Inferring health milieu geographies from Understanding Society and the UK Census

Presenter: Jens Kandt, University College London

Author: Jens Kandt

This paper presents theory, methods and results of combining the substantive depth of Understanding Society with the pervasiveness of 2011 Census neighbourhood statistics to estimate the local prevalence of so-called health behaviours in the context of lifestyle milieus. Drawing on Bourdieu’s theory of social practice, Wave 2 and 3 Understanding Society respondents were clustered based on a range of behavioural and attitudinal variables. Nine milieus were found, which differed strongly in their social and demographic profiles. In addition, the milieus diverged significantly on various age-standardised Understanding Society health variables; this divergence cannot be solely attributed to socio-economic position. There rather appear to be specific health pathways at work with differential impacts. Deterministic spatial microsimulation was used to infer the milieus’ geographic distributions. The technique matches demographics of survey respondents with those of small areas and has so far produced distinct and plausible spatial distributions of milieus in London. The findings suggest that a combined approach of sample segmentation and microsimulation offers opportunities to render subjective information of social surveys amenable to applied spatial statistics. The work is ongoing and will be extended to other UK cities with the additional objective to account for potential place effects in milieu specificity and associated socio-demographics and geographies. Integrating Understanding Society and population-wide datasets may be a promising way forward not only in inferring spatial distributions of survey responses but also in adding domain-relevant subjective orientations to neighbourhood statistics and supporting targeted, area-based policy interventions.