A study on the news media’s verbal and visual frames, and how these
influence the subjects’ perceptions of the danger.
Oda Brandal Erlandsen 11696710 Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Supervisor: Dr. R. Azrout 28th of June, 2019
Abstract
This research investigated if the verbal and visual frames of the media influences the public’s danger perceptions. To answer the research question, an online survey embedded experiment was executed, with a 2x2 full factorial design with two between-subjects verbal frames (terrorist vs. attack) and two between-subjects visual frames (small font vs. big font). Further, the research aimed to explore if the frames were mediated by negative emotions, and if the relationship between verbal frames and negative emotions was moderated by previous experience of a terrorist attack. Neither the verbal nor the visual frames did have a significant effect on the subjects’ perceptions of the danger, which also further lead to that there could not be determined a mediation nor interaction effect. However, previous experience of a terrorist attack significantly moderated the relationship between the verbal frames and emotions. This study contributes to previous studies of media effects, and demonstrates that negative emotions play an essential role in how the media content is perceived. In addition, if one has personally experienced a terrorist attack, the verbal frames of the media are likely to contribute to stronger emotional reactions.
1
Introduction
“The essence of framing is selection to prioritize some facts, images, or developments over others, thereby unconsciously promoting one particular interpretation of events.” (Norris, Kern and Just, 2003)
In today’s globalized world, the media have become an important institution to which millions of citizens turn to for information everyday. One influential way in which media can shape public opinion is by framing both events and issues in particular ways (de Vreese, 2005). Albertson and Gadarian (2014) believed that subjects are biased information processors, they remember and agree with threatening information. This was further studied by Soroka and McAdams (2015), who believe one should understand the nature of news content partly as a consequence of the asymmetry negativity bias observed in human behaviour, and that the propensity to over-represent negativity in the news is a product of profit-maximization. Therefore, the media frame their content towards what their audience is most likely to be reactive towards. After all, journalists and editors are humans as well, and inhabit the same negativity bias as their audience (Soroka and McAdams, 2015). This would explain why there is an overrepresentation of threatening and negative information in mass media content today.
Public knowledge on terrorism is massively influenced by news and other types of media. This further amplifies people’s fears and makes everyone feel as they could be the victim of an attack (Nacos 2003; Norris et al. 2003; Robertson, 2003; Picard 1993; Kasperson et al. 2003; Cohen 1983). Shoshani and Slone (2008) argue that there is a symbiotic relationship between the media and terrorism, where the
media have been referred to as “the theatre of terror”. However, while the terrorists may be the ones writing the script and performing the drama, “the theatre of terror” is only possible if the media provides a stage, and presents the news to a worldwide audience. What becomes important is that terrorism is aimed at the people watching, and not at the actual victims (Wiemann 1983). Media power is generally symbolic and persuasive, where they have to a certain extent the potential to control the minds of their readers (van Dijk, 1995). Furthermore, the perception of a political message, or news in this study, is not only shaped by textual information, but by its visual representation (Schindler, Krämer and Müller, 2017). For example, the font of the text can have a great influence on how the reader react to the information. Big headlines are more dramatic and contrarily to small headlines scream a message. Therefore, by including big headlines, the media can enhance the terrorist’s message and make it not only more visible but also more threatening. Moreover, when both visual and verbal frames are combined, impact on the public is further strengthened and their perception of danger is greater.
In addition, as presented by Small, Lerner and Fischoff (2006) causal attributions after an aggressive and tragic event, such as a terrorist attack, may be shaped by emotions. Being presented articles and information about dramatic events can therefore further be shaped by negative emotions, which will lead to a stronger perception of threat. As terrorism is both dramatic and threatening (Gans, 1979), journalists and editors have incentives to use emotionally powerful tools, visuals and storylines in what is referred to as a competitive media environment. This is to keep ever-shrinking news audiences, where these elements of media coverage can strengthen the public’s perceptions of the threat (Gadarian, 2010). In addition,
previous experiences can contribute to an increased emotional response, and how one finally perceives the threat presented in the news. If a person has themself been close to or been a direct victim of a terrorist attack it is likely that news concerning terrorism will arouse strong negative emotions.
This study intends to take the points above further, and research the extent to which textual and visual representation of an attack can influence subjects’ perceptions of the danger. By focusing on the framing of threatening news content, this paper intends to explore if this type of information and representation enhances negative emotions, and how it can influence the readers’ response. Additionally, when the audience have personally experienced a terrorist attack, if this can moderate the relationship between the verbal frames and emotions.
The research question will therefore be as follow:
RQ: To what extent do the media’s frame of an attack affect subjects’ perceptions of the danger?
● Does the use of the term terrorism in the news media affect subjects’ perceptions of the danger? And does font size independently have the same effect?
● Is there an interaction effect when combining the independent variables? ● Further, can it be moderated by previous experience and mediated by negative
emotions? 4
Theoretical framework
The face of war and political conflict have been through a change. While in the past war was fought in front-lines by the military, modern warfare occurs in the midst of civilian populations (Shoshani and Slone, 2008). These changes have been paralleled with rapid developments in the technology in the media, where these have come to reflect the environment by increasing the vividness and accessibility of news, where terrorism today is very much present (Carruthers, Cochrane and Pain, 2000). Cooper (2001) argued that to label a violent act as terrorism can be regarded as a political tool, but it is also a beneficial tool for the news media. Already in 1922, Walter Lippmann worried about the impact the media have on the democratic society, where he described the images that were created in the media as “pictures in our heads”. The pictures are employed by the public to make sense of the information they are presented. This is referred to by contemporary media as schemas, which people use as a frame of reference to organise both existing knowledge and incoming information (Coleman, 2010). As a consequence, an individual frame can possibly be reinforced, challenged, but also remain unaffected.
To follow the news was argued by Boukes and Vliegenhart (2017) to be a desirable behaviour, as it is crucial for what is today regarded as a functional democracy. It evokes everything from political interest, increasing knowledge about politics and motivates political participation. However, previous studies reveal that the news media is dominated by negative news, and that negativity is what humans are most interested in and also most reactive towards (Albertson and Gadarian, 2014). How news is negatively framed can therefore have a great impact on what citizens
decide to read and their further response to what they are exposed to (Soroka and McAdams, 2015).
Framing is always used when presenting information in the news media to the public. Frames are discussed and evaluated differently in a wide range of communication literature, and therefore it is essential for research to define and present a proper definition. Entman (1993) argues that to frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described. While frames may have a range of causes and effects, they exist first of all as words, images and symbols that appear on paper and in other media (Woods, 2014). Further, in political and mass communication, Reese (2011) argues that frames influence how people understand, remember, evaluate, and act upon a problem. This paper follows the understanding of frames that is evaluated and discussed by McQuail (1987), that a frame may refer to a characteristic of communication content. The news media is responsible for providing news in the “pseudoenvironment”, where one relies on experience and subjectively understands events that cannot be directly observed. Therefore, news frames are important in relation to how events are reported, because they reflect a process of recurring selection and emphasis in communication perceived reality (Entman 1993; Gitlin 1980). This research will focus on the power of discourse analysis, which studies the use of written, vocal, and/or sign of the language (Coulthard, 2014). The study at hand will focus on understanding framing based on visuals and words used in the media and evaluate the effect of these when communicating to the public.
Verbal Frames and Emotions
By framing their content in a certain way the media is able to construct and influence the perceptions of individuals in society ( Scheufele and Tewksbury, 2006 ). Previous research has revealed that terrorism has a negative effect on everything from economic performance, tourism to psychological well being (Dinesen and Jæger, 2013). In 2011, Woods argued that the terrorism frame is extremely powerful and that the salience of terrorism can affect subjects’ perceptions of the danger. Later on, Brinson and Stohl (2012) suggested that the use of verbal frames, such as ‘terror’, not only influences the public’s reaction to an event, but also their attitudes and how those are processed in the formation of judgements. The use of certain words can have a beneficial power to influence readers to react a certain way, as it can lead to different cognitive processes. Cooper (2011) suggested that by labelling a violent act as terrorism it can prime people’s concerns about the danger, attributions of their behavior and assumptions on how to stop future attacks. This is because terrorism is seemed as a threat people may have a strong emotional relation to, as it is uncertain and can occur at any time or place. Therefore, the terrorism frame is likely to increase the public’s perceptions of the danger, as opposed to if it was framed as an attack (Woods, 2011).
In addition, studies suggest that it is important to consider the role of
emotional responses in public opinion and political judgement in the field of political communication (Cho et al, 2010). A wide range of research supports the existence of a relationship between media coverage on terrorism and negative emotional consequences (Shoshani, and Slone, 2008). Emotions are a strong contributor to people’s reaction towards what they are exposed to; terrorism news stories are likely
to induce negative emotions as they evoke a fear of death (Landau et al, 2004). In fact, in a study, Slone, Shoshani and Baumgarten-Katz (2008) revealed that participants exhibited a higher anxiety level after they viewing terrorist related media broadcasts compared to non-terrorism-related political news. The power of emotions is commonly used in the news to affect how citizens process information. Further studies also concluded that terrorism has a negative psychological impact: it increases anxiety (Gunter, 1994; Slone, 2000), anger, (Newhagen, 1998; Yukawa, Endo, & Yoshida, 2001) and fear (Joshi & Kaschak, 1998). The study of Gadarian and Albertson (2014) revealed that strong and negative emotions led people to pay more attention to the news information presented to them. By using emotional appeals in news, one can influence citizens to pay more attention to a news story and make them willing to take action or behave a certain way based on what they read.
In addition, the commonality of anger and anxiety can be related to the theory of protection motivation, which argues that personally threatening stimuli is producing automatic anxiety responses that together with high motivation for defence, demonstrated in anger, increases the fear for potential future attacks (Shoshani, and Slone, 2008). By taking this theory into account, negative emotions will explain the relationship between the word terrorism in the news and the subjects’ perceptions of the danger. As a consequence, the hypotheses will be as follows:
H1: Individuals exposed to the word terrorism in news articles have higher perceptions of the danger compared to individuals exposed to news articles without the word terrorism.
H2: Emotions explain the effect between the word terrorism in the news and the subject’s perceptions of the danger.
Figure 1.
Verbal Frames, Emotions and Previous Experience
Norris et al (2003) argue that the terrorism frame does not represent a distinct form of the danger, but a frame that has been repeatedly described by the mass media as a new social problem. It was revealed in the study of Slone, Shoshani and Baumgarten-Katz (2008) that individuals that had high exposure to political violence showed much greater vulnerability to exposure of media content that included the life threatening events. As a consequence, the respondents had a higher level of anxiety related to terror, compared to the ones that had not experienced an actual terrorist attack. Further, Ahern, Kiehl, Sole and Byers (2006) argue that resilience is the most common response after one has experienced an attack at the scale of a terrorist attack, which indicates that one is prepared to recover after a trauma. This can further lead to an automatic response to news content related to threatening information. Woods, Ten Eyck, Kaplowitz and Shlapentokh (2008) suggested that proximity influences the public’s negative emotions towards a terrorist attack. Therefore, if one has
experienced terrorism, been close to a terrorist attack, or had friends or family that were victims of it, terrorism-related news content will likely trigger anger and fear. The hypothesis will be as follows:
H3: Previous experience strengthens the effect between the presence of the word terrorism and emotions.
Figure 2.
Visual Frames and Emotions
The perceptions of the messages in the news are not only shaped by the textual information, but also by its visual representation (Schindler, Krämer and Müller, 2017); studies reveal that political and visual communication can be strongly connected (Barnhust and Steele, 1997). Research suggests that the media offer what is referred to as a diet of visual material chosen to attract their audience’s attention. In addition, visual material is proven to be memorable and can be especially forceful depending on its perceptual salience (Joffe, 2008). People are more affected by the visuals than many would first think, and the way the media decide to visually present their news can be an influential type of representation. Further, in the study of Machin 10
and Niblock (208) they argue that the size of the font is a form of salience, as it takes up great space proportionally on a news page. Big fonts can be seen as being a part of sensationalism, which means that it is a vivid and dramatic presentation of an event, which further gives them a forceful impact on the mind of the reader (Stephenson and Bromley, 1998). As a consequence, the size of the font, when presenting negative information, has demonstrated to affect how people perceive the information they receive, and further how they process it. There are, however, different views regarding if big fonts actually signify the right values, and if one should use font size to draw attention (Coulter and Coulter, 2005). Hence, the news media are constantly making the headlines big, as this is often what readers are dragged towards.
Last, it is argued that visuals are used to send people in emotive pathways (Joffe, 2208). The font size is known as a powerful tool in relation to strengthening the public’s reaction towards what they read. If the font is big, the message will become more threatening and important to the reader, compared to if it is small (Stephenson and Bromley, 1998). In the study of Schindler, Krämer and Müller (2017), they argue that there is both effective and important with visual cues to influence citizens to process political information, or in this study news. This is confirmed in the study of Bayer, Sommer and Schacht (2012), where their results reveal that large font size leads to an increase for early emotions effect, which can be an indication of why the news media often uses big fonts in their headlines. It is not only to draw attention to the text, but also to evoke certain types of emotions. It can therefore be argued that when the news media present a type of layout/font, it can have priming/framing effects on a presented message (Schindler, Krämer and Muller, 2017). Taking this into account, the hypotheses will therefore be:
H4: When the size of the font is big in news articles the subjects’ perceptions of the danger will be higher, compared to if it is small.
H5: Emotions explain the relationship between the font size in the news and the subjects’ perceptions of the danger.
Figure 3.
Interaction between Verbal and Visual Frames
By taking the theories presented in this paper into account, it is likely to believe that there is an interaction effect if one combines the independent variables. By applying frames to a crisis situation like a terrorist attack, it may serve as a strategy to identify main causes and responsible agents, make moral judgments, and, finally, to suggest policy responses to the event (Papacharissa and de Fatima Oliveira, 2008: 55).
In addition, as already discussed by Bayer, Sommer and Schacht (2012) big font sizes increase emotional effect, which is an indication of why big letters are
popular and successful in the news media. It is evident that the word terrorism and big font size separately may influence the public’s perceptions of the danger, as it can evoke strong negative emotions. The importance of the interaction of the verbal and visual frames is that if the font is large and includes negative and threatening information, it is more likely to evoke a negative reaction, compared to if it is just a big font with neutral information. With combining strong visuals with threatening information, the message will automatically be regarded as a larger happening. Therefore, it is likely that the combination of these has an effect as well. The last hypothesis will therefore be:
H6: There is an interaction between font size and use of the term terrorism in news articles on the subjects’ perception of the danger.
Figure 4.
13
Methods
Experimental design
In order to answer the research question, an online survey embedded experiment has been conducted. A 2x2 full factorial design with two between-subjects verbal frames (terrorist vs. attack) and two between-subjects visual frames (small font vs. big font) has been administered. The following variables were examined: the independent variables verbal frames and visual frames; dependent variable subjects’ perceptions of the danger; the moderator previous experience, and the mediator emotions.
The true experiment was carried out by an online questionnaire software
named Qualtrics. The respondents were randomly assigned to one of four conditions through a randomization procedure that Qualtrics offers. The manipulation involved participants being exposed to a news article within four conditions. These four conditions combined the independent variables verbal frames (attack vs. terrorist attack) and visual frames (small font vs. big font). The respondents received either a news article that included a small font - attack framing, big font - attack framing, small font - terrorist attack framing, or big font - terrorist attack framing.
Sampling
A convenience sample was conducted, where the respondents were recruited through social media platforms such as Facebook, Whatsapp and Linkedin, and through the word of mouth. The respondents were required to be above the age of 18, and needed to have enough English qualification to fill in an English survey questionnaire. This study intended to collect respondents from all over the world and people of all generations. To access the questionnaire the participants received an anonymous
online link, and the survey was estimated to take around 5 minutes to complete. The online questionnaire was accessible from the 14th of May 2019 till the 28th of May 2019, and had in total N=208 valid respondents. Prior to this, 50 respondents had to be removed as they did not finish the survey completely. In addition, it was observed that the respondents were not equally distributed between the conditions, which is due to the randomization algorithm in Qualtrics (see table 5).
Table 5. Diagram of the number of participants in each condition.
Most of the respondents were female (68,8 %), followed by male (31,3%). The age of the respondents ranged from 19 to 88, where the average age was M= 31,75 (SD=12,78). 32 different nationalities participated in the survey, which made the sample quite diverse. However, the majority of the respondents were from Northern Europe. Most of the respondents came from Norway (56,3 %), and other countries such as the Netherlands (7,2 %) and Sweden (5,8 %). The highest level of education most of the respondents had received was a Bachelor’s Degree (51,4 %), followed by Master’s Degree (36,1 %) and High School Diploma (7,7 %). In terms of their current
employment status, most respondents were employed full-time (50 %), followed by students (33,7 %) and employed part-time (6,3%).
Procedure
After receiving an anonymous link through an online media platform, respondents gained an access to the online questionnaire. First, they were provided with information about the topic and aim of the experiment. However, the information was not too detailed in order to avoid priming the respondents. Further, information about the consent of participating was presented, where the participants had to agree on being sufficiently informed and if they wanted to proceed to the experiment. They additionally received some demographic questions such as gender, age, nationality, level of education and current employment status.
Furthermore, the respondents were exposed to the stimuli, where they received one of the conditions determined in the experimental design. After receiving the stimuli the participants were asked to what degree they have felt angry, anxious or frightened when reading the news about an attack. Further, questions concerning risk judgements, information seeking and on how to stop future attacks followed. One of the last steps was the manipulation check, where the respondents were asked if they remembered the title of the BBC article they read and if they considered the headline as big or small. Last, in order to measure the moderator previous experience, the respondents were asked if they or a friend or/and a family member had been close to or witnessed a terrorist attack, or if they just had experienced it through the news media. Last, the participants were thanked for their participation, and finally, the
researcher revealed the aim of the study. They were also informed that the BBC article provided in their stimuli was not real, and that it was only used in this research.
Measures
Independent Variables and Manipulation
The independent variables and stimuli in this study were two verbal frames (terrorism vs. attack) and two visual frames (big font vs. small font). As the sample was international it was decided to use a manipulated article from the BBC as the stimulus for each condition. This was to make sure that the respondents had to some extent knowledge about the newspaper, and because the BBC is additionally known to be a reliable newspaper (Jackson, 2015).
Dependent Variable
The dependent variable was subjects’ perceptions of the danger, adapted from the study of Woods (2011). However, by subjects’ perceptions of the danger, this paper intends to measure it through three dependent variables. These are if the frames in the media influence their attribution of behaviour (information seeking), increase their concern about the danger, and influence their assumptions on how to stop future attacks, which is similar to the variables in Woods (2011) study. In addition, this study includes the variable attribution of behaviour (information seeking) from the study of Boyle et al. (2004).
First, when the media apply frames to a crisis situation such as a terrorist attack, it can serve as a strategy to make moral judgements and to suggest policy responses to the event that occurred (Winter, 2008). When reading about an attack it
influences the public’s assumptions and thoughts on how to stop future attacks. This can be by agreeing to policies such as having stricter national border controls, make the police more present in the streets and that the authority should increase the surveillance to stop future attacks. These were therefore measured on a Seven-point Likert scale, where the respondents could agree on statements regarding these policies on a scale from (1) strongly agree to (7) strongly disagree. A factor analysis was conducted, and the results revealed that all the items loaded on one component, where the factors explain 73,15% of the variance in the original variables. The reliability scale was good, Cronbach’s alpha = .82(M=4,69, SD=1,65).
Second, the uncertainty reduction theory argues that individuals are motivated to seek more information and learn more about terrorism to reduce the uncertainty about what happened (Boyle et al, 2005). To measure the information seeking variable the respondents were asked if they search for more information about the attack when reading about it in the news and if they further search for information about attacks in general. This was measured on a Five-point Likert scale from (1) always to (5) never. In addition, it measured if they needed to do their own research to feel well informed on a Seven-point Likert scale from (1) strongly agree to (7) strongly disagree. A factor analysis was conducted, which revealed that all the items loaded on one component, where they together explain 72,15% of the variance in the original variables. The reliability scale was acceptable, Cronbach’s alpha = .75(M=3,10, SD=1,29).
Third, when reading about attacks it can influence the attributes of the subjects’ risk judgements. By searching for more information one increases the knowledge about what happened and receive an overview of the threat. By becoming
more informed it may influence citizens to be more worried. This will further make them become more concerned that an attack will happen in their country in the near future, feel concerned about themselves or a friend/family member being a victim of future attacks. In addition, it was asked if they have experienced an attack that shaken their own sense of personal safety and security. The measures were influenced by the variables from the study of Huddy, Feldman, Taber and Lahav (2005). This was measured on a Five-point Likert scale from (1) Very often to (5) never. A factor analysis was conducted, which further revealed that all the items loaded on one component, where they together explain 66,91% of the variance in the original variables. The reliability scale was acceptable, Cronbach’s alpha = .75(M=3,47, SD=1,05).
Moderator and Mediator
The moderator in this study is previous experience of terrorism. As argued in the study of Woods, Ten Eyck, Kaplowitz and Shlapentokh (2008) do proximity influence the public’s negative emotions towards a terrorist attack. At the end of the questionnaire in this experiment, the respondents were asked if they themselves, and/or a friend or family member had been a direct victim, been in close vicinity or been in the same city as a terrorist attack, or if they just read about it in the news. This was coded into No experience (0), Self (1), Family and/or friends (2), and self and family and/or friends (3). The descriptives revealed that 42,8 % did not have any experience at all, 23,1% had self experienced or been close to a terrorist attack, 29,3% had family/friends that had experienced a terrorist attack, and 4,8% had both themselves and had family/friends who experienced a terrorist attack. In addition, the
mediator in this study is emotions, where it measured negative related emotions such as anger, anxiety and fear. These were influenced by the study of Vasilopoulos, Marcus, Valentino and Foucault (2018) and Albertson and Gadarian (2014). The respondents were asked how often, if ever, they have felt angry, anxious and frightened when reading about attacks in the news on a Five-point Likert scale measured from (1) very strongly to (5) not at all. A factor analysis was conducted, which determined that all the items loaded on one component, where together they explained 72,36% of the variance in the original variables. The reliability scale was good, Cronbach’s alpha = .81(M=3,04, SD=1,14).
Results
Randomization Check
In order to reveal if it was a difference in the four conditions (terror vs. attack and small font vs. big font), it was important to execute a randomization check. This was tested with chi-square analyses and ANOVA on demographics such as gender, age, level of education, employment status and on the moderator previous experience. First, the chi-square analysis demonstrated that gender was equally distributed between the conditions (p=. 524). An analysis of variance was executed on age, which revealed that it was not significant and was therefore equally distributed between the conditions ( p =.597). An ANOVA demonstrated that the participants were distributed evenly in terms of the level of education between the four conditions (p=.451).
Furthermore, a chi-square analysis test on current employment status, revealed that there was not a significant difference between the participants in the four conditions (p= .648). Last, the chi-square tested on the previous experience revealed
that it was not significant (p=.740), which indicates that the moderator was also equally distributed between the four conditions. This demonstrates that the randomization worked for all the demographics and the moderator in this experiment, and it was consequently not necessary to control for any of the variables.
Manipulation Check
A manipulation check was carried out to explore if the stimuli worked in the manner they were supposed to. There were two questions, one asked if the name of the title was “Deadly terrorist attack in London” or “Deadly attack in London” and the second asked if the respondents considered the headline of the article as big or small. In order to determine if the stimuli worked, a chi-square test to reveal if the respondents remembered the manipulation they were presented. The chi-square revealed that 141 participants answered one or both of the questions wrong.
However, the results demonstrated that out of the respondents who were exposed to the small font stimuli, 40 considered the headline to be big, and of the respondents who received the big font article, 14 of them consider the headline to be small. The results revealed that there was a significant moderate association Φ=.49, p=.000. In addition, out of the respondents exposed to the attack frame, 12 believe that they read about a terrorist attack, and of the respondents exposed to the terror frame, 32 believed that it was an attack. The cross-tabulation revealed that there was a significant strong association Φ= .59, p= .000, as the majority of the respondents answered the questions correct for the specific manipulation. As a consequence, it can be argued that the manipulation was successful as the relationships were significant for the conditions.
Hypothesis 1
The first hypothesis suggested that individuals exposed to the word terrorism in news articles have higher perceptions of the danger compared to individuals exposed to news articles without the word terrorism. A one-way ANOVA was conducted to test the hypothesis.
The results revealed that the assumptions for the one-way ANOVA have been met, when testing the verbal frames on seeking for more information. The equal variance in the population has been met, Levene's (3, 204)=.630, p=.596. Additionally, there was not a significant difference between the verbal frames if the respondents would seek more information, F(1, 104) = .01, p = .923. Further, the one-way ANOVA tested on the influence of the verbal frames on the respondents assumptions on how to stop future attacks, revealed that the equal variance in the population has been met, Levene's (3, 204) = .195, p=.900. There was, however, not a significant difference between the verbal frames on the assumptions on how to stop future attacks, F(1, 41)= 2.30, p= .131. Last, when executing a one-way ANOVA to test the influence of the verbal frames on risk judgments, it revealed that the equal variance in the population was met, Levene’s (3, 204)=1.20, p=.313. Hence, it was not a significant difference between the verbal frames on the respondents risk judgements, F(1, 1)= .132, p= .716. In conclusion, H1 could not be confirmed (See table 11 for descriptives).
Hypothesis 2
The second hypothesis claimed that when people read about terrorism in the news negative emotions will more likely arouse, compared to if they read about an attack. The emotions will further lead to higher perceptions of the danger. Model 4 in the PROCESS macro by Hayes (2013) was executed to test the mediation effect. This was performed to explore if emotions had a mediated effect on the relationship between the verbal frames and the subjects’ perceptions of the danger. A bootstrapping procedure with a sample of 5000 to analyse the indirect effect of 95% bias-corrected bootstrap confidence interval. This analysis tested the verbal frames as the independent variable, emotions as the mediator, and information seeking, risk judgements and assumptions on how to stop future attacks as dependent variables. The PROCESS test was therefore executed three times to test each of the dependent variables.
As expected after the results revealed in the previous analysis, the verbal frames did not have a significant effect on risk judgements, also called path c’(b=-.13, p=.895), and did additionally not have an effect on emotions, path a(b=-.26, p=.522). Although not hypothesised, emotions had a significant effect on risk judgements, path b(b=.48, t(205)=7,83, p=.000). In addition, the verbal frames did not have a significant indirect effect on risk judgements, as the bias-corrected bootstrap confidence interval for the indirect effect included a zero (Indirect effect=-.12, 95% BCBI[-0.55 0.25]). Second, the verbal frames did not have a direct effect on information seeking, path c’ (b=-.00, p=.991). In addition, the results for a and c’-path did not reveal a significant mediation effect on information seeking and assumptions on how to stop future attacks. However, both analyses demonstrated that the b-path was significant (see figure 7 and 8). Further, the verbal frames did not have a
significant indirect effect on information seeking(Indirect effect=-.06, 95% BCBI[-0.28 0.13]) and assumptions on how to stop future attacks(Indirect effect=-.13, 95% BCBI[-1,90 0.29), because the bias-corrected bootstrap confidence interval for the indirect effect included a zero. As a consequence, the second hypothesis could not be confirmed.
Figure 6. Path model: Mediation effect on the relationship between the verbal frames and risk judgements.
Figure 7. Path model: Mediation effect on the relationship between the verbal frames and information seeking.
Figure 8. Path model: Mediation effect on the relationship between the verbal frames and assumptions on how to stop future attacks.
Hypothesis 3
The third hypothesis suggested that if friends and/or family of the respondents, or they themselves had experienced or been close to a terrorist attack, it will strengthen the effect between emotions and the word terrorism. In order to test this moderated mediation effect, model 7 of the SPSS macro PROCESS by Hayes (2013) was run. This was executed to test if previous experience of terror could moderate the effect between the verbal frames and emotions. PROCESS conducts an OLS regression-based path analysis, which estimates both the direct and indirect effects (Hayes, 2013) A sample of 5000 with a bootstrapping procedure was determined to analyse the indirect effect with a 95% confidence interval. As the moderator was a nominal variable, a dummy variable created, which was coded into No (0) and self, friends/family and both (1) in order to execute the analysis.
The results of the analysis revealed that there was an interaction effect
between the verbal frames and previous experience on emotions, as the results revealed are significant (b= 1,94, p= .018). In addition, the moderated mediation was
statistically significant because the bootstrap confidence interval did not include a zero (index of moderated mediation= .10, boot SE= .49, 95% BCBI [0.17 2.06]). In conclusion, H3 could be confirmed, previous experience of a terrorist attack moderates the relationship between verbal frames and emotions.
Hypothesis 4
The fourth hypothesis argued that if the size of the font is big the subjects’ will have higher perceptions of the danger, compared to if the font size of the article is small. As the subjects’ perceptions of the danger contains three variables, it was conducted a one-way ANOVA to explore if there is an effect of the visual frames on information seeking, risk judgements and assumptions on how to stop future attacks.
The results revealed that the assumptions for the one-way ANOVA have been met, when testing the visual frames on seeking for more information. The equal variance in the population has been met, Levene’s F(3, 204) = .63, p= .596. Hence, it was not a significant difference between the visual frames if the respondents would seek more information, F(1,19) = 1,71, p= .193. In addition, when conducting an one-way ANOVA analysis on the effect of the visual frames on respondents assumptions on how to stop future attack, it revealed that the equal variance in the population was met, Levene’s (3,204) = .20, p= .900. It was, however, not a significant difference between the visual frames on the assumptions on how to stop future attacks, F(1, 4) = .258, p= .612. Last, the one-way ANOVA tested on the influence of the visual frames on respondents risk judgements, revealed that the equal variance in the population was met, Levene’s (3,204) = 1,20, p= .313. Hence, it was
not a significant difference between the visual frames on the respondents risk judgements, F(1, 13)= 1,47, p= .227. In conclusion, hypothesis 4 could not be confirmed (see table 11 for descriptives).
Hypothesis 5
The fifth hypothesis expected that when subjects’ are exposed to articles with big fonts in the news, the negative emotions will be stronger, compared to if they are exposed to articles with small fonts. The negative emotions will further lead to higher perceptions of the danger. To test the mediation effect, model 4 in PROCESS macro by Hayes (2013) was executed. This analysis was run to explore if emotions had a mediated effect on the relationship between the visual frames and the subjects’ perceptions of the danger. This test included a bootstrapping procedure with a sample of 5000 to analyse the indirect effect of 95% bias-corrected bootstrap confidence interval. This analysis tested the visual frames as the independent variable, emotions as the mediator, and information seeking, risk judgements and assumptions on how to stop future attacks as the dependent variables.
As expected after the results determined in the previous analysis, the visual frames did not have a direct effect on risk judgements, path c’ (b= .12, p= .335), and did additionally not have an effect on emotions, path a ( b= -.31, p=442). Although not hypothesised, emotions did have a significant effect on risk judgements, path b (b=.16, p=.000). In addition, the visual frames did not have a significant indirect effect on risk judgements, as the bias-corrected bootstrap confidence interval for the
indirect effect did include a zero (Indirect effect= .15, 95% BCBI [-0.23 0.54]). In addition, the results for a and c’-path did not reveal a significant direct effect on information seeking and assumptions on how to stop future attacks. However, both analyses demonstrated that the b-path was significant (see figure 10 and 11). Further, the visual frames did not have a significant indirect effect on information seeking (Indirect effect = .07, 95% BCBI [-0.12 0.27]) and assumptions on how to stop future attacks (Indirect effect = .160, 95% BCBI [-0.24 0.65]), because the bias-corrected bootstrap confidence interval for the indirect effect included a zero. In conclusion, the fifth hypothesis could not be confirmed.
Figure 9. Path model: Mediation effect on the relationship between the visual frames and risk judgements.
Figure 10. Path model: Mediation effect on the relationship between the visual frames and information seeking.
Figure 11. Path model: Mediation effect on the relationship between the visual frames and assumptions on how to stop future attacks.
Hypothesis 6
The last hypothesis claimed that there is an interaction effect between visual frames and verbal frames on the subjects’ perceptions of the danger. An analysis of variance was conducted to explore if the hypothesis could be confirmed. Equally to the other hypotheses tested, the independent variables were tested on information seeking, risk judgements and assumptions on how to stop future attacks. First, the results revealed that it was not a statistically significant interaction between the effects of the visual frames and the verbal frames on information seeking, F(1,8) =.159, p= .691. Further, the second analysis revealed that it was not a statistically significant interaction between the effects of the visual frames and the verbal frames on risk judgements, F(1, 0)=.031, p=.861. Last, the analysis of variance additionally revealed that there was not a significant interaction effect between the visual frames and the verbal frames on assumptions on how to stop future attacks F(1, 25) = 1,43, p=.233. In conclusion, the sixth and last hypothesis could also not be confirmed.
Table 11. Mean values of Dependent Variables by Condition.
Discussion
In a media landscape with an increased focus on negativity in the news, it has become interesting to research the effect of negative and threatening frames on the media audiences (Soroka and McAdams, 2015). This study aimed to examine the impact of the visual and verbal frames in the media on the subjects’ perceptions of the danger. The research included the following dependent variables: - risk judgements, - assumptions on how to stop future attacks, and - information seeking. It also used negative emotions as the mediator variable and previous experience of a terrorist attack as the moderator variable. The results revealed that there was a significant moderated mediation effect, where previous experience strengthened the effect between the presence of the word terrorism and emotions. However, the other hypotheses tested did not reveal any statistically significant results.
Verbal Frames, Emotions and Previous Experience
The results in the first hypothesis revealed that people exposed to the attack frame had higher perceptions of the danger, compared to the subjects’ exposed to the terror frame. This contradicts what the first hypothesis suggested and what a large amount of previous literature suggest. By labelling a violent act as terrorism rather than a criminal act, or in this study attack, it primes the subjects’ perceptions of the danger (Cooper, 2011). Hence, as this study revealed, the attack frame lead slightly to higher perceptions of the danger, compared to the terrorist attack frame. This may be due to an attack can be regarded to be more uncertain than a terrorist attack, as terrorists often have a certain goal with their violence. This is interesting in relation to the risk perception research, where it is suggested that danger which is perceived as new, evokes greater levels perceived threat, compared to old and familiar hazards (Fischhoff et al. 1978, Covello 1992, Slovic 2004, Sjöberg 2005). When an attack is presented in the news it may indicate that it is a new and more uncertain type of danger, as it is not confirmed what the motivation for the violence was. This may influence the violent act to be regarded as more threatening than a terrorist attack. Consequently, terrorism has had an increased media coverage within the last decade (Powell, 2011), and therefore may not be regarded as that uncertain and threatening for today’s news audience.
This correspond with the study of Woods (2011), where his results revealed that by manipulating a term by labelling it as terrorism, did not have an effect on people’s worry. He suggested that the reason for this could possibly have been because the respondents conceptualize it as terrorism, even though it was not.
However, in this research paper more people who were exposed to the terror stimuli labeled the article they read as an attack, rather than the other direction. For future research, it may be interesting to change the manipulation into a criminal act, rather than an attack. Or narrow the focus by studying the difference between how the media label radical Islamic terrorism compared to extreme right terrorism and the effects of it, as this is something that has become evident in the media the last years.
In addition, when including emotions, it revealed that despite that the fact that the direct and indirect effects did not have a significant influence on the subjects’ perceptions of the danger, respondents with negative emotions may have higher risk judgements, information seeking and assumptions on how to stop future attacks. However, as the results of this analysis do not reveal causality, it can also indicate that higher danger perceptions lead to more negative emotions. This indicates that negative emotions and people’s danger perceptions may have an effect on each other, but that the framing of news events cannot be significantly demonstrated to evoke negative emotions and influence the subjects’ threat perceptions. It is, however, important for future research to continue to study the relationship between news content on negative emotions, especially with the negativity bias that is evident within the news content today (Soroka and Mcadams, 2015).
Further, previous experience did determine a significant moderated mediation effect, and therefore moderate the relationship between the verbal frames and the negative emotions. It is an interesting finding, as this has not been sufficiently studied in previous literature. However, it was revealed in the study of Slone, Shoshani and Baumgarten-Katz (2008) that individuals with a high exposure to political violence showed much greater vulnerability to exposure of media content that included life
threatening events. Therefore, this is confirmed in this study, as previous experience of a terrorist attack should be a factor that should be taken into further consideration when researching the effect of terrorist-related news content. As previous experience strenghtens the effect of verbal frames on negative emotions, further research should dive into which kind of emotions arise from it, and look more specifically at how this can further affect the respondents reactions towards terrorism-related news content. Lastly, psychological trauma is an interesting moderator for future research on the influence of threatening information on the audience of the news media.
Visual Frames and Emotions
Regarding the visual frames, the results revealed that the respondents had higher perceptions of the danger when the font of the article was big, compared to when it was small. This correlates with what previous research has revealed, that big fonts can be argued to be a part of sensationalism, which indicates that it is both a vivid and dramatic presentation of an event. That this gives a forceful impact on the mind of the media audience (Stephenson and Bromley, 1998). However, a great amount of the respondents failed the manipulation check, which may be the reason why the results were not significant for the direct effect. Hence, it is interesting if the respondents still were unconsciously manipulated by the stimuli, despite thinking that they were exposed to something else than what their manipulation intended to. When taking the manipulation check into account, the threatening information may have made the participants in the small font condition think that the font was big, as the information they read seemed both vivid and dramatic.
Further, it is interesting for future research to study the effect of the font, as it has become more evident that the media uses this as a visual tool to draw attention. However, there is a lot of research on visual images and audio, but there is a lack of studies that focuses on the power of the font size and visual appearance of text. The media have a huge influence on the public’s perceptions of the world, and therefore it is essential to study the tools the news media apply to influence their readers.
It is also interesting to relate the visual frames with emotions, as studies suggest that there is a strong link between visual information processing and the emotions of individuals. Research demonstrate that emotions likely play a role in individuals’ level of concern about terrorism (Fahmy, Cho, Wanta and Song, 2006). This can be related to the results of this study, where emotions had a significant role in the subjects’ perceptions of the danger. However, there were no significant direct effect or indirect effect of visual frames on emotions and risk judgements, information seeking and the assumptions on how to stop future attacks. It may still be interesting for future studies to take this further, as negative news has become more evident in the media today.
Last, the results were the same for the interaction effect, that when one is combining the visual and verbal frames the results did not reveal any statistically significant results. However, if future studies can demonstrate a main effect for each of the independent variables, it is likely that the combination of the variables will increase the subjects’ perceptions of the danger. A screaming threatening message can be argued as having a greater influence, than just a screaming neutral message or a threatening message alone.
Limitations
Some limitations have to be taken into account regarding this study. First, five out of six hypotheses could not be confirmed. This may be due to the participants not pay sufficient attention or remember what they saw in the manipulation or that the questions asked were not clearly formulated. For future studies, it is important to be cautious and clear when posing questions in an online survey.
Further, this study included one type of media, which was the British news outlet the BBC. This can be regarded as a limitation because most of the respondents were from Norway or other Northern European countries, which may have influenced the respondents’ reaction towards the content they read. Moreover, as it was an online questionnaire and the news articles were manipulated by the researcher, some of the participants may have recognised that the articles were not real. In addition, because this study had an international sample, the stimuli and the questionnaire were in English, which may have had an impact on the respondents’ answers, as English is not the first language of most of the participants. This was, however, because this paper executed a convenience sample, and can therefore not be argued to represent the general public. Therefore, the results have to be considered with caution as the respondents are not in their natural habitat. For future studies to focus on one nationality, instead of having an international sample, or if possible, do a field experiment
Conclusion
The media have great power in being the main provider of distributing information to the public in today’s society. It is essential for the future of democracy that audiences become more cautious and aware as to how the media frame their news content, since as this study showed, framing can impact their reaction and opinion regarding news information. This study sought to contribute to the research on media framing and negativity bias, and how it can influence the subjects’ perceptions of the danger presented. The findings revealed that previous experience can strengthen the relationship between the verbal frames of the media and the negative emotions. However, the majority of hypotheses tested did not result in any significant findings, this is still something that should be studied further, as verbal frames such as terrorism and visual frames such as font size are very much present in the news media today. This study demonstrated that to some extent the presentation of news content in the media can have an effect on the news audience. Additionally, previous experience of a terrorist attack moderated the relationship between the verbal frames of the media and emotions. This leads us to conclude that the way in which the media decides to frame their content, both visually and verbally, can prompt different cognitive processes, and as a result shape the public’s perceptions on what they read.
36
Reference List:
Bayer, M., Sommer, W., & Schacht, A. (2012). Font size matters—emotion and attention in cortical responses to written words. PloS one, 7(5), e36042.
Bradley, S.D., Angelini, J.R., Lee, S. (2007) Psychophysiological and Memory Effects of Negative Political ADS: Aversive, Arousing, and Well Remembered. Journal of Advertising, 36 (4), 115-127.
Boukes, M., and Vliegenhart, R. (2017) News consumption and its unpleasant side effect. Journal of Media Psychology, 29, 137-137.
Boyle, M. P., Schmierbach, M., Armstrong, C. L., McLeod, D. M., Shah, D. V., & Pan, Z. (2004). Information seeking and emotional reactions to the September 11 terrorist attacks. Journalism & Mass Communication Quarterly, 81(1), 155-167.
Brinson, M.E., and Stohl, M. (2012) Media Framing of Terrorism: Implications for Public Opinion, Civil Liberties, and Counterterrorism Policies. Journal of international and intercultural communication, 5(4), 270-290.
Cooper, H.A., 2001. Terrorism: the problem of definition revisited. American Behavioral Scientist, 44 (6), 881–893.
Dinesen, P.T., and Jæger, M.M. (2013) The Effect of Terror on Institutional Trust: New Evidence from the 3/11 Madrid Terrorist Attack. Political Psychology, 34(6), 917-926.
Entman, R. M. (2004). Projections of power: Framing news, public opinion, and US foreign policy. University of Chicago Press.
Woods, J. (2007). What we talk about when we talk about terrorism: Elite press coverage of terrorism risk from 1997 to 2005. Harvard International Journal of Press/Politics, 12(3), 3-20.
Gadarian, S.K., & Albertson, B. (2014). Anxiety, immigration, and the search for information. Political Psychology, 35 (2), 133-164.
Coleman, R. (2010). Framing the pictures in our heads. Doing news framing analysis:
Empirical and theoretical perspectives, 233-261.
Gadarian, S. K. (2010). The politics of threat: How terrorism news shapes foreign policy attitudes. The Journal of Politics, 72(2), 469-483.
Nacos, B. L., Bloch-Elkon, Y., & Shapiro, R. Y. (2011). Selling fear:
Counterterrorism, the media, and public opinion. University of Chicago Press.
Hayes, A. F., & Preacher, K. J. (2013). Conditional process modeling: Using structural equation modeling to examine contingent causal processes.
Chyi, H. I. and McCombs, M. (2004) Media Salience and the Process of Framing: coverage of the Columbine School Shooting. J&MC Quarterly, 81(1), 22-35.
De Vreese, C. H. (2005). News framing: Theory and typology. Information design
journal & document design, 13(1).
Hall, C. M. (2002) Travel Safety, Terrorism and the Media: The significance of the issue-attention cycle. Current Issues in Tourism, 5(5), 458-466.
Huddy, L., Feldman, S., Taber, C., & Lahav, G. (2005). Threat, anxiety, and support of antiterrorism policies. American Journal of Political Science, 49, 593-608
Lerner, J.S:, Gonzales, R.M., Small, D.A., and Fischhoff, B. (2003) Effects of Fear and Anger on Perceived Risks of Terrorism: A National Field Experiment. Psychological Science, 14(2), 144-150.
McQuail, D. (1987). Mass communication theory: An introduction. Sage Publications, Inc.
Nellis and Savage (2012) Does watching the News affect Fear of Terrorism? The Importance of Media Exposure on Terrorism Fear. Crime and Delinquency, 58(5), 748-7768.
Papacharissi, Z., & de Fatima Oliveira, M. (2008). News frames terrorism: A comparative analysis of frames employed in terrorism coverage in US and UK newspapers. The international journal of press/politics, 13(1), 52-74.
Schacht A, Sommer W (2009) Time course and task dependence of emotion effects in word processing. Cogn Affect Behav Neurosci 9, 28–43.
Scheufele, D. A., & Tewksbury, D. (2006). Framing, agenda setting, and priming: The evolution of three media effects models. Journal of communication, 57(1), 9-20.
Schindler, J., Krämer, B., and Müller, P. (2017) Looking left or looking right? Effects of newspaper layout style on the perception of political news. European Journal of Communication, 32(4), 348-366.
Slone, M. (2000) Response to Media Coverage of Terrorism. Journal of Conflict Resolution, 44(4), 508-522.
Soroka, S., & McAdams, S. (2015). News, politics, and negativity. Political Communication, 32 (1), 1-22.
Soroka, S., Young, L., and Balmas. (2015) Bad News or Mad News? Sentiment Scoring of Negativity, Fear and Anger in News Content. AAPSS, 659(1), 108-121.
Shoshanim, A. & Slone, M. (2008) The Drama of Media Coverage of Terrorism: Emotional and Attitudinal Impact on the Audience, Studies in Conflict & Terrorism, 31:7, 627-640
Van Dijk, T. A. (1995). Power and the news media. Political communication and
action, 6(1), 9-36.
Vasilopoulos, P., Marcus, G.E., adn Valentino, N.A. (2016) Fear, anger, and voting for the far right: Evidence from the November 13, 2015 Paris terror attacks. Political Psychology, 0(0), 1-24.
Vuilleumier P, Richardson MP, Armony JL, Driver J, Dolan RJ (2004) Distant influences of amygdala lesion on visual cortical activation during emotional face processing. Nat Neurosci 7, 1271–1278
Gabriel Weimann, “The Theater of Terror: The Effects of Press Coverage.” Journal of Communication, 33 (Winter 1983), pp. 38–45.
Woods, J. (2011) Framing terror: an experimental framing effects study of the perceived threat of terrorism. Critical Studies of Terrorism, 4(2), 199-217.
Woods, J., Eyck, T. A. T., Kaplowitz, S. A., & Shlapentokh, V. (2008). Terrorism risk perceptions and proximity to primary terrorist targets: how close is too close?. Human
Ecology Review, 63-70.
Joffe, H. (2008) The Power of Visual Material: Persuasion, Emotion and Identification. International Council for Philosophy and Human Studies. 55(84), 84-93.
Norris, P., Ken, M., and Just, M. (2003) Framing Terrorism: The News Media, the Government, and the Public. Routledge. 1-315.
Bromley, M. (1998). The “tabloiding” of Britain:“quality” newspapers in the 1990s.
Sex, Lies and Democracy: The Press and the Public, 25-38.
Coleman, R. (2010). Framing the pictures in our heads. Doing news framing analysis:
Empirical and theoretical perspectives, 233-261.
Rodriguez, L., & Dimitrova, D. V. (2011). The levels of visual framing. Journal of Visual Literacy, 30(1), 48-6.
Cho, J. (2008). Political ads and citizen communication. Communication Research,
35(4), 423-451.
Coulthard, M. (2014). An introduction to discourse analysis. Routledge.
Jackson, J. (2015, December 16). BBC rated most accurate and reliable TV news,
says Ofcom poll. The Guardian. Retrieved from
https://www.theguardian.com/media/2015/dec/16/bbc-rated-most-accurate-and-reliabl e-tv-news-says-ofcom-poll
Coulter, K. S., & Coulter, R. A. (2005). Size does matter: The effects of magnitude representation congruency on price perceptions and purchase likelihood. Journal of Consumer Psychology, 15(1), 64-76.
Appendix 1
Survey Stimulius 41Manipulation 1 - Small Font - Terrorist Attack
Manipulation 2 - Small Font - Attack