The Deep End:
How Different Visuals Affect Audience
Engagement and Attitudes in Coral Bleaching
News Stories
Master’s Thesis
Authored by Mathieu Dasnois (student number 11896582) Supervised by Andreas Schuck
Date of Completion: 01/06/2018
University of Amsterdam, Graduate School of Communication Erasmus Mundus Master’s in Journalism, Media and Globalisation
Introduction ... 4
Theoretical Background ... 5
Methodology ... 14
Stimuli ... 14
Sample and participants ... 15
Survey procedure ... 16 Dependent variables ... 17 Moderators ... 19 Analysis ... 19 Results ... 20 Manipulation check ... 20 Main Effects ... 21
Moderated and mediated effects ... 23
Moderated mediation ... 25
Discussion... 25
References ... 30
Abstract
Coral bleaching is a consequence of climate change that poses serious risk to biodiversity and
the economic health of multiple regions. Despite this, the visual communication of coral
bleaching is difficult and not well understood. This study examines the effects of various
image types in coral bleaching news stories. In an online experiment with 223 participants,
four image types were compared in terms of issue engagement, importance, risk perception
and efficacy. The image types are: (1) stock images, (2) contrasting images, (3), infographics
and (4) solution-oriented images of coral farming. Results suggest that solution-oriented
images and infographics have the best chance to engage the reader/viewer. Engagement is
increased when either collective efficacy is increased, which happened most with images of
coral farming, or when risk perception is increased, which happened most with infographics.
These findings are discussed within the academic context of visual climate change
communication. It is hoped that the use of two separate processes to increase issue
engagement may inform the communication strategies of newspapers and civil society.
Keywords: coral bleaching; visual communication; climate change communication; image
Introduction
Climate change is one of the greatest challenges of this and future generations. The
effects will only be fully felt in the future, but the scientific consensus is that they will be
severe to catastrophic (Schmittner, Oschlies, Matthews, & Galbraith, 2008). Yet while the
discussion has focused on carbon reduction, much less time has been dedicated to the effects
of climate change on the ocean, such as coral bleaching. Coral bleaching is the – eventually
fatal – rejection of symbiotic algae from coral organisms due to a climate-change-induced
rise in ocean temperatures (Hughes et al., 2017). Coral reefs support 25% of marine life, one
billion people, massive tourism revenue, and protection from storms for isolated island
communities (Hughes et al., 2017; Witschge, 2018). If coral reefs collapse, the ocean
ecosystem is compromised, with consequences far beyond the coastline (Hoegh-Guldberg,
2007). Why then, has this issue received comparatively less attention, both politically and in
the media? Part of the answer may lie with problems of communication.
Brenthel (2016) argues that climate change is hard to communicate because of its
scale. The effects are global, the changes are subtle, and the consequences of present-day
actions lie in the future (Schroth, Angel, Sheppard & Dulic, 2014). This makes climate
change hard to grasp on a personal, individual level (O’Neill and Nicholson-Cole, 2009). Yet
climate change communication has made great strides. Visuals have become a key
component of good global warming communication (O’Neill & Smith, 2014). Key pictoral themes have emerged, from the early days of the polar bear on a floating iceberg, to the
looming smoking chimney spewing black smoke, to positive images of renewable energy
technology and the juxtaposition of green nature and brown industrialism (Brenthel, 2016;
León & Erviti, 2015; O’Neill & Smith, 2014; Schroth et al., 2014). These images render a
broader, more abstract concept into something tangible, understandable on a personal and
Whitmarsh 2007 for the need to make climate change communication both local and
individually accessible). Coral bleaching has no such communication strategies. While a
chimney stack spewing black smoke is easy to interpret as ‘bad’, most people may not be able to recognise a bleached coral as a dying coral. Nor is it easy to grasp just how much
marine life relies on coral reefs, or how long corals take to grow. In academia, the situation is
similar. While communication science has done excellent work analysing visual climate
change communication (Metag et al., 2016; O’Neill, 2013; O’Neill, Boykoff, Niemeyer & Day, 2013; O’Neill & Smith, 2014; Hart & Feldman, 2016a), no research has been done on the visual communication of ocean-related climate change consequences such as coral
bleaching or ocean acidification, nor has research compared the effects of diverse image
types.
Coral bleaching is thus an example of a hard-to-communicate problem and is tied to
similar issues that are also hard to visually communicate, such as acidification of the ocean
(Anthony, Kline, Diaz-Pulido, Dove & Hoegh-Guldberg 2008) and pollution (Danovaro et
al., 2008). Understanding visual coral bleaching communication therefore has implications
for the communication of climate change, ocean acidification and ocean pollution.
This paper aims to examine how readers react to different visuals of coral bleaching.
The research question is thus: to what extent does the use of different image types amplify or
attenuate the effects of coral bleaching news stories?
Theoretical Background
Anthropogenic climate change has been widely accepted within the scientific
community (IPCC, 2007; Oreskes, 2018), and yet this acceptance has not permeated public
consciousness (Brulle, Carmichael & Jenkins, 2012; Lorenzoni & Pidgeon, 2006). Climate
Jenkins, 2012) and communication has largely failed to bring about individual behavioural
change on a global scale (Lorenzoni et al., 2007). Climate change feels temporally and
spatially removed (Benthel, 2016; Roeser, 2012; Schroth et al., 2014). One of the key
hurdles of climate change communication has therefore been to personalise and localise
climate change communication (O’Neill & Nicholson-Cole, 2009). For some aspects of
climate change, however, this is less feasible. Coral bleaching is a threat to tropical coral
reefs (and this is not the case for every such reef, the Red Sea’s corals have proven surprisingly resilient. See Osman et al., 2018). The effects are global (Hughes et al., 2017)
but neither the problem nor the solutions are easily localised.
A parallel solution for better climate change communication lies in better visuals
(Metag et al., 2016; O’Neill, 2013; O’Neill et al., 2013; O’Neill & Smith, 2014; Hart & Feldman, 2016a). Images tend to provoke more emotions than text alone (Childers &
Houston, 1984), are more likely to grab the attention of readers (David & Kang, 1998;
Jenkin & Harris, 2001) and to stay in their minds (Berry & Brosius, 1991; Gibson & Sargent,
1999). This study will compare the effects of different image types when embedded in the
same news article. These image types are: (1) a stock image, (2) contrasting images, (3) a
positive, solution-oriented image of a coral farm and (4) an infographic. The aim of this
study is to measure how much (a) issue importance, (b) risk perception and –most critically-
(c) issue engagement these images engender in the viewer/reader. These variables will be
discussed and defined before examining the literature on different image types.
Li, Kim and O’Boyle (2017) convincingly argued that personal engagement with an issue involves “perceived issue importance, risk, and involvement” (Li, Kim & O’Boyle,
2017, p.772). This triple approach will be adopted, but involvement will be referred to as
image types will therefore be compared in terms of issue importance, perceived risk and
issue engagement.
Issue importance has been convincingly tied to engagement (Baumgartner & Jones, 1993; Jones & Baumgartner, 2005; Weaver, 1991 all found in Hart & Feldman, 2016a).
When discussing issue importance, Alexander and Jetton’s (1996) definition of “reader determined or constructed importance” (p. 93) is useful, it defines a type of importance that is more personal and less abstract. According to Alexander and Jetton (1996), importance is
strongly tied to interest, and may be interchangeable with it. Li, Kim and O’Boyle (2017)
and Bernstein’s (2005) consider importance to be similar to personal relevance1.
Roeser (2012) argues that climate change communication is missing a sense of
urgency and that this lack of urgency stems from reduced risk perception among the public.
Risk has been relatively well researched in psychology (Rundmo, 2002 links images of risk
and risk perception) and health communication (Dillard et al., 2012 links risk perception and
behaviour), but the link between risk perception and climate change communication is less
evident (Xue et al., 2016). That being said, climate change communication includes a
message of risk or threat (Hart and Feldman, 2014) and DeBono, Vincenti and Calleja
(2010) found that higher perceived risks from climate change did lead to both support for
and engagement with climate change mitigation. Xue et al. (2016) link high-risk
environmental messages to what other researchers call impact-oriented communication, a
very common component of climate change communication. Much of climate change
communication is an attempt to communicate risk so as to change behaviour. Risk
perception is thus a crucial measure of the effectiveness of climate change communication.
1 Issue importance in visual climate change communication has been studied in the past
(Hart and Feldman, 2016a; O’Neill et al., 2013) and their findings will be used to form later hypotheses, but the definition of importance as the “perceptions of the importance of the issue” (Hart and Feldman, 2016a, p. 419) is likely to include a less personal, less involved, more abstract kind of importance than what I hope to measure.
Finally, the ultimate goal in most climate change communication is to incite
behavioural change (Chen, 2015; Lauren et al., 2016; O’Neill et al., 2013; O’Neill & Nicholson-Cole, 2010; Wang, 2018). Issue engagement is defined as how an individual
might personally get involved (referred to as personal political behaviour by Hart and
Feldman, 2016a and as intended political participation by Hart and Feldman, 2016b). Since
the issue of coral bleaching is quite far removed from individuals in (for example)
landlocked European countries, issue engagement in this study does not include actions such
as using less energy, but focuses on civil and political actions. Issue engagement is the key
dependent variable, the measure of behaviour that climate change communication ultimately
seeks to increase.
Having defined the variables we hope to measure, each image type will be discussed
in more detail. Stock images are generic, easy-to-source images of a bleaching coral reef,
often found in coral bleaching news stories. Very little research has been done on the use of
stock images in news stories, and none of that research applies to coral bleaching. While
some images of coral bleaching may make good use of lighting and dead space, others
simply feature a white, bleached coral. This may be confusing to many readers, for whom a
white coral may not imply a dead coral. Drawing from the field of pedagogy, when readers
were shown an article with/without an image of abstract and impressionist artwork (an
example of an image not accessible to some readers) the visuals alone had no significant
effect on understanding or recall (Brante, Olander & Nyström, 2013). This is a markedly
different result from news communication studies mentioned earlier, in which images
increased understanding and recall.
Contrasting images are also often used in coral bleaching news stories. Contrasting images show the reader a coral reef before and after coral bleaching. Since readers may not
presented next to each other. Interestingly, contrasting visual frames are also used within
climate change communication, for example green vs brown or nature vs. industrialism
(Brenthel, 2016). Unfortunately these contrasts have not been extensively studied, hence no
directional hypothesis can be formed. The first sub-research-question is therefore:
Sub-RQ 1: How do readers respond to a stock image compared to contrasting images and no
image?
In contrast to images that show the impact of coral bleaching like stock and
contrasting images, one aspect of visual climate change communication is far more established: people react very differently to positive, solution-oriented images. O’Neill and
Nicholson-Cole (2009) found that positive, non-fearful images increased engagement for
participants in the United States, United Kingdom and Australia. Metag et al. (2016)
replicated the study using the same images in Germany, Switzerland and Austria, finding
similar results. Hart and Feldman (2016a) found increased engagement with images of
renewable energy solutions. On the other hand, both O’Neill and Nicholson-Cole (2009) and
Metag et al. (2016) found that impact-oriented images were better at raising issue
importance. People exposed to negative images of impacts were shocked but also “powerless
and overwhelmed” (O’Neill & Nicholson-Cole, 2009, p. 374). People exposed to positive, solution-oriented images were more likely to want to get involved but less likely to view the
issue as crucially important.
The following hypothesis seeks to test whether this finding from climate change
communication can be translated to coral bleaching communication, and whether images of
H1a: An image of coral farming increases engagement
H1b: An image of coral farming reduces issue importance
Alongside increasing engagement and reducing importance, O’Neill et al. (2013) found that solution-oriented images were more likely to make people feel that they could do
something about climate change on a personal level. Such images therefore increased the
feeling of efficacy, i.e. the power to effect change, in readers/viewers (Hart and Feldman,
2016a,b). This thesis focuses on collective efficacy, the belief that a group (to which the
reader/viewer belongs) can effect change. Homburg and Stolberg (2006) directly compared
individual efficacy to collective efficacy and found that collective efficacy was a more
complete predictor of environmental behavioural change within cognitive stress theory. Chen
(2015) found a similar result among Taiwanese respondents. Collective efficacy is therefore
a key component of issue engagement. If people feel that we are powerless to make a
difference, they are unlikely to get involved in trying to make a difference. The hypothesis
predicts that solution-oriented images of coral farming will increase feelings of collective
efficacy.
H2: Images of coral farming increase feelings of collective efficacy
The last image type commonly used in environmental communication is the
infographic (Lazard & Atkinson, 2015). In one of the earlier studies of infographics in news articles, Pasternack and Utt (1990) found that infographics were successful in quickly
conveying a lot of information to the reader, but not for all readers. An update by Utt and
Pasternack (2000) found that editors consider infographics to aid understanding, a finding
studied environmental infographics in particular and found that infographics led to greater
elaboration, i.e. “volume of thoughts…. the vividness of these thoughts and the sensitivity to the message” (p. 17). While this has no direct parallel in terms of importance, risk and engagement, this thesis will measure whether respondents thought the image2 had increased their understanding of the issue. The next hypothesis is therefore
H3: Infographics generate greater understanding than a stock image or no image.
Before more complex models can be tested, moderators need to be considered. People
with first-hand knowledge of corals are more likely to be affected by a coral bleaching story.
As discussed in sub-RQ1, stock images are unlikely to resonate with non-divers, but people
who do know what a coral reef looks like, such as snorkellers and scuba-divers, might be
more moved by a stock image of a bleached coral. Readers/viewers with underwater
experience may therefore feel more engaged by a stock image than divers, while
non-divers might be more drawn to a visually clearer contrasting image. No research exists to
form a firm directional hypothesis, hence:
Sub-RQ2a: Are non-divers more likely to be engaged by a contrasting image than a stock
image?
2 It should be noted that infographics are also very useful at communicating the scale of the
problem and the possible solutions. For the sake of comparison with other image types, however, this thesis focuses on infographics that directly show and explain coral bleaching.
A corollary of this question also applies to coral farming. The young corals and their
future as a coral reef would presumably be most visually accessible to those respondents
with underwater experience.
Sub-RQ2b: Are divers more likely to feel engaged when seeing coral farming images?
Once the moderator is confirmed or discarded, more complex models of issue
engagement can be tested. Little research has been done directly measuring the effectiveness
of infographics on issue engagement. However, some research has been done
communicating risk via infographics3 (Damman et al., 2018; Scott et al., 2016; Shin, 2016) and risk perception to engagement (Li et al., 2017). Shin (2016) found that infographics are
regularly used in health and risk communication as well as to effect behavioural change (see
also Scott et al., 2016). Damman et al. (2018) found that cardiovascular information
conveyed in an infographic slightly increased risk comprehension, though not in all
participants. Lazard and Atkinson (2015) argue that environmental infographics increase
issue-relevant thinking and have potential in persuasion. In terms of the relationship between
risk and engagement, Li et al. (2017) found that, among students discussing gender violence,
the seriousness of an issue (in our case risk) is connected to the need to improve the situation
(engagement). Our hypothesis is therefore:
H4: When viewing infographics, risk perception increases which leads to more engagement
with the issue
3 While this is not entirely analogous, as the infographic of this thesis will communicate
‘what coral bleaching is’ so as to be comparable to the other image types, this research nevertheless serves as the beginning of a hypothesis.
Multiple researchers have also found that solution-oriented images lead to increased
efficacy (as discussed for H2, see Hart and Feldman, 2014; Hart and Feldman, 2016a; O’Neill and Nicholson-Cole, 2009) and that efficacy led to increased engagement (Chen, 2015; Wang, 2018; Xue et al, 2015). Hart and Feldman (2016b) conducted an online survey
of US respondents and found that several forms of efficacy mediated political participation.
Lauren et al. (2016) similarly found that efficacy led to more complex and difficult
pro-environmental behaviour in Australian respondents. Multiple studies have established a
direct link from efficacy to engagement (Xue et al., 2015 for personal and Wang, 2018 for
collective efficacy, with Chen, 2015 finding that collective efficacy was the more effective
measure). This study will seek to verify these results in the new context of coral bleaching,
using coral farming images.
H5: Images of coral farming increase collective efficacy which leads to more issue
engagement.
Thus this thesis explores two possible paths to issue engagement, using coral farming
images or infographics. Research has shown a link between efficacy and engagement. What
is less clear is whether an individual with a high belief in his or her own problem-solving
abilities will be more likely to believe that a larger issue can be solved. In other words,
whether an individual with high General Self-Efficacy (GSE) would be more likely to feel
increased feelings of collective efficacy in response to solution-oriented images. Homburg
and Stolberg (2014) as well as Chen (2015) showed a link between individual and collective
efficacy in pro-environmental behaviour, but both authors were focused on a direct
comparison. Moving beyond the field of environmental communication, Roos et al. (2014)
increased self-efficacy might lead to increased collective efficacy. This is unproven but
plausible. Koletsou and Mancy (2011) compared self-efficacy and collective efficacy on a
purely theoretical level vis a vis climate change mitigation, arguing that collective efficacy in
relation to climate change (or an associated problem such as coral bleaching) is essentially a
problem of cooperation. Cooperation is itself an individual action (Mancy, 2011). The
willingness to cooperate, therefore, might stem from or be influenced by personal efficacy.
The hypothesis is thus:
H6: Individual General Self-Efficacy (GSE) moderates the relationship between
solution-oriented coral farming images and collective efficacy, in such a way that the effect is
stronger for individuals with higher levels of GSE, leading to increased engagement.
Methodology
A controlled experiment was used to investigate the influence of various image types
on coral bleaching stories. Participants were randomly assigned to see a coral bleaching story
with either a stock image of a partly-bleached coral reef, a contrasting before-and-after
picture, an infographic explaining coral bleaching, an image of a coral farming station with
divers in the frame, or no visual element whatsoever (the control group). Respondents were
encouraged to take the survey on a laptop or computer to increase the image size and its
associated importance in comparison to the text (Wanta, 1988) though this instruction may
have been diluted by the snowball nature of the sample.
Each respondent was shown a news story about coral bleaching. The words of this
story did not differ between respondents, and included information from actual news stories,
lightly edited, intended to give a brief overview of coral bleaching both globally and in a
specific location (the Seychelles). The story included both dry factual information and human
elements, as well as a brief mention of coral farming, so that all possible images would fit the
story. The images in each category were chosen based on a pre-test of 41 participants. During
the pre-test, respondents were randomly shown three out of thirteen possible images (three
per category and four for infographics). Respondents were asked if the images made them
feel hopeful, sad, moved, whether they found the image to be positive or negative, whether
the image was visually powerful, whether the image was clear and accessible, how well the
image conveyed coral bleaching as a concept and how much they thought it increased their
understanding of the subject. Pre-test responses showed that stock images were least
engaging, contrasting images were saddest and most negative, infographics were clearest and
coral farming images produced the most hope. Based on these answers a final selection of
one image per category was made. The final selection shows similar corals in similar lighting,
with only the coral farming image having human subjects in the photo.
Sample and participants
Participants (N=223) were recruited through a mixture of convenience online
sampling and snowball sampling. Druckman and Kam (2009) as well as Mullinix, Druckman
and Freese (2014) have argued that convenience sampling does not create a statistically
are known and accounted for4. The mean age was 34 years old (SD=15.27) and varied from 19 to 79 years. 131 participants (59%) identified as female, with 89 male and 2 other. Most
participants had a Bachelor’s degree (36%), a Master’s degree (32%) or PhD (5%). 26% had either a diploma, a high school degree or had not completed high school. Most participants
identified themselves as being slightly to the left on a political philosophy scale, averaging a
mean value of 4.28 on an 11-point scale (SD=2.18). Respondents were on average unaware
of coral bleaching before the survey, with 53% of participants answering that they knew
nothing about coral bleaching or very little. 30% answered that they knew “a moderate amount” and 17% knew either “a lot” or “a great deal” about coral bleaching.
Randomisation was successful. There was no significant age difference between the
groups (F(4, 216)=1.76, p=.138), with the control group (no visual) having a slightly older
average age of 40. There was no significant difference in gender between the groups (ꭓ2(8, N
= 223) = 2.71, p = .951) nor in education levels (ꭓ2(20, N = 223) = 19.87, p = .473). Because participants who have seen corals underwater are more likely to be interested and engaged by
coral bleaching, scuba-divers (ꭓ2(4, N = 223) = 2.54, p = 0.689) and snorkellers (ꭓ2(4, N = 223) = 2.45, p = .654) were also tested for randomisation between the groups. The variation
was insignificant in both cases.
Survey procedure
Respondents were briefed about their rights and their ability to refuse to participate.
Respondents were told the topic of the survey (coral bleaching) but were not primed about
visuals or visual differences. Demographic information was sought first, including political
ideology, background knowledge of coral bleaching, and whether respondents were
4For this reason, additional potential moderators and mediators were gathered in the survey
divers, snorkellers, or neither. Individual efficacy was then tested as this was not expected to
vary based on the stimulus. Respondents were then randomly sorted into five groups based on
the four different visual elements and the control group. Once respondents had seen the
stimulus, they completed a manipulation check (detailed in the results section, below).
Respondents’ emotions were then measured by asking if they found the story moving, sad, hopeful, interesting, engaging, positive or negative, whether the story made respondents care
and whether the story made respondents feel they had an increased understanding. To reduce
the influence of the order in which the questions are asked (known as order bias, see
Tourangeau, Rips & Rasinski, 2000), the order of emotional questions was randomised. After
the questions pertaining to emotions, respondents were then asked about issue importance,
issue engagement, issue urgency and collective efficacy. The order of these various blocks of
questions was randomised alongside an attention test asking “What is the answer to 1+2=? Even though it is incorrect, please choose the answer ‘4’”. 96.4% of respondents correctly answered ‘4’. Finally, respondents were asked recall questions about the article, before being thoroughly debriefed about both the nature of the research and the veracity of the news
article.
Dependent variables
The operationalisation of issue importance started with a direct question adapted from
Hart and Feldman (2016a), asking respondents whether or not they agreed that “The article I
read made me feel that coral bleaching is an important issue” on a 7-point Likert scale. Furthermore, this thesis sought to test reader-determined or constructed importance
(Alexander & Jetton, 1996) rather than a feeling of abstract importance removed from the
personal sphere. This type of importance is tied to both interest (Alexander & Jetton, 1996),
also Casas & Williams, 2017 and Chen & Dredze, 2018 for visuals on social media). Hence
readers were asked how likely they were to seek out further information on coral bleaching
and share the story online. These three measures were combined into a single 7-point
measure of importance (α=.65, M=4.55, SD=1.21).
Issue Engagement was operationalised based on a battery of five questions, taken
from Hart and Feldman (2016a, b). Respondents were asked whether they would contact
government officials, participate in a rally, sign a petition, volunteer or donate to an
organisation working with either coral bleaching or climate change5, each on a seven-point Likert scale. These questions were combined into a single seven-point measure, with α=.82,
M=3.67, SD=1.44.
Risk perception measured the “seriousness of the threat” (Hart & Feldman, 2014, p.
334) and the operationalisation was loosely adapted from an experiment by DeBono et al
(2010) on the risks of climate change. Respondents were asked whether they thought it more
or less likely that coral bleaching would “occur more frequently”, that coral bleaching “would get worse”, that there would be “severe economic damages as a consequence of coral bleaching” and that “climate change will result in a situation that is out of control”. As with engagement, climate change substituted for coral bleaching (see Hughes et al., 2017 showing
the link between climate change and coral bleaching, also shown to respondents in the
stimulus) in two questions where coral bleaching was too far removed from the daily lived
reality of most respondents. Risk perception was then combined into a single scale from 1 to
7 (α=.80, M=6.25, SD=0.84).
Collective efficacy measures were taken from Wang, 2018, and Chen, 2015.
Respondents were asked seven questions on a Likert scale, asking if they agreed with
5 Many respondents are far removed from both corals and coral bleaching and thus government
contact, rallies etc in favour of coral bleaching would be nonsensical. For this reason the term of climate change was added, since climate change has been very strongly linked to coral bleaching by forty scientists (Hughes et al., 2017, also shown to respondents in the stimulus).
statements from “the world can reduce the effects of climate change and coral bleaching” to “my country can reduce the effects of climate change and coral bleaching” to “my community can make a difference”, “I am confident that together we can solve the problems of climate change” and “we can come up with creative ideas to solve these problems” as well as one negatively worded question asking if it is too late to address coral bleaching. These
questions were then combined into a single measure from 1 to 7 (α=.82, M=5.39, SD=0.95)
Understanding was measured by asking the question “To what extent do you think this story increases your understanding of coral bleaching?” The answers ranged on a 5-point
scale from “a great deal” to “not at all” and were then reverse-coded in SPSS. (M=3.11, SD=0.95)
Moderators
Snorkelling and scuba-diving activity were measured by asking respondents whether
they did one activity, both or neither. 42 participants engaged in scuba-diving activities, while
127 people snorkelled (the categories were not exclusive) and 89 engaged in neither activity.
The scale for individual efficacy was taken from Schwarzer & Jerusalem (1995)’s General Self Efficacy (GSE) scale. From the original ten questions of the GSE scale, six were
used, cutting out redundant questions for the sake of survey length. Although there is some
debate about efficacy scales and the GSE scale (Bandura, 2006), Scherbaum, Cohen-Charash
and Kern (2006) found the GSE scale to be adequately reliable and comparably effective to
the other two tested measures, in most parameters. GSE was combined into a single
four-point scale, with α=.79, M=3.10 and SD=0.43.
The analysis was done in two steps. The first was the use of Analysis of Variance tests
(ANOVA) and t-tests to investigate the main effects of the different image types while
simultaneously testing sub-research question one and hypotheses one to three. The second
step involved a mediation analysis to better understand the role of people with a particular
affinity for the subject matter, in this case snorkellers and scuba-divers6 and test two different paths to increased issue engagement. This mediation analysis was accomplished using the
SPSS PROCESS macro developed by Andrew Hayes (2013). For all tests, 5000 sampling
iterations were used to approximate a larger sample size and unless otherwise specified, all
confidence intervals represent a 95% probability. Since the PROCESS macro does not always
list probability values, when such values are not listed confidence intervals that do not
straddle zero may be taken as significant equivalent to p<0.05.
Results
Manipulation check
To confirm that manipulation was successful, respondents were asked if they had seen “an infographic (text overlayed on an image) describing coral bleaching”, “an image of a coral reef”, “two contrasting images before and after coral bleaching” or “an image of a coral
6Respondents were asked if they engaged in snorkelling and/or scuba-diving or neither. Underwater
experience was initially coded as ‘0’ for neither and ‘1’ for either snorkelling or scuba-diving (respondents who checked both answers were only counted once). Testing showed that the
difference between snorkelling and scuba-diving did prove significant on some measures, and hence another variable was created coded ‘0’ for neither, ‘1’ for snorkelling and ‘2’ for scuba-diving. Both variables were used for all tests. Unless otherwise specified, the latter ‘0-1-2’ variable is referred to by ‘underwater experience’ in this thesis. A variable coded ‘0’ for either no-experience or
snorkelling, and ‘1’ for scuba-diving was tried but proved insignificant, showing that there was a difference between ‘no experience’ and ‘snorkelling’.
farm”. Respondents in each group significantly more often (ꭓ2(12, N = 223) = 238, p < 0.001)
indicated7 that they had seen the correct image.
Main Effects
Sub-Research Question number 1 sought to compare contrasting images, a stock
image and the control group, in terms of issue importance, risk perception and engagement.
Measuring issue importance, there was no significant difference between the groups, at F(4,
218)=1.20, p=.313. The results showed no significant differences between all image types
when measuring risk perception, at F(4, 218)=1.09, p=.360. However, there was a significant
difference between the infographic image group (M=6.41, SD=0.54) and the control group
(M=6.05, SD=0.94) at t(79)=2.07, p=0.042, meaning that risk perception was increased by an
infographic image compared to the control group. Measuring engagement between all five
groups showed no significant difference at F(4, 218)=1.29, p=.275. The biggest difference
was between the infographic (M=3.93, SD=1.19) and the control group (M=3.43, SD=1.44)
and was significant in a one-tailed test at t(79)=1.72, p=.044, meaning that engagement was
increased by an infographic image compared to the control group, as expected (Lazard and
Atkinson, 2015).
In terms of our sub-RQ1, therefore, results showed no significant difference between
a stock image, contrasting images or no image when measuring engagement, issue
importance or risk perception. All three groups tended to perform less well than either
infographics or coral farming images though the differences were minor.
7 Respondents who saw a stock image (N=42) correctly identified it as an image of a coral reef 86%
of the time. Contrasting images (N=41) were either identified correctly (44%) or as an infographic (49%) which, due to the description of text over an image, cannot be considered entirely incorrect. Respondents who saw an infographic (N=43) correctly identified the type 88% of the time. Finally, respondents who saw an image of an underwater coral farm (N=59) either identified the image correctly (60%) or as a coral reef (32%) which, as a future coral reef, is not incorrect.
Hypothesis 1 posited that a solution-oriented image of coral farming would increase
engagement but reduce issue importance. Results showed no significant differences in
engagement between the coral farming image (M=3.74, SD=1.64) and either no image
(M=3.43, SD=1.44), a stock image (M=3.83, SD=1.33) or a contrasting image (M=3.35,
SD=1.46). Following up from the ANOVA listed in Sub-RQ1 showing the differences in issue importance (F(4, 218)=1.20, p=.313), post-hoc tests show the largest difference was
between the coral farming group and the contrasting group (Mdifference=0.46, p=.079) followed
by the control group (Mdifference=0.40, p=.086). These results were contrary to expectations
and did not merit a one-tailed test. Coral farming images therefore do not lead to increased
engagement (H1a) nor do they lead to decreased importance (H1b). H1 is therefore rejected.
Hypothesis 2 posited that coral farming images would increase feelings of collective
efficacy. The results did not show any significant difference between the image types at F(4,
217)=0.66, p=.618. Post-hoc tests between respondents who saw a farming image and
respondents who saw either a stock image (Mdifference=0.04, p=.841) or no image
(Mdifference=0.23, p=.255) were both insignificant. Coral farming images did not lead to
feelings of increased collective efficacy. H2 is therefore rejected.
H3 sought to test whether respondents exposed to an infographic would experience
increased understanding and recall. The different image types did not significantly differ with
regards to self-reported understanding, at F(4, 218)=1.17, p=.326. The difference between the
infographic group (M=3.33, SD=1.09) and the stock image group (M=2.93, SD=0.96) was
significant in a one-tailed test at t(83)=1.86, p=.033. The direction of the effect was as
expected (Lott, 1990). The difference between the infographic group and the control group
(M=3.00, SD=0.96) was insignificant at t(79)=1.42, p=.159. Infographics did lead to slightly
increased understanding when compared to the stock image group, but not convincingly
Moderated and mediated effects
Sub-RQ2a sought to examine whether stock images were more engaging for those
with more experience of the underwater world, particularly compared to contrasting images.
Contrasting images, being easier to grasp, might plausibly appeal to non-divers and those
with less experience in the water. A PROCESS model-1 test showed that an absence of
underwater experience had no effect on engagement, with b=0.09, SE=0.45, p=.800.
Snorkellers had a higher though insignificant effect with b=0.61, SE=0.32, p=.064, 95% CI:
[-0.04;1.26]. Divers had a larger, still insignificant effect with b=1.13, SE=0.62, p=.074, 95%
CI: [-0.11;2.37]. The results show that underwater experience has an insignificant effect at p<0.5 levels and is only significant at p<0.1 levels, though the difference between divers and snorkellers is not as large as between non-divers and snorkellers. By using a dichotomous
variable with ‘no experience’ on one side and both divers and snorkellers on the other8, the
moderating effect becomes significant, with b=0.79, SE=0.40, p=.0491, 95% CI: [0.0031;
1.58] for snorkellers/divers. Respondents who engaged in either snorkelling or diving were
more likely to be engaged by a stock image showing coral bleaching, than by two contrasting
images, as expected.
Sub-RQ2b sought to understand whether those with underwater experience would
react more favourably to images of a coral farm than non-divers. Using a PROCESS model-1
test between farming images and the control group, there was no effect for respondents with
no diving experience. Respondents who engaged in snorkelling were slightly more likely to
feel engaged by coral farming images, though the result was only significant at a p<0.1 level,
with b=0.60, SE=0.31, p=.0563, 95% CI: [-0.016; 1.221]. Respondents who were
scuba-divers were significantly more likely to feel engaged by coral farming images, with b=1.53,
SE=0.55, p=.0069, 95% CI: [0.431; 2.634], in line with expectations. Underwater experience is thus a significant moderating variable for scuba-divers seeing coral farming images.
Hypothesis 4 sought to test whether the mix of images and text of an infographic
would produce increased feelings of risk, and whether this would mediate a relationship with
engagement. A PROCESS model-4 test was used to compare the infographic image group
with the control group, with underwater experience as a co-variate (control) variable. The ‘A’
path between exposure to infographics and increased feelings of risk proved significant, at
b=0.359, SE=1.17, p=.0370, 95% CI: [0.022; 0.695]. The ‘B’ path between increased feelings of risk and increased issue engagement was only significant at a p<0.1 level, with b=0.351,
SE=0.30, p=.0749, 95% CI: [-0.036; 0.738]. The total indirect effect of infographics on engagement was b=0.126, SE=0.80, 95% CI: [-0.0016; 0.426]. The effect was therefore
insignificant at a p<0.05 level, and significant at a p<0.1 level. Infographics therefore did
lead to increased risk perception but increased risk perception only tentatively led to
increased issue engagement. Thus, we find tentative support for our expectations even though
we have to formally reject our hypothesis.
Hypothesis 5 sought to expand on the established link between efficacy and
engagement by testing whether increased feelings of collective efficacy would mediate issue
engagement. A PROCESS model-4 test was used, testing farming images against the control
group, with underwater experience as a co-variate (control) variable. The ‘A’ path between
farming images and collective efficacy showed an insignificant trend, with b=0.291,
SE=0.20, p=.149, 95% CI: [-0.106; 0.687]. The ‘B’ path between collective efficacy and increased engagement proved significant, at b=0.51, SE=0.15, p=.001, 95% CI: [0.209;
0.816]. The total indirect effect of the mediation was insignificant by conventional standards
of p<0.05 at b=0.15, SE=0.11, 95% CI: [-0.022; 0.447] but significant at a p<0.1 level. Thus
efficacy, does lead to significantly increased issue engagement. Hypothesis 5 is therefore
only tentatively and partially supported.
Moderated mediation
Hypothesis 6 is an extension of H5, adding a moderation effect of personal efficacy,
on the path between the farming image and collective efficacy. H6 therefore posits that when
seeing a farming image, those with high personal efficacy are more likely to feel higher
collective efficacy in response to the image. Collective efficacy in turn increases engagement.
H5 showed that higher collective efficacy does indeed lead to higher issue engagement, but
the ‘A’ path between an image of coral farming and feelings of collective efficacy, proved
insignificant. However, in line with expectations, using a PROCESS model 7 and controlling
for underwater experience, personal efficacy becomes a significant moderator at higher levels
(one standard deviation above the mean), with b=0.26, SE=0.16, 95% CI: [0.036; 0.634].
Respondents with higher-than-average personal efficacy were thus significantly more likely
to feel a higher sense of collective efficacy and in turn a higher sense of issue engagement,
when seeing a coral farming image. The null hypothesis is therefore rejected and H6 is
supported for those with higher than average GSE.
Discussion
This study adds to the growing literature on climate change visuals by incorporating a
less-known area of climate change: coral bleaching. While impact-oriented images,
solution-oriented images and infographics have been researched, they have not been directly
This thesis sought to better understand the effects of four different image types on
news stories related to coral bleaching. On average, images performed better than the control
group in terms of issue importance, engagement and risk perception. This confirms the body
of work within visual communications showing the engaging effects of images in a news
story (Childers and Houston, 1984; David and Kang, 1998; Jenkin and Harris, 2001; O’Neill, 2013; Zillmann, Gibson and Sargent, 1999). Comparing image types, the results show some
support for the effects of solution-oriented images (in this case coral farming) over
impact-oriented images (both stock images and contrasting images). Infographics also led to slightly
higher understanding and engagement, through increased risk perception. Both the coral
farming image and the infographic performed slightly above average in terms of issue
importance, engagement and risk perception.
This paper adds to the understanding of the complex mental path from seeing a
solution-oriented image to feeling increased issue engagement (Hart & Feldman, 2014; Hart
& Feldman, 2016a,b; Metag et al., 2016; O’Neill et al., 2013; O’Neill & Nicholson-Cole, 2009) even though the story shown to respondents focused more on the issue of coral
bleaching than on the solution of coral farming9. This finding supports existing literature that readers respond better to hope than to fear (O’Neill & Nicholson-Cole, 2009) and to solutions better than to impacts (Hart & Feldman, 2014). Contrary to expectations, however, the coral
farming image also produced more issue importance in viewers/readers. While importance
and engagement may seem intuitively linked, multiple studies (Metag et al., 2016; O’Neill &
Nicholson-Cole, 2009) have found that impact-oriented images increase importance in
comparison with solution-oriented images. Respondents who see an image of a flood may
9 The stimulus was based on three real stories, see the methodology section for details, and followed
journalistic principles. To create a story that could fit four different images, it was necessary to craft a story about coral bleaching rather than a story about coral farming, and then adapt that story to include a section on coral farming. The logical consequence was that the theme of the story remained coral bleaching and not coral farming.
rank the issue as important while not being motivated to get personally engaged. While Hart
and Feldman (2016a) found no such correlation, this study finds the opposite: coral farming
images lead to increased issue importance.
I suggest two possible explanations for this. It is plausible that seeing human beings
who clearly saw the issue as important enough to do something about (a key component of
positive climate change images according to Brenthel, 2016), created an emotional contagion
from the subject to the viewer (see Baberini,Coleman, Slovic, & Vastfjall, 2015 for more on
emotional contagion through news images). Impact-images in earlier studies often involved
people. Secondly, this study measured issue importance in a less abstract way. It is possible
that images of floods may make people feel that the issue is important but not important to
them. By expanding on the definition and operationalisation of issue importance, we arrive at the intuitive result that people who find an issue to be important are also likely to be engaged
by it.
Initially, coral farming images did not appear to lead to increased efficacy. Although
collective efficacy was highest among respondents who saw images of coral farming, the
result was not significant, in contrast to previous studies which found solution-oriented
images to produce personal efficacy (Hart and Feldman, 2014; Hart and Feldman, 2016b;
Xue et al., 2016) and/or collective efficacy (Chen, 2015; Wang, 2018). Upon reflection,
Collective efficacy measures the power of the “we”, in other words the individual is a part of
the group. Collective efficacy does not measure the power of an outsider group to fix the
issue. Neither the impact nor the solutions of coral bleaching are felt at a local level,
something O’Neill and Nicholson-Cole (2009) found to be key to issue relevance. It is therefore plausible that even a solution-oriented image of coral farming would not necessarily
have led individuals in remote countries to think ‘we can do something about this’. On the effects of collective efficacy, however, the link between collective efficacy and engagement
is very clear. Collective efficacy leads to increased engagement (in agreement with Chen,
2015 and Wang, 2018).
More complex tests showed that, for respondents with high degrees of general
self-efficacy (GSE), farming images did indeed produce significant collective self-efficacy, which in
turn led to increased engagement. In other words, people who see themselves as being
problem-solvers are more likely to think a major global issue can be solved by the
(international) community, and are significantly more likely to personally do something about
it. This only holds true for respondents with an above-average GSE score. Nevertheless, this
finding adds to the literature by expanding on the link between efficacy on an individual and
collective scale (as theorised by Roos et al., 2014). One way to increase engagement vis a vis
climate change is therefore to nurture the feelings of collective efficacy and target individuals
with high self-efficacy. Solution-oriented images are key, but this will not work equally for
everyone.
This paper also had expectations for the performance of environmental infographics.
Overall, infographics had high average mean values in terms of importance, risk and
engagement, though not enough to be significant. As expected, the use of an infographic
increased self-reported understanding by a small amount, significant (only) in a one-tailed
test. The perception of risk from coral bleaching also increased when viewing an infographic,
even more than for impact-oriented images such as the stock image or contrasting images.
When controlling for divers and snorkellers, a significant relationship emerged from
infographics to increased risk (p<0.05) which mediated increased engagement at a 90%
confidence interval (p<0.1). This finding agrees with health communication, and expands on
Roeser (2012) who argued that the path to risk perception and engagement lay in increased
emotion. The role of emotion and understanding is not fully understood but the dichotomy
possible that increased understanding via infographics leads to more, rather than less,
emotional arousal.
The greatest limitation in this study lies in the use of infographics. For the sake of
comparison between infographics and other image types, the infographic could not contain
more or different information than the article, and had to explain coral bleaching itself, much
in the way that a stock image or contrasting images do on a more visual and intuitive level.
Yet the potential of infographics is much broader, serving as a vehicle for complex
information (Lazard and Atkinson, 2015). An existing use of coral-bleaching infographics in
the news was a global map showing where coral bleaching does and does not take place, as
well as a more detailed map of bleaching sites within the Great Barrier Reef, for example.
Infographics are therefore under-utilised in this experiment, for the sake of comparison with
the other image types. Finally, the lack of a significant difference between stock images and
contrasting images may be generalizable, though it would have been easier to explore with a
greater sample of non-divers per condition.
In terms of further research, the two image types that led to the most engagement
were infographics and solution-oriented images of coral farming. More research is required to
see if infographics can lead to increased engagement in a statistically significant way. The
interaction between infographics and solution-oriented images also demands further research.
Using both in combination may have beneficial results, possibly increasing efficacy,
understanding and risk perception. In addition, no research has thus far been conducted on
solution-oriented infographics.
This study sought to examine the differences between four image types in coral
bleaching stories. The solution-oriented image of coral farming was overall most effective.
For individuals with high self-efficacy, there was a significant indirect effect when seeing
engagement. This finding adds to the existing literature by expanding on the link between
self-efficacy and collective efficacy. Infographics also had beneficial effects on risk
perception and engagement, though more research is required in this area. The significant
relationship between collective efficacy and engagement, and between risk perception and
engagement, was confirmed in this new visual context. It is hoped these findings may inform
the communication strategies of newspapers and civil society.
References
Alexander, P. A., & Jetton, T. L. (1996). The role of importance and interest in the processing
of text. Educational Psychology Review, 8(1), 89-121.
Anthony, K. R., Kline, D. I., Diaz-Pulido, G., Dove, S., & Hoegh-Guldberg, O. (2008).
Ocean acidification causes bleaching and productivity loss in coral reef
builders. Proceedings of the National Academy of Sciences, 105(45), 17442-17446.
Babbie, E. (2005). The basics of social research (3rd). Belmont, CA: Thomson Wadsworth
Baberini, M., Coleman, C., Slovic, P., Vastfjall, D. (2015). Examining the Effects of
Photographic Attributes on Sympathy, Emotions and Donation Behaviour. Visual
Communication Quarterly. Vol 22 (1). pp 118-128
Bandura, A. (2006). Guide for constructing self-efficacy scales. Self-efficacy beliefs of
adolescents, 5(1), 307-337.
Bas, O., & Grabe, M. E. (2015). Emotion-provoking personalization of news: Informing
citizens and closing the knowledge gap?. Communication Research, 42(2), 159-185.
Baumgartner, F. R., & Jones, B. D. (1993). Agendas and instability in American politics (1st
Bernstein, A. G. (2005). Gendered characteristics of political engagement in college
students. Sex roles, 52(5-6), 299-310.
Berry, C., & Brosius, H.-B. (1991). Multiple effects of visual format on TV news learning.
Applied Cognitive Psychology, 5, 519-528. doi:10.1002/acp.2350050607
Brante, E W., Olander, M H., & Nyström, M. (2013). Exploring the impact of contrasting
cases in text and picture processing. Journal of Visual Literacy, 32(2), 15-38.
Brenthel, A. (2016). The Drowning World: The visual culture of climate change. Lund
University (Media-Tryck).
Brulle, R. J., Carmichael, J., & Jenkins, J. C. (2012). Shifting public opinion on climate
change: an empirical assessment of factors influencing concern over climate change
in the US, 2002–2010. Climatic change, 114(2), 169-188.
Casas, A., & Williams, N. W. (2017). Images that matter: Online protests and the mobilizing
role of pictures. Available at
SSRN: https://ssrn.com/abstract=2832805 or http://dx.doi.org/10.2139/ssrn.2832805
Chen, M. F. (2015). Self-efficacy or collective efficacy within the cognitive theory of stress
model: Which more effectively explains people's self-reported proenvironmental
behavior?. Journal of Environmental Psychology, 42, 66-75.
Chen, T., & Dredze, M. (2018). Vaccine Images on Twitter: Analysis of What Images are
Shared. Journal of medical Internet research, 20(4).
Childers, T. L., & Houston, M. J. (1984). Conditions for a picture-superiority effect on
consumer memory. Journal of Consumer Research, 11, 643-654.
Cody, E. M., Reagan, A. J., Mitchell, L., Dodds, P. S., & Danforth, C. M. (2015). Climate
change sentiment on twitter: an unsolicited public opinion poll. PloS one, 10(8),
Danovaro, R., Bongiorni, L., Corinaldesi, C., Giovannelli, D., Damiani, E., Astolfi, P., ... &
Pusceddu, A. (2008). Sunscreens cause coral bleaching by promoting viral
infections. Environmental health perspectives, 116(4), 441.
David, P., & Kang, J. (1998). Pictures, high-imagery news language and news
recall. Newspaper Research Journal, 19(3), 21-30.
DeBono, R., Vincenti, K., & Calleja, N. (2010). Risk communication: climate change as a
human-health threat, a survey of public perceptions in Malta. The European Journal
of Public Health, 22(1), 144-149.
De Sousa, R. (1979). The rationality of emotions. Dialogue: Canadian Philosophical
Review/Revue canadienne de philosophie, 18(1), 41-63.
Dillard, A. J., Ferrer, R. A., Ubel, P. A., & Fagerlin, A. (2012). Risk perception measures'
associations with behavior intentions, affect, and cognition following colon cancer
screening messages. Health psychology, 31(1), 106.
Druckman, J. N., & Kam, C. D. (2009). Students as experimental participants: A defense of
the ’narrow data base’. SSRN eLibrary.
Feldman, L., Wojcieszak, M., Stroud, N. J., & Bimber, B. (2018). Explaining Media Choice:
The Role of Issue-Specific Engagement in Predicting Interest-Based and Partisan
Selectivity. Journal of Broadcasting & Electronic Media, 62(1), 109-130.
Grabe, M. E., Zhou, S., Lang, A., & Bolls, P. D. (2000). Packaging television news: The
effects of tabloid on information processing and evaluative responses. Journal of
broadcasting & Electronic media, 44(4), 581-598.
Hart, P. S., & Feldman, L. (2014). Threat without efficacy? Climate change on US network
news. Science Communication, 36(3), 325-351.
Hart, P. S., & Feldman, L. (2016). The impact of climate change–related imagery and text on
Hart, P. S., & Feldman, L. (2016). The influence of climate change efficacy messages and
efficacy beliefs on intended political participation. PloS one, 11(8), e0157658.
Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. Guilford Publications.
Homburg, A., & Stolberg, A. (2006). Explaining pro-environmental behavior with a
cognitive theory of stress. Journal of Environmental Psychology, 26(1), 1-14.
Hughes, T. P., Kerry, J. T., Álvarez-Noriega, M., Álvarez-Romero, J. G., Anderson, K. D.,
Baird, A. H., ... & Bridge, T. C. (2017). Global warming and recurrent mass bleaching
of corals. Nature, 543(7645), 373.
Intergovernmental Panel on Climate Change. (2007). Climate change 2007: Synthesis report.
Summary for policymakers. Cambridge, UK: Cambridge University Press. Jenkin, M., & Harris, L. (Eds.). (2001). Vision and attention (2001 ed.). New York, NY:
Springer.
Jones, B. D., & Baumgartner, F. R. (2005). The politics of attention: How government
prioritizes problems. Chicago, IL: University Of Chicago Press
Kim, Y. M. (2008). Where is my issue? The influence of news coverage and personal issue
importance on subsequent information selection on the web. Journal of Broadcasting
& Electronic Media, 52(4), 600-621.
Kennedy, L. (2012). Framing Compassion. History of Photography. 36(3), pp 306-314, DOI:
10.1080/03087298.2012.673312
Koirikivi, I. (2014). Measurement of affective empathy with Pictorial Empathy Test
(PET) (Doctoral dissertation, Helsingin yliopisto).
Koletsou, A., & Mancy, R. (2011). Which efficacy constructs for large-scale social dilemma
problems? Individual and collective forms of efficacy and outcome expectancies in
Lassen, I., Horsbøl, A., Bonnen, K., & Pedersen, A. G. J. (2011). Climate change discourses
and citizen participation: A case study of the discursive construction of citizenship in
two public events. Environmental Communication: A Journal of Nature and
Culture, 5(4), 411-427.
Lazard, A., & Atkinson, L. (2015). Putting environmental infographics center stage: The role
of visuals at the elaboration likelihood model’s critical point of persuasion. Science Communication, 37(1), 6-33.
León, B., & Erviti, M. C. (2015). Science in pictures: Visual representation of climate change
in Spain’s television news. Public Understanding of Science, 24(2), 183-199.
Li, Jo-Yun, Sei-Hill Kim, and Jane O’Boyle. "“I Believe What I See”: College Students’ Use
of Media, Issue Engagement, and Perceived Responsibility Regarding Campus Sexual
Assault." Journal of health communication 22.9 (2017): 772-782.
Lorenzoni, I., & Pidgeon, N. F. (2006). Public views on climate change: European and USA
perspectives. Climatic change, 77(1-2), 73-95.
Lorenzoni, I., Nicholson-Cole, S., & Whitmarsh, L. (2007). Barriers perceived to engaging
with climate change among the UK public and their policy implications. Global
environmental change, 17(3-4), 445-459.
Lott, P. (1990). Reader comprehension of infographics. Newspaper Research Journal, 11(2),
100-101.
Mah, D. N. Y., Hills, P., & Tao, J. (2014). Risk perception, trust and public engagement in
nuclear decision-making in Hong Kong. Energy Policy, 73, 368-390.
Metag, J., Schäfer, M. S., Füchslin, T., Barsuhn, T., & Kleinen-von Königslöw, K. (2016).
Perceptions of climate change imagery: Evoked salience and self-efficacy in
Mullinix, K. J., Leeper, T. J., Druckman, J. N., & Freese, J. (2015). The generalizability of
survey experiments. Journal of Experimental Political Science, 2(2), 109-138.
Nadeau, R., Niemi, R. G., & Amato, T. (1995). Emotions, issue importance, and political
learning. American Journal of Political Science, 558-574.
O’Neill, S. J. (2013). Image matters: Climate change imagery in US, UK and Australian
newspapers. Geoforum, 49, 10-19. doi:10.1016/j.geoforum.2013.04.030
O’Neill, S. J., Boykoff, M., Niemeyer, S., & Day, S. A. (2013). On the use of imagery for
climate change engagement. Global Environmental Change, 23, 413-421.
doi:10.1016/j.gloenvcha.2012.11.006
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, 355-379. doi:10.1177/1075547008329201 O’Neill, S. J., & Smith, N. (2014). Climate change and visual imagery. Wiley
Interdisciplinary Reviews: Climate Change, 5, 73-87. doi:10.1002/wcc.249
Oreskes N. (2018) The Scientific Consensus on Climate Change: How Do We Know We’re
Not Wrong?. In: A. Lloyd E., Winsberg E. (eds) Climate Modelling. Palgrave
Macmillan, Cham
Osman, E. O., Smith, D. J., Ziegler, M., Kürten, B., Conrad, C., El‐Haddad, K. M., ... & Suggett, D. J. (2018). Thermal refugia against coral bleaching throughout the northern
Red Sea. Global change biology, 24(2), e474-e484.
Pasternack, S., & Utt, S. H. (1990). Reader use & understanding of newspaper
infographics. Newspaper Research Journal, 11(2), 28-41.
Poortinga, W., & Pidgeon, N. (2003). Public perceptions of risk, science and
government. Main findings of British survey of five risk cases. University of East
Reinard, J. C. (2006). Communication research statistics. Thousand Oaks, CA: Sage
Publications.
Rebich-Hespanha, S., & Rice, R. E. (2016). Dominant Visual Frames in Climate Change
News Stories: Implications for Formative Evaluation in Climate Change
Campaigns. International Journal of Communication (19328036), 10.
Roos, S. M., Potgieter, J. C., & Temane, M. Q. (2013). Self-efficacy, collective efficacy and
the psychological well-being of groups in transition. Journal of Psychology in
Africa, 23(4), 561-567.
Rundmo, T. (2002). Associations between affect and risk perception. Journal of Risk
Research, 5(2), 119-135.
Schmittner, A., Oschlies, A., Matthews, H. D., & Galbraith, E. D. (2008). Future changes in
climate, ocean circulation, ecosystems, and biogeochemical cycling simulated for a
business‐as‐usual CO2 emission scenario until year 4000 AD. Global biogeochemical cycles, 22(1).
Schroth, O., Angel, J., Sheppard, S., & Dulic, A. (2014). Visual climate change
communication: From iconography to locally framed 3D visualization. Environmental
Communication, 8(4), 413-432.
Scott, H., Fawkner, S., Oliver, C., & Murray, A. (2016). Why healthcare professionals should
know a little about infographics. British Journal of Sports Medicine, 50(18),
1104-1105.
Scherbaum, C. A., Cohen-Charash, Y., & Kern, M. J. (2006). Measuring general
self-efficacy: A comparison of three measures using item response theory. Educational
and Psychological Measurement, 66(6), 1047-1063.
Shin, H. (2016). Epidemic and risk communication: An analysis of strategic and graphic