Virtual Expert Group Meeting on Statistical Confidentiality and Microdata Access in the Caribbean
Work area(s)
Event information
Date
3 Dec 2026, 10:00 - 12:30Event type
Participation
Address
Trinidad & Tobago
It is a well-established principle that statistical offices should seek to provide access to microdata for research purposes, to the extent that this is possible bearing in mind the need to protect statistical confidentiality. In 2007, the United Nations Statistical Commission approved Principles and Guidelines for Managing Statistical Confidentiality and Microdata Access (UNSD, 2007). Core principles included that access should only be provided for statistical purposes, that confidentiality should be protected, and that procedures for researcher access, as well as the uses and users of microdata should be transparent, and publicly available. In 2010, CARICOM’s Standing Committee of Caribbean Statisticians (SCCS) adopted a position on access to microdata, encouraging its provision under controlled conditions and highlighting the need for capacity building to enable the establishment of mechanisms for disclosure prevention.
Some capacity building activities have been undertaken, including regional workshops in the Bahamas in 2019 (led by Statistics Canada as part of the PRASC project) and in Trinidad and Tobago in 2024 (organized by CDB, OECS and the World Bank). ECLAC has also provided technical guidance and support in the anonymization of survey microdata on violence against women and census microdata (Jones and Fox, 2016). These activities were intended to support Caribbean statisticians in managing access to microdata, including the application of Statistical Disclosure Control (SDC) methods to prepare datasets for dissemination. Some Caribbean statistical offices have established policies and/or procedures for providing access to microdata, however, the majority operate on a more informal basis. The absence of transparent policies and procedures results in uncertainty, excessive caution, and generally hinders research access to microdata.
There are widely recognised tools and guidance available for the anonymization of microdata that have been developed and updated over many years. The most well-known are the open-source software packages, sdcMicro (an R-package developed with World Bank support) and μ-Argus (developed and supported by Statistics Netherlands). These tools are accompanied by extensive documentation on the theory and practice of SDC (Benschop and Welch, n.d.a, n.d.b and n.d.c; Hundepool and others, 2025). While some training has been provided to Caribbean statisticians, capacity constraints and the lack of practical experience in applying these methods and tools in Caribbean statistical offices continues to be a barrier to the dissemination of microdata.
However, it should be emphasized that minimizing disclosure risk is not only a matter of applying SDC methods to prepare safe microdata. Proponents of what has been called a “user-centred” approach to managing disclosure risk argue for greater focus on the rules and administrative procedures which govern the release of microdata and which determine, for example, who has access to microdata, how they access it, and under what conditions (Ritchie, Hafner and Lenz, 2019). This approach has been formalized as the Five Safes framework (safe data, safe projects, safe people, safe settings and safe outputs), which is a set of principles for managing access to microdata data which explicitly recognizes that the policies and procedures governing access to microdata are at least as important for minimizing risk as the preparation of the microdata itself. The Five Safes framework (United Kingdom Government, 2025) was originally developed by the Office for National Statistics and has since been adopted by the Australian Bureau of Statistics, Statistics New Zealand, Statistics Canada and Eurostat, among other statistical offices, data archives and research organizations. It is a framework that could usefully guide Caribbean statistical offices in the development of policies and procedures for microdata access.
In principle, this can include household survey and census microdata, business statistics microdata and administrative microdata. Disclosure risks are generally easiest to manage for household surveys because the sampling fraction provides a measure of protection against disclosure risk. This is not the case for census or administrative data, while business microdata presents its own challenges, primarily the difficulty of protecting against the re-identification of large businesses. Bearing in mind the limited resources of Caribbean statistical offices, the development of policies and procedures for providing access to survey and census data are likely to be the areas where there is the most potential for progress in the short term.
The meeting will take place online. ECLAC will present findings from a draft study on Statistical Confidentiality and Microdata Access in the Caribbean. Participants representing statistical offices, international organizations and academia will consider the issues raised and discuss recommendations to promote safe access to official microdata.
[email protected] , (868) 224-8071