April 5, 1999
Dear Mr. Charney:
This letter is to provide the comments of the Consortium of Social Science Associations (COSSA) on the proposed revisions to OMB Circular A-110 that were published in the Federal Register on February 4, 1999. The Consortium is supported by over 100 professional associations, scientific societies, universities, and research institutes, with an interest in promoting attention to and federal support for the social and behavioral sciences. COSSA serves as a bridge between the academic research community and the Washington policymaking community. (A list of our Members, Affiliates and Contributors is attached.)
OMB is to be commended for trying to deal with a complex and difficult situation created by the provision in the Omnibus Appropriations Act of 1998, Public Law 105-277. Your attempts to narrow the scope of the legislative mandate are helpful, but many difficulties still remain. We strongly believe that the best way to handle this problem is to enact H.R. 88 that would repeal the onerous provision. Then, the National Academy of Sciences' Committee on National Statistics should be given the mandate to update its excellent 1985 report Sharing Research Data (Fienberg, S., Martin, M. and Straf, M. eds.). The NAS, in consultation with all the players -- federal research agencies, non-profit institutions, the research community, and the public -- should be allowed to develop data sharing policies that meet the information needs of the 21st century.
The 1985 NAS report recommended that "Data relevant to public policy should be shared as quickly and widely as possible." However, it added a caveat: "This recommendation is not intended to support the public release of analyses prior to appropriate review." Our concern is that the new law and the proposed rules, while clearly favoring the first notion, do not pay enough attention to the second.
We agree that the goal of sharing data is a good one and the social sciences have a strong tradition of doing just that. The over-50 year history of our major data archive, the Inter-University Consortium for Political and Social Research, demonstrates that commitment. Many social science data collections funded by federal grants are on deposit there for use by social scientists to replicate, verify and conduct secondary analyses. These data however, are fully cleaned and documented.
The social and economic sciences division at the National Science Foundation took the lead in the mid- 1980s to foster data sharing policies that eventually were adopted by the entire agency. These policies, while stressing the need for sharing data, also recognize the complexity of the research enterprise and the need for flexibility to accommodate the different approaches to data gathering used by scientists. The new law and the proposed rules take a "blunderbuss approach" in treating all scientific methods the same.
We also feel that the Freedom of Information Act (FOIA) is the wrong vehicle for devising a data sharing policy for the whole government. Turning a researcher's work into "agency records" and putting the agency in charge of determining when and whether to release scientific data is not useful to the nation or the scientific enterprise. The research enterprise could be harmed and important research areas "chilled" by the effect of the new law. We agree with NSF Director Rita Colwell that using FOIA procedures creates "a complex web of expensive and bureaucratic requirements for individual grantees and their institutions."
The social sciences utilize a number of techniques that differ in their approach to collecting data and that could suffer from exposure under FOIA requests. In experimental research, individuals are subject to preplanned conditions and their responses are tabulated in some fashion. This research is governed by Human Subjects Protections and Institutional Review Boards. Would FOIA requests supersede these review bodies?
Researchers sometimes develop mathematical and computer models to aid them in the analysis of their data. Would these models be subject to release under FOIA, perhaps before they have been tested? Object based research could include situations where the data are often archaeological specimens or fossil remains. Because these often do not belong to the researchers, (e.g., they cannot remove them from a foreign country), how would the scientist provide such data under a FOIA request?
Qualitative information from microfilms, oral interviews, videotapes, ethnographic or linguistic field notes, recordings or transcriptions or hand written records of open ended interviews are other techniques utilized by social scientists. How would data collected under these conditions fare under FOIA requests? Would raw notebooks have to be turned over? Would videotapes, where privacy protections would be difficult, also be fair game?
Other major tools for gathering quantitative social and economic data sets are the survey questionnaire and the use of administrative data. How would these be treated? Could confidentiality protections afforded respondents be protected? None of these major considerations are addressed by either the law or the proposed rule.
That is because the proposed rule needs clarification in a number of areas. Such terms as "data," "publication," and "public policy and rules" require further definition. Does data mean "all data" as noted in the law and what does that entail? Publication in this electronic information age gets murky quickly. Is it only in peer-reviewed journals, which we would favor, or can it be a presentation at a professional meeting, an Internet or pre-print posted paper as NIH is contemplating, or exchanges of preliminary results with other scholars? The rule is unclear.
How direct does the connection have to be to "public policy or rules?" Is something like the Health and Retirement Survey, (see below) which could help decision makers determine policies regarding Social Security and Medicare, but not necessarily immediately providing a direct linkage to a policy decision, be subject to the data sharing requirements of the new rule?
Some specific examples:
- The Health and Retirement Survey (HRS) funded by the National Institute on Aging provides important information about the work histories, retirement plans, and health of an ever-increasing aging American population. Through the method of the sample survey and a unique arrangement with federal agencies to share administrative data, the HRS can help decision makers determine future policies on Social Security, Medicare, and other policies affecting the elderly. There is concern that under the new law and the rules covering it, protections of confidentiality afforded survey respondents, especially if detailed geographic and occupational data on individuals were forced to be revealed, could not be adequately maintained. In addition, the ability to use administrative data collections, will fall by the wayside, if all data would be released by the agency in response to a FOIA request.
- A number of important areas of economic research would also be compromised by the new law and the proposed rules. A significant area of economic research today involves the collection of microdata from individuals and individual businesses. Premature release of data under FOIA would compromise again confidentiality protections, especially with regards to businesses, since FOIA exemptions only protect "personal privacy," and not institutions. One economist suggested the new rules could cripple work in Industrial Organization research, particularly studies of innovative firms, as cooperation would cease if there was a danger that these firms could be identified from the data in the studies. This would also be true of research into the financing of start-up businesses.
- Painstaking compilation of databases that are costly and time consuming will be discouraged if the researcher is forced to prematurely release data already collected, but perhaps not yet integrated into a larger data base and not yet published in a peer reviewed journal.
- As part of the Adolescent Health Survey funded by the National Institute of Child Health and Human Development interviews were conducted in schools. Under the new law and proposed rules, would the names of the schools included in the study be protected from disclosure? Again, because FOIA exemptions do not include institutions, a request could force NICHD to reveal the names of the schools, which could then lead to individual identification, since some principals and teachers were respondents.
- A study examining congressional recruitment practices surveyed political leaders in 200 congressional districts. The results of this survey could have public policy implications for campaign finance laws. Is it subject to a FOIA request under the new rules? Again, a FOIA request could lead to disclosure of the districts, which could make it easy to identify the names of the political leaders who participated in the survey. This would be another example where assurances of confidentiality, the backbone of survey research, would be destroyed.
There are other concerns. Although the new law and rules refer to authorizing "a reasonable user fee equaling the incremental cost of obtaining the data," it is unclear who gets the bill and who gets the money? Does the researcher send a bill to the agency for costs incurred in preparing the data for the agency? One suggestion is that the funds would simply revert to the Treasury. This seems a bit unfair to the researcher and the agency.
The harassment factor is something that also must be considered. People or groups opposed to a particular line of research may file numerous FOIA requests that keep the researcher so busy responding that he/she never gets to conduct the research. This, of course, is just what the harasser hoped to achieve, and could create a "chilling effect" on certain types of research
Furthermore, how are researchers with federal and non-federal funding going to be treated? Since the dollars are often fungible, how can you determine what part of the data came from the federal funding and what from the private sources? This is probably why Kathy Casey, Legislative Director for Senator Richard Shelby, told an audience at a meeting at the American Association for the Advancement of Science meeting, that the Senator also favored similar data sharing requirements for privately funded studies if they affected public policy.
Finally, are there time limits to how long data would need to be kept by the researcher? By the agency? ? Are the rules retroactive for data that was collected years ago, but only now are being mined for publication or to affect public policy?
We raise all of these questions, because we believe the law and the proposed rules are not the way to get data shared. There are too many unanswered questions creating opportunities for too much mischief that would endanger the scientific enterprise that has made this country a world leader in science, and whose economy, driven by science and technology, is the envy of the world.
Therefore, we urge a postponement of the implementation of these rules. Then, get the National Research Council working on a study and let's try to do this right!
Sincerely,
Howard J. Silver, Ph.D.
Executive Director
Encl.