Volume 39 Issue 2, June 2004, pp. 15-28

Spatial analysis and mapping of georeferenced, individual-level data can help identify important geographical patterns or lead to knowledge significant for dealing with specific social issues in a particular area. However, given the need to protect personal privacy when using geospatial data, the possibility for undertaking geographical analysis on certain types of individual-level data is becoming increasingly circumscribed. This article addresses the need to protect geoprivacy while making georeferenced, individual-level data available in such a way that analytical results are not significantly affected. The effectiveness of three geographical masks with different perturbation radii (r) is examined using a data set for lung-cancer deaths in Franklin County, Ohio, in 1999. The findings reveal a rather consistent trade-off between data confidentiality and accuracy of analytical results. There seems to be a threshold r-value at which the results of analyses on masked data become substantially different from the original results. An r that produces an area about the average size of the study-area census-block groups achieves a desirable optimum trade-off between privacy protection and accuracy of results. The study shows that implementing appropriate geographical masks may help data managers or researchers establish the desirable trade-off, in a particular context, between privacy protection and accuracy of geographic information.