• No results found

You are the one to blame: How the attribution of responsibility influences the way the public reacts to Twitter messages about crises

N/A
N/A
Protected

Academic year: 2021

Share "You are the one to blame: How the attribution of responsibility influences the way the public reacts to Twitter messages about crises"

Copied!
27
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)
(2)

Introduction

In the last few years, public diplomacy has been responding to the changes in international and domestic environments and societies, with new communication techniques being applied (Hocking & Melissen, 2015, p. 9). With 86% of all 193 United Nations members having presence on Twitter, it has become an essential tool for governments to communicate with their publics (Twiplomacy, 2015). Considering there are 380 million monthly users active on Twitter and 1,65 billion monthly users active on Facebook, it has become easier for governments to reach a broad and targeted audience by using Social Networking Sites (SNS). While traditional public diplomacy focused on one-way communication, where governments mostly just transmit information, the new public diplomacy focuses on a dialogue with the public and enables direct interaction with the help of new tools like SNS (Melissen, 2005, p. 11-12). Although this new form of communication style does not mean that governments should negotiate with their publics about policy, SNS does give the opportunity to “increase the mutual understanding between governments and citizens worldwide” (Cercel & Saftescu, 2015, p. 386).

SNS seems particularly useful during a crisis, since “the public can be informed directly and quickly, and governments can engage in a dialogue with them” (Utz et al, 2013, p. 41). During the Nepal earthquake in 2015, which killed over 8000 people and injured over 21000 people, most foreign ministries throughout the world used digital diplomacy to respond to the earthquake and demonstrated the manner in which social media was being used during a crisis, with all foreign ministries offering emergency information on their Twitter accounts (Manor, 2015). While some foreign ministries only posted emergency information, other foreign ministries also replied to questions from their followers, or made use of trending hashtags to reach more people. This implies that foreign ministries are “still attempting to master the use of digital diplomacy” (Manor, 2015).

Whether diplomatic actors will engage with their public, depends on the the number of requests the diplomatic actor receives, and on the character of the requests. Looking at the character of the requests, positive requests will encourage interaction, while negative requests will discourage interaction (Tromble, under review). To reach the potential of digital diplomacy, that is, reciprocal engagement, the tone of requests has to be mostly positive, since this encourages interaction. In reality this will not be the case, since the public will not always praise the diplomatic actor. But why is some of the communication from diplomatic actors during a crisis perceived in a positive way, while other messages are perceived in a negative

(3)

way? Until now, no research has been done on why digital communication by diplomatic actors is perceived in a positive or negative way.

To answer this question, the Situational Crisis Communication Theory (Coombs, 2007) which is based on the Attribution theory, a social psychology theory, is used. It is argued that the attribution of responsibility, that is, whether the public holds the actor responsible for causing the crisis, influences the way the public reacts to Twitter messages that are send by this diplomatic actor about the crisis. While recent research has started to analyse the role of attribution of responsibility in combination with SNS in crisis communication, most of these studies mainly focus on organizations that are non-governmental and therefore do not focus on digital diplomacy (Jin et al., 2011; Schultz et al., 2011; Utz et al. 2013). The public will also experience an emotional response to the crisis communication after they have attributed responsibility (Weiner, 2001, p. 7). Although little research is done, it is also important to study this emotional response of the public, since this can “provide more insight for diplomatic actors on how to communicate with publics in a sensible way to achieve more effectiveness” (Jin, 2014, p. 81). In other words, if this insight is given, it will be easier to understand how diplomatic actors can use digital diplomacy in an effective way and reach the full potential of digital diplomacy.

To understand whether attribution of responsibility has an impact on the emotions the public experiences, the replies to tweets about four crises were analysed. In two of the cases, the diplomatic actor was responsible for causing the crisis, while in two of the other cases, the diplomatic actor was not responsible for causing the crisis. An overview of digital diplomacy and the use of digital diplomacy during a crisis will be offered first, followed by an explanation of Attribution theory and Situational Crisis Communication Theory. The methods that are employed to collect and analyse the Twitter data will then be presented, and this data will be analysed in a quantitative way, by using statistical techniques, and in a qualitative way.

Literature Review

Digital diplomacy

Public diplomacy is seen as one of the key instruments of soft power. Joseph Nye (2008, p. 94) defined soft power as “the ability to affect others to obtain the outcomes one wants through attraction rather than coercion or payment”. Unlike hard power, where behaviour is influenced by economic or military means, soft power affects people through cooperation and attraction (Gilboa, 2006, p. 718). Especially since access to global information has become a

(4)

lot easier, the battle for overseas “hearts and minds” has become increasingly important for states (Nye 2004, p. 1). To win the support from foreign audiences, public diplomacy is used as a strategic communication tool to project soft power (Sotiriu, 2015, p. 35; Gilboa, 2008, p. 57). While traditional diplomacy is about the relationships between the representatives of states, public diplomacy targets non-state actors and more specifically, the general public in foreign societies. Engagement with foreign audiences is increasingly accepted as “a condition of success in foreign policy” (Melissen, 2005, p. 13-14). Public diplomacy can therefore be seen as a strategic two-way communication tool to influence public opinion in foreign societies (Melissen, 2005; Gilboa, 2008)

More recently, social network sites (SNSs), such as Twitter and Facebook, are being incorporated in diplomatic efforts. Known as digital diplomacy, it is seen as an important tool that enables direct two-way communication and engagement with foreign audiences. Digital diplomacy can be used “to achieve foreign policy goals and to manage image and reputation”, and exists at two levels, namely that of foreign ministries and embassies worldwide (Manor & Segev, 2015, p. 94). According to Nye (2010, np), it is crucial that digital diplomacy remains a two-way process, since the projection of soft power “depends upon understanding the minds of others”. In reality, dialogic communication does not seem to be the standard when it comes to engagement on social media by diplomatic actors. Kampf et al (2015) found that occurrences of dialogic engagement between foreign ministries and their online followers are a rarity. In this research, the authors evaluated the content published on SNS by eleven foreign ministries over a six-week period. Occurrences of engagement turned out to be low and when engagement did occur, it was in the form of a Question and Answer session. Altogether, the dialogic engagement only represented a small fraction of the overall activity. This is also the case when it comes to the use of digital diplomacy by embassies. Bjola and Julang (2015) analysed the diplomatic strategies of the embassies of the US, EU Delegation and Japan in China on the Chinese social media website Weibo, and found that instead of using social media as a tool to engage with their audiences, the embassies mostly used digital diplomacy as “an instrument for information dissemination” (Bjola & Julang, 2015, p. 87). These studies confirm that digital diplomacy has not reached its full potential of reciprocal engagement yet, even though it is seen as an easy tool to communicate with the mass public.

While digital diplomacy can be used as a strategy of public diplomacy, Holmes (2015) also sees it as a form of change management. States use digital diplomacy to manage two sources of change: incremental change and exogenous shocks. Incremental change is seen as a bottom-up process and happens because of decisions that are being made in everyday life.

(5)

These day-to-day developments can be anything, such as “a change in public mood, emotions or affective states, or discourse surrounding a particular issue” (Holmes, 2015, p. 21). On the other hand, exogenous shocks are “events that change the conditions in a way that it triggers actors to intersubjectively interpret them as requiring change, and can therefore be seen as top-down, structural-level shifts” (Holmes, 2015, p. 22-23). Incremental change is about supplying knowledge and information, while exogenous shocks demand attention “to reputations, negotiations, shared understandings and relationship construction” (Holmes, 2015, p. 24). According to Holmes (2015, p. 24), this makes digital diplomacy less useful when change is the result of exogenous shocks, because unlike face-to-face diplomacy, digital diplomacy does not excel at understanding intentions and managing relationships.

This suggests that digital diplomacy is not very useful during a time of crisis, since a crisis can be seen as an exogenous shock. However, little research is done on the effect of a crisis on the way social media messages are received by the public, even though the use of digital diplomacy differs during a crisis. Although there are many ways to define a crisis, for the purpose of this research a crisis is defined as “anything that interrupts the normal flow of business”, such as “an emergency, disaster, calamity or catastrophe” (Hagan, 2007, p. 414). Crisis public diplomacy distinguishes itself from traditional public diplomacy in the way that it “often entails communicating simultaneously with multiple audiences, including hostile publics, in a rapidly changing, highly visible and competitive communication environment” (Zaharna, 2004, p. 145). In certain crisis types, such as a war or terrorist attack, diplomatic actors will try to gain more domestic support during the crisis by identifying a foreign enemy. In turn, this will alienate the foreign public who identifies with this foreign enemy (Zaharna, 2004, p. 145). This could mean that crisis public diplomacy hinders the projection of soft power, since the foreign public that identifies with the foreign enemy will be less open to cooperation.

Attribution theory

Another way to explain the impact of a crisis or conflict situation on the way social media messages are received by the public, is by using Attribution Theory. This theory has its roots in social psychology, and is concerned with how people search for causes of events and therefore make attributions, especially when these events are negative and unexpected (Weiner, 1985). The reaction of the public to social media messages from diplomatic actors could therefore be explained by who the public holds responsible for causing the crisis. Publics can attribute responsibility either internal or external (Heider, 1958). Internal

(6)

attribution occurs when performance is perceived as a result of personal abilities and actions of the actor, for example his intelligence. On the other hand, external attribution occurs when the performance of the actor is perceived as a result of an exogenous situation that is beyond the actor’s control, for example a natural disaster (Sirin & Villalobos, 2011, p. 336).

When it comes to attribution by the public, an actor-observer bias exists, which means that the public will mostly use internal attribution and will perceive the performance of the diplomatic actor as a result of his or her actions. According to Jones and Nisbett (1972, p. 80), the actor-observer bias means “that there is pervasive tendency for actors to attribute their actions to situational requirements, whereas observers tend to attribute the same actions to stable personal dispositions”. This means that if a crisis takes place, for example the terrorist attack on the American diplomatic compound in Benghazi in 2012 where the ambassador and three others were killed, the United States will more likely blame this on factors that are beyond its control, in this case the surprise terrorist attack. On the other hand, the public will more likely see the attack as the result of actions of the United States, since the United States denied requests for additional security on the compound before the attack happened.

Next to internal and external attribution, controllability is of importance when observers perceive someone as responsible (Weiner, 1995). If the outcome of an event is controllable by others, that is, if the actor could have altered the cause of the event, then the actor is perceived as responsible for the outcome. Contrary, if the outcome of an event was not controllable, the actor is not perceived as responsible (Weiner, 2001, p. 7-8). In case of the Benghazi attack, members of the consulate already feared for their safety and asked for additional security, and this request was not honoured. If additional security measurements were taken, it would have been less likely that the ambassador would have been killed. This means that the outcome was controllable and the United States will be perceived as responsible. However, if the attack against the consulate would have taken place in the Netherlands and there were no signals that safety was a concern, this would mean that there would have been less controllability and the United States would not be perceived as responsible. This is also the case for most natural disasters, such as earthquakes and volcanic eruptions, since natural disasters are usually not controllable by governments. However, if the government knew that a natural disaster was about to happen, for example a tsunami, and it did not warn its citizens to evacuate in time, the government will be held responsible for the outcome of the event.

(7)

Situational Crisis Communication Theory

The Situational Crisis Communication Theory (SCCT) by Coombs (2007) has used these assumptions and has laid the link between Attribution Theory and crises. SCCT is audience orientated, and seeks to explain how an organization’s publics will perceive the crisis and how they will attribute crisis responsibility. It also looks at the reputational threat for an organization, because the reputation of an organization, which is how the public perceives the organization, is threatened when a crisis happens (Coombs & Holladay, 2002, p. 167). This theory also gives crisis response strategies for crisis managers to protect the reputation of the organization. It is important for a state that its reputation is not damaged by a crisis, since the success of projecting soft power also depends on the reputation of the state within the international community (Wang, 2006, p. 91). Although SCCT has not been applied to international relations, it can be relevant since it explains the relationship between communication by organizations (which could be foreign ministries or embassies), the response of stakeholders (the public) and crises. The internal-external dimension and the intentional-unintentional dimension of Attribution theory influence how the public will interpret the crisis (Coombs, 1995, p. 454). According to Coombs (1995, 2007) initial crisis responsibility by stakeholders shapes the reputational threat, since the stronger the perceived responsibility, the more likely the crisis will have a negative impact on the reputation of the organization. The perception of crisis responsibility also increases when attribution of personal control intensifies, which is based on the internal-external dimension (Coombs & Holladay, 2002, p. 168). This means that if the actor had control over the cause of the event, attribution of responsibility will increase.

The news media and Internet play a critical role in framing a crisis, since most people will learn about a crisis through news reports and online social media. Framing means that information is presented in a message in a certain way, which can be described as frames in communication. This could be done by highlighting certain aspects of the situation or by using certain phrases or words, which will affect political judgements (Coombs, 2007, p. 167). Frames in communication also help to shape frames in thought, which is “how people define problems, causes of problems, attributions of responsibility and solutions to problems” (Coombs, 2007, p. 167). The crisis type, that is, “the frame that publics use to interpret the event”, needs to be identified to select the right crisis response strategy (Coombs & Holladay, 2002, p. 167). SCCT has therefore grouped crisis types into three crisis clusters, which help to identify how a crisis will be framed and also determine how much responsibility the public

(8)

will attribute to the organization for causing the crisis (Coombs & Holladay, 2002; Coombs, 2007).

The first type is the Victim cluster, where the organization is also a victim of the crisis along with the stakeholders. The organization had no intention to cause the event to happen or had any control over this. The attribution of crisis responsibility will therefore be low. The second type is the Accidental cluster; in this type of crisis the actions of the organization led to a crisis but were unintentional. Because the organization did not intend to create the crisis, but did have some sort of controllability, attribution of crisis responsibility will be moderate, since there was no intention to create the crisis. The third type is the Preventable cluster, where the organization purposely took inappropriate actions and violated regulation. The organization did have the intention to cause the crisis to happen and also had control over this. Attribution of crisis responsibility will therefore be strong (Coombs & Holladay, 2002; Coombs, 2007).

Looking at these types of crisis clusters, the response of stakeholders will depend on how the crisis is being framed. When observers have attributed responsibility, they will also experience an emotional reaction. In social psychology it is possible to examine emotions in term of valence (positive versus negative) or in discrete emotions (hope, sympathy, sadness, anger). Attribution theory mainly uses sympathy and anger as the discrete emotions (Weiner, 2001, p. 5). The public will therefore react with either sympathy or anger to the performance of others. Responsibility of the diplomatic actor for a negative outcome tends to prompt anger, while lack of responsibility gives rise to sympathy (Weiner, 2001, p. 7). The emotions used in Attribution theory, anger and sympathy, can also be used in crisis communication. Crisis responsibility is related to the affect created by the crisis, and stronger perceptions of crisis responsibility will strengthen negative affect, while positive affect will be strengthened when the perception of crisis responsibility is low (Coombs & Holladay, 2005, p. 276). This means that if the attribution of crisis responsibility is low, which is the case for the Victim cluster, sympathy will be the dominating emotion. If attribution of a crisis responsibility is high, which is the case for the Intentional cluster, anger will be the dominating emotion (Coombs, 2007, p. 166).

Schultz et al. (2010) have deepened the theoretical foundation of crisis communication by looking at the effects of traditional and social media on crisis communication and the reactions to this crisis communication. They found that the medium is more important than the actual message, and although people talk more about newspaper articles, crisis communication via Facebook and Twitter lead to a higher reputation of the organization than

(9)

crisis communication via newspapers. When the organization uses social media to inform its public about the crisis, it shows that the organization is “eager to inform its stakeholders quickly and directly and that it is willing to engage in a dialogue with them” (Utz et al., 2013, p. 41). In a recent study, Utz et al. (2013) used these conclusions and also incorporated crisis type to see the effect on secondary crisis communication. Secondary crisis communication could be either word-of-mouth, or when looking at social media, “the willingness to forward a message and to leave a comment or reaction” (Schultz et al., 2010, p. 23) The authors used the Fukishima Daiichi nuclear disaster as the crisis scenario, and found that when this crisis was framed as intentional, where the Japanese company that managed the nuclear power plants was held responsible, it led to more negative secondary crisis communication than when the crisis was framed as a victim crisis. This confirms that the crisis type will most likely have an influence on the reaction of the public on Twitter messages from diplomatic actors.

These considerations lead to the following hypotheses:

H1: When the crisis is framed to be part of the Victim cluster, it is more likely that the tone of the secondary crisis communication will be mostly positive.

H2: When the crisis is framed to be part of the Preventable cluster, it is more likely that the tone of the secondary crisis communication will be mostly negative.

H3: When the crisis is framed to be part of the Victim cluster, it is more likely that the dominating emotion of the secondary crisis communication will be sympathy.

H4: When the crisis is framed to be part of the Preventable cluster, it is more likely that the dominating emotion of the secondary crisis communication will be anger.

Research Design and Methodology

Categories of the independent variable

The independent variable for this study is the attribution of crisis responsibility. The categories used to measure this variable are based on two of the crisis clusters used by Coombs (2007). Since the crisis clusters used by Coombs are mainly based on crisis types that occur within businesses and not in politics, the accidental cluster will be left out. This accidental crisis cluster assumes that the actor has performed actions that lead to the crisis, but these actions were unintentional and the attribution of responsibility will therefore be low. It is very hard to find cases that fit the criteria, since crises that take place in world politics usually have a rich history with many factors that could have contributed to the emergence of a crisis. Most crises in world politics fall in either the Victim cluster or the Preventable cluster

(10)

and therefore only these two clusters will be used. The study of Utz et al. (2013), where SCCT was applied to crisis communication on social media, also only made use of these two clusters.

A crisis will fall in the Victim cluster if the diplomatic actor is a victim of the crisis and is in no way responsible for the crisis. The diplomatic actor does not have to be directly involved, as long as a large number of citizens of the country is affected. This also means that the crisis can happen anywhere in the world, but citizens of the country need to be a victim of the crisis. Since the diplomatic actor is not responsible for causing the crisis, the attribution of crisis responsibility will be low. A crisis will be part of the Preventable cluster when the diplomatic actor “knowingly placed people at risk, took inappropriate actions or violated law/regulation” (Coombs, 2007, p. 168). When the diplomatic actor is responsible for causing the crisis, and also intended the crisis to happen, the crisis will fall in this cluster. This means that there will be a strong attribution of crisis responsibility by the public.

Case selection

The case selection is based on the two crisis clusters. In total four cases, or crises, will be used. The first two cases are similar in two ways. In both of these crises Russia was somehow involved and in both cases a plane crashed in circumstances where technical issues did not play a role.

The first crisis is the crash of Metrojet Flight 9268 on 31 October 2015. This international passenger flight was on its way from Sharm el-Sheikh, Egypt to Saint Petersberg, Russia. The plane crashed in northern Sinai, Egypt after a bomb on board exploded, which killed all 224 people on board, of which 217 were Russian holidaymakers. Shortly after the crash, the Sinai branch of Islamic State (IS) claimed responsibility and further investigation led to the conclusion that terrorism was the cause of the crash. This crisis will therefore be part of the Victim cluster, since Russia had no responsibility in causing the crisis and mostly Russian holidaymakers were killed in the crash.

The second crisis is the crash of MH17 on 17 July 2014. The Malaysian Airlines flight was on its way from Amsterdam to Kuala Lumpur when it crashed in Ukraine after it was shot down and killed all 298 passengers. The plane was shot down at the height of the conflict between Ukrainian government troops and pro-Russian separatists and this crisis was fuelled by Russian support for the pro-Russian separatists, including support through arms, materiel and training (Murray, 2014). Several Western officials and the Ukrainian government claimed that the missile that shot down the plane was brought from Russia and was launched from a

(11)

rebel-held part in eastern Ukraine. The leader of the separatists, Russian colonel Igor Girkin, claimed responsibility for shooting down a military aircraft after the crash on a Russian social networking site, but took the post down after it became clear that it was a civilian plane. Russian officials have rejected these claims and have claimed instead that a Ukrainian missile brought down the plane. Nonetheless, most people still hold Russia responsible for the crash, since most evidence points towards Russia. Therefore, this crisis will fall in the Preventable cluster.

The third crisis that will be used is the bombing during the Boston Marathon on 15 April 2013. During the marathon, two bombs exploded near the finish line and killed 3 people and injured 264 others. The bombings were qualified as a terrorist attack by the United States, and two Chechen brothers Dzhokhar Tsarnaev and Tamerlan Tsarnaev were held responsible for the bombings. One of the perpetrators later said in a note that “the bombings were retribution for U.S. military action in Afghanistan and Iraq” (CBS News, 2013). Although this could mean that the public might hold the United States responsible, the way the government frames the crisis plays an important role in the attribution of responsibility (Canel & Sanders, 2010, p. 452). Since the crisis was framed with the United States being the victim of the terrorists, most of the public will attribute responsibility to the terrorists. Most of the public that follows US Twitter accounts will also identify with the victim and/or victim group and not with the perpetrators who killed innocent people, which in turn strengthens the attribution of responsibility to the perpetrator (Doosje et al., 2007, p. 139). Considering this and the fact that the United States had little controllability in preventing the crisis from happening, the crisis will be part of the Victim cluster since the attribution of responsibility to the United States will be low.

The fourth crisis will be the launch of Operation Protective Edge, an ongoing airstrike assault and ground invasion launched by Israel on 8 July 2014. This operation was a response to Hamas launching rockets towards Israel. During this operation, which took place between 8 July and 27 August, over 2100 Palestinians were killed and 8300 injured in the Gaza strip, with the vast majority being civilian. While it is open for discussion if Israel had the right to defend itself, clear is that Israel is responsible for launching an ongoing airstrike that could lead to the deaths of civilians. Therefore, this crisis will fall in the Preventable cluster, since Israel was responsible for this escalation of the Israel-Gaza conflict and had the control over the outcome of the crisis, since Israel could have chosen not to launch the airstrike.

(12)

Data collection and sampling

The main unit of analysis will be tweets about the crisis. To collect the data, Twiplomacy (http://twiplomacy.com) was consulted to identify which diplomatic actors held verified Twitter accounts. Twiplomacy is a global study of diplomatic actors on Twitter, and identifies to which extent these diplomatic actors use Twitter. If a diplomatic actor holds a verified Twitter account, this means that Twitter has verified the authenticity of the identity of this diplomatic actor. The verified accounts used for the crisis where Russia was involved, are the ministry of foreign affairs of Russia (@mfa_russia), the English account of the Russian Prime Minister Dmitry Medvedev (@MedvedevRussiaE) and the English account of the Presidency of Russia (@KremlimRussia_E). The official account of the Russian president Vladimir Putin is not used, because this account mostly posts tweets containing pictures of meetings with other diplomatic actors and almost nothing about crises. For the crisis where Israel was involved, the accounts of the ministry of foreign affairs (@IsraelMFA), the Office of the Prime Minister of Israel (@IsraeliPM) and former president Shimon Peres (@PresidentPeres) were used. The American accounts used were the U.S. Department of State (@Statedept), the White House (@Whitehouse) and President Obama (@BarackObama),

The extension Web scraper was then used to retrieve the tweets sent by those diplomatic actors during the crisis. The timeframe used for the cases is the day the crisis happened until a week after the start of the crisis. This time frame is chosen since this study is about the initial attribution of crisis responsibility, that is, right after the crisis. To make sure that there was enough data, the time frame was expanded by two weeks for the Russian cases, since the frequency of tweets per day was relatively low. Since the Israeli accounts each had an average of 10+ tweets per day, a random sample of 10 tweets for each account was conducted to make sure that the number of analysed tweets for each case laid between 20 and 30.

Only tweets that had content about the crisis were used. To measure the dependent variable, namely the secondary crisis communication, replies to the tweets that contain content about the crisis were analysed. The replies to a tweet can be found by clicking on this tweet, and the replies will appear underneath this tweet. Twitter users can also reply to these replies to the original tweet, and a separate conversation will start in the reply section. Replies that only retweet the content of the original tweet without adding extra text and replies that contain links were left out. Furthermore, only replies that were in English were used. Although this could be seen as a limitation, during the research it turned out that most of the replies (over 80%) were written in English.

(13)

Data coding and analysis

Replies were then coded “positive”, “negative” or “neutral” based on the tone explicitly directed to the diplomatic actor who sent the tweet or to the country he represents. For example, when a reply praises the actions taken after the crisis by the diplomatic actor, or expresses concern for the victims, it is coded as positive. When the reply accuses the diplomatic actor of being responsible or compares the crisis with other foreign policy mistakes, it is coded as negative. If the reply has nothing to do with the crisis, does not blame anyone or is directed at someone else but just @-mentions the diplomatic actor, the reply is coded as neutral. For the complete coding rules, see the codebook in the appendix.

According to the Attribution theory and SCCT, the public will also experience a discrete emotion to the crisis, which could be either anger or sympathy. These emotions should therefore also be notable in the replies send by the public and were also coded. Anger and sympathy differ in their valences, yet they are likely to be experienced by publics who are not a victim of the crisis (Weiner, 2001; Jin, 2014). As noted, anger is felt among non-victims when the perception of crisis responsibility increases and when the actor is held responsible for the suffering (Coombs & Holladay, 2005, p. 277). Sympathy is felt “among non-victims when exposed to the description of a crisis they do not directly involve but in which victims’ suffering is witnessed and depicted” (Jin, 2014, p. 82). Replies will therefore be coded as being dominated by the emotion “anger”, “sympathy” or “else”. Anger is present when a reply contains words like “blood on your hands, murderers, killed”, any cursing words, or when capital letters and/or exclamation marks are used. Sympathy is present when a reply contains words like “sorry, condolences, praying, support”. If a reply does not contain anger or sympathy, it will be labelled as “else”. For the complete coding rules, see the codebook in the appendix.

T-test and regression analysis

To see if there is a difference in means in tone and emotion based on the the crisis clusters, independent t-tests were conducted. An additional t-test with only the Russian cases was also conducted, since it is interesting to see what the results are when the country and type of crisis are similar. The dependent variable tone is used to test hypotheses 1 and 2, and will range between -1 (perfectly negative) and 1 (perfectly positive). The dependent variable emotion is used to test hypotheses 3 and 4, and will range between -1 (angry) and 1 (sympathetic).

A linear regression model was also used to measure the impact of the crisis cluster on the tone and emotion of the secondary crisis communication (both continuous variables). A

(14)

number of control variables that are also expected to have an impact on the dependent variables were added. The first control variable is the number of days, a continuous variable that indicates how many days after the crisis happened the tweet was send, with a tweet send on the date of the crisis being 0. The second control variable is the type of account, a dichotomous variable consisting of the account of the head of state, an account from the ministry of foreign affairs and a governmental account, with the governmental account as the reference category. The third control variable is country, a dichotomous variable consisting of Russia, the US and Israel, with the US as the reference category.

Findings

Quantitative analysis

Overall, 2280 reactions in response to 100 tweets were analysed, of which 1212 were replies to the cases in the Preventable cluster and 1068 were replies to the cases in the Victim cluster. 60% of the replies to the cases that fell in the Victim cluster was positive, while only 19% was negative. Of the replies that fell in the Preventable cluster, 61% was negative while 14% was positive (see figure 1). When looking at the dominating emotion in the replies, 37% of the replies to the victim cases contained sympathy, while 5% contained anger. For the preventable cases, 37% of the replies to the preventable cases contained anger, while 4% contained sympathy (see figure 2). However, most of the replies fall in the “else” category, with respectively 58% (victim) en 59% (preventable) and did not contain any sympathy or anger.

Figure 1: Tone of the secondary crisis communication

14% 61% 25% 60% 19% 21%

(15)

Figure 2: Dominating emotion of the secondary crisis communication

To see if there is a statistically significant difference (p<0,05) between the means of the two types of crisis clusters, an independent-samples t-test was conducted. The results can be found in table 1. The difference in the average of the tone for the two clusters is evident. The average tone for the Victim cluster was M=0,311 and therefore mostly positive, while the average tone for the Preventable cluster was M= -0,476 and therefore mostly negative. The difference between the mean for the Victim cluster and the mean for the Preventable cluster is statistically significant at the chosen 0,05 level (t (74) = -10, p<0,001). The difference in the average of the dominating emotion is also evident, although the means are lower than the difference in the average tone score. The average for the Victim cluster is M=0,267 and the dominating emotion is therefore mostly sympathetic, while the average for the Preventable cluster is M= -0,297 and the dominating emotion is therefore mostly angry. The difference between the two means for emotion is also statistically significant at the 0,05 level (t (68) = -11, p<0,001).

Since the same country was involved in the Russian cases and the crisis type was also similar, a separate t-test with just the Russian cases (n=41) was also conducted (see table 2). The results are similar to the previous results were all cases were included. Although the results are similar, there is one notable difference. The mean for emotion in the Victim cluster is higher (M=0,323) than the mean for tone (M= 0237), while this was not the case when the t-test was conducted for all the cases. This difference can be explained due to the fact that overall, there were more negative replies to the tweets in the Russian Victim case, which lowered the average, but these replies were not angry and therefore did not lower the mean for emotion. The average tone of the replies to tweets in the Russian Preventable case was also more negative (M= -0,537) than the average tone for all the cases combined (M= -0,476). The average for emotion in the Russian Preventable cluster (M= -0,285) is similar to the average

37% 5% 58% 4% 37% 59%

(16)

for emotion for all cases combined (= -0,297). The results for both tone (t (21) = -5, p<0,001) and emotion (t (20)= -6, p<0,001) in the Russian cases were also significant.

Table 1: Independent-samples t-test

Victim (N=45) Preventable (N=55) Mean Standard Deviation Mean Standard Deviation df t Sig. (2-tailed) Tone (N=100) 0,311418 0,443095 -0,475623 0,298892 74,452 -10,171 <0,001 Emotion (N=100) 0,267077 0,306522 -0,297417 0,178591 67,588 -10,929 <0,001

Table 2: Independent-samples t-test (only the Russian cases) Victim (N=17) Preventable (N=24) Mean Standard Deviation Mean Standard Deviation df t Sig. (2-tailed) Tone (N=41) 0,237055 0,565301 -0,536973 0,267511 21,109 -5,245 <0,001 Emotion (N=41) 0,323016 0,424043 -0,284856 0,176738 19,970 -5,577 <0,001

A multiple linear regression analysis with three control variables was also conducted and confirmed these findings. Table 3 provides the regression analysis for tone as the dependent variable, while emotion is the dependent variable in table 4. The crisis cluster has a significant effect (p<0,05) on the tone of the replies. If the crisis falls in the Victim cluster, it is more likely that the tone of the replies is positive (β=0,726, p<0,001). This also means that if the crisis falls in the Preventable cluster, it is more likely that the tone of the replies is negative. These findings confirm hypotheses 1 and 2. Country, type of account and days all had a non-significant effect on tone.

The crisis cluster also has a significant effect on the dominating emotion. The standardized coefficient of the Victim cluster in the emotion model (β=0,830, p<0,001) is even stronger than in the tone model. This also confirms hypothesis 3 and 4, that is, if a crisis falls in the Victim cluster, the dominating emotion is more likely to be sympathy. On the other hand, if a crisis falls in the Preventable cluster, the dominating emotion is more likely to be anger. In

(17)

this model type of account and country both had a non-significant effect. However, the number of days did have a significant effect on the dominating emotion (β= -0,163, p<0,05).

Table 3: Coefficients linear regression model tone (r2= 0,55) B (unstandardized) (standardized) β SE Victim 0,783 0,726*** 0,243 Personal -0,158 -0,118 0,117 MFA -0,143 -0,126 0,091 Russia -0,049 0,001 0,120 Israel 0,074 0,064 0,064 #days -0,020 -0,131 0,011 Constant -1,138 -1,138*** 0,243 *** p<0,001

Table 4: Coefficients linear regression model emotion (r2= 0,58) B (unstandardized) β (standardized) SE Victim 0,618 0,830*** 0,078 Personal -0,054 -0,058 0,068 MFA -0,039 -0,050 0,061 Russia 0,157 0,208 0,080 Israel 0,104 0,130 0,103 #days -0,018 -0,163* 0,008 Constant* -0,959 -0,959*** 0,162 * p<0,05 *** p<0,001 Qualitative analysis

To deepen the findings, the replies will also be analysed in a qualitative way. For a full overview of examples of replies, see table 5. Looking at the cases in the Preventable cluster cases, responsibility was already assigned by the public when the first tweet was send in both the MH17 crash and the attack on Gaza. The public is more likely to get information about the crisis through news media, and the public will assign responsibility based on how the crisis is

(18)

framed in the media. The first few tweets from the Russian actors mostly expressed sympathy and condolences, and while a few people replied with their own condolences, most people attributed blame right away. Replies contained words like “murderers”, “you shot down the plane”, “you killed”, “you did it” or variations on this.

The Israeli actors mostly send out tweets where they blamed Hamas for starting the war or that Israel was acting out of self-defence. Most of the replies were negative. These negative replies blamed Israel for killing innocent civilians by launching the airstrikes, for example “you kill innocents”, “tell us how many of your people died by those "rocket fires"?! I tell you how many innocents you killed so far: over 150! Murders!”, “you have no heart. Palestine doesn’t have any army, navy and ur attacking innocent children” [sic]. However, when comparing the amount of negative replies in the Israeli case to the Russian case, there seemed to be less negative replies overall. This could be due to the fact that blame is more likely to be attributed more evenly when the case is about starting a war, since two parties are involved and both are usually somewhat responsible. Nonetheless, the amount of negative replies dominated the amount of positive replies in both cases.

The tone of the tweets of the Israeli actors also seemed to have an influence on the tone of the replies. When the tweet blamed Hamas for being responsible for the war, the tone of the replies was more negative. When the tweet talked about self-defence, more people seemed to reply in a positive way, by agreeing that Israel indeed had the right to defend itself. For example: “Pay them in a similar manner”, “Every country has the right & responsibility to vehemently defend its territory and integrity! We support you”. Comparing this to the Russian case, it seemed that people already attributed responsibility in the Russian case before the tweet was send, and the content of the tweet did not alter the attribution of responsibility. It can not be said with certainty with the available data why this is the case for Israel and not for Russia. This could be due the type of crisis or the country who send the tweet, or other reasons.

Looking at the cases in the Victim cluster, the replies were mostly positive. Most positive responses also expressed sympathy. For example, the Boston case mostly had sympathetic responses, with the hashtags #prayingforboston and #bostonstrong being used in a lot of replies. A lot of people also expressed their condolences, and some mentioned explicitly that they were from a certain country, such as Canada, France, Egypt, India, New-Zealand etc. This makes it clear that responses were send from all over the world, and digital diplomacy does reach a broad public. The Russian plane crash in Egypt also had mostly positive replies that expressed condolences, with people from all around the world

(19)

commenting. For example: “condolences on behalf of the American people”, “#RIP all the victims of the plane crash... horrible tragedy”.

However, both cases in the Victim cluster also had some negative replies. Most of these negative replies in the Russian case compared the plane crash with the crash of MH17, for which Russia is held responsible, by calling it karma or comparing it to other foreign policy of Russia. For example: “Now you know how the Dutch people felt when their people died by Russian Buk-rocket!!!”, “It's a pity it wasn't shot down by either #Malaysia or the #Netherlands - because that would be so fitting”, “Now you need another 364 days for mourning everyone who dies due to Russian aggression in Ukraine, Syria etc.”.

When there were any negative responses to tweets of the US accounts about the Boston case, they were mostly aimed at other foreign policy of the United States, like the drone strikes in the Middle East. Examples of these replies are: “Do your prayers also go to the innocent women and children killed in Pakistan by US Drones?”, “Cool. Maybe you could send some prayers to Pakistan, Yemen, and Afghanistan instead of sending drones.”. One other crisis that was mentioned many times in the negative replies, was the terrorist attack on the American consulate in Benghazi, Libya by Islamic militants. The ambassador and three other Americans were killed during this attack and the United States is held responsible by many people for not providing enough security on the compound. This was also notable in the replies to tweets of the US actors that talked about finding out who was responsible for the Boston bombings. To these type of tweets, a lot of these negative replies compared the investigation of the Boston bombings with the lack of investigation in the Benghazi attack. For example: “Why didn't you treat the Americans that were tortured and murdered in #benghazi with such urgency???”, “Hope you use a better investigation team then you did for the Benghazi cover up”. Overall, a tweet that expressed condolences had more positive and sympathetic replies than tweets that talked about finding perpetrator and/or investigating the crisis for both cases in the Victim cluster. This case also points out that the content of the tweet has an influence on the tone of the replies.

When looking at the emotion expressed in the replies, it turned out that the replies to the Victim cases that were positive also mostly expressed sympathy. Like mentioned above, sympathy was mostly expressed in the form of condolences, prayers or hashtags containing words as “pray for” or “stay strong”. The ways sympathy was expressed was therefore quite limited, and most replies were similar. However, the replies to the Victim cases that were coded as negative did not necessarily express anger. While most of these negative replies criticized the country that send out the tweet in some way, the replies mostly contained other

(20)

emotions such as annoyance, frustration, indifference or suspicion. For example: “Will you move as quickly as you have on Benghazi?”, “Russia lives of "accidents" but even if the character has very dangerous and permission of espionage”.

The replies to the Preventable cases that were negative were also mostly angry. Most replies used either cursing, capital letters, exclamation marks or words like killers or murderers to express their anger. For clear examples see table 5. Although most negative replies also contained anger, some of the replies contained other emotions as the ones mentioned above. However, for a lot of replies that were coded “else”, it is hard to name a specific emotion that was used.

Overall, it seems that most replies to the Victim cases were indeed mostly positive and contained mostly sympathy in the form of condolences or prayers. Replies to the Preventable cases were mostly negative and contained mostly anger in the form of cursing and blaming of killing innocent people.

(21)

Table 5: Examples of tone and emotion expressed in replies

Tone Examples

Positive “@Whitehouse Thank You and May God guide U at this time in all the crucial incisive decisions pending #BostonMarathon”

“@PresidentPeres Americans understand and support Israel!”

“@IsraelMFA Every country has the right & responsibility to vehemently defend its territory n integrity! We support you.”

@KremlinRussia_E Sir We Stand by Russia and D People of Russia. “@Barackobama Thank you Mr. President”

“@KremlinRussia_E Prayers for the horrible event that took place over Egypt. Please stop them from any more attacks!”

Negative “@Presidentperes I really think you don't have a heart, killing innocent people. Reminds me of Hitler” “@Statedept Do your prayers goes to the innocent women and children killed in Pakistan by US Drones?” @KremlinRussia_E You send soldiers, terrorist and armory and it's your fault. Putin you killed them! You are liar! And your media lie too!

“@Whitehouse Why didn't you treat the Americans that were tortured and murdered in #benghazi with such urgency??????”

“@Russian_MFA Your government is responsible for this tragedy”

“@IsraelMFA Why should Palestinians not fight back? they are watching their civilians being killed. #Israel kills children. #FreePalestine”

Emotion Examples

Sympathy “@MedvedevRussia_E My deep condolences to my Russians friends for the plane crash in Sinai, RIP ALL THE VICTIMS”

“@Whitehouse Canadian are saying there prayers too”

“@Barackobama Tonight my prayers are for the USA people and for the Boston Peoples and families. God bless the USA”

“@KremlinRussia_E Condolences to the families of those killed. Expressions of solidarity to the Russian people. This is terrible news. Too sad”

@BarackObama not only american.. Frenchs, Germans... All people... #prayforboston

“@Whitehouse Deepest Condolences from the people of Pakistan, may the perpetrators get caught and are severely punished for this barbaric act”

Anger “@Israelpm Murderer and killer of innocent babies and adults .Every soul shall taste of death and HELL is your abode forever .#cowards”

“@Presidentperes Protect?!ARE YOU SERIOUS? WHAT THE FUCK SHOULD YOU PROTECT?YOU ARE KILLING HUNDREDS OF PEOPLE FOR WHAT? DIE”

“@MedevevRussia_E FUCK you MURDERER”

“@MedevevRussia_E Their blood is on your hands, I hope there's a hell so assholes like you can burn in it forever”

@IsraelMFA fuck off. You're the real terrorists here for killing innocent people and children of Gaza. “@MFA_Russia Excuse me! You blew up the fucking plane!!! #MH17”

(22)

Discussion

Altogether, the data does suggest that the attribution of responsibility has an impact on the way Twitter messages are received by the public. When the attribution of responsibility is high by the public, the tone of the tweets is more likely to be negative and the dominating emotion is more likely to be anger. When the attribution of responsibility is low, the tone of the tweets is more likely to be positive and the dominating emotion is more likely to be sympathetic. Although the results for emotion were significant, it has to be noted that the “else” category had a quite large percentage compared to the emotions used in this research, anger and sympathy. For both the clusters, the “else” category turned out to be almost 60%, which means that most of the replies contain other types of emotion. Further research should look at different types of emotions that are being used in secondary crisis communication. The number of days also turned out to be significant in the emotion model. It should be noted however, that the timeframe for all four cases was not the same, respectively one week for Israel and USA and two weeks for Russia. To see if the number of days really has an effect on how Twitter messages are received, a longer time span should be used.

The content of the tweet during the crisis also seemed to have an impact on the way tweets are received. When a tweet in the Victim cluster expressed condolences, most of the replies seemed to be positive and sympathetic. However, when a tweet in the Victim cluster talked about finding out who was responsible for the crisis, more negative reactions seemed to follow, although most of the replies were still positive. Expressing condolences in a tweet in the Preventable cluster seemed to have little effect on getting more positive replies, while claiming to protecting itself did seem to have an effect on getting more positive replies. Contrary, blaming someone else for the crisis seemed to lead to more negative replies in the Preventable cluster. This suggests that the content of the tweet plays an important role in how people respond to tweets.

Although digital diplomacy seems to become increasingly important for diplomatic actors, it seems that the goal of two-way communication is still a long way to go. When looking at the accounts used in this study, none of the accounts replied to any tweets from the public after the crisis took place. While doing research for the cases, it also became clear that many cases could not be used since a lot accounts had little to no replies, which means there is not even an opportunity for two-way communication. Especially smaller countries and accounts from foreign ministries seemed to lack replies.

This research does have several limitations. The number of cases and tweets is relatively low, with only 4 cases and 100 tweets. The level of subjectivity is also quite high,

(23)

since the coding was only done by one person in a short period of time. The influence of other factors on the attribution of responsibility, such as the history of politics of the country and the way the country handled previous crises, have not been taken into account. The last factor, the way the country handled previous crises, was for example notable in the Russian case. The crash of the plane in Egypt happened after MH17 was shot down, and many people referenced to this in the replies. The way previous crises were handled therefore does seem to have an influence on the way people respond to future crises, even when the country is the victim in the future crisis and should be taken into account.

Despite these limitations, this research showed that the attribution of crisis responsibility does seem to play an important role in the way the public responds to crisis communication. Although Attribution theory and SCCT have not been applied to political science yet, it seems that these theories are indeed useful for political science when it comes to crises. The advantage of this research is that it was conducted online and not in an experimental setting, unlike many other studies that used SCCT and social media. The reactions of the public are therefore completely natural and more valid.

If diplomatic actors are really willing to participate in two-way communication in the digital age, they should consider making use of crisis response strategies to get more positive reactions. Looking back at the Situational Crisis Communication Theory (SCCT), this theory also suggests crisis response strategies to protect the organization against negative reactions. For example, when responsibility is obvious, the actor should apologize or remind the public that he is a victim too. If it is a victim crisis, providing information is usually enough. Although the strategies are mainly for crises that take place in business, they could be adapted for political crises so diplomatic actors can use them. Further research should therefore develop crisis communication strategies for diplomatic actors. Like mentioned before, the content of the tweet also seems to have an impact on the way people react and therefore the content should be based on the crisis response strategies. If diplomatic actors are really interested in digital diplomacy and do not only use SNS for displaying information, they should anticipate on how responsibility will be attributed by the public and use SCCT combined with further research to use the right form of communication. Since it matters what the public says for opening up political exchange and a positive tone will more likely lead to responses from the diplomatic actors (Tromble, under review), countries should make use of these yet to be developed communication strategies to decrease the level of attribution of responsibility so they can reach the full potential of digital diplomacy, which is two-way communication.

(24)

References

Bjola, C. and Jiang, L. (2015). Social media and public diplomacy: a comparative analysis of the digital diplomatic strategies of the EU, US and Japan in China. In Bjola, C., & Holmes, M, Digital Diplomacy: Theory and Practice (71-88). London: Routledge. Canel, M. J., & Sanders, K. (2010). Crisis communication and terrorist attacks: framing a

response to the 2004 Madrid bombings and 2005 London bombings. The handbook of crisis communication, 449-466.

CBS News. (2013). Boston bombings suspect Dzhokhar Tsarnaev left note in boat he hid in, sources say. [online] Cbsnews.com. Available at:

http://www.cbsnews.com/news/boston-bombings-suspect-dzhokhar-tsarnaev-left-note-in-boat-he-hid-in-sources-say/ [Accessed 13 May 2016]

Cercel, M.O., & Saftescu, R.G.. (2015). Digital Diplomacy-Perspectives and Impact on Traditional Diplomatic Practices-Case Study: Digitization Impact on Romanian and Belgian National Diplomatic Systems. International Journal of Social Sciences and Education Research, 1(2), 385-398.

Coombs, W. T. (1995). Choosing the right words the development of guidelines for the selection of the “appropriate” crisis-response strategies. Management Communication Quarterly, 8(4), 447-476.

Coombs, W. T. (2007). Protecting organization reputations during a crisis: The

development and application of situational crisis communication theory. Corporate reputation review, 10(3), 163-176.

Coombs, W.T. and Holladay, S.J. (2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory, Management Communication Quarterly, 16, 165–186.

Coombs, W. T., & Holladay, S. J. (2005). An exploratory study of stakeholder emotions: Affect and crises. Research on emotion in organizations, 1, 263-280.

Doosje, B., Zebel, S., Scheermeijer, M., & Mathyi, P. (2007). Attributions of responsibility for terrorist attacks: The role of group membership and identification. International Journal of Conflict and Violence, 1(2), 127-141.

Gilboa, E. (2006). Public diplomacy: The missing component in Israel's foreign policy. Israel affairs, 12(4), 715-747.

Gilboa, E. (2008). Searching for a theory of public diplomacy. The annals of the American academy of political and social science, 616(1), 55-77.

Hagan, L. M. (2007). For reputation’s sake: Managing crisis communication. In E. Toth (Ed.), The Future of Excellence in Public Relations and Communication Management (pp. 413–440). Mahwah, NJ: Lawrence Erlbaum Associates.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hocking, B. L., & Melissen, J. (2015). Diplomacy in the Digital Age. Clingendael,

Netherlands Institute of International Relations.

Holmes, M. (2015). Digital Diplomacy and International Change Management. In Bjola, C., & Holmes, M, Digital Diplomacy: Theory and Practice (13-32). London: Routledge.

Jin, Y. (2014). Examining publics' crisis responses according to different shades of anger and sympathy. Journal of Public Relations Research, 26(1), 79-101.

Jin, Y., Liu, B. F., & Austin, L. L. (2011). Examining the role of social media in effective crisis management: The effects of crisis origin, information form, and source on publics’ crisis responses. Communication Research, 0093650211423918.

Jones, E. E. and Nisbett R.E. (1972). The actor and the observer: divergent perceptions of the causes of behavior. Morristown, N.J.: General Learning Press.

(25)

Comparison of Public Engagement in Facebook and Twitter. The Hague Journal of Diplomacy, 10(4), 331-362.

Manor, I. (2015). Digital Diplomacy during #NepalEarthquake. Exploring Digital Diplomacy. Retrieved from http://digdipblog.com/2015/05/21/digital-diplomacy-during-nepalearthquake/

Melissen, J. (2005). The new public diplomacy: Between theory and practice. In The new public diplomacy (3-27). Palgrave Macmillan UK.

Murray, W. (2014). MH17: Hillary Clinton says Russian-backed rebels likely shot down plane. [online] the Guardian. Available at:

http://www.theguardian.com/world/2014/jul/18/mh17-hillary-clinton-says-russian-backed-rebels-likely-shot-down-plane [Accessed 13 May 2016].

Nye, J. S. (2004). Soft power: The means to success in world politics. PublicAffairs. Nye, J. S. (2008). Public diplomacy and soft power. The annals of the American academy

of political and social science, 616(1), 94-109.

Nye, J. S. (2010). The new public diplomacy. Project Syndicate, 10.

Schultz, F., Utz, S., & Göritz, A. (2010). Is the medium the message? Perceptions of and reactions to crisis communication via Twitter, blogs and traditional media. Public relations review, 37(1), 20-27.

Sirin, C. V., & Villalobos, J. D. (2011). Where does the buck stop? Applying attribution theory to examine public appraisals of the president. Presidential Studies Quarterly, 41(2), 334-357.

Sotiriu, S. (2015). Digital diplomacy: between promises and reality. In Bjola, C., & Holmes, M, Digital Diplomacy: Theory and Practice (33-51). London: Routledge. Tromble, R. (under review). “Thanks for (Actually) Responding! How Citizen Demand

Shapes Politician’s Interactive Practices on Twitter”. Twiplomacy. (2015). Twiplomacy. Retrieved 15 May 2016, from

http://twiplomacy.com/blog/twiplomacy-study-2015/

Utz, S., Schultz, F., & Glocka, S. (2013). Crisis communication online: How medium, crisis type and emotions affected public reactions in the Fukushima Daiichi nuclear disaster. Public Relations Review, 39(1), 40-46.

Wang, J. (2006). Managing national reputation and international relations in the global era: Public diplomacy revisited. Public Relations Review, 32(2), 91-96.

Weiner, B. (1985) ‘An attributional theory of achievement motivation and emotion’, Psychology Review, 92, 548–573.

Weiner, B. (1995). Judgments of responsibility: A foundation for a theory of social conduct. Guilford Press.

Weiner, B. (2001). Intrapersonal and interpersonal theories of motivation from an attribution perspective. In Student motivation (pp. 17-30). Springer US. Zaharna, R. S. (2004). Asymmetry of Cultural Styles and the Unintended

Consequences of Crisis Public Diplomacy. Intercultural Communication and Diplomacy, 133-155.

(26)

Appendix

Codebook

The following variables should be coded for each reply:

- Tone – either positive, negative or neutral.

This regards the tone the reply takes toward the diplomatic actor itself, or to the country the diplomatic actor represents.

-­‐‑   Positive:

§   The reply does not blame the diplomatic actor for the crisis or someone else is blamed for the crisis.

§   The reply asks to find the perpetrators or the truth about the cause of the crisis. §   The reply is defending the country the diplomatic actor represents

§   When the diplomatic actor is thanked in the reply. §   When sympathy is expressed in the reply (see below). -­‐‑   Negative

§   The reply blames the diplomatic actor or country for causing the crisis, or when the tweet mentions that the diplomatic actor is guilty or responsible. This includes blaming Russia for supplying arms to the Russian rebels.

§   The reply compares the current crisis to a crisis that took place in the past where the country is held responsible for. For example, if the Russian plane crash in Egypt is compared with the crash of MH17, or if the Boston bombings are compared to Benghazi or drone attacks in the Middle East.

§   The reply accuses the diplomatic actor of telling lies or does not believe the sympathy the diplomatic actors expresses in his tweet.

§   When anger is expressed in the reply (see below). -­‐‑   Neutral

§   The reply is about another crisis or other foreign policy, and does not compare this with the current crisis.

§   The reply is not blaming anyone, is not sure who to blame yet or is just asking if the country is responsible.

§   The reply only mentions (non)actions of another country. For example, the reply mentions President Obama not expressing his condolences in the Russian plane crash, or just mentions the electoral fraud of the president in Venezuela.

(27)

- Emotion – either sympathy, anger or else.

This regards the dominating emotion in the reply toward the diplomatic actor itself, or to the country the diplomatic actor represents.

-­‐‑   Sympathy

§   The following words express sympathy:

•   “Condolences” (or thanking the diplomatic actor for his condolences) •   “I’m sorry”/”Mourning”

•   “Rest in peace”/”RIP”

•   “My thoughts are with”/”my heart goes out to”/”heartbroken” •   “Standing with”

•   “Praying for”/#prayingforboston/#bostonstrong •   A sad/heart smiley

-­‐‑   Anger

§   The following words express anger:

•   Any cursing (“fucking”, “asshole”, “dick”, “cunt”, etc.) •   “Blood is on your hands”

•   ”Muderers”/”you killed”/”you terrorists” •   “Hate you”

•   (In combination with) the use of exclamation marks and capital letters •   Angry smiley

-­‐‑   Else

Referenties

GERELATEERDE DOCUMENTEN

In the meantime, a textbook on Alevism has been published in Germany, co-pro- duced by an immigrant teacher and a promi- nent Alevi writer in Turkey.* As it

The EPP demands a determined application of the new instruments which have been developed in the framework of Common Foreign and Security Policy (CFSP), among which are recourse

The first secondary research question is: “Which variables positively affect the actual demand for public transport services among seniors in the rural area of the municipalities

Organising the process of writing a response to reviewers’ comments and making best use of the expertise of your co-authors increases your chances of being successful in getting your

Of note, the intention here is not to debar researchers from engaging in valuable research activities or essential travel but merely to encourage a culture in which we meticu-

examined the effect of message framing (gain vs. loss) and imagery (pleasant vs. unpleasant) on emotions and donation intention of an environmental charity cause.. The

Eisenbl&amp;ldquo;atter, et al., Multispectral optoacoustic tomography of the human breast: characterisation of healthy tissue and malignant lesions using a

We will use a local discontinuous Galerkin (LDG) fi- nite element method to solve systems modeling phase transitions in solids, Van der Waals fluids and the