Investigating behavioural impacts of poor neighbourhoods: towards new data and analytic strategies

Three challenges confront statistical researchers of neighbourhood impacts on individual behaviours: (1) operationalising 'neighbourhood processes'; (2) potentially non-linear relationships between neighbourhood characteristics and outcomes; and (3) the selection bias problem. To better comprehend these challenges and overcome them, the paper proposes an overarching conceptual framework wherein outcomes of interest are affected by neighbourhood interacting in a mutually causal fashion with housing tenure, housing wealth, household socio-economic status, and mobility behaviour. It advances a five-equation, simultaneous system for home ownership, mobility expectations, housing wealth, household socio-economic status and neighbourhood character. Although current US census data do not provide perfect proxies for neighbourhood processes, there is evidence that a battery of them could represent reasonable operationalisations. Tests for non-linearity could be conducted in this framework. This model could use a sufficiently robust set of instrumental variables to overcome the issue of neighbourhood selection bias, and thereby produce considerably more precise estimates of neighbourhood impacts on individual outcomes of interest. Implications for qualitative research approaches also are drawn.