Tozzi found even more reason to rejoice as Bush made a pivotal appointment to head the OIRA, Tozzi's old domain within OMB that would now handle data quality: John Graham, a risk-assessment specialist with a history of close ties to regulated industries.

"John Graham came in, and he did an unbelievable job," Tozzi said. "Better than I could have done had I been there myself."

Politicizing the Process

Graham had been the head of Harvard's Center for Risk Analysis, an institution funded primarily by contributions from more than 100 industry and trade association donors. While there, he had amassed a reputation as a skilled critic of the cost of regulation.

In one analysis, conducted with funding from the auto industry, he concluded that it would be a mistake to require side air bags in cars because they would cost $400,000 for every year of life saved. Independent experts reviewing his work found that the figure was actually about $60,000, and Graham had to rewrite his article -- and change his conclusion -- before it could be published in a prestigious medical journal.

When Bush nominated Graham to head the OIRA, many citizen and environmental groups vehemently objected and more than one-third of the Senate voted no. In his first few months, Graham sent many near-final regulations back to the agencies that had proposed them, often saying he was not convinced they were worth the cost.

Then he turned to the job of implementing the Data Quality Act.

By the fall of 2001, Graham's office had published detailed guidelines for implementing the act. A year later, federal agencies started accepting petitions requesting that they withdraw information that allegedly did not meet

Individual agencies are responsible for reviewing the challenged data and deciding whether they are indeed reliable. But the OMB, a part of the White House, oversees the process closely -- through involvement in the agencies' deliberations and by demanding annual reports describing how agencies dealt with each petition.

OMB staff members have been providing "extensive assistance to agencies in preparing responses to correction requests," Graham acknowledged. "OMB oversight is critical to make sure that agencies handle these requests in a diligent and consistent manner," he said.

Graham said the OMB's unprecedented foray into science is justified in part because the data in question often serve as a foundation for costly regulation, which the OMB oversees. To fulfill the new role, Graham hired the OMB's first nine career scientists, including six with PhDs.

The Data Quality Act, or at least something like it, "was absolutely needed," said Horner of the Competitive Enterprise Institute.

Yet Steinzor, the Maryland environmental lawyer, and other critics complain that the OMB's involvement politicizes the process. The expertise of the handful of scientists hired by Graham, they say, cannot match that of the thousands of experts on agency staffs.

And while Graham said the OMB still supports weight-of-the-evidence analyses, Steinzor and others contend that the Data Quality Act inherently focuses on individual snippets of data -- each of which is inevitably open to criticism -- instead of on overarching bodies of evidence.

"You can get lost in the minutiae, and that's exactly where they want you to go," Steinzor said. "They just pick, pick, pick, until you're so addled you can't protect people or the environment."


A Tool for Decreasing Regulation

A few environmental and public interest groups have tried to use the Data Quality Act. Public Employees for Environmental Responsibility, a Washington-based group that helps federal scientists who believe their data are being suppressed, has filed three petitions under the act.

One challenged the credibility of a Defense Department document supporting a proposed Army Corps of Engineers project; one contended that the Fish and Wildlife

"I'm not sure it is the sharpest tool in the environmental toolbox, but at least it is a tool," said executive director Jeff Ruch, adding that the swan petition lost and the other two are still under review.

Many citizen groups and environmental activists believe the Data Quality Act will always be more useful to those seeking to decrease government regulation. Newly proposed regulations must be justified with evidence, they note, and the act is designed specifically to challenge such evidence.

"What it really can do best is slow the regulatory process," said Sean Moulton, a senior policy analyst with OMB Watch, a government watchdog group. "And even a simple delay of a rule can mean a huge financial windfall for an industry."

In the first 20 months, a handful of petitions -- all from industry -- have been at least partly successful. In one, the Competitive Enterprise Institute had wording added to a multi-agency federal climate change report stating that the report's findings did not meet Data Quality Act standards.

In another, a law firm with corporate clients in asbestos litigation got the EPA to agree to make changes in its booklet that offers warnings and safety advice to brake mechanics.

Yet another, filed by a group that receives funding from the conservative Scaife Foundation, succeeded in getting the National Institutes of Health to downgrade warnings about the effects of smokeless tobacco. And then there was the atrazine challenge.


'Manufacturing Uncertainty'

That petition, filed by Tozzi, made a two-pronged attack on the effort to regulate atrazine more stringently. The first was to claim that the evidence for atrazine's gender-bending effects in frogs was not fully reproduced by other Syngenta-funded EcoRisk scientists. The second was to claim that the EPA did not have the proper test to prove atrazine had ill effects.

Tozzi said reliance on irreproducible results would violate the Data Quality Act because information that is not reproducible is "not accurate, reliable or useful."

As evidence of irreproducibility, he pointed to the dozen or so studies sponsored by Syngenta in addition to Hayes's study. An independent panel of experts convened by the EPA had already expressed exasperation over the conflicting results and mistakes they found in the design and implementation of those studies.

In at least two of the studies the "control" frogs that were supposed to be atrazine-free were later found to have been in water contaminated with atrazine, an error the scientists said was unintentional. Another set of Syngenta studies was found to be unreliable because 80 to 90 percent of the animals died, apparently as a result of inadequate care.

Essentially what Syngenta-funded scientists did "was produce a number of studies that were purposefully flawed and misleading, and that changed the weight of the evidence," Hayes said.

While the EPA review also found some flaws in Hayes's studies, his conclusions have been echoed by at least four other independent research teams in three countries.

"What a coincidence that everybody can find an effect of atrazine on gonads," Hayes said, "except [those] funded by Syngenta."

David Michaels, a professor of occupational and environmental health at George Washington University School of Public Health and Health Services, said even a good study will appear "not reproducible" if enough bad studies are thrown into the mix.

"I call this 'manufacturing uncertainty,' and there is a whole industry to do this," said Michaels, who was the Energy Department's assistant secretary for environment, safety and health under Clinton. "They reanalyze the data to make [previously firm] conclusions disappear -- poof. Then they say one study says yes and the other says no, so we're nowhere."

Pastoor of Syngenta said there was no conspiracy to create conflicting data.

"I don't think it's extending things too far to say atrazine may be one of the best studied chemicals on the face of the earth," he said. "Unfortunately -- or fortunately, depending on how you look at it," other EcoRisk team members "could not replicate what Tyrone had done."

But Hayes was not the only team member who at least privately agreed that atrazine was having some effect on frogs. Team member James Carr of Texas Tech told Hayes in an e-mail in February 2003: "I agree with you that the important issue is for everyone involved to come to grips with (and stop minimizing) the fact that independent laboratories have demonstrated an effect of atrazine on gonadal differentiation in frogs. There is no denying this."

The second prong of Tozzi's attack was that the EPA had not designated tests that would serve as the gold standard of proof of hormone disruption in frogs.

The EPA does have certain "guideline tests" that can automatically trigger regulation, including some that measure certain health effects of chemicals on wildlife. But not for hormone disruption.

Jennifer Sass, a scientist with the Natural Resources Defense Council, said Tozzi's position flies in the face of decades of regulatory science. She said the evidence on atrazine's effects was more than convincing by traditional standards. The act, she said, has "hamstrung EPA's ability to express anything that it couldn't back up with a mountain of data. It basically blocked EPA scientists from expressing an expert opinion."

Hayes said he supports efforts at the EPA to create a gold standard test. However, he said, "when we discover a pattern like this, we know we have a problem. Yes, we should work to validate it perfectly. But in the meantime, let's not keep using 80 million pounds of atrazine per year while we figure it out."


Avoiding Tighter Restrictions

The EPA ultimately agreed with Tozzi that the lack of such a test prevented it from regulating atrazine as a hormone disruptor -- a concession many environmentalists found surprising.

No one claims that Syngenta's Data Quality Act petition was single-handedly responsible for giving atrazine's renewed approval the green light. But coming at the end of an arduous 10-year review, the data quality challenge was "the final one-two-three punch," said Sass of the NRDC, which has sued the EPA repeatedly on atrazine.

In closed meetings -- details of which the EPA has declined to release -- company representatives and EPA officials worked out a plan to avoid tighter restrictions. Instead, the plan calls for Syngenta to track atrazine levels in 40 U.S. watersheds over the next three years to see how farmers are doing in their efforts to minimize contamination. If concentrations rise above a level that the company agrees is "of concern," then the company will work with the farmers to try to reduce the levels.

The company will also fund more studies on frogs and reanalyze its data on employee cancers.

The resolution, Sass said, was "basically negotiated instead of going with a scientific rationale."

Asked why other stakeholders, such as environmental groups or outside scientists, were not allowed to be part of the negotiations as they were in earlier stages of atrazine's review, James Jones, director of the EPA's office of pesticide programs, said opening the meetings "would be incredibly complicated and would create a disincentive for the company to come to the table."


Exempting Atrazine

In June, Tozzi filed his latest Data Quality Act petition.

This time it was directed at the National Toxicology Program. That is a part of the National Institutes of Health that reviews chemicals to see if they cause cancer.

The program had announced in the Federal Register that atrazine was among a long list of chemicals that it was considering for examination. In his petition, Tozzi seized on a few sentences from the program's description of its chemical review procedures. He claimed that those sentences contained discrepancies that violated the Data Quality Act.

Therefore, he wrote, the program should be barred from reviewing the cancer-causing potential of any chemicals. In particular, the petition noted, atrazine.