MA Erasmus Mundus Master
Journalism, Media and Globalisation
(joint degree)
THE DISTANT SUFFERER:
Measuring Spectatorship of Photo Journalism
by
Teresa Elena Weikmann
Student ID: 11300493
Master’s Thesis
Graduate School of Communication
Master’s programme Communication Science
Supervisor/Examiner: dhr. T.E. Powell MSc
Date of completion: 30th May 2017
Abstract
The way in which Western citizens perceive news about suffering humans in distant war and disaster zones through visual channels has received increasing attention in research on foreign reporting, whereas quantitative approaches are lacking. Recent studies on the effects of graphic news images indicate that pictures have the power to mobilize people via emotions. Yet, little is known about other potential reactions embedded in the theory of distant suffering such as apathy or voyeuristic pleasure. This study uses an experiment to, for the first time, quantify overlapping roles of the spectatorship of suffering. Via cluster analysis I explore in which combinations responses of empathy, voyeurism, protest and apathy co-occur within spectators. Furthermore, exposing participants to victimizing photographs of children and adults suffering during the Syrian civil war suggests that personal characteristics of participants like gender play a larger role in the processing of distant suffering than the pictures’ content. Besides shedding light on audience perception of suffering, this study provides empirical evidence for a fuller range of potential responses to photo journalism. The results are discussed in light of ethical difficulties with the visual depiction of war and the photo journalistic profession in general.
“The hunt for more dramatic (as they're often described) images drives the photographic enterprise, and is part of the normality of a culture in which shock has become a leading stimulus of consumption and source of value.” (Sontag, 2003)
Introduction
The picture of a naked girl running away from napalm explosions in the Vietnam war is one of the most iconic news images ever taken, gaining worldwide attention until today (Miller, 2004). Next to the Pulitzer Prize, it has won the World Press Photo Award for
outstanding photo journalism. Death, violence and pain are recurring patterns in the history of the prestigious photo contest (Zarzycka & Kleppe, 2013), but also in non-awarded photo journalism that reaches a Western audience on a daily basis. The opening quote by Susan Sontag summarizes an important claim: that photographers purposely seek for images that shock. But why? They justify the importance of their work by stating that the camera does not lie and that they are therefore fostering understanding of the dramatic sceneries of war (Zguri, 2014).
Moreover, studies on visual framing effects hint towards the essential role of emotions which are evoked by images (Iyer et al., 2014). Experiments exposing participants to graphic imagery found that they have in fact the power of mobilizing people to indicate the
willingness to protest, with the emotions of anger and sympathy being important drivers for action (Powell, Boomgaarden, de Swert & de Vreese, 2015). However, this mobilized role is just one of various ways in which a spectator might react.
Scholars have theorized about other effects of violent imagery and the portrayal of suffering, whilst claiming that it actually might turn spectators into passive voyeurs or leave them with indifference (Chouliaraki, 2006). These appear to be undesirable reactions towards images like the so-called Napalm girl. Chouliaraki (2008) further proposes that the media might function as “moral education” (p.838), as it engages spectators. The theoretical claim gains relevance when considering that many times children serve as the exemplification of the senselessness of war and are hereby being instrumentalised (e.g., Moeller, 2002;
Maciá-Barber, 2013). As previous experiments have found, spectators feel empathy when being presented with victims and might therefore show the intention to act. The single child victim can be an effective tool for raising money in humanitarian aid campaigns (Kogut & Ritov, 2005). At the same time, scholars argue that these images may leave people with indifference, or make them voyeuristic (Sontag, 2003), as they provide them with an exclusive glimpse into somebody else’s life. If the audience does not in fact care and even finds some sort of enjoyment in these images, how can the release and distinction of the photographs be justified?
In an increasingly visual media landscape, the understanding of visual effects on the spectator still remains surprisingly rudimentary (Coleman, 2010). Graphic images are found to influence citizen’s evaluation of war situations (Scharrer & Blackburn, 2015), but the ethical difficulties of graphic imagery and their potentially undesirable consequences have found relatively little attention in quantitative research. To what extent can images of
another’s suffering influence people and does the audience really empathise? This study aims to shed light on the complex effects of graphic news images on the spectator. I therefore set out to understand:
To what extent do victimizing images of children or adults in crisis photography influence the spectator’s moral perception by either turning them into an empath, voyeur, protester or leave them with apathy?
To answer this question, a survey-embedded online experiment was conducted in which participants were randomly exposed to images portraying child and adult victims. Afterwards they were asked about their perception of the images, with the aim to disentangle the various role profiles adopted by audiences when paying witness to distant suffering.
Theoretical Framework The audience of distant suffering
Every day images of suffering people are distributed in the news, bearing testimony of violent conflicts in a global setting. Professional war reporters deliver a collection of
wounded, crying and suffering humans, which “offer viscerally exciting and voyeuristic glimpses into theaters of violence” (Griffin, 2010, p. 8). Sontag (2003) dedicates the essay ‘Regarding the pain of others’ to the phenomenon of conflict photography and how it
influences its audience. The author sharply criticizes the inability of these images to help the spectator understand, claiming that the horrors of war may only be shown to those who are actually in the position to change the situation, as “the rest of us are voyeurs, whether or not we mean to be” (Sontag, 2003). Also Chouliaraki (2006) theorizes about the ‘Spectatorship of suffering’ and what impact it actually has on the audience, but unlike Sontag (2003) focuses on moving images in television. She condemns the discrepancy between “the comfort of spectators in their living rooms and the vulnerability of sufferers on the spectators’ television screens” (Chouliaraki, 2006, p. 4). For her, watching the spectacle of another’s misfortune creates an ethical dilemma, as spectators are limited in their action. Furthermore, the author expects the audience to react in various ways, as they might be turned into “voyeurs, philanthropists or [...] protesters” (Chouliaraki, 2006, p. 203). She hereby draws from Boltanski’s (1999) theory of distant suffering, who in line with Sontag (2003) claims that there is a limitation to the actions the spectator can take and suggests that there is a kind of pleasure in horrifying images, as the audience might find the aesthetics of it “sublime”
(Boltanski, 1999, p. 114). Distant suffering, as discussed by Chouliaraki and Boltanski, in this context means the presentation of the suffering of distant ‘others’ through the media (Joye, 2013).
Studying the effect of distant suffering on the audience is according to Joye (2013) urgent in the field of communication science. Drawing from semi-structured interviews with leading scholars like Cottle, Chouliaraki, Hanusch and Moeller, he concludes that audience perception of suffering needs to be tested empirically, stressing that “mediated meanings can have powerful social consequences” (Jensen, 2002, p. 273 as cited in Joye, 2013).
Audience engagement with distant suffering has so far found attention in qualitative research by Scott (2014) and Kyriakidou (2014) who conducted interviews with focus groups who were exposed to television news reports. Both agreed that there are different roles of spectatorship detectable. Scott (2014) found that participants often reacted with indifference and solitary enjoyment, stating that the suffering of others did not affect their everyday life, whilst appreciating their own safety. And this occurred more frequently for male participants. Kyriakidou (2014) identified four distinct types of audience engagement, namely affective
witnessing, ecstatic witnessing, politicized witnessing and detached witnessing. In her
interviews with a Greek focus group she found evidence for all positions and furthermore untangled the difficulties of emotional empathy and the actual limits of engagement in the real world as described in theory by, among others, Sontag and Boltanski (Kyriakidou, 2014). However, the qualitative research put forward by Kyriakidou and Scott is unable to objectively assess the identified roles of spectatorship and quantify their nature and prevalence in audiences at large.
Overall, previous research suggests that audience engagement with distant suffering is highly complex, as the spectator may take various distinct positions and therefore acts as a “’free’ agent” (Chouliaraki, 2008, p.833). The main roles identified so far can be described as (1) somebody who feels empathy for the distant other and is emotionally touched (e.g., Scott, 2014; Kyriakidou, 2014), (2) a passive voyeur, who simply contemplates in awe (e.g.,
Boltanski, 1999; Sontag, 2003; Kyriakidou, 2014; Scott, 2014), (3) somebody who is
mobilized to take some kind of action (e.g., Kyriakidou, 2014) or (4) a desensitized spectator reacting with indifference due to the inability to act (e.g., Scott, 2014; Sontag, 2003). In this study, these positions are measured within four key dependent variables, namely (1) empathy, (2) voyeurism, (3) protest and (4) apathy.
Extant research fails to show how frequent and in which combinations these detected positions occur. The role of the spectator as voyeur has so far only been theorized or
marginally observed in qualitative studies. Generally speaking, a voyeur is somebody
“[...]who seeks stimulation by visual means” (Blazer, 2006, p. 379). However, in this context, it matches more with what Kyriakidou (2014) identifies as affective witnessing, who claims that “[...] this focus on sensation and the bodily pain renders suffering into a spectacle [...]” (p. 222). This study aims to - for the first time - disentangle the different roles from one another and make them quantifiable. Therefore a sample of students (N=127) with no background in media and communication studies was chosen to identify roles and their prevalence, using cluster analysis. This analysis is useful to identify how the detected roles may overlap and what this implies normatively for the photo journalistic practice. Based on extant findings, it was not possible to hypothesize a prediction about role combinations and their frequency. The sub-research question for the study at hand is therefore as follows:
Sub-RQ: How prominent and in which combinations do the identified roles occur within spectators of distant suffering?
Photo journalism and the ethics of images
The representation of the pain of others in the media can be linked to the
photographing of war and disaster (e.g., Sontag, 2003). This concerns photo journalists, as they are the ones responsible for the creation of images. Since the Spanish Civil War was covered by the famous Robert Capa, hero-roles have been attributed to photographers depicting conflicts and therefore building the groundwork for modern photo journalism (Griffin, 2010, p. 10). Several awards such as the World Press Photo award, the Pulitzer Prize for photography or the Robert Capa Gold Medal have emerged to honour great photo
journalists. Their profession’s importance is based on the claim that they are truthfully documenting reality. However, there is an ongoing ethical discussion about this sort of reasoning, as photography is only ever an interpretation of reality and therefore highly subjective (Zguri, 2014). Furthermore, although ethical codices exist, they rarely address the handling of tragedy and violence. It is often unclear how photographers are supposed to
behave in war sceneries, as there are no precise guidelines that tell them when to remain an observer and when to intervene (Keith, Schwalbe & Silcock, 2006).
An improvement of ethical codices, also internationally, is therefore required
especially when considering that subjects of war reporting are highly vulnerable individuals who cannot control whether they are being photographed or not. Victims may hereby become instrumentalised to raise awareness. In this context, several scholars contributed to the ethical discussion of the child victim. Seu (2015) identifies children as ideal victims, as they have the potential to evoke empathy within aid campaigns and therefore can function as mobilizers. Höijer (2004) investigated this phenomenon in television reporting, stating that a child’s crying and sadness evokes pity, as it makes the spectator want to protect the subject (p.522). According to Moeller (2002), who focuses on the overall portrayal of children in foreign reporting, they are the perfect exemplification of the senselessness of war and therefore preferably used to create compassion.
It can be assumed that children are consciously portrayed, as they constitute the ideal victim. Yet again, this raises an ethical dilemma. Maciá-Barber (2013) analysed visual coverage of the Israeli-Palestinian conflict, finding that images often “violate children’s right to privacy, and frequently infringe settled principles of professional journalistic deontology” (Maciá-Barber, 2014, p.89). Other content analysis by Fahmy (2010) has also found that pictures from the Afghan war very frequently depicted women and children, which shows the dominance of human interest frames in visual conflict reporting. According to Seu (2015), who qualitatively interviewed viewers of humanitarian communication campaigns showing violated children, the exposure to the child victim can also lead to desensitization, which makes the justification of the intruding of their private sphere even more questionable.
Scholars agree that images of distant suffering create an ethical concern. On the one hand, it puts photographers in a delicate position, as they are being applauded for the depiction of suffering humans. On the other hand, the discussion focuses on the effect their
images eventually have on the distant spectator. This makes the study of different audience reactions highly relevant, especially concerning the usage of children. If the child victim evokes more compassion than a group of adults and motivates more people to take positive action, it could justify taking their photograph. If it, however, leaves spectators to a large extent with indifference, any justification of this photo journalistic practice lacks validity.
This study will therefore use award winning press photography that portrays children as victims as the key independent variable. According to the literature, children are an ideal exemplification of the negative impacts of foreign conflicts and therefore serve as adequate stimuli to portray distant suffering. This ought to be even more powerful when presenting the child victim alone, without a benefactor or in groups. According to Slovic (2007), single victims are more touching, as they constitute a psychologically coherent unit. This can lead to more distress and compassion within the spectator than when being presented with a group of victims (Slovic, 2007, p. 89). By defining the child as the ideal victim and therefore a suitable example of distant suffering, the following hypothesis is put forward:
H1: Pictures of child victims will create stronger emotional reactions and empathetic positions within the spectators than groups of adult victims.
As war photography is an important form of presenting pain through the media, it ideally exemplifies the underlying theory. When it comes to the effects of images, this furthermore has to be linked to the theory of visual framing and effects of news images in general. Consequently, in order to understand the influential effects of photography, a review of studies that have recorded audience response to images will be given.
Effects of news images
Earlier studies on the effects of images in news reporting focused mainly on the combination of text and pictures and their interaction but less on evoked emotions. According to Zillmann, Knobloch and Hong-sik Yu (2001), victimizing images can lead to a longer reading time of manipulated news reports, as they grab attention. Moreover, an early study by
Wanta (1988) found that photo size is crucial, as it can heighten the perceived importance of the content of a news report consisting of an image-text combination. Photo size should therefore be taken into consideration within any study on image effects.
According to Coleman (2010), news visuals affect consumers very differently to text: “[...] photographs, particularly highly evocative ones [...] will become part of the reader’s memory of the story, leaving an emotional trace that in time will supplant most of the written or spoken details (who, what, when, where) of the story. ‘Processing’ this story, then,
involves an emotional lamination that occurs mainly through visual cues that are processed peripherally” (p. 239). Emotions have been observed to function as mediators in visual
framing effects, which can in turn influence intended behaviour and policy support. When
exposing participants to photographs of the London bombing in 2005, Iyer, Webster, Hornsey and Vanman (2014) observed that fear, anger and sympathy lead to intended policy support, no matter if the photo showed a victim or a terrorist. Furthermore, Iyer and Oldmeadow (2006) found that people who witness extreme threat in visual news reporting react with anxiety rather than with sympathy. This emphasizes how highly complex emotional reactions towards images are. What they have in common is their tendency to influence behavioural intentions. Images can evoke an emotional response even without the support of textual information, therefore indirectly influencing actions. In a recent experiment, Powell, Boomgaarden, de Swert and de Vreese (2015) showed participants images of a relatively unknown conflict. Their findings suggest that pictures alone, without accompanying text, can trigger the spectator to show behavioural intentions (Powell et al., 2015).
In a non-news media context, Bradley, Codispoti, Cuthbert and Lang (2001) studied the psychological and physiological reaction of participants who were being exposed to pleasant, unpleasant and neutral pictures. The authors conclude that participants found
pleasant and unpleasant pictures more arousing than neutral pictures. Unpleasant pictures lead to larger blinks, skin conductance and heart rate changes (Bradley et al., 2001). Bradley et al.
(2001) furthermore stress that affective responses can translate into various reactions, reaching from mobilization for action to attention and social communication, which are dependent on the image’s context (p.296). This further underlines the complexity of reactions towards visuals and their mobilizing effects. As the feeling of empathy appears to be a
predictor for the eventual willingness to take action (Powell et al., 2015; Iyer et al., 2014), I therefore hypothesize:
H2: Participants that take the role of an empathetic spectator will be more likely to show intended behaviour for policy support and vice versa.
Hilmann, Rosengren and Smith (2003) also observed physiological responses of participants towards affective images. They found a gender difference in their study, stating that females showed an increased movement away from unpleasant pictures, such as scenes of mutilation (Hilmann et al., 2003). Gender appears to matter in regards to witnessing violence, as has already been observed within qualitative research on distant suffering (Scott, 2014; Höijer, 2004). According to Mrug, Madan, Cook and Wright (2014), violence can decrease perceived empathy. Using an empathic concern scale, the authors investigated to what extent participants feel empathy under exposure to real-life media violence. The ability to understand and empathize deteriorated at a high exposure level, especially for males (Mrug et al., 2014). Overall, men are found to enjoy violence in the media more than women (Hoffner & Levine, 2009). Hence I hypothesize the following:
H3: Male participants will more frequently take the role of the spectator voyeur and react with indifference than female participants.
Methodology Design
An online survey-embedded experiment was conducted, where participants were randomly assigned to one out of four different photographs. Two of the photos showed a child victim alone, the two others showed male adult victims in groups. The pairs were later added together and treated as one variable for the data analysis, thus improving the study’s
generalisability. The caption remained the same for all photographs. Randomization checks showed that there were no significant differences regarding knowledge about the conflict, interest in photo journalism, age or political orientation between the two condition groups. Participants were asked to fill out the survey on a laptop and not use a smartphone or tablet, to keep the size of the stimulus approximately even.
Participants
The sample of this study aspired to question a group of people with little experience regarding media and journalism, who do not have a strongly formed opinion on crisis photography or an unusually high exposure towards it. Otherwise answers might have been biased and unrepresentative of a wider audience reaction. Therefore a total number of 127 business students from a small University in Germany, aged 18-34, constitute the sample of this study. Two e-mails over the course of two weeks were sent to 520 of all currently enrolled business students. Additionally, the researcher visited several lectures personally, both to briefly present the study and to motivate students to participate. Out of 157 responses, 26 had to be excluded due to incompletion of the survey. Two more negated to be currently enrolled as students. As the gender of the respondent is crucial to test the hypotheses, two participants who didn’t indicate their sex were also excluded. The majority of participants is female (n=81). The political orientation of participants turned out to be moderately distributed (1=very left; 10=very right; M=4.55, SD=1.48). An attentiveness check was included at the beginning of the questionnaire, warning people to pay close attention in the case of failure (n=28). More information about the constitution of the sample can be found in the appendix (A).
Stimuli
The photographs constituting the child victim frame were retrieved from the World Press Photo Award website and are part of a photo essay called ‘Douma’s children’ by photo journalist Abd Doumany. It was awarded the second prize of the prestigious award in 2016 in
the category General News stories. Originally, more photos from the essay were aspired to have been included, but a pre-test revealed that only those were unambiguously identified as showing children. To contrast the single child victim, two photos of groups of adults were selected, also taken in Douma, Syria in the aftermath of the airstrike massacres that happened there in 2014 and 2015. These were retrieved from regular news websites. The two pairs showed a very similar content (child close-up with blood in their faces, adult males carrying a wounded person). At least two photos were picked per stimulus group to make the result more generalisable and not just draw conclusions from one single photograph.
The Syrian civil war is very present in the news nowadays. The specific event of the airstrike massacres in Douma however is a rather unknown event. Even though it would have been ideal to choose a less well-known conflict to minimize prior exposure, it is important for the ethical discussion to include prize-winning press photography in this study. As this is predominantly awarded to pictures covering relevant current conflicts, choosing the specific photos from Douma seemed to be an adequate compromise. Participants furthermore
indicated a rather low self-reported knowledge about both history of the civil war (M=3.16,
SD=0.78; 1=Not knowledgeable at all, 7=Extremely knowledgeable) and especially the
airstrike massacres (M=1.40, SD=0.69). The exact stimuli are included in the appendix (B).
Procedure
After the attention check, participants were asked to self-report their knowledge about the background of the photographs, their overall interest in press photography and the amount of exposure to pictures portraying war, conflict and disasters plus their political orientation. Upon encouraging them to take a careful look, the students were then randomly assigned to one of the photographs and the caption to focus on for at least 20 seconds. The pre-test
showed that this was the minimum time needed to capture the basic content of the photograph and scan the caption. After the forced exposure time participants were allowed to continue whenever they felt ready. The choice of looking at the picture longer than necessary was
expected to reveal additional information about the voyeur-role. Then the measures of dependent variables of the different roles of spectatorship were displayed, as specified in the Measures section. At the end of the survey, participants were asked to provide basic personal information, followed by a de-briefing.
Measures
Self-reported knowledge was measured on seven-point Likert scales (1=Not knowledgeable at
all, 7=Extremely knowledgeable), interest in photo journalism and exposure to war
photography on five-point scales (1=Not interested at all, 5=Extremely interested; 1=Never,
5=Every day).
After viewing the stimulus, the key dependent variables were measured (1=Strongly
disagree, 7=Strongly agree). This started with three items measuring the role of the empath
(α = .75), using questions that have been adapted from Escalas and Stern (2003), who define empathy as the “capacity to feel within or in another person’s feelings” (p. 567). One example of measurement is therefore: ‘Looking at this photo, I could imagine what the person on the
picture was feeling’. Secondly, voyeurism (α = .77) was measured. Here inspiration was
drawn from studies conducted by Nabi (2003) and Baruh (2010), who measured self-reported voyeurism within the audience of reality TV-shows. Even though this constitutes a different context, the studies appeared to be appropriate templates, as they also measure reactions towards visuals. The items combined the pleasure and appreciation of insight that the photograph was giving. An example is therefore: ‘I like the picture because it gives me an
insight that I normally wouldn’t get’. The role of the protester (α = .73), in theory somebody
who feels the urge for action, was measured within the same items used by Powell et al. (2015), including ‘intention to discuss, donate, sign a petition’. Lastly, self-designed
measures of indifference (α = .89) were included, drawn from the statements of participants in Scott’s (2014) study, combining the lack of affect due to spatial distance, e.g. ‘Looking at the
photograph leaves me with indifference, as all this is happening so far away’. See all exact
items in the appendix (C).
Additionally, participants were asked two questions to determine if they appreciated the photograph in itself, to examine if they recognize a difference between award-winning and regular news photos. Then appraised salience of the stimulus was measured (I found the
photograph attention-grabbing; 1=Strongly disagree, 7=Strongly agree), followed by
emotions of sympathy (sympathetic, compassionate, empathetic; α = .78), anger (angry,
outraged, furious; α = .89) and distress (worried, upset, anguished; α = .79) on five-point
scales (1=not at all, 5=extremely). Originally, only anger was included, but participants of the pre-test indicated that they would have liked to choose from a larger pool of emotions. A successful manipulation check question asked people whether there was a child in the picture they were presented with, for which the child was always clearly identified (p<.001).
Analysis
Cluster analysis was conducted to answer the proposed sub-research question, which classified participants into distinct groups with differing role combinations. Additionally, a Pearson’s correlation was performed to examine how the roles of empathy, voyeurism, protest and apathy are related to each other. Multivariate ANOVA was conducted, including gender (female; male) and the two image conditions (adult victims; child victim) as independent variables and the different spectator roles (empathy, voyeurism, protester, apathy) as
dependent variables. The effect of the victim images and participants’ gender on the various roles was hereby examined. In previous visual framing experiments, emotions have been found to be mediators that can influence behavioural intentions. Therefore the way in which the protester-role is mediated by various factors, including emotions (anger, sympathy,
distress) was assessed with ordinary least squares path analysis using Hayes PROCESS in
SPSS. The analysis was performed to examine the interplay of empathy and protest as well as to contribute to extant findings in the field.
Results Cluster analysis
Quantifying roles of distant suffering. Cluster analysis is an exploratory multivariate
method. Cases of a sample - in this study the students exposed to distant suffering - are being classified into a number of distinct groups. This happens based on a set of variables that should at least be measured on an ordinal scale (Cornish, 2007). Participants’ degree of empathy, voyeurism, protest and apathy were decisive here. After an initial hierarchical clustering using Ward’s method with squared Euclidean distances, a Dendogram indicated that there should be three to five clusters recognizable within the sample (appendix D). However, for this study, a k-means cluster was found to be most suitable, considering the moderately sized dataset and the highly interpretable outcome. The k-means method allows the researcher to decide how many clusters should be put out by the data analysis software. Hence the clustering was repeated several times for different numbers of clusters, whereas the option with four was found to be most meaningful, producing most dissimilar groups.
Repeating k-means and comparing options is a recommended procedure (Cornish, 2007). As all variables were measured on the same scale (1=strongly disagree; 7=strongly agree), there were no standardized Z-scores created. To double-check the choice of four clusters, two-step clustering in SPSS was conducted. This method indicates to what extent the different
variables work as predictors for classification into one of the clusters. The model summary indicated that the cluster quality is fair, whereas all variables had a high predictor importance (0.68-1.00), meaning they all had a meaningful impact on the classification of cases (for more information see appendix D). The visualization of the k-means cluster was found to be most straightforward to interpret and is therefore depicted in the graph below.
Figure 1: Bar chart of final cluster centres retrieved from k-means clustering
As can be seen from the bar graph, there are four different roles of spectatorship (clusters). For all clusters, protest and empathy behave very similarly - when somebody scores high or low on one, the person will also score equivalently on the other. The clusters can be summarized as follows:
(1) “The apathetic spectator” - people high on apathy, low on everything else
(2) “The conflicting spectator” - people who have approximately the same mean on all of the roles
(3) “The ideal spectator” - people very high on protesting and empathy including a noteworthy score on voyeurism, lowest on apathy
(4) “The moral spectator” - people higher on protesting and empathy in comparison to apathy and voyeurism
The last two labels mostly differ by their degree of voyeurism. The third is called ‘ideal’ as this group has the highest means of empathy and protest - desirable features of a distant sufferer. The fourth however seems to be more ‘moral’, as voyeurism and apathy are equally low in this group. The exact mean scores of final cluster centres as well as how frequently they occur is shown in the table below:
Table 1: Final cluster centres retrieved from k-means clustering
Clusters
Roles 1 2 3 4
apathetic spectator conflicting spectator ideal spectator moral spectator
Empathy 2.33 4.24 5.75 4.32
Voyeur 1.81 3.86 4.44 2.00
Protest 1.52 3.68 5.39 3.84
Apathy 5.29 3.72 1.63 2.12
Total 127(100 %) 7 (5.5 %) 35 (27.6 %) 51 (40.1 %) 34 (26.8 %)
Note: N=127; high scores indicate high values of attitudes (1=strongly disagree, 7=strongly agree). The variables that were most central in the formation of the clusters are shaded grey (the more important, the darker the shade).
The results of the k-means cluster answers the posed sub-research question, as it shows in which combinations and how frequently the roles of spectatorship occur: Overall, apathetic people constitute the smallest group (n=7; 5.5%). There are two large groups including empathetic protesters (n=51; 40.1% / n=34; 26.8%). Moreover, there is a
considerable group of people who appear to be conflicting in their attitudes towards distant suffering (n=35; 27.6%). When repeating k-means separately for both condition groups (1=adult victims; 2=child victim) these clusters did not change remarkably.
Correlations
There was a strong, positive correlation between empathy and protest, which was statistically significant (r=.613, N=127, p<.01). Also empathy and voyeurism correlated significantly (r=.396, N=127, p<.01) as well as voyeurism and protest (r=.424, N=127,
p<.01). Apathy correlated negatively with all other roles, very strongly with empathy
(r=-.509, n=127, p<.01) and protest (r=-.643, N=127, p<.01), least with voyeurism (r=-.219,
N=127, p=.013). When taking the results from k-means clustering and correlations together,
the second hypothesis H2 can be fully supported: participants that take the role of an empathetic spectator will be more likely to also show characteristics of a protester and vice versa.
Table 2:Bivariate correlations between the key dependent variables
Correlations
Dimension Mean SD Empathy Voyeurism Protest Apathy
Empathy 4.76 1.36 1 .396** .613** -.509**
Voyeurism 3.48 1.43 1 .424** -.219*
Protest 4.29 1.32 1 -.643**
Apathy 2.54 1.33 1
Note: N=127; values are Pearson’s correlation coefficients; marked values are statistically significant (*p<.05, **p<.01).
MANOVAs
Exploring responses towards victimizing images. The image conditions (1=adult
victims; 2=child victim) didn’t have a significant effect on the respondent’s indications of empathy, voyeurism, protest or apathy. Additionally to the role of the empath, emotions of sympathy were measured. A supplementary t-test revealed that overall people who were given the image of a child condition reported significantly stronger emotions of sympathy (M=3.47,
SD=0.92) compared to those exposed to the image of adults (M=3.10, SD=0.98);
t(125)=-2.159, p=.033. However, due to a lacking effect of the conditions on the role of the empath, the first hypothesis H1 can only be partly supported: participants under the child image conditions do not rather take the role of an empathetic spectator, but show more empathetic emotions.
There were no statistically significant interaction effects between gender and the conditions in regards to the dependent variables, however a weak interaction for voyeurism was observed F(1,123)=3.226; p=.075; Wilks’ Λ=.949 (based on a two-tailed test). Females were more voyeuristic towards the adult (M=3.76, SD=1.47) than males (M=3.03, SD=1.31), but men more voyeuristic when facing the child (M=3.58, SD=1.55) than women (M=3.36,
SD=1.37). Moreover, there was a statistically significant difference in self-reported empathy
apathy F(1,123)=7.126, p=.009; Wilks’ Λ=.879 based on the respondents’ gender. Females reported significantly more empathy towards the child (M=5.35, SD=1.09) compared to males (M=4.28, SD=1.38). Also, women who saw the adults indicated more behavioural intentions (M=4.43, SD=1.08) than males (M=3.73, SD=1.30). In regards to apathy, male respondents reacted significantly stronger with indifference (M=2.96, SD=1.47) than females (M=2.08,
SD=1.11) when seeing the child victim. As for voyeurism, no significant differences could be
found. Due to the gender gap regarding empathy and especially apathy, hypothesis H3 is partly supported: male participants are more likely to react with apathy, but not necessarily with voyeurism. Overall, gender is found to be a higher predictor for the taken role than whether a respondent saw a suffering child or adult victims. Mean differences are visualized in the figure below.
victim image protester sympathy anger distress a path = -.33* b path = .55*** .37*** .22** .12 -.17 c path = -.003 (p=.99) (adults = 1 child = 0) R²=.54 f (1.123) =36.593 sympathy ab=-.03 CI = -.41 to -.02 Mediation
Measuring the role of emotions. The exposure to victims had a significant indirect
effect of sympathy on the role of the protester. As can be seen from Figure 3, participants who saw the child victim reported stronger empathetic emotions (a= -0.33) and in turn displayed a higher willingness to act (b=0.55). There was no significant indirect effect of anger and distress, whereas the emotions had a direct effect on the role of the protester.
Figure 3: Indirect effect of the victim images on the role of the protester through emotions
Note: N=127; Mediation model showing the indirect effects of emotions. Unstandardized beta coefficients are shown, *p<.05, **p<.01.; ***p<.001; 95% bias-corrected bootstrap confidence intervals based on 5,000 bootstrap samples are shown for significant indirect effects.
These results contribute to extant findings of visual framing effects, as they demonstrate how emotions can function as mediators. Considering a victim frame was used in this experiment, the indirect effect of sympathy appears plausible. Furthermore, hypothesis H2 finds support in this analysis, as it shows how empathy as an emotion ultimately influences protest.
Discussion
This study investigated how varying depictions of victims of the Syrian civil war affect the viewer’s morality and which different roles of spectatorship an audience adopts when being presented with distant suffering. It is therefore the first study to tackle the underlying theory with a quantitative approach, identifying role combinations and making
positions measureable. Four types of distant sufferers have been identified: (1) the apathetic spectator, (2) the conflicting spectator, (3) the ideal spectator and (4) the moral spectator. This result matches with extant findings from focus group research, like the detected indifference by Scott (2014) or “politicised witnessing” described by Kyriakidou (2014), which complies with the ideal spectator role. This study is however the first to provide quantifiable evidence for these roles and shows that there may be conflicting positions united within spectators.
The experiment further contributes to research on the effects of graphic news images. There were small, but insignificant differences between participants who were presented with a suffering child and those who saw groups of adults in a desperate situation. Instead,
individual characteristics mattered for the moral position taken: there were significant gender differences found within the condition groups, as women overall reacted with more empathy and showed a willingness to act in great contrast to males, who indicated stronger
indifference. Besides being in line with results from focus group research by Scott (2014) or Höijer (2004), this further underlines the complexity of image effects as they are dependent on individual characteristics (Domke, Perlmutter & Spratt, 2002).
So what do these findings imply for photo journalists and their presentation of suffering in the light of media ethics? Overall, this study suggests that Boltanski’s (1999) theory is not about the type of suffering that is being witnessed, but about the witnesses themselves. When it comes to the presentation of another’s misfortune through the media, the audience seems to matter more than content. In this study, gender as a personal feature was taken into account and turned out to have a stronger impact than the stimulus. Also, items like interest in photo journalism and other covariates may be taken into account in future studies.
A reason for the lack of significant effects of the stimuli might be that participants were numbed by death toll numbers. The accompanying caption contained statistical
information about casualties of Douma’s airstrike massacres. When being reminded of all the other victims and the extent of the suffering, empathy for a single case may decrease (Small,
Lowenstein & Slovic, 2007). Overall, the interaction of image and text matters hugely when examining the effects of photographs, as it can dominate the response towards the visual (e.g. Powell et al., 2015; Maier, Slovic & Mayorga, 2016). However, as there were also no big differences regarding voyeurism, protest or apathy, this rather speaks for the overshadowing of media effects by personal characteristics. Nonetheless this study underlines that emotions play a role in the effects of news images, as sympathy was found to be a significant mediator. People who saw the child reported more sympathy and were in turn more likely to show behavioural intentions and take the role of the protester. This is in line with other experiments that used victim frames as stimuli (e.g. Iyer et al., 2014), which explains the significant indirect effect of sympathy in contrast to the other measured emotions.
The cluster groups of people indicating strong empathy and protest have been identified to be the biggest within this sample, regardless of the actual content of the
photograph. They are here called moral spectator and ideal spectator. Despite a noteworthy amount of voyeurism, the latter is called ideal, as it’s the role with the highest means
regarding honourable implications. The empathy and protest items even correlated positively with voyeurism, indicating that this may be a factor worth observing, but one that does not constitute a threat towards noble intentions. Does this make its condemnation by Susan Sontag (2003) and others obsolete?
Even if voyeurism does not appear to diminish empathy or protest, the detected roles in this study are to some extent worrisome. The fact that there was a group present in the identified clusters that reacted predominantly with apathy should concern journalists
responsible for the depiction of distant suffering. One could however also see the limited size of the cluster-group (5.5% of the sample) as a positive sign regarding the morality of
spectators. Replication is needed to provide perspective of this role in other contexts. Simultaneously, the role of the conflicting spectator is worth further observation. People might have become overwhelmed by their exposure to the suffering others, and confused
about what to feel. Alternatively, this group of spectatorship could simply be realistic about their scope to help, therefore unifying all potential reactions to the same extent, as all seem appropriate. However, the apathetic and conflicting spectator are two undesirable outcomes resulting from witnessing another’s pain, which pose a serious threat to the vindication of victimizing images and support the criticism put forward by various scholars (e.g. Sontag, 2003; Chouliaraki, 2008; Höijer, 2004).
Moreover, showing the intention to act and actual behaviour are two very distinct things. Chouliaraki (2006) claims that good journalism exceeds the act of purely reporting what happens, but engages people to take action. A trigger for this can be empathy. This study provides evidence that behaviour is indeed highly dependent on empathy, as also the positions behave simultaneously to another. Yet, the struggle remains how to translate a feeling into meaningful actions (Maier et al., 2016). According to Kyriakidou (2014) “cosmopolitan empathy” (p.288) has a limited effect, as “the viewer is positioned within the moral space of the national and local” (p.229). It is naturally easier to feel for those to whom we live in proximity. When already being limited in their ability to empathize, the spectator must be even more restricted in taking the threshold of actual behaviour.
Overall, this study largely questions the so frequently asserted ‘power of the picture’ and to what extent it can influence actions. The famous World Press Photo by Nik Ut of the Napalm-girl is credited to have ended the Vietnam War (Miller, 2004). In reality, political settlements have been made long before the picture was taken. In ‘Getting it wrong’, Campbell (2016) provides a clear analysis of historic events, claiming that the image might have lead to a public outcry, but didn’t change the process of political decision making. A similar occurrence is the photograph of Alan Kurdi, a refugee boy lying dead on the Mediterranean coast. The picture went viral on social media. But the compassion was fugacious. According to the Poynter Institute “[it] was a photo that shook the world – or at least the media – for a few days. But not much more than that” (Warren, 2015 as cited in
Maier et al., 2016). Images might have the power to move people, but they do not really start a movement (Domke et al., 2002). In this study, the assumingly more ‘powerful’ World Press Photo winning pictures did not heighten spectators’ empathetic positions significantly. Also participants did not appreciate its artistic composition more, nor did they find it increasingly attention-grabbing. Rather did personal features of the recipient have to do with how suffering was perceived by the audience.
It is crucial to note that there is likely to be a social desirability bias within the
measured dependent variables of this study, as participants will assume it more appropriate to indicate moral actions rather than admitting apathy or even voyeuristic enjoyment. Indeed empathy and protest were participants’ dominant response across the sample. Generally, the creation of the key independent variables was one of the biggest challenges for this study, as there is a limited body of literature to draw from, which makes this study all the more important. The developed items are overall reliable and yet the results have to be taken with caution. A replication of this methodological approach is urgently needed. Especially self-reported voyeurism is a challenging variable to measure. Openly admitting to ‘enjoy’ the viewing of a violently hurt child is not common behaviour and people might not even have thought about the factors that make them ‘like’ the picture before taking the survey. This may have made it harder for them to confess any appreciation of the image. Further levels of measurements of physiological nature should be taken into account in future studies (Bradley et al., 2001; Hillmann et al., 2004). For instance it will be relevant to observe whether people voluntarily exposed themselves to the picture for a longer time period than necessary.
Voyeurism was found to correlate negatively with the time spent on adult stimuli and positively with the child pictures’ timing. These findings were however not significant. Another limitation is the constitution of the sample. There were clearly more females taking the survey, which means that the conducted cluster analysis can only give an initial idea of how spectators may group together.
Despite these limitations, the creation of the items measuring empathy, voyeurism, protest and apathy can be described as successful. An additionally conducted factor analysis demonstrated that they all constitute separate dimensions (appendix E). They are therefore a valuable contribution to future research on distant suffering and should provide a stimulus to more quantitative approaches. Especially the inclusion of apathy and voyeurism in the study of graphic news visuals should be enhanced in upcoming studies, as these variables have been neglected before. Also, voyeurism is the only variable that evoked a weak interaction effect, which should be further investigated. This study’s attempt to measure this variable stands out in the research on image effects. The findings further contribute to a critical reflection of photo journalists hunting for shocking images, who can now be confronted with real numbers of how their audience clusters together. Considering the prominence of the Syrian civil war in today’s news, other and less known conflicts should also be included in the research on distant suffering. Participants may have already been desensitized by the daily exposure to news images from Syria or in contrast have an increased issue-awareness. A longitudinal approach would moreover be sensible, to explore how spectator-roles may change over time and how personal characteristics and media effects interact. Lastly, actual behaviour may be tested in the future, as has recently been claimed by Maier et al. (2016).
To conclude, this study adds to the understanding of the phenomenon of distant suffering and gives thought-provoking impulses on how to tackle the phenomenon in a quantitative way. It further underlines the complexity of visual effects and to some extent uncovers the limits of their power. Returning to the quote from the beginning, this study challenges the justification of shocking news images in a new way and takes an important step towards backing up Sontag’s claims empirically.
Acknowledgement
I would like to thank Prof. Dr. Walter Kiel, Prof. Dr. Barbara Hedderich and the Faculty of Business Studies at the Ansbach University of Applied Sciences for their support, in particular for enabling me to take a student sample for this study.
References
Baruh, L. (2010). Mediated voyeurism and the guilty pleasure of consuming reality television. Media Psychology, 13(3), 201-221.
doi:10.1080/15213269.2010.502871
Blazer, S. M. (2006). Rear window ethics: Domestic privacy versus public responsibility in the evolution of voyeurism. Midwest Quarterly: A Journal of Contemporary
Thought, 47(4), 379-92.
Bradley, M., Codispoti, M., Cuthbert, B., Lang, P., Davidson, Richard J., & Scherer, K. R. (2001). Emotion and motivation I: Defensive and appetitive reactions in picture processing. Emotion, 1(3), 276-298.doi:10.1037//1528-3542.1.3.276
Campbell, W. J. (2016). Getting it wrong: Debunking the greatest myths in American
journalism. Oakland, CA: University of California Press.
Chouliaraki, L. (2006). The Spectatorship of suffering. London: SAGE Publications.
Chouliaraki, L. (2008). The media as moral education: Mediation and action. Media, Culture
& Society, 30(6), 831-852. doi:10.1177/0163443708096096
Coleman, R. (2010). Framing the pictures in our heads: Exploring the framing and
agendasetting effects of visual images. In P. D’Angelo & J. A. Kuypers (Eds.), Doing news framing analysis (pp. 233–262). New York: Routledge.
Cornish, R. (2007). “Statistics: 3.1 Cluster Analysis.” Mathematics Learning Support Centre. Retrieved from http://www.statstutor.ac.uk/resources/uploaded/clusteranalysis.pdf.
Domke, D., Perlmutter, D. & Spratt, M.(2002). The primes of our times?: An examination of the 'power' of visual images. Journalism, 3(2), 131- 159.
doi:10.1177/146488490200300211
Escalas, J., & Stern, B. (2003). Sympathy and empathy: Emotional responses to advertising dramas. Journal of Consumer Research, 29(4), 566-578. doi:10.1086/346251
Fahmy, S. (2010). Contrasting visual frames of our times: A framing analysis of English- and Arabic-language press coverage of war and terrorism. International Communication
Gazette, 72(8), 695-717. doi:10.1177/1748048510380801
Griffin, M. (2010). Media images of war. Media, War & Conflict, 3(1), 7-41. doi:10.1177/1750635210356813
Hillman, C.H., Rosengren, K.S. & Smith, D.P. (2004). Emotion and motivated behavior: Postural adjustments to affective picture viewing. Biological Psychology, 66(1), 51-62. doi:10.1016/j.biopsycho.2003.07.005
Hoffner, C. A., & Levine, K. J. (2005). Enjoyment of mediated fright and violence: A meta-analysis. Media Psychology, 7(2), 207-237. doi:10.1207/s1532785xmep0702_5 Höijer, B. (2004). The discourse of global compassion: The audience and media reporting of
human suffering. Media Culture & Society, 26(4), 513-+. doi:10.1177/0163443704044215
Iyer, A., & Oldmeadow, J. (2006). Picture this: Emotional and political responses to photographs of the Kenneth Bigley kidnapping. European Journal of Social
Psychology, 36(5), 635-647. doi:10.1002/ejsp.316
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.
Joye, S. (2013). Research on mediated suffering within social sciences: Expert views on identifying a disciplinary home and research agenda. Interdisciplinary Science
Reviews, 38(2), 106-121. doi:10.1179/0308018813z.00000000039
Keith, S., Schwalbe, C., & Silcock, B. (2006). Images in ethics codes in an era of violence and tragedy. Journal of Mass Media Ethics, 21(4), 245-264.
doi:10.1207/s15327728jmme2104_3
Kogut, T. & Ritov, I. (2005). The singularity effect of identified victims in separate and joint evaluations. Organizational Behavior and Human Decision Processes, 97(2), 106-116. doi:10.1016/j.obhdp.2005.02.003
Kyriakidou, M. (2015). Media witnessing: Exploring the audience of distant suffering. Media,
Culture & Society, 37(2), 215-231. doi:10.1177/0163443714557981
Maciá-Barber, C. (2013). Ética, fotoperiodismo e infancia: Imagen del conflicto palestino-israelí en España. Cuadernos.info, 33(1), 89-98. doi:10.7764/cdi.33.528
Maier, S., Slovic, P., & Mayorga, M. (2016). Reader reaction to news of mass suffering: Assessing the influence of story form and emotional response. Journalism, Journalism, 08/11/2016. doi:10.1177/1464884916663597
Miller, N. (2004). The girl in the photograph: The Vietnam War and the making of national memory. JAC, 24(2), 261-290.
Moeller, S. (2002). A hierarchy of innocence: The media's use of children in the telling of international news. Harvard International Journal of Press/Politics, 7(1), 36-56. doi:10.1177/1081180X02007001004
Mrug, S., Madan, A., Cook, E., & Wright, W. (2015). Emotional and physiological desensitization to real-life and movie violence. Journal of Youth and Adolescence,
Nabi, R., Biely, E., Morgan, S., & Stitt, C. (2003). Reality-based television programming and the psychology of its appeal. Media Psychology, 5(4), 303-330.
doi:10.1207/S1532785XMEP0504_01
Powell, T. E., Boomgaarden, H. G., Swert, K. D., & Vreese, C. H. (2015). A clearer picture: The contribution of visuals and text to framing effects. Journal of Communication,
65(6), 997-1017. doi:10.1111/jcom.12184
Scott, M. (2014). The mediation of distant suffering: An empirical contribution beyond television news texts. Media, Culture & Society, 36(1), 3-19.
doi:10.1177/0163443713507811
Scharrer, E., & Blackburn, G. (2015). Images of injury: Graphic news visuals’ effects on attitudes toward the use of unmanned drones. Mass Communication and Society, 09 May 2015.doi:10.1080/15205436.2015.1045299
Seu, I. (2015). Appealing children: UK audiences’ responses to the use of children in
humanitarian communications. International Communication Gazette, 77(7), 654-667. doi:10.1177/1748048515601558
Slovic, P. (2007). "If I look at the mass I will never act": Psychic numbing and genocide.
Judgment And Decision Making, 2(2), 79-95.
Small, D. A., Loewenstein, G., & Slovic, P. (2007). Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims.
Organizational Behavior and Human Decision Processes, 102(2), 143-153.
doi:10.1016/j.obhdp.2006.01.005
Sontag, S. (2003). Regarding the pain of others. New York: Farrar, Straus and Giroux. Wanta,W. (1988). The effects of dominant photographs: An agenda-setting experiment.
Zarzycka, M., & Kleppe, M. (2013). Awards, archives, and affects: Tropes in the World Press Photo contest 2009–11. Media, Culture & Society, 35(8), 977-995.
doi:10.1177/0163443713501933
Zguri, R. (2014). Ethical issues in photojournalism. 61-70. Retrieved from goo.gl/xbYQve. Zillmann, D., Knobloch, S., & Yu, H.-S. (2001). Effects of photographs on the selective
reading of news reports. Media Psychology, 3, 301–324. doi:10.1177/0093650202239023
Appendix A Sample constitution Table1: Attention-check Frequency interested 29 excited 2 angry 2 enthusiastic 3 proud 2 ashamed 0 inspired 4 attentive 16
None of the above 99
Total 127
Warning was flashed 28 times, meaning that 99 participants passed the attention-check
Table 2:
Age Frequency Percent
18 1 .8 19 12 9.4 20 21 16.5 21 27 21.3 22 16 12.6 23 21 16.5 24 10 7.9 25 7 5.5 26 5 3.9 27 2 1.6 28 1 .8 29 2 1.6 31 1 .8 34 1 .8 Total 127 100.0 Table 3:
Country of origin Frequency Percent
Germany 112 88.2 Kazakhstan 2 1.6 Kosovo 1 .8 Austria 1 .8 Poland 2 1.6 Romania 1 .8 Russia 2 1.6 Slovenia 1 .8 Turkey 5 3.9 Total 127 100.0
gender: female (N=81; 62.8%); male (46; 35.7%); no report (N=2; 1.6%)
Table 4:
Gender Frequency Percent
Female 81 62.8
Male 46 35.7
No report 2 1.6
Appendix B
Caption:
“As the conflict in Syria entered its fifth year, the city of Douma, the largest opposition stronghold in Damascus province, was subject to sustained bombardment. Figures are hard to confirm, but the Violations Documentation Center in Syria (VDC), an independent NGO, puts civilian fatalities in and around Douma at 1,740 in the first half of 2015 alone” (World Press Photo, 2015).
Figure 1: child pictures 1 and 2
Doumany, A. (2015). Douma's children [Photograph found in General News, second prize stories, World Press Photo, Amsterdam]. Retrieved May 23, 2017, from
https://www.worldpressphoto.org/collection/photo/2016/general-news/abd-doumany (Originally photographed 2015, February 2).
(1) A wounded boy awaits treatment, following airstrikes on a market and a hospital, which, according to the SOHR, were carried out by government forces.
Figure 2: adults pictures 1 and 2
(1) [Photograph found in AFP/Getty Images]. (2015, November 7). Retrieved May 23, 2017, from http://www.ethiogrio.com/news/world-news/35858-suspected-russian-air-strikes- bombard-main-street-in-syrian-city-of-douma-leaving-at-least-23-dead-including-six-children.html
A man who is covered in blood is carried to safety after air strikes hit the town of Douma in Syria today, killing at least 23 people
(2) Mamet, B. (2014, September 14). [Photograph found in Reuters]. Retrieved May 23, 2017, from http://blogs.ft.com/photo-diary/category/middle-east/page/24/
Residents carry an injured man after what activists said were two air strikes by forces of Syria’s President Bashar al-Assad on a market in central Douma, eastern al-Ghouta, near Damascus
Appendix C Items measuring key dependent variables (English)
Empathy (1=strongly disagree; 7=strongly agree):
(1) When I saw this picture it made me feel compassionate for the person.
(2) Looking at this photo, I could imagine what the person on the picture was feeling. (3) When seeing this photo, I could imagine what it must be like if this would happen to
me.
Voyeurism(1=strongly disagree; 7=strongly agree):
(1) In a way, I enjoy looking at the picture, as it helps me get a peek into somebody’s life. (2) I like the picture because it gives me an insight that I normally wouldn’t get.
(3) I appreciate the picture for showing me something I usually wouldn’t see.
Protest (1=strongly disagree; 7=strongly agree):
(1) When I saw the picture I felt the urge to discuss the conflict in Douma.
(2) Looking at the photograph makes me want to donate for the victims of the conflict. (3) After looking at the photograph I would sign a petition that helps to stop the suffering
in Douma.
Apathy (1=strongly disagree; 7=strongly agree):
(1) Looking at the photograph leaves me with indifference, as all this is happening so far away.
(2) Looking at the photograph didn't really affect me as there is nothing I can do anyway. (3) The photograph gives me a feeling of apathy, as the events don't concern me.
Appendix D Additional information hierarchical cluster
Figure 1: Dendogram using Ward Linkage
The horizontal axis of a dendrogram entails the distance or dissimilarity between clusters (Eucledian distance). The vertical axis represents the cases and clusters. The splitting of one horizontal line into two horizontal lines indicates the joining or fusion of two clusters. The horizontal position of the split, shown by the short vertical bar, gives the distance
(dissimilarity) between the two clusters.
Clustering in NCSS, Hierarchical Clustering / Dendrograms. (n.d.). Retrieved May 23, 2017, from https://www.ncss.com/software/ncss/clustering-in-ncss/ [Documentation PDF]
Additional information 2-Step cluster
Table 1: Predictor importance of continuous variables determined by 2-step cluster
Nodes Importance
Voyeurism 0.6754
Empathy 0.8804
Protest 0.9662
Appendix E Factor Analysis
Table 1:
Dimensions Components Urge for action
(protest/apathy)
Voyeurism Empathy Degree of
Commonality
1/2 3 4
Apathy The photograph gives me a
feeling of apathy, as the events don't concern me.
-.900 .828
Looking at the photograph didn't really affect me as there is nothing I can do anyway.
-.893 .821
Looking at the photograph leaves me with indifference, as all this is happening so far away.
-.820 .677
Protest After looking at the photograph
I would sign a petition that helps to stop the suffering in Douma.
.625 .560
Looking at the photograph makes me want to donate for the victims of the conflict.
.589 .561
When I saw the picture I felt the urge to discuss the conflict in Douma.
.573 .422
Voyeurism I like the picture because it
gives me an insight that I normally wouldn’t get.
.884 .784
I appreciate the picture for showing me something I usually wouldn’t see.
.768 .645
In a way, I enjoy looking at the picture, as it helps me get a peek into somebody’s life.
.747 .649
Empathy When seeing this photo, I could
imagine what it must be like if this would happen to me.
.640 .771
Looking at this photo, I could imagine what the person on the picture was feeling.
.849 .773
When I saw this picture it made me feel compassionate for the person.
.640 .828 .622
Self-values (eigenvalues) 5.276 1.805 1.054
Variance explained (%) 43.969 15.093 8.782 67.790
Kaiser-Meyer-Olkin = .849
