<< News Scan | Main | SCOTUS Criminal Law Decisions >>

Transparency in Science

"Robert Hahn is a visiting professor at Oxford University's Smith School of Enterprise and the Environment and a non-resident senior fellow at the Brookings Institution. He recently served as a commissioner on the U.S. Commission on Evidence-Based Policymaking."

Dr. Hahn has an op-ed in the WaPo, titled "Many mocked this Scott Pruitt proposal. They should have read it first."  He notes a vehement reaction to the proposal but concludes it is more "the result of rhetoric surrounding the rule" than the rule itself.

Here's what the rule would actually do. First, it would require the EPA to identify studies that are used in making regulatory decisions. Second, it would encourage studies to be made publicly available "to the extent practicable." Third, it would define "publicly available" by listing examples of information that could be used for validation, such as underlying data, models, computer code and protocols. Fourth, the proposal recognizes not all data can be openly accessible in the public domain and that restricted access to some data may be necessary. Fifth, it would direct the EPA to work with third parties, including universities and private firms, to make information available to the extent reasonable. Sixth, it would encourage the use of efforts to de-identify data sets to create public-use data files that would simultaneously help protect privacy and promote transparency. Seventh, the proposal outlines an exemption process when compliance is "impracticable." Finally, it would direct the EPA to clearly state and document assumptions made in regulatory analyses.
Those are sensible proposals in all fields, not just environmental science. Transparency, including data-sharing, should be mandatory for all government funded research, including crime research.
For many kinds of studies, there is more than one way to slice the numbers.  A researcher-advocate with an agenda can choose the method that supports his preferred result.  If the study goes to court with opposing experts and an adjudication, it may be found that it really does not show that.  For example, the notorious Baldus Study on racial discrimination in capital punishment was found by the federal district court not to show what Baldus claimed but, if anything, just the opposite.  See McCleskey v. Zant, 580 F. Supp. 338, 368 (N.D. Ga. 1984).

Most studies don't get that kind of scrutiny, though.  Far too often, the researcher's interpretation is deemed conclusive as to what his study "shows."  Making the data available for others to slice in other ways can provide needed additional perspectives.

If the government paid for the data, then it should be available to all.  No one should have a monopoly on data that properly belongs to the general public.

Leave a comment

Monthly Archives