Affective Imagery, Risk Perceptions, and Climate Change Communication
Summary and Keywords
Affective imagery, or connotative meanings, play an important role in shaping public risk perceptions, policy support, and broader responses to climate change. These simple “top-of-mind” associations and their related affect help reveal how diverse audiences understand and interpret global warming. And as a relatively simple set of measures, they are easily incorporated into representative surveys, making it possible to identify, measure, and monitor how connotative meanings are distributed throughout a population and how they change over time. Affective image analysis can help identify distinct interpretive communities of like-minded individuals who share their own set of common meanings and interpretations. The images also provide a highly sensitive measure of changes in public discourse. As scientists, political elites, advocates, and the media change the frames, images, icons, and emotions they use to communicate climate change, they can influence the interpretations of the larger public. Likewise, as members of the public directly or vicariously experience specific events or learn more about climate risks, they construct their own connotative meanings, which can in turn influence larger currents of public discourse. This article traces the development of affective imagery analysis, reviews the studies that have implemented it, examines how affective images influence climate change risk perceptions and policy support, and charts several future directions of research.
Natural scientists have described global warming as perhaps the preeminent environmental risk confronting the world in the 21st century. Meanwhile, social scientists have found that people respond to hazards based on their perceptions of the risks and the efficacy of potential solutions. What the public perceives as a risk, why they perceive it that way, and how they will subsequently behave are thus vital questions for policymakers attempting to address global climate change, in which the effects are delayed, have inequitable distributions of costs and benefits, and are beyond the control of any one group. In this situation, public support for or rejection of proposed climate policies will be greatly influenced by the perceived risks of global warming.
Much risk perception research has contrasted expert versus public knowledge or mental models of climate change and found that Americans generally lack detailed conceptual understandings of this environmental risk. Further, mental model researchers have found that Americans inappropriately apply concepts from other environmental issues (e.g., Bostrom, Morgan, Fischhoff, & Read, 1994; Kempton, Boster, & Hartley, 1995; Read, Bostrom, Morgan, Fischhoff, & Smutts, 1994). For example, many Americans confuse global warming with ozone depletion, leading them to make inaccurate inferences about both the causes of and solutions to climate change (Kempton, Boster, & Hartley, 1995).
Risk perception researchers, however, have demonstrated that knowledge about risks, while important, is not sufficient to explain risk perceptions and behavior (e.g., Epstein, 1994; Finucane, Alhakami, Slovic, & Johnson, 2000; Slovic, 1997, 1999, 2001; Wildavsky, 1991). Further, environmental scientists, decision-makers, and risk communicators are increasingly aware that providing more detailed and accurate information, while important, is not sufficient to generate appropriate public concern for some risks or to allay public fears about others. Mental model researchers have analyzed how people cognitively reason, but not how risk perception and behavior are guided by emotion and affect. This review describes affective image analysis as a complementary approach, blending quantitative and qualitative elements, to examine public risk perceptions along both cognitive and affective dimensions.
Affect refers to the specific quality of “goodness” or “badness” experienced as a feeling state (with or without conscious awareness) or the positive or negative quality of a stimulus. Affect is distinguished from emotion, which generally refers to specific states, such as anger, fear, or happiness. Affect is also distinct from mood, which generally refers to transitory, low-intensity feelings, which are undirected and lack specific cognitive content. By contrast, affect refers to a person’s positive or negative evaluation of specific cognitive contents or images.
Imagery refers to all forms of mental representation or cognitive content. Images include both perceptual representations (pictures, sounds, smells) and symbolic representations (words, numbers, symbols) (Damasio, 1999, pp. 317–321). Affective images are thus “broadly construed to include sights, sounds, smells, ideas, and words, to which positive and negative affect or feeling states have become attached through learning and experience” (Slovic, MacGregor, & Peters, 1998, p. 3).
Affective images are evaluative feelings of good/positive or bad/negative associated with particular concepts or stimuli; e.g., “cancer” evokes negative images for most people, while “sushi” probably evokes either positive or negative images. Affective images occur rapidly and automatically; note how quickly you sensed your positive or negative feelings associated with the stimulus words cancer or sushi. Affective images also guide decision-making; seeing “sushi” on a dinner menu causes some people to react with disgust, while others salivate in anticipation, guiding their subsequent behavior (what to order for dinner). Further, affective images are part of an “affect heuristic,” a natural and efficient process for evaluating risk situations and making decisions (Finucane, Alhakami, Slovic, & Johnson, 2000; Slovic, Finucane, Peters, & MacGregor, 2002; Slovic, Finucane, Peters, & MacGregor, 2007).
Affective Image Analysis
… associations are simply a remarkably easy and efficient way of determining the contents of human minds without having those contents expressed in the full discursive structure of language.
— Szalay and Deese (1978, p. 9)
Cognitive scientists, psycholinguists, and social psychologists demonstrate that affective images link to one another in complex networks of association (e.g., Anderson & Bower, 1973; Deese, 1965; Kess, 1992; Sadoski & Paivio, 2001; Sloman, 1996; Slovic, Flynn, & Layman, 1991; Slovic, MacGregor, & Peters 1998). As Kess (1992, p. 213) describes: “The system is like a spider web, with words in the mental network related to other words via associative links of varying strengths.”
Psychologists have long used word association techniques (Galton, 1880; Wundt, 1883; Freud, 1924; Hollway & Jefferson, 2008; Joffe & Elsey, 2014). A wide range of fields use word association methods, including the study of memory and cognition (e.g., Altarriba, Bauer, & Benvenuto, 1999; Nelson, McEvoy, & Dennis, 2000; Zeelenberg, Shiffrin, & Raaijmakers, 1999), the structure of thought systems (McGuire & McGuire, 1991, 2001), and psychotherapy (MacMillan, 2001). Affective image analysis employs a particularly structured and systematic form of word association. Researchers have implemented affective image analysis in many different research domains. Szalay, Strohl, and Doherty (1999), for example, used affective image analysis of word associations to examine substance abuse for the National Institute on Drug Abuse, to develop cross-cultural communication guides for the U.S. Department of Education, and to improve minority training for the National Institute of Mental Health and Office of Naval Research. Risk perception researchers have used affective image analysis of word associations to examine public perceptions of global warming (Leiserowitz, 2005, 2006), nuclear power (Slovic, Flynn, Mertz, Mays, & Poumadere, 1996), and smoking (Benthin et al., 1995). Others have used it to predict vacation preferences and behavior (Slovic et al., 1998) and investors’ evaluations of initial public offerings on the stock market (MacGregor, Slovic, Dreman, & Berry, 2000).
In gathering word associations for affective image analysis, researchers typically present subjects with a key word or concept and ask them to provide the “first word or image” that comes to mind. Some instruments gather multiple images in addition to subjects’ first answers. Example questions include:
Q1: What is the first word or image that comes to mind when you think of nuclear energy?
Q2: What is the second word or image that comes to mind when you think of nuclear energy?
Affective image analysis involves content analysis of respondents’ open-ended answers to the stimulus words. The results enable researchers to describe the range, structure, and salience of respondent associations with the stimulus term. We describe this methodology in greater detail below.
Besides images, some instruments also gather subjects’ directional, affective evaluations of both the key word stimulus and respondents’ images of the stimulus. An example affective evaluation question is:
Q3: When you think of nuclear energy, is your impression very negative, somewhat negative, somewhat positive, or very positive?
If a survey respondent answered Q1 with “cancer,” an example follow-up question to obtain an affective evaluation of the image might be:
Q4: When you think of cancer, is your impression very negative, somewhat negative, somewhat positive, or very positive?
Unique Strengths of Affective Image Analysis
A first unique strength of affective image analysis is that it minimizes the researcher bias that can occur in closed-ended questionnaires. Researchers often construct closed-ended questionnaires based on a particular theory or set of academic or scientific categories. This process risks using terms, concepts, and categories unfamiliar to, or unused by, the general public. As Cox explains (quoted in Robinson, 1998, p. 389):
The categories supplied by the investigator may have quite different meanings to different individuals in the sample. Likewise the responses supplied by the investigator on the assumption that they may be exhaustive may be far from so. As a consequence, one may learn more about the behaviour of the sample in responding to a set of categories the investigator attempts to impose on them, than about the behaviour under investigation itself.
Affective image analysis, however, uses an open-ended method of free association, with no imposition of researcher concepts or categories, with the exception of the stimulus term itself. This feature makes affective image analysis a particularly useful tool for examining public risk perception, as scientific analysis of risks often uses a highly technical lexicon, which members of the public may not understand or be comfortable with.
A second unique strength of this technique is that subjects’ free associations are unfiltered, relatively context-free, and spontaneous. In this way, they represent a powerful way to study sensitive topics. Because they are immediate and spontaneous, they tend to bypass self-censorship or social desirability bias, which can occur in the fully articulated assertions required by standard survey questions. As Szalay and Deese (1978, p. 9) explain:
… the association method reveals the content of minds in a way that propositional language does not. We can and do reveal ourselves in associations in ways that we might find difficult [or unpleasant] … if we were required to spell out the full propositions behind our associations.
Third, affective image analysis provides a much richer dataset than closed-ended questionnaires. This technique efficiently collects the range and diversity of respondent interpretations of a stimulus, thus allowing researchers to map the distribution and saliency of subjective meanings held by a population. Further, this technique helps investigators identify, compare, and contrast interpretive communities—groups of individuals who share a relatively coherent and consistent set of meanings while often differing greatly from one another.
Fourth, affective image analysis provides a less costly method for researchers to acquire some of the data richness of qualitative approaches. The time, labor, and resource costs of in-depth interviews, participant observation, and ethnography, for example, greatly exceed the relatively simple, quick, and inexpensive collection of affective images.
Fifth, when conducted with representative samples, affective image survey results are generalizable to entire populations and can be quantitatively correlated to behavior, attitude, opinion, and sociodemographic variables.
Affective image analysis thus combines the strengths of qualitative and quantitative methods—providing some of the richness of qualitative methods with the generalizability of quantitative approaches. Word associations elicit short, open-ended responses (without imposing researcher categories) from a representative sample using standard survey methods. The qualitative meanings can then be coded into empirically determined categories and subsequently correlated with other survey measures (e.g., sociodemographics).
Affective Imagery in Decision-Making
Many theories of risk decision-making are “cognitive and consequentialist” (Loewenstein, Weber, Hsee, & Welch, 2001). They are based on the assumption that individuals assess the desirability and likelihood of possible outcomes, then rationally calculate or weigh the pros and cons of these expected outcomes to arrive at a decision. This assumption also underlies the expected utility model that informs much of economic and psychological theory. Thus, past research has attempted to model how people make logical, rational choices. Affect and emotions are typically ignored in these models or viewed as epiphenomena of the decision-making process. Thus, many theorists have assumed that decision-making about risk is essentially a cognitive activity.
In a critique of this predominantly cognitive paradigm, Zajonc (1980) argued that affective reactions to stimuli occur automatically and guide subsequent information processing and judgment. Affect typically arises prior to cognition and plays a crucial role in subsequent rational thought. A large and growing literature in cognitive science, social psychology, and neurology has confirmed affect’s key role in cognition, decision-making, and behavior (e.g., Alhakami & Slovic, 1994; Epstein, 1994; Isen, 1993; Janis & Mann, 1977; Johnson & Tversky, 1983; Kahneman & Snell, 1990; Loewenstein, 1996; Loewenstein et al., 2001; Mellers, Schwartz, Ho, & Ritov, 1997; Peters & Slovic, 1996; Rozin, Haidt, & McCauley, 1993; Slovic, Flynn, & Layman, 1991; Slovic et al., 1998, 2007; Wilson et al., 1993).
For example, cognitive neuroscientists’ discoveries of a direct link between affect and cognition illustrate affective imagery’s importance for risk perception and decision-making research. Damasio (1994, 1999) found that patients with brain damage (specifically, the ventromedial frontal cortices) retain their intelligence, memory, and logic but lose their ability to feel, including the ability to associate feelings with their actions’ anticipated consequences. Through several experiments, Damasio found that such emotional deficits dramatically impair decision-making processes and seriously compromise decision quality. Based on these and other studies, Damasio concludes that human thought largely consists of images, broadly construed to include ideas, words, sounds, smells, and real or imagined visual impressions. Through life experience these images become “marked” by positive or negative feelings and linked directly or indirectly to somatic (bodily) states. A negative image sounds an alarm; a positive image becomes a beacon of incentive. Damasio argues that these positive or negative “somatic markers” greatly improve decision-making efficiency and accuracy and guide both slow, deliberative choices and quick, intuitive responses.
Affective Imagery in Risk Perception and Behavior
Using affective image analysis, Slovic, Flynn, and Layman (1991) reported four surveys conducted at the national, regional, state, and local levels to examine public risk perceptions of the proposed national nuclear waste storage facility underneath Yucca Mountain, Nevada. The surveys collected 10,000 affective images from 3,334 respondents to the stimulus “underground nuclear waste repository.” Respondents’ associations coalesced around images like “death,” “cancer,” “mushroom cloud,” and “nuclear war.” Respondents evaluated these images extremely negatively, with positive imagery almost entirely absent. Slovic, Flynn, and Layman (1991, p. 1604) concluded that, “What these responses reveal are pervasive qualities of dread, revulsion and anger—the raw materials of stigmatization and political opposition.” Further, the researchers found that these affective images predicted public risk perceptions of nuclear energy, and correlated strongly with intended voting behavior and opposition to new nuclear power plant construction (Peters & Slovic, 1996; Slovic, Flynn, & Layman, 1991).
Finucane et al. (2000) used affective image analysis to assess public risk perceptions of blood transfusions with a nationally representative survey in the United States (n = 385) conducted in 1997–1998. At the beginning of the survey, respondents were asked: “When you hear the phrase ‘blood transfusions,’ what is the first word or brief image that comes to your mind?” Up to three associations were collected from each respondent, who then rated their images on a 5-point, positive-to-negative affect scale. The images were then coded into four main categories: health and safety, functional considerations, personal considerations, and other. Health and safety was the largest overall category and the largest subcategory was “risk of AIDS/HIV” reflecting the crisis of the prior decade. Overall average affect for images associated with “blood transfusions” was positive, with the exception of AIDS/HIV images, which were more negative. The study concluded that most Americans had neutral to positive associations with blood transfusions but that a substantial minority of the public associated this critical medical practice with the dangers of AIDS/HIV infection. Thus the raw materials of stigmatization were present in the connotative meaning of “blood transfusions,” a somewhat dormant set of meanings that might become activated in the event of a new blood safety crisis, thus posing a risk to the medical field, which depends on volunteers to donate blood and sick individuals to accept transfusions.
Siegrist (2003) used affective image analysis to assess public acceptance of “gene technology” with a telephone survey in Switzerland (n = 1,001), conducted in 1997. Responses were coded “as positive (e.g., progress, drugs, benefits, no risks), as neutral (e.g., benefits and risks, cloning, new food, manipulation), or as negative (e.g., frog without head, risks, superfluous, environmental disasters” (Siegrist, 2003, p. 49). Most responses (55%) were neutral, while 34% were negative and only 11% were positive. People with negative or neutral associations were much more likely to perceive gene technology as a risk than people with positive associations.
Siegrist, Earle, Gutscher, and Keller (2005) then used affective image analysis to assess public risk perceptions of mobile phone and base station risks with a telephone survey in Switzerland (n = 1,015), conducted in 2002. The researchers provided respondents with introductory information about how mobile phones rely on base stations, then asked for their first free association to “base stations.” Respondents then rated their own associations on a -5 (negative) to +5 (positive) affect scale. The study found that the associations evoked by the term “base station” were, on average, more negative than positive. Dominant negative associations included electromagnetic radiation pollution (34%), health-related images (18%), and aesthetics (9%). Respondents with negative associations to base stations perceived them as a greater risk than respondents with positive associations.
Siegrist, Cusin, Kastenholz, and Wiek (2007) also assessed public acceptance of nanotechnology foods and food packaging with a convenience sample of 153 grocery shoppers in Switzerland. The researchers provided respondents with introductory information about nanotechnology foods, then asked them to provide as many free associations as they wished in 60 seconds. The respondents then rated their first five associations on a 5-point negative to positive affect scale. The affect ratings were averaged and used as a proxy for respondents’ nanotechnology food affect and used as a variable in a structural equation model predicting willingness to buy. Image affect was found to play a significant role in shaping respondents’ perceptions of nanotechnology benefits, risks, and their willingness to pay for such products.
Keller, Visschers, and Siegrist. (2012) used affective image analysis to assess the relationships between public associations to “nuclear power plants,” acceptance of new nuclear power plants, trust, gender, and region with a representative sample of Switzerland (n = 1,221), conducted in 2009. At the beginning of the survey, respondents were asked for their first association to the phrase “nuclear power plant.” Respondents then rated their own associations on an 11-point, -5 (negative) to 0 (neutral) to +5 (positive) affect scale. Respondents were then asked how much they favored or opposed the replacement of current nuclear power plants with new ones, their level of trust in nuclear power plant operators, and environmental organizations. The associations evoked by the term “nuclear power plant” were, on average, more negative than positive. Dominant negative associations included the categories of risk, accident, negative affect, and waste problems. Positive associations included the categories of energy and simple descriptions of the appearance or location of nuclear power plants. Overall, negative images were correlated with opposition to new nuclear power plants, while positive images were correlated with support. Further, opponents had a more diverse set of negative images than supporters had positive images, indicating they had more reasons to oppose nuclear power. Men were more accepting of new nuclear power plants than women. Negative image affect was significantly correlated with opposition to nuclear power plants, distrust of nuclear plant operators, and trust in environmental organizations.
Truelove (2012, p. 478) used affective image analysis to “assess affective, emotional and cognitive perceptions of coal, natural gas, nuclear and wind energy and the relationship between these perceptions and support for the energy sources,” with a small sample (n = 94) of respondents drawn from a university-based consumer panel. Respondents were asked for their first five associations to “coal power,” “nuclear power,” “natural gas power,” and “wind power” and then asked to rate their own associations on a 5-point, 1 (very negative) to 5 (very positive) affect scale. The most common coal images were of mines/miners, coal, and trains, which were relatively neutral in affect. But there were also more negatively charged images, including mine disasters, unhealthy, and environmentally harmful. Nuclear power images were often quite negative and included associations with Chernobyl, bombs, and explosions. Natural gas power images included neutral images like pipes, blue flame, and gas appliances and more negative images like fires/burns and explosions. Finally, wind power generated almost entirely positive images, including windmills/turbines, wind, and eco-friendly. Image affect was correlated with support for each power source, especially nuclear power, although a separate scale of cognitive beliefs about the impacts of each power source on dangerous waste, the U.S. economy, the environment, energy independence, risk to humans, and benefits vs. costs explained more variance in support in a combined regression model.
Overall, these studies demonstrate that affective images, or connotative meanings, are important predictors of public risk perceptions and support for or opposition to a wide variety of potential hazards.
Affective Imagery, Risk Perception, and Climate Change
Other researchers have used affective image analysis to investigate the role of these free associations, or connotative meanings, in public climate change risk perceptions and policy support. Leiserowitz (2005) first used affective imagery analysis as part of a nationally representative survey (n = 673) conducted in 2002–2003 on American climate change risk perceptions, policy preferences, and behaviors. Respondents were asked: “When you hear the words ‘global warming,’ what is the first thought or image that comes to mind?” Each self-reported image was then rated by the respondent on a Likert affect scale ranging from −5 (very negative) to +5 (very positive). The images were then content-analyzed by two independent coders to identify common themes (Figure 1).
Americans were most likely to associate global warming with impacts on places or natural ecosystems distant from everyday experience. Associations to melting ice (e.g., melting polar ice caps, Antarctica melting) were most frequent, followed by generic references to rising temperatures (e.g., temperatures increasing) and impacts on nonhuman nature (e.g., upset ecological balance). These were followed by associations to the ozone hole, alarmist images of disaster (e.g., death of the planet, like after nuclear war), flooding and sea level rise, climate change, and then “naysayer” images, indicating doubt or denial of the reality of climate change (e.g., unproven theory, a hoax).
One of the most important findings was what was not found—almost no respondents associated global warming with impacts on human health. Instead global warming was interpreted as a risk largely affecting non-human nature (e.g., ice and polar bears) far away in space and time and removed from the daily concerns of most people. Nonetheless, mean affect scores indicated that “global warming” had negative connotations for almost all respondents. Alarmist images of disaster produced the strongest negative affect, while naysayers reported minimal negative affect.
The predominance of associations to melting ice was likely due to the prevalence of melting ice images accompanying news stories about climate change. Satellite images and maps of declining sea ice on the Arctic Ocean, footage of calving icebergs and ice shelves breaking off Antarctica, and before-and-after photographs of retreating glaciers became iconic symbols of a warming world and played an important role in shaping the public’s imagination. On the one hand, these media images helped make the abstraction of a warming world visible and concrete as they activated embodied mental models held by most of the public. Through repeated experience (e.g., a glass of ice outside on a summer day), people have learned what happens when ice is exposed to warm conditions—it melts. Media images of melting ice from around the world helped people understand the reality of global warming intuitively, without the abstractions of statistical data, model results, or scientific explanations.
On the other hand, very few Americans live on the shores of the Arctic Ocean, in Antarctica, or next to a melting glacier. Thus these same images tended to reinforce the notion that the impacts of global warming are far away in space—primarily happening at the poles or at high altitudes. And importantly, most of these images of melting ice lacked any human figures, thus reinforcing the (incorrect) inference that global warming will primarily impact non-human nature, not people. While the images of melting ice often depicted dramatic events, they lacked the visceral or emotional punch of other risks, e.g., terrorist attacks or human beings in distress. The public is still largely unaware of the broader systemic consequences of melting ice—e.g., rising sea levels and flooding of coastlines worldwide, shifts in global weather patterns, or the decrease in the earth’s albedo as reflective snow and ice is replaced by darker ocean and land, absorbing and trapping more heat at the earth’s surface, leading to a spiral of amplified warming, among many other impacts.
Associations to ozone depletion were also relatively common. The ozone hole became a public issue a decade before global warming was widely reported in the media. Since individuals tend to assimilate new information into already existing mental models (Kempton et al., 1995; Levy, 1997), this led to several important misconceptions and confusions between the two environmental issues. Many people mistakenly believe that ozone depletion is a cause of climate change. Many reason that if there is a “hole” in the ozone layer and a global “greenhouse” effect, then there must be a “hole” in the “greenhouse.” This “hole” either allows more solar radiation into the biosphere—warming the planet—or, alternatively, is allowing heat to escape—cooling the planet. This metaphorical reasoning is logical, but incorrect. It is also a valuable demonstration of how people make inferences based on what they already know or believe. In this case, many members of the public already had a mental model about the ozone hole, which they used to reason about climate change.
Leiserowitz (2005) also identified two distinct “interpretive communities” or groups within the American public (alarmists and naysayers), each of which held a relatively coherent and consistent set of associative and connotative meanings about global warming while differing greatly from each other. First, 11% of Americans provided “alarmist” associations of disaster and catastrophe to the stimulus term “global warming.” Many of these respondents associated global warming with social and environmental collapse, mass extinctions, and the end of the world. By contrast, 7% of Americans provided “naysayer” associations of profound skepticism and denial about global warming. Five different types of naysayers were identified: those who flatly deny that global warming exists; those who believe that global warming is just part of a normal, natural phenomenon; those that believe the scientific case is still unproven; those who believe that global warming is real but overly hyped by the media or environmentalists; and conspiracy theorists who believe global warming is a hoax perpetrated on the public by scientists, liberals, environmental groups, the UN, or other countries.
Naysayers were much more likely to be conservative, better educated, Republican males and preferred to get their news from the radio. These naysayers had lower global warming risk perceptions, strongly opposed climate change policies, and were less likely to have taken mitigation actions. They also tended to have pro-individualist and anti-egalitarian worldviews and to hold anti-environmental attitudes. They were also much more likely to prefer protection of the economy over protection of the environment. By contrast, Alarmists tended to have pro-egalitarian and anti-individualist worldviews, were politically liberal, strongly supported government policies to mitigate climate change (including raising taxes), and were significantly more likely to have taken personal action to reduce greenhouse gas emissions.
A subsequent study (Leiserowitz, 2006) conducted a series of regression analyses to examine the relative power of these affective images as predictors of climate change risk perceptions and policy preferences. The analyses found that several image categories (and in particular the naysayer images) were consistently more powerful predictors of climate change risk perceptions and policy preferences than sociodemographic or even political variables (party ID or ideology). Thus the connotative meaning of global warming played a significant, independent role in shaping public responses to the issue.
Smith and Leiserowitz (2012) conducted a time-series analysis to investigate how image associations and their accompanying affect had changed over time. Using nationally representative data from four national surveys in the United States (n = 673, 1,014, 2,164, and 1,001, completed in 2003, 2007, 2008, and 2010, respectively), they found a sharp increase in naysayer image associations, with decreases in other major categories like melting ice, heat, and ozone depletion. Naysayer images spiked especially after the “climategate” episode, in which scientists’ emails were stolen and then individual fragments were taken out of context by opponents of climate action to allege scientific fraud. Americans became significantly more likely to report naysayer associations such as “the biggest scam in the world to date,” “unscientific theory,” and “there really is no such problem,” indicating greater doubt and dismissal of the reality or seriousness of the issue. Replicating the results in Leiserowitz (2006), images and their associated affect also explained higher proportions of variance in climate change risk perceptions and policy preferences than other predictors, including cultural worldviews, political ideology, and other sociodemographics, further indicating the important role affective imagery and connotative meaning can play in public opinion, judgment, and decision-making.
Exploring religious reactions to climate change, Smith and Leiserowitz (2013) investigated the role of affective imagery in how American evangelicals engage with the issue. Using a nationally representative survey (n = 2,164) in the United States collected in 2008, they found that evangelicals who associated climate change with alarmist, flooding, sea level rise and weather images perceived global warming as a greater risk, whereas those who associated the issue with naysayer and politics-based imagery had lower risk perceptions. Additionally, affective images were also significant predictors of support for or opposition to a range of climate policies. Evangelicals who associated climate change with alarmist imagery were the most likely to support policy, whereas those who held naysayer images were more likely to oppose policy. Concern about global warming impacts and the cultural worldviews of egalitarianism and individualism were stronger predictors of risk perception and policy support among American evangelicals, yet the image categories were more predictive than a variety of sociodemographic and political variables.
Smith and Leiserowitz (2014) then used affective image analysis to examine the links between affect and emotion and the role discrete emotions play in how people respond to climate change. Using a nationally representative survey in the United States (n = 1,001) completed in 2010, respondents were asked to provide affective images of global warming, were asked to rate the intensity of different emotions (anger, fear, worry, hope, interest, etc.) felt when thinking about global warming, and completed other measures including cultural worldviews and sociodemographics. Regression analyses revealed that discrete emotions were the strongest predictors of policy support, followed by cultural worldviews. Affective images were more predictive than sociodemographic variables including political party identification and political ideology. Thus affective images were again found to play an important role in public climate change policy support, even controlling for discrete emotions.
The majority of research using the affective image approach to investigate public responses to climate change has utilized American samples. However, a couple of studies have examined affective images of climate change in other countries. Lorenzoni, Leiserowitz, Doria, Poortinga, and Pidgeon (2006), for example, collected affective images from a nationally representative sample in the United Kingdom to compare with those collected in the United States (see Leiserowitz, 2005, 2006). Using the stimulus term “climate change” instead of “global warming,” the study found that Britons had different associations to this stimulus term than Americans did to “global warming.” For example, Britons were most likely to provide associations to weather compared to melting ice in the United States. Americans were more likely to provide associations with rising temperatures and impacts on nonhuman nature, whereas Britons were more likely to provide associations to the ozone hole. Despite these differences, however, mean affect ratings were negative for all images in both the American and U.K. samples, indicating that global warming and climate change had negative connotations for both populations.
Leviston, Price, and Bishop (2014) collected affective image associations from a nationally representative sample of the Australian public, using a variety of elicitation exercises. In an initial survey study participants were instructed to provide the first three images that come to mind when thinking about climate change along with accompanying affect each image evoked. The top three image associations of rising sea levels, drought, and melting ice caps somewhat mirrored those produced by Americans. However, this study extended the investigation of the affective dimensions of the images beyond “goodness” and “badness.” Though a series of workshops, new participants were asked to sort through a set of 82 images, broadly representing initial image associations, and place those they most associated with climate onto an emotion grid of different discrete emotions positioned along two axes: active vs. inactive and positive/pleasant vs. negative/unpleasant. Imagery about climate impacts was found to be the most salient category of associations for these participants. Imagery associated with disasters and extreme events evoked feelings of alarm, anger, and fear, “icemelt” imagery evoked feelings of upset and frustration, and drought and denuded landscapes evoked feelings of upset and misery. Using qualitative analysis of the workshop discussions, Leviston et al. (2014) argued that these different emotional reactions prompt a range of different adaptive responses.
In a final study, Leiserowitz et al. (2014) explored the different affective images the American public associated with the terms “global warming” versus “climate change.” While scientists often prefer the term “climate change” to “global warming,” the latter is still more widely used in public discourse, and communicators often question which term is more tactically effective with the general public and specific audiences. In a nationally representative survey, participants were randomly assigned to receive an extensive survey using one or the other term. Affective images were also collected. The analysis found that “global warming” evoked more alarmist and “icemelt” images, while “climate change” evoked more associations with weather. Moreover, different responses were reported between those who differed in political orientation and other demographic characteristics. For example, “global warming” was more likely to evoke alarmist imagery among liberals, males, and African-Americans, and “icemelt” images among moderates, conservatives, men and women, and those under 60. “Climate change” however, was more likely than “global warming” to elicit naysayer images from moderates and Hispanics, whereas conservatives, females, and Hispanics were more likely to provide weather images in response to “climate change” than in response to “global warming.” The study found important similarities and differences in the connotations each term evokes, indicating that one term may be more effective than the other, depending on the goal of the communication.
The studies reviewed contribute to the growing scientific literature and understanding of how public risk perceptions and policy preferences are influenced by affective imagery and some indication of how these images of risk can vary across cultural contexts. Much work remains to be done, however, especially in less developed and non-English speaking countries. Also not clearly understood is how well affective imagery analysis can detect and track shifts in connotative meaning over time. Each of the studies reviewed offers a snapshot of the dominant image associations among a particular population at a specific time (with the exception of Smith & Leiserowitz, 2012). Below we provide the first long time-series analysis of how affective imagery, or the connotative meaning of global warming, has evolved in Americans’ consciousness from 2003 until 2016.
Changes in Climate Change Affective Imagery Over Time
A total of 14 datasets were analyzed for the time-series analysis spanning a 14-year period between 2002 and 2016 (see Table 1 for dates of data collection, sample sizes and methodology employed). Survey 1 utilized a mail-out, mail-back methodology using the tailored design method (Dillman, 2000). Survey 2 was conducted in collaboration with Gallup using adults drawn from Gallup’s household panel that had originally been recruited using random selection criteria. Both of these surveys were nationally representative of the American public, and data was weighted to match U.S. Census Bureau population parameters for each time period. All the remaining surveys utilized either Knowledge Networks’s KnowledgePanel® or GfK’s KnowledgePanel®. For both panels, prospective members were recruited using a combination of random digit dial and address-based sampling techniques that cover virtually all (non-institutional) resident phone numbers and addresses in the United States. Those contacted who do not have access to the Internet are loaned computers and given Internet access so they may participate. The sample includes a representative cross section of American adults—irrespective of whether they have Internet access, use only a cell phone, etc. Key demographic variables were weighted, postsurvey, to match U.S. Census Bureau norms.
Table 1. Survey Data Collection
Dates of data collection
Method of data collection
November 2002–February 2003
Gallup household telephone panel
October 7–November 12, 2012
Knowledge Networks KnowledgePanel®
December 24, 2009–January 3, 2010
Knowledge Networks KnowledgePanel®
May 14–June 1, 2010
Knowledge Networks KnowledgePanel®
April 23–May 12, 2011
Knowledge Networks KnowledgePanel®
October 20–November 16, 2011
Knowledge Networks KnowledgePanel®
March 12–March 30, 2012
Knowledge Networks KnowledgePanel®
August 31–September 12, 2012
Knowledge Networks KnowledgePanel®
April 10–15, 2013
November 23–December 9, 2013
Split sample used. Participants were randomly assigned to otherwise identical questionnaires that either used the term “global warming” (n = 830) or “climate change” (n = 827).
April 15–22, 2014
October 17–28, 2014
March 18–31, 2016
Affective image data collected in each survey contained two elements: a cognitive component (the image category) and associated affective rating (a goodness or badness evaluation). Slightly different wording was used in the 2002 survey: “What is the first thought or image that comes to your mind when you think of global warming?” versus all subsequent surveys: “When you think of ‘global warming,’ what is the first word or phrase that comes to your mind?” Responses took the form of single word associations (e.g., “apocalypse”) or short narrative statements (e.g., “the end of the world”). Once collected, respondents were then asked to provide an affective rating for their own images using a 10-point scale in 2002 (+5 = a very good thing and –5 = a very bad thing) and a 6-point scale in each subsequent survey (+3 = a very good thing and –3 = a very bad thing). This procedure produced a series of rich data sets that were analyzed using inductive content analysis. Ten percent of the images from each survey were also double coded to ensure transparency of the coding frame, and agreement between both coders was satisfactory (80% or higher). Differences were resolved following discussion between the two coders. As image coding was conducted by different coders across this 14-year time period, Nicholas Smith also cross-checked every image association to ensure image coding was consistent across the whole time-series analysis. The mean affect of each image category was calculated and affective ratings were normalized to a +1 to –1 scale (+1 = very good and –1 = very bad) to enable comparability analyses.
Alarmist and Naysayer Trends
The first measures of Americans’ affective images associated with “global warming” were conducted in early 2003. At the time, alarmist images accounted for 11%, while naysayer images accounted for 7% of all images (Figure 2). From 2003 to 2008, however, both categories doubled in size to 20 and 13%, respectively. These increases likely reflected the increasing media attention and agenda-setting effect the issue received over this time period (Downs, 1972; Luedecke et al., 2017), including the release of the movies The Day After Tomorrow and An Inconvenient Truth, the subsequent Academy Award for best documentary going to Al Gore (former vice president and 2000 presidential nominee of the Democratic Party), and the Nobel Peace Prize awarded to Al Gore and the Intergovernmental Panel on Climate Change.
In 2009, however, naysayer images rose to a record high of 23%, while alarmist images dropped to 14%. This likely reflected several factors, including the rise of the Tea Party and cues from conservative political elites claiming that global warming wasn’t happening or was a hoax (Carmichael, Brulle, & Huxster, 2017), and the climategate event (Leiserowitz, Maibach, Roser-Renouf, Smith, & Dawson, 2013). Since the fall of 2011, the two categories have oscillated near 15% each.
Another interesting trend is Americans’ associations of “global warming” with “weather” over the time series. Starting in 2007, the proportion of weather-related images increased more than four times, reaching 13% of all images in early 2016. Moreover, the changes in these weather associations to global warming correlate reasonably well with shifts in the seasonal (3-month periods) Climate Extremes Index (Gleason, Lawrimore, Levinson, Karl, & Karoly, 2008), suggesting that extreme weather events were influencing how some Americans were interpreting “global warming” (Figure 3). These results suggest that extreme weather events are potential “teachable moments” (Hart & Leiserowitz, 2009) when Americans are paying attention to events that, in some cases, have strong links to climate change (National Academies of Sciences, Engineering & Medicine, 2016).
Discussion and Conclusion
Collectively, these studies demonstrate that affective imagery, or connotative meanings, play an important role in shaping public risk perceptions, policy support, and broader responses to climate change and other threats. These simple “top-of-mind” associations and their related affect help reveal how diverse audiences understand and interpret complex issues. As a relatively simple set of measures, they are easily incorporated into representative surveys, making it possible to identify, measure, and monitor how connotative meanings are distributed throughout a population and how they change over time. Likewise, the technique of affective image analysis can help identify distinct interpretive communities of like-minded individuals who share a set of common meanings and interpretations, while differing greatly from other groups. Affective imagery also provides a highly sensitive measure of changes in public discourse. As scientists, political elites, advocates, and the media change the frames, images, icons, and emotions they use to communicate or contest issues, they can influence the interpretations of the larger public. Likewise, as members of the public directly or vicariously experience specific events or learn more about risks, they construct their own connotative meanings, which can in turn influence larger currents of public discourse.
Regarding climate change, affective imagery studies have helped identify and track several important indicators of how the public conceptualizes and responds to this issue. First, Americans (and the citizens of several other developed countries) hold a diverse set of associations to global warming; these associations tend to be dominated by images of climate change impacts, such as melting ice, warmer temperatures, and impacts on non-human species. Yet these same dominant associations tend to be psychologically distant in time and space, helping to explain why global warming remains a relatively low priority even among people who accept that it is real. Second, despite this diversity of associations, almost all of them evoke feelings of negative affect, e.g., most members of the public conceptualize global warming as a bad thing. Very few conceptualize global warming as a good thing.
Third, this research has identified the existence of distinct “interpretive communities”—each strongly predisposed to interpret global warming and its potential consequences in very different ways. In particular, the groups in the United States identified as alarmists and naysayers interpret global warming through very different lenses, with alarmists tending to believe that unchecked global warming will have existential, catastrophic consequences, while naysayers express doubts or denial about the reality or seriousness of the problem, ranging from questioning the science to outright conspiracy theories. These groups are the most prone to engage in various forms of motivated reasoning (e.g., Leiserowitz et al., 2013), including ignoring, discounting, or denying evidence that contradicts their strongly held prior beliefs.
Fourth, when utilized in nationally representative surveys, the technique finds remarkable consistency in public associations to global warming, including the same dominant image categories over time. Yet the technique also finds that these connotative meanings of global warming change over time in response to shifts in political elite cues and specific events, including “scandals” like climategate and extreme weather events. Thus, affective images are malleable, changing, and dynamic. Fifth, affective images are strongly correlated with public climate change risk perceptions and policy support, even controlling for other important variables like political party and ideology. In short, meaning matters.
Implications for Risk Communication
For scientists, policymakers and risk communicators concerned about global climate change, these findings indicate that multiple strategies are needed to communicate about global warming. First and perhaps most importantly, this literature has identified the near complete absence of public associations to the projected human health consequences of global climate change. Risk communicators need to articulate and emphasize these health impacts, which are among the most serious consequences of projected climate change. Human health impacts are also more likely to elevate public concerns about global warming (Myers, Nisbet, Maibach, & Leiserowitz, 2012), especially compared to the associations currently dominant (melting ice, generalized heat, and impacts on non-human nature).
Likewise, helping the public understand that the impacts of global warming are here and now, not just far away in space and time is important. Americans and other people around the world are already experiencing flooding from sea level rise, more frequent and severe extreme weather events, crop damage, water shortages, species extinctions, and many other climate change consequences. Communicators should articulate and seek to strengthen these “here and now” associations in the public mind.
A different strategy would be needed, however, for those people who confuse global warming with the ozone hole. Risk communicators should probably target the source of the confusion—the inappropriate application of knowledge about the ozone hole to the problem of global warming—by explicitly attempting to disassociate these two environmental problems. Prior studies have found that these Americans are already inclined to believe government and media sources and exhibit pro-environmental attitudes and behaviors; thus they should be relatively open to risk communication about global climate change. As many already have negative affect associated with ozone depletion, risk communicators could try to link this negative affect to a more accurate image and elaborated mental model of global climate change. As the ozone layer recovers, however, media attention has turned to other issues and fewer news articles mention global warming and the ozone hole together (Google Trends, 2017). Interestingly, American associations of global warming with the ozone hole have also decreased over the past decade, indicating that some affective images may simply fade from public memory when they are not repeatedly reinforced.
An entirely different set of strategies, however, would be needed to convince naysayers that global warming is a serious concern. Naysayers will be difficult to engage constructively, as they tend to strongly distrust government, experts, and the media. This may help explain why increased amounts of scientific information and media exposure have not successfully persuaded this vocal minority of the public. In fact, increased science and media coverage may serve only to strengthen some naysayers’ disbelief, to the point of conspiracy theory. Further research is needed with this subpopulation to identify the arguments, values, and information sources that they do trust, but some will likely never be convinced that global warming is happening, human-caused, or a serious threat (Dunlap & McCright, 2010; McCright & Dunlap, 2011).
By contrast, alarmists already exhibit grave concern about global warming. They strongly support policies to mitigate climate change and are already predisposed to be attentive to and believe scientific, government, and environmentalist messages regarding climate change risks. Yet other research indicates that many alarmists lack an understanding of what they as individuals or society collectively can do to reduce these risks (Leiserowitz, Maibach, Roser-Renouf, & Smith, 2011). Thus, the communication of affective images of climate change solutions is probably vital for these otherwise highly motivated individuals.
These are exciting times for scholars and practitioners interested in measuring and tracking the origins and flow of connotative meanings through mass societies, with many new sources of data available (e.g., representative surveys, experiments, social media), collected in real time, at massive scale and increasingly around the world. Affective imagery studies using surveys still have significant advantages, especially population representativeness. But new data mining techniques using large-scale datasets (e.g., Twitter, Facebook, and other online platforms) offer a range of complementary capabilities.
One suggestion for future studies would be to develop more comprehensive “maps” of the “ecosystems of meaning” surrounding different concepts. For example, a more elaborated map of the connotative meanings surrounding “global warming” would include the connotations associated with other related key concepts, like “climate change,” “carbon dioxide,” “sea level rise,” “carbon tax,” and “clean energy,” among many others. How many and how strong are the connections of these other concepts with “global warming”? How does activation of the affective images associated with one concept shape or color the interpretation of another? Are there typical pathways of affective image cascades, where the activation of one concept strongly tends to activate others in a common pattern?
Likewise, most of the studies above analyzed the prevalence of different affective images associated with a single concept (e.g., “nuclear power” or “global warming”) within a national society. But each society is comprised of many different subgroups—political, economic, racial/ethnic, religious, etc. Thus another future direction is to map these “ecosystems of meaning” for different audiences. For example, the network of connotations associated with “global warming” is likely to be very different for Democrats and Republicans, or Latinos and Anglos. How are these maps similar and overlapping or different and distinct from one another? Could they identify connection points where different systems of meaning intersect and affective images are shared? Could they help explain differences in perception and behavior between different groups?
Other studies should investigate how affective images are created, interpreted, reinterpreted, and diffused locally, nationally, or globally through different communication channels. In a given domain, when new concepts are introduced to the world, what affective images are first deliberately or unintentionally associated with the idea? Which ones continue to “travel” with the key term as it diffuses through a society? How do these original connotations change as they are reinterpreted by new audiences? What new connotations adhere to the concept? And ultimately how do some connotations come to dominate? In the case study of global warming, most Americans were unaware of the term until the testimony of Dr. James Hansen in front of a congressional subcommittee in 1988 (Corfee-Morlot, Maslin, & Burgess, 2007), at which point “global warming” became front page news. How and why did the concept then come to be strongly associated with images of melting ice rather than countless other possibilities? What social actors deliberately tried to change or influence the connotative meaning of “global warming,” how successful were they, and among which audiences? How do these different meanings diffuse through old, new, and evolving communication channels, ranging from face-to-face conversations to global social media platforms?
Finally, the field badly needs more studies in non–English-speaking counties, especially countries representing very different cultures, socioeconomic, and geographic circumstances. The connotative meanings of “global warming” in the United States are different but still relatively similar to those in the United Kingdom or Australia. But how similar are they to the connotative meanings of “global warming” in China, India, Malawi, Argentina, or Vanuatu?
Finally, existing affective imagery studies have barely scratched the surface of the enormous, dynamic, complex, and growing system of human meanings. Global warming, nuclear power, blood transfusions, and the handful of other concepts explored thus far represent tiny forays into a vast world. There have been countless explorers and mapmakers of the world of human meaning over the centuries, from philosophers to historians, poets to anthropologists, and many, many others. Affective imagery studies, however, offer a unique contribution to the exploration of human meaning as a set of scientific techniques that offer systematic, empirical, replicable, predictive, and generalizable findings at scale.
Alhakami, A. S., & Slovic, P. (1994). A psychological study of the inverse relationship between perceived risk and perceived benefit. Risk Analysis, 14(6), 1085–1096.Find this resource:
Altarriba, J., Bauer, L. M., & Benvenuto, C. (1999). Concreteness, context availability, and imageability ratings and word associations for abstract, concrete, and emotion words. Behavior Research Methods, Instruments, & Computers, 31(4), 578–602.Find this resource:
Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Hillsdale, NJ: Lawrence Erlbaum.Find this resource:
Benthin, A., Slovic, P., Moran, P., Severson, H., Mertz, C. K., & Gerrard, M. (1995). Adolescent health-threatening and health-enhancing behaviors: A study of word association and imagery. Journal of Adolescent Health, 17(3), 143.Find this resource:
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.Find this resource:
Carmichael, J. T., Brulle, R. J., & Huxster, J. K. (2017). The great divide: Understanding the role of media and other drivers of the partisan divide in public concern over climate change in the USA, 2001–2014. Climatic Change, 141(4), 599–612.Find this resource:
Corfee-Morlot, J., Maslin, M., & Burgess, J. (2007). Global warming in the public sphere. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 365(1860), 2741–2776.Find this resource:
Damasio, A. (1999). The feeling of what happens. New York: Harcourt.Find this resource:
Damasio, A. R. (1994). Descartes’ error: Emotion, reason, and the human brain. New York: Grosset/Putnam.Find this resource:
Deese, J. (1965). The structure of associations in language and thought. Baltimore, MD: Johns Hopkins Press.Find this resource:
Dillman, D. A. (2000). Mail and Internet surveys: The tailored design method (2d ed.). New York: J. Wiley.Find this resource:
Downs, A. (1972). Up and down with ecology—issue-attention cycle. The Public InterestNo. 28, 38–50.Find this resource:
Dunlap, R. E., & McCright, A. M. (2010). Climate change denial: Sources, actors and strategies. In C. Lever-Tracy (Ed.), Routledge handbook of climate change and society (pp. 240–259). Abingdon, Oxon, U.K.: Routledge.Find this resource:
Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. American Psychologist, 49(8), 709–724.Find this resource:
Finucane, M. L., Alhakami, A., Slovic, P., & Johnson, S. M. (2000). The affect heuristic in judgments of risks and benefits. Journal of Behavioral Decision Making, 13, 1–17.Find this resource:
Freud, S. (1924). Collected papers. London: Hogarth.Find this resource:
Galton, F. (1880). Psychometric experiments. Brain, 2, 149–162.Find this resource:
Gleason, K. L., Lawrimore, J. H., Levinson, D. H., Karl, T. R., & Karoly, D. J. (2008). A revised US climate extremes index. Journal of Climate, 21(10), 2124–2137.Find this resource:
Google Trends. (2017). Global warming + ozone hole news search. Retrieved from http://bit.ly/2nuW5fb.
Hart, P. S., & Leiserowitz, A. A. (2009). Finding the teachable moment: An analysis of information-seeking behavior on global warming related websites during the of the day after tomorrow. Environmental Communication-a Journal of Nature and Culture, 3(3), 355–366.Find this resource:
Hollway, W., & Jefferson, T. (2008). The free association narrative interview method. In L. M. Given (Ed.), The SAGE Encyclopedia of Qualitative Research Methods (pp. 296–315). Sevenoaks, CA: SAGE.Find this resource:
Isen, A. M. (1993). Positive affect and decision making. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 261–277). New York: Guilford Press.Find this resource:
Janis, I. L., & Mann, L. (1977). Decision making: A psychological analysis of conflict, choice, and commitment. New York: Free Press.Find this resource:
Joffe, H., & Elsey, J. W. B. (2014). Free association in psychology and the grid elaboration method. Review of General Psychology, 18(3), 173–185.Find this resource:
Johnson, E. J., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of Personality and Social Psychology, 45, 20–31.Find this resource:
Kahneman, D., & Snell, J. (1990). Predicting utility. In H. J. Einhorn & R. M. Hogarth (Eds.), Insights in decision making: A tribute to Hillel J. Einhorn (pp. 295–310). Chicago: University of Chicago Press.Find this resource:
Keller, C., Visschers, V., & Siegrist, M. (2012). Affective imagery and acceptance of replacing nuclear power plants. Risk Analysis, 32(3), 464–477.Find this resource:
Kempton, W., Boster J. S., & Hartley, J. A. (1995). Environmental values in American culture. Cambridge, MA: MIT Press.Find this resource:
Kess, J. F. (1992). Psycholinguistics. Amsterdam: John Benjamins.Find this resource:
Leiserowitz, A. (2005). American risk perceptions: Is climate change dangerous? Risk Analysis, 25(6), 1433–1442.Find this resource:
Leiserowitz, A. (2006). Climate change risk perception and policy preferences: The role of affect, imagery, and values. Climatic Change, 77(1–2), 45–72.Find this resource:
Leiserowitz, A., Feinberg, G., Rosenthal, S., Smith, N., Anderson, A., Roser-Renouf, C., et al. (2014). What’s in a name: Global warming vs. climate change. New Haven, CT: Yale University and George Mason University.Find this resource:
Leiserowitz, A., Maibach, E., Roser-Renouf, C., & Smith, N. (2011). Global warming’s six Americas, May 2011. New Haven, CT: Yale Project on Climate Change Communication.Find this resource:
Leiserowitz, A. A., Maibach, E. W., Roser-Renouf, C., Smith, N., & Dawson, E. (2013). Climategate, public opinion, and the loss of trust. American Behavioral Scientist, 57(6), 818–837.Find this resource:
Leviston, Z., Price, J., & Bishop, B. (2014). Imagining climate change: The role of implicit associations and affective psychological distancing in climate change responses. European Journal of Social Psychology, 44(5), 441–454.Find this resource:
Levy, D. A. (1997). Tools of critical thinking: Metathoughts for psychology. Boston: Allyn and Bacon.Find this resource:
Loewenstein, G. (1996). Out of control: Visceral influences on behavior. Organizational Behavior and Human Decision Processes, 65, 272–292.Find this resource:
Loewenstein, G., Weber, E., Hsee, C., & Welch, E. (2001). Risk as feelings. Psychological Bulletin, 127(2), 267–286.Find this resource:
Lorenzoni, I., Leiserowitz, A., Doria, M. D., Poortinga, W., & Pidgeon, N. F. (2006). Cross-national comparisons of image associations with “global warming” and “climate change” among laypeople in the United States of America and Great Britain. Journal of Risk Research, 9(3), 265–281.Find this resource:
Luedecke, G., McAllister, L., Nacu-Schmidt, A., Andrews, K., Boykoff, M., Daly, M., & Gifford, L. (2017). United States Coverage of Climate Change or Global Warming, 2000-2017. Retrieved from http://sciencepolicy.colorado.edu/icecaps/research/media_coverage/usa/index.html.
MacGregor, D. G., Slovic, P., Dreman, D., & Berry, M. (2000). Imagery, affect, and financial judgment. Journal of Psychology and Financial Markets, 1(2), 104–110.Find this resource:
Macmillan, M. (2001). The reliability and validity of Freud’s methods of free association and interpretation. Psychological Inquiry, 12(3), 167–175.Find this resource:
McCright, A. M., & Dunlap, R. E. (2011). Cool dudes: The denial of climate change among conservative white males in the United States. Global Environmental Change, 21(4), 1163–1172.Find this resource:
McGuire, J., & McGuire, C. V. (1991). The content, structure, and operation of thought systems. In R. S. Wyer & T. K. Srull (Eds.), Advances in social cognition: The content, structure, and operation of thought systems (Vol. 4, pp. 1–78). Hillsdale, NJ: Lawrence Erlbaum.Find this resource:
McGuire, W. J., & McGuire, C. V. (2001). Dimensions of the social mind: Size, asymmetries, congruence, and sex differences in thought systems focused on self or other persons. In J. P. Forgas, K. D. Williams, & L. Wheeler (Eds.), The social mind: Cognitive and motivational aspects of interpersonal behavior (pp. 25–45). Cambridge, U.K.: Cambridge University Press.Find this resource:
Mellers, B. A., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision Affect Theory: Emotional reactions to the outcomes of risky options. Psychological Science, 8(6), 423–429.Find this resource:
Myers, T. A., Nisbet, M. C., Maibach, E. W., & Leiserowitz, A. A. (2012). A public health frame arouses hopeful emotions about climate change. Climatic Change, 113(3–4), 1105–1112.Find this resource:
National Academies of Sciences, Engineering, and Medicine. (2016). Attribution of extreme weather events in the context of climate change. Washington, DC: The National Academies Press.Find this resource:
Nelson, D. L., McEvoy, C. L., & Dennis, S. (2000). What is free association and what does it measure? Memory & Cognition, 28(6), 887–899.Find this resource:
Peters, E., & Slovic, P. (1996). The role of affect and worldviews as orienting dispositions in the perception and acceptance of nuclear power. Journal of Applied Social Psychology, 26, 1427–1453.Find this resource:
Read, D., Bostrom, A., Morgan, M. G., Fischhoff, B., & Smuts, T. (1994). What do people know about global climate change? Survey results of educated laypeople. Risk Analysis, 14(6), 971–982.Find this resource:
Robinson, G. M. (1998). Methods and techniques in human geography. New York: John Wiley.Find this resource:
Rozin, P., Haidt, J., & McCauley, C. R. (1993). Disgust. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 575–594). New York: Guilford.Find this resource:
Sadoski, M., & Paivio, A. (2001). Imagery and text: A dual coding theory of reading and writing. London: Lawrence Erlbaum.Find this resource:
Siegrist, M. (2003). Perception of gene technology and food risks: Results of a survey in Switzerland. Journal of Risk Research, 6(1), 45–60.Find this resource:
Siegrist, M., Cousin, M. E., Kastenholz, H., & Wiek, A. (2007). Public acceptance of nanotechnology foods and food packaging: The influence of affect and trust. Appetite, 49(2), 459–466.Find this resource:
Siegrist, M., Earle, T. C., Gutscher, H., & Keller, C. (2005). Perception of mobile phone and base station risks. Risk Analysis, 25(5), 1253–1264.Find this resource:
Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin, 119, 3–22.Find this resource:
Slovic, P. (1997). Public perception of risk. Journal of Environmental Health, 59(9), 22(23).Find this resource:
Slovic, P. (1999). Trust, emotion, sex, politics and science: Surveying the risk-assessment battlefield. Risk Analysis, 19(4), 689–701.Find this resource:
Slovic, P. (2001). The perception of risk. London: Earthscan.Find this resource:
Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D.G. (2002). The affect heuristic. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 397–420). New York: Cambridge University Press.Find this resource:
Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D.G. (2007). The affect heuristic. European Journal of Operational Research, 177, 1333–1352.Find this resource:
Slovic, P., Flynn, J., Mertz, C. K., Mays, C., & Poumadere, M. (1996). Nuclear power and the public: A comparative study of risk perception in France and the United States. Eugene, OR: Decision Research.Find this resource:
Slovic, P., Layman, M., Kraus, N., Flynn, J., Chalmers, J., & Gesell, G. (1991). Perceived risk, stigma, and potential economic impacts of a high-level nuclear waste respository in Nevada. Risk Analysis, 11, 683–696.Find this resource:
Slovic, P., Flynn, J. H., & Layman, M. (1991). Perceived risk, trust, and the politics of nuclear waste. Science, 254(5038), 1603–1608.Find this resource:
Slovic, P., MacGregor, D. G., & Peters, E. (1998). Imagery, affect, and decision-making. Eugene, OR: Decision Research.Find this resource:
Smith, N., & Leiserowitz, A. (2012). The rise of global warming skepticism: Exploring affective image associations in the United States over time. Risk Analysis, 32(6), 1021–1032.Find this resource:
Smith, N., & Leiserowitz, A. (2013). American evangelicals and global warming. Global Environmental Change-Human and Policy Dimensions, 23(5), 1009–1017.Find this resource:
Smith, N., & Leiserowitz, A. (2014). The role of emotion in global warming policy support and opposition. Risk Analysis, 34(5), 937–948.Find this resource:
Szalay, L. B., & Deese, J. (1978). Subjective meaning and culture: An assessment through word associations. Hillsdale, NJ: Lawrence Erlbaum.Find this resource:
Szalay, L. B., Strohl, J. B., & Doherty, K. T. (1999). Psychoenvironmental forces in substance abuse prevention. New York: Kluwer Academic/Plenum.Find this resource:
Truelove, H. B. (2012). Energy source perceptions and policy support: Image associations, emotional evaluations, and cognitive beliefs. Energy Policy, 45, 478–489.Find this resource:
Wildavsky, A., & Dake, K. (1990). Theories of risk perception: Who fear what and why? Daedalus, 119(4), 41–60.Find this resource:
Wildavsky, A. (1991). Risk Perception. Risk Analysis, 11(1), 15–16.Find this resource:
Wilson, T. D., Lisle, D. J., Schooler, J. W., Hodges, S. D., Klaaren, K. J., & LaFleur, S. J. (1993). Introspecting about reasons can reduce post-choice satisfaction. Personality and Social Psychology Bulletin, 19(3), 331–339.Find this resource:
Wundt, W. (1883). Über psychologische Methoden. Philosophische Studien, 1, 1–38.Find this resource:
Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35, 151–175.Find this resource:
Zeelenberg, R., Shiffrin, R. M., & Raaijmakers, J. G. W. (1999). Priming in a free association task as a function of association directionality. Memory & Cognition, 27(5), 956–961.Find this resource: