Geodemographic classifications represent the multidimensional socio?economic characteristics of people living within defined areas. They have been successfully applied in both private and public sector applications across a range of industries since the 1970s. In UK geodemographic classifications, neighbourhoods are predominantly defined using a full postcode or the 2001 Census Output Areas.
Little research to date has been completed on: the potential loss of information one may experience using classifications derived at courser geographical aggregations; the most appropriate scale through which concepts of 'neighbourhood' can be measured; or how the characteristics of private sector data sources affect classification structure. This paper presents the results of a comparative analysis between the National Statistics Output Area Classification (OAC) and the commercial postcode level classification Mosaic. It assesses the degree to which commercial classifications constructed at the scale of unit postcode leverage greater insight about the composition of neighbourhoods over those built at output area. It is argued here that the inclusion of additional data at finer spatial granularity in commercial classifications do present added information about the composition of neighbourhood areas; however, the information loss experienced by switching to classification using output area is reasonably low. These losses are also shown to be socially and spatially heterogeneous. Finally, concerns are raised over the spatial and temporal reliability of those data which are used to inform commercial unit postcode classification, and solutions are offered as to how public sector users may maintain methodological robustness whilst retaining the advantages of finer scale neighbourhood classification.
This working paper is available as a PDF. The file size is 1544KB.
Authors: Alex Singleton
Publication Date: 1/12/2007