Cumulative risk communication: comparing practitioner perspectives with public interpretation and response

Abstract

Background

Cumulative risk assessment has become a key framework for understanding how multiple environmental and public health stressors interact, yet little is known about how these risks are communicated to the public. This study examined both professional practices and public responses to cumulative risk communication.

Methods

Fifteen semi-structured interviews were conducted with federal risk communicators, and a nationally representative online survey of 2,531 U.S. adults was fielded. The survey included an experiment comparing responses to cumulative versus single-risk messages. Qualitative data from interviews were thematically analyzed, and survey responses were analyzed using descriptive statistics and experimental contrasts.

Results

Most practitioners acknowledged the importance of cumulative risk but reported rarely communicating it, citing time pressures, institutional silos, and perceived audience cognitive limitations as barriers. Public survey results contradicted this latter assumption: a majority of respondents preferred messages that addressed multiple, interacting risks, and experimental data showed no significant differences in comprehension, prioritization, or behavioral intentions between cumulative and single-risk messages.

Conclusions

Findings reveal a disconnect between practitioner assumptions and public responses to cumulative risk communication. Closing this gap will require organizational support, accessible communication tools, and continued research to align professional practice with public preferences and capabilities.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-25206-6.

Background

Decades of research in environmental health shows that people are rarely exposed to a single hazard in isolation. Instead, multiple stressors, such as environmental pollutants, social and economic conditions, and natural hazards, often overlap and interact over time. These combined exposures shape health, safety, and environmental outcomes in complex and dynamic ways. Recognizing this reality, many researchers, practitioners, and government agencies have developed tools to better assess and address the cumulative nature of risk. For instance, the U.S. Environmental Protection Agency (EPA) has adopted formal methods for evaluating combined exposures, known as cumulative risk assessment.

In a seminal 2003 report that provides a framework for cumulative risk assessment, the EPA defined cumulative risk as “the combined risk from aggregate exposures to multiple agents or stressors” [7]. Other researchers have expanded on this definition. For example, Callahan and Sexton [5] describe cumulative risk as “the combined threats from exposure via all relevant routes to multiple stressors including biological, chemical, physical, and psychosocial entities.”

Cumulative risk assessment, by extension, refers to “an analysis, characterization, and possible quantification of the combined risks to health or the environment from multiple agents or stressors” [7]. Similarly, Callahan and Sexton describe it as “a tool for organizing and analyzing information to examine, characterize, and possibly quantify the combined adverse effects on human health or ecological resources from multiple environmental stressors” (2007).

Since the EPA’s formal definition in 2003, a growing body of research has explored how to apply and refine cumulative risk assessment. Scholars have contributed both practical examples and methodological guidance for conducting these assessments (e.g., [1–4, 6, 10, 15, 21, 22, 24]).

Many examples of cumulative risk have received attention from scholars in the last two decades. In recent years, for example, researchers have increasingly focused on how indoor and outdoor air pollution, combined with extreme heat, interact to produce cumulative impacts. Studies show that outdoor pollutants can infiltrate indoor environments, compounding exposure for people who spend much of their time inside (e.g., [12, 13]). Extreme heat can further degrade air quality and amplify the health impacts of pollution (e.g., [9, 18]). When these stressors converge, they can pose serious threats, particularly for vulnerable populations. In response, some scholars have called for the development of warning systems that explicitly address cumulative risk rather than treating risks in isolation (e.g., [17]).

While research has been essential for advancing how cumulative risk is measured and analyzed, far less attention has been given to how it is communicated. Very little work to date has focused on the challenges and best practices of cumulative risk communication, defined here as the process of sharing information about the combined effects of multiple stressors in ways that are understandable, credible, and actionable for affected populations, decision-makers, and other stakeholders. Most existing risk communication practices, and the studies that inform them, focus on individual risks in isolation, rather than on their cumulative effects.

This disconnect between risk assessment and communication raises important research questions that motivated the current project:

1. To what extent, and in what ways, do risk communicators incorporate cumulative risk into their communication strategies?

2. How do members of the public interpret, prioritize, and respond to information about cumulative risk?

These questions point to the importance of understanding both how practitioners communicate about cumulative risks and how the public interprets and reacts to that communication. Answering them can strengthen the clarity, relevance, and effectiveness of environmental risk messages, particularly in communities facing multiple, compounding stressors.

Methods

To address these questions, researchers at the University of Oklahoma’s Institute for Public Policy Research and Analysis (IPPRA), in collaboration with partners at the U.S. Environmental Protection Agency (EPA), designed a mixed-methods study examining both professional practices and public responses related to cumulative risk communication. The study included two main components: (1) qualitative interviews with risk communication professionals across several federal agencies, and (2) a nationally representative survey of the U.S. public. Both components were conducted with approval and oversight from the University of Oklahoma’s Institutional Review Board.

Practitioners perspectives on cumulative risk communication

The first component explored how risk communicators within different federal agencies understand and approach cumulative risk. The research team conducted 15 semi-structured interviews with key personnel from the EPA (8), the National Oceanic and Atmospheric Administration/National Weather Service (NOAA/NWS) (2), the Federal Emergency Management Agency (FEMA) (3), the Centers for Disease Control and Prevention (CDC) (1), and the National Institutes of Health (NIH) (1). Interviews were conducted between October 28 and December 18, 2024, and lasted an average of 36 min (ranging from 27 to 50 min). Participants were selected based on their active roles in communicating risks to the public across a variety of environmental and public health hazards. Informed consent was obtained from all participants prior to their participation in the study.

All interviews were recorded, transcribed, and analyzed using thematic coding to identify key patterns across participants. The coding process was primarily inductive and informed by principles of grounded theory. Rather than applying a predefined coding scheme, researchers used open and axial coding to allow themes to emerge directly from the data. Transcripts were reviewed independently by multiple members of the research team to enhance reliability, and discrepancies in interpretation were resolved through discussion and consensus. This approach ensured that the analysis was closely aligned with participants’ own language, experiences, and perspectives.

To avoid overstating the precision of findings from this relatively small number of interviews, the results discussed below do not include exact numbers or percentages, as doing so may give the impression of unwarranted accuracy given the interpretive nature of thematic analysis and the limited sample size. Instead, the results are reported using general descriptors to convey how commonly a theme appeared across participants. Themes mentioned by 2 to 4 (of 15) participants are described as raised by “some” participants; those raised by 5 to 7 participants are described as mentioned by “several.” Themes noted by 8 to 11 participants are described as mentioned by “many,” suggesting a widespread pattern. If a theme was raised by 12 to 14 participants, it is described as mentioned by “most,” indicating near consensus. When all 15 participants raised a particular theme, it is described as mentioned by “all” participants.

Public interpretation and response to cumulative risk communication

The second component focused on how members of the public interpret and respond to information about cumulative risk as compared to information about single risks. To investigate this, the research team developed and fielded a nationally representative web survey of 2,531 U.S. adults between May 15 and May 29, 2025. The survey achieved a completion rate of 98.7%. The only noticeable data quality issue was that some respondents completed the survey very quickly; to address this, respondents who finished in less than five minutes (n = 309) were removed from the dataset.

The sample was provided by Verasight, a survey research firm that maintains a large panel of survey respondents recruited through a combination of probability-based and non-probability-based sampling methods including random address-based sampling, random person-to-person text messaging, and dynamic online targeting. The full survey questionnaire is provided in Appendix A. Informed consent was obtained from all participants prior to their participation in the study, and respondents were compensated for their time through the survey vendor’s standard incentive system.

Like most nonprobability samples, the opt-in nature the survey panel used for this study introduces the possibility of selection bias, though invitations to this study did not disclose the survey’s purpose, which reduces concern about topic-specific bias. Nevertheless, to enhance generalizability and further limit the potential impact of selection, the data were weighted to match the March 2025 Current Population Survey on key demographic variables, including age, race/ethnicity, sex, income, education, region, and metropolitan status. Additional adjustments incorporated a running three-year average of partisanship distributions from Pew Research Center’s National Public Opinion Reference Survey (NPORS) benchmarking surveys, as well as population benchmarks from the 2024 vote, to ensure alignment with salient characteristics that may shape attitudes toward environmental and health risks in the U.S. electorate. Table 1 reports the population benchmarks alongside the unweighted and weighted sample proportions for these demographic and political variables. As shown in the table, the unweighted sample distributions were already reasonably close to the population benchmarks, and the application of survey weights further improved alignment, yielding even stronger representation of the target population.

Table 1.

Population benchmarks vs. survey sample distributions

Category	Benchmark	Population	Sample (Unweighted)	Sample (Weighted)
Gender	Male	48.7%	45.5%	48.3%
Gender	Female	51.3%	54.5%	51.7%
Age	18–29	19.9%	16.0%	19.1%
Age	30–49	33.7%	36.8%	33.2%
Age	50–64	23.2%	26.4%	23.8%
Age	65 +	23.2%	20.8%	23.9%
Education	HS or less	37.1%	33.4%	36.3%
Education	Some college/2-yr degree	26.2%	29.0%	26.6%
Education	4-yr/post-graduate degree	36.7%	37.6%	37.2%
Race/Ethnicity	White	59.8%	62.0%	60.3%
Race/Ethnicity	Black	12.1%	12.6%	11.7%
Race/Ethnicity	Hispanic	18.4%	17.1%	18.2%
Race/Ethnicity	Other	9.6%	8.3%	9.7%
Census Region	Northeast	17.3%	16.5%	17.1%
Census Region	Midwest	20.4%	19.5%	20.7%
Census Region	South	38.6%	41.2%	38.3%
Census Region	West	23.6%	22.8%	23.9%
Metro Status	Metropolitan	87.5%	87.5%	88.0%
Metro Status	Non-metro	12.5%	12.5%	12.0%
Household Income	< $50,000	29.0%	33.2%	28.9%
Household Income	> $50,000	71.0%	66.8%	71.1%
Party ID	Democrat	30.4%	31.0%	30.7%
Party ID	Independent	39.4%	39.6%	38.9%
Party ID	Republican	30.2%	29.5%	30.4%
Vote 2024	Kamala Harris	28.3%	31.2%	28.3%
Vote 2024	Donald Trump	29.2%	31.8%	28.7%
Vote 2024	3rd or didn't vote	42.5%	37.0%	43.0%

Understanding, experience, and preferences

The first part of the survey included a variety of background questions, such as general risk perceptions, prior experiences with environmental hazards, and trust in government agencies. It also featured items designed to assess respondents’ understanding of cumulative risk, their past experiences with cumulative risk communication, and their preferences for how such information should be communicated in the future. Two short scenario-based questions were included to evaluate respondents’ understanding of cumulative risk.

In the first scenario, respondents were asked to imagine: “two people with the same underlying health conditions. One person is exposed to outdoor air pollution and extreme heat at different times. The other person is exposed to outdoor air pollution and extreme heat at the same time.” They were then asked, “Which person do you think is more likely to experience serious health problems over time?” Response options included: (1) The person exposed to the hazards at different times, (2) The person exposed to both hazards at the same time, (3) Both are equally likely, and (4) Not sure.

In the second scenario, respondents were told: “Imagine that health experts estimate that exposure to outdoor air pollution increases your chance of developing a serious health problem by 10%, and exposure to extreme heat increases your chance by 15%.” They were then asked: “Now imagine being exposed to both air pollution and extreme heat at the same time. What do you think your combined risk would be? In other words, what is the total chance of a serious health problem from being exposed to both at the same time?” Response options included: (1) 15% (the highest single risk), (2) 25% (sum of the two risks), (3) More than 25% (more than the sum of the two risks), and (4) Not sure.

The next set of items in this section of the survey measured respondents’ past experiences with cumulative risk communication and their preferences for how such information should be communicated in the future. To ensure that all respondents had a shared understanding of the concept, this portion of the survey began with a brief explanation of cumulative risk and one tradeoff between single and cumulative communication.

The explanation stated: “Risk communication can be challenging because people are often exposed to multiple environmental hazards at the same time. In some instances, the health impacts of those hazards don’t just add up—they interact in ways that make health effects worse. For example, breathing polluted air during a heatwave can put more stress on the body than either air pollution or heat alone.” It continued: “This makes it difficult to develop concise messages that clearly explain hazards, their impacts, and the protective actions you can take. As a result, communicators often face a tradeoff: Some prioritize simplicity—focusing on one risk at a time (such as air pollution) to provide a clear and straightforward message. Others prioritize completeness—focusing on how multiple risks interact to provide a detailed and comprehensive message.”

After the explanation, respondents were asked to reflect on their past exposure to different types of risk communication. Specifically, they were asked: “In your experience, which type of message is more common in the public health or safety information you typically see? (1) Simple messages that focus on a single risk, or (2) More complex messages that focus on multiple interacting risks.” They were then asked to indicate their personal preference for future communication approaches. The question read: “In your opinion, which type of message should be more common in future public health and safety communication? (1) Simple messages that focus on a single risk, or (2) More complex messages that focus on multiple interacting risks.”

Interpretation and responses to risk messages

The second part of the survey featured a between-subjects experimental vignette designed to test how respondents interpret and respond to different types of risk communication messages. Figure 1 illustrates the flow of the experiment. At the start of the vignette, respondents were told: “Here is an example message developed for a public information campaign. In the actual campaign, the message would appear in a more visually engaging and attention-grabbing format. For now, please focus on the text alone.” They were then randomly assigned to one of six message conditions. Each respondent saw only one of the experimental messages. Three focused on a single risk and three focused on cumulative risk.

Fig. 1.

Flow chart of the survey experiment

The single-risk messages addressed either indoor air pollution, outdoor air pollution, or extreme heat. Respondents in these conditions were shown information about only one specific risk, reflecting how many public health and environmental campaigns currently communicate hazards in isolation. In contrast, the cumulative risk messages presented information about the combined or interactive effects of multiple hazards. One message focused on indoor and outdoor air pollution, another on outdoor air pollution and extreme heat, and a third included all three: indoor air pollution, outdoor air pollution, and extreme heat. Each message explained how these risks can interact and lead to greater harm than when considered individually. In all six cases, the messages highlighted the identified hazard(s), potential health impact(s), and recommended protective actions. To illustrate, Fig. 2(A) shows the single-risk message presented to respondents assigned to the outdoor air pollution condition, while Fig. 2(B) shows the cumulative risk message presented to respondents assigned to the condition that provided information about indoor and outdoor air pollution and extreme heat. The full text of all messages is provided in Appendix A.

Fig. 2.

Example (A) single and (B) cumulative risk messages from the experiment

After reading their assigned message, all respondents answered the same set of follow-up questions. These outcome questions were designed to assess whether people respond differently to information about multiple, related risks (cumulative risk) compared to information about a single risk. Drawing from past research on the effectiveness of health risk messages (e.g., [8, 16]), the outcome questions fell into two main categories: message characteristics and message response.

The first set focused on how respondents perceived the characteristics of the message itself—specifically, whether it was memorable, simple, relevant, and actionable. Memorability was measured by asking: “Which of the following risks were highlighted in the message you saw on the previous page? Please select all that apply.” Simplicity, relevance, and actionability were measured by asking respondents how much they agreed or disagreed with the following statements:

• “The information in the message was simple and easy to understand” (simplicity)

• “The message felt relevant to me and my situation” (relevance)

• “The message gave me clear steps I could take to protect myself if necessary” (actionability)

Responses were collected using a five-point Likert scale ranging from strongly disagree to strongly agree. Importantly, respondents answered these questions based on their initial reading of the message. They were not allowed to go back and re-read the message while completing this set of items.

The second set of questions, drawn from Protection Motivation Theory (PMT) [14, 19, 20], was designed to evaluate message response and assess the underlying factors that, according to the theory, influence whether individuals are motivated to take protective action. Specifically, the survey measured perceived severity, perceived vulnerability, response efficacy, self-efficacy, and behavioral intention by asking respondents how much they agreed or disagreed with the following statements:

• “The risks mentioned in the message could have serious consequences for my health and safety” (perceived severity)

• “I believe I am personally at risk from the hazards described in the message” (perceived vulnerability)

• “Taking the actions recommended in the message would effectively reduce my health and safety risks” (response efficacy)

• “I am capable of taking the actions recommended in the message if needed” (self-efficacy)

• “I intend to follow the actions recommended in the message if needed” (behavioral intention)

Each response was recorded using a five-point Likert scale ranging from strongly disagree to strongly agree. Unlike the earlier items on message characteristics, respondents were allowed to view the message while completing this set of questions.

The third set of questions, drawn from the Planned Risk Information Seeking Model (PRISM) [11], was designed to measure respondents’ likelihood of seeking additional information about the risks described in the message they saw. While these questions offered valuable insights, their responses are not analyzed in this report, as the primary focus is on how participants interpret and respond to messages, not on their subsequent information-seeking behaviors.

Results

Practitioners perspectives on cumulative risk communication

To what extent, and in what ways, do risk communicators recognize and incorporate cumulative risk into their communication strategies? Thematic analysis of the interview transcripts revealed that most practitioners were familiar with the general concept of cumulative risk and agreed that it is important, even if they were not always familiar with the specific term or its formal definition. For example, one participant explained, “when I think about cumulative risk and cumulative impacts, I think of it within the context of total human exposures to things that might cause negative outcomes” (Participant 5). Another noted, “once you couple [one hazard] with another hazard, it could increase the impacts or increase the vulnerabilities and vice versa” (Participant 13).

In some cases, this lack of specific familiarity with the concept of cumulative risk led to overlapping or blurred interpretations. Many practitioners referenced related but distinct ideas, such as compounding risk, where multiple hazards interact simultaneously to amplify overall impacts, or cascading risk, where one hazard triggers a sequence of subsequent hazards. As distinct from these concepts, cumulative risk emphasizes the combined effects of multiple exposures or stressors that accumulate over time, whether or not they occur simultaneously or sequentially. One interviewee explicitly noted the confusion, stating, “So, you know, cumulative risk, I usually confound it with compounding risk” (Participant 8). Another remarked, “I think in our domain of weather we often talk about cascading risks, which I’d say are related, not the same, but I think that’s often how we think about it in this domain” (Participant 10). Reflecting more broadly on the distinction between these concepts, one participant said, “I’m not sure that I necessarily differentiate between them just because I think we were trying to kind of build this holistic approach to how we think about risk and protecting communities” (Participant 9). These examples illustrated how the boundaries between related concepts were not always clearly defined in practice.

Despite recognizing its importance, many practitioners reported that they rarely incorporate cumulative risk into their communication strategies. Instead, their communication efforts tend to focus primarily, or even exclusively, on one risk at a time. As one participant explained, “in my experience is that there will be one or two issues that the community wants to know, and wants answered, and it behooves us to focus on those” (Participant 1). Others expressed similar views, stating, “I would say primarily we focus on a single thought, a single hazard” (Participant 12), and “often when we’re communicating risks, we often take them one at a time” (Participant 10).

This raised an important question: if practitioners recognize the importance of cumulative risk, why is it so rarely incorporated into their communication strategies? Thematic analysis of the interview transcripts identified three recurring themes in participants’ responses to this question. These themes were often described as barriers or constraints that make it difficult to communicate about cumulative risk in practice.

The first theme was time constraints. Most practitioners explained that their work is often driven by urgent or crisis-related events, which require immediate attention and fast communication. In high-pressure, event-driven settings, they tend to prioritize the most immediate or visible threat, leaving little room to address additional or interacting risks. For example, one participant noted that during major wildfire events, communication efforts typically focus on the risks of poor outdoor air quality, especially when that concern is top of mind for the affected community, even though the same community may also face related risks, such as poor indoor air quality or compromised water quality. One participant explained, “If we're in the middle of a crisis like a wildfire, then we want to focus on how do we get people to reduce their exposures to smoke right away” (Participant 2). Practitioners described this as a necessary tradeoff, emphasizing the importance of addressing the most pressing risks first, much like a triage approach in an emergency room, before turning to other, less immediate concerns.

The second theme was perceived cognitive constraints. Many participants described cumulative risk as complex and difficult to explain. Several expressed concern that discussing multiple risks at once could confuse or overwhelm their audiences. This perceived cognitive burden was viewed as a major barrier to more integrated communication strategies. According to participants, the technical nuance required to clearly explain a single risk becomes even more challenging when additional risks and their interactions are introduced. They worried that members of the public might struggle to understand the information or feel overwhelmed, ultimately making the message less effective. For example, participants expressed concern that “it’s always challenging to speak to communities that don’t have a familiarity and comfort with science and data” (Participant 1), and that “we find that it’s hard for people to take in a lot of messages at once” (Participant 15). Another said, “The barriers are translating the science into plain spoken, human action-oriented solutions and relating to the humans who are affected” (Participant 11). One participant added, “Your message needs to be short and easily understood” (Participant 12). When asked about barriers to communicating multiple risks, the same participant simply stated, “Confusion, information overload, information paralysis” (Participant 12). This perception likely reinforces the first theme, time constraints, by making communicators even more reluctant to address cumulative risks during urgent or high-stakes events, when clarity and speed are prioritized.

The third theme was institutional constraints. Many practitioners pointed to agency structures, such as internal policies, reporting hierarchies, and limited resources, that discourage or complicate cumulative risk communication. For example, communicators who focus on air quality often work in separate units from those responsible for extreme heat or water contamination. Similarly, there are administrative divides between teams focused on preparedness, response, mitigation, and recovery. One participant noted, “the fact that resilience and recovery sit in different offices often become a little bit of a struggle just because they're siloed, and yet they're directly intertwined when you're trying to recover in a way that's going to make you more resilient in the future” (Participant 9). Many practitioners noted that these organizational silos can make it difficult to develop coordinated messages that reflect the combined effects of multiple risks. As one participant explained, “I think the regulatory approaches may make it challenging because we tend to regulate on a pollutant-by-pollutant basis” (Participant 2). Another noted, “Programs are given directives [along with the] funding and statutory authority to deal with sets of issues, not the full cumulative issue” (Participant 4). Taken together, these examples illustrate how institutional structures and mandates can fragment communication efforts and limit attention to cumulative risks.

Public interpretation and response to cumulative risk communication

As highlighted in the interview findings, one factor that appeared to limit cumulative risk communication was the perception among several practitioners that cumulative risk was too complex for members of the public to understand. As a result, some believed that the public was better able to interpret and respond to single risk messages. These perceptions served as useful, testable hypotheses that helped organize the analysis of the survey results.

The first hypothesis was that most members of the public would not understand the basic insight that risks are often interactive—that is, multiple risks can combine to produce greater harm than any one risk alone. Responses to the short scenario-based questions described in "Understanding, experience, and preferences" section provided a direct test of this hypothesis.

Figure 3(A) shows the distribution of responses to a question comparing the health risks of being exposed to outdoor air pollution and extreme heat at the same time versus at different times. Nearly half of respondents (45%) correctly recognized that simultaneous exposure posed greater risk. Another 46% said both exposures were equally risky, indicating a partial understanding of risk interaction. Very few (9%) chose one of the other two options.

Fig. 3.

Intuitive understanding of cumulative risk: (A) Question 1 and (B) Question 2

Figure 3(B) shows responses to the question asking respondents to estimate the combined risk of simultaneous exposure to outdoor air pollution and extreme heat. A plurality (43%) correctly answered that the combined risk would be more than 25%, indicating an intuitive understanding that risks can interact synergistically. Another 33% selected the 25% option, recognizing additive risk but not interaction. Only 11% chose the 15% option, and 13% were unsure.

While not definitive, the results from these two questions challenge the hypothesis that most members of the public lack a basic understanding of cumulative risk. Rather, a substantial portion of respondents, despite receiving no prior explanation, demonstrated either full or partial comprehension of how multiple risks can interact to increase harm.

A second, related hypothesis that emerged from the interview findings was that most members of the public prefer single-risk messages over cumulative risk messages. Responses to the second set of questions described in "Understanding, experience, and preferences" section provided a basis for testing this hypothesis. Figure 4(A) shows that, when reflecting on their past experiences, most respondents (60%) reported that they typically encounter simple messages that focus on a single risk at a time. However, as shown in Fig. 4(B), an even larger majority (76%) expressed a preference for more complex messages that address multiple interacting risks. Taken alone, the second result clearly refutes the hypothesis that the public prefers single risk messages. More revealing, however, is the contrast between experience and preference: combining these results into a two-by-two matrix (Table 2) shows that 42% of respondents typically receive simple messages but would prefer more information about cumulative risk. Only 5% reported the opposite, that they usually receive more complex messages addressing multiple risks, but would prefer simpler, single-risk messages.

Fig. 4.

(A) Experience with and (B) preferences for cumulative risk messages

Table 2.

Experience vs. preferences for cumulative risk messages

		Preference
		Single risk messages	Multiple risks messages
Experience	Single risks messages	17%	42%
	Multiple risks messages	5%	35%

The third hypothesis that surfaced in the interviews with practitioners was that most members of the public are better able to interpret and respond to single risk messages than to cumulative risk messages. The survey experiment described in "Interpretation and responses to risk messages" section was explicitly designed to test this hypothesis by comparing responses to single and cumulative risk messages across two primary outcomes: perceived message characteristics and protective response.

The first dimension of perceived message characteristics was memorability, assessed by asking respondents to recall which hazards were mentioned in the message they were shown. Correct recall was evaluated in two ways. The first metric (total recall) calculated the percentage of respondents who correctly identified all the hazards in their assigned message. While this is a straightforward metric, it represents a relatively difficult test, especially for those in the cumulative risk conditions who were required to recall and select multiple hazards accurately. To account for this, the second metric (partial recall) calculated the percentage of respondents who correctly identified at least one of the hazards mentioned in their assigned message. This more lenient measure may be a better indicator of basic message retention, particularly in real-world settings where audiences are unlikely to retain every detail but may still walk away with a general understanding of the risks involved. It is important to note, however, that this partial recall metric carries the opposite limitation as the total recall metric—it is inherently easier for respondents to recall one of multiple hazards, rather than one, and it may be more susceptible to correct guesses.

Figure 5(A) shows the results for the total recall metric, and Fig. 5(B) shows the results for partial recall. The points on each plot indicate the percentage of respondents in each experimental condition, with 95% confidence intervals represented by error bars. The horizontal lines display the average percentages for the broader single and cumulative risk conditions, with shaded areas indicating the corresponding 95% confidence intervals.

Fig. 5.

A Total and (B) partial recall of the hazards mentioned in the experimental messages

Beginning with Fig. 5(A), the results indicate that total recall was generally higher among respondents who viewed single risk messages. On average, 91% of those in the single-risk conditions correctly recalled the hazard mentioned in their message, compared to 75% of respondents in the cumulative risk conditions who correctly recalled all hazards. It is worth noting, however, that recall also varied across the specific message conditions. Among respondents in the single risk condition, those assigned to the outdoor air pollution message were less likely to recall the hazard than those who received messages about indoor air pollution or extreme heat. While the first pattern is notable, its cause is not immediately clear. In the cumulative risk condition, recall was highest among respondents who saw the message about outdoor air pollution and extreme heat, slightly lower for those in the indoor and outdoor air pollution group, and lowest among those who received the message covering all three hazards. This pattern is more straightforward. Total recall declined as the number of hazards increased, though it remained surprisingly high (68%) even in the most complex condition.

Figure 5(B) shows the opposite pattern: partial recall was higher among respondents who saw cumulative risk messages. About 96% of these respondents correctly recalled at least one of the hazards mentioned, compared to 91% in the single risk condition. This estimate is unchanged because, in the single risk condition, partial and total recall are effectively the same. In contrast to the total recall results, where there were differences across specific risk conditions within the cumulative risk group, partial recall was remarkably consistent across conditions. This is somewhat surprising. One might have expected higher partial recall among participants who viewed the message combining indoor air pollution, outdoor air pollution, and extreme heat, simply because there were more hazards to potentially guess correctly.

If one adopts the total recall metric, these results offer some support for the hypothesis that members of the public are better able to interpret and respond to single risk messages than cumulative risk messages. If retaining all the information is the primary goal, single risk messages appear to be slightly more memorable. However, an argument can be made that remembering at least one hazard and the appropriate response is nearly as important, if not more important, than recalling the entire message. If that is the goal, then the hypothesis, at least in terms of memorability, is not supported by the data.

The second, third, and fourth dimensions used to assess message characteristics across the experimental conditions were perceived simplicity, relevance, and actionability. Figures 6(A) through 6(C) present the results for each dimension, respectively. In each plot, the points represent the mean rating on the five-point agreement scales by experimental condition, with 95% confidence intervals shown as error bars. Horizontal lines indicate the mean ratings for the broader single and cumulative risk conditions, and the shaded areas around these lines represent their respective 95% confidence intervals.

Fig. 6.

A Simplicity, (B) relevance, and (C) actionability the experimental messages

Figure 6(A) shows that, on average, respondents who viewed single risk messages rated them as slightly simpler and easier to understand (M = 4.33) than those who viewed cumulative risk messages (M = 4.22). Although the difference was small, an independent samples t-test indicated that it was statistically significant, t(2426) = 2.95, p < 0.05. As with recall, it is worth noting that the outdoor air pollution single risk message was somewhat of an outlier. Respondents in that condition rated their message as less simple than those in the other two single risk conditions, with ratings comparable to those who were assigned cumulative risk messages.

With respect to relevance and actionability, Figs. 6(B) and 6(C) show that the difference in mean ratings between respondents who viewed single and cumulative risk messages was effectively zero. Neither type of message was perceived as more relevant or more actionable than the other. As with perceived simplicity, the outdoor air pollution message stood out as a slight outlier, receiving somewhat lower ratings for both relevance and actionability compared to the other single risk messages. On the cumulative risk side, the three messages were rated as roughly equivalent in terms of relevance but showed some variation in actionability. On average, respondents who viewed the message about outdoor air pollution and extreme heat rated it as less actionable than respondents who viewed the other cumulative risk messages.

Figures 7(A) through 7(E) follow the same format to display results from the PMT-based items used to measure intended behavioral response and its precursors, collectively referred to here as protective response outcomes. Looking across the precursors (perceived severity, perceived vulnerability, response efficacy, and self-efficacy) there was only one notable difference between respondents who viewed single risk messages and those who viewed cumulative risk messages: perceived vulnerability. On average, respondents who viewed cumulative risk messages reported slightly higher perceived vulnerability (M = 3.61) than those who viewed single risk messages (M = 3.48). Though relatively small in magnitude an independent samples t-test indicated that this difference was statistically significant, t(2430) = 2.43, p < 0.05. For all other precursors, the single and cumulative risk messages were functionally equivalent in their impact.

Fig. 7.

A Perceived severity, (B) vulnerability, (C) response efficacy, (D) self-efficacy, and (E) intended response to the experimental messages

Even more importantly, the same null pattern held for behavioral intentions. As shown in Fig. 7(E), respondents who viewed single risk messages were no more likely to report that they would follow the recommended actions than those who viewed cumulative risk messages. In other words, broad message type (single vs. cumulative risk) had no meaningful impact on respondents’ stated likelihood of engaging in protective behaviors, and very limited effect on the underlying perceptions that often motivate protective action.

There were, however, a few notable differences within the broad message conditions, particularly among respondents who viewed single risk messages. For example, those who saw the extreme heat message reported slightly higher levels of intended action compared to those who saw the indoor or outdoor air pollution messages. This pattern (with extreme heat consistently rated highest) was also reflected across all the precursor measures, although the differences were generally smaller and often not statistically significant. In contrast, there were no notable differences among respondents who viewed the different cumulative risk messages, perhaps because the presence of multiple hazards led to a blending or perceptions across message types.

With the possible exception of message recall and perceived simplicity—where only very small differences were observed—the collective results from the survey experiment provide relatively strong evidence against the third hypothesis that emerged from the practitioner interviews. There is no evidence that most members of the public are better able to interpret and respond to single risk messages than to cumulative risk messages. Instead, the findings suggest that the two message types are roughly equivalent in their impact on commonly used outcome measures of message effectiveness.

Discussion

This study was motivated by an apparent divide between risk assessment and risk communication. On the risk assessment side, both researchers and practitioners have long recognized the cumulative nature of risk. Yet relatively little attention has been devoted to how cumulative risk is or should be communicated to the public. To help bridge this divide, the study addressed two research questions using a combination of interviews with risk communication practitioners and a nationally representative survey of the general public. First, to what extent, and in what ways, do risk communicators incorporate cumulative risk into their communication strategies? Second, how do members of the public interpret, prioritize, and respond to information about cumulative risk?

In response to the first question, the interviews revealed that although most practitioners acknowledge the importance of cumulative risk, they rarely incorporate it into their communication strategies, instead focusing on single risks one at a time. This tendency appears to stem from a combination of reported time, cognitive, and institutional constraints. Time constraints arise from the urgency of crisis-driven communication that many practitioners experience, pushing them to prioritize the most immediate or visible threat rather than the broader set of risks a person or community may be facing at any given time. Perceived cognitive constraints arise from the reported complexity of cumulative risk, which many practitioners believe may confuse or overwhelm the public. This perception reflects what some scholars describe as reliance on a ‘mythical public’, an assumed audience that is easily confused and incapable of processing complex information (e.g., [23, 25]). By framing the public in this way, practitioners can justify prioritizing single, clearly defined hazards rather than doing the harder work of communicating about multiple, interacting threats, even though evidence suggests that people can engage with complex information when it is conveyed clearly and accessibly. Institutional constraints arise from organizational silos and relatively rigid administrative structures that often separate responsibility for different hazards or phases of risk management. These divisions make it difficult for practitioners to deliver coordinated messages that reflect cumulative risk involving multiple hazards and phases. In other words, many current institutions encourage practitioners to ‘stay in their lane,’ which often translates to focusing on a single risk at a time rather than addressing the broader, interconnected nature of cumulative risk.

In response to the second question, the survey results revealed that members of the public appear to recognize the tendency of practitioners to focus on a single risk at a time. When given the option, most (though not all) respondents expressed a preference for more complete messages that include information about cumulative risk when it is relevant. Perhaps more importantly, the survey provided a direct test of the perceived cognitive constraint identified in the interviews. The common belief among practitioners that cumulative risk is too complex for the public to understand, and that, as a result, people are better able to interpret and respond to single risk messages. The data collected for this study offer little support for this perception. Instead, the data suggest that many members of the public intuitively grasp the concept of cumulative risk, even though it is complex and can be challenging to communicate. Moreover, there appears to be little or no difference in how people interpret and respond to single versus cumulative risk messages. Across a range of outcomes designed to assess message characteristics and behavioral responses, cumulative risk messages performed comparably to single risk messages. This finding directly challenges the cognitive constraint theory often cited by practitioners. The public appears less constrained than the ‘mythical public’ practitioners sometimes imagine, demonstrating a capacity to engage with complex messages when they are communicated clearly.

While these findings provide useful insights, they should be interpreted in light of several limitations. First, the interview sample included a relatively small and non-random subset of risk communication practitioners. Although participants were selected for their experience across different sectors, the range of perspectives may not fully capture the diversity of communication strategies used across agencies, organizations, and local contexts.

Second, although the survey sample was designed to be nationally representative based on key demographic characteristics, it was not drawn using probability-based sampling. As a result, the findings may not be fully generalizable to the broader U.S. population, and they should be interpreted with some caution. This limitation is common to most opt-in online panels, but it underscores the importance of future research that leverages probability-based designs or mixed-method approaches to validate and extend these results.

Third, although the survey included a range of demographic and socioeconomic measures, it did not capture all participant characteristics that may shape responses to cumulative risk messages, such as literacy, first language, pre-existing health conditions, pregnancy status, or outdoor occupation. As a result, the findings may not fully account for important differences across subpopulations. The research team examined proxies such as education and income and did not find meaningful differences in experimental results across these groups, but we recognize that these are not substitutes for direct measurement. This limitation highlights the need for future research that explicitly examines how these characteristics influence the impact of messaging strategies.

Finally, the study focused on a limited set of hazards and message formats, raising questions about how these findings might extend to other risks or communication contexts. We also acknowledge that the cumulative risks practitioners encounter in practice are often more complex than those tested in our public survey, which further constrains generalizability. Future work should expand the range of message types, include more varied and representative samples of practitioners and publics, and test cumulative risk communication in real-world settings. Doing so would deepen our understanding of when and how cumulative risk messages can be most effectively designed and deployed.

Implications and conclusions

Overall, the results of this study suggest that cumulative risk communication deserves greater attention. While the findings cast doubt on concerns that cumulative risk messages confuse or overwhelm the public, they also present a puzzle for application: the data provide no clear evidence that cumulative risk messages are more effective than single risk messages. What does this mean in terms of practical advice to practitioners?

The answer to this question is, in part, subjective. From the authors' perspective, the lack of evidence showing that cumulative risk messages are more effective should not be interpreted as a rationale for avoiding them. Rather, the predominantly null results presented throughout the study should offer reassurance that such messages can be used without compromising the public’s ability to interpret and respond to them. Practitioners should feel encouraged to explore cumulative risk communication in contexts where it reflects the realities that communities face. In doing so, communicators may foster a more comprehensive understanding of risk and help build public capacity to navigate complex environmental and public health challenges.

While easing practitioners’ concerns is an important first step, many other barriers to cumulative risk communication remain. Addressing them will require sustained, rigorous research and meaningful organizational change. Recall that cognitive constraints were only one of several barriers identified in the interviews. Time constraints and institutional barriers remain significant obstacles that scholars and practitioners have yet to fully address.

On the time side, additional research is necessary to explore the relative effectiveness of cumulative risk communication in crisis contexts, particularly when communities are most impacted by or concerned about a single hazard. It is possible that in these settings, the coalescence of limited time and limited public attention requires a more narrowly focused communication strategy. Future work should examine whether and how cumulative risk messages can be adapted for these high-pressure environments without sacrificing clarity, relevance, and public trust.

With respect to institutions, many practitioners continue to operate within single-risk silos that are deeply embedded in organizational structures and hierarchies. In such environments, coordinating across programs or agencies to communicate cumulative risk can be challenging and, in some cases, may even be actively discouraged. In the long run, meaningful organizational change may be necessary to fully address this challenge. In the short term, initiatives such as intra- and interagency working groups focused on cumulative risk communication may serve as an important first step—helping to build shared understanding, improve coordination, and support the development of integrated messaging across organizational boundaries. Notably, one interview participant indicated that such a working group was already underway within their organization and reported that it had been highly beneficial.

Last but certainly not least, a meaningful shift from single to cumulative risk communication will require practical resources and support for risk communicators. Practitioners often depend on well-tested message templates and outreach materials that are specifically designed to communicate individual risks effectively. Comparable resources for cumulative risk communication, however, remain largely unavailable, leaving practitioners without the tools needed to implement these strategies at scale. As the researchers who developed the cumulative risk messages for the experimental component of this study can attest, it is neither easy nor immediately obvious how to craft concise and effective messages that convey cumulative risks, their potential impacts, and the actions individuals can take to reduce those impacts. Just as it took decades of research, trial, and refinement to develop effective single risk messages, it will likely take sustained effort and collaboration to build a comparable foundation for cumulative risk communication.

Abbreviations

EPA

U.S. environmental protection agency

IPPRA

Institute for public policy research and analysis

NOAA

National oceanic and atmospheric administration

NWS

National weather service

FEMA

Federal emergency management agency

CDC

Centers for disease control and prevention

NIH

National institutes of health

IRB

Institutional review board

U.S.

United States

CPS

Current population survey

NPORS

National public opinion reference survey

PMT

Protection motivation theory

PRISM

Planned risk information seeking model

Authors’ contributions

All authors contributed to the conception and design of the study, data collection, analysis, and interpretation. All authors were involved in drafting the manuscript and revising it critically for important intellectual content. All authors read and approved the final manuscript.

Funding

This research was supported by the U.S. Environmental Protection Agency under the project titled Know Your Audience: Cumulative Climate Risk. The views expressed are those of the authors and do not necessarily reflect the official policies or positions of the U.S. Environmental Protection Agency.

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to restrictions related to participant confidentiality, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This research was reviewed and approved by the University of Oklahoma Institutional Review Board (IRB #17626) and conducted in accordance with the ethical principles outlined in the Belmont Report (U.S. National Commission for the Protection of Human Subjects, 1979) and consistent with the Declaration of Helsinki. Informed consent was obtained from all participants prior to their participation in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.