G-Confid (Disclosure Avoidance - Generalized System)

G-Confid is a Statistics Canada generalized system which prevents the release of confidential information in tabular data using the method of cell suppression. Preventing the release of this information is done by identifying the sensitive cells to be protected and finding appropriate complementary cells to suppress in order to protect the sensitive cells. In addition G-Confid may be used to audit cell suppression patterns, find sensitive aggregates and round tabular data.

To determine the level of protection needed by a given cell, the cell’s sensitivity is calculated. G-Confid may be used to calculate a cell’s sensitivity according to commonly used sensitivity measures including the p-percent rule and the (n,k) rule, as well as the sets of rules used within Statistics Canada. Arbitrary linear sensitivity measures may also be calculated. In addition G-Confid may be used to determine sensitive aggregates of cells that are not necessarily the margins of any table.

To protect the sensitive cells, additional cells may need to be suppressed. G-Confid uses a linear programming algorithm to find complementary cells to suppress that protect the sensitive cells while attempting to minimize a cost function. G-Confid will also take into account cells which are forced to be published or suppressed. The weights used in the cost function may also be specified.

G-Confid consists of six components: one SAS procedure and five SAS macros. All components are built in-house and work the same way as other SAS procedures and macros. The components are easy to use and are well documented. They offer a methodology that is known and approved at Statistics Canada. The software is fully supported by a team of programmers and methodologists who continuously work at improving the software and are ready to offer technical and methodological support.

Statistics Canada also offers a version of G-Confid for use with SAS Enterprise Guide.