Promoting Remedial Response
to the Risk of Radon:
Are Information Campaigns Enough?

Coping with environmental hazards has become a fact of life for residents of communities everywhere (Kasperson 1977). Although we have begun to identify the factors that shape perceptions of risk and that influence the actions people take, (Note: 1) less is known about the types of information most needed by the public and the best ways of communicating that information (Kasperson 1986),

In discussing environmental risk communication, it is necessary to distinguish between hazards managed by direct government action (including regulation) and those that require action by individual citizens Most pollution problems – the kinds that environmental agencies normally handle – fall into the “government action” category. Toxic waste cleanups, air pollution regulation, and landfill monitoring are examples of steps taken by government authorities to protect the public. When the government deals directly with hazards, the principal goal of communication is typically to explain what it is doing or proposes to do. Suggestions may be solicited from the community, but more often the communicator seeks nothing from citizens but acquiescence.

For hazards requiring individual action, on the other hand, the goal of communication is to help people respond in appropriate ways. Hazards of this type differ in the degree of government involvement in the decision process. For example, individuals facing risks are sometimes given instructions as to how they should act, instructions based on decisions made by government authorities. When risk reduction requires the individual implementation of such institutional decisions – as when civil defense officials order the evacuation of homes in the path of a hurricane – communication efforts aim to encourage citizen compliance.

Other hazards are less amenable to institutional decision making, or even to standard setting or blanket recommendations. In situations characterized by strong individual differences in risk factors or risk preferences, individual citizens may be left to draw their own conclusions about both the degree of personal risk and the appropriate precautionary behaviors. Asbestos, urea-formaldehyde foam, lead from water pipes, and geological radon are all examples of hazards in the home about which the government does not tell people exactly what they should do. In each case, it is private citizens who are responsible for determining to what extent the hazard is present in their homes, for deciding whether it poses a substantial risk, and for choosing and carrying out an appropriate course of action.

Yet even with respect to this latter class of hazards, there is a role for government – a communication role. Two basic communication strategies are available: (1) Government can simply provide information to interested individuals to help them make and implement their own decisions; or (2) government can provide guidance or recommendations along with information and take other steps to promote specific responses.

Of course, there are no clear boundaries between situations in which the government attempts to compel citizen action and situations in which it relies on voluntary action. When recommendations for voluntary hurricane evacuation are ignored, for example, residents may be ordered to leave their homes. Similarly, states can set standards for indoor radon levels and require home builders, home sellers, or even homeowners to prove that the standards are met. But at least with respect to existing housing not involved in real-estate transactions, the states have so far chosen to rely on voluntary action, putting themselves in the role of communicator rather than regulator. Environmental agencies have much less experience with the former role and may be less adept at playing it.

Government communicators see themselves as responsible for providing people at risk with the information they need to cope effectively with the problem. But is information alone sufficient? If people use the information to arrive at accurate assessments of their risk, to make appropriate decisions about the need for remedial action, and, when necessary, to choose suitable remedial measures, an information-only strategy is enough. However, this strategy may fail for a variety of reasons. People may not seek out or attend to information; they may misunderstand the information; or they may prove unable to use the information to arrive at a decision. Furthermore, if the course of action they select is at all complex, they may need additional assistance in carrying out their plans.

When information programs are unsuccessful, communicators may have to be more aggressive in distributing information, may need to examine the comprehensibility of the materials provided, and may have to refocus communication efforts on issues people seem to misunderstand (such as who is and who is not at risk). Nevertheless, experience with a wide range of hazards has shown that even well-designed information and education programs are seldom adequate to bring about appropriate protective behavior (Weinstein 1987). Communicators who wish to promote self-protective action may have to consider approaches that are not primarily informational and use the power of interpersonal influence, incentives, vivid hazard images, repeated reminders, and the like. They may have to develop messages offering explicit recommendations or warnings to convince those who are at risk that they should act, and they may need parallel messages to reassure those whose risks are minimal.

In the remainder of this article, we assess the efficacy of an information program in promoting appropriate individual responses to a specific hazard – geological radon. We focus on a target audience of special interest: citizens who have voluntarily monitored their homes for radon and have found nontrivial levels, and who now face the decision of what if anything to do to reduce the problem. The data presented concern the accuracy of judgments such people make about the radon in their homes, and provide insight into the problems they encounter in their decision-making efforts. Finally, we raise the question of whether the needs of individuals exposed to the threat of radon were met by the information strategy employed.

Radon Pollution in New Jersey

Geological radon is a radioactive gas produced by the decay of naturally occurring uranium. This gas seeps into the basements of houses built over uranium-bearing soil or rock and becomes trapped. If homes have elevated levels of radon, radioactive dust particles may become lodged in residents’ lungs, increasing the risk of lung cancer.

Radon is a potentially serious problem for residents in many areas of the country; it is estimated that in the United States between 5,000 and 20,000 lung cancers each year are the result of radon exposure (A Citizen’s Guide to Radon 1986). Since about two million people in this country die each year, these figures suggest that between one in 400 and one in 100 deaths are caused by radon.

In northern New Jersey, experts estimate, roughly 30% of all houses have radon levels that exceed the action guideline of 4 picocuries of radiation per liter of air (pCi/l) currently accepted by the U.S. Environmental Protection Agency (EPA), the New Jersey Department of Environmental Protection (DEP), and the New Jersey Department of Health (DOH). Since 1984, when radon was first recognized as a serious home pollution problem in the United States, government agencies have been faced with the task of ensuring that residents have the information they need to cope with this hazard.

It is tempting to suppose that the only real communication task is to encourage people to test their homes to find out if they are at risk. Since a good deal is known about radon risks (more than is known about most environmental carcinogens), one might expect that once individuals have monitored, their decisions should follow logically from their test results. Individuals with negligible radon concentrations, for example, should not waste time and money trying to achieve still lower levels. Those with higher levels should weigh their risk against the cost and feasibility of mitigation in deciding whether to take action. It is true that encouraging citizens to monitor their homes is itself a formidable communication task. However, the results presented in the remainder of this article indicate that communicators cannot count on appropriate responses following radon monitoring.

Officials in New Jersey, while desiring to reduce the risk faced by state residents, were also concerned about avoiding panic. Past research indicates that panic is not a common response to hazards (Drabek 1986) and is usually found only in situations in which time is thought to he running out (Lazarus 1966). But panic seemed possible in this case because radon involved high risk levels, radiation, cancer, and a difficult remediation problem – hazard characteristics that have produced fear among members of the public in the past.

In an effort to forestall panic while still educating residents, DEP (the state agency given principal responsibility for dealing with radon in New Jersey) adopted a strategy characterized by passive diffusion of information. A radon hot line was set up to answer questions residents might have about radon pollution and radon testing. When requested by local authorities, DEP and DOH representatives spoke about radon at public meetings, but no attempt was made to distribute information directly to residents in areas of the state where radon was thought to be a problem. (Note: 2) Copies of EPA radon information sheets were also available to hot-line callers on request. These February 1985 fact sheets provided an explanation of the radon problem, but did not contain information about the risk at different radon levels or about the actions that should be taken by homeowners, nor did they mention the suggested action level of 4 pCi/l. (Note: 3) Another fact sheet prepared by the Pennsylvania Department of Health (1985) contained mortality probabilities for different radon levels and different lengths of exposure. It was available from DEP and from some commercial testing companies and was also distributed in large numbers to homeowners in high-risk areas of New Jersey by an elected official.

For this communication strategy to prove successful, concerned individuals would need to seek out information and share what they learned with friends and neighbors. These steps would lead to the eventual dissemination of information throughout the population at risk, and to widespread individual decisions to monitor and to take remedial action, where appropriate.

As an additional service to New Jersey residents, a program was established for those who had tested their homes and had found radon levels in excess of 4 pCi/l. If they wished to confirm their initial test results, these individuals could obtain a second measurement from the state at no charge. Participants in this program were informed that if test results were above 4 pCi/l in any frequently used portion of the home, DEP recommended they take action.

Sample and Method

In July 1986 we conducted a survey to assess the responses of homeowners who had monitored their homes for radon, found levels in excess of 4 pCi/l, and participated in the DEP confirmatory monitoring program. (Note: 4) This was part of a research program to guide DEP radon communication efforts. Other parts included a general survey of the population at risk and a comparison of the effectiveness of different ways of explaining radon risk levels. If we found evidence that even those who had monitored and found a problem still failed to understand the nature of the radon risk or were responding inappropriately, the results would suggest strongly

The 359 individuals who had requested confirmatory monitoring of their radon levels from the DEP as of June 1986 were invited to participate in the survey. Because the identities of residents in the monitoring program were confidential, we could not recruit them directly. Instead, a letter describing the project was mailed by DEP to all program participants, along with a postcard they could return to us indicating their willingness to take part. This method of recruitment, although necessary for ethical reasons, resulted in a relatively low response rate; 47.3% returned postcards. Had we been able to contact potential participants directly or to send them reminders, the return rate probably would have been higher.

Those who responded to our request were very well educated (64% had completed college), but no different from the randomly selected sample of homeowners who participated in a general survey of attitudes toward radon and radon testing (Weinstein, Sandman, and Klotz 1987). However, respondents to this survey were clearly different from other New Jersey homeowners in three important ways: (1) They voluntarily monitored their homes for radon at a time when most of the population at risk had not even considered monitoring; (2) when they found levels in excess of 4 pCi/l, they sought confirmation from DEP; and (3) when asked to participate in the survey, they gave their consent. These differences suggest that, compared to the average New Jersey homeowner, survey respondents were probably more aware of radon and more concerned about its risks, more interested in seeking additional information about radon and their own risk levels, and more trusting of the ability of government and academe to provide appropriate help while maintaining confidentiality.

Although the participants in the confirmatory monitoring program were not a representative sample of the population at risk, they were an important group to study for two reasons. First, because members of this sample were more knowledgeable about radon than the remaining population and because they had demonstrated their willingness to deal with the radon issue by getting initial and follow-up measurements, they can he viewed as a best-case indicator of the effectiveness of communications available at that time. If members of this group were misinformed or had difficulty responding in a manner appropriate for their radon levels, the adequacy of the available information for the rest of the public would be doubtful. Second. these were people who had found appreciable radon levels in their homes, so the questions of how they would evaluate the seriousness of their risk and what actions they would decide to take were highly relevant.

Questionnaires were mailed to the 170 eligible volunteers, and 141 returned completed forms. We also conducted lengthy face-to-face interviews in the homes of 16 survey participants between October and December 1986. Although the number of these interviews is small, they proved enlightening, and will be discussed where appropriate.

Correlates of Radon Levels

A major goal of radon communication efforts is for individuals to understand the risk they face, so that those with low levels will relax and those with high levels will remediate. Consequently, an indicator of the effectiveness of radon communication is the degree of association between the objective risk – indicated by home radon levels – and mitigation actions. Less direct indicators of the appropriateness of individuals’ responses were also used, including their intentions regarding future remediation, their evaluations of the risk posed by their radon levels, and the amount of radon-related distress they experienced.

Strong, positive correlations between radon levels and these response measures would indicate that communications were accurately explaining the magnitude of the risk and that people were able to use this risk information in judging their need for action. As the results presented in Table 1 show, however, strong correlations were not found. Correlations between radon levels (self-reported for basement and first floor) and responses to a variety of survey questions were sometimes statistically significant, but were modest in size. (Note: 5) Correlations with a crude, dichotomous measure of risk – whether or not a reading was above the 4 pCi/l guidance level – were larger than with the actual radon test result. The remainder of this article explores some of the factors that diminish the size of these correlations and the implications of these findings for communication.

Table 1. Correlations Between Radon Levels and Selected Variables

General Knowledge about Radon

Survey participants had acquired a good deal of information about radon. On a quiz covering 17 “facts,” the average score was 85% correct. Incorrect answers were infrequent; instead, respondents were more likely to acknowledge, “I don’t know.” People were especially well informed about the physical characteristics of radon (e.g., that it is a colorless, odorless gas). They were less familiar with its health effects (e.g., that it does not cause skin cancer).

The high scores on the radon quiz are not surprising, given that the sample was restricted to people who had both monitored their homes and contacted the state for a confirmatory test. Indeed, members of this sample were unusual in their fervor for radon information. Most of the people we interviewed maintained a “radon file” containing newspaper clippings, correspondence, and other radon-related material they had obtained. Nevertheless, the radon quiz results do show that factual information was available to people who sought it out. Other research indicates that the general population in New Jersey was also absorbing this information, although at a somewhat slower pace (Weinstein, Sandman, and Klotz 1987).

The acquisition of general information, however, did not improve decision making. Multiple regression equations containing knowledge (as measured by the radon quiz), radon level, and the interaction of these two variables were no better at predicting actions, plans, and risk perceptions than equations containing the radon level alone. This finding is consistent with previous research in this area (Note: 6) and argues against the strategy of relying on general education to move people in the desired direction.

The personal interviews shed some light on the lack of association between knowledge and reactions to radon. Many respondents we interviewed expressed frustration with their search for information, especially for information that would help them evaluate their levels and make decisions about remediation. There are several possible reasons for their dissatisfaction. First, at the time of this survey the radon problem was still so new that detailed information was not generally available, especially information regarding remediation procedures and companies able to carry out these procedures. Second, many people lacked the background needed to make use of the rather technical information that was available, a situation that can lead people to misunderstand or even to disregard information (Friedman 1981; Hanley 1980). For example, the units in which radon levels are usually reported, picocuries of radiation per liter of air, were unfamiliar to residents. Many of the people we interviewed found their test results virtually meaningless. Furthermore, explanations of the risk associated with various levels were often presented in terms of probabilities, which many people find hard to understand (Tversky and Kahneman 1974).

Finally, some people felt that the available information was inadequate because it offered too little guidance. The desire for help in choosing remediation strategies was especially strong. Many of the people we interviewed wanted a prescribed course of action, not an explanation of the various options available to them. It may be that the technical nature of the problem, the uncertainty associated with both risk assessment and remediation, and the high stakes involved (in terms of both money and health) combined to make the situation too taxing for the average homeowner to deal with without expert help.

Negative Emotions

Respondents were quite concerned about, but not distraught over, the level of radon in their homes. Six 5-point scales asked about the amount of concern, worry, fear, depression, helplessness, and anger experienced (e.g., 1 = not at all concerned, 2 = slightly concerned, 3 = moderately concerned, 4 = very concerned, 5 = extremely concerned). Respondents reported a higher average rate of concern (M = 3.6) than of any other emotional reaction (worried = 2.7; frightened = 2.2; helpless = 2.1; depressed 1.9; angry = 1.8). When ratings of these six adjectives were summed to yield a single index of distress, its correlation with home radon levels was not significant (see Table 1). Several other variables did prove to be associated with distress: concern about health effects appeared to be the major force behind negative emotional reactions. Respondents who reported greater distress tended to feel that their radon levels would produce illness (r = .56, p < .001); that the risk of illness would remain elevated even if levels were lowered (r = .43, p < .00l); and that their radon levels posed a serious problem (r = .48, p < .001).

Although the average level of distress was relatively low, we should not ignore the fact that some people in our sample were quite upset – 35% rated themselves as at least moderately frightened, and 53% were at least moderately worried. Furthermore, the nonsignificant correlations between distress and radon levels indicate that people with low levels were as likely to become upset as those with high levels.

The interviews we conducted suggest an additional caveat. Several individuals indicated that they had forgotten about radon over the summer months when the house was open, but were experiencing increased anxiety with the onset of winter, when windows are kept closed and radon levels are higher. One person remarked that he became alarmed each time he heard his forced-air heating system come on (bringing contaminated air up from the basement). Since the survey was conducted in July, ratings of negative affect might have been attenuated temporarily by the decreased salience of the radon problem.

Accuracy of Risk Perceptions

Although they did not deny the presence of radon in their homes – only 10.4% of the sample said their levels were below or much below average – respondents were not very accurate in their assessment of the risk involved. The correlation between the perceived seriousness of their radon problem and the basement radon level, although significant, was only moderate in size, and the correlation with the first-floor radon level (which is a better indicator of the actual risk) was not significantly different from zero. Associations between radon levels and other variables that indicate the perceived seriousness of the radon threat, such as beliefs about the likelihood of effects on health and the size of effects on property values, were modest in size (see Table 1).

Clearly, the belief that radon poses a serious home problem did not accurately reflect actual radon test results. In a similar study conducted in Maine, Johnson and Luken (1987) found no relationship between perceived and actual risk.

Discrepancies between actual and perceived risk were not totally random; people were more likely to underestimate than overestimate the hazard (Weinstein, Klotz, and Sandman 1988). In fact, responses to several questions in this survey indicate an optimistic bias, a common finding in research on risk perception. (Note: 7) For example, respondents were asked how serious it would be if someone in their families became ill because of radon. Only 53% of the sample said that it would be serious or very serious. Even among respondents who were well informed about the health risks from radon (those who knew that it can cause lung cancer and does not cause skin cancer or eye and throat irritation), only 55% were willing to acknowledge that radon-caused illness would be serious or very serious for a family member.

Participants were also asked to compare the risk from their radon levels with the risk from other common health hazards. They judged whether the risk from their own level of radon was less than, the same as, or greater than the risks from smoking, not wearing a seat belt, eating a high-cholesterol diet, exposure to hazardous waste, and being 20 pounds overweight. Very few respondents perceived their radon concentrations to be more risky than these other hazards. For example, only 13% said their radon level was more risky than smoking a pack of cigarettes a day, although 35% of this sample had first-floor radon levels that posed a greater lung cancer risk (based on EPA figures). The actual risk from living near a hazardous waste site appears to be quite small in most cases, but only 4% of the sample believed their radon levels were a greater risk. Furthermore, as radon levels go up, survey participants should more often judge that their risk is greater than that presented by the comparison risk, but all the correlations between radon levels and these comparative risk judgments were nonsignificant (see Table 1).

Several characteristics of geological radon may predispose people to underestimate the degree of risk it presents (Sandman, Weinstein, and Klotz 1987). First, people tend to underestimate the seriousness of natural hazards, especially when there is no “villain” to blame (Burton and Kates 1964; Burton, Kates, and White 1978). Second, the lack of sensory cues indicating that a hazard is present will lower the perception of risk (Mileti and Sorenson 1987). Radon is an invisible threat, and it produces no early warning signs of illness. Finally, it is impossible to identify specific instances of lung cancer that have been caused by radon.

Communicators could adopt several strategies to minimize excessive optimism. They could (1) emphasize the magnitude of the threat posed by radon, perhaps comparing it to more familiar risks; (2) stress that the lung cancer resulting from radon exposure would produce no early warning symptoms; (3) emphasize that even though radon is not visible it has a cumulative effect; and (4) provide evaluative information – for example, label radon concentrations as “low threat,” “moderate threat,” or “high threat” – rather than giving only quantitative risk information.

Of course, not all people in our sample were optimistic. Some respondents appeared excessively concerned about their low or moderate levels. Although this type of inaccuracy does not pose the same health risk as underestimation, anxiety is itself a health problem, and those who overestimate the threat may waste both time and money trying to reduce their radon risk to zero. They should not be forgotten in designing risk communications.

Behavioral Responses

Only 19% of those surveyed had radon-mitigation projects completed or in progress. Another 43% said they planned to take action. Still, 27% were undecided as to whether remediation was necessary, and 11% had decided not to act. While remedial action was not necessary for everyone in the sample, 74% of the sample had basement levels at or above the EPA’s recommended action level, and 45% had first-floor radon levels that high.

Remediation actions that were completed or in progress and plans concerning remediation were only weakly related to actual radon levels (people with higher levels being slightly more likely to have acted or to say they intended to take action). One might attempt to explain such modest correlations by referring to other variables that should affect mitigation decisions – but likely candidates, such as the respondents’ incomes and their beliefs about the effectiveness of radon-reduction methods, yielded nonsignificant correlations. (Note: 8)

People who had not yet taken action were asked to give reasons for their delay. Of the small number who had decided that home modifications were unnecessary, the majority (80%) cited low radon levels. Uncertainty about the best mitigation method was the most common reason given by those who were undecided or still in the planning stage (49% and 53%, respectively), followed by lack of time (40% and 32%). These results suggest that when more is known about remediation techniques, more people will take action. The interviews reinforce this conclusion. Most people visited expressed a desire to do something, but few had a clear idea of what to do. Those individuals who lacked “handyman” skills were especially confused about the course of action they should take.

Dichotomizing The Risk Continuum

Interestingly, a very crude measure of risk status – being above or below the 4 pCi/l guidance level – was just as strongly correlated with response variables as was the actual degree of risk. In fact, as seen in Table 1, the correlation with this dichotomous risk measure was often considerably stronger than the correlation with the actual level, and multiple regression analyses showed that in no case did the radon level add any explanatory power to predictions of risk perceptions, actions, and plans beyond that provided by this simple, two-level variable. (Note: 9) People with radon readings ten times the guidance level were no more likely to plan action than those just above 4 pCi/l. Although there was a wide range of responses both above and below 4 pCi/l, this guidance level did influence homeowners’ perceptions and plans. However, available materials clearly failed to communicate an appropriate sense of the degree of risk.

Additional Information Concerns

The face-to-face interviews allow us to examine three other questions regarding communication of information about radon.

Summary

The results of this study show that communicators had achieved only limited success in their efforts to meet the needs of survey respondents. Our sample was well informed in terms of general knowledge about radon, but such knowledge did not lead to appropriate reactions. There were modest relationships between objective risk levels and both risk perceptions and remediation plans, but these were entirely due to the distinction of being above or below the guidance level of 4 pCi/l. People much above this level did not react differently than those only slightly above it. Underestimation of risk appeared to be a frequent problem; overestimation of risk and distress at low radon exposures was less common but nonetheless present for some respondents.

The people we surveyed seemed to be taking the issue of radon seriously. However, they had been hampered by a lack of information and by difficulties in using the available information to formulate a course of action. Pamphlets published by EPA since these data were gathered provide more information than the materials that were available to our respondents. The new pamphlets give more advice about the seriousness of the risk and the need for action, and they give explicit recommendations for people with different radon levels, avoiding the suggestion that 4 pCi/l is a magic dividing line. It remains to be seen whether such materials will provide sufficient guidance to meet the needs of New Jersey homeowners.

Communication with peers through informal social comparison has been inadequate, leaving homeowners unsure of whether their responses (both behavioral and affective) are “normal.” A tendency toward secrecy about radon problems is likely to limit social comparison in the future, contributing to an underestimation of the extent of the problem and slowing the general diffusion of information about radon. In view of the limited availability of peer interactions and the inability of government agencies to provide individual mitigation advice, nonstandard communication approaches, such as community self-help groups, may need to he considered.

The same considerations must be addressed for other environmental hazards that require individuals to take action. We suspect that the problems to be faced in communicating about similar hazards – lead, polluted home wells, and so on – will not be much different from those we have uncovered here. In that case, campaigns that aim at “educating the public” and that are reluctant to provide explicit recommendations may have relatively little impact on the majority of the population, and those who do act may not be those exposed to the greatest risk.

The results of this study suggest that environmental hazards calling for individual remediation will require active government promotion of appropriate protective behaviors. The decisions individuals must make in coping with these hazards are difficult. Merely providing them with the facts will not be sufficient to ensure decisions consistent with the actual level of risk. It is important that communicators become aware of the special needs of citizens faced with such individual hazards and that they search for communication strategies that will provide sufficient guidance to meet these needs.

Notes

1. See, for example, Edelstein and Wandersman (1987); Johnson and Tverskv (1983); Kasperson and Pijawka (1985); Kates (1985); Slovic, Fischhoff, and Lichtenstein (1985); Whyte and Burton (1988).

2. A mass mailing to all homeowners in the areas of greatest risk was undertaken in November 1988, two years alter the research reported here.

3. An updated information booklet, A Citizen’s Guide to Radon, became available in August 1986. This booklet contains a wide range of risk information, including estimates of the lung cancer risk posed by different levels of radon (presented both graphically and as probabilities), comparisons to other risks (e.g., smoking), and general suggestions for mitigation at various levels of exposure. Also available in August was a booklet describing basic remediation techniques.

4. Additional information about this study is contained in Weinstein, Sandman, and Klotz (1987).

5. The questions concerning risk, emotion, and plans in Table 1 refer to the radon level at the time the individual answered these questions (e.g., “I feel that the radon problem in my home is…” ). To see whether these responses are strongly associated with the objective risk, one should examine their correlations with the basement and first-floor radon levels present at that time. These are the same as the radon levels reported by the confirmatory monitoring program, but only if respondents have not made any home changes subsequent to the DEP test. Therefore, the correlations in Table 1 are based only on those individuals who had not yet acted (except for the variable “remediation completed or in progress” which refers to the full sample) Furthermore, nearly all confirmatory test results fell between 0 and 60 pCi/l. A few homeowners with unusually high values (one case with a first-floor level of 80 and five cases with basement levels between 96 and 174) were excluded from the calculations because they would have unduly influenced results.

6. See, for example, Burton, Kates, and White (1978); Mileti (1975); Sorensen (1983).

7. See, for example, Mileti and Sorensen (1987); Weinstein (1980, 1988).

8. Similar results were reported in Johnson and Luken (1987).

9. In a 1988 study, 123 of the respondents in the survey reported here were recontacted (Weinstein, Sandman, and Roberts, 1988). The follow-up focused on the success of remediation efforts, but two of its findings are relevant here. First, most respondents did undertake some action in the intervening two years. Second, those with higher radon readings were more likely to act. The biggest difference as before was between those with readings above and below 4 pci/l. Still, the amount of radon above 4 pCi/l did make a difference. For example, the dichotomous basement level (above or below 4 pCi/l) explained 9.8% of the variance in mitigation action, but adding the actual basement level significantly (p < .02) increased the total to 17.0%. The first-floor dichotomous level variable accounted for 9.2% of the variance in mitigation action; adding the actual level increased this figure (p < . I) to 11.8%. Many factors may account for why remedial action by 1988 was more frequent and the relationship with exposure level was stronger than one might have expected from the 1986 data: more explicit advocacy of remedial action above the 4 pCi/l level by government agencies and literature, more information available on remediation strategies, more contractors who were able to perform remediation work, increased attention to radon risks, publication of the EPA’s A Citizen's Guide to Radon, and increasing numbers of testers to provide social comparison information and to make remedial action appear more consonant with neighborhood norms.

10. Edelstein and Wandersman (1987) also found that people in homes contaminated with toxic chemicals felt ostracized.

References

Burton, Ian, and Robert Kates. 1964. The perception of natural hazards in resource management. Natural Resources, 3(l):412–41.

Burton, Ian, Robert Kates, and Gilbert White. 1978. The environment as hazard. New York: Oxford Univ. Press.

A citizen's guide to radon: What it is and what to do about it, 1986. U.S. Environmental Protection Agency and U.S. Centers for Disease Control Pamphlet. OPA–86–004.

Drabek, Thomas E. 1986. Human system response to disaster. New York: Springer–Verlag.

Edelstein, Michael R., and Abraham Wandersman. 1987. Community dynamics in coping with toxic contaminants. In Neighborhood and Community Environment, ed. I. Altman and A. Wandersman, 69–110. New York: Plenum.

Friedman, Sharon M. 1981. Blueprint for breakdown: Three Mile Island and the media before the accident. Journal of Communication 31 (Spring) 116–29.

Goethals, 0. R., and John Darley. 1977. Social comparison theory: An attributional approach. In Social comparison processes: Theoretical and empirical perspectives, ed. J. Suls and R. Miller. Washington, DC: Hemisphere.

Hanley, J. 1980. The silence of scientists. Chemical and Engineering News 58(12):5.

Johnson, Eric, and Amos Tversky. 1983. Affect, generalization, and the perception of risk. Journal of Personality and Social Psychology 45(1):20–31.

Johnson, F. Reed, and Ralph A. Luken. 1987. Radon risk information and voluntary protection: Evidence from a natural experiment. Risk Analysis 7(l):97–l07.

Kasperson, Roger E. 1977. Societal management of technological hazards. In Managing technological hazard: Research needs and opportunities, ed. R. W. Kates, 49–80. Boulder: Institute of Behavioral Science, University of Colorado.

______ 1986. Six propositions for public participation and their relevance for risk communication. Risk Analysis 6(3):275–81.

Kasperson, Roger E., and K. David Pijawka. 1985. Societal response to hazards and major health events: Comparing natural and technological hazards. Public Administration Review 45(1):7–18.

Kates, Robert W. 1985. Success, strain, and surprise. Issues in Science and Technology 2(1):46–58.

Lazarus, Richard. 1966. Psychological stress and the coping process. New York: McGraw-Hill.

Mileti, Dennis S. 1975. Natural hazard warning systems in the United States. Boulder: Institute of Behavioral Science, University of Colorado.

Mileti, Dennis, and John Sorensen. 1987. Natural hazards and precautionary behavior. In Taking care: Understanding and encouraging self-protective behavior, ed. Neil D. Weinstein, 189–207. New York: Cambridge Univ. Press.

Pennsylvania Department of Health. 1985. Health fact sheet: Radon gas exposure. Harrisburg: PDH.

Sandman, Peter M., Neil D. Weinstein, and M. L. Klotz. 1987. Public response to the risk from geological radon. Journal of Communication 37(3): 93–108.

Slovic, Paul, Baruch Fischhoff, and Sarah Lichtenstein. 1985, Characterizing perceived risk. In Perilous progress: Managing the hazards of technology, ed. R. W. Kates, C. Hohenemser, and J. Kasperson 91–125 Boulder, CO: Westview.

Sorensen, John II. 1983. Knowing how to behave under the threat of disaster: Can it be explained? Environment and Behavior 14(4):438–57.

Tversky, Amos, and Daniel Kahneman. 1974. Judgment under uncertainty: Heuristics and biases. Science 185:1124–31.

Unger, D., and Abraham Wandersman. 1985. The importance of neighboring: The social, cognitive and affective cornponents of neighboring. American Journal of Community Psychology 13:139–69.

Weinstein, Neil D. 1980. Unrealistic optimism about future life events. Journal of Personality and Social Psychology 39:806–20.

______ 1987. Taking care: Understanding and encouraging self-protective behavior. New York: Cambridge Univ. Press.

______ 1988. Unrealistic optimism about illness susceptibility: Conclusions from a community wide sample. Journal of Behavioral Medicine 10:481–500.

Weinstein, Neil D., Mary Lou Klotz, and Peter M. Sandman. 1988. Optimistic biases in public perceptions of the risk from radon. American Journal of Public Health 78:796–800.

Weinstein, Neil D., Peter M. Sandman, and M. L. Klotz. 1987. Public response to the risk from radon, 1986. Report to the Division of Environmental Quality, New Jersey Department of Environmental Protection, Trenton.

Weinstein, Neil D., Peter M. Sandman, and Nancy E. Roberts. 1988. Homeowner radon mitigation. Report to the Division of Environmental Quality, New Jersey Department of Environmental Protection, Trenton.

Whyte, Anne, and Ian Burton, eds. 1980, Environmental risk assessment. New York: John Wiley.

Authors’ Note: The research discussed in this article was conducted under Contract C29543 from the Division of Environmental Quality, New Jersey Department of Environmental Protection. The interpretations and opinions expressed in this article are those of the authors alone.

Copyright © 1989 by Science, Technology, & Human Values

For more on precaution advocacy: