Restricted Data Access: Principles and Standards

Governments mandate their National Statistics Offices to collect empirical data to determine appropriate policies. Re-use of this data for research can provide input regarding the effectiveness of government action. In Western Europe and North America policies and institutions support the efficient collection and sharing of official data for research purposes. In Africa the sharing of government microdata is constrained by several obstacles. African National Statistics Offices have limited resources to curate microdata and ensure its long-term availability. Consequently many African data producers do not follow international best practice with regard to survey data management or share the microdata from the surveys they conduct. This was confirmed by a survey conducted in order to investigate the availability of survey microdata from African National Statistics Offices for research. A further obstacle to access to government microdata in Africa is inadequate producer-user communication channels. Concerns around the confidentiality of respondent information also present a barrier to data usage for research, as does the bureaucratic nature of government institutions involved in data production in African countries. Access to official microdata for research requires sound data usage policies driven by African decision-makers who appreciate the role of information utilisation in national development.

The European Union has long faced the problem of how overcome the challenge of sharing data across borders for effective cross-national research. One of the key issues is a harmonisation of standards and protocols. The ESSNet project funded by Eurostat has developed a protocol for statistical disclosure that can be implemented by all european member states. This presentation will outline some of the challenges faced, the standard developed, and will look how the standard might be used to change the european data infrastructure in the future.

Access to restricted microdata for research is increasingly part of the data dissemination strategy within countries, made possible by improvements in technology and changes in the risk-benefit perceptions of NSIs. For international data sharing, relatively little progress has been made. Recent developments in Germany, the Netherlands and the US are notable as exceptions. This paper argues that the situation is made more complex by the lack of a general coherent risk-assessment framework. Discussions about whether something should be done become sidetracked into discussions about how procedural issues would constrain implementation. International data sharing negotiations quickly become bilateral, often dataset-specific, and of limited general value. One way forward is to decouple implementation from principles. A principles-based risk-assessment framework could be designed to address the multiple-component data security models which are increasingly seen as best practice. Such a framework allows decisions a out access to focus on legal-procedural issues; similarly, secure facilities could be developed to standards independent of dataset-specific negotiations. In an international context, proposals for classification systems are easier to agree than specific multilateral implementations. The paper concludes with examples from the UK and cross-European projects to show how such principles-based standards could work in practice.

Where access to data are controlled by a data service or depositor, their use may be restricted to individuals who are able to adhere to certain conditions regarding their use. These conditions may require the applicant to store, use or limit sharing in particular ways. In the UK such conditions have become more commonplace with the growth of special licences and securing settings for government microdata. In order to assess the potential impact of these conditions on the usability of data, a survey was conducted to obtain data from a representative sample UK social scientists in ten disciplines to better understand their working environments and practices during the autumn of 2009. A 61% response rate was achieved resulting in over six hundred completed questionnaires. The survey covered a range of questions which included, access to computing facilities, including data and printout storage, home working, data transportation, awareness of institutional policies on personal data, and attitudes to a range of access conditions.