Probability Information in Risk Communication: A Review of the Research Literature

Communicating probability information about risks to the public is more difficult than might be expected. Many studies have examined this subject, so that their resulting recommendations are scattered over various publications, diverse research fields, and are about different presentation formats. An integration of empirical findings in one review would be useful therefore to describe the evidence base for communication about probability information and to present the recommendations that can be made so far. We categorized the studies in the following presentation formats: frequencies, percentages, base rates and proportions, absolute and relative risk reduction, cumulative probabilities, verbal probability information, numerical versus verbal probability information, graphs, and risk ladders. We suggest several recommendations for these formats. Based on the results of our review, we show that the effects of presentation format depend not only on the type of format, but also on the context in which the format is used. We therefore argue that the presentation format has the strongest effect when the receiver processes probability information heuristically instead of systematically. We conclude that future research and risk communication practitioners should not only concentrate on the presentation format of the probability information but also on the situation in which this message is presented, as this may predict how people process the information and how this may influence their interpretation of the risk.     

Varying Risk Comparison Elements: Effects on Public Reactions

Expanding a limited empirical base on effects of risk comparisons, a pilot experiment explored how varying elements of such comparisons might influence public response. The scenario used was a hypothetical trial of an asbestos-installing firm for putting students and staff at a junior high school at "unreasonable risk," first used by Slovic et al. (1990). Study participants played the role of jurors in the trial, asked to rate the risk, the firm's guilt, whether it should be made liable for future health effects, whether the school should be occupied in its current condition, and whether the asbestos should be removed at a cost of 3 million dollars. Opportunity samples of New Jersey residents (n= 309) received information intended to vary four comparison attributes: number of dimensions of comparison; single versus multiple ways of expressing mortality; a narrative to explain and justify the risk comparison; and the hypothetical role of the participant (juror vs. parent). The dependent variable was a hazard scale constructed from four of the five postcomparison judgments. ANOVA found format variations swamped in their effect by concern about asbestos, with the absence of a narrative and the parental role being the only attributes that increased negative risk reactions. Multiple regression analyses found that multidimensionality, narrative, role, and the presence of any risk comparison at all had significant effects when these (plus death formats) were the only independent variables. However, only multidimensionality (which increased negative reactions) retained significance when asbestos concern, risk beliefs (no safe level of exposure to a carcinogen; any exposure leads to cancer), and demographic variables were added to the analysis. Concern and risk beliefs alone explained 33% of variance in hazard scores; adding demographics and risk comparison variables only raised explained variance to 36% (having children at home and being exposed to multidimensional risk comparisons raised scores; age and income lowered them). The results underline the potentially small effect of risk comparison information on risk views (at least between subjects), but offer some insight into aspects of message design left unexplored in empirical literature to date.     

Incorporating Message Framing into Narrative Persuasion to Curb E-Cigarette Use Among College Students

This study examines the interaction effect of message format (narrative vs. nonnarrative) and message framing (gain vs. loss) in e-cigarette prevention targeting young adults. Results of a two-way experiment (N = 439) revealed that transportation and discrete emotions mediated message effect on risk perception and behavioral intention. Compared to the gain-framed nonnarrative, the gain-framed narrative reduced feelings of guilt, and guilt was negatively related to risk perception and positively related to behavioral intention. Thus, the gain-framed narrative achieved desirable persuasive outcome through guilt—increasing risk perception and decreasing intention to use e-cigarette. Similarly, the loss-framed narrative evoked greater sadness, which also led to increased risk perception and decreased behavioral intention. Transportation and discrete emotions mediated message effect in a serial order. This research not only contributes to the literature on narrative persuasion and emotion, but also provides insight for health communication designed for e-cigarette prevention.     

Some Criteria for Evaluating Risk Messages

Seven criteria are presented for use in evaluating communications designed to explain the magnitude of a risk. The criteria are: (1) comprehension (Does the audience understand the content of the communication?); (2) agreement (Does the audience agree with the recommendation or interpretation contained in the message?); (3) dose-response consistency (Do people facing a higher dose of a hazard perceive the risk as greater and/or show a greater readiness to take action than people exposed to a lower dose of this hazard?); (4) hazard-response consistency (Do people facing a hazard that is higher in risk perceive the risk as greater and/or show a greater readiness to take action than people exposed to a hazard that is lower in risk?); (5) uniformity (Do audience members exposed to the same level of risk tend to have the same responses to this risk?); (6) audience evaluation (Does the audience judge the message to have been helpful, accurate, clear, etc.?); and (7) types of communication failures (When different types of failures are possible, are the failures that occur generally of the more acceptable variety?). Each of these criteria is illustrated with data collected in a test of message formats designed to explain the risk presented by radon gas in a home.     

Risk Perception and Risk Talk: The Case of the Fukushima Daiichi Nuclear Radiation Risk

Individuals’ perceptions and their interpersonal communication about a risk event, or risk talk, can play a significant role in the formation of societal responses to the risk event. As they formulate their risk opinions and speak to others, risk information can circulate through their social networks and contribute to the construction of their risk information environment. In the present study, Japanese citizens’ risk perception and risk talk were examined in the context of the Fukushima Daiichi nuclear radiation risk. We hypothesized and found that the risk information environment and risk literacy (i.e., competencies to understand and use risk information) interact to influence their risk perception and risk talk. In particular, risk literacy tends to stabilize people's risk perceptions and their risk communications. Nevertheless, there were some subtle differences between risk perception and communication, suggesting the importance of further examination of interpersonal risk communication and its role in the societal responses to risk events.     

Communicating Risks and Benefits from Fish Consumption: Impact on Belgian Consumers' Perception and Intention to Eat Fish

Communicating about the health effects of fish and seafood may potentially result in a conflict situation: increasing intake is desirable because of health and nutritional benefits, but higher consumption may also lead to an increased intake of potentially harmful environmental contaminants. In order to anticipate the communication challenge this conflict may pose, the research presented here aimed to assess the impact of risk/benefit communication on Belgian consumers' fish consumption behavior and fish attribute perception. Data were collected in June 2005 from a sample of 381 women, aged between 20 and 50 years. An experimental design consisting of four message conditions (benefit‐only; risk‐only; benefit‐risk; and risk‐benefit) combined with three information sources (fish and food industry; consumer organization; government) was used. Exposure to the benefit‐only message resulted in an increase from a self‐reported fish consumption frequency of 4.2 times per month to an intended fish consumption frequency of 5.1 times per month (+21%), while fish attribute perceptions only marginally improved. The risk‐only message resulted in a strong negative perceptual change in the range of two points on a seven‐point scale. This translated into an 8% decrease of behavioral intention (from eating fish 4.5 times per month to an intention of eating fish 4.1 times per month). Balanced messages referring to both risks and benefits yielded no significant change in behavioral intention, despite a significant worsening of fish attribute perception. The presentation order of benefits and risks in the balanced message showed a tendency to affect both behavioral intention and attribute perception, with the first message component being most influential. Information source did not yield any significant impact either on behavioral intention or on attribute perceptions, independent of the message content. The results from this study provide valuable insights for future risk/benefit and balanced communication about seafood.     

Does Iconicity in Pictographs Matter? The Influence of Iconicity and Numeracy on Information Processing,Decision Making,and Liking in an Eye‐Tracking Study

Researchers recommend the use of pictographs in medical risk communication to improve people's risk comprehension and decision making. However, it is not yet clear whether the iconicity used in pictographs to convey risk information influences individuals’ information processing and comprehension. In an eye‐tracking experiment with participants from the general population (N = 188), we examined whether specific types of pictograph icons influence the processing strategy viewers use to extract numerical information. In addition, we examined the effect of iconicity and numeracy on probability estimation, recall, and icon liking. This experiment used a 2 (iconicity: blocks vs. restroom icons) × 2 (scenario: medical vs. nonmedical) between‐subject design. Numeracy had a significant effect on information processing strategy, but we found no effect of iconicity or scenario. Results indicated that both icon types enabled high and low numerates to use their default way of processing and extracting the gist of the message from the pictorial risk communication format: high numerates counted icons, whereas low numerates used large‐area processing. There was no effect of iconicity in the probability estimation. However, people who saw restroom icons had a higher probability of correctly recalling the exact risk level. Iconicity had no effect on icon liking. Although the effects are small, our findings suggest that person‐like restroom icons in pictographs seem to have some advantages for risk communication. Specifically, in nonpersonalized prevention brochures, person‐like restroom icons may maintain reader motivation for processing the risk information.     

Cancer Anxiety and the Perception of Risk in Alarmed Communities

Cancer is a significant public health topic and is frequently a factor in public reactions to environmental hazards. It may be reasonable to suggest that a unique form of health anxiety exists - one specific to cancer. In this article, we explore a measure of cancer anxiety that has applicability to risk perception in the specific context of communities that are alarmed over suspect cancer rates thought to be associated with environmental hazards. A literature search was used to identify survey questions previously used to measure cancer worry, fear, anxiety, etc. A list of 24 items was employed in a mail survey sent to 30 communities in which cancer rates were under investigation (part of a broader study). An analysis of the dataset of 1,111 responses yields two versions of a cancer anxiety scale: one an additive combination of 21 items (alpha= 0.77) and the other a two-factor model consisting of nine- and four-item subscales (alpha= 0.74 and 0.69). The resulting scales are evaluated for their ability to predict perception of risk from the environmental hazards in these cases. Controlling for age, sex, and cancer status, the scales explain between 2% and 10% of risk perception (full R(2) values ranging from 0.17 to 0.24). Given the range of concepts required to model risk perception, we conclude that this measure of cancer anxiety is sufficiently reliable and robust to be recommended for use in circumstances involving hazards associated with cancer. Other uses and further development of the measure are discussed.     

Designing Graphs to Communicate Risks: Understanding How the Choice of Graphical Format Influences Decision Making

Previous research suggests that the choice of graphical format for communicating risk information affects both understanding of the risk magnitude and the likelihood of acting to decrease risk. However, the mechanisms through which these effects work are poorly understood. To explore these mechanisms using a real‐world scenario, we examined the relative impact of two graphical displays for depicting the risk of exposure to unexploded ammunition during potential land redevelopment. One display depicted only the foreground information graphically (a bar graph of the number of people harmed), and a second depicted the foreground and background graphically (a stacked bar graph representing both the number harmed and at risk). We presented 296 participants with either the foreground‐only or the foreground and background graphical display and measured a broad set of outcome variables, examining (1) the graphical display effect on each of the outcome measures and (2) the pathways by which any display effects work to influence decision making. We found that the foreground‐only graphical display increased perceived likelihood and experienced fear, which produced greater worry, which in turn increased risk aversion. In addition, a positive evaluation of the communication materials increased support for policies related to land redevelopment, whether those policies were risk taking or risk mitigating. Finally, the foreground‐only graphical display decreased understanding of the risk magnitude, showing that approaches to accomplish one risk communication goal (promoting risk‐averse decisions) may do so at the expense of another goal (increasing understanding).     

Affective and Cognitive Factors Influencing Sensitivity to Probabilistic Information

In study 1 different groups of female students were randomly assigned to one of four probabilistic information formats. Five different levels of probability of a genetic disease in an unborn child were presented to participants (within‐subject factor). After the presentation of the probability level, participants were requested to indicate the acceptable level of pain they would tolerate to avoid the disease (in their unborn child), their subjective evaluation of the disease risk, and their subjective evaluation of being worried by this risk. The results of study 1 confirmed the hypothesis that an experience‐based probability format decreases the subjective sense of worry about the disease, thus, presumably, weakening the tendency to overrate the probability of rare events. Study 2 showed that for the emotionally laden stimuli, the experience‐based probability format resulted in higher sensitivity to probability variations than other formats of probabilistic information. These advantages of the experience‐based probability format are interpreted in terms of two systems of information processing: the rational deliberative versus the affective experiential and the principle of stimulus‐response compatibility.     

Designing Risk Communications: Completing and Correcting Mental Models of Hazardous Processes, Part I

Many risk communications are intended to help the lay public make complex decisions about risk. To guide risk communicators with this objective, a mental models approach to the design and characterization of risk communications is proposed. Building on text comprehension and mental models research, this approach offers an integrated set of methods to help the risk communication designer choose and analyze risk communication content, structure, and organization. An applied example shows that two radon brochures designed with this approach present roughly the same expert facts as a radon brochure widely distributed by the U.S. EPA but meet higher standards on other content, structure, and organization criteria.     

Communicating Low Risk Magnitudes: Incidence Rates Expressed as Frequency Versus Rates Expressed as Probability

The study investigated the effects of incidence rates stated as a probability (e.g., 0006) and incidence rate information expressed in terms of frequency (e.g., 600 in 1,000,000) on risk-avoidant behavior. Subjects were informed about the risks associated with an old and a new, improved medication. They were asked how much they were willing to pay for the safer medicine. Risk information was given either in a frequency or a probability format. The second factor manipulated was the level of risk, either high or low. As expected, analysis of variance yielded a significant interaction. Subjects confronted with high risk in the frequency format were willing to pay the highest prices for the improved medication. The choice between frequency or probability format can be made according to the goal of the communication of risk.     

Knowledge-Based Representations of Risk Beliefs

Beliefs about risks associated with two risk agents, AIDS and toxic waste, are modeled using knowledge-based methods and elicited from subjects via interactive computer technology. A concept net is developed to organize subject responses concerning the consequences of the risk agents. It is found that death and adverse personal emotional and sociological consequences are most associated with AIDS. Toxic waste is most associated with environmental problems. These consequence profiles are quite dissimilar, although past work in risk perception would have judged the risk agents as being quite similar. Subjects frequently used causal semantics to represent their beliefs and "% of time" instead of "probability" to represent likelihoods. The news media is the most prevalent source of risk information although experiences of acquaintances appear more credible. The results suggest that "broadly based risk" communication may be ineffective because people differ in their conceptual representation of risk beliefs. In general, the knowledge-based approach to risk perception representation has great potential to increase our understanding of important risk topics.     

Agreement Between Scales in the Measurement of Breast Cancer Risk Perceptions

The objective of this article is to compare the accuracy and numeric responses of breast cancer risk perception as measured by a frequency scale and percentage scale. A cross-sectional survey was conducted. Perceptions of five-year and lifetime breast cancer risk were measured using a frequency and a percentage scale. Estimation error was calculated as the absolute difference between actual breast cancer risk as determined by the Gail model and perceived risk. Agreement between scales was determined by calculating the mean and standard deviation of the difference between numeric responses. The study was conducted among women enrolled in two primary care clinics associated with an academic medical center. Two-hundred-fifty-four participants were recruited from one of the two participating internal medicine clinics. Inclusion criteria included female gender and age 40-84 years. Exclusion criteria included a history of breast cancer, dementia, or a life expectancy of less than two years. The frequency scale was more accurate than the percentage scale in estimating lifetime risk (p= 0.05), but less accurate in estimating five-year risk (p < 0.02). Only 79 participants (31%) were considered consistent scale users, providing identical responses when using the frequency and percentage scale for a given risk estimate. Although the mean difference (percentage-frequency scale) for estimates of breast cancer lifetime risk was only 2.4, the empirically determined 90% limits of agreement between the frequency and percentage scale for lifetime risk were wide, from -30 to 40. Higher numeracy was associated with consistent use of scales (OR 1.61, 95% CI; 1.09-2.37). We report disagreement in breast cancer risk perceptions when measured by a frequency and a percentage scale. The accuracy and direction of bias associated with each scale varies according to the time frame of risk being assessed.     

Why Near‐Miss Events Can Decrease an Individual's Protective Response to Hurricanes

Prior research shows that when people perceive the risk of some hazardous event to be low, they are unlikely to engage in mitigation activities for the potential hazard. We believe one factor that can lower inappropriately (from a normative perspective) people's perception of the risk of a hazard is information about prior near‐miss events. A near‐miss occurs when an event (such as a hurricane), which had some nontrivial probability of ending in disaster (loss of life, property damage), does not because good fortune intervenes. People appear to mistake such good fortune as an indicator of resiliency. In our first study, people with near‐miss information were less likely to purchase flood insurance, and this was shown for both participants from the general population and individuals with specific interests in risk and natural disasters. In our second study, we consider a different mitigation decision, that is, to evacuate from a hurricane, and vary the level of statistical probability of hurricane damage. We still found a strong effect for near‐miss information. Our research thus shows how people who have experienced a similar situation but escape damage because of chance will make decisions consistent with a perception that the situation is less risky than those without the past experience. We end by discussing the implications for risk communication.     

Communicating uncertainty in national security intelligence: Expert and nonexpert interpretations of and preferences for verbal and numeric formats

Organizations in several domains including national security intelligence communicate judgments under uncertainty using verbal probabilities (e.g., likely) instead of numeric probabilities (e.g., 75% chance), despite research indicating that the former have variable meanings across individuals. In the intelligence domain, uncertainty is also communicated using terms such as low, moderate, or high to describe the analyst's confidence level. However, little research has examined how intelligence professionals interpret these terms and whether they prefer them to numeric uncertainty quantifiers. In two experiments (N = 481 and 624, respectively), uncertainty communication preferences of expert (n = 41 intelligence analysts in Experiment 1) and nonexpert intelligence consumers were elicited. We examined which format participants judged to be more informative and simpler to process. We further tested whether participants treated verbal probability and confidence terms as independent constructs and whether participants provided coherent numeric probability translations of verbal probabilities. Results showed that although most nonexperts favored the numeric format, experts were about equally split, and most participants in both samples regarded the numeric format as more informative. Experts and nonexperts consistently conflated probability and confidence. For instance, confidence intervals inferred from verbal confidence terms had a greater effect on the location of the estimate than the width of the estimate, contrary to normative expectation. Approximately one-fourth of experts and over one-half of nonexperts provided incoherent numeric probability translations for the terms likely and unlikely when the elicitation of best estimates and lower and upper bounds were briefly spaced by intervening tasks.     

From the Inside Out: Environmental Agency Views about Communications with the Public

Relatively few studies have examined risk communication with the public from the viewpoint of the staff of institutions attempting such communications. This paper reports results of interviews with managers and staff of the New Jersey Department of Environmental Protection, on their programs' current and ideal communications with the public. Q analysis revealed two orthogonal perspectives on current program communications, the Enthused and the Constrained views. The primary focus of their divergence was on the matter of commitment and support for such communications. The Enthused group felt that they had attitudinal support from program culture and managers; the Constrained group focused on the lack of concrete operational support in terms of time, money, and expertise. These differences between the two groups did not appear to be associated with gender, managerial status, education, communication training, or organizational unit. When ideal program communications were discussed, the focus was on the need for all kinds of commitment and support (i.e., culture, managers, time, money, building expertise through training), as well as on more proactive and responsive communication. Both perspectives agreed that communication with the public is essential to their programs' success, trying new ways to communicate is worthwhile, scientific bases for decisions are not compromised by communication, and communication is not delegated to specialists. While varying ideas were held of the public's capacity and interest in communication, citizens were largely not held responsible for communication problems.     

Perception of Risk and Credibility at Toxic Sites

This study integrates previous research methodologies to compare the risk perceptions and responses to risk messages of agency personnel and neighbors of Superfund sites in Michigan. The integration attempted and the focus on risk messages are shaped by a critical review of the social amplification conceptual framework. The study involved all four agency groups and three groups of site neighbors actively involved in Superfund planning across the state. The first part of the study utilized the psychometric techniques of hazard rating and hazard profiles that had not previously been used in studies involving stakeholders. While agency personnel responded similarly to experts in previous studies, the responses of individuals in the neighbor groups reflected experience with toxic sites and were dissimilar to previous ratings by the general public. The second part of the study consisted of a hypothetical toxic site scenario that focused on specific risk messages at different times in the site history. Results indicate that the difference in perception of risk occurs after the first testing at a site, and that dramatic differences arise between agency and resident groups regarding the credibility of information sources and the need for independent testing. A general lack of trust in the Superfund program was demonstrated by all groups. The results indicate that problems of institutional credibility and program adequacy cannot be addressed by better risk communication.     

Toward an Understanding of the Influence of Cultural Background and Domain Experience on the Effects of Risk‐Pricing Formats on Risk Perception

Risk‐information framing can be a powerful tool for aiding the communication of risk and improving decision making. However, little work has investigated the extent that these framing effects depend on the characteristics of the perceiver. In our study, we examine whether the effects of different risk‐pricing formats on risky choices are the same for all individuals, no matter their domain experience or cultural background, or whether there are interactions between these factors. Survey 1 revealed that three risk‐pricing formats of the same choice problem resulted in the same individuals making different risky choices (preference reversal), suggesting that risk perception was distorted by the risk‐pricing format manipulation. In Survey 2, the effects of the risk‐pricing formats were shown to differ by the participants’ cultural background (Asian vs. European) and the extent of their domain experience. The fact that there were no differences between the cultural or domain experience groups in their overall tendency to select riskier (cf. safer) choices indicates that risk behavior differences between groups are often closely linked to perceptual, rather than simply attitudinal, cognitive processes. The results demonstrate the complex, interactive cognitive processes that are used to encode risk information, involving the framing of the information and the cultural background and previous experiences of the individual. We conclude that it is important to consider the characteristics of the individual (e.g., culture, domain experience, etc.) when manipulating risk‐information framing with the aim of improving risk communication.     

The Measurement of Subjective Probability: Evaluating the Sensitivity and Accuracy of Various Scales

The RISK of an event generally relates to its expected severity and the perceived probability of its occurrence. In RISK research, however, there is no standard measure for subjective probability estimates. In this study, we compared five commonly used measurement formats—two rating scales, a visual analog scale, and two numeric measures—in terms of their ability to assess subjective probability judgments when objective probabilities are available. We varied the probabilities (low vs. moderate) and severity (low vs. high) of the events to be judged as well as the presentation mode of objective probabilities (sequential presentation of singular events vs. graphical presentation of aggregated information). We employed two complementary goodness‐of‐fit criteria: the correlation between objective and subjective probabilities (sensitivity), and the root mean square deviations of subjective probabilities from objective values (accuracy). The numeric formats generally outperformed all other measures. The severity of events had no effect on the performance. Generally, a rise in probability led to decreases in performance. This effect, however, depended on how the objective probabilities were encoded: pictographs ensured perfect information, which improved goodness of fit for all formats and diminished this negative effect on the performance. Differences in performance between scales are thus caused only in part by characteristics of the scales themselves—they also depend on the process of encoding. Consequently, researchers should take the source of probability information into account before selecting a measure.