There is widespread interest in the social sciences in the research potential of two emerging sources of data: administrative data from government agencies and transaction data from online services. In both cases, these data are likely to be covered by non-disclosure agreements that will prevent researchers from sharing the data. We believe that there are ways to share confidential data by combining analysis of disclosure risks, data security measures, and legal agreements that bind researchers to high standards of ethical conduct.
To maintain norms of open access to data in the face of legitimate concerns by public and private data owners, we must develop standards and practices that will make data more widely available. As a step toward developing such standards and practices, we proposed to learn more about the problem.
First, we identified publications based on administrative and transactional data likely to be covered by non-disclosure agreements. ICPSR’s bibliographer searched journals to discover publications likely to be based on data from each kind of source. Second, we contacted the authors of these articles, requested the terms of their agreements with the data owners, and when possible conducted brief interviews during which we elicited information about the processes by which researchers got access to the data, negotiations around publication of findings, etc. Third, we worked with legal advisors at the University of Michigan to analyze the texts of these agreements in order to identify key concerns and provisions most likely to hinder open access.
As a final step, a report was prepared by University of Michigan legal counsel summarizing the findings of their review of 36 agreements and recommending ways that such contracts might be improved. In addition, a model was developed for others to customize when creating data sharing agreements.