Improving Risk Communication (1989)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

6 Problems of Risk Communication In this chapter we address what we consider to be the principal problems of risk communication. First we describe problems deriv- ing from the structure of the political and administrative system. These are problems for which little can be done by those involved in risk communication beyond understanding them. They must be confronted and accommodated, since they cannot be done away with. Next we describe problems of risk communicators and recipients. These problems, in contrast, are much more amenable to improve- ment or solution. The problems of these two groups are presented together because many things are problems for risk communicators because they are problems for the recipients of risk messages. For example, the risk communicator needs to pay attention to the under- standability of risk messages because most recipients have difficulty comprehending the technical terms typically found in risk assess- ments and other technical analyses. PROBLEMS DERIVING PROM THE INSTITUTIONAL AND POLITICAL SYSTEM . As we have seen, scientific and technical information is of central importance to decisions about how to respond to risks and thus is an important element in risk messages. But risk management decisions also take place as part of a democratic process, and risk analysis 108

PR OBLEMS OF RISK COMMUNI CA TI ON i 109 s only one of several sources of relevant information. Furthermore, politics can, and often does, assert control over decisions otherwise delegated to experts. The intrusion of politics can result in consid- erable frustration to risk managers as well as to others involved in the process. It is thus important to consider problems posed by the institutional and political system for risk communication. For the most part the problems of the institutional and political system are part of the context within which risk managers and risk communicators operate. Even though these problems are largely beyond the ability of the principals in risk communication to affect, they nevertheless can have considerable impact on actions and events. Legal Considerations Risk communicators may be constrained because legal consider- ations influence the options available to risk managers and therefore the content of risk messages. Several kinds of legal provisions may provide such constraints, including (1) statutory mandates, (2) lia- bility, and (3) informed consent and "right-to-know" requirements. . Statutory Prescriptions and Proscriptions Statutory language may, in effect, force the risk manager to take certain kinds of actions, some of which have important consequences for the content of risk messages or their dissemination. This is per- haps most obvious with respect to units of the Public Health Service. The major goal of the Centers for Disease Control (CDC), for exam- ple, is to lead public health efforts to prevent unnecessary disease, disability, and death. The CDC pursues this goal through programs aimed at prevention and control of infectious and chronic disease and of disease, disability, and health associated with environmental and workplace hazards (Department of Health and Human Services, 1986~. These programs include not only regular publications such as the Morbidity and Mortality Weekly Report (MMWR) but also emergency advisories. CDC officials are often quoted in the news, and their statements can have considerable impact. CDC's concern about long-term contact with soil in the Times Beach, Missouri, area was an important factor in the government's decision to purchase homes and permanently relocate the residents. CDC's mandate to lead public health efforts thus goes a Tong way toward establishing the tone and approach of risk messages emanating from the CDC.

110 IMPROVING RISK COMMUNICATION Such prescriptions and constraints and their impact on action may in their turn have a strong impact on whether the public views the agency as an advocate and on the credibility of the organization with respect to public health issues. Liability One principal reason that legal constraints constitute a prob- lem for risk communication is that these considerations may make difficult or impossible the crafting and presenting of messages that effectively address the issues that may be most relevant to the in- tended recipients of the message. For example, following the 1985 release of aldicarb oxime from its plant in Institute, West Virginia, Union Carbide had to decide about what information to make public about the accident (Coppock, 1987~. Communications and com- munity relations experts usually advise making available everything that is known about an accident as quickly as possible, in terms that laypeople can readily understand. I.egal advice is almost always ex- actly the opposite: give out as little information as possible so as to avoid providing ammunition for use in court. Given the prevalence of large court awards in product liability and toxic exposure cases, con- cern with liability is in the minds of many business people. The final message probably involves a compromise between these perspectives. Informed Consent and Right-to-Know Issues of informed consent have changed the way the health pro- fession interacts with patients. Attempts are made, for example, to hold physicians to fairly stringent standards in obtaining con- sent prior to initiating experimental therapy. But "right-to-know" issues are having equally important impact in many other areas. Em- ployees are to be informed about the hazards of the materials they handle under Occupational Safety and Health Administration rules, communities are to be informed about inventories and emissions of hazardous substances under the community right-to-know provisions in Title Ill of the Superfund Amendments and Reauthorization Act of 1986, and CaTifornia's Proposition 65 provides for provision of in- formation about any product containing carcinogens or teratogens. The overall eject of such developments is that there are many more legal requirements that result in the preparation and dissemination of risk messages than in the past.

PROBLEMS OF RISK COMMUNICATION Sharing of Power 111 Communicating with citizens about risk issues can increase their desire to participate in or otherwise influence decisions about the control of those risks. These demands can change the dynamics of the situation for the risk manager, the risk communicator, and the citizen. The motivation for citizen involvement becomes even stronger when a decision process appears to result in an outcome with which the individual disagrees. The interests of citizens and their motivation to participate can be especially problematic when the implementation of risk control measures is necessarily decentralized and local preferences preclude solutions in the broader interest. The sharing of power is a central facet of representative democ- racy. Citizens transfer decision-making power to elected officials. However, citizens have the responsibility to hold both the legislative and the executive branches of government account To ~'r~rri,;:~ _ O ~ I_ ~^v~_ ~ ~ ~^ ~-=0 ~ 1 · ~ · ~ - . ~ · . ,, ~ ants responslulllty and Judge the delegation of authority, citizens need information. Demonstration of thin a~,nt.nhililv i; rune of tl~" ~ ~^ ~ A^^ V) ^~ ~ ~^ ~~ · ~ ~ ~ Important functions ot risk communication. Holding government ~.~mint.~hl~ rn~z~n~ in ~ Ah; Q^-C!^ by_ V Cat -~ ''^-~ ~ 44 ~ ~ ~ ~A-O~ ~ ~11 . ~ . ~ . . _ _ surlng that government policies and actions correspond to public preferences. The difficulty, of course, is aggregating across the pref- erences of the many people involved. Most people believe they have a right not to be subjected by others to unreasonable risks. Some people believe not only that risk to life and health should be mini- mized but also that three kinds of unfairness should disqualify, say, siting a hazardous waste facility: (1) imposing costs on those who have not voluntarily agreed to bear them. (2) imposing costs on _ those who oppose availability of and avoid use of the products and services generating the hazard, and (3) imposing disproportionately large burdens on those who benefit least (Simmons, 1984~. When people believe that any of these three hold, they may fee! imposition of a hazardous waste site to be unfair regardless of the no. used to derive that particular site. ouch convicts are at the core of many instances of "locally unwanted land uses" (LULUs). Local preferences often run counter to solutions that would otherwise seem to be in the broader public interest. Here risk is only one part of the problem, and thus risk communication per se cannot be expected to resolve all the issues. v If _ _ · . . . ~ . . . Communication research suggests that risk messages will be more easily understood when the risk communicator not only incorporates language familiar to the recipients but also genuinely respects and

112 IMPROVING RISK COMMUNICATION incorporates their views (Covello et al., 1987b). This perspective suggests that effective risk communication should begin before im- portant decisions have been made. If all the important choices have already been determined, it will be difficult to reflect the views of the recipients. Risk messages will become little more than attempts to "sell" a predetermined conclusion, which may create considerable alienation among the intended recipients. But as suggested in the above discussion, people may reject the attempt no matter when it occurs if they are unwilling to compromise their position. The risk manager may thus face a difficult task in seeking advice from people but excluding them from the resulting decision. People naturally want to see their views affect the outcome, and they may have difficulty differentiating the risk communication process from the risk management process. The American political culture puts a premium on procedures that offer a wide variety of interest groups and citizens the opportunity to participate in decision making (MeInick, 1988~. Fragmentation Risk control decisions can be made or influenced by several dif- ferent political actors. At the federal level, Congress, the executive branch, and the courts all shape health and environmental regula- tions. Despite the dominant role of the federal government, state and local governments also remain important. This fragmentation may make communicating about risks more difficult because of dispersion of responsibility, incentives for each actor to gain as much leverage as possible from the limited portion he or she controls, and difficulty in determining who is responsible for the eventual outcomes. Dispersion of Responsibility Fragmentation of risk control decisions derives from a central feature of the structure of American political institutions: dispersion of power. A basic tenet of the American political system is separa- tion of powers, but power is also dispersed to a remarkable degree. For example, one source claims that Edmund Muskie, though only chairman of a Senate subcommittee, had at least as much influence on environmental policy from 1969 to 1979 as Richard Nixon, Jimmy Carter, William Ruckelshaus, or Douglas Costle (MeInick, 1988~. And several environmental groups, most notably the Natural Re- sources Defense Council and the Environmental Defense fund, have

PROBLEMS OF RISK COMMUNICATION 113 used their success in litigation to become major players in national policymaking (MeInick, 1988~. In the United States each level and branch of government pro- vides access to a variety of groups. Thus corporations, trade asso- ciations, labor unions, professional associations, intergovernmental lobbies, and environmental groups all influence regulation and its implementation. To the extent that individuals and organizations participate, they can also be held responsible for the overall out come. Dispersion of responsibility can lead various executive agencies to take different positions with respect to the same issue. The U.S. Department of Energy, for example, views the hazards associated with radioactive contamination of groundwater differently than does the EPA. Officials of the two agencies say quite different things about the risks involved in specific instances of contamination in Idaho. Differing positions can also be found within different parts of the same government organization. This derives in part from the organization of large bureaucracies into separate divisions but also in part from the belief that separation of power yields greater benefit than cost. When fragmentation leads various parts of government to differ- ent positions or approaches with respect to the same risk, it can lead to problems for risk communicators. Incentives to Gain Leverage The extensive dispersion of responsibility among parts of gov- ernment means that there are often jurisdictional conflicts and over- lapping responsibilities among different governmental organizations. The existence of these overlaps can provide the opportunity for par- ticular organizations to apply leverage beyond their organizational boundaries. The "crisis" involving groundwater and contamination of foodstuff with ethylene dibromide (EDB) is one example. At the federal level the EPA, the Food and Drug Administration, the De- partment of Agriculture, and the Occupational Safety and Health Administration all had responsibility for some aspect of exposure to EDB. They had more or less reached agreement as to the han- dling of the pesticide. But action by state government agencies in Florida and Massachusetts brought EDB to the public attention and forced changes in the response from the federal government. The Massachusetts EDB team leader, Dr. Havas, summed up the state

114 IMPROVING RISK COA{:MUNICATION Department of Public Health experience like this: "It was a success . . . particularly how quickly we got EDB out of the Massachusetts food supply. What we did drove EDB out of the food supply for the entire nation . . . not just for Massachusetts . . . everybody got the benefit" (quoted in Krimsky and Plough, 1988~. Diffi:culty in Determining Responsibility for Outcomes Dispersion of responsibility and the actions of various individ- uals and organizations to obtain as much leverage as possible can mean that the recipient of risk messages has a difficult time know- ing exactly which organizations have jurisdiction over the hazard in question. Various organizations may have competing aims or goals with respect to the hazard in question and the control of the associ- ated risks. The resulting confusion can constitute a problem for the risk communicator because he or she needs to clarify the organiza- tional responsibilities as well as the risk involved. The fragmentation of risk control decision making thus contributes to the difficulty of communicating about risks. Unbalanced Access to Formation If the group of people that a risk communicator is trying to reach thinks that the system for generating information relied upon by that source does not consider its concerns, it may reject the information from that source as a basis for Sections about risks. Rejection of its information can be a considerable problem for a risk communi- cator. Organizations disseminating risk messages need to be aware of the effects of uneven access to information by those affected by or requesting the organization's action. Information is not free. It is expensive to develop empirical data, and there are not enough research funds to examine all questions that might be relevant to particular issues. Thus the amounts of information about all considerations relevant to such decisions are unequal and may therefore introduce imbalance into the information base for risk decisions. Government and industry spend large amounts of money on re- search. This not only encourages their concerns to be reflected in research projects but also establishes patterns of information flow and interactions that reinforce this effect. Environmental groups or trade unions do not have equal amounts of money to fund research

PROBLEMS OF RISK COMMUNICATION 115 and may be at a disadvantage in justifying their positions in con- flicts about regulatory decisions or other risk management strategies. However, they can often serve the valuable function of criticizing the information developed by other organizations. Science tends to be conducted in institutional settings with strong incentives the amateur scientist was, for the most part, a character of the last century. Researchers at universities and other independent research facilities are subject to powerful influences, both through budgetary constraints and the need to publish their results in peer-reviewed journals. There never will be enough re- search funds to pursue all questions relevant to particular hazards. Funds that do exist may be inappropriately allocated. Issues that are popular in particular disciplines may thus introduce imbalance into the information base for risk decisions. Even when information has been created, it may not be equally accessible to everyone. The research community can be reached more easily by those with resources to support, follow, and interpret its activities. Local citizens' groups are likely to have even less contact with relevant research communities. They will probably be unfamiliar with the language of science and may not formulate their questions in ways that scientists can use. This may detract from the usefulness of public hearings and other settings where exchange might take place between the providers of information and concerned citizens. Systematic Interests and Biases Those most strongly motivated to communicate about risk are of- ten also those with the strongest interest in the decisions. So whenever a personal or a social dLecision may affect interested groups, conpicting messages that repect the conpicting interests may be expected. The beliefs of risk communicators, and their interests, create incentives to slant or even distort or misrepresent information. This can skew messages in many different directions on the same issue. The American Cancer Society and the Tobacco Institute offer conflicting messages about the health effects of smoking, the National Agricultural Chemicals Association and the National Farmworkers Union are in conflict about the health risks of pesticides, and the Sierra Club and the Edison Electric Institute take different positions about the dangers of acid rain. The reasons for these differences may be complicated, but smokers contemplating quitting, farmers

116 IMPROVING RISK COMMUNICATION considering the adoption of integrated pest management, and citizens taking positions on the regulation of air pollution are confronted with making judgments about the risks by weighing messages from obviously interested sources and messages from other sources whose biases are not so obvious. Consider the experience with messages about AIDS. Fearing the response to the epidemic of a traditionally homophobic society, var- ious groups representing the gay community have at different times underplayed and exaggerated the risk of AIDS (Shilts, 1987~. Ini- tially, the gay community denied that there were special risks asso- ciated with homosexual practices and sought to protect bathhouses and other gathering places from interference by public health offi- cials. As the toll has increased, the tendency has been to claim rapid spread of the disease among heterosexuals. Gay community groups tended to describe AIDS as a societal affliction not concentrated in an isolatable and stigmatized group. When everyone is a potential victim, both compassion and resources are likely to become more plentiful. For their part, blood banking organizations have consistently sought to underplay the risks of AIDS contracted through transfu- sions. A prime motivating factor has been their need to maintain an adequate supply of blood for the nation. If blood is linked to a new and highly dangerous disease, the public might, as has happened, curt ail donations in a mistaken belief that there is a risk to donors. Until 1982 the blood banking community rejected epidemiological evidence that AIDS could be transmitted through banked blood and told the public the blood supply was "safe," when all that was known was that the risk of AIDS had not been convincingly demonstrated. The overriding concern was a desire to reassure the donating public (Holland, 1987~. The blood banking community continues to claim that the blood supply is "as safe as it possibly can be for AIDS," although some recommend that additional screening procedures be used (Holland, 1987~. The point is that on matters of public controversy risk messages tend to be flavored by the positions taken by the sources of the various messages. Moreover, these biases are not necessarily obvious to those who receive the messages and use them to make personal decisions or to inform their political positions.

PROBLEMS OF RISK COMMUNICATION PROBLEMS OF RISE COMMUNICATORS AND RECIPIENTS 117 Examinations of risk communication have tended to focus on the preparation, presentation, and transmission of messages about the nature of risks and risk reduction measures and on their receipt and interpretation by the intended recipients (Covello et al., 1987b; Davies et al., 1987~. Most of this attention has been directed at the problems of the individual or office preparing and disseminating risk messages. Here we describe many important problems the risk communicator will face in these tasks as well as the special problems of the recipients of such messages. We also examine aspects of the interactions between the risk communicator and other groups: other people within his or her organization, other groups or organizations, and the intended recipients. One of the central aspects of risk communication is that risk messages are not created and transmitted in a vacuum. The policy, administrative, or political arena within which the communication process occurs is an important influence on what eventually happens. We describe problems that derive from within the risk communica- tor's organization or group as well as those that characterize the broader setting of interactions with other individuals, groups, and organizations. Debates between risk managers and experts, or between experts and members of the informed and involved public, are often poorly understood by the general public. Although such debates are not particularly well attended to, they are also not ignored. Risk debates often are interpreted by the general public in two ways: the world is a dangerous place, and risk managers either do not know what they are doing or do not understand what they are supposed to be doing. In other words, risk debates often generate fear, which is unpleasant and generally not helpful for making decisions. Neither heightening of public fear nor heightening of public distrust of risk management can be considered constructive as such. But even though risk com- munication may engender at least some fears in the public regardless of content and procedural safeguards, we fee! that it is a necessary and important part of risk management in a democracy. The risk communicator attempts to present information in such a way that the intended recipients will receive and attend to its message. Usually, the risk communicator presents this information in the hope of influencing the recipients' attitudes or actions. But the recipient may not particularly care about the issues raised by a

118 IMPROVING RISK COMMUNICATION message. Many messages are likely clamoring for his or her attention, and those about the same issues are likely to be contradictory. The recipient is faced with the difficult task of making sense out of a confusing mess of information from many different sources. To some extent, the problems of the recipient of a risk message are the mirror image of those of the risk communicator. This is the reason we address them together. The communicator worries about credibility because the recipient judges messages on the basis of the reputation of the source as well as the content of the mes- sage. Risk messages not only need to be, but must also appear to be, accurate and responsible representations of the issues because the skeptical recipient will be on the lookout for incompetence, inac- curacy, misrepresentation, and deceit. Similarly, the communicator tries to be clear and easily understandable because most recipients have difficulty with complex technical material. Occasionally recipients of risk messages become risk communi- cators. When an individual becomes motivated to join or create a group with the aim of influencing decisions about risks, he or she generally disseminates oral or written messages to others. In these situations that individual will experience not only the problems of interpreting risk messages from other sources but also many of the problems of risk communicators. The risk communication process then becomes interactive in its most fundamental sense. In exarn~ning the problems of risk communicators and risk recip- ients, we describe several things that make easily understandable risk messages difficult to achieve. We present general conclusions about mistakes to be avoided. The attributes of risk communicators and risk messages we identify and their impact on the risk communication process are derived from this general research base and our collective judgment. The research is much weaker, however, in giving guidance about what will work in specific situations. The only way to be sure is to pretest communications materials with representatives of the intended audience. Establishing and Recognizing Credibility Lack of credibility alters the communication process by adding distrust and acrimony. The most important factors detracting from the credibility of a risk message relate to the accuracy of the mes- sage and the legitimacy of the process by which its contents were determined, as perceived by the recipients.

PROBLEMS OF RISK COMMUNICATION 119 The perceived accuracy of a message is hampered by the follow- ing: real OT perceive] advocacy of a position not consistent with a careful assessment of the facts; reputation for deceit, misrepresenta- tion, or coercion; previous statements or positions that do not support the current message; self-serving framing of messages; contradictory messages from other sources; and actual OT perceived professional incompetence and impropriety. The perceived legitimacy of the process by which the contents of the message were determined depends on the following: legal standing, justification of the communication program, access of affected parties to the decision-making process, and fair review of conficting claims. ~. ~. ~ . Real or Perceived Advocacy of Unjustified Positions Perhaps the most critical element of credibility for a source is the degree to which intermediaries and the ultimate recipients of the risk message believe that source to be justified in the position reflected in the message. As we have already pointed out, it is extremely unlikely that the recipients will be in the position to judge the accuracy, balance, and fairness of a risk message from the content of that message alone. One result is that recipients tend to judge the messenger as well as the message. The reputation of the source, . In terms of past record with regard to accilra~v elf ~f~nt.~nt. zany _ . · . . . ~.. . . ~. legitimacy of the processes by which it is developed, will be an important influence on the way recipients view particular messages. ~t Is Important that an organization ensure that its positions are technically competent. An unfortunate example of the failure to T. · · . . . ~ . . . _ do so involves the EPA and its decision to conduct a chromosome damage study of the residents of Love Canal in New York (Levine, 1982~. At the outset the decision was made, apparently by EPA lawyers to restrict the number of people studied because of the high cost of studying chromosome damage (Davis, 1987~. Individuals were therefore selected to maximize the likelihood of damage, following the reasoning that comparison of what should be a high-risk group to a group from an uncontaminated neighborhood should make it easier to determine whether a larger study would be justified (Levine, 1982~. However, funds were further curtailed, so that a total of only 36 cases could be included and the control group from the uncontaminated neighborhood was eliminated. This ensured that the results would be difficult to interpret, which was recognized by the scientist conducting the study. Given the extreme emotions

120 IMPROVING RISK COMMUNICATION surrounding the events at Love Canal, the result was considerable controversy. Five independent reviews of the chromosome study were submitted to the EPA, two requested by federal agencies and three by the scientist conducting the study. All emphasized the limited inferences that could be drawn due to the lack of a control group. The reviews commissioned by federal agencies criticized the interpretation of the data on chromosome damage in the study, while those requested by the scientist conducting the study were more favorable concerning the data interpretation. Although this example is extreme, scientific studies are subject to strict examination of their methods of data collection and interpretation. This examination is usually severe when the studies are used to support controversial public policy decisions. Reputation for Deceit, Misrepresentation, or Coercion Perhaps the most difficult problem for credibility is a past record of deceit, misrepresentation, or coercion. For example, as was ac- knowledged to us by officials from the U.S. Department of Energy (DOE), one of the biggest problems confronting the civilian radioac- tive waste program at DOE is the legacy of the Atomic Energy Commission and even earlier government programs (Isaacs, 1987~. The attribution during the 1950s of fallout in St. Louis to Russian sources when in fact it was known to come from tests in Nevada was a blatant abuse of public trust, the repercussions of which the DOE must live with today. When the responsible government organiza- tions have been proven to lie, it is not surprising that people want independent verification. One year of being honest with the people is not enough. Given the knowledge today of the cavalier treatment of facts concerning its activities in the past and the tremendous op- portunity for uncertainty to enter its analyses and for its analyses to be skewed, the DOE faces tremendous credibility difficulties. Even the slightest indication of less than complete candor and honesty will probably lead many people to reject whatever position the agency takes. Given the highly politicized issues that DOE's program ad- dresses, this legacy adds to an exceedingly difficult challenge. The situation is somewhat different for nongovernment organi- zations. Private corporations, advocacy groups, and private citizens are commonly expected to interpret the facts of the situation in ways that support their aims and goals. This is part of the reason cor- porations and their messages are distrusted. Despite the difficulty

PROBLEMS OF RISK COMMUNICATION 121 that recipients of risk messages have in recognizing misreading or deceitful messages, a reputation for consistently bending the facts to fit one's purposes will undermine one's credibility to many recipients, although one's direct constituency may become more supportive. Contradiction of Previous Positions Establishing and defending credibility is difficult when the mes- sage represents a departure from previous positions. In large part credibility derives from the demonstration over time of consistent competence and fairness. Both scientific incertitude and changes in policy can serve to undermine credibility to the lay public. The ne- cessity of correcting mistaken statements or positions can undermine credibility with the public. Care must be taken to demonstrate why the interpretation of scientific or policy conclusions has changed. The rapidity with which new scientific findings about the AIDS virus that counter or revise previous positions are being presented undermines the credibility of the experts. At least one response to the rapid-fire changes in estimates and contradictory conclusions about heterosexual transTn~ssion of the HIV virus is to conclude that the experts really do not understand what is happening and that every imaginable precaution is thus justified. Inconsistencies can also result from changes in administration, as was clearly illustrated by the treatment of formaldehyde by the Occupational Safety and Health Administration (OSHA). In Decem- ber 1980, OSHA, acting in conjunction with the National Institute for Occupational Safety and Health, released a Current Intelligence Bulletin recommending that formaldehyde be considered a potential carcinogen and that appropriate controls be implemented to reduce employee exposure to the chemical. In March 1981 the new Assis- tant Secretary of Labor for Occupational Safety and Health rescinded OSHA's sponsorship of the bulletin. In January 1982, OSHA denied a petition by labor unions for an emergency temporary standard to reduce formaldehyde levels in the workplace (Ashford et al., 1983~. Such reversals of position based on the same evidence can only rein- force the appearance of inconsistency and undermine the credibility of the source. Self-Serving Framing of Messages The leeway that exists in the collection of data, its interpreta- tion, ant! the final crafting of messages provides ample opportunity

122 IMPROVING RISK COMMUNICATION to present information in ways that support the position of the organization more strongly than the evidence itself might justify. Such framing of messages need not involve direct deception or lying. For example, on August 27, 1986, the Washington Post reported on the Soviet government analysis of the potential health effects of the Chernobyl disaster under the heading, "Chernobyl Report Sur- prisingly Detailed but Avoids Painful Truths, Experts Say" (Smith, 1986~. The "painful truth" was that the disaster may result in 35,000 to 45,000 cancer deaths in the Soviet Union and that "as many as 90,000 people could be affected by the recent explosion." But the Soviet report said that fatalities would be "less than 0.05 percent in relation to the death rate due to spontaneously arising cancer." Since this percentage works out to be 35,000 to 45,000 premature deaths over the lifetimes of the people exposed, both reports are equally accurate. But the two messages stimulate very different responses in most people. The incentives to slant the presentation of information to sup- port an issue one believes to be important can be strong. But in order to strengthen their credibility, public service organizations, and especially those in government, must resist this temptation. Contradictory Messages from Other Sources The adversarial nature of the American regulatory system is of- ten cited as one of its strengths. But it also can help undermine the credibility of sources of risk messages. Parties with potential gain or Toss are motivated to develop the best evidence and strongest arguments for their respective positions. When this is accomplished according to strict rules of evidence and scientific review, it can be expected to produce a reasonably complete picture of the issues in question. However, it also encourages competing interpretations of the evidence whenever there is uncertainty in the data, in appli- cable methods, or in the models for interpreting empirical results. Sometimes science is claimed to support all sides of a conflict about risk. Conflicting messages also can derive from sources that are usu- ally not associated with different "sides" in an adversarial situation. For example, state officials in Florida and Massachusetts sent clear signals that ethylene dibromide (EDB) should be of serious concern at exactly the same time that the EPA administrator was attempting to reassure the public (Krimsky and Plough, 1988~.

PROBLEMS OF RISK COMMUNICATION .. . . 123 in most situations the risk communicator will have to deal with conflicting messages from other sources. Contradictory messages can be a central part of the issue as seen by the public or may be a relatively minor side point. If the communicator does not deal with these effectively, it can undermine his or her credibility. The extremes the recipient needs to watch out for are intentional manipulation and outright dishonesty. Although it is generally dif- ficult for the layperson to determine from the message itself when it is manipulative or dishonest, some are so poorly crafted as to be blatantly miscreant. Less extreme instances are exceedingly difficult to identify. One strategy is to see how other individuals and or- ganizations with a stake in the issue respond to the positions and statements of that source. The other parties in an issue have in- centives to sort out misrepresentations and unsupported findings of their opponents and make them known. The difficulty of deterrn~ning the degree to which a risk mes- sage reflects advocacy of a particular position can be illustrated by reference to the experience with EDB in the early 1980s (Krimsky and Plough, 1988~. Environmentalists critical of the basic federal regulatory treatment of pesticides found the debate on EDB to be an excellent opportunity to press their argument. They were joined by some state officials, who used the issue as a lever to force the federal agencies to act. Industry and trade associations thought the issue illustrated the need to weigh economic considerations in the regulation of hazardous chemicals. The EPA had to both defend its previous decisions with regard to EDB and quickly evaluate new data on exposure levels. The news media attempted to package all this in a way that would be newsworthy. And the recipient had to weigh all these positions against each other. Another problem for the recipient of risk messages is to differ- entiate conflicts based on scientific disagreement about the facts or their interpretation from conflicts based on advocacy of policy aims. Sometimes this requires determination of the extent to which the facts reflect advocacy. If the source has identified its aims and pur- poses, this task can be made easier. Unfortunately, many sources are much more concerned with the outcome of the issue under question than with helping the recipient understand all its ins-and-outs. In fact, some sources deliberately confuse such questions and obfuscate rather than clarify because such actions contribute to the possibility that their position may hold sway. The recipient of risk messages

124 IMPROVING RISK COMMUNICATION may thus have considerable difficulty separating scientific conflicts from policy conflicts. One of the effects of "dueling experts" may be to reduce the im- portance of expertise in the minds of the recipients of risk messages. If scientific evidence is inconclusive, then why not use whatever fac- tors seem to suggest a solution? One result of such an approach would probably be to shift the nature of the problem definition away from the aspects of risk that have been quantified to other factors that may or may not be measurable. This has probably happened with respect to nuclear power. The technical questions are no doubt less critical to most people than their beliefs about the overall impact of increasing or decreasing reliance on nuclear reactors. Professional Incompetence and Impropriety A major element helping determine the credibility of risk mes- sages is the perceived competence of the individual or organization concerning the subject under question. An individual with special training about the phenomena involved is often accorded greater credibility than someone whose training is less relevant. The state- ments of a physician, for example, might be given greater credence with respect to a public health question than those of a dentist or health economist, even though each is a health professional. Simi- larly, organizations enhance the credibility of their messages about technical issues when they have professional staffwith training in the areas covered. Into strong criticisms of government agencies are that they do not have sufficient stab with the necessary professional com- petencies to fully and completely understand the phenomena they regulate and that the understanding that the professional staff does have is not shared by the nontechnical personnel at the apex of the · - organization. Private organizations, especially business corporations, are more often suspected on account of propriety than competence. Many people believe that corporations have, or hire, the best expertise available but that these experts present only information that is in the interest of the corporation. Since the decisions of private organizations are not subject to the constraints of due process as are those of government, there is a strong presumption by many that the messages of private organizations emphasize only the information they believe to best serve their interests. Similar descriptions apply to many public interest or public advocacy groups.

PR OBLEMS OF RISK COMMUNI CA TI ON 125 There are thus strong incentives for an organization disseminat- ing risk messages to be as open and clear as possible about the way it gathers and interprets information. This is especially true if it incor- porates scientific peer review or other technical review procedures. Demonstrating the professional competence of the organization and the propriety of its procedures will likely enhance the way its mes- sages are received. Legal Standing An important influence on perceived credibility is legal standing and involvement in the issue under question. Being lawful or corre- sponding to the dictates of the law is a fundamental component of legitimacy. As we have indicated above, a special sense of responsibility attaches to government. Government officials and agencies are ex- pected to act in the public interest. At the most fundamental level the legitimacy of a federal agency to speak to an issue derives from its statutory mandate and from the exercise of due process under the Administrative Procedures Act. Determining this is relatively straightforward, although there may be questions of jurisdiction among different agencies. For example, there may be disagreement about whether the EPA or the OSHA should regulate airborne toxic contaminants in particular situations. On occasion, however, government agencies are expected to act and, if appropriate, interact with the public even when the topic is not strictly within their statutory responsibilities. For example, the EPA feels obliged, correctly many would agree, to deal with radon exposures in homes even though radon is not strictly within its statutory responsibilities. The obligation of public officials and organizations to act in the public interest even when their charge is less than perfectly clear can make the risk communicator's job more difficult. When the EPA publicizes the standards it sets, the reasons for them, and the penalties for violation for topics within its statutory responsibilities, there is little objection. But some people question the agency's justification for disseminating messages where due process has not given the agency that responsibility. Establishing legal standing and involvement for nongovernmen- tal individuals or organizations is more complicated. Whether a nongovernmental entity is "lawful" in the sense of corresponding to

126 IMPROVING RISK COMMUNICATION statutory requirements is less important than, for example, the con- cept of determination of "standing" in court, that is, of who has the right to bring suit. This concept includes notions of being materially affected by the outcome or being a representative of those who at least potentially are so affected. Being materially affected usually ensures the right to have a say in issues of public choice. In addition, personal involvement in or knowledge of the activity or events under question generally increases interest in what that individual has to say about that question. Such justification is used with respect to environmental interest groups. Finally, there is the expectation that any citizen should have the right to speak to any public issue. Of course, being party to the creation of a hazard automatically grants standing in determining how to control that hazard. There was no question, for example, that Union Carbide should be heard from following the release of methyl isocyanate from its plant in Bhopal, India. Needless to say, however, most people expect that an organization that contributes to the creation of a hazard will make statements they believe will be in that organization's own best interest. Justification of Communication Campaigns People may view what it means to be a responsible commu- nicator somewhat differently for government and nongovernmental organizations. Many people feel that government agencies should never "advocate" in the ordinary sense, that the job of public offi- cials is to determine the factual situation, identify the impacts on affected parties, and lay out the options. The selection of what is to be done should be left to elected bodies or to the relevant indi viduals and organizations. In Chapter 4, however, we discusser] the acceptability of influence in risk messages. We observed there that there are no clear lines distinguishing techniques that are appropri- ate from those that are not and that it is important to be able to demonstrate that the effort derives from a social decision supporting the communications program. Most people would agree that an important part of the activities of federal agencies is communication about the standards they set, the reasons for them, and the penalties for violations. However, it is not always obvious when government ought to undertake programs of risk communication and how far it should go in persuading the

PROBLEMS OF RISK COMMUNICATION 127 populace to undertake particular actions. For at least some pro- grams there are serious questions about the conditions under which persuasive risk communication should be permitted. The central issue here is the compatibility of government-spon- sored programs with individual autonomy and related values (Fader, 1987; Faden and Beauchamp, 1986~. In the first place the basis for the government promoting certain life-styles over others is not clear. With major problems that affect a large portion of the populace and inflict substantial monetary and social costs, most people are prepared for government agencies to be proactive and persuasive. But exactly what makes an issue a major problem justifying advocacy on the part of government is not clear. This, of course, is not to claim that government should never conduct risk communication programs. Some problems addressed by collective programs can be most efficiently dealt with by the affected individuals. But some question how the EPA can judge the risk of household radon exposure to be sufficient to warrant an extensive campaign of communication to homeowners, while that risk is not deemed sufficient to warrant establishment of exposure standards. One problem that may confront a government source is to justify conducting a risk communication effort rather than devoting its effort directly to reducing the risk in question. However, this can usually be justified in terms of improved efficiency in implementation of risk reduction programs. People are generally more tolerant about communication from private enterprises or interest groups than from government agencies. These organizations are not expected to exhibit the same impartiality as government, and their attempts to present persuasive information are not viewed with the same suspicion. - Access of Affected Parties to the Decision-Making Process Alienation of citizens due to the difficulty of getting government officials to listen to them or due to the judged inappropriateness of the officials' response has been repeatedly described as a major moti- vation of individuals who have become active in controversial issues. The homemaker/activist/media spokesperson appears frequently in community disputes over environmental issues (Levine, 1982; Mazur, 1987; Spain, 1984~.

128 IMPROVING RISK COMMUNICATION An important part of the legitimization of government activities that has been codified in the Administrative Procedures Act and else- where concerns guaranteeing the affected parties access to regulatory decisions. Government is, however, sometimes criticized for being too passive in this respect, for not actively seeking out those affected and informing them about proposed actions. This has become more ob- vious with respect to siting decisions and cleanup of hazardous waste facilities, where local hearings and extensive community involvement · · . are mcreasmgly necessary. Nongovernmental organizations are not under the same obliga- tion to grant the affected parties access to their decision-making processes. Nonetheless, it often appears to be advantageous to in- volve the public in appropriate ways. The Chemical Manufacturers Association, for example, organized a program called Community Awareness and Emergency Response (CAER), which helps chemical plant managers provide information to their communities on a regu- lar basis and involve the community in emergency response planning, including the chemical plant emergency response plans. The CA-ER program, which has been implemented at more than 1500 facilities, does not suggest that citizens be involved in the actual operating de- cisions but rather that plant managers regularly interact with them. It emphasizes the importance of treating citizen concerns as impor- tant and providing careful and accurate responses to the public. Fair Review of Conflicting Claims Expectations of fair and impartial treatment by government or- ganizations are of central importance. Nothing undermines the legit- imacy of government positions more quickly than the demonstration that dispensations have been unfairly granted. It is important that all claims are genuinely listened to and treated fairly. Some criticized the EPA under Anne Burford because it allegedly attempted to rid the Scientific Advisory Board of scientists holding views of which it did not approve (Marshall, 1983b). EPA officials were also ac- cused of inappropriately using portions of an industry publication in a "cut-and-paste" review of toxicology data (Marshall, 1983a: 12003. The charge that EPA officials were not treating the issues fairly, but serving industry interests better than environmental interests, became a serious challenge to the agency's credibility (Rushefsky, 1984; Sosenko, 1983~. This tarnished image caused problems for the agency for several years.

PROBLEMS OF RISK COMMUNICATION 129 Again, there is not the same expectation that nongovernmental organizations will treat conflicts fairly. Many believe, on the ba- sis of past experience, that industry and citizen action groups will present information that supports their interests in the most effec- tive way. But treating the positions taken by others respectfully and thoughtfully, and carefully and clearly laying out the premises and assumptions of one's position, are likely to enhance the reputation of nongovernmental organizations as well. Making Messages Underetandable The risk communicator needs to present information in language and concepts that recipients already understand:, that use magnitudes common in ordinary experience, and that are sensitive to the psycho- [ogical needs of the recipients. 1 Unfamiliar language In Chapter 2 we described the scientific information needed for risk decisions and the difficulty of presenting that information in simple terms that do not overwhelm the recipient. Here we will point out some of the ways the terminology of science, and of risk assessment in particular, interferes with understanding by laypeople. For those who are not familiar with it, the technical terminol- ogy of risk assessment is very difficult to understand. Research has shown, for example, that even for seemingly familiar terms such as probabilistic precipitation forecasts there is a high degree of mis- understanding in lay interpretations. People were equally likely to interpret a "who chance of rain" as "rain 70~o of the time," "rain over 70~o of the area," and "who chance of some measurable rain" (the official definition) (Murphy et al., 1980~. In this case people apparently had difficulty understanding not the probabilities being used but the events to which they were ap- plied. In other cases risk assessments confuse people because they use concepts of probability theory that are not intuitive. For example, a committed communicator can usually convey the meaning of a sim- ple probability of an event occurring (as long as it is not too small). But the notion of conditional probability is much more difficult to get across, as is that of the probability of the conjunction of several events. In the EPA's experience with EDB, people were confused by the notion of the aggregate risk of this pesticide to the exposed

130 IMPROVING RISK COMMUNICATION population. They did not know how to interpret this information to answer the question, "Should ~ eat the bread?" (Sharlin, 1987~. Much has been written about the intuitive properties of proba- bilistic events and how they differ from probability theory (see the summary in Fischhoff et al., 1981a). These include the "gambler's fallacy" (after a series of heads in flips of a fair coin, most people expect a tail) and the tendency to impose order on the results of random processes. The skeptical recipient of risk messages will be on the lookout for such influences in the material he or she receives. Recipients of risk messages also need to be wary of "framing effects" differences that can result from the way information is presented. For example, one guide for chemical plant managers points out the following ways that the annual fatalities resulting from emission of an air toxic might be presented (Covello et al., 1988~: deaths per million people in the population, deaths per million people within miles of the facility, deaths per unit of concentration, deaths per facility, deaths per ton of the airborne toxic substance released, deaths per ton of the airborne toxic substance absorbed by people, deaths per ton of chemical produced, and deaths per million dollars of product produced. The authors point out that depending on the circumstances different expressions will strike the recipients as more or less ap- propriate, more or less frightening, or more or less credible. The recipient needs to be aware that simply changing the way a piece of information is presented can alter its effect on many people and be aware of their pattern of response. Each of these ways of presenting has framing effects, which suggests that using more than one might be useful in some circumstances. Misunderstanding can also result from inconsistency in the use of the same term among different disciplines. For example, the term "risk" has been used with a variety of somewhat different meanings. Uses include the total number of deaths, deaths per person exposed or per hour of exposure, loss of life expectancy due to exposure, and loss of the ability to work (Fischhoff et al., 1986~. The recipient also needs to look for the sources of uncertainty In the analysis. As we pointed out in Chapter 2, at least four types .

PROBLEMS OF RISK COMMUNICATION 131 of uncertainty may be found in risk messages: (1) weaknesses of the available data, (2) assumptions and models on which estimates are based when data are missing or uncertain, (3) sensitivity of the estimates to changes in the assumptions or models, and (4) sensitivity of the decision to changes in the estimates. Unfamiliar Magnitudes Risk communicators are generally well aware that most people have difficulty comprehending magnitudes that are exceedingly small or exceedingly large. Often they utilize analogies to convey such magnitudes. For example, a risk of 0.05 may not mean much to most people; the statement that about 5 people in an auditorium of 100 would be affected is much easier to comprehend. A cancer risk of 4.7 x 10-6 is difficult for most people to relate to. But it may be more understandable to imagine 10 cities of 100,000 people each, all exposed to the hazard. In 5 of the cities there would be no eject, and in 5 cities there would be 1 additional cancer as a result of the exposure (Covello et al., 1988~. Such mental aids must be used with extreme care, however. For instance, in this example there could be 5 additional cancers in a single city instead of spread across the cities, and there could be an overall total of 10 or of 0. Recipients need to look for the magnitudes in risk messages and how they are presented. These numbers are subject to the same kinds of presentation effects as the concepts described in the previous section. Recipients need to be especially wary of misrepresentation that can be introduced in comparing risks. In particular, magnitudes do not always represent the level of hazard. For example, characterizing the magnitude of ash emitted as filling an olympic-sized swimming pool or covering a football field to a depth of 6 inches omits any reference to the potency of the material. Because of differences in potency, a fairly small amount of one substance may present the same risk as a much larger amount of another. Insensitivity to Psychological Needs of the Recipient Factors influencing understandability intermingle with certain psychological components of decision making. Motivational factors can be involved. For example, it may be difficult to determine whether the intended recipients have not understood the message or whether they have understood but decided for whatever reason

132 IMPROVING RISK COMMUNICATION not to heed its contents. People may want simple yes-or-no answers, and they may want to know what they as individuals should do. When expecting information in such a format, they are likely to have trouble understanding information presented in some other format. The first image people receive about a problem also tends to be the strongest and longest lasting. If they make up their minds on the basis of that image, it will be hard to get them to change. But the risk communicator can use such psychological attributes to his or her advantage, as well. If the risk communicator is timely and presents a vivid image, he or she can have considerable impact. With certain issues and certain parts of the population, commu- nication may be especially difficult. There may be, for example, a climate of mistrust in parts of the population about anything that can be labeled toxic. These people may automatically reject a mes- sage and oppose the production, use, or disposal of products labeled toxic regardless of the risk estimates of experts. For them it may be that risk messages would elicit little differentiation of response regardless of their format, message content, or the organization from which they emanate. People are unlikely to be interested in risk information that they cannot use. A risk communicator wishing to change the recipient's thinking (even if only to make him or her better informed) thus needs to try to understand how that person receives, processes, and acts on information. Elsewhere we discussed the psychology of risk perception and social factors that influence perceptions of risk. Here we review only the more important of those psychological and social influences. People differ. Their interests, life-styles, and living conditions vary. What they do in their private lives and how they interact with others in their public lives will strongly influence how they are likely to use risk information. The risk communicator will be most effective if these attributes of recipients can be reflected in the risk message. For issues that affect large numbers of people, it will nearly always be a mistake to assume that the people involved are a ho- mogeneous group. It is therefore generally necessary to segment the population into groups with similar needs. It is often useful to craft separate messages that are appropriate for each segment. Prepar- ing messages appropriate for different segments of the population requires determining what the recipients already know or think they know, what is necessary for a full and sufficient understanding of the risk and risk reduction measures, and how they would be able to use

PROBLEMS OF RISK COMMUNICATION 133 new information. Depending on the numbers involved, this can be expensive and time consuming. The purpose the message serves can dominate people's informa- tion needs and therefore the content of an effective risk message. For example, information needs will probably be quite different when the situation calls for providing emergency instructions, for alerting peo- ple to a previously unrecognized risk, or for providing information that a risk is actually less serious than was previously thought. A common mistake is to expect quantitative risk assessments to include everything people are concerned about. Affective states (those involving or appealing to emotion) are equally or more impor- tant than physical conditions to many people. Since risk assessments are usually limited to physical events and consequences, they can be expected to speak to only part of what concerns most people. These other aspects of risks that concern people are sometimes called qual- itative risk factors. Thus, in order to present information that is relevant to the intended recipients, it may be necessary to expend some effort to find out what is bothering people. To be effective, a risk message needs to refer both to information about risk and risk reduction and to the psychological or affective factors that influence the intended recipients. Unfortunately, it can be difficult and expensive to de- ve~op empirical Information about recipients, especially if they are geographically or culturally dispersed. It is important not to expect too much from risk communication efforts. Advertising campaigns are considered successful if they result ~ .. . . ~ .. in shifts of a few percentage points in the market for a product. It took decades of multiple messages from many different sources to create major shifts in public attitudes about smoking. It is hard enough to make risk messages understandable to laypeople. It is harder still to know whether risk messages have an impact on their thinking. Preparing Messages with Few Data and No Time Sometimes the risk communicator must disseminate messages when there are not enough relevant data to draw satisfactory conclu- sions and; there is no time to obtain better information. This usually OCCURS in one of the following situations: f1) an emergency requires immediate action OT (29 events lead to requests for information prior to the completion of study or analysis.

134 IMPROVING RISK COMMUNICATION Responding in an Emergency Emergencies occur when external events take control and require action by an individual or organization. They often, but not always, require immediate issuance of warnings, instructions, or advice about what to do. Examples of such emergencies include Three Mile Island, the Tylenol poisonings, and emergency releases from chemical or other industrial facilities. The problem is most extreme in a true emergency when no preparation has been made in advance of the event. For example, the Nuclear Regulatory Commission was almost totally unprepared for an accident at the time of Three Mile Island (Ahearne, personal communication with National Research Council staff, 1988~. There was no effective management structure to support emergency decision making, and time was lost figuring out who should do what. The lack of preparedness permitted the involvement of too many who lacked the technical competence to grapple with the emergency, thereby slowing the rate at which necessary information could be generated and interpreted. No one with the technical background to explain what was happening had been assigned the role of spokesperson, and it was a couple of days before a credible source of information emerged. Nor did the agency, much less the electric power company involved, appreciate the importance of timely and accurate news releases. Finally, the agency had no notion of how to deal with the electric company or the news media in such an emergency. Emergency situations are likely to expose the risk communicator to conflicting motivations. For example, a company dealing with an emergency release of toxic substances into the air, such as that from the Union Carbide plant at Institute, West Virginia, in August 1985, will probably balance several competing factors in deciding what messages to give out (Coppock, 1987~. After the initial emergency response, when the overriding concern is what to do to contain and stop the release, almost every business person immediately wonders who will bring suit. The common view among legal advisors is almost always to give out as little information as possible so as to avoid providing ammunition for use in court. This is in almost direct conflict with what communications and community relations experts advise, which is to say everything that is known, as quickly as possible, in terms the layperson can easily understand. Advice of company scientists and engineers usually falls somewhere between these two views. They caution against attributing cause and eject

PROBLEMS OF RISK COMMUNICATION 135 before being reasonably certain about what happened. The message that is finally sent out probably involves compromises between these three points of view. Situations involving emergency response are often governed by special considerations not shared with other kinds of risk communica- tion. We have chosen to focus primarily on the other, more prevalent situations of risk communication and therefore do not discuss emer- gency response in detail. Communicating on the Basis of Incomplete Information It is often difficult to estimate risks, consequences, and possible risk reduction measures with any precision. One result is that the risk communicator may be left with very little information that can be presented with confidence. As one scientist at the EPA put it, "One of the nice things about the environmental standard setting business is that you are always setting the standard at a level where the data is lousy" (quoted by MeInick, 1983:244~. The poor quality of relevant information is also often involved in pressing issues. When the concern about EDB shifted from ground- water contamination in a few isolated wells to residues in food prod- ucts, EPA administrator William D. Ruckelshaus sent a letter to the governors of the 50 states requesting data on residue levels in food products. He hack to answer queries by admitting that his agency did not have the answers. "If they tthe public] want absolute informa- tion, we can't give it to them." For a period of nearly a month, the best he could do was say, "l don't want to unduly alarm the pub- lic, nor do ~ want them not to know about it" (Sharlin, 1987:192~. The risk communicator may often fee] as if the world wants to know definitive answers to questions about which he or she has no adequate information. External demands can also force an organization to make state- ments on the basis of limited data. Examples include Love Canal and transmission of the AIDS virus. Another form in which this problem can be found ~ the decision about whether to release prelirn~nary information or tentative results. In 1986 the EPA began cooperating with the New York State Energy Research and Development Author- ity (NYSERDA) on a program monitoring radon levels in geographic regions thought to have radon problems (Smith et al., 1987~. Three monitors were placed in each home, one in the basement for 2 to 3

136 IMPROVING RISK COMMUNICATION months, a second in the basement for 12 months, and a third in the living area for 12 months. Originally, the plan was to give the home- owners the readings from all three monitors at the end of the study. But in the spring of 1986, when radon became very much a public issue, NYSERDA became concerned that they would be accused of withholding public health information if they kept the short-term reading until the end of the study. It was decided to provide the initial basement reading to the homeowners, even though the full research design called for confirmation of annual exposure levels and living-area exposures with the other monitors (Fisher, 1987~. Capturing and Focusing Attention Many other things compete with risk messages for attention, and the risk communicator often has difficulty getting intended recipients to attend to the issues. There are two separate aspects to this prob- lem: (1) stimulating the attention of recipients and (2) interacting with the news media and other intermediaries. Stimulating Recipient Interest It is not always easy to capture the attention of people who re- ceive risk messages. Most information campaigns share the following attribute: the people most likely to receive messages and to attend to them are those who already possess some information about the issues under question; those who may be characterized as relatively uninformed are less likely to receive and pay attention to messages. The people who need information most seem to be the least likely to pay attention. One contemporary example might be the very low likelihood that intravenous drug users will attend to messages about AIDS transmission via dirty hypodermic needles. Involvement in community affairs has been characterized as a pyramid. At the bottom is the broad base of most people who are uninvolved in any personal sense and basically are uninterested. A somewhat smaller number of people are aware of issues but do not go to much effort to obtain additional information. A still smaller group actively seek information on particular issues. The number of people actually participating in organized efforts is smaller still. Finally, some individuals seem to participate in, and often lead, every activity in a community (Verba and Nie, 1972~.

PROBLEMS OF RISK COMMUNICATION 137 One of the consequences of this differential interest and involve- ment among various parts of the public is that information readily made available will generally be taken up much more readily by some people than by others. Educated and involved people usually absorb information much more quickly. But conflict can motivate otherwise uninterested people to gain more information. In some situations, when the principal aim is to stimulate understanding in a broader sector of the public, it might be useful to stimulate conflict. Very often journalists and the media seek out such conflicts and serve as information channels as conflicts play out. The majority of people, however, will probably not be interested in the issues addressed in a particular risk message. When a signif- icant number of people are similarly affected, a champion for that group is likely to emerge, especially when the impact is undesirable. Such people can be engaged in risk communication activities as de- scribed elsewhere in this report. When trying to affect the behavior of uninterested, uninvolved people, however, the risk communicator will need to find ways of attracting the attention of intended recipi- ents and making the message meaningful to those people. This will probably be easier if the risk is one that is perceived to directly im- pinge upon people and for which there are clear control measures that do not substantially interrupt their private lives. For example, people have tended not to heed messages about seat belt use, maintenance of automotive emission control devices, and radon contamination of homes. However, it is difficult to determine whether they simply paid no attention or whether they received the information, understood it, and decided not to act in accordance with the proffered advice. Different people rely on different information channels. They read different newspapers and magazines and listen to different radio and television stations. They may turn to different information channels for different purposes. Young people, for example, may rely on mass media sources to learn about the AIDS epidemic and its spread. But they may turn to their friends in determining whether to be worried and alter their behavior. Risk communicators need to know what channels their intended audience uses for what aspects of risk information. One example of this is the use of music television spots by the National Cancer Institute to convey the message to teens that it is not "sexy" to smoke, rather than providing information about the undesirable health effects of cigarette smoke.)

138 IMPROVING RISK COMMUNICATION Interacting with the News Media and Other Intermediaries The mass media are widely perceived as playing a powerful role in constructing laypeople's understanding of and attitudes about risk. Journalists and the media help identify conflicts about risk and are important channels of information during the resolution of those conflicts. There are both critics and defenders of the effects of the news media. In any case the risk communicator must deal with.the fact that some journalists tend to treat risk issues differently from the way technical and scientific people do. Some conflict between risk communicators and journalists and other intermediaries is probably inevitable. But this conflict can be reduced, and there are approaches the risk communicator can use toward achieving this aim. An important part of this is to recognize the typical differences in the way risk communicators, as sources of information, and journalists approach information gathering and dissemination. Organizations involved in risk issues typically seek to centralize and restrict the flow of information, hoping to prevent the publication of damaging information. But reporters expect access not only to public information specialists but also to experts and managers and what they know (Sandman et al., 1987a). This is especially true in emergency situations. The price of not providing that access may include suspicion, anger, and sometimes damaging coverage. Despite the legal and technical constraints, it is important to consider meeting the needs of the news media. Many journalists are proud of their ability to flesh out a story with the views of uninvolved experts, dissident insiders, and others whose perspective on an event is likely to be different from the official one (Sandman et al., 1987a). Specialized reporters are proud of their contacts and investigative reporters of their skill at finding those who know and of persuading them to talk. Tying to stop reporters from talking to people within an organization is sure to encourage them to investigate further. Differences of opinion as to what should ideally be presented are likely to exist between risk communicators and journalists. Sources sacrifice all credibility in the eyes of reporters when they lie or mislead, and they lose much of it when they err, omit, or delay (Sandman et al., 1987a). Different sources are commonly held to different standards of credibility. Industry spokespeople, for exam- ple, are often discounted as opinionated even when they are providing

PROBLEMS OF RISK COMMUNICATION 139 facts, while academics and public interest groups often are accorded the credibility of neutral sources even when they are offering opinions. Journalists, too, have a problem with credibility (Sandman et al., 1987a). A botched story not only misleads the reader or viewer but it also diminishes the source's willingness to cooperate with that reporter next time and perhaps with other reporters as well. The two most important complaints about reporters' treatments are misquotation and inaccuracy. Technical stories have greater chance of misquotation, simply because they involve terms and concepts less familiar to the reporter than nontechnical stories. Nevertheless, incompleteness and misemphasis quoting out of context" are more frequent than direct misquotation. Complaints about inaccuracy are also generally about being incomplete or misreading. Sometimes the complaint is that too much credence is given to other sources who, in that source's judgment, are wrong or intentionally misreading the journalist. These questions are commonly sources of conflict between risk communicators and journalists, especially because the journalist does not see his job as discovering the truth, but rather as reporting accurately what others with some claim to attention consider to be true. Risk messages are often routed to their intended recipients through health professionals or other intermediaries. In addition, the views of influential members of the community, such as county or local public health officers, prominent physicians, fire chiefs, and politicians, often provide valued guidance to citizens as they form their opinions about controversial issues. Sometimes executive offi- cers of professional or volunteer organizations serve as "gatekeepers," controlling the distribution of information, and their approval or dis- approval can be a critical factor in the dissemination of some risk messages. In some circumstances the intermediaries are even more important than the news media in reaching the intended recipients. Interacting with non-news-media intermediaries can also involve problems. Health departments, public libraries, professional asso- ciations, and voluntary organizations all have their own aims and purposes. They may or may not offer relevant messages to the in- tended recipients that are appropriate in terms of format and style. It may be quite time consurn~ng and costly to establish working rela- tionships with such intermediaries, however, and there is the danger of losing control over the content of the messages. Nevertheless, es- tablishing links with such institutions and organizations can shortcut the development of routes of influence with the target recipients.

140 IMPROVING RISK COMMUNICATION In this context it is important to realize that there are several different ways that messages can reach the final recipients: face to face (physician to patient, friend to friend, within the family), in groups (work sites, cIassrooms), within organizations (professional or volunteer), through the mass media (radio, television, magazines, newspapers, direct mail, billboards, and transit cards), and within the community (libraries, malls, fairs, and local government). Each of these channels offers advantages and disadvantages in specific situations. Interpersonal channels like physicians or pharmacists are likely to be trusted and influential. But messages relying on interpersonal channels require the intermediaries to be thoroughly familiar with the message and may thus require expensive and slow Tong-term contact. Community channels such as libraries and community organiza- tions can reinforce and expand upon media messages. Establishing links with community organizations can require less time than re- liance on interpersonal channels. Using celebrities can be effective if they are directly associated with the message (e.g., they have been a cancer patient, are pregnant, or successfully altered a hazardous habit). But they speak for them- seIves, and it is important to have firm agreement about what they will and will not say. The appearance of a celebrity may compete with the content of the message for attention, and some recipients may not react favorably to some celebrities. Finally, celebrities live in the public eye and a change in their popularity or personal life style could affect their impact. Working with intermediaries is essential in many situations. In- termediaries can help by providing special access to the intended recipients, credibility because the recipients consider them to be a trusted source of information, and additional tangible or intangible resources. Working with these individuals and organizations, how- ever, can also have drawbacks. It can be time consuming to locate them, convince them to participate, gain their approval, and develop and agree on their role. It can require adjustment in order to match the priorities and programs of intermediary organizations. It can result in loss of control of the risk message because they may change the time schedule, functions, or even the content of messages and take credit for part or all of the effort.

PROBLEMS OF RISK COMMUNICATION Getting Information 141 Recipients of risk messages may have difficulty deciding what to do because they cannot get information that satisfactorily answers their questions. This can result from one or both of the following: authorities who do not listen or who respond inappropriately and ~ifflcuity in finding trusted focal sources of information. Authorities Who Do Not Listen or Respond The story of the concerned citizen motivated to organize protest groups because of the cold or indifferent response of public officials is common in the literature of environmental and citizens' organizations (Fitchen et al., 1987; Institute for Environmental Negotiation, 1984; Krimsky and Plough, 1988; Mazur, 1987~. A citizen who had spent several years as an activist opposing the construction of a hazardous waste facility in her community told us of the frustration her group experienced in trying to get the authorities to take their concerns seriously and in attempting to obtain materials they could use to inform themselves and their neighbors (Smith, 1987~. She spoke of the anger generated by the lack of respect given her group's questions by government officials. At a public hearing the company proposing to construct the facility was allowed to speak freely. But questions from the public had to be submitted in writing. Nor was the citizens' group able to find support from the traditional national environmental organizations. Finally, they turned to other citizens' groups who were opposing the same company in other locations. This may be a common experience for citizens' groups focusing on locally unwanted land uses. The number of Superfund sites around the country and the pressing necessity for finding ways of dealing with hazardous wastes will make this kind of difficulty likely to reappear many times. Diff;icullies in Finding Trusted Sources of Information Other developments will result in citizens or citizens' groups seeking additional information. For example, Title ITI of the Su- perfund Amendments and Reauthorization Act of 1986, also called the Community Right-to-Know Act, includes provisions for creating emergency response plans and for reporting data about hazardous substances stored and regular ern~ssions to the EPA. The EPA must

142 IMPROVING RISK COMMUNICATION make these data available to the public. The Community Right-to- Know Act will make a tremendous amount of information about po- tential hazardous situations available to citizens who wish to obtain it. But this information is likely to be in highly technical form, most of which would require considerable interpretation to appreciate. A citizen wishing to make sense of this information about a facility in his or her community will need to interpret data from material safety data sheets developed for occupational exposures and estimate peak or periodic exposures from annual emission totals. He or she may wish to seek additional interpretations to those provided by facility personnel, and finding trusted and qualified people to interpret this information will be an important part of the process. SUMMARY We distinguish two major types of problems in risk communi- cation. Those involved in risk communication can do little about problems deriving from the institutional and political system beyond understanding them and their influence. These problems can have considerable impact on events, and if they are ignored it may be quite difficult to understancl why things happen the way they clot Prob- lems of risk communicators and recipients can be addressed more directly and are more amenable to improvement or solution. In most instances the problems of risk communicators and the recipients of risk messages are mirror images of each other. In the next chapter we describe conclusions and recommendations that are intended to improve risk communication in ways that will address the problems of risk communicators and of the recipients of risk messages. NOTE 1. This is unlikely to meet our criteria of informing or of accuracy of the message.