socio-Economic Disadvantage: Evidence from the Millennium Cohort Study in Wales. British Journal of Educational Studies. 66(1): 29-51
3. Overview of methodology
This research linked three administrative data sets held in the Secure Anonymised Information Linkage (SAIL) Databank.
- Statutory homelessness data from the City and County of Swansea
- Education attainment data (National Pupil Database and National Data Collection)
- the Welsh Demographic Service (WDS)
The homelessness data used in this study relates to applications, or ‘cases’, made by adults seeking help under homelessness legislation. In this study, an adult is defined as anyone aged 18 years or older. This analysis relates to children and young people in homeless households, rather than children and young people who have made a homeless application themselves.
Data initially available for analysis related to all assessments under Welsh housing legislation where the case for help ended between January 2012 and March 2017. This equated to 16,971 cases or applications under homelessness legislation to the local authority.
One of the study’s key challenges was that the statutory homelessness data available for this project only related to ‘adult’ applicants for help. This is problematic given the need to identify children’s education outcomes. Anonymised data from the WDS were used to identify children living in homeless households.
The WDS provides a history of residences where a person has registered as living, updated every time they sign up to a General Practitioner (GP) in Wales or change their address with their GP. The WDS can be used to identify children and young people living with an adult applicant on the date of the application for homelessness assistance.
However, rather than taking all children and young people who were identified as being ‘co-resident’ with a homeless applicant, only those who registered or de-registered from an address on the same date as the main applicant were included. The assumption was that these individuals have some form of association, more so than simply living in the same residence. This constraint on who was deemed co-resident improved the accuracy of the method for identifying children and young people associated with a case. As a final measure to ensure that the children and young people were correctly identified as homeless, we only retained children living in households where local authority data also indicated that the household contained dependents.
The final sample, after data cleaning and processing, contained 2,099 cases. When compared to the original population of cases, 44% of cases with dependents were successfully identified through the administrative data method. Although the sample size is reduced, the measures taken to ensure that homeless children were being correctly identified limits any validity issues from including non-homeless children and young people in the homeless group.
The education outcomes for these children and young people for the period 2012 to 2016 were removed from the education data held in the SAIL Databank. As foundation phase assessments replaced Key Stage 1 in 2011/12, the time period 2012 onwards was used. This reduced the need to account for changes in assessments at the age of 6 to 7 years old. Education data used in this analysis related to foundation phase, up to and including Key Stage 4 (or GCSEs), and therefore covered the period of compulsory schooling.
In the United Kingdom there is scant quantitative evidence of the links between education and homelessness; insight mainly comes from small qualitative samples of the experiences of homeless households.
This study used linked administrative data to begin to explore the educational achievement of children and young people living in homeless households.
Preliminary analysis has the potential to show that achievement against expected outcomes in the core areas of learning for children and young people living in homeless households appears to be lower than the national average. However, their achievements are potentially in-line with young people who are eligible for FSM. This finding will be looked at in further analysis.
This project makes a methodological contribution in providing a method for the identification of children and young people in homeless households where only the adult head of household is recorded. Though this methodology is a pragmatic approach for analytical purposes, the use of indirect means of inferring household structure and identifying other householders is far from perfect. This is especially true for the purposes of accurately assessing the scale of homelessness in Wales.
The main recommendation from this report is to encourage service providers, if they do not do so already, to collect personal data on all household members. Having personal data on all householders has both practical benefits, in enabling services to identify people returning to them, in addition to opening up the potential for data linkage research. Until such point as homelessness services record data on all household members, this methodological development will be key to any future homelessness data linkage research.
5. Future releases
This report has been supported by funding from the Welsh Government Homelessness Prevention Fund.
Two further projects are planned within the area of education and homelessness; the first directly builds upon the exploratory study to compare education attainment of housed (deprived/not deprived) and homeless young people. The second project will explore the relationship between school attendance rates and homelessness status. This is of interest because there is a large body of mainly international evidence that suggests homelessness is associated with school disengagement, including absence. It may therefore be possible for the school to act as a site of early prevention and as a venue through which work can be done with families experiencing housing precarity.
This work has been carried out by Dr Ian Thomas and Dr Pete Mackie at WISERD, as part of Administrative Data Research (ADR) Wales housing and homelessness body of work.
ADR Wales is part of the Economic and Social Research Council (part of UK Research and Innovation) funded ADR UK.