The authors of It Ain't Necessarily So are here to enlighten us. David Murray, Joel Schwartz, and S. Robert Lichter promise to offer "insight into the choices and judgments that journalists make" to create the news, and to provide new ways of empowering savvy news consumers to "defend themselves against misleading coverage" of science. And they offer a word of caution to consumers of the mainstream news: Caveat lector. Let the reader beware.

It would be good advice for readers of this book.

If the authors had delivered on their promises, the book would be an excellent addition to the literature of media criticism. They all are trained as social scientists. Murray is an anthropologist and media critic at the Washington, D.C.-based Statistical Assessment Service (STATS). Lichter is a political scientist at the Washington, D.C.-based media watchdog group, the Center for Media and Public Affairs (CMPA). And Schwartz is a political scientist at the Hudson Institute, an Indianapolis-based think tank with an office in Washington, D.C. They promise an "objective" look into the culture of journalism, functioning "as anthropologists do in approaching a foreign culture." No one can accuse them of aiming low. "We can for the first time answer the question of what can and cannot be proclaimed with confidence," they announce early in the book.

The authors offer some useful, if not earth-shattering, insights into the way research becomes news. Based on information from the U.S. Centers for Disease Control and Prevention, newspapers, for instance, reported an increase from 1994 to 1995 in the percentage of women with AIDS. The reports, however, didn't mention that the absolute numbers of cases among women actually had dropped. The increase in the percentage among women was due, in fact, to a drop in the number of AIDS cases in men. Therefore, good coverage should include both percentages and raw numbers, particularly if they lead to different conclusions.

Coverage of research, they remind us, is influenced by the choices journalists make to cover some studies and ignore others. Research goes through several filters, consisting of researchers, public relations specialists, and journalists, before it reaches our homes as news. Journalists sometimes rely too much on the way research news is spun in press releases when they should be analyzing the science themselves.

The trio also offers an interesting discussion of social science research techniques and some practical advice on what to look for in good coverage. Not all of their criticisms are off target: Some journalists tend to portray clarity when the truth is ambiguous; some cast health and environmental stories as morality plays, and some overplay preliminary findings.

Many of the authors' gripes, however, reflect their own biases and naïveté about the proper role of the press, rather than any inherent failings of reporters and editors. For example, it's ridiculous to blast the New York Times for covering a government whistle-blower's theory, later discredited, that plutonium at the Yucca Mountain nuclear waste dump might trigger a thermonuclear blast. The Times was doing its job, sounding an alarm on a possible danger that may have been suppressed, then later covering the news when it didn't pan out. And despite the authors' fault-finding, it's silly to expect reporters to wait months or years for peer-reviewed studies when important but early results are announced. That's like asking a statehouse reporter to wait until the law is passed before reporting it. The science writer's job is to shine light into the world of science, with all its warts and brilliance.

The authors use patronizing and inaccurate arguments stating that most journalists don't understand the science they cover, routinely "mistaking correlation for causation and coincidence for conviction." They claim that scientists, but not journalists, regard reality as "contingent" on the next finding, despite the fact that journalists routinely cover developing stories that change daily. And revealingly, the stories they choose to criticize almost invariably concern social statistics that conservatives don't like, or else they reveal health or environmental hazards that might encourage profit-reducing regulations. For example, they strain credulity by arguing that reporters should have ignored or downplayed a 1996 Nature paper that offered the first evidence that a northward shift in a butterfly's species range was due to global warming.

Throughout the book they make a mistake that even cub reporters know to avoid. Although they have no experience as working journalists, they claim to know reporters' motivations, even though they conduct their entire analysis from press clippings and research articles; they fail to cite a single interview with a working journalist. (So much for their anthropological look at the culture of journalism.) Margaret Mead spent years in Samoa, but the authors couldn't phone across town to the Washington Post. Worst of all, they argue with as much innuendo as evidence. For example, they write of the butterfly coverage, "One can suspect that research conforming to the global warming scenario is greeted more favorably in many newsrooms than research contradicting it." One can suspect instead that the authors have their own agenda.

The suspicion grows as one starts looking closely at their arguments, which are tendentious and weak, and at their supporting data, which are anecdotal and cherry-picked. A critical look at the authors' recommendations supports this case. They argue seriously that reporters should not reveal who funds a study, even if it's the industry that has the most at stake from the results. Instead, a study should be judged entirely on its merits. For example, they argue, coverage of global-warming skeptic Robert Balling of Arizona State University should ignore the fact that he's funded by an oil- and coal-industry trade group called the Western Fuels Association. Such advice clashes with that of most experts in research and journalism ethics, who argue for full disclosure.