The HONEST Act is this year's version of a piece of legislation formerly called the "Secret Science Reform Act." Its sponsor is Rep. Lamar Smith of Texas, the chair of the House Committee on Science, Space, and Technology — the same Congress member who, by the way, said that President Donald Trump "might be the only way to get the unvarnished truth."
The HONEST Act stipulates that the EPA can't make any assessment or analysis based on science that not openly accessible to the public. Specifically, the text states the EPA can't cite research that isn't:
publicly available online in a manner that is sufficient for independent analysis and substantial reproduction of research results, except that any personally identifiable information, trade secrets, or commercial or financial information obtained from a person and privileged or confidential, shall be redacted prior to public availability.
Sounds reasonable, right? If passed by the Senate, it would mean the EPA would have to make all the data it uses in its decision-making freely available online so that public and independent researchers could more easily scrutinize its decisions. For sensitive health data, the bill has provision that would give the Food and Drug Administration the power to redact.
On the face of it, the bill is in line with what a lot of researchers argue for: open access not just for journal papers but for data too. The big idea is that this will make science more transparent and replicable, and decrease the friction for one lab to evaluate the work of another. (Psychology and a number of other fields have been dealing with an ongoing "crisis"in which they're finding past research doesn't replicate. Open access is a way to rectify it.)
But that's the ideal world. In our real world, tons of high-caliber science is still locked behind journal paywalls. And so the "radical transparency" Smith is arguing for would perversely have the immediate effect of limiting the number of studies the EPA can cite in its decision-making.
There are two big reasons why.
1) The "HONEST" Act increases the costs of compiling research.
The EPA conducts much of its own research, and collects its own data. And under this law, it would have to make that data publicly available (the agency already does this with much ofits data). Under the bill, the EPA would have to get the underlying data it cites from any other scientific research institution.
"It is not clear how EPA would meet the bill's requirements," the Congressional Budget Office wrote about last year's version of the law, which is nearly identical, save for the new provision on redacting sensitive information.
And the EPA would have to pay for it from its own budget -— which seems likely to shrink. The CBO estimated the bill would cost the EPA around $250 million a year — due to the cost of ensuring that the nearly 50,000 scientific studies the agency cites annually all have publicly accessible data. But the current legislation only outlines $1 million for compliance.
To save costs, the CBO suggested, the EPA "could instead rely on significantly fewer studies each year in support of its mission." Plainly: To save on these costs, the EPA could just cite less scientific research in its actions. That's a clear roadblock to making decisions on the totality of evidence on an important question.
And it's especially not helpful considering that President Trump is seeking to cut the EPA budget by 31 percent.
The new provision to redact personal information from studies will likely add to the cost of the current bill (which the CBO has not yet scored). A study on asthma, for example, could have data on 10,000 people. Redacting that information "is a huge job that can occupy entire offices for thousands of hours," Ed Yong writes at the Atlantic.
2) The bill's definition of "replicable research" is very dubious.
The stated intent of the law is "to prohibit [EPA] from proposing, finalizing, or disseminating regulations or assessments based upon science that is not transparent or reproducible."
That last bit — "reproducible" — may also cause problems.
Scientists endlessly debate what makes for a good reproduction of a result. A huge point of contention: Does a study have to use the exact same methods, or can a "conceptual" reproduction that tests the same question with a different approach suffice?
Science has not found a clear answer, and to leave that term undefined in the text of the law is an opening for courts to weigh in on the question. If a court took the word "reproduce" literally, that too would limit the data the EPA can use.
"Cancer studies that took 30 years to compile evidence for, studies from natural disasters and chemical explosions — we can't reproduce those events," Lamdan says. "The way we regulate against our exposure to radioactive waste is we look at studies from Nagasaki and Hiroshima — clearly we're not going to reproduce those."
The EPA also collects data from one-time situations like the Deepwater Horizon oil spill. When the next oil spill happens, what information should the EPA rely on then? "We're concerned that in these situations the EPA could be constrained from using important or relevant research in making decisions," Lexi Shultz, director of public affairs at the American Geophysical Union, wrote in a letter to Congress.