Homelessness and rough sleeping risks and drivers in Great Britain: a composite survey modelling approach

Despite growing levels of homelessness and increased policy attention to it across Britain, modelling to highlight individual risk factors alongside broader drivers has been hampered by the obvious fact that many people experiencing homelessness are not clearly in a household and hence not captured by mainstream household surveys. Framed within a critical realist ontology, a composite survey approach is developed, combining a specialized survey targeting people at risk of severe deprivation with a major national household panel dataset, to enable predictive models to be developed using data which includes significant representation of hard-to-reach and non-household populations. Models predicting homelessness and rough sleeping are derived and compared, highlighting the roles of key individual and structural factors, appropriately sequenced in time and able to interact. Vignettes are used to show how the risks vary dramatically between households in different situations, while the potential role of such models in micro-simulation or prediction of impacts of different scenarios is discussed.