On January 28, student, staff and faculty from across UMass gathered to reflect on the changing expectations and implications they face in regard to data management--exploring the tensions and opportunities embedded in competing calls for transparency and confidentiality in social science research.
In her framing remarks, Krista Harper (UMass Amherst/Anthropology) noted that new NSF data management and archiving requirements require researchers to make much more of their data available. "At the same time," she noted, "research participants’ expectations of confidentiality--or, conversely, recognition--are changing in a moment where 'public social science' and engaged research are on the rise. In everyday life, ubiquitous social media and mobile devices create new data as well as quandaries for privacy. These transformations affect all social science research--from quantitative and computational social science projects drawing on 'big data' produced as an artifact of new technologies to qualitative research with conflicting demands to share data with the public while maintaining confidentiality."
Three panelists with deep and varied experiences in developing and evaluating data management plans offered rich perspectives to spark participant reflection. ISSR Director Laurel Smith-Doerr (UMass Amherst/Sociology) offered a historical context of rising calls for transparency in public-funded research, and an inside look at the debates among scholars advising the NSF Program on Science, Technology and Society as it develops its new guidance for data management planning (DMP). Smith-Doerr emphasized the imperative for approaches to DMP to reflect the realities of each research project--protecting confidentiality and relations of trust where necessary, and rewarding researchers who invest resources in producing and publishing original data that others may subsequently use. Janet Vertesi (Princeton/Sociology) took up this thread, offering the concept of "data economies" to mark science's social relations, which shape whether and how researchers might share data. When the production of data relies on the currency of trust (among researchers or with subjects), it circulates (or does not) with a value intrinsic to those relationships. Without data management standards that recognize diverse data economies--and support a tiered approach to making data public--scientists struggle with contradictory imperatives enshrined in IRBs and DMPs, and in the pursuit of academic publishing and careers. Brian Schaffner (UMass Amherst/Political Science) shared his own experience as an NSF reviewer trying to apply general guidelines across divergent methodologies and disciplines, and as a principal investigator balancing these demands himself. As more and more projects draw on multiple (and not always public) data sources, Schaffner noted that the increasing trend among top-tier journals to require publication of data in the name of scientific progress can disadvantage researchers who cannot--for ethical or professional reasons--disclose their data.
These messages resonated strongly with the event's participants, including scholars in the social and natural sciences, as well as library and foundation relations staff who support grant preparation and data management. Key concerns related to the need for flexible and appropriate standards for data management as well as for adequate resources in grant budgets and time frames for the additional work needed to prepare data for public release. Others called attention to the particular squeeze that transparency requirements place on ethnographic and participatory researchers, as well as on early-career scholars under pressure to develop several publications from the data they collect. The panel thus called for active engagement with IRBs and funders, to develop responsible and feasible approaches to data management and sharing, and to align funding, technologies, and career pathways in a supportive institutional architecture for social science research. Vertesi noted opportunities to get involved in multiple groups grappling with these and other issues, including the Council for Big Data, Ethics and Society, which is currently calling for new members and offers a recent post and report on federal funders' data management plan requirements.