New York Times
August 31, 2005

OP-ED: INTERROGATIONS

Thanks to a little-known piece of legislation, scientists at the EPA
and other agencies find their work questioned not only by industry,
but by their own government

By Chris Mooney

The longstanding fault line in American life between politics and
science has become increasingly unstable of late, drawing headlines on
divisive issues ranging from stem cell research to evolution. But
there's a subterranean aspect of this conflict that rarely makes the
news: the fight over how science is used by government to protect us
from health and environmental risks--in short, to regulate.

Some time early next year, a federal appeals court in Virginia is
expected to decide a pivotal lawsuit concerning the uses of scientific
information in the regulatory arena. Brought by the US Chamber of
Commerce and the Salt Institute, an industry trade association, the
suit challenges a National Institutes of Health study showing that
reduced salt intake lowers blood pressure. But there's far more at
stake here than government dietary advice or salt industry profits. At
a time when science itself has increasingly become the battleground of
choice for determining what regulatory actions the government will
take, the case turns on whether such fights will ultimately find their
resolution in the courtroom, at the hands of non-expert judges.

In the process, the suit will define the scope of the five-year-old
Data Quality Act, a below-the-radar legislative device that defenders
of industry have increasingly relied upon to attack all range of
scientific studies whose results or implications they disagree with,
from government global warming reports to cancer research using animal
subjects. On its face, the act merely seeks to ensure the "quality,
objectivity, utility, and integrity" of government information. In
practice, as interpreted by the Bush administration, it creates an
unprecedented and cumbersome process that saddles agencies with a new
workload while empowering businesses to challenge not just government
regulations--something they could do anyway--but scientific
information that could potentially lead to regulation somewhere down
the road. The Data Quality Act, Chamber of Commerce vice president
William Kovacs explained in an interview, allows industry to influence
the regulatory process from "the very beginning."

Whether companies can sue agencies that reject their "data quality"
complaints, thereby dragging individual studies into the courtroom, is
the legal question at the core of the Salt Institute and Chamber of
Commerce lawsuit. If the judge in the case writes a precedent-setting
opinion, and if higher courts agree, a brand-new body of law could
emerge, consisting largely of corporate lawsuits against scientific
analyses.

But even if the lawsuit fails to establish judicial review of
government regulatory science, Republicans in Congress (who are
watching the case) are considering legislation that would ensure it
anyway. And lawsuits are just the beginning. The Bush administration
has also used the Data Quality Act's language as an excuse for
implementing a government-wide, industry-friendly "peer review"
system for scientific information--one that scientific societies,
including the American Public Health Association and the Association
of American Medical Colleges, regarded with serious misgivings when it
was first proposed.

In the three decades since the regulatory revolution that created the
Environmental Protection Agency and other fixtures of the federal
bureaucracy, industry has become increasingly adept both at weighing
down the rulemaking process with years of preliminaries and at
challenging regulations once they are announced. Now, each regulatory
decision seems to devolve into a "science" fight--not because we
don't have enough qualified scientists in federal agencies, or because
they're doing a poor job, but rather because those seeking to avoid
regulation constantly seek to raise the burden of proof required for
action.

The Data Quality Act provides an ideal means of achieving this goal--
all under the guise of an attempt to improve the quality of
information.

In truth, calling the "Data Quality Act" an "act" smacks of over-
glorification. In 2000, a flamboyant industry lobbyist named Jim
Tozzi, who had worked for tobacco and chemical companies, and other
interests, helped to draft two sentences of legalese that
Representative Jo Ann Emerson, Republican of Missouri, then tucked
into a massive appropriations bill signed into law late in the Clinton
administration.

But to understand where the Data Quality Act comes from, and how it
facilitates a veritable war on government science conducted by private
industry, you first need some historical context. Over the past
several decades, companies subject to government regulation have
increasingly sought procedural tools to allow them to challenge and
dispute the scientific data forming the basis for such regulation. The
overarching strategy--sometimes dubbed "paralysis by analysis"--has
been to make the fight turn on the validity of information, rather
than moral or policy questions, such as how much of a precautionary
approach our government should take in endeavoring to protect us from
risks.

One key proponent of challenges to government scientific information
was the tobacco industry. In 1998, in a dress rehearsal for the Data
Quality Act, lobbyist Tozzi worked on the so-called Shelby Amendment--
named after its official "author," Republican senator Richard Shelby
of Alabama--for Philip Morris. Also a brief insert to an
appropriations bill, and equally loathed by the scientific community,
the one-sentence amendment allows for the use of the Freedom of
Information Act to obtain "all data produced" by any publicly funded
scientific study.

Though subsequently limited in scope by a wary Clinton administration,
the Shelby Amendment helps companies conduct their own audits of
studies they don't like--as many firms, including tobacco, have often
sought to do. In the process, they can reanalyze the data and put a
different spin on it.

The Data Quality Act has been called the "daughter" of Shelby. Sure
enough, Philip Morris documents show that Tozzi circulated language
for a "data quality" regulation to the company and received comments
in response. While Tozzi agrees there was "definitely" a connection
between the tobacco industry and the Data Quality Act, in an interview
he insisted it was "not a cause and effect." But he added that if
tobacco companies had "continued challenging stuff, data quality
would have been a big help to them."

With the Data Quality Act, industry had finally found a means of
disputing agency science in the earliest stages of the regulatory
process. At the merest public mention that a government agency might
be looking at a particular study and wondering if it could compel
regulatory action, industry interests could file a complaint
challenging the agency's "dissemination" of the information.

"Industry tends to think there's some magic bullet somewhere that's
going to protect them from regulation," says Gary Bass, executive
director of OMB Watch, a government watchdog group that has monitored
the Data Quality Act closely, and a kind of counterpart to Tozzi on
the left. "If they can just get there earlier." Now, Bass continues,
the strategy is to look "in the science, in the data," for a way of
avoiding regulation.

In fact, both the Shelby Amendment and the Data Quality Act represent
successful under-the-radar attempts to pass bite-sized pieces of
legislation highly reminiscent of a massive, but failed, "regulatory
reform" bill pushed by Republicans during the Gingrich Congress in the
mid-1990s. "In the end," the Chamber of Commerce's Kovacs says,
"what we're going to get is far more than we could have ever gotten
by having a comprehensive regulatory law passed."

As soon as the Data Quality Act came into effect in October 2002,
corporate interests took it for a test drive. Tozzi's Center for
Regulatory Effectiveness teamed up with the Kansas Corn Grower's
Association and a group called the Triazine Network to challenge an
EPA risk assessment that had discussed evidence suggesting that
atrazine, a herbicide in widespread use in US cornfields and on other
crops, causes gonadal abnormalities in male frogs (possibly
contributing to declining amphibian populations worldwide). Sure
enough, citing the lack of an accepted test for measuring endocrine
effects, the EPA ultimately concluded that hormone disruption did not
constitute a good reason to restrict atrazine's use. Tozzi's petition
very likely had an effect on the decision. "You put the sequence
together and I think it's common sense to say it had an influence,"
notes Sean Moulton of OMB Watch.

And the atrazine challenge is just one example. The Data Quality Act
has been employed in a wide variety of lesser known industry-friendly
campaigns: by a chemical company challenging EPA information on
barium, its product; by a law firm that has reportedly represented
companies involved in asbestos lawsuits, objecting to an EPA document
on asbestos risks from work on vehicle brakes; and by paint
manufacturers questioning the basis for regulations on air emissions
from their products, to name just a few examples.

All three of these challenges targeted the Environmental Protection
Agency. But other agencies have also seen an influx of complaints.
Logging interests, for instance, challenged several US Forest Service
studies and documents relating to habitat protections for the Northern
Goshawk, submitting a massive 281-page petition outlining their
objections.

Some liberals and environmental groups have also attempted to employ
the Data Quality Act. The libertarian-leaning Tozzi--a jazz aficionado
with regulatory "nerd" expertise --has himself teamed up with
Americans for Free Access, which used the act to dispute government
claims that marijuana has no medical uses. "My musician friends will
finally appreciate my work," cracked Tozzi in an e-mail alerting me to
the pot petition.

But mid-2004 analyses by The Washington Post and OMB Watch showed that
the act, in effect for almost two years, had been overwhelmingly used
by industry groups. In any all-out war over regulatory science,
corporations will ultimately have the financial edge. In effect, then,
the Data Quality Act tilts the regulatory playing field still further
in industry's favor.

But as successful as Tozzi's and his clients' efforts have been, they
might not have gone very far without the help of John Graham. The
controversial head of the OMB Office of Information and Regulatory
Affairs (and former head of the Harvard Center for Risk Analysis),
Graham is responsible for the review of government regulations in the
Bush White House.

Not long after coming into office, Graham seized upon the Data Quality
Act and instructed federal agencies to draw up their own guidelines
for implementing it by October 2002, when the law would go into
effect. Then, in September 2003, claiming the Data Quality Act gave
his office a newfound role in improving the quality of government
science, Graham proposed using the act's thin language to justify an
unprecedented government-wide "peer review" system for agency
science.

As legal scholar Wendy Wagner of the University of Texas argued in a
recent article entitled "The 'Bad Science' Fiction," there's little
real evidence to support the notion that government agencies churn out
"junk science"--a frequent industry accusation--or that their
existing peer review protocols are inadequate. So it's no surprise
that scientific heavyweights like the American Public Health
Association and American Association for the Advancement of Science
announced their concerns over Graham's initial proposal, which would
have required review of all "significant regulatory information," and
an additionally laborious review process for data with "a possible
impact of more than $100 million in any year."

Most significantly, perhaps, the proposal would have blocked academic
scientists from serving as reviewers if they had obtained or were
seeking "substantial" research funding from the government agency in
question--a condition likely to exclude leading academic experts in a
field--yet showed little concern about the participation of industry
scientists. Graham's office subsequently softened the proposal, and
removed this most objectionable of requirements.

But the administration has still failed to adequately explain why such
a "peer review" system was needed in the first place. After all, no
government-wide standard for peer review existed in the past--and that
may have been a good thing. Different agencies have different needs,
just as different scientific disciplines employ different
methodologies and standards of proof.

Furthermore, concern about onerous and unnecessary intrusions into the
regulatory process remain warranted. The process for vetting "highly
influential scientific assessments" under the new peer review plan
remains quite burdensome, requiring the preparation of a peer review
report that must be made public and a written response from the
agency.

Such procedures will only further ossify an already sluggish
regulatory process. And as "peer review" critic Sidney A. Shapiro, of
the Wake Forest University School of Law, has observed, these
procedures are required even for "routine information" that is not
"complex, controversial, or novel."

Such objections notwithstanding, in December of last year Graham's
office finalized the peer review plan. Its provisions for "highly
influential scientific assessments" took effect on June 16. The media
hardly noticed. Below the radar, as always, expansion of the Data
Quality Act continues apace.

Chris Mooney is Washington correspondent for Seed magazine and author
of the forthcoming book "The Republican War on Science" (Basic), from
which this article is adapted.

Copyright 2005 The New York Times Company