Search

When publishing official population statistics, many national statistical organisations define small areas termed 'building blocks' which are then aggregated into larger 'output zones' for data release. While output zones are known to have enormous influence on spatial analysis, there has not been any systematic analysis of the effect of building blocks on characteristics of output zones. This paper evaluates current international practice in building block design, identifying key conceptual and practical issues. Using the example of six local authorities in England and Wales, it employs automated zone design to evaluate the influence of two sets of building blocks (street blocks and postcodes) on output zone characteristics. Household-level census data, accessed under secure conditions, are used to evaluate the impact on both newly designed and maintained output zones. Postcodes are shown to be more effective building blocks than street blocks, providing more uniform population and household sizes, greater precision for matching postcoded data to census data, and enabling more zones to be maintained. However, street blocks deliver more compact output zones and greater internal homogeneity of tenure and accommodation type. The scale effect of the modifiable areal unit problem and the specific geographical patterning of variables are both shown to be important factors when designing building blocks. These findings have directly informed policies and processes for the 2011 Census in England and Wales and provide useful conceptual and practical guidance for any national statistical organisation or analyst designing their own building blocks. The paper concludes that some aspects of international building block design practice could be more effectively harmonised but that such design should always be nationally specific to incorporate locally varying conceptual and practical issues. Further research should extend this analysis to other building block types, notably grid squares. [ABSTRACT FROM AUTHOR]

Automated zone-design methods are increasingly being used to create zoning systems for a range of purposes, such as the release of census statistics or the investigation of neighbourhood effects on health, Inevitably, the characteristics originally underpinning the design of a zoning system (eg, population size or homogeneity of the built environment) change through time. Rather than designing a completely new system every time substantive change occurs, or retaining an existing system which will become increasingly unfit for purpose, an alternative is to modify the existing system such that zones which still meet the design criteria are retained, but those which are no longer fit for purpose are split or merged. This paper defines the first generic methodology for the automated maintenance of existing zoning systems. Using bespoke, publicly available, software (AZTo0I), the methodology is employed to modify the 2001 Census output geographies within six local authority districts in England and Wales in order to make them suitable for the release of contemporary population-related data. Automated maintenance of an existing system is found to be a more iterative and constrained problem than designing a completely new system; design constraints frequently have to be relaxed and manual intervention is occasionally required. Nonetheless, existing zone-design techniques can be successfully adapted and implemented to automatically maintain an existing system. The findings of this paper are of direct relevance both to the Office for National Statistics in their design of the 2011 Census output geographies for England and Wales and to any other countries or organisations seeking to maintain an existing zoning system. [ABSTRACT FROM AUTHOR]