Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Climate Risk Perception and Environmental Threat Communication

Abstract. Climate risk perception concerns subjective judgments about the likelihood, severity, distribution, and personal or collective consequences of climate hazards, while environmental threat communication concerns how information about these hazards is conveyed, interpreted, trusted, and used. This critical narrative monograph integrates historical theories of risk and decision-making, mental-model research, psychological distance, affect, identity, source credibility, and action efficacy with contemporary preregistered experiments, cross-national studies, field trials, and meta-analyses. It distinguishes hazard exposure from perceived vulnerability, improved scientific comprehension from changed attitudes, and information-seeking behavior from verified preparedness. The synthesis finds that risk appraisals are multidimensional, that scientific-consensus messages reliably improve consensus understanding but have smaller downstream effects, that localized cues can increase engagement under specified conditions, and that climate communication interventions have modest heterogeneous average effects. It identifies major limits of single-wave self-report, behavioral proxy measures, population generalization, and causal mediation. The final research agenda emphasizes calibrated local risk, equitable actionability, independent replication, and transparent human–artificial information systems. This monograph presents a critical narrative synthesis of published evidence; it reports no original experimental data or independent systematic-review protocol.

Keywords: climate risk perception; environmental threat communication; climate change communication; environmental psychology; risk appraisal; mental models; psychological distance; extreme-weather experience; risk-as-feelings; political polarization; scientific consensus; misinformation inoculation; risk communication efficacy; preparedness; field experiments; environmental justice; climate warning design; Artificial Era

Climate Risk Perception as a Scientific Research Object

Climate risk perception concerns judgments about the likelihood, severity, temporal distribution, geographic reach, and personal or collective consequences of climate-related hazards. These judgments include beliefs about processes already underway, probabilities assigned to future events, perceived vulnerability, the capacity to respond, and trust in institutions supplying information. Environmental threat communication concerns the production, interpretation, circulation, and consequences of messages about these hazards. The two processes are reciprocal: experience, social affiliation, and prior belief shape the reception of climate information, while communication changes the categories through which later evidence is interpreted. The psychological object is therefore a measurable sequence linking hazard information, appraisal, affect, comprehension, intention, and action rather than a single attitude toward the environment Slovic (1987) Weber (2006) Pidgeon (2012).

A scientifically usable definition separates an independently characterized hazard from a respondent's assessment of it. Heat exposure, inundation likelihood, fire weather, and air-quality warnings require physical measurements and modeled uncertainty; perceived likelihood, severity, controllability, and confidence require psychological measurement. A person may correctly identify the existence of anthropogenic warming while underestimating local flood exposure or overestimating the precision of a seasonal forecast. Conversely, pronounced worry may coexist with accurate beliefs about global hazards and uncertainty about feasible protective action. Construct validity depends on distinguishing these outcomes and tracking the reference period, location, target population, and comparison group implied by every question Bostrom et al. (1994) Sandlund et al. (2026).

The present monograph integrates original experimental work, longitudinal and cross-national surveys, conceptual and methodological research, and recent preregistered communication interventions. It assesses the inferential strength of each design, contrasts personal risk with societal risk, and distinguishes message comprehension from behavioral effects. Its scope covers chronic climate change and climate-related acute hazards insofar as they involve risk appraisal and communication. The clinical assessment of climate anxiety, the engineering of adaptation infrastructure, and disaster recovery as a collective institutional process are neighboring research objects. The final synthesis treats information systems and generative artificial agents as emerging parts of the communication environment and specifies testable research questions rather than claiming settled psychological effects.

Historical Emergence of Climate Risk Research

The history of climate risk perception joins decision research with the communication of environmental science. Tversky and Kahneman's account of judgment under uncertainty identified heuristics that can influence probability estimates when direct statistical evidence is difficult to use. Slovic's psychometric risk research subsequently demonstrated that judgments of risk include qualitative dimensions such as dread, perceived control, and familiarity. These frameworks were developed across different problems and populations, and their application to climate change requires specifying how a diffuse, cumulative, and geographically heterogeneous hazard differs from an immediately encountered threat Tversky & Kahneman (1974) Slovic (1987).

Bostrom, Morgan, Fischhoff, and Read's 1994 mental-models study examined public understandings of global climate change and documented gaps and confusions in causal explanations. Its lasting contribution was methodological: it made comprehension of causal mechanisms a research problem rather than assuming that awareness of a named phenomenon implies an accurate explanatory model. In 2006, Weber analyzed the differences between experience-based and description-based representations of long-term risks, while Leiserowitz examined affect, imagery, and values in American climate risk perceptions. These works helped place lived experience, symbolism, and social interpretation alongside formal probabilities in psychological accounts Bostrom et al. (1994) Weber (2006) Leiserowitz (2006).

The following research period introduced psychological distance, partisan identity, personal experience, and climate-message experimentation as central variables. Whitmarsh investigated flood experience; O'Neill and Nicholson-Cole tested reactions to visual representations; Spence and colleagues distinguished temporal, spatial, social, and hypothetical distance; Kahan and colleagues examined the interaction of cultural identity with scientific literacy and numeracy. Later multicountry interventions and meta-analyses tested which patterns extend beyond particular samples and contexts. This trajectory moves the field from descriptive public-opinion measurement toward experimental designs that compare specific communications, isolate mechanisms, and recognize heterogeneous outcomes Whitmarsh (2008) O’Neill & Nicholson-Cole (2009) Spence et al. (2012) Kahan et al. (2012) Većkalov et al. (2024).

Climate Hazards, Exposure, Vulnerability, and Risk Appraisal

Environmental risk is produced through relations among physical hazards, exposure, vulnerability, and protective capacity. Psychological studies need to preserve these distinctions because a respondent can interpret a warning as describing a severe hazard while believing that personal exposure is low. For example, a map may communicate a high projected frequency of extreme heat in a region, yet the experienced personal threat depends on housing insulation, occupation, age-related vulnerabilities, access to cooling, and opportunities for relocation. Risk appraisal is consequently situated within material circumstances. A survey that asks only whether climate change is dangerous cannot distinguish low hazard awareness from low exposure, high coping confidence, or a divergent meaning of the question Pidgeon & Fischhoff (2011) Nobel & Hiscox (2026).

The temporal character of climate change adds another difficulty. Climate risk encompasses shifts in distributions of events rather than a prediction that every particular storm or fire must occur. A message describing an increase in the probability of an event can be misread as certainty about individual outcomes, while communication of uncertainty can be mistaken for ignorance about the underlying process. The psychological distinction between probability, confidence, variability, and the quality of evidence is crucial. Researchers should test whether recipients understand the denominator of a risk estimate, the time window, and the difference between a regional forecast and an address-level assessment Bostrom et al. (1994) Pidgeon & Fischhoff (2011).

Perceived severity is also multidimensional. A threat can implicate personal health, family safety, property, local ecological conditions, livelihood, and collective institutions; each becomes salient under different framing and social conditions. Anticipated harm, perceived irreversibility, moral concern, and feasibility of mitigation or adaptation may contribute differently to a reported risk score. For that reason, complete risk assessment records hazard type and psychological target rather than treating 'climate concern' as a unitary measure. Comparing outcomes across studies requires attention to whether they measured belief in warming, probability of harm, subjective worry, support for a policy, or adoption of a protective behavior Leiserowitz (2006) van der Linden (2015).

Mental Models, Causal Understanding, and Scientific Literacy

A mental model is an organized representation of how a system works, including beliefs about causes, intermediate processes, feedback, and possible interventions. Climate models held by non-specialists may combine accurate insights with simplified analogies and associations unrelated to greenhouse forcing. Bostrom and colleagues used interviews and elicitation to investigate such representations. Their work illustrates why a communication study should assess causal explanation directly—for example, how carbon dioxide affects heat retention—rather than infer understanding from agreement with a general statement that the climate is changing Bostrom et al. (1994).

Scientific literacy can improve access to probabilistic and mechanistic reasoning, yet its relationship with climate risk is moderated by social context. Kahan and colleagues reported stronger polarization on climate risk among more science-literate and numerate respondents in the particular American samples and cultural contexts they studied. This is evidence about an interaction among reasoning resources, identity, and the topic's social meaning; it does not establish that knowledge causes denial or that scientific literacy is uniformly counterproductive. Communication research must identify what kind of knowledge is measured, how it is embedded in affiliation, and whether the outcome is factual comprehension, personal risk appraisal, or public-policy preference Kahan et al. (2012) Hornsey et al. (2016).

Knowledge correction and conceptual learning can be studied using open explanations, targeted misconception tests, diagrams, structured interviews, and delayed recall. A sound design compares baseline causal models with post-message understanding and tests whether gains transfer to unfamiliar scenarios. Participants should be able to explain a risk mechanism in their own words, recognize a misleading causal inference, and distinguish scientific confidence from forecast precision. Repeated measurement also helps detect answers that reflect short-term agreement with the experimenter's language. Mental-model research is thus a necessary complement to standard agreement scales and message-liking measures Bostrom et al. (1994) Moser (2010).

The Architecture of Subjective Climate Risk

Risk appraisal is a structured psychological outcome with multiple components: perceived probability, anticipated loss, vulnerability, personal relevance, controllability, efficacy, affective salience, and judgments of institutional credibility. Individuals can give different answers to questions that emphasize any one of these components. The risk-as-feelings tradition places affective imagery and intuitive evaluation within a broader decision process, while descriptive judgment research shows why numerical probabilities and personally encountered episodes need not be weighted alike. Weber's distinction between description and experience explains a central climate communication difficulty: statistical knowledge about a future distribution may compete with the salience of familiar weather and recent episodes Weber (2006) Leiserowitz (2006).

A comprehensive explanatory model should specify paths rather than present a list of correlates as equivalent causes. Prior experience may change perceived vulnerability, attention, and emotional salience; perceived vulnerability may increase information seeking; social identity may influence the credibility assigned to a messenger; perceived ability to act may condition whether concern is translated into preparation. Survey models can estimate associations among these variables but generally cannot establish the ordering of causal processes. The integrated model proposed by van der Linden organizes experiential, sociocultural, cognitive, and sociodemographic predictors, offering a useful research map whose predictive performance must be assessed outside its original measurement context van der Linden (2015) Sambrook et al. (2021).

The scientific synthesis therefore treats subjective climate risk as a family of operationally distinct appraisals with partially shared determinants. When investigators report an effect on risk perception, the claim is complete only after identifying the measurement scale, baseline distribution, message exposure, follow-up interval, and experimental contrast. The same treatment might increase perceived seriousness of a collective hazard while leaving personal vulnerability, protective action, and trust unchanged. Such differential effects are theoretically informative; they identify which component of appraisal has actually moved and constrain claims about practical effectiveness Većkalov et al. (2024) Herberz et al. (2026).

Heuristics, Availability, and Experience of Extreme Events

Extreme weather episodes make climate hazards perceptually concrete through vivid images, disrupted routines, and memories of loss. Availability-based judgment predicts that easily retrieved examples can influence perceived frequency and risk, particularly when a relevant reference class is difficult to enumerate. Yet an experienced storm, wildfire, or heat wave is evidence of the event experienced; attribution of that particular event and estimation of long-term climate trends require additional scientific information. Research therefore distinguishes exposure to an event, recollection of it, the subjective appraisal of consequences, and beliefs about the contribution of climate change Tversky & Kahneman (1974) Weber (2006) Sambrook et al. (2021).

Whitmarsh's investigation of flood experience indicated that the relation between direct experience and generalized climate beliefs is contingent rather than automatic. Myers and colleagues subsequently examined perceived personal experience in relation to belief in global warming, while van der Linden studied the interrelations of personal experience, affect, and risk appraisal. Reviews of the literature emphasize variation in measurement, attribution, hazard type, and baseline belief. Surveys that ask respondents whether they have personally 'experienced climate change' partly measure an interpretive judgment and can inflate associations between experience and belief because both are reported by the same person Whitmarsh (2008) Myers et al. (2013) van der Linden (2014) Sambrook et al. (2021).

Improved research links objectively dated exposure records with repeated surveys gathered before and after events, accounts for migration and selective residence, and uses independent climate attribution where relevant. Immediate postevent reactions should be compared with longer-term trajectories because salience can decay, become politicized, or be integrated into lasting memories. Experience may motivate precaution when respondents judge protective action to be feasible, but it may also produce habituation, fatalism, or confidence derived from having survived earlier events. The sign of the effect is an empirical question for each hazard and population Wong-Parodi & Garfin (2022) Sambrook et al. (2021).

Psychological Distance, Temporal Horizons, and Geographic Relevance

Climate change is often presented through distant places, future generations, and abstract global averages. Psychological distance describes the perceived remoteness of events along spatial, temporal, social, and hypothetical dimensions. Spence, Poortinga, and Pidgeon investigated these dimensions in climate change judgments and found meaningful associations with concern and willingness to respond. The dimensions nevertheless require separate operationalization. A person may regard climate harms as happening now, view them as most severe abroad, and believe personal control is limited. Such a combination has no single numerical distance Spence et al. (2012).

Messages that make risk local can increase relevance when their geographic details correspond to the recipient's actual conditions. The scientific requirement is fidelity: a neighborhood flood map with coarse resolution may convey an impression of precision the underlying models cannot justify. Proximity also interacts with economic position, mobility, residence, social identity, and past disaster experience. Simply replacing 'the planet' with 'your neighborhood' does not establish the efficacy of a communication strategy; investigators need to compare comprehension, credibility, perceived vulnerability, and real decisions under randomized conditions Pidgeon & Fischhoff (2011) Nobel & Hiscox (2026).

The preregistered field work of Nobel and Hiscox examined location-proximal cues in personalized disaster-preparedness messaging to Australian homeowners. Localized material increased the odds of engagement with a preparedness resource compared with a more generic presentation. This is informative behavioral evidence because it records an observable click in a natural setting. It is also bounded evidence: clicking a resource does not demonstrate that a household revised its emergency plan, purchased equipment, reduced exposure, or experienced less damage. Behavioral outcome definitions determine how much practical significance a localizing effect can support Nobel & Hiscox (2026).

Self–Other Comparisons and the Interpretation of Optimism

Climate risk assessments frequently distinguish anticipated consequences for oneself, one's community, one's country, and people around the world. Sandlund and colleagues' meta-analysis synthesized self–other comparisons across 83 effect sizes, encompassing more than 70,000 participants from 17 countries. Its pooled difference showed that respondents tended to judge climate-related risk as greater for others than for themselves. This pattern documents a reliable average discrepancy in subjective comparative reports. It does not alone identify an incorrect estimate: personal and others' real exposures, wealth, resilience, and adaptation opportunities can differ in legitimate ways Sandlund et al. (2026).

A claim of unrealistic optimism requires an objective comparison with relevant individual risk, not merely a difference between assessments for self and a vaguely defined 'average other.' Respondents may imagine particularly exposed populations when asked about other people; the wording and scale of questions also determine which reference group is mentally represented. Cross-national samples introduce variation in household vulnerability, insurance, infrastructure, geographic location, and confidence in warnings. A meta-analytic mean can conceal substantial heterogeneity, and prediction intervals may permit weaker or reversed effects in some settings Sandlund et al. (2026).

The methodological priority is to couple calibrated hazard probabilities and individualized exposure data with a standardized target comparison and a repeated measure of perceived risk. Researchers can then examine whether discrepancies persist after controlling for objective differences and whether a message about personalized exposure produces better calibration. A message that equalizes two survey ratings is not necessarily an improvement if it compresses meaningful variation. Useful communication aligns perceived threat with a defensible evidence base, conveys irreducible uncertainty, and supports actions whose expected benefits are plausible Pidgeon & Fischhoff (2011) Nobel & Hiscox (2026).

Affect, Imagery, Worry, and the Regulation of Attention

Affective appraisal gives uncertain climate threats motivational and attentional significance. Leiserowitz's research linked climate-related images, evaluative feelings, and values with perceived risk and policy preferences. Imagery may include personal experiences of drought, a threatened coastline, destroyed homes, or a more abstract conception of planetary change. Affective responses are informative components of judgment, but research must separate the intensity of worry from the estimated probability of an event, from knowledge about its causes, and from action intentions Leiserowitz (2006) Weber (2006).

Fear-based communication illustrates why psychological mechanisms and outcomes should be distinguished. Witte's extended parallel process model proposed that threat information interacts with beliefs about response efficacy and self-efficacy: credible severity and vulnerability may focus attention, while low perceived ability to act can redirect coping toward disengagement. O'Neill and Nicholson-Cole's climate visualization research found that threatening images can attract attention yet may be less constructive for sustained engagement than representations connecting risk with meaningful responses. These results support studying threat and efficacy jointly, while their designs and contexts do not justify a universal rule that fear always helps or always backfires Witte (1992) O’Neill & Nicholson-Cole (2009).

Affective communication can also obscure disparities when dramatic imagery turns distant suffering into a symbol without specifying practical obligations and local consequences. Strong emotion may increase recall but impair the intended outcome if recipients interpret the message as manipulative, repetitive, or irrelevant. Evaluation should therefore include message comprehension, emotional response, perceived agency, delayed retention, avoidance behavior, and feasible protective actions. Distress intensity is not a surrogate endpoint for successful communication, and clinically significant climate anxiety is addressed by its own specialist research literature Herberz et al. (2026) O’Neill & Nicholson-Cole (2009).

Trust, Source Credibility, and Epistemic Authority

Climate risk information is received through relationships with scientists, public agencies, local leaders, news organizations, community networks, and acquaintances. Perceived expertise, integrity, shared values, transparency, and procedural fairness shape which evidence is accepted as relevant. Trust can facilitate accurate interpretation of complex model-based warnings when citizens cannot independently reconstruct the analyses; it can also create vulnerability to persuasive misinformation from trusted but poorly informed sources. An adequate psychological account specifies trust in which source, on which subject, with what expectation of competence and accountability Pidgeon & Fischhoff (2011) Moser (2010).

Risk communication entails more than transmission of accurate statements. Recipients must identify what the source knows, which conclusions follow from the evidence, what remains uncertain, and whether the recommended action is within their control. Deliberative and participatory approaches invite questions and local knowledge, potentially improving the relevance and acceptability of risk information. Their effects on comprehension and preparedness require evaluation rather than inference from a favorable conception of participation. A transparent source can state uncertainty and still be credible when its claims and revisions are traceable Pidgeon & Fischhoff (2011) Pidgeon (2012).

Trust measurement should distinguish confidence in climate science, trust in a particular messenger, belief that local authorities will issue timely warnings, and confidence that institutions will protect vulnerable populations. Such beliefs may show different relationships to household behavior. A useful intervention randomizes messenger identities while keeping factual content constant, tests interactions with prior identity and experience, and monitors whether gains persist once the recipient encounters conflicting claims. The long-term criterion is informed and usable decision-making, including the capacity to recognize changed evidence McCright & Dunlap (2011) van der Linden et al. (2017).

Political Identity, Cultural Worldviews, and Motivated Reasoning

Climate change is a scientifically investigated physical phenomenon whose public interpretation is also embedded in political and cultural life. Political parties, identity groups, media environments, and cultural worldviews provide cues about what claims are credible and which actions appear legitimate. McCright and Dunlap documented increasing partisan polarization in United States views of global warming over the 2001–2010 period. Their analysis establishes a historically situated pattern in public opinion and illustrates why a message that is effective in one cultural environment may have a different reception in another McCright & Dunlap (2011).

Motivated reasoning offers a mechanism linking social affiliation to the selection, interpretation, and recall of evidence. Kahan and collaborators reported that polarization concerning climate risk could be greater among respondents with higher science literacy and numeracy within particular cultural groupings. This finding challenges a simplistic information-deficit account that assumes additional facts alone will close every disagreement. It is equally important to retain the competing possibility that education, topic knowledge, trusted local experience, and institutional environments exert positive effects under other conditions. Both mechanisms must be tested with designs capable of measuring change rather than inferred from cross-sectional correlations Kahan et al. (2012) Hornsey et al. (2016).

Causal studies of identity-sensitive messaging should vary the framing or source while holding the factual claim constant and measuring attitudes before exposure. Researchers must examine effects separately by prior beliefs, political identity, relevant local hazards, and baseline trust; a pooled average can hide opposing directional responses. Ethical communication does not require concealing a scientific conclusion to make it acceptable. It requires explaining evidence accurately, identifying sources, allowing legitimate questions, and evaluating whether communication supports factual understanding and constructive decisions across groups McCright & Dunlap (2011) Pidgeon (2012).

Public Belief, Concern, and the Limits of Attitude Indicators

Belief that global temperatures are rising, attribution of warming to human causes, perceived societal danger, personal worry, and support for climate policy are related but different psychological indicators. Hornsey and colleagues' meta-analyses combined research on antecedents and consequences of climate-change belief and found that ideological commitments, values, and other individual differences are associated with belief patterns. Their synthesis clarifies the dimensions of the literature while remaining subject to the limitations of correlational designs and variable questionnaire formulations Hornsey et al. (2016).

Measurement equivalence becomes especially important across countries and languages. A survey question asking whether climate change is 'a threat' can evoke immediate mortality, economic disruption, institutional instability, or harm to ecosystems. Variations in political language and hazard exposure complicate comparisons of mean scores. Researchers should evaluate translation, item interpretation, factor structure, and response styles before treating rankings as direct differences in underlying psychological risk. A high mean for collective concern cannot be read as individual knowledge of local adaptation procedures van der Linden (2015) Većkalov et al. (2024).

Communication research should use multidimensional outcome batteries including factual understanding, risk probability estimates, perceived severity, trust, policy support, readiness, and documented action where feasible. Registering primary outcomes in advance helps distinguish expected effects from exploratory subgroup findings. A study that finds a significant change in one of many attitude measures while all behavioral outcomes remain unchanged supplies evidence of a limited attitudinal effect, not proof that the intervention caused social transformation. Such precision is essential when findings inform public institutions Herberz et al. (2026) Vlasceanu et al. (2024).

Scientific Consensus as a Communication Target

The scientific consensus about human-caused climate change is an empirical object of public knowledge. Communication studies can test whether accurately representing expert agreement changes estimates of consensus, belief about causation, perceived risk, or support for collective action. Van der Linden and colleagues' experimental work on the gateway belief model found that consensus messages could influence perceived agreement among scientists and related attitudes. The proposed mechanism is a pathway from perceived scientific agreement to other climate judgments; its causal and practical scope must be evaluated across designs van der Linden et al. (2015).

A major cross-national test by Većkalov and collaborators used a preregistered experiment across 27 countries. It found a sizable positive shift in accurate perceptions of scientific consensus, accompanied by much smaller average shifts in some beliefs, and little evidence of direct improvement in climate worry or policy support. This comparison matters because recognition that scientists agree is not equivalent to changing one's perceived capacity to act or a concrete policy preference. It also shows why a valid message may achieve an informational goal even when an ambitious behavioral endpoint does not change Većkalov et al. (2024).

Rode and colleagues' 2025 meta-analytic structural equation analysis addressed modeled paths between consensus communication, gateway beliefs, and public action support. Indirect pathways in an aggregate model can coexist with experiments reporting weak direct downstream effects; they answer different inferential questions. A mediation model depends on assumptions about variable ordering, confounding, and measurement and should not be taken as an automatic estimate of real-world policy change. Scientific consensus messages are most defensible when evaluated for accurate knowledge acquisition and when linked to separately tested mechanisms for action Rode et al. (2025) Većkalov et al. (2024).

Misinformation, Prebunking, and Corrective Communication

Climate communication takes place in an information ecosystem in which misleading claims can circulate alongside legitimate scientific disagreement, uncertainty, and public debate about policy. Misinformation can distort judgments by replacing a causal account with an inaccurate alternative, by selectively displaying evidence, or by presenting isolated uncertainty as if it undermined an entire body of research. Psychological work therefore studies not only whether a false statement is believed but also its effect on confidence, trust, causal interpretation, and subsequent correction. Corrective communication is strongest when the error is named precisely and the scientific explanation is supplied in an accessible form Bostrom et al. (1994) van der Linden et al. (2017).

Inoculation or prebunking exposes recipients to a weakened misleading technique and an explanation of why it misleads before later exposure to a stronger claim. Van der Linden and colleagues experimentally investigated inoculation against climate misinformation, finding that forewarning and counterargument material could protect aspects of the understanding of scientific consensus. This supports the plausibility of a preventive process in a specified task. Its ecological effectiveness depends on the novelty and volume of later misinformation, timing, source credibility, and the outcome used to define resistance van der Linden et al. (2017).

Future trials should distinguish correcting an immediate quiz response from preserving beliefs over weeks or months in ordinary media environments. Randomized studies can compare short factual corrections, causal explanations, inoculation, and simple repetition of accurate information, recording trust and possible unintended effects. Researchers should preregister moderation by political identity and baseline belief while guarding against excessive exposure to false claims. A transparent intervention specifies which part of the message is demonstrably false and which portion reflects scientific uncertainty or a legitimate disagreement about values and policy Pidgeon & Fischhoff (2011) van der Linden et al. (2017).

Framing, Narrative, Messenger, and Audience Segmentation

A climate risk message can frame the same factual evidence in terms of health, local livelihoods, economic losses, ecological consequences, community safety, or future generations. Framing alters which consequences are made salient and which comparisons seem appropriate. Narrative communication can give a temporal structure to causes and effects, while numbers and visualizations communicate magnitudes and uncertainty. Message design therefore constitutes a set of experimental variables rather than a single technique called 'better communication' Moser (2010) O’Neill & Nicholson-Cole (2009) Pidgeon & Fischhoff (2011).

Audience segmentation attempts to organize variation in prior beliefs, concerns, trust, and readiness for engagement. The Six Americas tradition classified American audiences into differing climate-attitude segments, illustrating one practical approach to tailoring communication. Such groupings are descriptive and may change over time; the labels are not stable psychological diagnoses and do not establish that a chosen message is effective for every member of a segment. Segmentation requires replication, attention to within-group diversity, and independent tests of whether targeting outperforms a broad evidence-based message Leiserowitz et al. (2021) Hornsey et al. (2016).

Message tailoring also raises equity questions. Digital systems may offer fine-grained personalization, but highly individualized content requires reliable data, meaningful consent, and safeguards against opaque persuasion. Researchers should distinguish personalization by actual hazard exposure from targeting based on inferred vulnerability or identity. A scientifically responsible communication system makes the underlying facts, uncertainties, and proposed protective actions inspectable. Intervention success is determined by accurate comprehension and beneficial decisions in different populations, including those less represented in online experimental panels Nobel & Hiscox (2026) Herberz et al. (2026).

Risk Communication Design and the Psychology of Efficacy

A warning has practical value when it enables a recipient to recognize a relevant hazard and identify an available course of action. Efficacy is therefore a separate scientific construct from perceived threat. Response efficacy concerns whether a recommended action is expected to reduce harm; self-efficacy concerns whether the recipient believes that they can perform it under actual conditions. A resident may regard extreme heat as very dangerous yet lack safe access to cooling or employment flexibility. Communication that heightens concern without describing attainable options can leave the causal barrier untouched Witte (1992) O’Neill & Nicholson-Cole (2009).

Climate communication strategies vary in the information they prioritize: projected severity, personal exposure, a feasible protective response, descriptive norms, institutional responsibility, or collective benefits. These features may interact. For example, a personalized flood probability can increase attention while a trusted local source increases acceptance and explicit steps support preparedness. Each mechanism should be measured independently. Messages directed at structural hazards should also identify institutional responsibilities, because assigning an individual the task of overcoming unavailable infrastructure can misrepresent the actual capacity for harm reduction Pidgeon & Fischhoff (2011) Nobel & Hiscox (2026).

The extended parallel process model predicts that threat and efficacy jointly affect engagement. This is a general theoretical model with applications to climate messaging; whether a particular danger/efficacy combination succeeds is an empirical question. A factorial experiment can vary severity information and response guidance separately, with immediate and delayed outcomes for comprehension, perceived control, information seeking, and observed protection. Measuring these variables permits explanation of divergent findings that would otherwise be summarized vaguely as fear appeals working or failing Witte (1992) Herberz et al. (2026).

Visualization, Probabilistic Information, and Forecast Literacy

Climate risks frequently reach the public as probability ranges, hazard maps, model ensembles, color scales, scenarios, and photographs. Each representation demands different interpretive operations. A numerical probability of flooding within a specified time horizon is a statement about an event class and model assumptions; a colored map may invite an interpretation of precise neighborhood boundaries; a photograph can evoke vivid consequences without conveying frequency. Visual format is therefore part of the scientific content that recipients actually process Bostrom et al. (1994) O’Neill & Nicholson-Cole (2009) Pidgeon & Fischhoff (2011).

Comprehension tests should ask respondents to reconstruct a risk statement rather than only rate the graphic as convincing. For probabilistic messages, investigators can test denominator understanding, time horizon, comparison with baseline probability, and confidence interval interpretation. For maps, they can assess the user's ability to find the relevant region, distinguish model resolution from address-specific certainty, and recognize the limits of the underlying exposure data. Alternative displays should be compared with randomized assignment and task-based accuracy outcomes, including performance among people with lower numeracy or visual accessibility needs Pidgeon & Fischhoff (2011) Weber (2006).

Visual communication must also account for emotional meaning. O'Neill and Nicholson-Cole reported that dramatic or threatening climate images could produce attention without fostering constructive engagement. Their findings motivate a design principle of linking legitimate hazards to context, agency, and specific interpretation tasks. Yet imagery research is heterogeneous: photo selection, message text, population, exposure duration, and chosen endpoint can all change measured effects. The strongest guidance emerges from comparisons that preserve factual equivalence while varying graphic form, assess delayed comprehension, and document harms such as confusion or unwarranted certainty O’Neill & Nicholson-Cole (2009) Herberz et al. (2026).

Personalization, Local Warnings, and Field Experiments

Personalization can make a hazard forecast meaningful to the location and decisions of a particular household. Its effectiveness depends on both the accuracy of the underlying exposure estimate and the actions to which a person can reasonably respond. An address-based reminder about heat or flood risk differs from a generic global warming message because it makes an immediately relevant decision available. Research must therefore distinguish the informational intervention from the infrastructure and resources that make protective action possible Nobel & Hiscox (2026) Wong-Parodi & Garfin (2022).

Nobel and Hiscox's preregistered field experiment in Australia compared localized and more generic climate-disaster preparedness messaging among homeowners. A proximal cue increased the odds of clicking through to a preparedness resource, with the reported odds ratio exceeding two in that specified comparison. This is stronger evidence of observed attention or engagement than a post-message intention rating because it records a choice in the field. Its substantive boundary remains the measured click; later readiness, property protection, insurance behavior, or emergency outcomes require additional data Nobel & Hiscox (2026).

A program evaluation can extend this logic with staged outcome measurement: message delivery, opening or viewing, accurate interpretation, preparation planning, completion of recommended steps, and performance during an actual warning. Randomized or quasi-experimental designs should record reach and attrition, compare socioeconomic and hazard-exposure groups, and avoid conditioning the main analysis only on participants who voluntarily engage. Personalized systems also require a clear account of data provenance, privacy, update frequency, and what happens when forecasts change. The defensible objective is accurate, equitable, and actionable local risk communication Pidgeon & Fischhoff (2011) Nobel & Hiscox (2026).

Intervention Meta-Analysis and the Size of Communication Effects

The cumulative effectiveness of climate communication cannot be established by selecting a single vivid demonstration. Herberz and colleagues' systematic review and meta-analysis synthesized experimental communication interventions across numerous publications, studies, and outcome measures. The average standardized effect was small and heterogeneity was substantial, indicating variation in designs, populations, interventions, and measurements. Such an aggregate provides a useful benchmark while leaving open the identity of the communication mechanisms that reliably produce consequential change in specific settings Herberz et al. (2026).

The evidence base also presents reporting and generalization challenges. One-time online exposures are convenient for experimental control, but ordinary risk communication unfolds across repeated messages, relationships, conflicting information, and changing hazards. An intervention can have a measurable effect immediately after exposure that dissipates before a protective decision arises. Meta-analytic averages may blend belief change, intentions, message ratings, and observed behavior; an intervention that improves one category cannot be credited with improvements in every other category. Reviews also require attention to publication bias, preregistration, dependence among effect sizes, and the date range of the literature search Herberz et al. (2026).

The practical unit of synthesis is therefore a clearly defined intervention–population–outcome–time contrast. Effects on beliefs may be worth reporting independently from effects on preparedness or policy support. Researchers should interpret standardized effects alongside absolute changes, cost, reach, expected damage reduction, and equity consequences whenever those quantities are measured. Consistent small effects across large populations might matter, whereas a seemingly large self-reported intention effect may lack demonstrated real-world transfer. The best-supported recommendation is to build evidence for individual communication decisions rather than presume a universal persuasive formula Herberz et al. (2026) Vlasceanu et al. (2024).

Global Behavioral Interventions and Cross-National Heterogeneity

The International Climate Psychology Collaboration coordinated a multicountry experimental test of behavioral interventions across 63 countries. Vlasceanu and colleagues evaluated multiple interventions against distinct climate outcomes, including beliefs, policy support, willingness to share information, and a more effortful behavioral task. The research provides unusual breadth and an opportunity to examine treatment-by-context variation within a standardized program. The results do not support a simple ranking in which one message universally improves every intended outcome Vlasceanu et al. (2024).

Some treatments affected particular attitudinal or sharing outcomes, while stronger behavioral endpoints were less responsive. Differences by country, baseline belief, and outcome complicate general claims about global transportability. Even with many participants, an intervention delivered through an online platform differs from municipal warning systems and sustained community programs. The correct inference is that design choices may shift measured responses under some conditions; adaptation to a different setting requires validation of the recipient population, implementation channel, and target decision Vlasceanu et al. (2024) Većkalov et al. (2024).

Large coordinated experiments also illustrate why theoretical interpretation must be disciplined. An effect on information sharing could arise from improved comprehension, identity congruence, heightened emotion, or perceived social approval; the experimental assignment identifies the total intervention effect but cannot separate these mediators without additional measurements and assumptions. Cross-national differences may arise from unequal exposure, language translation, political context, or measurement noninvariance. Subsequent research should integrate field behavioral follow-up with preregistered mechanism tests and transparent country-level uncertainty Vlasceanu et al. (2024) Herberz et al. (2026).

Survey Measurement, Construct Validity, and Risk Calibration

Survey research is indispensable for estimating the distribution of climate beliefs and for studying differences among groups. Validity nevertheless depends on the construct named by each question. 'How concerned are you about climate change?' measures a broad evaluation; 'What is the probability that your home will flood within the next decade?' requests a temporally and geographically specified forecast. Respondents can answer either question consistently while using different reference classes, probability interpretations, and experiential cues. Measurement programs should report item wording, response anchors, translation procedure, time horizon, and whether respondents can state that they lack sufficient information van der Linden (2015) Sandlund et al. (2026).

Psychometric analysis can examine reliability, factor structure, test–retest stability, item response characteristics, and measurement invariance across languages and groups. However, internal consistency among similar worry questions cannot establish that the score measures calibrated hazard risk. Calibration requires comparing perceived probabilities with an independently justified benchmark, itself subject to uncertainty. In a changing climate, what counts as the reference distribution, forecast horizon, and ground-truth event rate needs explicit definition. Without a defensible benchmark, a statistically significant mean difference between groups is a difference in reported appraisal, not necessarily a difference in accuracy Bostrom et al. (1994) Pidgeon & Fischhoff (2011).

Mixed measurement systems offer stronger inference. A questionnaire can gather subjective probability, confidence, worry, local exposure belief, and perceived response efficacy; an independent geospatial or meteorological record can characterize relevant hazards; interviews can identify misunderstood causal mechanisms; subsequent behavior can indicate whether the information entered real decisions. The method supports separation of information deficiency, warranted vulnerability differences, and action barriers. Longitudinal studies can then distinguish temporary salience from durable recalibration Sambrook et al. (2021) Wong-Parodi & Garfin (2022).

Experimental Identification, Mediation, and Longitudinal Designs

Randomized message experiments permit causal inference about the assigned communication under the conditions studied. A simple design compares a clearly specified message with a control while holding other features constant. Factorial designs can separately vary local relevance, source identity, uncertainty wording, and response guidance, thereby identifying interactions that a packaged intervention obscures. Pretests should establish whether recipients notice the intended manipulation and whether factual meaning remains equivalent across experimental conditions. Outcomes must be chosen before inspection of the results to reduce flexible analytic choices van der Linden et al. (2015) Većkalov et al. (2024).

Mediation claims require special care. A message may change perceived expert consensus, which correlates with subsequent policy support, but observational associations among post-treatment variables can be confounded by prior belief and unmeasured interpretive processes. Structural equation models offer formal hypotheses about causal ordering; experiments with intermediate measures test whether those hypotheses are compatible with the data. They do not eliminate assumptions about mediator–outcome confounding and measurement error. Comparing model-derived indirect effects with direct experimental treatment effects helps identify where causal conclusions remain limited Rode et al. (2025) Većkalov et al. (2024).

Longitudinal and field designs add temporal and environmental realism but introduce different vulnerabilities. Participants may move, forget earlier messages, encounter competing explanations, or leave a panel selectively. Severe hazards generate nonrandom exposure that complicates comparisons between those affected and unaffected. Researchers can strengthen inference through prior baseline observations, pre-specified event windows, independent hazard records, sensitivity tests for attrition, and transparent missing-data methods. Strong climate communication evidence requires convergence across randomized experiments, observational cohorts, and verified behavioral traces, with each design contributing a distinct inferential strength Whitmarsh (2008) Wong-Parodi & Garfin (2022) Nobel & Hiscox (2026).

Preparedness, Mitigation, and the Belief–Action Relationship

Climate risk communication is often expected to encourage action, but the relevant actions differ in mechanism and timescale. Preparedness may involve emergency planning, warning subscriptions, home modifications, or obtaining cooling resources. Mitigation can involve energy use, transport, voting, collective organization, or institutional change. A belief about anthropogenic warming may be relevant to all of these, yet each action is also constrained by costs, habits, infrastructure, social norms, policy, and opportunities. Studies must specify their behavioral criterion before claiming that increased climate concern produces beneficial action Wong-Parodi & Garfin (2022) Vlasceanu et al. (2024).

Wong-Parodi and Garfin examined adaptation behavior in Florida and Texas in relation to hurricane experience and climate attribution. Their observational findings help frame hypotheses about the link between experienced hazards, causal beliefs, and preparedness, while the study design limits strong causal interpretation. The multi-country intervention tournament provides a complementary experimental contrast: changes in some attitudes did not consistently produce changes on a more demanding behavioral outcome. Together these designs illustrate why the evidence for belief change and the evidence for completed protective action should be reported separately Wong-Parodi & Garfin (2022) Vlasceanu et al. (2024).

Behavioral measurement can progress from verbal intention to a verified action, repeated practice, and ultimately protection under actual hazard conditions. Each transition introduces new confounding factors. An evacuation intention may fail to become evacuation because transport is unavailable; installing cooling may require a landlord's permission; voting for climate policies depends on electoral institutions and the option set. Communication should be evaluated together with the social and technical conditions that permit its recommended actions. This approach gives psychological evidence an appropriate place within wider risk governance Pidgeon & Fischhoff (2011) Herberz et al. (2026).

Social Inequality, Environmental Justice, and Communication Access

Exposure to climate hazards and capacity to respond are distributed unequally. Housing quality, outdoor work, disability-related access needs, income, insurance, mobility, language, and local infrastructure can shape who faces high risk and who can act on a warning. Psychological research should therefore interpret an apparent lack of preparedness in relation to available choices. Communication studies that measure only intention or personal responsibility may overlook the material constraints that determine behavior. Environmental justice also concerns whose hazard experience counts as knowledge in forecasting, planning, and public communication Pidgeon & Fischhoff (2011) Moser (2010).

Risk communication can be evaluated for distributive reach and procedural fairness. Who received the message, in what language, through which device or institution, and at what time? Could recipients understand the information and obtain the recommended protection? Did the warning identify uncertainty honestly without masking risks for less visible populations? Inclusive design tests readability, visual and auditory accessibility, culturally situated examples, and the actionability of advice for people with constrained resources. Equity assessment is a property of communication performance rather than an ornamental addition to an otherwise generic intervention Pidgeon (2012) Nobel & Hiscox (2026).

Participatory methods, community interviews, and co-designed message evaluation can reveal hazards that administrative averages fail to represent. These methods should be combined with credible exposure estimates and independent outcome assessment so that local testimony and physical risk models inform each other. Research also needs to avoid treating underserved groups as internally homogeneous; variations in location, age, occupation, disability, and trust require disaggregated analysis when sample sizes permit. A communication strategy succeeds when factual accuracy, comprehension, accessibility, and attainable protection are evaluated together Pidgeon & Fischhoff (2011) Herberz et al. (2026).

Age, Culture, and Developmental Differences in Climate Appraisal

The meanings of climate risk differ across life stages and social roles. Adolescents may understand climate threats through schooling, intergenerational responsibility, identity, and anticipated futures, while older adults may interpret them through cumulative local experience, health vulnerability, and memories of historical weather. Caregivers may consider threats to dependents, and occupational groups may focus on heat exposure, crop yields, or emergency-response duties. These are hypotheses about mechanisms that require age-specific measurement; they should not be transformed into universal descriptions of what members of an age group believe Spence et al. (2012) Hornsey et al. (2016).

Cross-national experiments show the value and limitations of standardized global assessment. Većkalov and colleagues found that consensus information was understood in numerous national settings while downstream attitudinal responses were smaller; Vlasceanu and colleagues found heterogeneous responses to different behavioral interventions. Such projects expand the evidence beyond one country yet cannot automatically represent populations excluded from the sampling frame. Online literacy, translation, and platform selection can alter measured susceptibility to a message. Country should be analyzed as a contextual variable with uncertainty, not treated as an explanation by itself Većkalov et al. (2024) Vlasceanu et al. (2024).

Developmentally and culturally informative studies combine age-appropriate comprehension tasks, careful translation, local exposure records, and multiwave observation. They compare the same psychological construct across groups only after assessing measurement equivalence. Special attention is warranted when information concerns hazards that disproportionately affect children, people with chronic health vulnerabilities, or communities facing limited infrastructure. The scientific aim is to determine which communication elements make accurate climate-risk information accessible and actionable under diverse life circumstances, rather than to assume that one emotional appeal or numerical display has a universal audience Pidgeon & Fischhoff (2011) Herberz et al. (2026).

Risk Communication in Digital and Artificial Information Environments

Digital communication alters the routes by which climate information reaches the public. Automated feeds, search systems, geographic alerts, social networks, and conversational artificial agents can change message order, credibility cues, personalization, and perceived authority. These mechanisms invite psychological research because users may treat a fluent explanation as an indication of expertise even when its underlying evidence has not been checked. Climate information presented through such systems should retain clear source attribution, forecast limitations, dates, and links to the relevant scientific and public-safety institutions Pidgeon & Fischhoff (2011) van der Linden et al. (2017).

A generative system can be experimentally compared with static materials on explanation accuracy, calibration, clarification-seeking, recall, and observed preparedness choices. Important manipulations include whether the agent explains probability with a defined reference period, marks uncertainty, identifies the jurisdiction of a warning, and refuses to invent address-specific forecasts. Research should also examine whether conversational personalization creates misplaced confidence, whether corrections update the user's mental model, and how users interpret an artificial messenger's apparent empathy or certainty. These are prospective research questions for the Artificial Era, not established conclusions from climate communication trials Bostrom et al. (1994) Nobel & Hiscox (2026).

The relevant unit of analysis is an accountable human–information-system relationship: data providers generate physical risk estimates; computational interfaces select and present information; users form appraisals; institutions retain responsibility for warning quality and feasible protections. Public agencies and platform designers can make this chain inspectable through stable source provenance, versioned warnings, known geographic resolution, transparent explanations, and independent outcome auditing. AI-mediated climate communication should be judged by the same scientific criteria as other communication: accuracy, comprehension, equity, calibrated confidence, and consequential behavior Pidgeon (2012) Herberz et al. (2026).

Contradictory Results, Replication, and Research Limitations

A recurrent problem in climate risk research is the apparent contradiction between positive effects on a proximal psychological endpoint and weak effects on a more demanding downstream behavior. Experiments on consensus messaging demonstrate that accurate recognition of scientific agreement can increase while policy support or worry changes little. Field studies of local warning cues can show more resource engagement without establishing actual preparedness. Meta-analyses report small average communication effects alongside substantial heterogeneity. These observations are compatible once researchers distinguish the stages and causal requirements of an information-to-action pathway Većkalov et al. (2024) Nobel & Hiscox (2026) Herberz et al. (2026).

Another tension concerns the relation between personal experience and climate beliefs. Some studies report associations between perceived experience and concern, yet experience may be defined through respondents' own interpretation of an event. Direct exposure is not randomly assigned, and people differ in prior beliefs, housing, local news, and hazard probability. Disentangling the components requires pre-event measurements, externally validated exposure, multiple follow-ups, and careful controls. Single-wave self-report comparisons can motivate hypotheses but should not be treated as experiments about the causal effect of extreme weather Whitmarsh (2008) Myers et al. (2013) Sambrook et al. (2021).

Replication priorities extend beyond repeating significant tests. Researchers should compare exact stimulus content, survey wording, treatment implementation, participant recruitment, geography, baseline belief, and follow-up delay. Negative and null outcomes require publication alongside positive effects; otherwise, the apparent literature overstates efficacy. Meta-analysts should handle non-independent effect sizes and outcome mixing explicitly, while reviewers should disclose that evidence retrieved from abstracts may not permit the same methodological assessment as evidence inspected in full text. These practices enable cumulative scientific knowledge about communication mechanisms Vlasceanu et al. (2024) Herberz et al. (2026).

A Comparative Evidence Synthesis Across Research Designs

The strongest established claim is that climate risk perception has multiple determinants and that the effectiveness of a message depends on what it is meant to change. Mental-model interviews demonstrate interpretive complexity; surveys identify broad patterns in values, imagery, ideology, and experience; experiments isolate the immediate effect of a specified message; field trials measure some real engagement; systematic reviews estimate average patterns and variability across studies. The value of an integrated review lies in assigning each claim to the design capable of supporting it Bostrom et al. (1994) Hornsey et al. (2016) Vlasceanu et al. (2024) Herberz et al. (2026).

Evidence for psychological distance and comparative risk shows that respondents can locate climate harms in socially and geographically differentiated ways, and often rate risks to others more highly than risks to themselves. The magnitude and normative interpretation of such comparisons depend on the question and the populations compared. Evidence for scientific-consensus communication supports improved recognition of expert agreement, with smaller and less consistent downstream consequences. Evidence for localized preparedness communication supports engagement under specific field conditions. The research corpus therefore yields a mosaic of mechanisms and bounded effects rather than one universal causal explanation Spence et al. (2012) Sandlund et al. (2026) Većkalov et al. (2024) Nobel & Hiscox (2026).

Across this mosaic, the recurring limitation is measurement of the wrong endpoint for the intended practical conclusion. Attitudes and intentions are legitimate psychological outcomes. When a study seeks to improve survival, housing protection, or institutional adaptation, those outcomes must be measured directly or be linked through validated intermediate processes. Reports should separately state the population, baseline hazard exposure, treatment, comparator, timing, effect uncertainty, and accessibility of the recommended action. This evidence discipline prevents a message's rhetorical force from becoming a substitute for demonstrating its function Pidgeon & Fischhoff (2011) Wong-Parodi & Garfin (2022) Herberz et al. (2026).

Future Directions in Climate Risk Perception Research

The next stage of research should test the calibration of personal and community risk judgments against increasingly localized, uncertainty-aware physical evidence. A longitudinal design can measure beliefs before a forecast, after receipt of a warning, and after the corresponding hazard season, while maintaining the distinction between what was forecast and what actually occurred. Combining physical exposure data with cognitive interviews can identify whether errors arise from misunderstanding a probability, distrusting the source, misidentifying the geographic reference, or lacking experience with comparable events. Such work would make psychological diagnosis more informative for communication design Bostrom et al. (1994) Sandlund et al. (2026).

Large-scale message trials should increasingly incorporate meaningful action outcomes, heterogeneous treatment effects, independent replication, and durable follow-up. Message components can be randomized separately to estimate the contributions of uncertainty explanation, local relevance, efficacy guidance, source trust, and accessible format. Studies should oversample populations exposed to high risk and low communication reach while reporting the limits of selection and generalizability. Trial registration, openly specified primary endpoints, and complete reporting of null effects are necessary for credible evidence accumulation Vlasceanu et al. (2024) Većkalov et al. (2024) Herberz et al. (2026).

Artificially mediated risk communication adds research questions about automated source selection, interactive explanation, personalized risk expression, and shifting epistemic authority. Evaluation should compare human-authored, algorithmically tailored, and conversational messages with equivalent scientific content, while monitoring hallucinated figures, excessive certainty, unequal targeting, and user understanding. Institutional governance and public accountability remain essential endpoints because the most sophisticated message cannot create cooling, shelter, transport, or protection where those resources are absent. The scientific program should connect accurate knowledge to material opportunities for safe action Nobel & Hiscox (2026) Pidgeon (2012).

Conclusion

Climate risk perception and environmental threat communication constitute a distinct scientific research field at the intersection of environmental psychology, judgment and decision-making, social psychology, risk analysis, and the communication of uncertain scientific knowledge. The foundational finding is that subjective risk is a differentiated appraisal: people evaluate likelihood, severity, relevance, responsibility, and response capacity through cognitive representations, experience, affect, values, social identity, and confidence in information sources. The causal history of any one judgment cannot be recovered from an isolated concern score. Valid explanation requires a defined hazard, a defined psychological outcome, and a method suited to the causal claim Slovic (1987) Bostrom et al. (1994) van der Linden (2015).

The contemporary evidence shows why communication must be evaluated at several levels. Accurate consensus messages can improve factual recognition of scientific agreement; personally relevant warnings can increase engagement with preparedness information; carefully designed interventions can shift some attitudes. Average effects on stronger behavioral endpoints are less uniform, and large multicountry studies and meta-analyses document substantial contextual variation. Scientific progress depends on reporting these differences precisely, with due attention to sampling, replication, baseline beliefs, measurement validity, and behavioral follow-through Većkalov et al. (2024) Nobel & Hiscox (2026) Vlasceanu et al. (2024) Herberz et al. (2026).

A mature climate-risk communication system joins verified physical information, understandable uncertainty, legitimate public trust, accessible pathways to action, and evaluation of actual protective outcomes. Environmental injustice and institutional capacity shape which recommendations are feasible; artificial information systems reshape how risk explanations are encountered and interpreted. In the Artificial Era, these transformations extend the psychological field toward the study of human–artificial communication without changing its evidentiary obligations. The durable task is to improve the correspondence among what is known, what is understood, and what people and institutions can do to reduce harm Pidgeon & Fischhoff (2011) Sandlund et al. (2026).

References

  • Bostrom, A., Morgan, M. G., Fischhoff, B., & Read, D. (1994). What do people know about global climate change? 1. Mental models. Risk Analysis, 14(6), 959–970. https://doi.org/10.1111/j.1539-6924.1994.tb00065.x
  • Herberz, M., Uenal, F., Gouiran, M. D., Krainz, M., Pagel, L., Schmid, D., Spampatti, T., Vallaeys Mora, I., & Brosch, T. (2026). A systematic review and meta-analysis of communication strategies to promote climate action. Journal of Environmental Psychology, 114, 103176. https://doi.org/10.1016/j.jenvp.2026.103176
  • Hornsey, M. J., Harris, E. A., Bain, P. G., & Fielding, K. S. (2016). Meta-analyses of the determinants and outcomes of belief in climate change. Nature Climate Change, 6, 622–626. https://doi.org/10.1038/nclimate2943
  • Kahan, D. M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L. L., Braman, D., & Mandel, G. (2012). The polarizing impact of science literacy and numeracy on perceived climate change risks. Nature Climate Change, 2, 732–735. https://doi.org/10.1038/nclimate1547
  • Leiserowitz, A. (2006). Climate change risk perception and policy preferences: The role of affect, imagery, and values. Climatic Change, 77, 45–72. https://doi.org/10.1007/s10584-006-9059-9
  • Leiserowitz, A., Roser-Renouf, C., Marlon, J., & Maibach, E. (2021). Global warming’s Six Americas: A review and recommendations for climate change communication. Current Opinion in Behavioral Sciences, 42, 97–103. https://doi.org/10.1016/j.cobeha.2021.04.007
  • McCright, A. M., & Dunlap, R. E. (2011). The politicization of climate change and polarization in the American public's views of global warming, 2001–2010. The Sociological Quarterly, 52(2), 155–194. https://doi.org/10.1111/j.1533-8525.2011.01198.x
  • Moser, S. C. (2010). Communicating climate change: History, challenges, process and future directions. WIREs Climate Change, 1(1), 31–53. https://doi.org/10.1002/wcc.11
  • Myers, T. A., Maibach, E. W., Roser-Renouf, C., Akerlof, K., & Leiserowitz, A. A. (2013). The relationship between personal experience and belief in the reality of global warming. Nature Climate Change, 3, 343–347. https://doi.org/10.1038/nclimate1754
  • Nobel, N., & Hiscox, M. (2026). Enhancing climate resilience with proximal cues in personalized climate disaster preparedness messaging. Nature Human Behaviour, 10, 505–513. https://doi.org/10.1038/s41562-025-02352-w
  • O’Neill, S. J., & Nicholson-Cole, S. (2009). “Fear won't do it”: Promoting positive engagement with climate change through visual and iconic representations. Science Communication, 30(3), 355–379. https://doi.org/10.1177/1075547008329201
  • Pidgeon, N. (2012). Climate change risk perception and communication: Addressing a critical moment? Risk Analysis, 32(6), 951–956. https://doi.org/10.1111/j.1539-6924.2012.01856.x
  • Pidgeon, N., & Fischhoff, B. (2011). The role of social and decision sciences in communicating uncertain climate risks. Nature Climate Change, 1, 35–41. https://doi.org/10.1038/nclimate1080
  • Rode, J. B., Remshard, M., Groot, L. J., & van der Linden, S. (2025). A meta-analytic structural equation analysis of the Gateway Belief Model: Highlighting scientific consensus increases support for public action on climate change. Current Opinion in Behavioral Sciences, 63, 101521. https://doi.org/10.1016/j.cobeha.2025.101521
  • Sambrook, K., Konstantinidis, E., Russell, S., & Okan, Y. (2021). The role of personal experience and prior beliefs in shaping climate change perceptions: A narrative review. Frontiers in Psychology, 12, 669911. https://doi.org/10.3389/fpsyg.2021.669911
  • Sandlund, I., Bjälkebring, P., & Bergquist, M. (2026). Meta-analytical evidence of a self–other discrepancy in climate change-related risk perceptions. Nature Sustainability, 9, 377–384. https://doi.org/10.1038/s41893-025-01717-3
  • Slovic, P. (1987). Perception of risk. Science, 236(4799), 280–285. https://doi.org/10.1126/science.3563507
  • Spence, A., Poortinga, W., & Pidgeon, N. (2012). The psychological distance of climate change. Risk Analysis, 32(6), 957–972. https://doi.org/10.1111/j.1539-6924.2011.01695.x
  • Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/science.185.4157.1124
  • van der Linden, S. (2014). On the relationship between personal experience, affect and risk perception: The case of climate change. European Journal of Social Psychology, 44(5), 430–440. https://doi.org/10.1002/ejsp.2008
  • van der Linden, S. (2015). The social-psychological determinants of climate change risk perceptions: Towards a comprehensive model. Journal of Environmental Psychology, 41, 112–124. https://doi.org/10.1016/j.jenvp.2014.11.012
  • van der Linden, S., Leiserowitz, A. A., Feinberg, G. D., & Maibach, E. W. (2015). The scientific consensus on climate change as a gateway belief: Experimental evidence. PLOS ONE, 10(2), e0118489. https://doi.org/10.1371/journal.pone.0118489
  • van der Linden, S., Leiserowitz, A., Rosenthal, S., & Maibach, E. (2017). Inoculating the public against misinformation about climate change. Global Challenges, 1(2), 1600008. https://doi.org/10.1002/gch2.201600008
  • Većkalov, B., Geiger, S. J., Bartoš, F., White, M. P., Rutjens, B. T., van Harreveld, F., Stablum, F., Akın, B., Aldoh, A., Bai, J., Berglund, F., Bratina Zimic, A., Broyles, M., Catania, A., Chen, A., Chorzępa, M., Farahat, E., Götz, J., Hoter-Ishay, B., . . . van der Linden, S. (2024). A 27-country test of communicating the scientific consensus on climate change. Nature Human Behaviour, 8, 1892–1905. https://doi.org/10.1038/s41562-024-01928-2
  • Vlasceanu, M., Doell, K. C., Bak-Coleman, J. B., Todorova, B., Berkebile-Weinberg, M. M., Grayson, S. J., Patel, Y., Goldwert, D., Pei, Y., Chakroff, A., Pronizius, E., van den Broek, K. L., Vlasceanu, D., Constantino, S., Morais, M. J., Schumann, P., Rathje, S., Fang, K., Aglioti, S. M., . . . Van Bavel, J. J. (2024). Addressing climate change with behavioral science: A global intervention tournament in 63 countries. Science Advances, 10(6), eadj5778. https://doi.org/10.1126/sciadv.adj5778
  • Weber, E. U. (2006). Experience-based and description-based perceptions of long-term risk: Why global warming does not scare us (yet). Climatic Change, 77, 103–120. https://doi.org/10.1007/s10584-006-9060-3
  • Whitmarsh, L. (2008). Are flood victims more concerned about climate change than other people? The role of direct experience in risk perception and behavioural response. Journal of Risk Research, 11(3), 351–374. https://doi.org/10.1080/13669870701552235
  • Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59(4), 329–349. https://doi.org/10.1080/03637759209376276
  • Wong-Parodi, G., & Garfin, D. R. (2022). Hurricane adaptation behaviors in Texas and Florida: Exploring the roles of negative personal experience and subjective attribution to climate change. Environmental Research Letters, 17(3), 034033. https://doi.org/10.1088/1748-9326/ac4858