Exposing the Vulnerability of Almighty Data   

Tainted Truth: The Manipulation of Fact in America
By Cynthia Crossen
Simon & Schuster

Tainted Truth: The Manipulation of Fact in America
By Cynthia Crossen
Simon & Schuster
272 pages; $23

This book falls just short of being a classic, raising (but unfortunately not resolving) dead-on questions about modern society's groveling at the heel of Almighty Data.

Cynthia Crossen, a veteran Wall Street Journal reporter and editor, documents with wicked precision how we glorify and objectify numbers, as if, either by fearing or worshiping them, we could somehow exempt numbers from the fallibilities of any other human analysis.

In short, she argues, we can't. Figures lie, and liars figure, and we've known that for centuries. Still we continually fall for the aggressively asserted studies and surveys that besiege us, to take just one example, via the 12 billion display ads and 300,000 TV commercials presented every day.

But Crossen doesn't limit herself to easy targets like ads in considering "the increasingly disturbing trend of inaccurate and corrupt information." She also scrutinizes political polls, food and nutrition research, public policy studies, and medical and legal research, repeatedly finding problems in areas where consumers have little personal expertise and much at stake.

"People know enough to be suspicious of some numbers in some contexts, but we are at the mercy of others," Crossen writes. "..[M]ore and more of the information we use to buy, elect, advise, acquit and heal has been created not to expand our knowledge but to sell a product or advance a cause."

Dissecting one study after another, she exposes the vulnerability of data, from flagrant chicanery and sleight of hand to the unsteady assumptions that underlie nearly all research.

She traces what she calls "the three conflicts of interest inherent in research — the sponsor's, the researcher's and the media's."

Researchers' problems are the most familiar, from poor design to sloppy analysis. Her book is filled with interesting examples of loaded survey questions ("Would you support universal health care if it would mean the loss of thousands of jobs, particularly in Connecticut?"); manipulation of samples (one soda company's taste test used only people who lived within 100 miles of its bottling plants); and the unreliability of human subjects anyway (researchers once found that the health of two-thirds of patients receiving placebos improved).

Crossen makes a more timely contribution in documenting a trend toward sponsored research that, inevitably it seems, produces results favorable to the sponsor. A study that walnuts help unclog arteries was sponsored by the California Walnut Commission. A conclusion that bald men are three times likelier to have heart attacks was sponsored by the Upjohn Co., which makes a hair growth agent.

Once, the public could rely on university scientists for neutral assessments of such work. But more and more scholars do sponsored research and depend on interested parties for their financing. "It's not the client dictating the results, it's the researcher self-censoring, unconsciously finagling the results to please the client, to get a grant renewed," one observer explains.

Meanwhile the media, she contends, do little to probe the methods and connections of researchers or to help consumers separate the credible from the flaky.

Crossen focuses in detail on several issues (Are disposable diapers more or less harmful to the environment than cloth? Will the family leave bill cost money or save money?) in which the truth, wherever it lies, is clouded by contradictory evidence and self-interested manipulation.

The result? Public confusion, then cynicism, leading to loss of faith in all information.

The key question then becomes, "If all information is not equally believable, how is the good distinguishable from the bad?"

Here, Crossen's otherwise excellent book founders. Perhaps the question can't be answered, but Crossen's readers could hardly be blamed for concluding that the research process is so polluted that nothing can be believed.

She does offer a set of "solutions," but they amount to asking everyone involved to do the right thing. She wants reporters to examine research more closely, universities to weed out conflict of interest, and pollsters and researchers to police themselves. This seems akin to asking folks to send in their taxes without an IRS to nudge them. Surely there are many honest researchers out there, but the problem she describes goes beyond altruistic self-policing.

Since a new government Ministry of Research Purity probably isn't the answer, where should we start? I wish Crossen had written more forcefully about the larger matter of weighing claims and evaluating knowledge. At the very least, she could have said that we need to recognize that numbers mean nothing by themselves. They are only as good, or as fallible, as the people who process them.