Powerful Politicians – Powerful Pictures?
The Visualization of Political Power across Time and Media Systems
Master’s Thesis
Lydia Weber
University of Amsterdam Graduate School of Communication
Erasmus Mundus Master’s Journalism, Media and Globalization Supervised by Tom Powell
Abstract
The study examines how political power is constructed by press photography in news media. More specifically, this is done by studying published pictures of Angela Merkel in four European countries and four points of time. Drawing upon visual communication, psychology and photography literature, the analysis is based on a theoretical framework consisting
production techniques, nonverbal behavior and structural environment. By analyzing 400 images published in Germany, France, Denmark and Portugal during the German election years 2009, 2013, 2017 and the refugee crisis 2015, the study finds that there is no universal depiction of political power. Angela Merkel is portrayed differently across countries and time. Although time has an effect on the depiction of the German chancellor, the study shows that the visualization of political power is mainly affected when comparing the depiction among the four countries. The results add to the understanding of visual political communication and show that images are an important tool to highlight or degrade political power.
Keywords: political power, visual communication, Angela Merkel, media systems, events, press photography
Powerful Politicians – Powerful Pictures? The Visualization of Political Power across Time and Media Systems
Introduction
Angela Merkel is one of the most powerful politicians in the world. She was reelected among the most influential persons by Forbes Magazine and is repeatedly ranked among the most powerful women (Forbes, 2018). Her approval rates in Germany show that are her popularity seems to be unbroken although she governs the country for over a decade now (Zeit Online, 2017; Nardelli, 2015b). Although during that time she faced substantial crises like the Euro crisis or the refugee crisis, she not only managed to stay in power but she even sticks out as female world leader in a male dominated world (Davidson Sorkin, 2017).
Merkel’s power is reflected by her speeches and actions but also by her representation in the media. However, when looking at pictures of the German chancellor she does not seem to be one of the most important politicians in the world. She lacks mimic and compared to politicians like Barack Obama, Merkel would always look stiff and unnatural. “In the past she has faced much ridicule for her looks and presentation”, the Guardian writes (Connolly, 2013). However, it might be worth to risk a closer look. This study wants to research how political power is visualized across time and in different media systems exemplified by the case of Angela Merkel.
Literature Review and Justification
When reviewing visual communication literature, it becomes clear that most studies on visualization of political power deal with the influence of politician’s depiction on voting behavior (e.g., Marcinkowski, Lünich., & Starke, 2017; Verser & Wicks, 2006; Hall, Goren, Chaiken, & Todrov, 2009), their appearance during televised material of election candidates or debates (e.g., Nagel, Maurer, & Reinemann, 2012; Bucy & Grabe, 2008; Verser, 2007), or on election posters (e.g., Vigso, 2017). These scholars suggest that pictures are a strong tool
to ultimately affect voter behavior. Besides, visual bias and visual framing have been widely discussed in this context and most studies show visualization patterns in news media (e.g., Frosh, 2011; Barrett & Barrington, 2005).
Only a hand full of studies focuses on visual communication across different countries (Krogstad & Storvik, 2012). Since foreign audiences don’t have an immediate effect on election outcomes or a politicians’ rating this field is often overlooked but it can be argued that the depiction in foreign newspapers is as crucial for politicians as the depiction in
national media. Most people get their information about international politics from the media because they can’t experience the news event first hand. Thus, the image of foreign politicians is often influenced by their media appearance. The depiction in foreign media matters to politicians because it might be able to change the country’s overall perception abroad
(George, 2016). This is especially true for the head of state or prime minister since they stand for the country’s image. That idea is closely connected to two concepts of the International Relations literature: soft power and nation branding (George, 2016). The paper argues that perceived power abroad can possibly lead to more support on a bilateral level. Thus, it is crucial to compare the visualization of political power across different countries. Building up on the classification of Western media systems by Hallin and Mancini (2004) and the recent adaptation by Brüggemann, Engesser, Büchel, Humprecht and Castro (2014) pictures are sampled in Germany, France, Portugal and Denmark.
Another important aspect is the depiction across different time frames. Empirical research in this direction mostly focuses on visualization of specific events in a broader sense, for instance the work of Fahmy (2005) that deals with cases of war and conflict photography, and compares visuals dealing with 9/11. This research among others (e.g. O’Brien, 1993; Rössler, Bomhoff, Haschke, Kersten, & Müller, 2011), suggests that visuals differ
of politicians across various election cycles (e.g. Bucy & Grabe, 2008). However, this study’s focus is on the comparison between the 2009, 2013 and 2017 German national elections as well as the European refugee crisis 2015. During election years, politicians and campaign teams have more possibilities to influence which photos are published (Schill, 2012). It is not clear on the other hand, how the visualization of political power is influenced during non-election events.
I will follow the method of quantitative image analysis, based on the discussion by Geise and Rössler (2013). The authors argue that pictures can be analyzed in a systematic and quantitative way if a couple of considerations are taken into account (2013). In order to identify patterns and structures that are inherent of the analyzed images an image analytics tool is introduced. The authors argue that the picture has to be split in three different layers to make sure that manifest and latent interpretations are recorded: Oberflächenstruktur (will be referred as macro level), Binnenstruktur (will be referred as meso level) and Tiefenstruktur (micro level) (Geise & Rössler, 2013). These levels form the overall framework for my analysis of visualization of power.
Traditional power theories access the notion of political power in a broad sense (e.g. Weber, 1973). More recent studies discuss the implications of related concepts such as status, authority and dominance in relation to each other (e.g. Bailey & Kelly, 2015). Scholars agree that apart from fine nuances between the concepts all terms refer to a hierarchical relationship (Bailey & Kelly, 2015). Since the fine nuances between each concept cannot be distinguished during the analysis of visuals, my study will follow the same line when referring to political power.
This study compares 400 photos of Angela Merkel published in four countries – Germany, France, Portugal and Denmark, and during four different years – the election
periods 2009, 2013 and 2017, and the refugee crisis 2015 to answer the main research question:
To what extent do published photos of Angela Merkel differ across media systems and for different events in their visualization of political power?
Visualization of political power
After reviewing a large body of literature, it became clear that the visualization of political power manifests in various aspects of a picture. These include nonverbal behavior cues of depicted people, production technique and structural environment. Keeping in mind the three picture levels, as highlighted by Geise and Rössler (2013), each of the pictures’ traits can be grouped under the macro, meso and micro layer of the photograph.
Macro level: Production techniques
This group of variables focus on the way the picture was taken and photo techniques. According to Geise and Rössler (2013) the macro level of the picture includes formal
characteristics of the picture arrangement, like camera angle and shot. Especially, these two aspects play an important role in the depiction of power (Verser, 2007).
Camera angle. Several scholars found that persons depicted from a low angle appear
to be taller and more powerful (Giessner, Ryan, Schubert, & van Quaquebeke, 2011; Verser & Wicks, 2006; Mandell & Shaw, 1973). Adding to that, Giessner and Schubert (2007) find that tall people are perceived to have a higher status, salaries and leadership competences. However, it has to be stressed that the angle of the picture should not be too extreme either, since “extreme angles tend to produce negative evaluations” (Graber 1996 as cited in Schill, 2012, p. 128). In general, a low camera angle reinforces visualization of power whereas a high camera angle has the opposite effect.
Camera shot. Scholars share a common ground on the effects of the camera shot on
the perception of power. Longer shots often make people appear overwhelmed and
diminished by the background and their surroundings. Thus, they are used to make people appear less powerful whereas medium shots are widely used for a neutral depiction of people (Mamer, 2000). Close camera shots on the other hand, make the person appear more likeable, friendlier, approachable and dominant, since others are able to identify themselves easier with the depicted person (Schill, 2012; Zuckerman, 1986; Calogero & Mullen, 2008, Mamer, 2000). In conclusion, a long camera shot weakens the perception of political power, while close-ups and portraits make the person appear approachable.
Meso level: Nonverbal behavior cues
Studies found that voters’ perception of a politician’s personality is highly shaped by nonverbal communication (Coleman & Banning, 2006) and research shows that “complex social judgements can be influenced by quick inferences made from faces” (Hall, Goren, Chaiken & Todorov, 2009, p. 84). These implications are especially important for notions of leadership and power, as several studies show (e.g. Trichas, Schyns, Lord & Hall, 2017; Haumer & Donsbach, 2009).
The variables classified as nonverbal behavior cues can be found on the meso level of the photograph, which helps to identify tendencies and categorization of the picture (Geise & Rössler, 2013).
Eye gaze. Regarding the gaze, scholars seem to be divided on what is perceived as
dominant and powerful. Verser (2007) for instance, in line with other authors (e.g. Brooks, Church, & Fraser, 1986; Burgoon et al., 1984; Burgoon & Le Poire, 1999), suggests that the “lack of eye gaze can indicate uncertainty or even a lack of confidence, which are two qualities that are unwanted in a political candidate, especially a presidential candidate” (p.
20). However, Exline and Messick (1967) found that the opposite was true: persons that gazed less while listening to others were perceived more dominant.
Apart from studies on interpersonal communication, Van Leeuwen (2008) interpreted the gaze of a depicted person in photography. The author calls direct gazing a social interaction with the viewer of the picture while he points out that a lack of gaze puts the viewer in the state of an observant or voyeur (2008). Although, Exline and Messick (1967) describe a different effect in social interaction, this study follows the indication that political power is emphasized by a direct gaze in the camera.
Facial expression. Most of the literature on implications of facial expressions deals with
the perception of different emotions. Several authors highlight that happy emotions result in an overall higher leadership perception but also expressions of anger or pride are rated as powerful. (e.g. Trichas at al., 2017; Tiedens, Ellsworth, & Mesquita (2000); Conway,
DiFazio, & Mayman, 1999). On the other hand, people expressing sadness, guilt, appreciative expressions and fear are perceived powerless (Tiedens et al., 2000; Conway et al., 1999).
Other research finds, that relaxed faces in contrast to nervous-looking faces (Aguinis, Simonsen, & Pierce, 1998) and lowered eyebrows in contrast to raised eyebrows (Keating, Mazur, & Segall, 1977) signify dominance and power. Besides, Verser (2007) implies that politicians tend to stick to a serious facial expression to emphasize their role as upright and trustworthy leader. This goes in line with findings of Trichas et al. (2017) that highlight the importance of emotional stability expressed for a higher leadership rating.
Overall, political power is supported by a relaxed facial expression, positive emotions or expressions of proudness or anger and lowered eyebrows. Besides, seriousness and emotional stability are important factors that contribute to the perception of political power.
Body posture. Many scholars agree that an open body posture, meaning the person is
standing upright, leaning slightly forward and is oriented towards the conversation partner, signifies power (Mehrabian & Deese, 1969; Bailey & Kelly, 2015; Cashdan, 1998; Carney, Hall, & LeBeau, 2005). Moreover, Mehrabian and Deese (1969) point out that seated women with an open arrangement of their arms are perceived more positive.
Although, it is worth noticing that many scholars make a difference between the power display of men and women when it comes to body posture. Research shows that women, unlike men, “can use counter-stereotypical nonverbal displays, dominant poses, to change how they are initially perceived in terms of power” (Bailey & Kelly, 2015, p. 317). However, Cashdan (1998) reinforces that even for women an open body posture is associated with power. In conclusion, an open body posture is an important indicator of political power for men as well as for women.
Gesture. Schill (2012) finds that forceful hand gestures result in perception of strength and
passion. Other research suggests that vast hand gestures are frequently used by politicians and can signal self-assurance and dominance (Nagel, Maurer & Reinemann, 2012; Verser, 2007). Furthermore, the most powerful and aggressive gestures are the ones including one or two hands forming fists (Morris, 2002; Fischer, Fischer, Englich, Aydin & Frey (2011).
But this might not be true for female politicians. As Dumitrescu (2016) highlights, “female candidates who make agentic hand gestures, such as assertive, expressive, or choppy
movements, are likely to be punished for them” (Dumitrescu, 2016, p.1663). Thus, it can be concluded that forceful gestures like fists reinforce political power in visual communication but choppy very confident movements can have the opposite effect for female politicians.
Clothing. Regarding the clothing and political power there are some general findings
pointed out by scholars. Verser (2007) for instance, finds that formal clothing is predominant in pictures of two presidential candidates in the US. But depending on how the candidates
want to be seen, they choose a more informal clothing, such as Barack Obama staging himself as a man of the people. Besides Feltman and Elliot (2011) find that red clothing has a positive effect on the perception of dominance and threat among recipients. Less research has been done on this field but it can nevertheless be concluded that formal and red colored clothing emphasize a politicians’ power.
Meso level: Structural environment
The third category I will focus on, is the setting or structural environment of the picture. Together with nonverbal behavior cues, this category forms the meso level of analysis and helps identify underlying patterns and tendencies in the depiction (Geise & Rössler, 2013). It includes location, symbols, formality and appearance of other people appear in the shot (Verser, 2007). This category is not necessarily connected with the notion of political power but is an inevitable part of the picture itself and might give interesting insights in the depiction of Angela Merkel in different media systems and across events.
Setting formality, location, symbols, groups. In former studies, the setting has
especially played a role in constructed pictures of politicians, such as posters or campaign visuals (Goodnow, 2013). This category is closely connected to visual framing since the setting can influence the perception of politicians and create a specific image for the audience, like the “man of the people” (e.g. Coleman, 2010; Entman, 1993). Goodnow (2013) shows that politicians are likely to be framed in a certain way in order to appear as statesman or populist campaigner, for instance by picturing the politician sitting at a desk (location) or together with large audiences (groups).
Less research has been done on the concrete effects of the structural environment on the perception of power. However, since some scholars stress the importance of this category and its potential influence on the perception of power (Graber, 1996; Schubert, 2005; Spezio, Loesch, Gosselin, Mattes, & Alvarez, 2012; Reh, van Quaquebe, Giesmer, 2017), it shall be
included in the study. That way, not only the person in focus would be in focus but also its surroundings.
Research Question
After developing the theoretical framework on the visualization of political power, the following part focuses on the main research question as well as a set of sub-questions. The main aim of this study is to find to what extent photos of Angela Merkel differ across media systems and events.
RQ 1: To what extent is it possible to identify differences in the depiction of Angela
Merkel across different media systems?
RQ 2: How far is it possible to identify similar patterns in the depiction of Angela
Merkel across different media systems?
As mentioned before, this study will include four countries, representing the four media systems: Germany (Central), France (Southern), Portugal (Western) and Denmark (Northern) (Brüggemann et al., 2014). Other research shows, that content is affected by the membership to a certain media system (e.g. Esser & Umbricht, 2013). The different variables Hallin and Mancini (2004) use for their classification seem to cause substantial differences in media content. Furthermore, regarding the field of visual communication, Krogstad and Storvik (2012) find differences in the depiction of politicians in Norway and France. Thus, it is to expect that Angela Merkel is portrayed differently in each media system.
On the other hand, one could argue that due to journalistic norms, routine pictures published abroad are mainly picked from news agencies. These underlying media routines that dominate in large newsrooms nowadays might lead to similar patterns in the depiction of Angela Merkel across media systems (Frosh, 2003). According to Reese and Shoemaker (1991), media content is influenced by various levels in the media system. These levels
include journalistic norms and decisions but more importantly media routines as well as organizational, extra media and ideological forces (Reese & Shoemaker, 1991). The influence of the latter levels might lead to a similar depiction of Merkel.
Besides, the chosen European countries are located in the same part of the world and are culturally not very different from each other. Media content would underlie strong influences caused by cultural differences in visual communication. Thus, it is to expect that the published pictures of Merkel are similar.
RQ 3: How does the depiction of Angela Merkel differ between election years and the
refugee crisis?
The refugee crisis 2015 marks a crucial year for Merkel. Not only in Germany but also in Europe she was the main political figure during that time, trying to overcome differences in the European Union, negotiating with Turkey and proclaiming a positive attitude towards refugees (Nardelli, 2005a). However, she was also highly criticized for her actions during that time since not all European leaders agreed on her position towards refugees (Nardelli, 2005a). All four countries in the sample, were affected by the crisis and had their own political agenda influencing the respective media content. France being the target of some terrorist attacks and seeing an influx in migration did not back Merkel up in her course (Hewitt, 2015), and neither did Denmark, since it was one of the main destinations for the incoming refugees and it is known for its harsh migration policies (Gammeltoft-Hansen & Malmvig, 2016). Portugal on the other hand, was welcoming refugees and supporting them due to political and economic reasons (Costa & Sousa, 2017). Thus, it is to expect that there are differences in the depiction between the refugee crisis and election cycles.
Methodology
This study follows the method of quantitative image analysis as suggested by Geise and Rössler (2013). Apart from the structure that was already discussed throughout the paper,
the approach was used to develop the codebook. The emphasis lies on inter-coder reliability and validity of the methodological instrument as this is crucial for the study. Pictures are highly interpretative and can be “read” in many different ways, depending on the specific context, social background and culture of the person looking at it (Geise & Rössler, 2013). By referring to the levels of comparison that the authors introduce (macro, meso, micro), the study ensures to reduce subjectivity and personal interpretation as far as possible. However, the micro level shall be excluded for this study since it aims at underlying information and the deeper meaning of pictures, which is highly interpretative (Rössler & Geise, 2013).
Reliability and validity are weak spots of quantitative image content analysis since the interpretation of pictures is influenced by situation, space, time, individual, social and
medium context (Geise & Rössler, 2013). Therefore, during the development of the codebook I made sure that the unit of analysis was chosen carefully and that the categories for analysis were clear. Besides, an image-based codebook was developed and detailed descriptions were added in order to maximize validity (Geise & Rössler, 2013; see Appendix D for the
codebook).
The codebook consists of 39 variables measured along the categories: general codes, macro level and meso level. The latter was subdivided into nonverbal characteristics and structural environment (see Appendix D). Most variables are measured on a nominal or ordinal scale to record all nuances necessary and avoid reliability problems. Some variables are measured as open categories or in multiple-answers categories to seize all possible outcomes (see Appendix D).
Sampling. The sample was retrieved in four countries, representing the four media
systems (Brüggemann et al., 2014). France, Portugal and Denmark were chosen representing because of accessibility to archives. The analysis in Germany on the other hand, might allow
interesting insights because Angela Merkel is a German politician which affects media content (Nossek, 2004).
Among the newspapers with the highest circulation numbers, the study chose one conservative paper for each country (Die Welt (GER), Le Figaro (FRA), Expresso (PT), Jyllands Posten (DK)). By keeping the political leaning of the media outlets similar, the sample is less prone to political bias that could interfere with the general comparison. The data from Die Welt and Jyllands Posten was accessed via the online archive of the
newspapers. For Expresso and Le Figaro there was no archive material available and thus the online version of the newspaper was included. Unfortunately, this step was unavoidable due to limited resources.
The study identified suitable time frames for the sampling process. Photos were retrieved in the years 2009, 2013 and 2017 during German elections and in 2015 during the refugee crisis, representing almost 10 years of Merkel governing the country. For the three election years, the election day was taken as a reference point. The sampling was retrieved three months before that day and one month after (27.06.2009 – 27.10.2009; 22.06.2013 – 22.10.2013; 24.06.2017 – 24.10.2017).
The refugee crisis was picked based on various factors. Firstly, it affected all European countries equally and was highly discussed in the media. Secondly, Merkel was one of the key figures during the crisis since she represented a strong position to support the incoming
refugees and find a European solution (Nardelli, 2005a). Although, the refugee crisis has no fixed dates, the study will focus on a period of time that includes the influx of migrants arriving in Europe (September and November 2015) as well as the refugee summit with EU member states and Turkey (November 2015) (1.8.2015 – 1.12.2015) (Süddeutsche Zeitung, 2016).
The time frames were searched for pictures of Angela Merkel. The keyword “Merkel” simplified the search for suitable articles. Cartoons and caricatures were excluded from the sample since the emphasis was on press photography.
Table 1
Number of pictures published per country and time frame countries
Total Portugal France Denmark Germany
Time frames 2009 3 24 17 32 76 2013 13 32 29 32 106 2017 18 32 25 32 107 Refugee crisis 33 32 17 29 111 Total 67 120 88 125 400
It was aimed for a sample size of 100 pictures per country. However, this was not possible for Portugal and Denmark since not enough pictures were published in the respective years during the chosen time frames (see table 1). Thus, for both countries the whole
population is included in the study. Consequently, the sample size for France and Germany was raised slightly in order to meet the total sample size of 400. For these two countries, I used the method of random sampling to select pictures from the whole population. This results in a total number of 400 pictures (N = 400).
Coding. The codebook was applied to the sample. When deemed necessary, variables
were slightly modified. After every change, the whole sample was coded again for the respective variables. As preparation for analysis, nominal and ordinal variables were transferred into bivariate variables. For example, the variable perspective was previously coded on a 5-point scale from low to high and transferred into three independent bivariate variables (perspective low, perspective neutral and perspective high; see Appendix A for the
complete list of recoded variables). This was necessary for the statistical analysis in order to run bivariate logistic regressions.
Inter-coder reliability (ICR) was ensured by coding 10% of the total sample by a second coder. The reliability was calculated using Krippendorff’s alpha (Hayes & Krippendorff, 2007). I excluded variables that resulted in obvious answers, like date or photographer as suggested by Lacy, Watson and Lovejoy (2015). This decision was based on my impression of which variables are most important in conveying power across the countries and timeframes. Despite some scholars suggesting a minimal ICR score of 0.80 (Lacy et al., 2015) the value of 0.60 for this study due to limited time for training and retraining, and the relatively small sample size.
After being instructed, the second coder applied the codebook to a few pictures and had the possibility to ask questions. The second coder continued independently and coded in total 40 pictures which makes about 10% of the total sample (N = 400) (see Appendix B for an overview of ICR scores). A set of variables showed insufficient ICR scores below 0.60. After a re-training the second coder, these variables passed the reliability test with a sufficiently high Krippendorff’s alpha (see Appendix B).
Results
Media Systems. For the analysis among the four media systems the pictures were
separated into four country groups (PT: N = 67; DK: N = 88; FR: N = 120; GER: N = 125). The results of logistic regressions show significant differences for 24 variables (see table 2; for a comprehensive list of all variables see Appendix A). In general, it can be concluded that most differences between the countries were found regarding Angela Merkel’s nonverbal behavior.
For the variable quality (high) the logistic regression model was statistically significant. France had a 3.32 times higher chance to have a picture with high quality
published compared to Germany and a 2.95 times higher chance compared to Denmark. This is supported by the variable that describes medium quality since pictures have this feature significantly less in France (see table 2).
Table 2
Significant results logistic regression, country comparison countries PT FR DK GER general codes quality medium 0.13 0.07c* d** 0.16 0.19 high 0.87 0.93c* d** 0.83 0.81 macro level brightness medium 0.88 0.93c*** d* 0.75a*b*** 0.84 high 0.09 0.04c*** d* 0.22a*b*** 0.14 shot close 0.27 0.38 0.44a*d*** 0.22b** c*** meso level portrait neutral 0.35 0.48d* 0.28 0.60a* c*** emotional 0.41b* d** 0.23 0.57b** d*** 0.19
other depicted person is
famous 0.67 0.90a* c* d* 0.72 0.72
distance others is far 0.18 0.28d** 0.21 0.10
others are facing another
person/ point 0.33 0.30 0.15 0.11a** b**
others are facing the
camera 0.30d* 0.40 0.54 0.56 nonverbal behavior impression positive 0.30 0.17a* c*** d*** 0.41 0.35 negative 0.34 0.41c* d* 0.26 0.26 gaze direct 0.09c* d** 0.10c** d*** 0.24 0.27
towards something/
someone 0.82 0.88c** d*** 0.72 0.69a** b**
facial expression relaxed
0.73c** d** 0.82 0.91 0.89
emotion seriousness 0.46c*** d*** 0.53c*** d*** 0.14 0.08 eyebrows neutral 0.72c** d* 0.79 0.94a** b** 0.87
holding a speech 0.21c* 0.24c** d* 0.10 0.12 gesture 0.64c* d*** 0.51d* 0.42a* 0.34a*** b* Structural environment location outdoor 0.28d** 0.22c* d*** 0.40 0.54 party event 0.24d* 0.17 0.17 0.14 background is neutral 0.27b** c** d** 0.43 0.44 0.45 symbols 0.12 0.20c* d** 0.13 0.09
Note. N = 67 (PT). 120 (FR). 88 (DK). 125 (DE) (N = 34 (PT). 60 (FR). 47 (DK). 53 (DE) for meso level). All variables: 0 = no. 1 = yes. Each row displays one binomial logistic regression model. Reference categories are: a = Portugal, b = France, c = Denmark, d = Germany. Mean scores are shown. * < .05. ** < .01. *** < .001.
On the macro level, pictures differed significantly in their length of shot. In Germany, fewest pictures showed Merkel in a close shot, followed by Portugal, France and Denmark. The bivariate logistic regression showed that in France pictures have a 2.08 times higher chance to show Merkel in a close shot (close-up or portrait shot) compared to Germany. The results furthermore show that Denmark publishes the most pictures with a close shot and significantly more than Portugal and Germany. In this case, Germany and Denmark are located on the two sides of the scale: in Denmark pictures are rather close while in Germany the shots are rather long.
The meso level contained the most variation between the data. The binomial logistic regression showed statistically significant results for emotional portraits. In Germany, Merkel was depicted remarkably less emotional – especially when compared to Portugal and
Denmark. The same tendency is true for France. Connected to that, in France the German chancellor is shown displaying significantly more serious emotion. Same goes for Portugal. In Germany and Denmark on the other hand, photos show Merkel less serious.
Regarding the nonverbal characteristics, Merkel was found to be portrayed in a different way for additional variables. First, her eye gaze directed towards the camera was found to be statistically different. The direct comparison with Germany shows that in France and Portugal it is less likely that the published photos show Merkel with a direct gaze (see table 2). In regard to that, in both countries, chances are about four times higher that Merkel’s gaze is directed towards something or someone else. Moreover, in these two countries, chances are significantly higher that Merkel is shown holding a speech. This goes in line with the fact that she is also shown gesticulating more in Portugal and France compared to
Denmark and Germany.
The overall impression of Angela Merkel shows significant differences between the countries as well. The chances are higher that Merkel is depicted more favorable in Portugal, Denmark and Germany compared to France. To complement this result and looking at the contradictory variable, in Germany and Denmark chances are higher that the impression of Merkel is positive in contrast to France.
It can be summarized that the country comparison shows some patterns. Denmark and Germany seem to depict Merkel similarly. France and Portugal show similar depictions that differ from those in Denmark and Germany. Moreover, the photos published in Germany are rather neutral, for instance supported by the variable’s neutral portrait or the fact that Merkel was shown in less close shots. In France, on the other hand, Angela Merkel was depicted least favorable (see variables impression negative or relaxed facial expression).
Regarding the second sub-question that tries to find similarities between the depiction of Merkel across different media systems it is to stress that all variables that did not show
significant differences in the regression analyses support this assumption. This was found to be true specially for the group of general picture codes and the macro level (see Appendix A), like format, sharpness, contrast or brightness of the photographs. These are variables that are supposed to describe the overall structure of the picture and its features. This finding will be discussed further in the discussion of results below.
Time comparison. The third research question is directed towards similarities and
differences across different time frames. For this comparison, all pictures (N = 400) were separated into four groups: election year 2009 (N = 76), election year 2013 (N = 106), election year 2017 (N = 107) and the refugee crisis 2015 (N = 111). Several chi-square tests of independence helped to explore the data as a first step. In total 25 variables were found to have significant differences between the four time frames (see table 3).
Table 3
Significant results logistic regression, time comparison
Time frames
2009 2013 2017 2015
general codes
format landscape 0.71 0.75 0.82 0.87a** b*
quality medium 0.22 0.17 0.06a** b* 0.13 high 0.76 0.83 0.94a*** b* 0.87 macro level perspective low 0.46 0.39 0.40 0.23a*** b* c** high 0.14 0.09 0.08 0.22b* c** shot close 0.29 0.34 0.40d* 0.26 meso level
portrayal with others 0.51 0.39d*** 0.51 0.62
portrait
emotional 0.59b** c*** d*** 0.34 0.22 0.23 other depicted person is
famous 0.79 0.83 0.82 0.65b* c*
reaction of others visible 0.27 0.32 0.33 0.12b* c** nonverbal behavior impression positive 0.33 0.40c* d* 0.25 0.23 medium 0.25c**d** 0.35 0.43 0.45 negative 0.42b* 0.25 0.32 0.32 gaze direct 0.22 0.25c* 0.12 0.15 towards someone or something 0.76 0.71c* 0.85 0.78
relaxed facial expression 0.89 0.88 0.77d* 0.87
Merkel smiling 0.63c** d** 0.49 0.41 0.39 Merkel displays happiness 0.62b* c*** d*** 0.47 0.37 0.37 negative 0.16b* c*** d** 0.28 0.32 0.26 clothing red 0.20 0.10 0.24 0.21b* structural environment
location outdoor 0.44 0.45 0.19a** b** d** 0.38
background neutral 0.42 0.31d** 0.43 0.49
speech 0.18 0.16 0.22d* 0.11
symbols visible 0.09 0.07c** d* 0.18 0.19
Note. N = 76 (2009), 106 (2013), 111 (2015), 107 (2017). All variables: 0 = no. 1 = yes. Each row displays one binomial logistic regression model. The reference category is: a = 2009, b = 2013, c = 2017, d = 2015. Mean scores are shown. * < .05. ** < .01. *** < .001.
Regarding the general codes, table 3 shows that quality and landscape format of the photos differ significantly. The binomial logistic regression highlights that during the refugee crisis 2015 more pictures with a landscape format have been published. Regarding the quality, the analysis shows that in 2017 the photos reached their best. Now compared to other years, the regression reveals that the chances are around three to four times higher for a high-quality
picture to be published in 2017 in contrast to 2009 and 2013. Complementing this finding, a Pearson’s correlation between years and the quality judgement reveals that there is a strong positive correlation between ascending years (2009 – 2017) and quality of the photos, r = .96, p = .041 (see appendix C for correlation results). Photographs that were published earlier have less quality than the ones published recently.
Regarding the macro level, binomial logistic regressions show that the odds are significantly higher that a photo showing Merkel from a low angle was published in one of the election years (2009, 2013 or 2017) compared to the refugee crisis (see table 3). The reverse result is true for a high perspective. It is less likely that pictures are showing Merkel from a high angle in the years 2013 and 2017 (see table 3).
On the meso level some interesting differences are observed among the time frames. Logistic regressions show statistically significant differences among the four time frames, regarding Merkel displaying happiness. In 2009, Angela Merkel was depicted significantly happier and displaying less negative emotions compared to all other years (see table 3). In addition to that, in the year 2009, she is depicted in more emotional portraits and less neutral. But at the same time, the impression of Merkel was not automatically perceived more
positively during 2009.
In fact, during that year, the impression of Angela Merkel was the most negative. However, the difference to the other years is only significant for 2013. The chances are over two times higher that a picture with a negative impression was published in 2009 compared to 2013 (see table 3). Adding to that cluster of results, the results of Pearson’s correlation show that with ascending years Merkel was depicted less positive. Moreover, there are less photos published in which Merkel is depicted happy (see Appendix C for correlation results).
In conclusion, the country comparison showed some interesting results. The year 2009 seems to be a particular “outlier” for many categories which will be further discussed in the
following section. The year was found to be significantly different to all other three years for neutral portraits as well as for emotional ones and for the display of happiness and negative emotions. In the year of the refugee crisis, Angela Merkel was depicted differently from election years most importantly regarding the angle but also whether the other person was famous or showing an emotional reaction.
Generally, it should be pointed that it was found that the year 2009 differed substantially from the others. Unlike for the media system comparison, where all groups showed significant differences for one variable or the other, the differences in time comparison come to a head for 2009. Apart from 2009, less variables show significant differences than the country comparison, suggesting that the visual portrayal of Angela Merkel differs more across countries than over time. These findings will be discussed in the next section of this study.
Discussion
The aim of this study was to explore the field of visual political communication taking the comparison of Angela Merkel’s depiction across (1) time and (2) media systems as
reference points. After considering all results, it becomes clear that the visualization of political power across media systems and for different events differs to a moderate extent. This adds to previous research suggesting that content is influenced by media systems
(Brüggemann et al., 2014; Esser & Umbricht, 2013) or underlies changes in time (e.g. Bucy & Grabe, 2008).
Focusing on the media systems (Brüggemann et al., 2014), it can be noted that in general the depiction of Merkel in Germany was more neutral compared to the other countries. Connecting these findings with the theoretical framework on visualization of political power the photos published in Germany signal emotional stability (Trichas et al.,
2017; Verser, 2007) but they make Merkel appear less approachable or dominant by using less close shots (e.g., Schill, 2012).
This tendency can be explained by drawing from Brüggemann et al. (2014) and their classification of media systems. Germany belongs to the central European cluster, standing out due to a low level of political parallelism. Thus, the newspaper Die Welt depicts Merkel rather neutral despite having a conservative line just like Merkel’s political party.
Furthermore, taking into account that Merkel is a German politician, the neutral depiction in Die Welt might also be connected to journalistic norms of the newspaper (Strömbäck, 2008). The strive to fulfil the watchdog role and avoid any kind of bias might result in a neutral visual portrayal.
In France, Merkel is depicted rather negative and powerless, for instance supported by a less direct gaze in the camera. Connecting the findings to the theory, it shows that a less direct gaze in the camera makes the person appear less powerful as well as less approachable (Verser, 2007; van Leeuwen, 2008). In addition, the fact that Angela Merkel is portrayed more distant to others and that these people are facing another point or person in the picture are drawing the attention from her and show her isolated from groups. Additionally, this is supported by Merkel’s gaze towards others directing the focus away from her. As a last point, the more negative impression of Merkel adds to the overall picture, Merkel being portrayed less powerful in France.
This negativity can be explained from a political point of view. The relationship between France and Germany was foremost under question during the Euro crisis since the countries had a different approach for Europe’s austerity policy (Hewitt, 2015). “Merkel was pictured as Europe’s new ‘Iron Lady’ outside Germany, imposing hardship on Greece as well as other European countries” (Crespy & Schmidt, 2014, p. 1085). The crisis had a long-lasting
effect on the image of Germany in Europe but also of Angela Merkel herself. This narrative of Merkel as ‘Iron Lady’ is supported by serious facial expression.
Another explanation might lie in the classification of media systems. France belongs to the Southern European cluster where according to the authors political parallelism is very strong (Brüggemann et al., 2014). Apart from being a conservative paper as well, Le Figaro seems to be critical about Merkel – perhaps showing that, in this case, French foreign policy is reflected more strongly in visual content than the political leaning of the paper. The
assumption can be supported by empirical research (Nossek, 2004) emphasizing that national identity effects professional norms of journalists.
The comparison between years reveals less consistent patterns in the data. In general, Merkel is portrayed slightly less powerful during the refugee crisis supported by a portrayal from a high camera perspective. The camera angle is a decisive factor for the perception of power, with a high angle making the person appear smaller and less powerful (Giessner et al., 2011). This can be explained by the general criticism Angela Merkel faced during the refugee crisis for her migration policy all over Europe (Gammeltoft-Hansen & Malmvig, 2016; Hewitt, 2015). Moreover, the criticism seems to be reflected in the depiction of her during the 2017 election period, too. The overall impression of her was rather negative during that time unlike the other two election cycles.
The election period in 2009 should be mentioned as well, since the positive and powerful depiction of Merkel sticks out. Although, political reasons may play a role, it has to be emphasized that the sample size during that time was particularly low (N = 76). Including, about 30 pictures less than the other years might have had an effect on the results.
Overall, it is hard to find strong variations concerning the years. Thus, I draw the conclusion that the visualization of political power is less influenced by time but more by different media systems. It was pointed out above that these variations might be caused by the
structural differences in the media system itself (e.g. political parallelism), by cultural
differences between the countries, or from the perspective of international relations, bringing political reasons forward. Indeed, although this study selected countries based on differences in their media systems, it cannot pull apart these competing explanations. A visual framing approach could provide further insights that are not connected to visualization of political power but helpful in this regard.
Finally, it should be stressed that this study supports the assumption that there is no universal depiction of political power. This is supported by the fact that only a few variables did not show significant differences between the groups (sharpness, contrast, depth of field, resolution, type of group portrait, head position, standing). Moreover, this is a strong indicator that the important differences can be found on the meso level, including nonverbal behavior - like smiling, display of emotion or eye gaze - and structural environment – such as location, background or symbols.
Implications, future research and limitations
Even in times when politicians are able to control which photos are taken and a lot of newspapers get their pictures from online sources or agencies it is possible for editors to make a choice and influence the way a politician is portrayed (Rössler et al., 2011). This might sometimes even be used to transport a hidden message when open critique in form of written words is not appropriate or when support of a specific politician is supposed to be
emphasized. This study therefore shows that visuals are an important way in which political journalists can level critique at powerful politicians and possibly influence perceptions of them abroad. According to theories of soft power and nation branding in a broader sense, visualization of political power might even contribute to the overall impressions of countries abroad (George, 2016).
It is crucial to point out that this discussion and interpretation of pictures is influenced by my social and cultural background (Geise & Rössler, 2013) and thus should be tested further. This could be done by carrying out an effect study trying to show how depiction of political power is perceived for a large number of participants. Moreover, in this context, a framing study would be beneficial in order to examine the frames that can be drawn from this study. This would help to understand the specific narratives in each country. Besides, an interesting aspect for further investigation is the question whether powerful depiction and positive perception always go hand in hand.
It has to be mentioned that some bivariate variables could not be examined at all because the sample size was not high enough per group (see Appendix A). On the one hand, that is an important indicator which variables are not as influential for the depiction of
political power. On the other hand, it points to an important limitation of this study - the small number of cases as well as the unequal group size in my sample. Although everything was done to keep the amount of cases as high and as comparable as possible, it is crucial to repeat the study with a larger sample.
Another critical point regarding the sample is the fact that conservative newspapers might not have the same degree of conservativeness across all countries included in the sample. However, having included this factor in my sampling process made the sample as little influenced by the political leaning as possible. Adding to the criticism, it should be mentioned that online and offline sources alike are included in the sample. This aspect was not avoidable for my research due to reasons of accessibility but should be avoided for further research or could be subject of a future study.
Regardless, this study makes an important contribution in the field of visual political communication by adding findings from comparisons across media systems and time frames. Being one of the first studies that takes a closer look into the visualization of political power it
was shown that even such a universal concept like power is underpinned by diverse depictions in press photography.
References
Aguinis, H., Simonsen, M. M., & Pierce, C. A. (1998). Effects of nonverbal behavior on perceptions of power bases. The Journal of social psychology, 138(4), 455–469. https://doi.org/10.1080/00224549809600400.
Bailey, A. H., & Kelly, S. D. (2015). Picture Power: Gender Versus Body Language in Perceived Status. Journal of Nonverbal Behavior, 39(4), 317–337.
https://doi.org/10.1007/s10919-015-0212-x.
Barrett, A. W., & Barrington, L. W. (2005). Bias in Newspaper Photograph Selection. Political Research Quarterly, 58(4), 609. https://doi.org/10.2307/3595646.
Brooks, C., Church, M., & Fraser, L. (1986). Effects of Duration of Eye Contact on
Judgments of Personality Characteristics. The Journal of Social Psychology, 126(1), 71-78. Bucy, E. & Grabe, M. (2008). “Happy warriors” revisited Hedonic and agonic display
repertoires of presidential candidates on the evening news. Politics and the life sciences: the journal of the Association for Politics and the Life Sciences, 27, 78-98.
https://doi.org/10.2990/27_1_78.
Burgoon, J. K., & Koper, R. J. (1984). Nonverbal and Relational Communication Associated with Reticence. Human Communication Research, 10(4), 601-26.
Burgoon, J. K., & Le Poire, B. (1999). Nonverbal cues and interpersonal judgments: Participant and observer perceptions of intimacy, dominance, composure, and formality. Communication Monographs, 66(2), 105-124.
Brüggemann, M., Engesser, S., Büchel, F., Humprecht, E., & Castro, L. (2014). Hallin and Mancini Revisited: Four Empirical Types of Western Media Systems. Journal of
Communication, 64(6), 1037-1065.
Calogero, R.M. & Mullen, B. (2008). About face: Facial prominence of George W. Bush in political cartoons as a function of war. The Leadership Quarterly, 19(1), 107-116.
Carney, D., Hall, R., & LeBeau, J. (2005). Beliefs about the nonverbal expression of social power. Journal of Nonverbal Behavior, 29(2), 105-123.
Cashdan, E. (1998). Smiles, Speech, and Body Posture: How Women and Men Display Sociometric Status and Power. Journal of Nonverbal Behavior, 22(4), 209-228.
Coleman, R. (2010). Framing the pictures in our heads: Exploring the framing and agenda-setting effects of visual images. In P. D’Angelo & J. Kyupers (Eds.), Doing news framing analysis: Empirical and theoretical perspectives (233-261). New York: Routledge.
Coleman, R., & Banning, S. (2006). Network TV News' Affective Framing of the Presidential Candidates: Evidence for a Second-Level Agenda-Setting Effect through Visual Framing. Journalism & Mass Communication Quarterly, 83(2), 313-328.
Connolly, K. (2013, Sept 3). 'Merkel diamond' takes center stage in German election campaign. The Guardian. Retrieved from https://www.theguardian.com/world/german-elections-blog-2013/2013/sep/03/angela-merkel-diamond-german-election-campaign Conway, M., DiFazio, R., & Mayman, S. (1999). Judging Others' Emotions as a Function of
the Others' Status. Social Psychology Quarterly, 62(3), 291-305.
Costa, P., & Sousa, L. (2017, Feb 10). Portugal’s Openness to Refugees Makes Demographic and Economic Sense. Refugees Deeply. Retrieved from
https://www.newsdeeply.com/refugees/community/2017/02/10/portugals-openness-to-refugees-makes-demographic-and-economic-sense
Crespy, A., & Schmidt, V. (2014). The clash of Titans: France, Germany and the discursive double game of EMU reform. Journal of European Public Policy, 1-17.
Davidson Sorkin, A. (2017, Sept 22). Angela Merkel, the Most Powerful Woman in a World of Unstable Men. The New Yorker. Retrieved from
https://www.newyorker.com/news/daily-comment/angela-merkel-the-most-powerful-woman-in-a-world-of-unstable-men
Dumitrescu, D. (2016). Nonverbal Communication in Politics. American Behavioral Scientist, 60(14), 1656–1675. https://doi.org/10.1177/0002764216678280.
Entman, Robert M. (1993). Framing: Toward Clarification of a Fractured Paradigm. Journal of Communication, 43(4), 51-58.
Esser, F., & Umbricht, A. (2013). Competing models of journalism? Political affairs coverage in US, British, German, Swiss, French and Italian newspapers. Journalism, 14(8), 989– 1007.
Exline, R. V., & Messick, D. (1967). The effects of dependency and social reinforcement upon visual behavior during an interview. British Journal of Social and Clinical Psychology, 6, 256–266.
Fahmy, S. (2005). Photojournalists' and Photo Editors' Attitudes and Perceptions: The Visual Coverage of 9/11 and the Afghan War. Visual Communication Quarterly, 12(3-4), 146-163.
Feltman, R., & Elliot, A. (2011). The influence of red on perceptions of relative dominance and threat in a competitive context. Journal of Sport & Exercise Psychology, 33(2), 308-14.
Fischer, J., Fischer, P., Englich, B., Aydin, N., & Frey, D. (2011). Empower my decisions: The effects of power gestures on confirmatory information processing. Journal of Experimental Social Psychology, 47(6), 1146-1154.
Forbes (2018). The World’s 100 Most Powerful Women. Forbes Magazine. Retrieved from https://www.forbes.com/power-women/#7a8247325e25.
Frosh, P. (2011). Framing Pictures, Picturing Frames: Visual Metaphors in Political Communications Research. Journal of Communication Inquiry, 35(2), 91–114. https://doi.org/10.1177/0196859911410242.
Frosh, P. (2003). The image factory: Consumer culture, photography and the visual content industry. Oxford: Berg.
Gammeltoft-Hansen, T. & Malmvig, H. (2016, May 4). The Ugly Duckling: Denmark’s Anti-Refugee Policies and Europe’s Race to the Bottom. Huffington Post. Retrieved from
https://www.huffingtonpost.com/thomas-gammeltofthansen/denmark-refugee-europe_b_9574538.html
Geise, S. & Rössler, P. (2013). Standardisierte Bildinhaltsanalyse. In W. Möhring & D. Schlutz (Eds.), Handbuch standardisierte Erhebungsverfahren in der
Kommunikationswissenschaft (307-326). Wiesbaden: Springer Vs.
George, C. (2016). Soft power: Media influence and its limits, Media Asia, 43 (2), 69-75, https://doi.org/10.1080/01296612.2016.1209716.
Giessner, S. R., Ryan, M. K., Schubert, T. W., & van Quaquebeke, N. (2011). The power of pictures: Vertical picture angles in power pictures. Media Psychology, 14(4).
Goodnow, T. (2013). Facing Off. American Behavioral Scientist, 57(11), 1584–1595. https://doi.org/10.1177/0002764213489013.
Graber, D. A. (1996). Say it with pictures. The Annals of the American Academy of Political and Social Science, 546, 85-96.
Hall, C.C., Goren, A., Chaiken, S., & Todorov A. (2009). Shallow Cues with Deep Effects: Trait Judgments From Faces and Voting Decisions. In E. Borgida, C.M. Federico, & J. L. Sullivan (Eds.), The Political Psychology of Democratic Citizenship (chapter 4)
https://doi.org/10.1093/acprof:oso/9780195335453.003.0004.
Hallin, D., & Mancini, P. (2004). Comparing media systems: Three models of media and politics. Cambridge, New York: Cambridge University Press.
Haumer, F. & Donsbach, W. (2009). The Rivalry of Nonverbal Cues on the Perception of Politicians by Television Viewers, Journal of Broadcasting and Electronic Media. Journal of Broadcasting & Electronic Media, 53(2), 262–279.
Hayes, A. F. & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures 1, 77-89.
Hewitt, G. (2015, Nov 15). Paris attacks: Impact on border and refugee policy. BBC News. Retrieved from: http://www.bbc.com/news/world-europe-34826438
Keating, C. F., Mazur, A., & Segall, M. H. (1977). Facial gestures which influence the perception of status. Sociometry, 40(4), 374–378.
Krogstad, A., & Storvik, A. (2012). Picturing Politics: Female Political Leaders in France and Norway. Historical Reflections/Reflexions Historiques, 38(3).
https://doi.org/10.3167/hrrh.2012.380308.
Kulish, N. (2009, Sept 27). Merkel’s Party claims Victory in Germany. The New York Times. Retrieved from https://www.nytimes.com/2009/09/28/world/europe/28germany.html Lacy, S., Watson, B., Riffe, D., & Lovejoy, J. (2015). Issues and Best Practices in Content
Analysis. Journalism & Mass Communication Quarterly, 92(4), 791-811.
Mamer, B. (2000). Film production technique: Creating the accomplished image. Australia: Wadsworth.
Mandell, L., & Shaw, D. (1973). Judging people in the news — unconsciously: Effect of camera angle and bodily activity. Journal of Broadcasting, 17(3), 353-362.
Marcinkowski, F., Lünich, M., & Starke, C. (2017). Spontaneous trait inferences from candidates’ faces: The impact of the face effect on election outcomes in Germany. Acta Politica, 34(1), 213. https://doi.org/10.1057/s41269-017-0048-y.
Mehrabian, A., & Deese, J. (1969). Significance of posture and position in the communication of attitude and status relationships. Psychological Bulletin, 71(5), 359-372.
Morris, D. (2002). Peoplewatching. London: Vintage.
Nagel, F., Maurer, M., & Reinemann, C. (2012). Is There a Visual Dominance in Political Communication? How Verbal, Visual, and Vocal Communication Shape Viewers' Impressions of Political Candidates. Journal of Communication, 62(5), 833–850. https://doi.org/10.1111/j.1460-2466.2012.01670.x.
Nardelli, A. (2015a, Nov 8). Angela Merkel’s stance on refugees means she stands alone against catastrophe. The Guardian. Retrieved from
https://www.theguardian.com/commentisfree/2015/nov/08/angela-merkel-refugee-crisis-europe
Nardelli, A. (2015b, Sept 18). Angela Merkel: 10 years in 10 charts. The Guardian. Retrieved from https://www.theguardian.com/news/datablog/2015/sep/18/angela-merkel-10-years-germany-chancellor-economy-cdu
Nossek, H. (2004). Our News and their News: The Role of National Identity in the Coverage of Foreign News. Journalism: Theory, Practice & Criticism, 5(3), 343-368.
O'Brien, S. (1993). Eye on Soweto: A Study of Factors in News Photo Use. Journal of Mass Media Ethics, 8(2), 69-87.
Reh, S., van Quaquebeke, N., & Giessner, S. R. (2017). The aura of charisma: A review on the embodiment perspective as signaling. The Leadership Quarterly, 28(4), 486–507. https://doi.org/10.1016/j.leaqua.2017.01.001.
Reese, S., & Shoemaker, P. (1991). Mediating the Message: Theories of Influence on Mass Media Content. New York: Longman.
Rössler, P., Bomhoff, J., Haschke, J. F., Kersten, J., & Müller, R. (2011). Selection and impact of press photography: An empirical study on the basis of photo news
factors. Communications: The European Journal of Communication Research, 36(4), 415-439. https://doi-org.proxy.uba.uva.nl:2443/10.1515/comm.2011.021.
Schill, D. (2012). The Visual Image and the Political Image: A Review of Visual Communication Research in the Field of Political Communication. Review of Communication, 12(2), 118–142. https://doi.org/10.1080/15358593.2011.653504. Schubert, T. W. (2005). Your highness: Vertical positions as perceptual symbols of power.
Journal of Personality and Social Psychology, 89(1), 1–21.
Spezio, M., Loesch, L., Gosselin, F., Mattes, K., & Alvarez, R. (2012). Thin‐Slice Decisions Do Not Need Faces to be Predictive of Election Outcomes. Political Psychology, 33(3), 331-341.
Strömbäck, J. (2008). Four phases of mediatization: An analysis of the mediatization of politics. The International Journal of Press/Politics, 13(3), 228-246.
Süddeutsche Zeitung (2016, Aug 31). Etappen der Flüchtlingskrise in Deutschland: Eine Chronologie. Süddeutsche Zeitung. Retrieved from
http://www.sueddeutsche.de/news/politik/migration-etappen-der-fluechtlingskrise-in-deutschlandeine-chronologie-dpa.urn-newsml-dpa-com-20090101-160831-99-279450 Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000). Stereotypes About Sentiments and
Status: Emotional Expectations for High- and Low-Status Group Members. Personality & Social Psychology Bulletin, 26(5), 560-574.
Trichas, S., Schyns, B., Lord, R., & Hall, R. (2017). “Facing” leaders: Facial expression and leadership perception. The Leadership Quarterly, 28(2), 317–333.
https://doi.org/10.1016/j.leaqua.2016.10.013.
van Leeuwen, T. (2008). Discourse and practice: New tools for critical discourse analysis. Oxford studies in sociolinguistics. Oxford, New York: Oxford University Press.
Verser, R. (2007). The 2004 presidential election between George W. Bush and John F. Kerry: An analysis of visually comparative televised advertisements. Retrieved from https://mospace.umsystem.edu/xmlui/bitstream/handle/10355/4696/research.pdf?sequence =3&isAllowed=y
Verser, R., & Wicks, R. (2006). Managing Voter Impressions: The Use of Images on Presidential Candidate Web Sites During the 2000 Campaign. Journal of
Communication, 56(1), 178-197
Vigsø, O. (2017). The visual construction of personal ethos in election posters. The Poster, 4(1), 31–57. https://doi.org/10.1386/post.4.1-2.31_1.
Weber, M. (1973): Soziologie, universalgeschichtliche Analysen. Stuttgart.
Zeit Online (2017, Jul 12). Jeder Dritte kann sich für Merkel begeistern. Die Zeit. Retrieved from https://www.zeit.de/politik/deutschland/2017-07/bundestagswahl-angela-merkel-martin-schulz-umfragewerte
Zuckerman, M. (1986). On the meaning and implications of facial prominence. Journal of Nonverbal Behavior, 10(4), 215–229.
Appendix A
The table shows the list of recoded variables from the codebook (see Appendix D) into
bivariate versions. Moreover, it indicates which variables were dropped for the analysis due to low number of cases (<5). The respective variable names are in italics.
Table 1
List of bivariate variables after recoding process
General picture codes
Size (h) Size (w)
Format landscape Insufficient number of cases (country) Format square
Format vertical Insufficient number of cases (country) Knowledge of the picture Insufficient number of cases (country)
Low quality Insufficient number of cases (country)
Insufficient number of cases (time) Medium quality
High quality
Macro level
Resolution high Insufficient number of cases (country) Insufficient number of cases (time) Resolution medium Insufficient number of cases (country)
Insufficient number of cases (time) Resolution high
Sharpness medium Insufficient number of cases (country) Insufficient number of cases (time) Sharpness high Insufficient number of cases (country)
Insufficient number of cases (time) Contrast low Insufficient number of cases (country) Contrast medium
Contrast high
Brightness low Insufficient number of cases (country) Insufficient number of cases (time) Brightness medium Insufficient number of cases (time) Brightness high Insufficient number of cases (time) Depth of field shallow
Depth of field medium Depth of field deep Perspective low Perspective neutral Perspective high Shot close Shot neutral Shot far Meso level Type of portrait Portrait neurtal Portrait talking
Portrait emotional
Portrait group Insufficient number of cases (country) Portrait situational Portrait chat Others famous Distance close Distance neutral Distance far Facing Angela Facing other
Emotional reaction Insufficient number of cases (country) Insufficient number of cases (time) Reaction hapiness Insufficient number of cases (country)
Insufficient number of cases (time) Weak emotional reaction Insufficient number of cases (country)
Insufficient number of cases (time) Medium emotional reaction
Strong emotional reaction Insufficient number of cases (country) Insufficient number of cases (time)
Nonverbal characteristics
Positive Impression Neutral impression Negative impression Gaze direct
Gaze towards someone or something
Eyes closed Insufficient number of cases (country)
Insufficient number of cases (time) Relaxed facial expression
Smiling Happiness
Negative emotion Seriousness
Degree emotionality
Eyebrows raised Insufficient number of cases (country) Insufficient number of cases (time) Eyebrows neutral
Eyebrows lowered Insufficient number of cases (country) Insufficient number of cases (time)
Head up Insufficient number of cases (country)
Insufficient number of cases (time) Head neutral
Head down
Bodyposition open Insufficient number of cases (time) Standing
Gesture
Gesture quotable Insufficient number of cases (country) Insufficient number of cases (time) Gesture iconic
Gesture pointing Insufficient number of cases (country) Insufficient number of cases (time) Gesture pragmatic
Clothing Color red
Color blue Insufficient number of cases (time)
Structural environment
Location press conference Location outside
Location party event Insufficient number of cases (time) Location parliament Insufficient number of cases (country)
Insufficient number of cases (time) speech
Background crowd Background neutral Symbols
Symbol EU flag Insufficient number of cases (country)
Appendix B Inter-coder reliability test
The sufficient results of the Inter-coder reliability test are presented in Appendix B, table 1. Some variables did not show a sufficient Krippendorff’s alpha ( > 0.60) after the first test. Thus, it was repeated a second time for the failed variables only (see Appendix B, table 2).
Table 1
ICR test for relevant variables
Variable names Krippendorff’s alpha
shot = .73
portrait type = .90
single portrait type = .85
distance = .87
impression = .63
gaze = .94
emotion Merkel = .80
degree emotion Merkel = .86
eyebrows = .87 head = .68 body position = 1.0 standing = 1.0 gestures = .75 type of gestures = .75 colors = .79 speech = .73 background = .85 symbols = .90 type of symbol = .74 smile = .75
Table 2
Second ICR test for relevant variables
Variables names Krippendorff’s alpha
perspective = .67
portrait type with others = .80
facing Angela Merkel = .69
emotional reaction = .72
facial expression = .74
Appendix C Table 1
Significant results correlation, time comparison
Time frames 2009 2013 2015 2017 r General codes Quality 4.04 4.21 4.58 4.73 ,959* Macro level Resolution 4.12 4.38 4.77 4.83 ,968* Depth of field 3.65 3.48 3.46 3.43 -,958* Meso level Portrait emotional 0.59 0.34 0.23 0.22 -,969* Nonverbal behavior
Merkel displays happiness 0.62 0.47 0.37 0.37 -,971*
Eyebrows lowered 0.05 0.09 0.11 0.12 ,993*
Head directed upwards 0.08 0.06 0.05 0.03 -,984*
Note. N = 76 (2009), 106 (2013), 111 (2015), 107 (2017). Quality & resolution: 1 = very low, 2 = rather low, 3 = medium 4 = rather high, 5 = very high. Depth of field: 1 = very shallow, 2 = rather shallow, 3 = medium, 4 = rather deep, 4 = deep. All variables meso level and
nonverbal behavior: 0 = no. 1 = yes. Each row displays one Pearson’s correlation model. Mean scores are shown. Significance is two-tailed.
Appendix D Codebook
Unit of analysis are photographs of Angela Merkel published in newspapers. The main focus of the analysis shall lie on Angela Merkel itself, unless specified differently.
Please note: Code 99 should only be used as last resort.
General codes
1 CODE
Note down the picture code following this format:
YYYYMMDD + media outlet code + picture number
Note: The picture number is assigned to every picture in the sample (per newspaper, starting from 1).
For example:
The picture 04 was published on 16th of March 2018 in the newspaper Die Welt. The picture code is: 201803163104
2 DATE
Note down the date the picture was published, following this format: YYYYMMDD
3 MEDI
Note down the medium the picture was published in
31 = Die Welt 32 = Le Figaro 33 = Jyllands Posten 34 = Expresso 99 = Unknown
General picture codes
4 SIZE
Note the size of the picture in the following format:
width x height
Note I: The size of the picture is measured in pixel. Note II: If unknown, please code 99.
5 FORMAT
Format of the picture
1 = strong landscape format
2 = = simple landscape format
