(FORTUNE Magazine) – LET'S START with a short quiz. The Environmental Protection Agency estimates that radon, an odorless gas that seeps up naturally from the ground and gets trapped in homes, may be causing as many as 20,000 lung cancer deaths a year. The EPA also figures that hazardous waste dumps are responsible for, at most, 500 cancer deaths. The U.S. is spending less than $100 million a year to reduce one of these risks and some $6.1 billion annually on the other. Guess which problem gets the big budget. You're right! Hazardous dumps. Irrational? Sure, but it's hardly an isolated instance. Two decades after America launched a crusade to control pollution, we still haven't figured out what is really worth worrying about. Our environmental spending, says EPA chief William Reilly, ''cries out for discipline and rationality, for someone to put things in a reasonable hierarchy.'' Instead, reason and scientific evidence are more often drowned out by the voices of Hollywood celebrities, outraged columnists, environmental activists, fearful citizens, and self- promoting politicians, all eager to inflate the pollutant-of-the-momen t into a national threat. The costs keep climbing. As a result of last fall's Clean Air Act amendments, public and private spending to clean up the environment, adjusted for inflation, is projected to rise over the next decade from $100 billion to nearly $150 billion. That translates into almost 3% of GNP by 2000. Says Reilly: ''If we're going to spend that much, then in an era of increasing international competitiveness, we'd better get it right.'' Reilly has joined the ranks of those championing a different way to get America's environmental priorities straight -- risk assessment. By examining the effects of massive doses of a pollutant in the workplace or the lab, scientists extrapolate how much harm tiny levels of the same pollutant might cause among the populace at large. The answers are often surprising. Americans, as various polls attest, worry most about oil spills, acid rain, pesticides, nuclear power stations, and hazardous wastes. But Reducing Risks, a recent report by the EPA's Science Advisory Board, suggests that these are low- or medium-level dangers. The experts see greater hazards from radon, lead, indoor air pollution, and fumes from some chemicals like benzene and formaldehyde. Using risk assessment to set priorities could produce big savings. Consider radon vs. waste dumps. It costs only $25 to test a home for radon and an average $1,100 to clean up one found contaminated. The entire national housing stock could be made largely radonproof for just a few billion dollars. By contrast, the bill for cleaning up all America's toxic waste dumps -- a project that Congress has passed laws to ensure -- could eventually exceed $500 billion, even though the government's Agency for Toxic Substances and Disease Registry reports that only 11% of such sites pose a measurable threat to health. But in other cases rationality can lead a risk assessor to conclude that meeting unaddressed dangers requires sizable new investments. Reducing Risks puts global warming and depletion of the ozone layer -- two problems that now generate little government spending -- at the top of its worry list. SCIENTISTS can't put numbers on the threat from global warming, as they can, say, in relating radon exposure to cancer. But we should worry anyway, argues Reilly, because ''a four- to six-degree increase in temperature would be so vastly influential on agricultural productivity, economic systems, tree species, and the like.'' Ozone depletion ranks high because a measurable reduction in the ozone layer is thought to generate a measurable rise in skin cancer cases -- possibly 100,000 more a year by 2050. Reflecting its own agenda, the White House's Office of Management and Budget is pressing the EPA to downgrade its estimates -- and thus lower the government's potential costs. Battling over budgets partly explains why risk assessment is likely to have trouble winning wider political and public acceptance. The larger difficulty is that this is still an emerging science, one whose numbers can be as squishy as they are controversial. We know cigarette smoking causes lung cancer because the universe of smokers is so huge -- and one in four die from smoking. But epidemiology, the study of human diseases, has little relevance to assessing pollutants regulated by the EPA, since it's so hard to isolate significant numbers of people who have been made ill by a particular one. Only the workplace provides enough case studies. Uranium miners who got lung cancer from radon or shipyard workers who died from cancers caused by asbestos, however, were exposed to massive dosages over a long time. If the exposures to radon in the basement or asbestos in school ducts are a thousand times lower, how do you calculate the hazard? James Wilson, a chemist who estimates risks for Monsanto, argues that such threats are ''so small as to be undetectable. You can't validate the arithmetic.'' Environmentalists, not surprisingly, disagree. The dispute revolves around what to make of tiny doses when your only evidence comes from massive ones. Because human epidemiology can't provide enough evidence, scientists have turned to animals. Bioassays, as such tests are called, have the advantage of being quick and dirty. You can feed animals massive doses of carcinogens, and since their lives are so short, the results show soon. The trouble starts when scientists try to extrapolate results from mice and rats to humans. Like epidemiologists, biochemists have to deal with huge differences in dosages, but they also have to take into account differences in size and species. Rats get tumors in the zymbal gland in their middle ear. Humans don't have such a gland. The biggest problem is size. If a certain amount of benzene causes cancers in a rat or a mouse weighing a few ounces, how much is risky for a man weighing 160 pounds? Do you extrapolate by weight or by skin surface? Thereby hangs a fine but notable bureaucratic contradiction. The Food and Drug Administration opts for weight, while the EPA uses skin surface. Geometry tells us that as an object gets larger, its surface area increases more than its weight -- seven times more in the case of mouse to man. So by extrapolating from the skin surface, the EPA scares a lot more people. Its estimates of the chances of getting cancer from formaldehyde, benzene, or pesticides are roughly ten times higher than the FDA's. When such uncertainty and confusion get codified into inflexible regulations and legislation, the results can be pretty silly. The dumbest is the well- known Delaney clause of the Food, Drug, and Cosmetic Act, which requires the FDA to ban foodstuffs with any trace, however minuscule, of a substance that causes cancers in animals, no matter what the dosage. In the late 1970s this legislation forced the FDA to block sales of saccharin. After it was shown that tumors in the test rats' bladders were caused by sheer irritation from the doses they received -- the equivalent of a man drinking 150 cases of diet cola a day -- Congress passed an exception to its own law and saccharin went back on the shelves. But despite several attempts to repeal it, the Delaney amendment still stands. Bruce Ames, a biochemist who runs the National Institute of Environmental Health Sciences Center at Berkeley, worries that the techniques he had a big hand in developing are being pushed too far. Says Ames: ''We are scaring everybody about incredibly tiny levels of chemicals that are trivial compared to what's in your diet.'' He points out that the EPA sets the safe daily dose limit of dioxin at six femtograms, or one quadrillionth of a gram, per kilogram of body weight. That dose of dioxin is the carcinogenic equivalent of drinking one beer -- if you nurse it for 345 years. Instead of worrying about pesticide residues, Ames observes, we ought to be trying to reduce cancer risk by eating more fresh produce. After all, a single cup of coffee contains more carcinogens than you'd normally get from the residues in a year's intake of fruit and vegetables. Not all scientists agree that risks have been overstated. James Fouts, the senior scientist at the government's National Institute of Environmental Health Sciences, believes that as research methods are refined, pesticide residues will eventually show up as a factor in increased non-Hodgkins lymphoma, a form of cancer. The green lobby's main worry is that increased use of risk assessment might provide excuses for inaction. Says Joel Hirschhorn, a longtime government pollution analyst who now runs a consulting firm in Washington: ''Environmentalists see this as something big guys use to get us to accept things we don't want.'' He also raises methodological objections. Suppose an EPA study concludes that a given exposure to benzene will result in one cancer in a million 70-year lifetimes. Typically that one-in-a-million risk applies to the population in general. If the risk for a narrower population, such as allergy sufferers, is one in a hundred, Hirschhorn argues, shouldn't that determine the acceptable level of exposure to benzene?

For business, risk assessment offers a tidy way of allocating costs efficiently and cutting out endless legal battles over environmental dangers. Companies are also increasingly using it as a management tool. Union Carbide has developed an elaborate method to avoid what it calls another ''major acute incident,'' an analgesic way of referring to the accident at its Bhopal subsidiary in India that killed as many as 3,700 people in 1984. Carbide has set up an independent auditing department to make a detailed assessment of all its 550 plants and distribution terminals. These can be rated safe, made safer, or closed down. The hardest job may be selling this approach to John Q. Citizen. He knows the government has lied in the past about some acute dangers -- the exposure of servicemen to nuclear tests leaps immediately to mind. Still, most people are wondrously illogical about risk. If it's voluntary, they will take huge chances by smoking, drinking, and eating fatty foods -- each of which causes more death and disease than all present pollutants combined (see table). But tell John Q. that he is passively inhaling a shmillionth of a part of benzene in his air, and he erupts with outrage. Never mind that he stands only one chance in a hundred million of getting leukemia. Just get rid of the stuff, and don't try to incinerate it near him or truck it by his house to a dump. NOR WILL greater acceptance of risk assessment rid the U.S. of the environmental hazard sometimes labeled ''litigen'' -- the gas emitted from law offices when partners catch sight of a long-running environmental class action suit. Take dioxin, one of the most carefully studied substances in recent times. Though dioxin is still described as one of the world's most potent carcinogens, researchers have found an increased incidence of cancer only among workers who were heavily exposed to it for long periods. That hasn't stemmed a flood of legal actions with no basis in science. Still, among Washington policymakers risk assessment is gradually winning adherents beyond the EPA's Bill Reilly. When Congress approved the Clean Air Act last fall, it set up a new Risk Assessment and Management Commission to study the applications of the technique to government policy. The White House's 1992 budget, released in late January, features a chapter on ''Managing Risk Reduction Sensibly.''

For all its faults, risk assessment is clearly preferable to what we have now -- a hubbub of self-interested partisanship with no agreed goals and standards. As it matures and gets more accurate, this technique should prove increasingly useful as a way to set sensible spending priorities, identify new dangers, and rein in legislators and lawyers with something approaching a rational consensus. In any event, it sure beats setting the nation's environmental agenda every day in the press.

CHART: NOT AVAILABLE CREDIT: FORTUNE CHART/SOURCE: EPA CAPTION: THE COST OF CLEANING UP IS SOARING

CHART: NOT AVAILABLE CREDIT: FRANCES FIFIELD FOR FORTUNE/SOURCES: EPA, THE BUDGET FOR FISCAL YHEAR 1992, NATIONAL ASBESTOS COUNCIL, NATIONAL CANCER INSTITUTE CAPTION: ENVIRONMENTAL RISKS DRAW THE HEADLINES . . . . . . BUT EVERYDAY CHOICES KILL MORE PEOPLE Though risk assessment provides a useful way to compare potential threats, the cancer incidences shown above are often very rough estimates. The figure for aflatoxin, a natural carcinogen in a mold that grows on peanuts, assumes that every American eats four tablespoons of peanut butter a day for 70 years.