1
Master’s Thesis
Fortunate attraction or fatal distraction?
The influence of visual aesthetics in deforestation news
reporting on political attitudes, behavior and memory
Erasmus Mundus:
Journalism, Media and Globalisation
Graduate School of Communication
Master’s Programme Communication Science
Student: Lena Lynge Rosing (11300396)
Supervisor: dhr. T.E. Powell MSc
2
Abstract
News visuals highly influence how the political messages of news reports
are perceived by readers. Much is known about the effects of certain visuals on viewers’ attitudes and behavior both when visuals stand alone as well as in their
combination with textual content. While both the framing and composition of visuals
in news media have been studied extensively, less is known about how aesthetics
influence the relation between visual frames and framing effects. The author used an
experimental study design to expose readers to deforestation news with different
levels of visual aesthetics. The study found that ‘hyper’-aesthetics visuals distract
readers attention from contextual content and that this affects readers attitudes and
behavioral intentions towards to item of deforestation. These effects are moderated
3
Introduction
Today, visual elements in online newspaper articles have become more of the rule
than the exception. Photographs, cartoons, graphs and other visuals are contributing to the
general communication of complex political messages across all types of media. Therefore,
visuals have become increasingly important in the study of political communication and
journalism. Even though the use of visuals in news media communication is not a new
phenomenon, significant technological and cultural developments in news production and
consumption presents new dimensions of interest to the study of media visuals. As with
textual content, visuals contain information, which significantly contribute to the
understanding of the news item in question (Zillmann, Gibson, & Sargent, 1999; Iyer,
Webster, Hornsey, & Vanman, 2014).
In communication science research, the significance of visuals as venues for
communication has become increasingly important in the understanding of the 21st century
news media landscape. Even in traditional print media, visuals have taken a core position in
the reporting of social issues (Geise & Baden, 2014). Visuals have truly taken center stage in
the communication of complex political issues, and with news media heavily influencing new
consumption habits, the importance of visuals in news communication cannot be
underestimated.
In academic literature, the effects as well as the selection of these visuals has been
studied extensively (Messaris & Abraham, 2001; Rössler, et al., 2011). Also the importance
of visuals in image-texts relations has received remarkable attention (Powel, Boomgaarden,
De Swert, & de Vreese, 2015). In a news media context however, research on how aesthetic
dimensions of news visuals affect consumption, issue perception, memory for news and
4 Often, aesthetic value has been the core argument for cultural heritage or landmark
conservation policies (Abramson, 2007), and for nature preservation and protection policies
in particular (De Young, 2000; Bengtson & Xu, 1997). The pure beauty and aesthetic value
of a place makes it worth protecting, and this might shed new light on the use of visual
aesthetics in social science and media studies. Raising awareness and affecting public
opinion is a crucial tool for social issues to be considered important in the public eye and
discourse. For environmental news, receiving this awareness, and thereby being considered a
social problem, is crucial in order for policymaker to take action. A social problem exists
primarily in terms of how it is defined and conceived in society. A social problem is a
putative condition or situation that (at least some) actors label a "problem" in the arenas of
public discourse, defining or framing it as harmful (Hilgartner & Bosk, 1988). As such,
reaching the status of a social problem is crucial for issues portrayed in the media, as it can
ultimately provoke crucial social resonance and following political change. Furthermore, Blumler & Kavanagh (1999) identified ‘aestheticization’ as one of the main characteristics of
what they call the third age of political communication, suggesting that political
communication process are increasingly influenced by the aesthetic demands of the 21st
century. There is no denying that visual aesthetics has taken center stage in media production
and consumption during recent years, and although notorious for its attributes to neutrality
and objectivity, news media is no exception.
In relation to environmental news reporting in particular, visual aesthetics are
especially interesting in relation to communicating the need for nature or climate preservation
and protection due to the link between human emotions and aesthetic responses to the natural
5 aesthetics in news media reports affects readers issue perception, political behavior and
memory.
Considering the above, this study aims to address the following research questions:
What role do visual aesthetics play in the production of framing effects when presented in news reports? How do visual aesthetics affect image recognition and memory for text? How do previous exposure to environmental news affect readers’ attitudes and behavior when exposed to either aesthetic or non-aesthetic visuals?
Theory Measuring visual aesthetics
As with many social problems (Hilgartner & Bosk, 1988), perhaps the greatest
challenge for environmental issues today is to make people care (Rebich-Hespanha & Rice,
2016). As this kind of news is inherently difficult to report on due to its often invisible
implications and far away nature, it is interesting to investigate how to make people care.
Here, aesthetic visuals might play a role due to its attention-grabbing nature, and its
emotion-provoking capabilities (Tello, 2004) as well as in memory and recall processes (Gilbert &
Schleuder, 1990; Kätsyria, Ravajaa, & Salminena, 2012). Aesthetics have dominantly been
associated with the world of art. In social research however, as well as in marketing research
(Huhmann, 2003), the effects of aesthetics on human beings have been studied extensively.
In the world of branding and advertising, a study found aesthetics to be the most
important determinant for overall attractiveness of an image, whereas content, brand and
usability were more important for overall preference (Sutcliffe, & Namoune, 2008). In
academia, the visual framing of news have mainly focused on visualization of conflict and
disaster and how these frames have affected public opinion and ultimately the outcome of a
particular conflict (Fahmy & Neumann, 2011). Most of these studies have been devoted to
6 in question, or how differences in the compositional framing such as angle, social distance
and point of view, can affect viewers’ perception of the issue (Fahmy, 2004; Kress & van
Leeuwen, 1996). In social science research however, far less attention has been given to the
study of other aspects of compositional framing such as aesthetic expression. But how does one scientifically measure something, which is often intuitive and subjective? Aydın and
Smolic (2014) found five key dimensions when assessing the aesthetic attributes of images,
namely: sharpness, colorfulness, depth, tone and clarity. In addition, through a small-scale experiment Sun, Yamasaki, and Aizawa (2013), found that human beings’ judgement on
complexity levels are congruous, hence complexity levels of photos are measurable. In their
study of photo aesthetics, Kong et al. (2016) found eleven attributes, which serve as crucial
measurements related to image aesthetic judgements: interesting content, object emphasis,
good lighting, color harmony, vivid color, shallow depth of field, motion blur, rule of thirds,
balancing element, repetition, and symmetry. The exact effect of aesthetics is said to be
somewhat individual but scholars also agree that there is a certain amount of consistency in
aesthetic assessment and therefore aesthetics are also to a certain extend objectively
measurable (Haas, et al., 2015; Wang, Datta, & Li, 2009).
These image quality assessment methods seek to estimate “image quality” without
requiring user involvement. However, even though the field of computerized aesthetic image
evaluation is studied extensively these years, there are still no definite algorithm or equation
able to determine with a 100% accuracy which images are considered aesthetic by human
beings.
News visuals and effects on attitude and behavior
Solaroli (2015) discussed technological, professional, and cultural shifts that have
been affecting news photography over the last decade in this context of digital
7 increasingly difficult differentiation between photojournalism and photography art.
Furthermore, Abrams (1995) pointed to the disconnect between objective journalistic
reporting of reality and the increasing possibilities in photographic post-production
(Schwartz, 1990). Although retouching is indeed a ‘touchy’ subject in journalism practices,
there is no denying that the technological development of audio-visuals proposes new
possibilities to the communication of social problems.
In photojournalism, there are two questions, which determine the potential framing
effect of a particular image, namely what is portrayed in the image and how it is portrayed.
These questions can be categorized as the frame and the composition of the image in
question. Aesthetic value however, is perhaps a combination of the two, but attributes from
the composition aspect such as color intensity, relative size, object distribution and texture,
are the strongest indicators of aesthetic value (Kong et al., 2016).
In the study of news media, the effects of visuals on attitude and behavior depends on
their combination with several other factors. When visuals are presented in combination with
text, as it is the case with most news media content, both image and text contributes to the
readers understanding of the issue portrayed. However, there are crucial differences in how
images and text are ‘read’ by the reader (Kress & van Leeuwen, 1996). The power of visuals
in political communication is connected to both sociological and psychological factors.
Studies suggest that images are the first entry point into newspapers for the readers because they attract people’s attention far more than textual content (Garcia & Stark, 1991).
Furthermore, readers attributes meaning to an image in a matter of seconds (Todorov et al.,
2005), which makes images a convenient tool for newspapers and other media channels.
In news production, the way an issue is defined and presented affects the
understanding and assessment of the issue and the actors involved. Because no news story
8 certain way, presenting some aspects of the issue over others (Entman, R. M., 1993).
Therefore, the framing of an issue is extremely important as it has been shown to
significantly affect public opinion, general understanding and behavior (Schuck & de Vreese,
2006) and can therefore ultimately affect how an issue is dealt with politically. Also the
agenda-setting effects of images have been studied (Fahmy, Cho, Wanta, & Song, 2006), as
well as the ability of images in provoking emotions in human beings (Zillmann, Gibson &
Sargent, 1999). Several scholars also argue for the enhancement of newsworthiness of an event if it implies a ‘picture-opportunity’ (Harcup & O'Neill, 2001) or if it involves
audio-visuals (Harcup & O'Neill, 2016). Every issue presented in the news reflects a specific
interpretation of reality from a plurality of possibilities, ultimately affecting whether the issue is ‘selected’ as a social problem worthy of public attention or not (Hilgartner & Bosk, 1988).
Both text and visual carry meaning separate from each other but do also carry
meaning in their combination (Geise, & Baden, 2014; Powell, Boomgaarden, De Swert & de
Vreese, 2015). The framing of issues in both text and visual content, and its effect on
agenda-setting, citizen mobilization and opinion formation has been studied extensively (e.g.,
Coleman, 2010; Fahmy, Wanta & Song, 2006; Schuck, & de Vreese, 2006; Schuck,
Vliegenthart, & de Vreese, 2016).
To the knowledge of the author, there exists no scientific evidence that visual
aesthetics alone should significantly influence political attitude or behavior. However, there
are plenty of research suggesting that visual aesthetics, especially when combined with
environmental preservation media frames, does provoke emotional reactions in human beings
(Crang, 2009; Christianson, 1992; Crowther, 1989; Coleman, 2010). In addition, emotional
reactions to news content, whether is being caused by textual or visual frames, most
definitely affects both behavior and political attitude (Fahmy, Cho, Wanta, & Song, 2006;
9 According to Zillmann, Knobloch, and Yu (2001), visuals do not only carry more explicit meaning, but also ‘capture’ readers attention and thereby affects their selective
reading of online news reports. Visualization in news reports is also highlighted as one of the
10 factors of newsworthiness proposed by Harcup and O'Neill (2016). They argue that the
newsworthiness of an item increases with the presence of audio-visual material. This is
particularly interesting in the study of environmental news reporting.
Environmental news visuals and environmental preservation policies
Especially in the communication of far-away conflict or unfamiliar topics, news
visuals offer an additional dimension to the portrayal of foreign issue, significantly increasing
the information given to the receiver of the news item.
It is important to recognize the dual and complex nature of visual aesthetics and
emotional responses. Amongst other abstract concepts, death and preservation are crucial
concepts to consider when trying to understanding the effect of visual aesthetics on human
emotions and behavior. In the study of photography, politics and waste in the global
imaginary, Crang (2009) found that counter-images to global capitalism were produced by
making waste ships work aesthetically in photo-documentary pieces. Interestingly, critics of
global trade have latched upon these counter-images of mighty ships carcasses being broken
in South Asia. Although not portraying living things, this narrative or frame, latched itself
upon the story of global trade, significantly and negatively changing the attitudes towards the
industry.
When relating this to nature preservation, and specifically the preservation of the
Amazon, one can expect equally strong emotional reactions to the damages done to this old,
silent and life-giving rainforest. According to Bengtson & Xu (1997), aesthetic, moral and
spiritual values of national forests have become increasingly important to forestry
10 years. They point to the paradox that the social values on which humans base environmental
protection attitudes are those which are the most difficult to measure (Bengtson & Xu, 1997).
As suggested by Rebich-Hespanha & Rice (2016), effectively communicating climate
change issues of often complex, long-term, and distant phenomena, is particularly
challenging because of the invisibility of many key aspects, as well as the geographical
disconnect between causes and effects. Although several studies have been dedicated to the
study of environmental news visuals (Hansen & Machin, 2013; Hart, & Feldman, 2016; O’Neill, 2013), the effects of aesthetic visuals within this particular journalistic beat remain
rather unexplored.
In a recently published report on climate change visuals in the media, there are several
key findings suggesting which images and frames to use, in order for audiences to respond
with action and attention. The study found that certain frames are more powerful than others
in provoking action responses from audiences, such as for an example showing ‘real people’ instead of ‘staged photo-ops’(Corner, Webster & Teriete, 2015). However, in contrast to
visual framing effects in conflict and war reporting, where it has been firmly established that
portraying victims will increase support for intervention (Fahmy & Neumann, 2011; Powell,
2015; Greenwood & Jenkins, 2015), the exact effect of certain environmental news visuals
on attitude and behavior remains more ambiguous.
In the context of climate change visuals in the US news media, Rebich-Hespanha &
Rice (2016) found several dominant visual frames. Two of these frames have been selected
for the purpose of this study namely the food and agriculture frame and the wilderness and
nature recreation frame. The rationale behind this choice is found in the imagery possibilities
of encountering both aesthetics and less aesthetic imagery within these specific frames. This
will be further discussed in the methods section. In relation to aesthetics, this study suggests a
Rebich-11 Hespanha & Rice (2016), food supply system imagery might have two conflicting effects on
readers. On the one hand, some audiences might experience emotional arousal such as a
desire for stronger connectedness to natural systems while at the same time striking other
audiences as romanticized and unrealistic. One the other hand, images displaying factory
farms or newly cleared agriculture terrain, might distract audiences from considering other
aspects of the commercial food systems, a very unfortunate effect limiting audiences
understanding of the issue (Rebich-Hespanha, S. & Rice, R. E., 2016). Unfortunately, the
same can be said for the wilderness and nature recreation frame. As the frame emphasizes the
value of the mere existence of the places portrayed, ambiguously, the beauty and aesthetic
value of this type of imagery might distract viewers, not only from the actual vulnerability of
these places, but also from the text accompanying the image which often carries important
contextual meaning. Hence, aesthetic images might distract readers attention and make it difficult to ‘see’ the severity of the situation and why measures should be taken against it, on
the other hand, aesthetic beauty of these places can contribute to the aesthetic reasoning
behind preservation attitudes. Suggesting that nature photography, which is prominent within
the wilderness and nature recreation frame, can simply become ‘too beautiful’ (Chianese,
2014) proposes interesting opportunities for the study of visual aesthetics in environmental
media consumption.
Based on the relation between the power of visual aesthetics in provoking human
emotion and following attitude and behavior in relation to the topic of environmental news
frames, it is hypnotized that:
“H1: Compared to non-aesthetic visuals, aesthetic visuals will positively affect readers
support for intervention and make them more likely to engage in political behavior against deforestation.”
12 According to Geise & Baden (2014), previous knowledge about the issue in question
affects how a frame is perceived and therefore to which extent framing effects occur or not.
Furthermore, studies suggest that social identities predict pro-environmental behavior, but the
strength may depend on whether the behavior is visible to others (Brick, Kim & Sherman,
2017). According to Johnson, R. D. (1987), both previous knowledge of the subject and the
frame presented affects information perception. Hence, these attributes affects understanding
and attitude towards the subject in question.Nelson, Oxley, and Clawson (1997) argue that
framing effects are not reducible to the new information that the framed message provides
but instead, frames operate by activating information already at the recipients' disposal,
stored in long-term memory (p. 225).
Based on these assumptions, a second hypothesis is proposed:
“H2: Previous knowledge of and exposure to environmental news will moderate the effect of visual aesthetics on measures of attitude and behavioral intentions.”
Memory and visual aesthetics
In neuroscience studies the relation between attention, emotion and visuals may
provide insight into why visuals trigger emotions in human beings. Talmi, D. (2008), found
that when participants were told to be attentive, not only was the picture remembered better,
but also emotion related areas of the brain were activated. This suggests an interesting
relation between attention-grabbing picture, emotions and people’s ability to retrieve or
remember the picture in question. Thus, when people were asked to be attentive, the parts of
the brain controlling emotion is activated, further suggesting that emotions and attention are
closely interlinked. As attention-grabbing attributes to images include specific attributes such
as color intensity, relative size, object distribution and texture, there is reason to suggest that
13 The power of visuals is also attributed to the neurological effect of visuals on human
beings. As painter, Joan Miró has been quoted: ‘You can look at a picture for a week and
never think of it again. Or you can look at a picture for a second and think about it all your life’. The quote by Miró perfectly describes the ambiguous effect of imagery on the human
mind and memory. Remembering, or forgetting for that matter, is a curious thing. Attention
to detail can increase the possibility of remembering and as such, visual attentiveness can
affect memory (Talmi, 2008). According to Broadbent and Broadbent (1981), attention
allocation is a prerequisite of encoding, storing and retrieval from memory.
In the world of journalism and political communication, attraction is key when
computing messages to fast-moving audiences across non-personal channels such as online
news websites. Here, readers cannot be asked for their attention explicitly, but they can however be ‘caught’ by factors pleasing to their unconsciousness, thereby maintaining their
interest and curiosity. Hence, visual aesthetics theory, in combination with visual framing
theory, might propose new insight into the study of news visuals and their effects on political
attitude, behavioral intentions and memory.
Interestingly, imagery can also distort memory, reduce understanding, and pose
barriers to communication, especially when words and imagery contain contradictory
messages (Harp & Mayer, 1997; Zillmann, Gibson, & Sargent, 1999). Several scholars
address their concern for a hyper-visual news culture as image alone stories can generate
stereotyped attitudes with a lack of a more complex understanding (Mendelson &
Darling-Wolf, 2009). Furthermore, Messaris and Abraham (2001) found that images reinforce
cultural stereotypes, even though they are not referred to in textual context, pointing to a
central role for visual images in framing research. In image-text relations, as often presented
14 presented alone, images generate stronger framing effects on opinions and behavioral
intentions than text.
Based on these assumptions and in particular the dual nature of the effects of visual
aesthetics on human responses, the author proposes the following hypothesis:
“H3: Compared to non-aesthetic visuals, aesthetic visuals will increase visual memory and
recognition but decrease memory for text”.
Method Design
As frames have been shown to heavily influence how readers perceive and understand
issues in the media, the author decided to limit and control the frames represented in each of
the experimental conditions. Therefore, images were selected within two main categories
derived from previous studies of frames in environmental news media (Rebich-Hespanha &
Rice,2016). As seen in their study of climate change visuals, the Food and Agriculture frame
as well as the Wilderness and Nature Recreation frame, are both some of the most dominant
visual frames presented in news media when reporting on environmental and climate change
issues. Furthermore, examining two frames improves generalizability of the study results by
going beyond one specific type of image.
In order to test the effects of visual aesthetics in environmental news articles, an
online-embedded experiment was conducted (Bock, Isermann, & Knieper, 2011; Babbie,
2005). Respondents of the experiment were assigned to one of four conditions in a two by
two factor design. After being assigned to one of the four conditions: (frame: wilderness and
nature recreation, food and agriculture) and (expression: aesthetic, non-aesthetic),
respondents were asked about several questions concerning: attitude, emotional response,
15 compare the effects of text-alone. The textual content was held constant throughout all four
conditions.
Research participants
A total of 223 adults were recruited for the study using traditional snowball sampling
via e-mail and social media. A total of 63 participants were excluded from the final sample as
they dropped out before concluding two-thirds of the experiment leaving exactly 160
participants for the final sample. Although participants ranged significantly on their
demographic characteristics (38 countries of origin), the distributions of age (M = 25-34, SD
= 0.83) and gender (119 females) resulted in a predominant participant profile which was
female, between 25-34 years of age, frequently exposed to environmental news and Danish
of nationality. In general, participants reported an unusually high level of previous exposure
to environmental news, suggesting that most participants had a high level of personal interest for environmental news and therefore could be considered ‘very knowledgeable’ about this
exact field of news reporting.
As explained in the theory section, the experimental condition and visual perception
should not differ significantly between social groups (eg. gender, race, age), and therefore the
sample was highly diverse in these demographic measures. In general, previous exposure to
environmental news was exceptionally high for the final sample. A total of 66.8 % of the
sample reported that they were exposed to news about environmental issues on a very regular
basis (daily or two to three times pr. week).
Stimuli
Stimuli was constructed using media coverage of deforestation issues related to
agriculture and nature recreation. This topic was chosen due to its obvious potential in
displaying extreme aesthetic imagery while at the same time relating to aesthetics and
16 In order to guarantee that human participants would agree with the algorithm-based
automatic aesthetic assessments, a total of 72 participants were recruited for the pre-test
through e-mail and social media. The sample were assured to be completely independent
from the main experiment sample, in order to prevent any previous exposure for participants
participating in the main experiment. For more information concerning the rationale behind
the pre-test see appendix C.
Automatic Image quality assessment methods seek to estimate “image quality”
without requiring user involvement. The images in this study were selected based on the
EyeEm and the Sensiphi technology. Measuring attributes associated with aesthetics, such as
color harmony, vivid color, shallow depth of field, motion blur and rule of thirds, these
algorithm dependent programs are used for this study in order to provide a clear distinction
between aesthetic and non-aesthetic visuals.
Out of twelve pre-test images, eight images was selected for the main experiment with
the following distribution: two aesthetic wilderness and nature recreation images, two
non-aesthetic wilderness and nature recreation images, two non-aesthetic food and agriculture images,
and two non-aesthetic food and agriculture images.
The text stimuli, which was held constant throughout all five conditions, was taken
from several real news media sources (The Guardian, The New York Times, The
Independent, The Economist) and manipulated to fit the experimental condition and design.
Each of the eight images was parred with the same headline, text and image capture. One
condition with text only was constructed in order to check for the effects of images as a
whole.For examples of the final stimuli see appendix A and B.
Procedure
When entering the study, participants were asked to indicate their previous exposure
17 randomly assigned to one of the five experimental conditions. In order to strengthen the
validity of the study, both the aesthetic and non-aesthetic condition was pared with two
different visual ‘frames’, so that an eventual effect could not be said to only be true for one
particular frame within the environmental news paradigm. This resulted in eight different
conditions with a two by two factorial design, and one no image condition. Then, participants
were asked to indicate their attitude towards deforestation, their emotional response and their
probability to engage in political actions, all measured on several separate items.
Demographics such as gender, age and country of origin was reported and finally, memory
for both text and image was measured through multiple choice questions.
Measures
Upon entering the main study, participants were asked to indicate their previous
exposure using two measures (‘how often would you say you are exposed to images of environmental issues?’, ‘how often would you say you are exposed to news about
environmental issues?’), both on a 7-point interval scale (α = .83). After exposure to the
randomized experimental condition, participants were asked about their perception of the
news article on several measurement elements.
Five statements measured participants attitude towards the issue (1 = strongly agree, 7 = strongly disagree, α = .31). As the scale for attitude did not reach a sufficient level of
reliability according to the Cronbach’s alpha coefficient, an exploratory factor analysis was
conducted. One factor was extracted from the analysis measuring Support for Intervention
(against further deforestation). A strong correlation between three of five items and a
semi-reliable scale was extracted combining three of the initial five items (α = .67).
One item measured appraisals of political urgency (eg. ‘Deforestation of the Amazon is a problem in need of urgent political action’). One item measured support for deforestation
18 assessed appraised severity (eg. ‘Nature recreation and preservation is crucial for the well-being of future generations’) while two items measured the direction of appraised importance
(eg., ‘The Amazon must be protected from deforestation - because it is one of nature’s great
wonders’, ‘The Amazon must be protected from deforestation – because it is important for
the ecosystem’). The support for deforestation items was later reverse coded to fit with the
other measures for attitude. Combining urgency, gravity and one of the measures of direction
of appraised importance, a satisfactory scale was constructed (1= strongly agree, 7= strongly disagree, α = .67).
Afterwards, participants were asked to indicate how strongly they felt various
emotions after being exposed to the experimental condition (1 = not at all, 7 = extremely).
One item measured anger, one item measured sadness and one item measured optimism.
Finally, two items measured perceived sentimentality (eg., nostalgic, sentimental). As the
emotions items did not form a reliable scale, a factor analysis was conducted in order to
check for significant clusters. A semi-reliable scale was found in combining the sadness item with the sentimentality items (1= not at all, 7= extremely, α = .69).
Four items assessing participants’ self-reported probability to engage in political
activities within the following 12 months were constructed (eg measured behavioral intentions. ‘Sign a petition in support of taking action to reduce deforestation’). This was
measured on a scale of 1-7 (1 = very likely, 7 = very unlikely, α = .72).
Finally, one item assessed participants’ perceived aesthetic evaluation of the image
they had seen in the article (1= Not beautiful, 7= Very beautiful). Memory for text alone was
measured using factorial questions about the textual content of the article (eg. ‘Based on the
article you just read, what did the study made by the Potsdam Institute Examine?’). This
question followed by a multiple choice opinion displaying three options where one was
19 would happen if deforestation intensified; What would happen if climate changes
intensified). Memory for image alone was measured using visual recognition measures as a
multiple choice item displaying all eight images from the final sample as well as an
additional four images not displayed in any condition but each presenting one of the four
conditions.
Manipulation checks
In order to address the concern for ambiguity of individual aesthetic assessment of
participants as mentioned in the theory section, one question assessed to which extend
participants perceived the image stimulus as intended. At the end of the survey participants
were asked to rate the image they had seen together with the article on a 5-point scale
identical to the one that pre-test participants were displayed, in order to check whether the
intended image manipulation was indeed perceived by participants (1= not beautiful, 5= very
beautiful). Here, there was a significant difference in perceived aesthetics/aesthetic score
between the aesthetic and non-aesthetic condition (p < .001).
Analysis
A series of ANOVAs were conducted to investigate the mean differences between
measures of support for intervention and behavioral intentions by aesthetic condition, as well
as the mean differences between memory for text and memory for images. The interaction
between frame conditions and measures of attitude and behavior by previous exposure was
explored using Hayes PROCESS-macro in SPSS. According to Hayes (2013), a moderation
occurs when the relationship between two variables differs in magnitude and direction based
on the level on another variable. For both moderation models, the frame conditions aesthetic,
non-aesthetic and no image were included as the independent variables. The level of previous
exposure to image and the level of exposure to text were each examined as moderators, in
20
Results
No direct significant effects of visual aesthetics on neither behavioral intentions nor
support for intervention were found in the experiment. As displayed in table 1, when exposed
to the aesthetic condition, participants reported slightly higher intentions to engage in
political behavior when exposed to an aesthetic image, regardless of whether the frame
displayed a food and agriculture frame or a wilderness and nature recreation frame. A non-significant difference in behavioral intentions between the two chosen ‘frame’ conditions
(agriculture and wilderness recreation) served the purpose of the experiment, F(4,155) = .26,
p = .89.
Table 1. Mean Behavioral Intentions and Support for Intervention pr. Visual Frame Condition.
The same was true for the support for intervention item between frame conditions
both in the aesthetic and non-aesthetic condition, F(4,155) = 1.19, p = .31. The relatively low
Aesthetic Non-Aesthetic No Image
Wilderness and Nature Recreation Food and Agriculture Wilderness and Nature Recreation Food and Agriculture Behavioral Intentions 3.50 (1.13) 3.50 (1.33) 3.54 (1.04) 3.76 (1.41) 3.59 (1.54)
Support for Intervention 1.36
(0.48) 1.44 (0.49) 1.50 (0.58) 1.30 (0.42) 1.25 (0.37) N 35 33 36 40 16
Note. Standard deviations in parenthesis Political behavioral intentions scale: 1 = very likely to engage in political behavior, 7 = very unlikely to participate in political behavior. Support for Intervention scale: 1 = very concerned, 7 = not at all concerned.
21 maximum reported support for intervention (3.33, 1 = high, 7 = low) suggests that something
within the text stimuli probably implied factors leaning towards an intervention frame. This
will be further discussed in the discussion section.
Due to not reaching statistical significance on any of the differences in the dependent
variables, these results reject the first hypothesis that: compared to non-aesthetic visuals,
aesthetic visuals will positively affect readers support for intervention and make them more likely to engage in political behavior against deforestation.
Implications on memory for text and image
As for memory, a series of ANOVA’s were conducted to check for both image
recognition and text recall. Memory for image was measured by a 1= correct vs. 0= incorrect
scale. Memory for text was measured on a 4 point scale (0= No correct answers to 3= All
correct answers).
Table 2. Means of memory for text and memory for image.
Memory for text was slightly higher in the non-aesthetic condition, which suggests
that aesthetics actually does distract from the contextual textual content of a new articles,
ultimately partly supporting the third hypothesis. Contrary to the reviewed literature
Aesthetic Non-Aesthetic No Image
Memory Image 0.88 (0.32) 0.92 (0.27) _ _ Memory Text 1.85 (0.77) 1.97 (0.78) 2.06 (0.77)
22 concerning memory and attentiveness to imagery, the study found that images with a high
aesthetic value were remembered less than images with a low aesthetic value. Although not
reaching statistical significance, the effect of aesthetic images on memory for text resulted in
confirming the second hypothesis that aesthetic images would distract participant’s attention
for textual content. These results partly confirm the third hypothesis that; compared to
non-aesthetic visuals, non-aesthetic visuals will decrease memory for text, but strangely rejects the
hypothesis that; compared to non-aesthetic visuals, aesthetic visuals will increase visual
memory. This might be due to more complex neurological process when exposed to
‘hyper’-aesthetics, where aesthetic attention might even distract from sufficient object processing
within the image frame This could be an interesting point of further study.
Moderation effect of previous exposure
The moderating effect of previous exposure to environmental news and images on
measures political attitude and behavior by aesthetic condition showed dual results. Both the
exposure to news and the exposure to images item were dummy coded into groups of high
exposure (Daily, two to three times pr. week), to low exposure (One a week, two to three
times pr. month, less often).
23 A significant overall moderation model was found for measures of support for
intervention F(3, 151) = 3.59, p = .01, 𝑟2 = 0.6. , but not for measures of behavioral
intentions. The interaction effect between variables was also significant b = .41, t(151) =
-3.1, p = .001.
Slopes for aesthetic condition predicting support for intervention at each level of
previous exposure to news showed, that the moderating effect was only significant for the
low exposure to news participants (p = .002), with an effect of - .34. Hence, the aesthetic
condition negatively affected low exposure to news participants in their support for
intervention. For high exposure participants, the condition had no significant effect on their
support for intervention (p = .28). As seen in figure 1, the exact same trend was observed in
the moderating effect of exposure to images on support for intervention by aesthetic
condition F(3, 155) = 3.88, p = .01, 𝑟2 = .06. Hence, probing the interaction showed that the
moderating effect of previous exposure took opposite directions on the effect on support for
intervention.
Although not statistically significant, participant frequently exposed to environmental
news reported higher levels of support for intervention against deforestation when exposed to
an aesthetic image (M = 1.29) than when in the non-aesthetic condition (M = 1.47). On the
contrary, for participants less frequently exposed to this type of news, the aesthetic condition
resulted in lower levels of support for intervention (M = 1.58), compared to the non-aesthetic
condition (M = 1.19).
Discussion
This study investigated the use of visual aesthetics in environmental news reports and
their implications for recipients’ political attitudes and behavioral responses, as well as on
24 aesthetics and readers support for intervention in, and general attitude towards, the issue of
sustainability, nature recreation and preservation.
In a political communication perspective, the findings imply several interesting points
of interest related to policymaking and public opinions theory. Influencing public opinion
through media messages is not only a strategic goal of media professionals and politicians
alike (Stromback & Nord, 2009), but also often an unconscious process where the text and
images are put together without any further reflection as to their combined (Rössler,
Bomhoff, Haschke, Kersten & Müller, 2011) nor their effects on attitude formation and
political behavior. As news frames have been found to significantly affect opinion and
behavior (Azrout, van Spanje & de Vreese, 2012), and as visual have been found to influence
this relationship (Geise & Baden, 2014) and even affect opinion while standing alone
(Powell, Boomgaarden, De Swert, & de Vreese, 2015), the role of visual aesthetics might
have a greater impact on public opinion and policy support than previous anticipated.
. As oppose to visual frames, defined as the existence or absence of certain objects
within a photographic image (Iyer, Webster, Hornsey & Vanman, 2014), aesthetic value
might be considered a ‘hidden influencer’ on unconscious human image decoding and visual
information processing (Kress & van Leeuwen, 1996). This, in combination with
environmental and preservation news in particular, presents new venues of research in the
field of public opinion and political communication.
Taken together, these results lead to suggest that hyper-aesthetic visuals in
environmental news reporting can do more damage than good to the appraised urgency and
support for intervention against deforestation from readers. However, this is only true for
people less frequently exposed to environmental news reporting. Despite results of this study suggesting the opposite, majority of research suggest that aesthetic or ‘attention-grabbing’
25 Parikh, Torralba & Oliva, 2014). Combined with the findings that aesthetic visuals decreased
memory for text, these results further confirm the distracting power of hyper-aesthetic
images, at least in deforestation reporting.
With this study, the author has identified the theoretical implications for why visual
aesthetics in the context of environmental news, could have an effect on readers issue
perception and political behavior, revealing relevant observations for further studies of the
effect of aesthetics in environmental news media production and consumption.
In response to the title question of whether visual aesthetics presents a fortunate
attraction or fatal distraction to the effective communication of environmental news
messages, the study concludes that the answer depends on reader’s previous exposure to
news and news visuals. This aligns with previous research in the field which suggests that
participants with a high level of previous knowledge about a certain topic are less affected by
news frames in their opinion formation towards social problems (Fahmy, Cho, Wanta &
Song, 2006). In the very specific case of deforestation reporting, trends suggests that visual
aesthetics do have a slight effect on readers issue perception, support for intervention against
deforestation and their likelihood of participating in political behavior, independently of who,
what and how objects are portrayed in the image. For participants with low exposure to
environmental news, ‘hyper’-aesthetic visuals do indeed pose a fatal distraction, making this
group less likely to take action against deforestation than when exposed to an aesthetic
image. Considering the suggestion made by Rebich-Hespanha & Rice (2016) that over
exaggerated visuals or hyper-aesthetic visuals might literally ‘distract’ viewers from seeing
the real gravity of the situation, this presents an interesting result which should be considered
in the production of environmental news. Although previous research already indicates that
26 this study has shed light on how visuals aesthetics in specific affects readers’ memory for
both text and visuals.
In practical terms, this suggests that producers of environmental news should be
cautious to use hyper-aesthetic images (at least when portraying a wilderness or food and
agriculture frame in environmental or deforestation reporting), as this decreases appraised
urgency and gravity, in turn leading to decreased support for intervention and probability to
engage in political behavior in favor of decreasing deforestation.
Limitations
Several limitations are mentionable in the study and therefore some measures and
results should be interpreted with caution. As the randomizer did not display conditions as
intended, this resulted in an uneven distribution of participants to each experimental
condition (aesthetic n= 68, non-aesthetic n= 76, no image n= 16).
As for the poor significance in the effect of visual aesthetics on attitude and behavior,
a better option would have been to increase the number of images exposed to each
participant. Only displaying one single image per article might not have been enough for
participants to receive the stimulus as strongly as intended, and this might very well have
affected the results.
Another mentionable limitation which might have influenced the low reported
measures of support for intervention might have been due to the framing of the textual
content of the stimulus. Although the construction of a predominantly neutral text was
intended, this might not have been achieved entirely. Furthermore, the insignificant
differences in aesthetic framing effects might have been due to a poor choice of stimulus
imagery. Even after checking for objective and subjective aesthetic scores, the initial sample
of images could potentially have been more extreme in their expression and level of
27 displaying one image per article was probably too little to see any significant effect of
aesthetics in news reports. A short series of images, as seen in an increasing number of online
articles from both the BBC, The Guardian and The CNN, would have been interesting and
might would have produced more significant results (Schuck & de Vreese, 2006).
Conclusion
To conclude, the implications of these findings on public opinion formation and
policy-making processes is dual of nature, and indicates a strong call for further
investigations into the aesthetic expression of news visuals. Conforming with the results of
previous studies, that visual framing affects opinion and behavior, and that visual aesthetics
affects emotional arousal, the study extracted interesting results, which are applicable for
further research in the field.
Possibly the most important finding of the study suggests is that previous exposure to
a certain type of news (also indicating personal interest for the item) affects how visual
framing and aesthetic expression is perceived. This in turn affects how affected or ‘trigged’ readers are by ‘hyper’-aesthetic visual in their news consumption. Here it is mentionable, that
perception and attitude towards complex far-away issues is usually formed in combination
with other factors also mentioned in the theory section (personality, exposure to other news
and previous knowledge). However, as visual aesthetics moves to the center stage of news
production, this area of study needs further investigation into how aesthetic dimensions of
28
References
Abrams, J., (1995). Little Photoshop of Horrors: The Ethics of Manipulating Journalistic
Imagery. Print 49, 24-42.
Abramson, D. B. (2007). The aesthetics of city-scale preservation policy in Beijing. Journal
Planning Perspectives, Volume 22, 2007 - Issue 2, University of Washington. Plan Perspect; 22(2): 129-66.
Aydın, T., Smolic, A., and Gross, M. (2015). Automated Aesthetic Analysis of Photographic
Images. IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, VOL. 21, NO. 1.
Azrout, R., van Spanje, J. & de Vreese, C. H. (2012). When news matters: Media effects on
public support for EU enlargement in 21 countries. Journal of Common Market Studies, 50, 691-70
Babbie, E. (2005). The basics of social research (3rd). Belmont, CA: Thomson Wadsworth,
237-259, 120153.
Bengtson, D., N. & Xu, Z. (1997). Trends in national forest values among forestry
professionals, environmentalists, and the new media, 1982–1993. Soc. Natur. Resour, 10, 43–59
Blumler, J., and Kavanagh, D., (1999). The Third Age of Political Communication: Influences
and Features. Political Communication, Vol.16(3), p.209-230.
Bock, M., A., (2008). Together in the Scrum: Practice News Photography for Television, Print,
and Broadband. Visual Communication Quarterly, Volume 15, 2008 - Issue 3. 169-179.
29 Bock, A., Isermann, H., & Knieper, T. (2011). Quantitative Content Analysis of the Visual. In
Margolis, E. & Pauwels, L. (Eds.). The Sage Handbook of Visual Research
Methods. (pp. 265-282). London, Sage Publications.
Borkin, M.,A. et al. (2013). What Makes a Visualization Memorable? IEEE Transactions on
Visualization and Computer Graphics, Dec. 2013, Vol.19 (12), 2306-2315.
Brick, C., Kim, H.S., Sherman, D.K (2017). “Green to be seen” and “brown to keep down”:
Visibility moderates the effect of identity on pro-environmental behavior.
Journal of Environmental Psychology 51, 226–238.
Brady, E. & Haapala, A. (2003). Melancholy as an Aesthetic Emotion. Contemporary
Aesthetics Journal 1.
Broadbent, D. E. & Broadbent, M. H. P. (1981). Recency effects in visual memory. The
Quarterly Journal of Experimental Psychology Section A, Human Experimental
Psychology, Volume 33, Issue 1:1-15.
Carlson, M. (2009) THE REALITY OF A FAKE IMAGE News norms, photojournalistic craft,
and Brian Walski's fabricated photograph. Journalism Practice Volume 3, Issue 2, 125-139.
Chianese, R. L. (2014). Is Nature Photography Too Beautiful? American Scientist; Research
Triangle Park 102.1: 64-67.
Chong, D., & Druckman, J., (2007). A Theory of Framing and Opinion Formation in
Competitive Elite Environments Department of Political Science. Northwestern University, Evanston, Journal of Communication 57 (2007) 99–118.
Christianson, S. (1992) The handbook of emotion and memory: Research and theory. Hillsdale
NJ : Erlbaum.
30 setting effects of visual images. In P. D'Angelo & J. A. Kuypers (Eds.), Doing news framing analysis (pp. 233– 262). New York: Routledge.
Corner, A., Webster, R. & Teriete, C. (2015). Climate Visuals: Seven principles for visual
climate change communication (based on international social research). Oxford: Climate Outreach
Crang, M. (2009). The death of great ships: photography, politics, and waste in the global imaginary. Department of Geography, Durham University, Environment and
Planning A 2010, volume 42, pages 1084-1102.
Crowther, P., (1989). The Aesthetic Domain: Locating the Sublime. British Journal of
Aesthetics 29: 2131.
de Vreese, C. H., van der Brug, W., & Hobolt, S. (2012). Turkey in the EU: How cultural and
economic frames affect support for Turkish membership. Comparative European Politics, 10(2), 218-235.
De Young, R. (2000). Expanding and evaluating motives for environmentally responsible
behavior. J. Soc. Issues, 56, 509–526.
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of
Communication, 43(4), 51–65.
Fahmy, S., (2004). Picturing Afghan Women: A Content Analysis of AP Wire Photographs
during the Taliban Regime and after the Fall of the Taliban Regime. International Communication Gazette, 66 (2), 20-30.
Fahmy, S., Bock, M. A., & Wanta, W. (2014). Visual communication theory and research: A
mass communication perspective. New York: Palgrave Macmillan. Fahmy, S., Cho, S., Wanta, W., & Song, Y. (2006). Visual agenda-setting after 9/11:
Individuals' emotions, image recall and concern with terrorism. Visual
Communication Quarterly, 13(1), 4–15.
31 Media Studies.
Garry, M., Strange, D., Bernstein, D. M., & Kinzett, T. (2007). Photographs can distort
memory for the news. Applied Cognitive Psychology, 21(8), 995–1004.
Geise, S., & Baden, C. (2014). Putting the image back into the frame: Modeling the linkage
between visual communication and frame-processing theory. Communication
Theory, 25(1), 46–69.
Gibson, Rhonda, and Dolf Zillmann (2000). Reading Between the Photographs: the Influence
of Incidental Pictorial Information on Issue Perception. Journalism and Mass
Communication Quarterly 77: 355-367. 24 Sept. 2007.
Gilbert, K. & Schleuder, J. (1990). Effects of Color and Complexity in Still Photographs on
Mental Effort and Memory. Journalism Quarterly 67. 24. Sep 2007.
Grabe, M., E., Lang, A., & Zhao, X. (2003) Implications for Memory and Audience
Evaluations. News Content and Form, Communication Research.
Greenwood, K., & Jenkins, J. (2015). Visual framing of the Syrian conflict in news and public
affairs magazines. Journalism Studies, 16(2), 207–227.
Haas, A. F., et al. (2015). Can we measure beauty? Computational evaluation of coral reef
aesthetic. PeerJ 3: e1390.
Hansen, A., & Machin, D. (2013). Researching visual environmental communication.
Environmental Communication, 7(2), 151–168.
Harcup, T., & O'Neill, D. (2001). What Is news? Galtung and Ruge revisited. Journalism
Studies, 2(2), 261-280.
Harcup, T & O'Neill, D. (2016): What is news? Journalism Studies. Published by Taylor &
Francis
32 and liberal democracy. Chicago, IL: University of Chicago Press.
Hart, P., S. & Feldman, L. (2016). The Impact of Climate Change - Related Imagery and Text
on Public Opinion and Behavior Change. Science Communication, Vol.38 (4), pp.415-441.
Hayes, A. F. (2013). Fundamentals of moderation analysis. In Introduction to Mediation, Moderation, and Conditional Process Analysis (pp. 207-244). New York: NY: Guilford Press.
Hilgartner, S., & Bosk, C. L. (1988). The rise and fall of social problems: a public arenas model. American Journal of Sociology, 94(1), 53-78.
Huang, Y., & Fahmy, S. (2013). Picturing a journey of protest or a journey of harmony?
Comparing the visual framing of the 2008 Olympic torch relay in the US versus the Chinese press. Media, War & Conflict, 6(3), 191–206.
Huhmann, B, A., (2003). The role of color, photography, and animation: Visual complexity in
banner ads. Visual Communication Quarterly Volume 10, Issue 3, 10-17.
Isola, P., Xiao, J., Parikh, D., Torralba, A. & Oliva, A. (2014). What Makes a Photograph
Memorable? IEEE Transactions on Pattern Analysis and Machine Intelligence, July 2014, Vol.36 (7), 1469-1482.
Iyer, A., Webster, J., Hornsey, M. J., & Vanman, E. J. (2014). Understanding the power of
the
picture: The effect of image content on emotional and political responses to
terrorism. Journal of Applied Social Psychology, 44(7), 511–521.
Jewitt, C., & R. Oyama (2001). Visual Meaning: A Social Semiotic Approach. In van Leeuwen
T., and C. Jewitt (eds), Handbook of Visual Analysis, London: Sage Publications, p. 134-156.
33 Johnson, R. D. (1987). Making judgements when information is missing: Inferences, biases,
and framing effects. Acta Psychologica. Volume 66, Issue 1: 69–82.
Katz, E., Adoni, H. & Parness, P. (1977). Remembering the News: What the Picture Adds to
Recall. Journalism Quarterly, Vol. 54 (2), 231-239.
Kätsyria, J., Ravajaa, N. B., & Salminena, M. (2012). Aesthetic images modulate emotional
responses to reading news messages on a small screen: A psychophysiological investigation. International Journal of Human-Computer Studies Volume 70, Issue 1, 72–87.
Kress, G., and van Leeuwen, T., (1996). Reading images: The grammar of visual design.
London, UK: Routledge.
Lang. A., Potter, R., F., & Bolls, P., D., (1999). Something for Nothing: Is Visual Encoding
Automatic? Journal Media Psychology. Volume 1, Issue 2, 145-163
Mendelson, A. L. (2004). For Whom is a Picture Worth a Thousand Words? Effects of the
Visualizing Cognitive Style and Attention on Processing of News Photos. Journal of Visual Literacy, Volume 24, 1, 1-22.
Mendelson, A., L. and Darling-Wolf, F. (2009). Readers’ interpretations of visual and verbal
narratives of a National Geographic story on Saudi Arabia. Vol 10, Issue 6, 2009.
Messaris, P., & Abraham, L. (2001). The role of images in framing news stories. In S. D. Reese,
O. H. Gandy, Jr., & A. E. Grant (Eds.), Framing public life: Perspectives on media and our understanding of the social world (pp. 215–226). Mahwah, NJ: Lawrence Erlbaum.
Murray, S. (2008). Digital Images, Photo-Sharing, and Our Shifting Notions of Everyday
Aesthetics. Journal of visual culture, Vol 7, Issue 2.
Nahmod, S., H., (1987). Artistic Expression and Aesthetic Theory: The Beautiful, the Sublime
34 and the First Amendment. Wisconsin Law Review 1987: 221-263.
Nelson, T.E., Oxley, Z.M. & Clawson, R.A. (1997). Toward a Psychology of Framing Effects.
Political Behavior Volume 19, Issue 3: 221–246.
O’Neill, S. J. (2013). Image matters: Climate change imagery in US, UK and Australian
newspapers. Geoforum, 49, 10–19.
Powell, T. E., Boomgaarden, H. G., De Swert, K., & de Vreese, C. H. (2015). A clearer picture:
The contribution of visuals and text to framing effects. Journal of Communication, 65 (6).
Rebich-Hespanha, S. & Rice, R. E. (2016). Dominant Visual Frames in Climate Change
News
Stories: Implications for Formative Evaluation in Climate Change Campaigns.
University of California, Santa Barbara, USA, International Journal of
Communication 10, 4830–4862.
Rössler, P., Bomhoff, J., Haschke, J. F., Kersten, J., & Müller, R. (2011). Selection and impact
of press photography: An empirical study on the basis of news factors.
Communications, the European Journal of Communication Research 36, 415– 439.
Schuck, A. R. T., Vliegenthart, R., & de Vreese, C. H. (2016). Who's afraid of conflict? The
mobilizing effect of conflict framing in campaign news. British Journal of Political Science, 46, 177-194.
Schuck, R. T., & de Vreese, C. H. (2006). Between risk and opportunity: News framing and
its effects on public support for EU enlargement. European Journal of Communication, 21, 5–31.
35 and photographic art. Journal Visual Sociology, Volume 5, Issue 2.
Stromback & Nord (2009). Who Leads the Tango? A Study of the Relationship Between
Swedish Journalists and Political Sources. International Communication Association, Sheraton New York.
Sturken M. (1997). Tangled Memories: The Vietnam War, the AIDS Epidemic, and the Politics
of Remembering. London: University of California Press.
Sutcliffe, A. & Namoune, A. (2008). Getting the Message Across: Visual Attention, Aesthetic.
Design and What Users Remember. Proceedings of the 7th ACM conference on Designing interactive systems. Manchester Business School. 11-20.
Talmi, D., et al. (2008). Immediate memory consequences of the effect of emotion on attention
to pictures Learning Memory Journal, Mar; 15(3): 172–182.
Tello, V. (2004). The Aesthetics and Politics of Aftermath Photography. Rosemary Laing's
welcome to Australia. Third Text Journal, Volume 28, Issue 6.
Todorov, A., Mandisodza, A. N., Goren, A., and Hall. C. C., (2005). Inferences of Competence
From Faces Predict Election Outcomes. Science 308: 1623–1626.
Wang, J., Datta, R., & Li, J., (2009). Algorithmic inferencing of aesthetics and emotion in
natural images: An exposition. The Pennsylvania State University, UniversityPark, PA 16802, USA.
Wanta, W. (1988). The effects of dominant photographs: An agenda-setting experiment.
Journalism Quarterly, 65(1), 107–111.
Zettl, H. (1973). Sight, sound, motion; applied media aesthetics. Belmont, Calif: Wadsworth
36 Zillmann, D., Gibson, R., & Sargent, S. L. (1999). Effects of photographs in news-magazine
reports on issue perception. Media Psychology, 1(3), 207–228.
Zillmann, D., Knobloch, S., & Yu, H.-S. (2001). Effects of photographs on the selective
reading of news reports. Media Psychology, 3, 301–324.
Appendix A
Example stimulus. An aesthetic image condition with a food and agriculture frame accompanied by the text.
37 Appendix B
Example stimulus. A non-aesthetic image condition with a wilderness frame accompanied by the text.
38 Appendix C
39 A 40 image sample was taken from online news media sources (The Guardian, The New York Times, The Independent, The Economist) the same news media, as well as from other media sources notorious for their use of vivid imagery (National Geographic, WWF), specifically portraying the two frame conditions.
In order to guarantee that human participants would agree with the algorithm-based automatic aesthetic assessments, participants in the pre-test were asked to rate the images on their perceived aesthetic value on measures of salience (“the image was attention-grabbing”), visual aesthetics (“the image was beautiful”) and emotional arousal (eg. “the image made me feel sad”, “the image made me feel angry”), all using a 1-7 scale. Images generally conveyed with the algorithm scores, however the few images which scored significantly lower or higher in the human aesthetic score test were excluded from the experiment so that only the images with the most extreme score of aesthetics and emotional response were selected for the main experiment.
A total of 72 participants were recruited for the pre-test through social media. The sample was assured to be completely independent from the main experiment sample, in order to prevent any previous exposure for participants participating in the main experiment. As the aesthetics evaluation were useful even when participants only rated one image, all answers were included in the sample, resulting in a total of 72 participants even though only 41 completed the survey. (Distribution numbers and info here). In general, images were rated similar to the score received through the automatic aesthetics analysis systems. Nevertheless, some images were rated slightly contradictory and were thus excluded from the final sample, leaving only images with a ‘perfect aesthetic appeal’ (either very low or very high) to the final sample and the main experiment, making them a ‘perfect fit’ as experimental stimulus. Thus, for the final experiment, 8 out of the initial 16 images were chosen as they provoked the most extreme emotional responses as well as the most extreme aesthetic scores. The 8 images were distributed in a two by two factor design combining an aesthetic condition and a frame condition with each experimental condition, making 4 final conditions. This was done to strengthen the validity of the study by eliminating the possibility that a possible relation would not be due to the specific content of the image within a certain condition. Thus 2 images were displaying the exact same condition only with a different image content. An initial interesting observation was made which suggested a contradictory effect of aesthetics on emotional responses. (Participants were ambiguous when reporting emotions on both aesthetic and non-aesthetic images. In other words, participants sometimes reported more extreme emotions towards non-aesthetic images than towards aesthetics images).
The aesthetic scores of all 40 images was determined by comparing score of algorithm-based online aesthetic analysis software EyeEm and Sensiphi. As the two systems were occasionally ambiguous in their rating of ‘medium’ aesthetic score images, only images with a clear agreement
40 between systems were chosen for the study. For the EyeEm image aesthetics recognition system, images which had an aesthetic score >75 where considered as highly aesthetic, whereas images scoring <25 were considered non-aesthetic or at least very little aesthetically pleasing. The same images where uploaded to the Sensiphi system, where a second round automatic aesthetic analysis of all images were conducted. Relying on 109 distinct features associated with aesthetic value, grouped into three general feature groups namely texture, color and objects, the Sensiphi system ranks aesthetic value on a scale of 0-50 where lower numbers represents high aesthetic values. Therefore, only images with a score <18 very considered for the aesthetic condition, while only images with a score >32 was chosen for the non-aesthetic condition.
Finally, 16 images were selected for the final pre-test with the following distribution: 4 aesthetic wilderness and nature recreation images, 4 non-aesthetic wilderness and nature recreation images, 4 aesthetic food and agriculture images, and 4 non-aesthetic food and agriculture images.
