What is the research?
A project is being carried out to investigate further the differences between the Census and some administrative data sources. The aim of the exercise is to evaluate the accuracy of the various household lists considered, with a view to improving intercensal population estimates in the future. A matching study is currently being conducted between ONS and Manchester, with a forthcoming study also planned for Westminster.

What is the framework of this exercise?
The work is in two phases. Phase 1 will match individual addresses from three main sources:

1. Public sources, e.g. the Royal Mail "postal address file", OS "addresspoint"
2. Administrative sources held by two local authorities (Manchester and Westminster), e.g. council tax registers, electoral registers
3. ONS census addresses

The results from phase 1 will then require further investigation to understand whether discrepancies are due to definitional differences or real differences in numbers of households. This work is yet to be defined in detail, but will involve extensive manual investigation.

It is important to note that a comparison exercise of this nature is not trivial. A report was compiled following an initial comparison between the Census household list and the address list supplied by Westminster City Council. The report has highlighted the fact that straightforward numerical comparisons between Local Authorities Property Lists and the Census are misleading and unreliable. The definitional differences, particularly with regard to communal establishments, mean that even large differences between the two datasets can be explained and do not necessarily reflect inaccuracies on either part.

Who will be carrying out this work?
For the Manchester study, phase 1 is to be undertaken by Manchester Geomatics under contract to ONS, and phase 2 will be undertaken by ONS. Manchester City Council are funding Manchester Geomatics for preparation and analysis of datasets provided by them.

What are the timescales for this work?
Phase 1 for the Manchester study began at the end of March 2003 and will last two months. The duration of phase 2 depends on the findings of phase 1 and is likely to last another two months. This gives an estimated completion time of the end of July.

This page last updated 01 November 2005