Politician or party?
The process of personalization in international war coverage
in Dutch broadsheet newspapers.
Master’s Thesis
Master’s program Communication Science
Graduate School of Communication
Lisanne Huisman (10142908)
Supervisor: Y. Lelkes
2 Abstract
Personalization in news coverage has been researched before, but this study adds to the research in that it is about personalization in a specific type of news coverage: international war coverage. Through a quantitative content analysis of Dutch broadsheet newspapers this study tries to answer the question if individual actors are used more over time in war
coverage. For this content analysis articles of Dutch newspapers were used on the Gulf War, Iraq War of 2003 and the war in Syria in 2014 to be able to tell something about a possible increase of personalization over time. This study found no significant results for
personalization over time in war coverage. Future research could look at the process of personalization in war coverage in tabloid newspapers instead of broadsheet or in coverage of other countries that have different political/economic interests.
Introduction
With the picture of a boy found dead on the beach of Greece which was posted in lots of media outlets the refugee crisis in Europe seemed to take on a new stance and the crisis seemed to be rated as more important and receive more attention because of this personal story. If personal stories in the media can have a real effect on attitudes and on the action that is to be taken in a situation, it is therefore important to research whether the media have become more personal or not. Different researchers have focused on the question if news media have become more personal or not, mostly with regard to politics (Oegema & Kleinnijenhuis, 2000; Boumans, Boomgaarden & Vliegenthart, 2013). The personalization theory that has derived from this focuses on the question whether news media give more attention to individual politicians and their private lives than they give to political parties and if this attention to individual politicians has increased over time. This study has used this
3
same theory of personalization, but has applied it in a very different type of coverage: not political elections, but international war coverage. Since international conflicts play a big role in today’s society and politics and newspapers still have an important function of informing the public about these two topics, it is important to study coverage on international conflicts as well. Although in the Netherlands there is evidence that supports the personalization theory for national political news coverage (Kriesi, 2011), this does not mean that it also holds for international war coverage. In this type of coverage it is understandable that journalists have more difficulties doing their job and creating a story if sources are geographically far away and war propaganda is made making it more difficult to find trustworthy sources and to know if well-known sources are still to be trusted. Due to the geographical distance and the different interests it is questionable if the process of personalization also holds for international war coverage in Dutch newspapers. This quantitative content analysis therefore tries to answer this line of thought by looking at the amount of attention that is given to individuals and groups, the frames that are used within the coverage (personal frames or context frames) and the sources that are used. This leads to the following research question:
RQ: How does the news coverage of the Gulf War differ from the Iraq War in 2003 and the war in Syria in 2014 in terms of personalization and use of sources in Dutch broadcast newspapers?
From a political point of view this is an important study to research since interactive and personalized online political communication increases the political involvement of citizens (Kruikemeier, Van Noort, Vliegenthart & De Vreese, 2013). If newspapers or other media become more personalized this therefore can have an influence on the public that has an important position within politics and a democratic society. From the journalistic point of view it is good to know more about possible changes that are occurring in the news coverage, in order to be critical and self-reflective on the articles that are presented in the media. The
4
following sections explain more about the theoretical background of the study and the method after which the results of the study are explained and a conclusion is given.
Theoretical framework
The term personalization has been used by multiple researchers in different ways, which is a cause for confusion in the research on personalization. Brants and Voltmer (2011) have
defined personalization in a very straightforward way in their research, which is useful for this study as well. They define personalization as ‘the relative share of coverage of individual politicians versus political parties’ (Brants & Voltmer, 2011, p. 95). In light of this study it is important to focus on this relative share of coverage. In absolute numbers there does not have to be more coverage on individual politicians than on political parties, but the relative share has to grow in order to say that there is a process of personalization going on in the news coverage. Van Aelst, Sheafer and Stanyer (2011) furthermore also differentiate between two types of personalization in their literature study: individualization and privatization.
Individualization in this case refers to an increase in the amount of news coverage on individual political actors compared to political parties, whereas privatization refers to an increase in the amount of news coverage that is given to the private lives of political actors compared to coverage on political aspects of the individual or party. This study focuses on individualization since the first step is to understand if more attention is given to individual actors rather than group actors.
From the research that has already been done on personalization, an hypothesis can be drawn on whether a process of personalization/individualization can be expected in this study. Vliegenthart (2011) for example has looked at campaign posters between 1946 and 2006 and how they have changed overtime with regard to professionalization. What he found is that party leaders have become more important and are more present on campaign posters
5
nowadays than in the early years. This already hints to some sort of personalization in the Dutch political field in which party leaders have become more important over time in
comparison to political parties. Boumans, Boomgaarden & Vliegenthart (2013) have basically come to the same conclusion within their article on the prominence of Cabinet members in British and Dutch newspapers. Between 1992 and 2007 parties have become less visible in the news coverage compared to individual politicians. However in the Netherlands there is more focus on the non-prominent Cabinet members, whereas in the UK the focus is more on prominent members and the Prime minister (Boumans et al., 2013). This research was conducted over routine political periods, meaning no election periods or crises happened within the periods of news coverage that have been included in the sample of the research. The fact that this research has shown that the process of personalization in political news coverage can be proven in a routine period of political news, makes it very much believable that there really is a process of personalization going on in Dutch news coverage in
newspapers. However, it is interesting that in Dutch newspapers non-prominent Cabinet members have become more important and visible. This shows that the Dutch newspaper landscape offers a wider spectrum of individual politicians than the British newspapers for example that focus mostly on the Prime minister and the most important or prominent Cabinet members (Boumans et al., 2013).
In another study by Vliegenthart, Boomgaarden and Boumans (2011) it was
disconfirmed that a process of personalization is happening in the Dutch newspapers. Though, the method used to determine this personalization might be the cause of that. Vliegenthart et al. (2011) selected articles for their sample based on the presence of a political party that was represented at that time in parliament and not based on topic. This means that all articles refer to main parties, and all other political parties are excluded from the research. Also, the articles included in the sample all ready refer to a political party, which makes it harder to determine
6
if the relative share of attention on individual politicians compared to parties has increased. In the current study this is not the case, since articles are selected for the sample based on the topic of the article and any political party or individual will be recognized and coded in this study. Vliegenthart et al. (2011) thus did not find an increase on focus on individual
politicians, which is a contradiction to the earlier study of Oegema & Kleinnijenhuis (2000) in which it was proven that party leaders play an important role in news coverage in the
Netherlands, however the parties still dominate. Balmas (2012) furthermore found that most research on personalization is mixed however, within unpaid newspapers in the Netherlands there is support for the process of centralized personalization, meaning that there is a stronger focus on party leaders and heads of executives. A study on personalization in Austria,
Germany, France, the Netherlands, Switzerland and the UK found no general trend of personalization (Kriesi 2011). Except for the Netherlands, where support was found for an increase of political personalization in broadsheet newspapers.
Most of the research that has been done on personalization as described above has focused on comparing personalization in multiple countries and within news coverage of national politics and the difference between election periods and non-election periods. This study derives from this research in that it only focuses on one country, the Netherlands, and on international war coverage and not on national politics or elections. However, overall from the studies mentioned above it can be concluded that a process of personalization is
happening in Dutch news coverage. Even though this study focuses on international war coverage, it still uses news articles of Dutch broadsheet newspapers which suggests that the same process of personalization should be going on in this type of coverage. Balmas, Rahat, Shaefer and Shenhav (2012) add to this that newspapers from Canada, Germany, France, Italy, the UK and the U.S. increasingly focus on leaders instead of countries within their international news. Also, international conflicts have a high news value and are covered in the
7
newspapers numerously due to the impacts of the events, the unexpectedness and negativity with which it is usually encompassed (Harcup and O’Neill, 2001). In that regard, it would be logical if this type of coverage is treated like any other topic with a high news value and journalistic routines that lead to more personalization in other types of coverage should therefore also apply to this type. This leads to the following hypothesis:
Hypothesis 1: The relative share of individual actors increases over time (personalization) in news coverage on international conflicts in Dutch broadsheet newspapers.
Not only the type of coverage or the presence of individual actors defines the process of personalization, also the use of certain frames can mark this process. “News frames refer to interpretive structures which set particular events within their broader context, they bundle key concepts, stock phrases and stereotyped images to reinforce certain common ways of interpreting developments.” (Norris, 1995, p. 2). By connecting information within an article to other information or displaying the information from a certain angle, frames are created that can influence the public’s image on the subject. The news coverage of the 2003 Iraq war for example was news for a lot of different news website all over the world. However, the war was framed differently on news websites outside the U.S. than on news websites from the U.S. due to different political and economic interests (Dimitrova, Kaid, Williams, Trammell, 2005). The U.S. websites used a frame that relied more on military conflict while foreign websites used the responsibility frame much more (Dimitrova et al., 2005). This shows how countries can differ in frames used on international conflicts, however the use of frames for different news media is not set in stone and can change over the years. Since a lot of research has been done on the comparison of frames used in conflict coverage between countries (Dimitrova, Connoly-Ahern, 2007; Camaj, 2010; De Vreese, Peter & Semetko, 2001), for this study the use of frames in international conflict news coverage will only be used to see if
8
there are changes over time with regard to personalization. The frames that this study will look at are therefore whether the article uses an episodic frame and/ or a thematic frame (Iyengar, 1991). “A thematic frame places an issue in some general context and usually takes the form of an in-depth backgrounder report. […] Episodic framing, on the other hand, depicts issues in terms of individual instances or specific events.”(Iyengar, 2005, p. 6). Episodic framing in this regard would focus more on individuals and could therefore be regarded as personalization, whereas a thematic frame is more objective and uses more background to create an article in which multiple sides are proposed and the reader can make up his own thoughts about the matter. Iyengar (2005) mentions in his article that episodic framing is the most used in the U.S. which therefore could be expected to be true for Dutch news coverage as well, since the U.S. and the Netherlands are both western countries and the press freedom and professionalization of journalism in both countries is relatively high. In line of the personalization theory it is not only expected that episodic frames are used more than thematic frames in the war news coverage, but also that the relative share of episodic frames increases. This leads to the following hypotheses:
Hypothesis 2: Episodic frames are used more often than thematic frames in news coverage on international conflicts in Dutch broadsheet newspapers.
Hypothesis 3: The use of episodic frames increases over time (personalization) in news coverage on international conflicts in Dutch broadsheet newspapers.
An important distinction that has to be kept in mind while focusing on the coverage of newspapers and not on other media like television is that newspapers have shown to be more independent of the government than television that showed more support to official sources like the government in the U.S. (Peer & Chestnut, 1995). Although this might have changed over time and it might also be different in the media landscape of the Netherlands, the results
9
of this study on the use of sources will make this more clear. The use of domestic sources, international sources, official sources or citizens as sources will show how independent Dutch newspapers are and how much a process of personalization might come from domestic and elite sources for example or from international citizens sources. Due to the rise of the internet and the growing technological possibilities to stay in contact with people all over the world, journalists might have a broader network to get to sources and increase their knowledge when writing the articles. Lecheler and Kruikemeier (2015) explain in their article that the
possibility of online sources has changed the relationship between journalists and their sources. Even though the news agenda in Dutch news coverage up to now has not changed because of online sources, the articles themselves and the stories told in the articles do change (Lecheler & Kruikemeier, 2015) and this might have an impact on the amount of
personalization found in the coverage too. Personalization could be dependent upon the use of sources. If a journalist relies more on individual political sources than official government sources for example it is highly likely that the coverage will be more personalized too since the journalist has to refer to an individual politician instead of a political party.
In the book ‘Reporting War’ Stuart Allan and Barbie Zelizer (2004) describe how news agencies perceive sources. Sources outside Europe were perceived as less reliable during the Iraq War for both the U.S. and the Arab news agencies (Allan & Zelizer, 2004). If this is true for Dutch media as well for multiple international conflicts, this could alter the outcome of this study. One journalist has proven on Twitter to rely more on alternative sources than elite sources (Hermida, Lewis & Zamith, 2014), however in traditional media as newspapers it is more likely to presume that journalists still rely more on the elite sources and their traditional media routines to retrieve information from sources. Broersma and Graham (2013) support this theory as well as Moon and Hadley (2014), that even though Twitter is used a lot more in newspapers nowadays as a source just like weblogs (Messner & Distaso,
10
2008), primarily it is still the Twitter accounts of official sources that make it into the news and not the popular accounts of citizens. The assumption will therefore be that within Dutch news coverage on international conflicts journalists will nowadays rely more on online sources, but still more on elite sources than alternative sources. This leads to the following hypothesis and sub-research question:
Hypothesis 4: Internet sources are used more over time in news coverage on international conflicts in Dutch broadsheet newspapers.
Sub-research question: What type of sources (elite vs. non-elite) are used most in the news coverage on international conflicts in Dutch broadsheet newspapers.
Method
The method of this study is a quantitative content analysis with which is determined whether personalization occurs in international war coverage in Dutch newspapers and which sources are referred to in the coverage. A quantitative content analysis is used because this method makes it possible to compare lots of articles with each other and to draw objective conclusions from that through the use of statistical analyses. Because all articles are coded the same a big part of the population can be included in the sample and this method therefore gives the opportunity to get clear conclusions that can be generalized to the entire population of the study.
Selection of research units
The population for the content analysis exists of all articles found in Dutch quality
newspapers on the three international conflicts used in this study in specific time periods. To create a sample for the content analysis LexisNexis is used to find all relevant articles. Multiple keywords have been used to find all articles about the three different conflicts (table
11
2). The two newspapers used for news coverage are Trouw and NRC Handelsblad. First of all is chosen for Dutch newspapers since it is expected to find positive results to the research question in the Netherlands, which can be a first step to understanding international conflict news coverage in multiple western democratic countries. These two Dutch newspapers can also be regarded as broadsheet or quality newspapers. Quality newspapers are chosen since tabloid newspapers usually have stronger personalization than broadsheet newspapers (Kriesi, 2011), therefore if this process can be found in quality newspapers then it is easier to presume this also is the case for tabloid newspapers. The reason that Trouw and NRC Handelsblad are chosen to represent the quality newspapers in the Netherlands is because they have the highest circulation rates in the Netherlands, they are national newspapers and they are described as quality newspapers amongst most researchers. De Volkskrant is also a national quality newspaper with high circulation rates, but articles from this newspaper were not available in LexisNexis until 1995 which made it hard to include this paper in the analysis.
The three international conflicts that are used as the topic of the articles are Gulf war 1990-1991, the Iraq War 2003 and the war in Syria of 2014-2015. These conflicts are chosen based on a most-similar design with regard to their geographical proximity, rate of importance within the world politics, duration of the war, the time periods and involvement of the Dutch. Table 1 shows the characteristics of the three wars.
Table 1
Characteristics of wars used for news coverage
Conflict Gulf War Iraq War Syria War
Geographical proximity to the Netherlands
Middle East (Iraq) Middle East (Iraq) Middle East (Syria)
12
type Iraq cause, then
invasion of 34 countries initiated by U.S. in Security Council of UN other Western countries to topple government Saddam Hussein international support from Western countries for the rebels in Syria
Involvement of NL Yes, though small contribution
No, not at the time of the invasion
No, not in the time period under study
Duration war August 1990 – January 1991
Invasion from March-May 2003, war continued until 2011
Uprisings started in March 2011, war still going on
Time period used in this study
Aug 1990- Jan 1991 March 2003 – Aug 2003
September 2014 – February 2015
As is clear in table 1, all three wars were important in world politics, though the Netherlands is quite distanced from all three wars, geographically the wars are all in the Middle East and far away from the Netherlands and the support given by the Netherlands in the form of military support or other physical support is in the periods under research small or does not exist at all. For this study a time period of 6 months per war has been chosen, since that is how long the Gulf war lasted and the period is long enough to be able to compare it with the other wars and newspapers and draw conclusions from it. Articles from the beginning of the invasion in Iraq up till 6 months later will be used since it was an international conflict at that time, but there was no involvement from the Netherlands. For the Syria war is chosen to use the time period of 6 months starting from the moment the U.S. started to strike targets and the war became an international conflict, since this would even all wars out in matter of
13
importance in world politics and at that point there was no Dutch involvement yet. Another reason that these wars are chosen for this study is because they started all about ten years apart from each other, making it the perfect opportunity to conduct a longitudinal study. In LexisNexis however the articles from Trouw do not date back to 1990, therefore a time period shortly later is used (first six months of 1992) for articles about Gulf War in this newspaper. Even though internal validity is less strong due to the use of these different time frames for the Gulf War, the articles are still about the same topic and the time frames are close enough to be able to compare them with the other wars and have around 10 years in between each war.
Table 2
Search terms and time periods used in LexisNexis
NRC Handelsblad Trouw
Gulf War Golfoorlog OR Irak OR Koeweit (From 1st August 1990 – 31st January 1991)
Golfoorlog
(From 1st January 1992 – 30th June 1992)
Iraq War Irak AND oorlog
(From 1st March 2003- 31st August 2003)
Irak AND oorlog
(From 1st March 2003- 31st August 2003)
Syria Syrië AND oorlog
(From 1st September 2014 – 27th February 2015)
Syrië AND oorlog
(From 1st September 2014 – 27th February 2015)
14
Variables
The variables that are used in this study are: individual actors headline, group actors headline, individual actors article, group actors article, role of actors, episodic frame, thematic frame and internet source. Personalization is measured through the first four variables. These
variables all count the number and type of actors used in the headline as well as the rest of the article to determine objectively if there is an increase over time in the amount of actors used in the articles. Individual actors headline and group actors headline are dichotomous and only answer the question whether an individual or group actor is present in the headline or not. Individual actors article and group actors article are both interval variables that count the amount of actors used in the entire article excluding the headline and can range from 0-5. If more than 5 actors are present in the article it is coded as 5 actors like Adriaansen and Bos did in their project on the campaigns of the elections for the House of Representatives in the Netherlands in 2006 in cooperation with the University of Amsterdam. For each actor it was further coded what the role of that actor is to be able to see if more elite actors are used in news coverage or non-elite actors. Episodic frame and thematic frame are also dichotomous and answer the question if the frame is present in the article or not. These questions for these frames are derived from the research by Iyengar (2005). Internet source is also dichotomous and answers the question if there is a reference in the article to the internet as a source. More specifications on the variables can be found in the codebook (appendix 1).
End Sample
The coding process resulted in a total of 431 coded articles of which the division per war and newspaper can be found in table 3. All relevant articles were coded for the time periods of the newspapers that did not have over a 100 relevant articles. For the coverage on wars in both newspapers that had more than a 100 relevant articles, stratified sampling was used.
15
Depending on the total amount of articles every nth article was coded to come to a sample of 100 articles per newspaper per war. The articles were first arranged based on date in order to have the entire range of the time period represented in the sample. For the news coverage on the Gulf War in the Trouw only 21 articles were coded, since there was no coverage of Trouw available on LexisNexis within the time period that the Gulf War took place. However, taken all articles together, there are similar amounts of coverage available for analysis on all three wars. Table 4 shows the characteristics of the variables of the sample together with the inter-coder reliability (Krippendorff’s Alpha).
Table 3
Amount of articles in sample per war and newspaper
Trouw NRC Handelsblad Total
Gulf War 21 100 121 Iraq War 100 86 186 Syria War 51 73 124 Total 172 259 431 Table 4
Characteristics variables with inter-coder reliability.
N M SD Min. - Max. KALPHA
Individual actors headline 431 .27 .44 0-1 .95
16
Individual actors article 431 2.28 1.57 0-5 .89
Group actors article 431 1.52 1.30 0-5 .86
Episodic frame 431 .63 .48 0-1 .83
Thematic frame 431 .45 .50 0-1 .89
Internet source 431 .04 .21 0-1 .99
Results
Personalization
To determine whether there is a process of personalization going on in Dutch broadsheet news coverage on conflicts, analyses are done on the use of individual and group actors in the headlines of the articles and the rest of the articles. Table 5 shows the percentages of the amount of articles that have used individual actors and group actors in the headline.
Table 5
Amount of actors per war in percentages
Gulf War Iraq War Syria War
Individual actor Yes 25.6 29.6 25
No 74.4 70.4 75
Group actor Yes 21.5 17.7 51.6
No 78.5 82.3 48.4
From table 5 it is clear that most headlines do not contain an individual actor and this does not appear to change much over time. Group actors also do not seem to make it a lot in the
17
headlines, however in the latest news coverage, the coverage on the Syria war, there has been a rise in the use of group actors in the headline of an article. A One-Way analysis of variance has been conducted on the use of individual actors and group actors over time and there is indeed a significant effect found on the use of group actors in the headlines of war coverage over time, F(2, 428) = 25.54, p < .001, η² = .11. However, Levene’s test of homogeneity of variances shows that the different groups of the variable wars do not have equal variances with regard to the group actor variable (L(2, 428) = 39.64, p < .001) and therefore the ANOVA test is not as reliable as it should be. The posthoc test Bonferroni shows that the significant differences of the ANOVA test are between the articles on the Syria war and the other two. The headlines of the articles on the Syria war have significantly made more use of group actors than articles on both the Gulf War and the Iraq war, though these results as said are slightly unreliable. On the other hand the analyses also show that the variances within the coverage of the different wars with regard to individual actors can be perceived as equal (L(2, 428) = 1.92, p = .148) indicating that the ANOVA for this variable is reliable F(2, 428) = 0.49, p = .613. This means that there has not been a significant increase in the use of
individual actors in the headlines of news articles on war. This shows that the personalization theory does not hold for Dutch coverage in broadsheet newspapers. Both Trouw and NRC Handelsblad support the results mentioned above on all news coverage and they do not show any interesting differences.
When looking at the actors used in the rest of the article for both individual actors and group actors there are significant differences found in the news coverage between the three wars. However, there is not a clear line in these differences. The table in appendix 2 shows the percentages of the division of amount of actors found in the bodies of the articles. One thing that can be concluded from the table is that there are more individual actors found in the bodies of the articles than group actors. Though, when looking at the use of individual and
18
group actors separately there is no clear increase or decrease of the use of actors over time. An ANOVA test shows that both variables significantly differ with regard to the three wars (F(2, 428) = 11.00, p < .001, η²= .05; F(2, 428) = 18.72, p < .001, η² = .08). A Bonferroni test on the Analysis of Variance shows that the news coverage on the Iraq War uses significantly less individual actors than the Gulf War and Syria War (Mdifference = -0.77, p < .001; Mdifference =
-0.61, p = .002). The same goes for group actors that are less used in the coverage on Iraq War than in coverage on Gulf War (Mdifference = -0.85, p < .001) and Syria War (Mdifference =
-0.57, p < .001). These findings are true for Trouw as well as for NRC Handelsblad. No mentionable differences are found between the use of actors in the articles from both
newspapers. Overall, hypothesis 1 is rejected. The relative share of individual actors does not increase over time in news coverage on international conflicts.
Frames
Table 6 shows the percentages of articles that use an episodic and/or a thematic frame in the news coverage per war. The table clearly shows that the episodic frame is used a lot more than the thematic frame, however there is no clear increase or decrease over time in the use of both of these frames, meaning that the use of episodic frames over time does not support the personalization theory. Hypothesis 2 is therefore supported, international war coverage in Dutch broadsheet newspapers make more use of episodic than thematic frames, but hypothesis 3 is rejected. The use of episodic frames does not increase over time in news coverage on international wars in Dutch broadsheet newspapers.
19
Table 6
Articles sorted by frame and war in percentages
Gulf War Iraq War Syria War
Episodic frame Yes 73.6 58.1 58.9
No 26.4 41.9 41.1
Thematic frame Yes 38 49.5 46
No 62 50.5 54
Internet as a source
As could be predicted of all articles about the Gulf War and Iraq War none included a reference to the internet as a source. This is logical since the internet was on the rise during both these wars. Even though in 2003 during the Iraq War it would have been possible that within the news articles the internet was referred to as a source for the article, this is not the case. But even in today’s society in which the internet is used a lot for information, only 19 of the 124 articles (15%) contained a reference to the internet as a source for the article. This implies that journalists are not (yet) relying on the internet to get there information, but still stick to the news routines they have known for a long time. Even war coverage apparently does not make journalists rely on internet websites. However this number might go up in the future. Out of the articles that did contain a reference to the internet, 16 were a reference to social media websites, 2 to other news agencies websites and 1 to a regular website. Social Media in that regard do have an influence in Dutch newspaper war coverage. Hypothesis 4 is supported in that the internet is used more as a source over time, however it is not strongly
20
supported, since even news coverage on Syria did not include a lot of references to the internet as a source.
Type of actors
The sub-research question was about the kind of actors that were used in the articles and whether more elite actors were used than non-elite. All actors defined in the coding process were elite actors except for the actor citizen or group of citizens that can be assumed as a non-elite actor. The total amount of individual actors in all articles used was 982 and the total amount of group actors used was 653. It does have to be noted though that of all articles with more than 5 actors only the first five are coded. The percentages of the different roles that the actors had are shown in graph 1. The graph shows that the articles mad the most references to incumbent politicians for the individual actors and to the government or government
representatives for the group actors. 12% of the individual actors were citizens and 21% of the group actors, meaning that more than 80% of the actors used in the articles were elite actors.
Graph 1
Amount of actors used in articles by role in percentages
0 5 10 15 20 25 30 35 40 45 Individual actor Group actor
21
Graph 2 shows the same percentages but then only for all the actors from the articles on the Syria War to see whether there is more reference to non-elite actors in this time period. However, for both the individual actors and the group actors the non-elite actors, the citizens, still only made up about 19% of all actors present in the articles, 309 and 208 respectively. Thus, also in the news coverage during the Syria War more than 80% of the actors were elite.
Graph 2
Amount of actors used in articles about war in Syria by role in percentages
Conclusion and Discussion
In conclusion the theory of personalization does not hold up for Dutch news coverage on international wars. This study has included news coverage on three wars in the Middle East over ten years apart, but no significant results were found for an increase in the relative share of individual actors in the coverage. The headlines of the articles did not mention a lot of individual or group actors, also not over time, except for the news coverage on the war in
0 5 10 15 20 25 30 35 40 Individual actor Group actor
22
Syria that included a lot of group actors in the headline. This is probably due to the fact that ISIS is an important factor of the war in Syria and as a terrorist group it was counted as a group actor in the coding process, whereas within the coverage on the Gulf War and the Iraq War the headlines mainly referred to the countries involved in these wars which were not coded as actors in this study. In the bodies of the articles there were significantly more individual and group actors used in the coverage on the Gulf War and the Syria War than in the coverage on the Iraq War. This however does not proof any personalization theory, since there is no constant increase of the use of individual actors over time. The episodic frame did prove to be used more than the thematic frame, however also this factor did not prove to increase over time, therefore disproving the personalization theory. Furthermore, the war in Syria was the only war coverage in which was referred to the internet as a source, however this was only in a small portion of articles the case. Therefore it is not really possible to speak of an increase over time in the use of internet sources in war coverage. Lastly, the actors used in the articles of all three wars were mostly still elite actors compared to non-elite actors. Even in articles on the war in Syria there was no increase in the use of non-elite actors which expected a little bit. Journalists and news agencies apparently still rely for the most part on elite sources.
This study also holds a few limitations. As mentioned before, there was one notable difference between the three wars that had an influence on the results of this study, even though this study tried to take upon a most-similar approach with regard to the three wars. The difference is that the Gulf War and the Iraq war were mostly about a fight between different states and their coverage therefore held a lot of references to different countries, whereas the war in Syria was mainly about ISIS which resulted in a lot of references to this terrorist group that were all coded as such. This resulted in the high percentage of group actors used in the headlines of the articles on the war in Syria. Yet, this does not seem to have
23
influenced the outcomes for the individual actors and as such the outcome of the process of personalization. That is, the amount of individual actors did not differ significantly from the amount of individual actors in the other wars and therefore this limitation did not result in possible false findings of individualization.
Another limitation of this study is that for one newspaper, Trouw, another period for the sample of articles for the Gulf War was used as was used for NRC Handelsblad. For NRC Handelsblad the time period was used of the actual war, whereas for Trouw, the time period started one and a half year later. This had as a result that there were a lot fewer articles that could be coded for the Gulf War coverage of Trouw. Although this might have influenced the outcome of the study, the total amount of articles on Gulf War was still a good number that did not differ from the amount of articles coded on the Syria War. At the same time, there were no real differences found between the two newspapers within the rest of the coverage therefore a valid assumption can be made that this did not affect the finding that no
personalization is found over time in war coverage.
Besides the points mentioned above, a shortcoming of this study is that the amount of articles per war were not equal. This could result in an unfair display of results. Yet, this problem was countered by showing percentages in the results which made it easier to compare the groups. For the analysis the fact that the groups were not equal was constantly kept in mind and for the ANOVA’s also Levene’s test of homogeneity of variances was used to see whether the variances within the different groups could be treated as equal or not.
Although there was a clear reason for using only broadsheet newspapers, the results might have been different if tabloid newspapers were included in the sample. Tabloid newspapers are expected to show more signs of personalization and for that reason it is recommended for future research to include tabloid newspapers in the study. That should
24
determine whether there is no personalization found in war coverage within broadsheet newspapers or if there is no process of personalization in war coverage of newspapers at all. In addition to including tabloids, it would be interesting to see what happens if future research would look at pictures used in war coverage within newspapers and maybe even TV. The introduction of this study started off with the example of the picture of a boy that influenced the public debate on the refugee crisis and it would be interesting for future research to have a look at the pictures used in war coverage and see if there are changes with regard to
personalization over time. Pictures might have a very different effect than text.
Another suggestion for future research is to use different wars for news coverage. The wars used in this study were international conflicts in the Middle East, however if different wars are subject of research different interests come in to play that might have an effect on how the war is covered. That can be due to different interests of the country with the dominant actors of the war, economically or politically, but also the extent to which the
country is involved in the war or historical relations between the countries. At the same time it is interesting to do research on the war coverage of different countries. The degree to which a country has press freedom and to which the news agencies are professionalized differ per country, but can be of importance when covering international conflicts. It might be more difficult for instance for a country that has limited press freedom to use controversial sources in the articles that can shine a different light on the situation. Also the media system of a country can have an effect on the way journalists cover wars, since a public broadcaster could have different interests than a commercial one for example and newspapers that have strong ties with certain political parties have different interests than independent newspapers.
25 Reference list
Allan, S. & Zelizer, B. (2004). Reporting war. London: Routledge.
Balmas, M. (2012). When fake news becomes real: Combined exposure to multiple news sources and political attitudes of inefficacy, alienation, and cynicism. Communication
Research, 41(3), 430-454. Doi: 10.1177/0093650212453600
Balmas, M., Rahat, G., Sheafer, T., & Shenhav, S. (2012). Two routes to personalized
politics: Centralized and decentralized personalization. Party Politics, 20(1), 37-51. Doi: 10.1177/1354068811436037
Boumans, J., Boomgaarden, H., & Vliegenthart, R. (2013). Media personalization in context: A cross-national comparison between the UK and the Netherlands, 1992-2007. Political
Studies, 61, 198-216. Doi: 10.1111/1467-9248.12013
Brants, K. & Voltmer, K. (2011). Political communication in postmodern democracy. Basingstoke: Pakgrave Macmillan.
Broersma, M. & Graham, T. (2013). Twitter as a news source. Journalism Practice, 7(4), 446-464. Doi: 10.1080/17512786.2013.802481
Camaj, L. (2010). Media framing through stages of a political discourse: International news agencies' coverage of Kosovo's status negotiations. International Communication
Gazette, 72(7), 635-653. Doi: 10.1177/1748048510378147
De Vreese, C., Peter, J., & Semetko, H. (2001). Framing politics at the launch of the euro: A cross-national comparative study of frames in the news. Political Communication, 18(2), 107-122. Doi: 10.1080/105846001750322934
26
Dimitrova, D. (2005). War on the web: The immediate news framing of Gulf War II. The
Harvard International Journal Of Press/Politics, 10(1), 22-44. Doi:
10.1177/1081180x05275595
Dimitrova, D. & Connolly-Ahern, C. (2007). A tale of two wars: Framing analysis of online news sites in coalition countries and the Arab World during the Iraq War. Howard
Journal Of Communications, 18(2), 153-168. Doi: 10.1080/10646170701309973
Harcup, T. & O'Neill, D. (2001). What is news? Galtung and Ruge revisited. Journalism
Studies, 2(2), 261-280. Doi: 10.1080/14616700118449
Hermida, A., Lewis, S., & Zamith, R. (2014). Sourcing the Arab Spring: A case study of Andy Carvin's sources on Twitter during the Tunisian and Egyptian revolutions. J
Comput-Mediat Comm, 19(3), 479-499. Doi: 10.1111/jcc4.12074
Iyengar, S. (1991). Is anyone responsible?. Chicago: University of Chicago Press.
Iyengar, S. (2005). Speaking of values: The framing of American politics. The Forum, 3(3). Doi: 10.2202/1540-8884.1093
Kriesi, H. (2011). Personalization of national election campaigns. Party Politics, 18(6), 825-844. Doi: 10.1177/1354068810389643
Kruikemeier, S., van Noort, G., Vliegenthart, R., & de Vreese, C. (2013). Unraveling the effects of active and passive forms of political Internet use: Does it affect citizens' political involvement?.New Media & Society, 16(6), 903-920. Doi:
27
Lecheler, S. & Kruikemeier, S. (2015). Re-evaluating journalistic routines in a digital age: A review of research on the use of online sources. New Media & Society, 18(1), 156-171. Doi: 10.1177/1461444815600412
Messner, M. & Distaso, M. (2008). The source cycle. Journalism Studies, 9(3), 447-463. Doi: 10.1080/14616700801999287
Moon, S. & Hadley, P. (2014). Routinizing a new technology in the newsroom: Twitter as a news source in mainstream media. Journal Of Broadcasting & Electronic Media, 58(2), 289-305. Doi: 10.1080/08838151.2014.906435
Norris, P. (1995). The restless searchlight: Network news framing of the post‐Cold War world. Political Communication, 12(4), 357-370. Doi: 10.1080/10584609.1995.9963084
Oegema, D. & Kleinnijenhuis, J. (2000). Personalization in political television news: A 13-wave survey study to assess effects of text and footage. Communications, 25(1). Doi: 10.1515/comm.2000.25.1.43
Peer, L. & Chestnut, B. (1995). Deciphering media independence: The gulf war debate in television and newspaper news. Political Communication, 12(1), 81-95. Doi:
10.1080/10584609.1995.9963056
Van Aelst, P., Sheafer, T., & Stanyer, J. (2011). The personalization of mediated political communication: A review of concepts, operationalizations and key
findings. Journalism, 13(2), 203-220. Doi: 10.1177/1464884911427802
Vliegenthart, R. (2011). The professionalization of political communication? A longitudinal analysis of Dutch election campaign posters. American Behavioral Scientist, 56(2), 135-150. Doi: 10.1177/0002764211419488
28
Vliegenthart, R., Boomgaarden, H.G. and Boumans, J.W. (2011). Changes in political news coverage: Personalisation, conflict and negativity in British and Dutch newspapers. In Brants, K. and Voltmer, K. (Eds), Challenging the Primacy of Politics (pp. 92-110). London: Palgrave.
29 Appendix 1: Codebook
1. In which newspaper is the article published? a. Trouw (1)
b. NRC Handelsblad (2)
2. Date of article. Write in following form (in numbers): dd/mm/yy. 3. Title of the article. Copy complete title from dataset.
4. Amount of words used in article (answer in numbers).
5. Number of unique actors total mentioned in entire article excluding headline.
Actors are references in the text to individuals or groups. To be coded as an actor, the reference has to have the function of a source (recognizable by paraphrasing or citing) or it should be mentioned as a subject in a sentence more than once. There are individual actors (a reference to a specific person, not necessarily referred to by name) and group actors (a reference to anything other than a specific person including reference to two individuals mentioned as one group and a reference to a government as a whole. References to countries should not be coded at all. For this variable all individual and group actors should be counted in the article.
6. Does the headline of the article include an individual actor? a. Yes (1)
b. No (0)
7. Does the headline of the article include a group actor? a. Yes (1)
b. No (0)
8. How many individual actors are mentioned in the article (without headline)? a. 0
30
c. 2 d. 3 e. 4
f. 5 or more
9. [Only show if answer to 8 is not a] What is the role of the first mentioned individual actor? (choose most specific role)
a. Journalist b. Citizen
c. Government representative (includes military generals and mayors and
representatives of United Nations since it is an intergovernmental organization)
d. Incumbent politician e. Other politician
f. Company representative or business worker (only includes elitist professions, a
citizen that is defined as a baker for example should be coded as citizen)
g. Expert or independent official (includes professors or researchers) h. Royal
i. Terrorist or associate of rebel group j. Other, [namely]
k. Unspecified
10. [Only show if answer to 8 is c,d,e or f] What is the role of the second mentioned individual actor? (Same answers apply as question 9)
11. [Only show if answer to 8 is d,e or f] What is the role of the third mentioned individual actor? (Same answers apply as question 9)
31
12. [Only show if answer to 8 is e or f] What is the role of the fourth mentioned individual actor? (Same answers apply as question 9)
13. [Only show if answer to 8 is f] What is the role of the fifth mentioned individual actor? (Same answers apply as question 9)
14. How many group actors are mentioned in the article (without headline)? a. 0 b. 1 c. 2 d. 3 e. 4 f. 5 or more
15. [Only show if answer to 14 is not a] What is the role of the first mentioned group actor? (choose most specific role)
a. News agency or journalistic organization (includes group of journalists) b. Group of citizens
c. Government representatives or reference to government as whole (includes
United Nations since it is an intergovernmental organization)
d. Incumbent political party
e. Other political party (includes opposition in general)
f. Company or business groups (includes good causes like Amnesty
International)
g. Group of experts or independent bureau h. Royalty
i. Terrorist group or rebel group j. Other, [namely]
32
k. Unspecified
16. [Only show if answer to 14 is c,d,e or f] What is the role of the second mentioned group actor? (Same answers apply as question 15)
17. [Only show if answer to 14 is d,e or f] What is the role of the third mentioned group actor? (Same answers apply as question 15)
18. [Only show if answer to 14 is e or f] What is the role of the fourth mentioned group actor? (Same answers apply as question 15)
19. [Only show if answer to 14 is f] What is the role of the fifth mentioned group actor?
(Same answers apply as question 15)
20. Does the article focus on specific incidents or cases? (episodic frame) a. Yes (1)
b. No (0)
21. Does the article focus on larger trends or context? (thematic frame) a. Yes (1)
b. No (0)
22. Does the article refer to internet as a source? Any reference to websites or parts of the
internet like Facebook, news sites or Witter should be seen as a reference to internet and should be coded with ‘Yes’if it is presented as a sources where information is from.
a. Yes (1) b. No (0)
23. [Only show if answer to 22 is Yes] To which type of website is referred in the article? a. Social Media
b. News agencies c. Regular websites
33
d. Any combination of three types mentioned above e. Unspecified
34 Appendix 2: Amount of actors in articles per war in percentages
Gulf War Iraq War Syria War
Individual actors 0 14.3 17 7.8 1 23.8 36 29.4 2 9.5 17 11.8 3 38.1 18 23.5 4 4.8 6 13.7 5 or more 9.5 6 13.7 Group actors 0 28,6 41 27.5 1 42.9 42 29.4 2 23.8 10 31.4 3 4.8 5 5.9 4 0 1 3.9 5 or more 0 1 2
