Introduction: data sharing and disclosure limitation techniques

THIS SPECIAL SECTION OF JERHRE is in response to the needs of institutions to develop advanced data sharing capabilities. On October 1, 2003, the National Institutes of Health (NIH) initiated a requirement that investigator-initiated proposals for grants with direct costs over $500,000 in any year incorporate plans to accommodate sharing research data. The requirement stipulates that such plans describe the procedures through which shared data would be rendered "free of identifiers that would permit linkages to individual research participants and variables that could lead to deductive disclosure of the identity of individual subjects." (http://grants2.nih.gov/grants/policy/data_sharing). We expect that many researchers who deal with human research data are unfamiliar with the procedures presented in the ensuing articles. These sophisticated procedures have been developed to help protect confidentiality of subjects' data in files that are shared, while simultaneously preserving the analytic value of data for secondary users. Among these are procedures developed by government statisticians that include innovative methods to prevent deductive disclosure of identities. More recently, academic researchers and data experts have adapted or extended these methods. Together, these methods aim to achieve both disclosure limitation and retention of key analytic usefulness of the shared data.