Methods for statistical disclosure limitation have appeared in the published literature for over twenty years and in many cases have been incorporated into the practices of national statistical organizations. This paper presents the authors' views on establishing a research agenda that combines agency needs and potential advantages to respondents and data users. Separately for tabular data and microdata, the disclosure problem is discussed, methods and current practice are reviewed, and an agenda for research is presented. Issues regarding the effects of disclosure limitation on data quality and usefulness are examined. Disclosure limitation methods regarded as superior and suitable for general use are identified and the development and dissemination of corresponding statistical software is recommended.
