Microaggregation and model based methods for disclosure limitation and their application to business microdata
Luisa Franconi
Servizio della Metodologia di Base per la Produzione Statistica, Istituto Nazionale di Statistica, Rome, Italy
and Julian Stander
Department of Mathematics and Statistics, University of Plymouth, UK
Summary.
We briefly define disclosure limitation. We outline a new method for disclosure limitation,
illustrating our approach on microdata from the Community Innovation Survey. We review the
microaggregation approach to disclosure limitation. We explain how microaggregation does
not itself offer protection to the variable geographical area, whereas the new method
suggests how to define broader categories for releasing this variable. We discuss how
to assess both the amount of protection offered and the error induced by a disclosure
limitation method. We find that for the four NACE main economic activities considered
the new method offers more protection than microaggregation, and often leads to a smaller
error.
Keywords:
Area effects; Community Innovation Survey; Microdata for research;
Performance assessment for disclosure limitation methods; Prediction intervals
Address for correspondence:
Julian Stander, Department of Mathematics and Statistics, University of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK.
Publication:
This paper has been submitted to the JRSS (Journal of the Royal Statistical Society) series D for publication
|