Dissemination has been supported by Leverhulme award 20050099.

Introduction

Methodology

Summary of outputs

How to order custom datasets

Online data for electoral wards

Many analyses of change over time are simply not possible without a consistent and detailed set of population estimates using the same age and area classifications for 1991 and 2001.

Since census statistics are neither wholly complete nor comparable over time with more severe distortions as the scale of area units become smaller, members of the University of Manchester (Cathie Marsh Centre for Census and Survey Research) and the University of Leeds (School of Geography) have prepared new census-based datasets which allow consistent comparisons for sub-national areas with age and sex in England and Wales, Scotland and Northern Ireland and by ethnic group in England and Wales. 

These new estimates can help you tell how population age structure and ethnic composition changed in the 1990s for any up-to-date community, neighbourhood, postal or ward boundaries. Apart from analysis of past change they serve as the base for your projections and forecasts. There are no other consistent population estimates across time for areas smaller than districts, nor for ethnic composition. 

The change in population between 1991 and 2001 can be different from that shown by census counts, because of the enhancements implemented (see methodology). The differences between census and population estimates in 1991 and 2001 reflect changes in the population definition, addition of non-response not estimated within census output and harmonisation of geographical units.

Population studies and a broad range of social, employment and health studies gain from the improved data source. Many applications require population denominators in epidemiology, employment, crime and other substantive studies. Any study of service demand or of discrimination relies on an appropriate population for each group compared. 

The following table shows an example of the impact of adjustments to census output for the District of Birmingham as a whole, where a gain in population according to the census as published is in fact a slight decrease when complete estimates are used. 

Birmingham District total resident population: Census and population estimate: 1991, 2001 and decadal change

By following the links below, you can display the enhancements to the 1991 and 2001 Census output for the 2004 electoral wards in the District of Birmingham for the total population. These link to Excel woorkbooks containing three worksheets (Graph, Pivot and Data).

The Graph worksheet for the enhancements to the 1991 Census displays the following: (1) the impact of timing and non-response after the revision of mid-1991 population estimates in the light of the latest available data post-2001 Census; and (2) the impact as a result of transferring students from vacation address to term-time address.

Enhancements to the 1991 Census: 2004 electoral wards, Birmingham District

The Graph worksheet for the enhancements to the 2001 Census displays the following: (1) the adjustment for timing from Census day to mid-year; and (2) extra non-response not included in the 2001 Census output.

Enhancements to the 2001 Census: 2004 electoral wards, Birmingham District

Note that by selecting the Pivot worksheet you can choose any ethnic group you like and this will create a new graph on the Graph worksheet. Using the drop-down list box click on any group and the graph on the Graph worksheet will change accordingly.