| This study addresses the problem of temporally unstable areal units when attempting to conduct spatio-temporal analyses. Between the 1991 and 1996 censuses Enumeration Area definitions changed dramatically, frustrating many analysts including the Toronto Inner City Time Trends Working Group. The problems faced by this health research group are outlined and reviews of some potential solution methodologies are given.; Block-face data are introduced as a solution to the problem of incompatible zonal systems in Canada. In addition the relative accuracy of two simple areal interpolation methods, which are universally available in commercial GIS software, are examined. The effects of employing these different interpolators are analysed with reference to population data and the potential effects that may be incurred in subsequent health analyses. This study concludes that the choice of interpolator can have dramatic impact on the re-distribution of data across zonal networks, and thus the accuracy of subsequent analyses is questioned. |