Statistical Disclosure Control

School of Geography, University of Leeds


Principal investigators

Stan Openshaw
Phil Rees

Other researchers

Oliver Duke-Williams

Dates

1st February 1996 - 31st July 1998

Grant

EEC

Summary

This project saw the development of a software system, DIFFGEOG, assessing the confidentiality risks of publishing more than one set of small area statistics (SAS) from a UK census.

The research showed that previous Census Office practice of publishing SAS for postal sectors as well as for enumeration districts did not create "differenced areas" with sub-threshold numbers of persons and households. However, production of regular grid square did post risks, but theses could be avoided by using quadtree grid squares adjusted so that every grid cell had more than the threshold numbers. The research invented new spatial concepts such "inner and outer haloes" around standard and requested geographies: these are zones formed by intersecting two systems of geographical areas for which statistics can be produced through "differencing". The methods are currently being used in a project for the Northern Ireland Statistics and Research Agency, which intends to publish both output area and grid square data from the 2001 Census and needs to develop grid square designs for which it is safe to release SAS.

Main results

For further information see...

Duke-Williams, O. and Rees, P.H. (1998) 'Can Census Offices publish statistics for more than one small area geography? An analysis of the differencing problem in statistical disclosure' International Journal of Geographical Information Systems, 12, 6, 579-605. ISSN 1365-8816.


[School of Geography homepage] [Leeds University homepage]