Reporting two Gaza wars
A longitudinal comparison of media bias in war coverage in Dutch newspapers
Master Thesis 2018 Danielle Schweitzer
Graduate School of Communication University of Amsterdam
Master’s Program Communication Science: Political Communication Supervisor: Erika van Elsas
Student-ID: 11362804 Date of Submission: 02-02-2018 Word Count: 7102 excl. footnotes
MEDIA BIAS IN DUTCH WAR COVERAGE 2
Abstract
The aim of this research is to provide a better insight in the concept of biased reporting in newspapers on foreign wars. It is an attempt to understand if newspapers with a different political orientation cover wars differently and to what extent war coverage changes over time. The method used was a manual content analysis on Dutch war coverage of the
2008/2009 and 2014 Gaza wars. For this purpose, three Dutch newspapers are selected that can be classified as positioned on the left (Volkskrant), in the center (NRC Handelsblad) or on the right (De Telegraaf) of the political spectrum. A total of 250 articles were analyzed; N = 77, N = 85, N = 88, respectively. The results showed only minor, albeit sometimes
significant, differences between the analyzed newspapers over time. The overall findings suggest that Volkskrant paid more attention to the Palestinian side. Oppositely, De Telegraaf paid more attention to the Israeli side. Finally, the data did not show that NRC reporting was clearly more balanced than Volkskrant and/ or De Telegraaf. In all three newspapers,
balanced war reporting did not substantially change comparing 2008 to 2014. In the time frame studied, it was shown that apart from small differences, overall war coverage did not strongly reflected political polarization. The outcome of this research contributes to the understanding of war coverage over time in a single country not involved in the conflict that was analyzed.
Keywords: Media bias, War Coverage, Newspapers, Political Orientation, Over Time Developments
MEDIA BIAS IN DUTCH WAR COVERAGE 3
Introduction
Reporting of foreign wars in national news can influence public opinion in a country (Schwalbe & Dougherty, 2015). Even though foreign wars are geographically distant, some foreign wars are salient in Western national discourse and media coverage. Most Western citizens rely on their national news outlets for information regarding foreign wars (Schwalbe & Dougherty, 2015). Consequently, national news outlets shape their knowledge and
opinions. Foreign war reporting is expected to represent a true image of a conflict to citizens (Kuypers, 2006). It is the media’s task to produce accurate and fair war coverage that informs
citizens about foreign wars (Weaver, 1998; in Maslog et al., 2006). Credible and reliable war coverage requires objective reporting without media bias (Van Der Wurff & Schönbach, 2011). Furthermore, balanced and fair war reporting should provide citizens with context and understanding of a conflict (Neumann & Fahmy, 2012).
Past studies analyzing foreign war coverage have found evidence of media bias (Ruigrok, 2008; Thomas, 2011). Moreover, media bias is more likely to take place during times of conflict, as conflict enhances political polarization (Etman, 2003). Furthermore, political polarization is characterized by its spiral effect (Stroud, 2010). Partisan selective media exposure leads to polarization, and polarization leads to partisan selective exposure (Stroud, 2010). All in all, giving room to media bias. Most scholarly literature agrees that politics has become increasingly polarized over the past decades (Stroud, 2010; Pew Research Center, 2017). The influence of polarization is not limited to domestic affairs. Opinions concerning foreign conflicts have become more ideologically polarized over the past ten years (Smith & Doherty, 2016). On European level, opinions are sharply divided with regard to conflict resolution and humanitarian aid. Also in The Netherlands political polarization increased over time. Discourse surrounding foreign conflicts, such as the
Israel-MEDIA BIAS IN DUTCH WAR COVERAGE 4
Palestine/Arab conflict, have been characterized by extreme viewpoints (Abdullah & Chehata, 2011).
This research aims to study biased reporting in newspapers in The Netherlands on foreign wars. I attempt to understand how newspapers from a single country that is not directly involved in the conflict that was analyzed cover that conflict. This study analyzes first if a newspaper’s political orientation influences the Dutch news coverage on the 2008/2009 and 2014 Gaza wars. Second, this study analyzes if political polarization is reflected in Dutch news coverage on the 2008/2009 and 2014 Gaza wars. With this study I aim to answer two research questions: ‘Does a newspaper’s political orientation influences
balanced war reporting?’ and ‘To what extent did balanced war reporting change during the last decade?’
This study contributes to the knowledge about balanced war reporting in The Netherlands in two ways. Firstly, a substantial amount of the literature on foreign war reporting has analyzed war converge in the US and the UK media (Murray et al., 2008; Thomas, 2011; Fahmy & Eakin, 2014; Schwalbe & Dougherty, 2015; Neureiter, 2017). These countries are both characterized by high press partisanship. Nevertheless, literature concerning foreign war reporting in countries with low press partisanship is limited. One can expect biased war coverage from strong partisan news outlets (Vliegenthart & Mena Montes, 2014). However, war coverage in media systems with low press partisanship is expected to be relatively unbiased. This study seeks to identify the influence of Dutch newspaper’s political affiliation on media bias. The Netherlands is characterized by a low press partisan media system (Hallin & Mancini, 2004). Consequently, if the results indicate Dutch media are biased, it is most likely to assume media in countries characterized by high press partisanship also biased. However if the results indicate Dutch media report balanced, it could be due to low press partisanship in The Netherlands.
MEDIA BIAS IN DUTCH WAR COVERAGE 5
Secondly, scholarly literature on war coverage in Western countries has studied wars involving the country in question (Walgrave & Verhulst, 2005; Lindner, 2009; Leitz, 2011). A number of studies for example analyzed media coverage in the US and the UK during the Iraq war (Murray et al., 2008; Lindner, 2009). However, a country’s media are less likely to remain neutral in a conflict involving its own country (Lee & Maslog, 2005). Little is known about war coverage in countries whose government is not directly involved in a conflict. Since distance allows journalists to adopt a more neutral role (Maslog et al., 2006), media are less likely to take sides in a conflict not involving its country or government. Consequently, if the results indicate Dutch media are biased, it is most likely to assume media in countries directly involved in a conflict are also biased. However if the results indicate Dutch media reports balanced, it could be because media are less likely to take sides in a conflict not directly involving its country.
Finally, past studies on media bias in war coverage conducted cross-national comparisons (Philo & Berry, 2004; Vliegenthart & Schröder, 2009; Thomas, 2011). These studies only look at one point in time. Over time comparisons have been neglected. However passing of time is crucial when studying development of media bias. Only an over time comparison can detect changes in war reporting. During the last decade the Western world has seen increased political polarization (Stroud, 2010). The question to what extent war coverage reflects such developments remains. I aim to study newspapers over time to detect the reflection of political polarization on media bias.
The remainder of the article is structured as follows. First, I review previous literature. Second, I will explain the connection between newspaper coverage and political leaning, and between newspaper coverage and polarization, and introduce the hypotheses. Third, I shall explain the case and provide background information. Fourth, I elaborate on the
MEDIA BIAS IN DUTCH WAR COVERAGE 6
methodology, define key terms and subsequently present the results. Finally, I will discuss the findings and add suggestions for future research.
Theory Media bias
Media bias finds its roots in the contested principle of journalistic objectivity.
According to Ryan (2001) “objective journalists gather facts and opinions that conflict, verify
information carefully, seek to determine why accounts conflict and which most accurately reflect reality, and evaluate and fully identify sources” (p. 5). Furthermore, objective
reporting is not value free. Journalists hold on to the journalist code of ethics (Van Der Wurff & Schönbach, 2011). To strive for absolute objectivity is impossible, but journalists should aim for balance and fairness (D'Alessio and Allen, 2000).
There is not one way to define either balanced reporting or media bias. McQuail (1992) defined media bias as “a consistent tendency to depart from the straight path of objective truth by deviating either to left or right … In news and information it refers to a
systematic tendency to favor (in outcome) one side or position over another” (p. 191).
Brandenburg (2006) defined media bias as an outcome of press partisanship and considered it a measurement for press-party parallelism. Balanced or biased reporting can be influenced by news outlets political orientation (Vliegenthart & Schröder, 2009). Additionally, war
reporting can be influenced not only by news outlet, but also by the country and form of government the outlet operates in (Baum & Zhukov, 2015).
Scholars also disagree about the definition of balanced war reporting. Traditionally, scholars defined balanced war reporting as equal coverage of both sides of a conflict (Fico & Drager, 2001). However, when minority viewpoints receive equal attention as regular
viewpoints, balanced reporting is also biased due to overrepresentation (Boykoff & Boykoff, 2004). Others therefore argue balanced war reporting is not about an equal representation of
MEDIA BIAS IN DUTCH WAR COVERAGE 7
matters, but about a proportional representation of reality (Bischop, 2003).The present study, drawing on McQuail (1992), defines media bias as giving relatively more attention to one side of the conflict over another. Media bias is considered as the opposite of balanced reporting (Schwalbe & Dougherty, 2015). However, because the Israel-Palestine/Arab conflict is currently characterized by an asymmetrical number of casualties,1 this study also takes into account proportional representation of reality between newspapers (Bischop, 2003), especially in respect to victims. Bearing in mind the above mentioned, this research therefore does not study absolute media bias, but relative media bias between newspapers.
Media bias and war coverage
Previous studies have analyzed media bias in foreign war coverage for different types of news media. Studies have analyzed news magazines (Schwalbe & Dougherty, 2015), Internet coverage (Choi et al., 2006), televison coverage (Thomas, 2011), and most
extensively newspaper coverage (Ruigrok, 2008; Vliegenthart & Schröder, 2009; Neureiter, 2017). Other research combined the analysis of different media types (Van Belle, 2000; Maslog et al., 2007). Each media type, through their variety of characteristics, contributes differently to the understanding of news coverage (Nijens & Voorveld, 2015), and to the field of media bias. Newspapers in particular play an influential role in foreign war coverage (Vliegenthart & Schröder, 2009). Past research analyzing foreign war coverage has found evidence of media bias in newspapers (Ruigrok, 2008; Neureiter, 2017).
Next to differences between media types, there are also differences between news outlets, depending on country characteristics and news outlet characteristics (Brandenburg, 2006; Vliegenthart & Schröder, 2009). First, the influence of country characteristics on media bias will be discussed, and second the influence of news outlet characteristics.
1
In 2008/2009 there were approximately 1440 Palestinian casualties, of whom at least half were civilians, and 13 Israeli casualties (Zanotti et al., 2009). In 2014 there were approximately 2100 Palestinian casualties, of who approximately half were civilians, 6 Israeli civilian casualties and 67 Israeli soldier casualties (Human Rights Council, 2015).
MEDIA BIAS IN DUTCH WAR COVERAGE 8
The country in which news outlets operate influences media bias in foreign war coverage to some extent (Baum & Zhukov, 2015). Cross-national comparisons therefore contribute substantially to the field of media bias in foreign war coverage (Lee & Maslog, 2005; Maslog et al., 2006; Vliegenthart & Schröder, 2009; Fahmy & Eakin, 2014; Baum & Zhukov, 2015). Cross-national comparisons allow researchers to either look beyond national bias or detect possible country bias. For example, Maslog et al. (2006) compared Iraq war coverage in various Asian newspapers and concluded coverage and sourcing between Muslim and non-Muslim countries varied. Fahmy and Eakin (2014) confirmed that war coverage in countries directly involved in a specific conflict was more war orientated compared to non-involved countries. Baum and Zhukov (2015) argued there are media coverage differences between news outlets operating in democratic and undemocratic countries. Additionally, the authors argued cultural proximity to a subject or topic influences media coverage.
Vliegenthart and Schröder (2009) compared newspaper coverage in the US, the UK, Germany and The Netherlands, looking at tone, conflict involvement and attention. The authors emphasized the influence of over time fluctuation in content characteristics.
Walgrave and Verhulst (2005) also analyzed country involvement between eight countries. They concluded a country’s involvement in a conflict did influence media coverage to some
extent.
In sum, past studies have looked at media bias between countries based on country characteristics. They concluded that factors such as religion, form of government, proximity to a war, and involvement in a war contribute to a general viewpoint from which a war is reported in national news outlets. Because this study analyzes a single country, influencing factors, such as form of government, are assumed fixed factors. I expect all news in a single country to have a certain bias because of these fixed factors. In this research I will analyze within-country variation.
MEDIA BIAS IN DUTCH WAR COVERAGE 9
The relationship between newspapers and political orientation
Newspaper characteristics influence war coverage (Vliegenthart & Schröder, 2009). Newspapers have a variety of characteristics such as their organizational level and their quality. This study however focuses on newspapers’ political orientation. Countries as the US or the UK have liberal media systems with independent media (Hallin and Mancini, 2004) and strong press partisanship (Brants, 2002, in: Bardoel et al. 2002; Vliegenthart & Mena Montes, 2014). The news outlets clearly and openly associate with certain political
viewpoints. In contrast, The Netherlands has a democratic corporatist media system (Hallin and Mancini, 2004) and low press partisanship (Brants, 2002, in: Bardoel et al. 2002;
Vliegenthart & Mena Montes, 2014). News outlets in The Netherlands are independent from political parties. Nevertheless, newspapers are perceived to have a nuanced political leaning, but editors do not openly address that (Vliegenthart & Mena Montes, 2014).
Political parties in The Netherlands are divided regarding the Israel-Palestine/Arab Conflict. Right-wing parties and Christian parties generally tend to support pro-Israel viewpoints, and left-wing parties generally tend to support viewpoints critical of Israel (Neureiter, 2017). Looking at politicians voting behavior in parliament, these viewpoints are reaffirmed (The Rights Forum, 2017). With regard to media outlets, ideological theories assert that liberal media outlets are generally more critical of Israel and at the same time more supportive of Palestine compared to their conservative counterparts (Neureiter, 2017).
In this research, three Dutch newspapers are selected that can be classified as positioned on the left, in the center or on the right of the political spectrum (Vliegenthart & Schroder, 2009). With regard to the newspapers’ political orientation, I expect a
straightforward relationship: the more leftist the newspaper, the more attention it will give to the Palestinian side, reflecting the general political position of left-wing politicians and citizens (Walgrave and Verhulst, 2005; Ruigrok, 2008; Hirsh, 2017; Neureiter, 2017). A
MEDIA BIAS IN DUTCH WAR COVERAGE 10
newspaper positioned in the center of the political spectrum is expected to have most articles in which both conflict sides receive attention. Bearing in mind the above-mentioned, the following three hypotheses were formulated.
H1: A left leaning newspaper gives more attention to the Palestinian side, than a right leaning newspaper.
H2: A right leaning newspaper gives more attention to the Israeli side, than a left leaning newspaper.
H3: A newspaper positioned in the center of the political spectrum gives more balanced attention than newspapers with a left/right-wing leaning.
Media bias in war journalism over time
The political world is constantly changing. During the last decade, the Western world has seen growing support for populism and controversial viewpoints (Goodwin, 2011). More in general, there is a wider trend of public and political polarization in the Western world (Goodwin, 2011; Pew Research Center, 2014; Pew Research Center, 2017) enhanced by events such as terrorist attacks and refugee crises. At first sight terrorist attacks and increased immigration into Europe seem unrelated to public opinion about foreign conflicts. However, political polarization, enhanced by societal developments, can be reflected in foreign war coverage (Vliegenthart & Schröder, 2009), as foreign war coverage is not independent from domestic affairs. For example, political discussion concerning (diplomatic) interference and humanitarian aid for conflict areas include domestic interests. Besides, foreign conflict can also cause clashes between groups domestically. The Israeli-Palestine/Arab conflict is especially silent in European discourses because its fits poplar viewpoints concerning right-wing populism, anti-immigration and Islamophobia. The Israeli-Palestine/Arab conflict is used to project the popular right-wing narrative of the battle between the Judaeo- Christian heritage and the Muslim world (Kahmann, 2017).
MEDIA BIAS IN DUTCH WAR COVERAGE 11
Journalists tend to be influenced by public opinion (Vliegenthart & Schröder, 2009). Furthermore, the effect is reverse since selective media exposure tends to enhance
polarization (Stroud, 2010). Even though this study cannot measure the effects of increased polarization on media bias, increased polarization does provide an incentive to analyze over time developments in war coverage. The aim of this study is therefore not to study the effects or influence of polarization, but to detect if polarization is reflected in foreign war coverage over time. My expectations regarding media bias are straightforward. I expect newspapers to reflect the trend of polarization and report less balanced in 2014 compared to 2008/2009.
H4: A left leaning newspaper gives increased attention to the Palestinian side in 2014 compared to 2008.
H5: A right leaning newspaper gives increased attention to the Israeli side in 2014 compared to 2008.
The Case
This study examinesDutch newspaper reporting about the two most severe outbreaks of the Israel-Palestine/Arab conflict during last decade: the Gaza war in 2008/2009 and in 2014. To understand the origins of Israeli-Palestine/Arab conflict, there are two key historical events one needs to be aware of (Philo & Berry, 2004). Firstly, when Israel was established in 1948, large numbers of Palestinians were displaced from their homes and land, which is the origin of the Palestinian refugee case (Philo & Berry, 2004). Secondly, during the Six-day war in 1967, Israel captured the Gaza Strip. This event has been key to the Palestinian fight against Israel; Palestinians see Israel as their oppressor (Philo & Berry, 2004).
From the second intifada2 in 2000 onwards, Palestinians have fired rockets from Gaza targeted at Israel to fight the Israeli occupation. In September 2005, Israel disengaged from Gaza. Nevertheless, the United Nations still considers Gaza to be occupied by Israel
2
Intifada is the term for Palestinian uprising against Israeli occupation in the West Bank and Gaza Strip. The first intifada lasted from 1987 to 1993, and the second lasted from 2000 to 2005.
MEDIA BIAS IN DUTCH WAR COVERAGE 12
(Thomas, 2011). In 2006 Hamas, a Palestinian Sunni-Islamic fundamentalist organization that seeks to destroy Israel and is internationally seen as terrorist organization, won the elections in Gaza. This, for the time being, marked the end of any possible peace agreements between Israel and Palestine. To safeguard security, Israel put down a blockade around its Gaza borders (Thomas, 2011). This meant that Israel took control over both people and goods entering and leaving Gaza. When Egypt also put down a blockade in 2015, Gaza was
completely isolated. Living circumstances in Gaza deteriorated rapidly.
December 27, 2008, Israel launched a major military campaign against Hamas in Gaza to stop rockets targeted at Israel. The war ended January 18, 2009 (Fahmy & Neumann, 2012), after much international pressure to stop the conflict. After the conflict both Israel and Hamas were accused of violating international law (Amnesty International, 2009). There were approximately 1440 Palestinian casualties, of whom at least half were civilians, and 13 Israeli casualties (Zanotti et al., 2009). Israel held up the blockade to weaken Hamas.
By 2014, the economy in Gaza had completely collapsed due to the border blockade. The rights of the Palestinians were severely restricted (Human Rights Council, 2015).
Palestinian armed groups increasingly launched rockets during June and July 2014. Tensions, protests and violent clashes between the Israeli army and Palestinians increased.3 July 7, 2014, Israel launched operation ‘Protective Edge’ in Gaza. The goal was to stop the rocket
attacks by Hamas, destroy the cross-border tunnels Hamas had built, and seriously weaken Hamas. Firstly, the operation focused on airstrikes, but on July 17 Israel launched a ground operation. The operation concluded on 26 August when both Israel and Palestinian armed groups 4 adhered to an unconditional ceasefire (Human Rights Council, 2015). There were approximately 2100 Palestinian casualties, of whom approximately half were civilians. Six
3
In June, three Israeli teenagers were kidnapped and murdered in the West Bank. Israel responded by means of a search and arrest operation. Early July, a Palestinian teenager from East Jerusalem was murdered being burned alive in West Jerusalem as act of revenge.
MEDIA BIAS IN DUTCH WAR COVERAGE 13
Israeli civilians had been killed and 67 Israeli soldiers lost their lives (Human Rights Council, 2015). Navi Pillay, human rights official to the UN, condemned Israel for not doing enough to protect civilians and Hamas for "indiscriminate attacks" on Israel (BBC, 2014). The international community fiercely criticized Israel on grounds of disproportionate use of violence (Human Rights Council, 2015).
In conclusion, the 2008/2009 and the 2014 Gaza wars are relatively comparable. Both Gaza wars are characterized by large numbers of (Palestinian) civilian casualties (Amnesty International, 2009; Human Rights Council, 2015). Furthermore, during both wars the fighting parties received substantial international critique. Israel in particular received substantial international critique since they represent a democracy (Amnesty International, 2009; BBC, 2014.
Method Sample
The data are derived from a manual content analysis of 319 articles on two Gaza wars that appeared in 3 newspapers in The Netherlands. The 2008 Gaza war was from December 27, 2008 until January 18, 2009, and the 2014 Gaza war was from July 8 till August 26. These periods are highlighted because these were the most severe outbreaks of violence between the Israelis and Palestinians since the Israeli disengagement from the Gaza strip in 2005. The selection of the three newspapers was guided by the following consideration. I was interested in selecting papers skewed in opposite directions on the left/right political
continuum (Vliegenthart & Schroder, 2009). The first newspaper de Volkskrant is considered left leaning, NRC Handelsblad (NRC) is considered in the center of the political spectrum, and De Telegraaf is considered right leaning (Vliegenthart & Schroder, 2009). LexisNexis was used as search engine. Articles containing the words “Gaza” or “Israel” or “Palesti” or “Hamas”, during the research period were selected from LexisNexis. Using elaborate search
MEDIA BIAS IN DUTCH WAR COVERAGE 14
terms prevents selection bias. Besides, I aim to analyze newspaper content in general; also articles that refer to only one aspect of the war contribute to the overall coverage of the conflict.
Since it was not feasible to code all newspaper articles, a sample was taken. Past research showed that systematic sampling is the most efficient and representative method to study conflict because of its representative distribution of the course of the conflict (Fahmy & Neumann, 2012). In total I retrieved 1167 articles, with considerable cross-newspaper
differences regarding the number of articles and the length of the articles. To get a sample of roughly 300 articles, I used different calculation methods for each newspaper on the basis of the number of articles available. To get a representative sample, I used systematic sampling. For the first timeframe I selected every 4th article from the list of Volkskrant search results, every 3th article from NRC search results, and every 2th article from De Telegraaf search results. For the second time frame, I selected every 5th article from the list of Volkskrant search results, every 4th article from NRC search results, and every 5th article from De Telegraaf search results. In total 319 articles will be coded, which is approximately one quarter of the articles from our chosen population.5
Table 1 – Total Number of articles for each newspaper Newspapers Articles total
2008/ 2009 Articles sample 2008/ 2009 Articles total 2014 Articles sample 2014 Volkskrant 210 52 257 52 NRC 158 53 216 54 De Telegraaf 102 51 224 56 Total 157 157 697 162 5
In total 79 articles were deleted from the sample because they were not about the Gaza wars. Articles that were initially coded as ‘relevant’ (N = 38), but only referred to the Gaza war or an event related to the war once, were also deleted from the sample later on. In the end 38 extra articles were coded. The first article with more than one hit following the deleted article was chosen to safeguard representative sample with regard to the course of the war.
MEDIA BIAS IN DUTCH WAR COVERAGE 15
Operationalization
The present study defines media bias as giving relatively more attention to one party to the conflict than the other party to the conflict (McQuail (1992). Media bias is considered the opposite of balanced reporting (Schwalbe & Dougherty, 2015). The proportionality of the representation of reality between newspapers is taken into account (Bischop, 2003). I aim to determine media bias quantitatively looking at three concepts that measure attention:
historical and situational context, accessed voices, and the representation of victims and perpetrators (Ruigrok, 2008; Thomas, 2011; Gaisbauer, 2012). Historical context is defined as presence and absence of background information preceding the war. Historical context includes all historical events until the analyzed war. Situational context is defined as presence and absence of the direct motivation and direct incentive for war and violence for both Israel and Hamas (Thomas, 2011). Oppositely, to historical context, situational context is not only about events, but also about sentiments. Furthermore, situational context is also concerned with events during the war. Accessed voices are individuals or groups with an overt opinion, directly and indirectly related to the conflict that are explicitly mentioned, cited or referred to in the article (Thomas, 2011). Representation identifies how actors directly involved in the conflict are represented in a text. I distinguish between the roles of victims and perpetrators (Ruigrok, 2008; Gaisbauer, 2012). I have chosen these three concepts (context, accessed voices, representation of victim and perpetrator) for several reasons that will be explained in the following paragraphs.
Firstly, war coverage often accentuates a single event (Fahmy & Eakin, 2014). It lacks context, background or historical perspective. Consequently war coverage is often simplified into a superficial narrative with stereotypes and too little explanation (Galtung 2005;
Ruigrok, 2008). Past studies found high focuses on single events, concerning the Israel-Palestine/Arab conflict, and context was often missing (Philo & Berry, 2004; Fahmy &
MEDIA BIAS IN DUTCH WAR COVERAGE 16
Eakin, 2014). By providing historical and situational context one reduces the chances of stereotyping and simplifying, and increases the chances of balanced coverage. Therefore, context is a useful indicator of balanced war reporting. In general, the presence of context in the article is considered a characteristic of balanced reporting. However context in itself can be favorable towards one side. Context thus needs to be interpreted. If an article mentions consequences and motivations for a party it is considered to give attention to that party.
Secondly, previous research showed that media bias occurs when one party to the conflict is being heard and interviewed more often than the other party (Thomas, 2011). Equal presence of accessed voices in articles should prevent such bias resulting in balanced reporting. Accessed voices are dichotomized into either ‘critical of Israel or pro Palestine’ and ‘critical of Hamas or pro Israel’. For example, an article is coded as pro-Palestine if it
conveys either a Palestine or an anti-Israel statement or both. If an article is coded as pro-Palestine, the article is regarded as giving attention the Palestinian side, and vice versa. In the remainder of this article, ‘critical of Israel or pro Palestine’ is referred to as ‘pro Palestine’ and ‘critical of Hamas or pro Israel’ is referred to as ‘pro Israel’.
Thirdly, war reporting accentuates “us versus them” representations and utilizes words that are demonizing and victimizing (Maslog et al., 2006; Fahmy & Eakin, 2014). Unequal media representation of victims belonging to the respective conflict parties could strengthen bias (Gaisbauer, 2012). Equal media representation of victims thus contributes to a balanced reflection of the conflict. In contrast to past studies (Gaisbauer, 2012), this research analyzes both victim and perpetrator representation. No distinction is being made between victims resulting from the enemy’s versus from one's own actions (Gaisbauer, 2012). The
Gaza conflict in particular is characterized by victimized civilians on the one hand, and aggressive fighters on the other hand. A Hamas member is not necessarily a victim, whereas an under aged citizen from Gaza-city is more likely to be victim even though those two actors
MEDIA BIAS IN DUTCH WAR COVERAGE 17
represent the same side of the conflict. Differentiating between victim and perpetrator therefore contributes to a more complete understanding. In general, civilians form both sides are considered as victims, and fighters from both sides are considered as perpetrators.
References to Palestinian victims are considered as attention to the Palestinian side and vice versa. References to Palestinian perpetrators are interpreted as attention to the Israeli side and vice versa.
Finally, all variables are transformed into binary categories. For example historical context can either be ‘mentioned’ or ‘not mentioned’, and victims are either ‘Palestinian victims’ or ‘not Palestinian victims’.
Reliability
A randomly selected subsample of 9.4% (n = 30) of the articles was double-coded to allow reliability assessments. Five of the articles were not relevant. The reliability coding was done independently, by one additional coder (Van Zoonen et al., 2016). Three indices were used to calculate reliability scores of nineteen variables. First percent agreement, which is a simple procedure since all values consist only of two answer options, and the number of coders is two. Kappa is also calculated since it accounts for agreement by chance and Krippendorff’s alpha (Lombard et al., 2002). Krippendorff’s alpha should be greater than .677 to be
acceptable to draw a tentative sample in most cases (Riffe et al., 2014), and percent
agreement should be higher than .80 (Lombard et al., 2002). All variables but three, have a percent agreement of >.80 and a Krippendorff’s Alpha of > .677.6
Of these three, two
variables are dropped because they are not reliable.7 One variable with a percent agreement of
6
See Appendix A
7
The variables that were dropped were: for the question asking if the article referred to accessed voices answer option ‘ambiguous’ was dropped, which had a percent agreement of 84% and a Krippendorff’s Alpha of .570; for the question asking who was being identified as perpetrator answer option ‘IDf’ was dropped, which had a percent agreement of 64% and a Krippendorff’s Alpha of .434. Answer option ‘Israel (entity)’ and ‘IDF’ were merged into one variable, which makes sense since the army is under government control.
MEDIA BIAS IN DUTCH WAR COVERAGE 18
>.72 and a Krippendorff’s Alpha of .60 is kept.8 This low score is due to the variable’s dependence on a filter question. If coding error in the preceding (filter) question is taken into account, the variable had 86% percent agreement and Krippendorff’s Alpha is .758.
Results
The first research question asked whether a newspaper’s political orientation
influences balanced war reporting. The relationship between newspaper type and attention is being analyzed by means of a cross table. First of all, I will analyze the extent to which newspapers provide historical and situational context in their articles. Table 2 shows the percentages of articles that mentioned context of the total number of articles per newspaper. In general, 55.2% (N = 138) of the total number of articles contributed space to historical context. Volkskrant, as expected in hypothesis 1, paid more attention to Palestinian historical context (39%) compared to NRC (29.4%) and De Telegraaf (28.4%), but the relation is not significant. Oppositely to hypothesis 2 and 3, De Telegraaf did not pay most attention to Israeli historical context, nor did NRC reported most balanced.
Regarding situational context, 54.8% (N = 137) of the total number of articles contributed space to situational context. All newspapers mentioned Israel's motivation for violence in 50% of the articles, and Hamas motivation for violence in 33.2% of the articles. Volkskrant (35.1%) and NRC (38.8%) paid more attention to Hamas’ motivation compared
to De Telegraaf (26.1%), as expected in hypothesis 1. De Telegraaf (51.1%) roughly equal attention to Israel’s motivation compared to Volkskrant (50.6%) and NRC (48.2%), oppositely hypothesis 2. NRC most often referred to both parties’ motivation for violence (31.8%), as predicted by hypothesis 3. In sum, there are differences in terms of percentages between newspaper type and attention to context, but none are significant.
8
This variable is ‘historical context for Israel’ which is dependent on ‘historical context’. There was a skip pattern included in the codebook that automatically skipped the question regarding ‘historical context for Israel’ if the coder answered there was no historical context mentioned.
MEDIA BIAS IN DUTCH WAR COVERAGE 19
Table 2 – Frequency and percentages of articles that mentioned context of the total number of articles per newspaper
Volkskrant NRC Telegraaf Total Chi-square Lambda (N=77) (N=85) (N=88) (N=250) Sig. Historical context 47 (61%) 41 (48.2%) 50 (56.8%) 138 (55.2%) .244 .027 Historical context Israel 26 (33.8%) 22 (26%) 28 (31.8%) 76 (30.4%) .518 .000 Historical context Palestine 30 (39%) 25 (29.4.%) 25 (28.4%) 80 (32%) .287 .000
Articles per newspaper that mentioned historical context for both Israel and Palestine¹
19 (-) 16 (-) 19 (-) 54 (21.6%) .665 .000
Situational context 43 (55.8%) 47 (55.3%) 47 (53.4%) 137 (54.8.%) .946 .000
Motivation Israel 39 (50.6%) 41 (48.2%) 45 (51.1%) 125 (50%) .921 .024
Motivation Hamas 27 (35.1%) 33 (38.8%) 23 (26.1%) 83 (33.2%) .191 .000
Articles per newspaper that mentioned situational context for both Israel and Palestine²
23 (29.9%) 27 (31.8%) 21 (23.9%) 71 (28.4%) .485 .000
*p<.05, **p<.01, ***p<.000
¹This variable overlaps with the previous two variables; ² id..
Second, accessed voices are being analyzed. Table 3 shows that overall 84.4% (N = 210) of the articles contributed space to accessed voices. Furthermore, table 3 shows that NRC (71.8%) gives most attention to accessed voices pro Palestine compared to Volkskrant (62.3%) and De Telegraaf (65.9%), which is contrary to what was expected in hypothesis 3. Moreover, table 3 shows that NRC pays noticeably more attention to voices pro Palestine (71.8%) compared to voices pro Israel (57.6%). This outcome is contrary to hypothesis 3 that predicted NRC would report most balanced. Finally, De Telegraaf (73.9%) paid significantly more attention to accessed voices pro Israel (p = .045, lambda = .000) compared to
MEDIA BIAS IN DUTCH WAR COVERAGE 20
Table 3 – Frequency and percentages of articles that referenced to accessed voices of the total number of articles per newspaper
Volkskrant NRC Telegraaf Total Chi- square Lambda (N=77) (N=85) (N=88) (N=250) Sig.
Accessed voices 62 (80.5%) 71 (83.5%) 77 (87.5%) 210 (84.4%) .470 .000
Critical of Israel or pro Palestine 48 (62.3%) 61 (71.8%) 58 (65.9%) 167 (66.8%) .434 .000
Critical of Hamas or pro Israel 45 (58.4%) 49 (57.6%) 65 (73.9%) 159 (63.6%) .045* .000
Articles per newspaper that are both critical of Israel or pro Palestine and critical of Hamas or pro Israel¹
31 (40.3%) 39 (45.9%) 46 (52.8%) 116 (46.4%) .304 .034
*p<.05, **p<.01, ***p<.001
¹This variable overlaps with the previous two variables
Third, victim and perpetrator representation is being analyzed. Table 4 shows that overall 62.4% (N = 161) of the articles identified victims. There is a very weak significant association between newspaper type and victim identification (p = .021, lambda = .000). Volkskrant paid more attention to victims (76.6%) compared to NRC (56.5%) and De
Telegraaf (61.4%). Furthermore, in line with hypothesis 1, Volkskrant paid more attention to Palestinian victims (64.9%) compared to NRC (52.9%) and De Telegraaf (50%). All
newspapers paid roughly equal attention to Israeli victims. NRC, in line with hypothesis 3, most often referred to both parties as victims per single article (38.8%). This outcome
underlines NRC as most balanced compared to other newspapers. Besides, there is an overall significant moderate association (p < .000, Lambda = .340) between attention paid to Israeli victims and Palestinian victims.All newspapers paid more attention to Palestinian victims (55.6%) compared to Israeli victims (40.4%).9
Furthermore, table 4 shows that overall 49.4% (N = 123) of the articles represented casualties. There is no significant association between newspaper type and attention paid to
9
NRC paid significantly more attention to Palestinian victims compared to Israeli victims (p <.000, Lambda =.649), and De Telegraaf paid significantly more attention to Palestinian victims compared to Israeli victims (p <.000, Lambda =.450).
MEDIA BIAS IN DUTCH WAR COVERAGE 21
Israeli or Palestinian casualties. However in line with hypothesis 1, Volkskrant paid more attention to Palestinian casualties (48.7%) compared to De Telegraaf (40.9%). Furthermore, NRC represented Israeli and Palestinian casualties per simgle article more often (31.8%) compared to Volkskrant (N = 12) and De Telegraaf (27.3%), which is in line with hypothesis 3. Besides, there is an overall significant weak association (p < .000, Lambda =.111) between the attention paid to Israeli casualties and Palestinian casualties. All newspapers paid more attention to Palestinian casualties (45.4%) compared to Israeli casualties (31.8%).10
Finally, there is a weak significant correlation between newspaper type and Israeli perpetrator representation (p = .029, lambda = .138). As expected in hypothesis 2, De
Telegraaf represented Israeli actors less often as perpetrator (39.8%) compared to Volkskrant (61%) and NRC (48.2%). The representation of Palestinian actors as perpetrator, and the representation of Israeli and Palestinian actors as perpetrator per article do not significantly differ between the three newspapers.
In sum, Volkskrant paid the most attention to victims in general, but the least attention to Israeli casualties compared to NRC and De Telegraaf, which underlines hypothesis 1. De Telegraaf paid the least attention to Israel as perpetrator, which underlines hypothesis 2. Finally, all three newspapers paid more attention to Palestinian victims and casualties compared to Israeli victims and casualties.
Table 4 – Frequency and percentages of articles that represented victims and perpetrator of the total number of articles per newspaper
Representation Volkskrant NRC Telegraaf Total Chi-square Lamba (N=77) (N=85) (N=88) (N=250) Sig.
Victims 59 (76.6%) 48 (56.5%) 54 (61.4%) 161 (62.4%) .021* .000
10
NRC paid significantly more attention to Palestinian casualties compared to Israeli casualties (p <.005, Lambda =.214), and De Telegraaf paid significantly more attention to Palestinian casualties compared to Israeli casualties (p <.019, Lambda =.167)
MEDIA BIAS IN DUTCH WAR COVERAGE 22
Israeli victims 31 (40.2%) 34 (40%) 36 (40.9%) 101 (40.4%) .992 .000
Palestinian victims 50 (64.9%) 45 (52.9%) 44 (50%) 139 (55.6%) .130 .000
Articles per newspaper that identify both Israeli and Palestinian victims¹
23 (29.9%) 33 (38.8%) 30 (34.1%) 86 (34.4%) .487 .000
Casualties 41 (53.9%) 41 (48.2%) 41 (46.6%) 123 (49.4%) .621 .029
Israeli Casualties 17 (-) 28 (32.9%) 28 (31.8%) 73 (31.8%) .276 .000
Palestinian Casualties 37 (48.7%) 40 (47.1%) 36 (40.9%) 113 (45.4%) .565 .000
Articles per newspaper that identify both Israeli and Palestinian casualties ²
12 (-) 27 (31.8%) 24 (27.3%) 63 (25.2%) .052 .018
Israeli perpetrator 47 (61%) 41 (48.2%) 35 (39.8%) 123 (49.2%) .029* .138
Palestinian perpetrator 46 (59.7%) 34 (40%) 43 (48.9%) 123 (49.2%) .051 .122
Articles per newspaper that identify both Israeli and Palestinian perpetrator ³
35 (45.5%) 33 (38.8%) 27 (30.7%) 95 (38%) .162 .122
*p<.05, **p<.01, ***p<.001
¹This variable overlaps with the previous two variables; ² id.; ³ id..
The second Research question asked if balanced war reporting changed during the last decade. In order to answer this research question, a two-way ANOVA was conducted to analyze the interaction between type of newspaper and over time changes on the attention given to the Israel side and the Palestinian side.11 First, I will look at references to historical and situational context for each war. Table 5 shows that in 2008, 60 % (N total = 133) of the
articles mentioned historical context compared to 50% (N total = 117) in 2014. Table 6 shows
that the interaction effect between type of newspaper and over time changes on the attention for historical context of Palestine was significant F (2, 244) = 3.306, p = .038, η² = .026.12 Volkskrant paid significantly more attention to Palestinian historical context in 2014, as
11
Usually, a two-way ANOVA requires a continuous dependent variable. However research showed it is acceptable to conduct a two-way ANOVA if the sample proportions for the cells lie between .25 and .75 and there are at least 20 degrees of freedom (D’Agostino, 1971).
12
It should be noted that the assumption of equal variances in the population has been violated, Levene's F (5, 244) = 11.481, p = .000.
MEDIA BIAS IN DUTCH WAR COVERAGE 23
expected in hypothesis 4; NRC also paid increased attention to Palestinian historical context in 2014 (2008 = 27%; 2014 = 32%). Oppositely, De Telegraaf paid the least attention to Palestinian historical context in 2014 (15%) compared to 2008 (40%). Furthermore, table 7 shows a significant interaction effect between type of newspaper and over time changes regarding the number of articles that mentioned historical context for both parties in one article F (2, 244) = 3.421, p = .034, η² = .027.13 In 2014, De Telegraaf less often (7%) reported historical context for both parties compared to 2008 (34%). The interaction effects between type of newspaper and over time changes on the attention were not significant for: Israeli historical context; situational context of Israel; situational context of Hamas; and situational context of both parties per single article.14
Table 5 – Context per newspapers for both wars over time
Volkskrant Volkskrant NRC NRC Telegraaf Telegraaf Total Total
2008 2014 2008 2014 2008 2014 2008 2014 M (SD) N= 42 M (SD) N=35 M (SD) N=44 M (SD) N=41 M (SD) N=47 M (SD) N=41 M (SD) N=133 M (SD) N=117 Historical context .67 (.56) 28 .54 (.51) 19 .48 (.51) 21 .49 (.51) 20 .66 (.48) 31 .46 (.51) 19 .60 (.49) 80 .50 (.50) 58 Historical context Israel .36 (.49) 15 .31 (.47) 11 .27 (.45) 12 .24 (.44) 10 .45 (.50) 21 .17 (.38) 7 .36 (.48) 48 .24 (.43) 28 Historical context Palestine .36 (.49) 15 .43 (.50) 15 . 27 (.45) 12 .32 (.47) 13 .40 (.50) 19 .15 (.36) 6 .35 (.48) 46 .29 (.47) 34 Articles that mentioned historical context for both Israel and Palestine¹
.24 (.43) 10 .26 (.44) 9 .18 (.39) 8 .20 (.40) 8 .34 (.48) 16 .07 (.26) 3 26. (44) 34 .17 (.38) 20 Situational context .52 (.51) 22 .60 (.50) 21 .59 (.50) 26 .51 (.51) 21 .51 (.51) 24 .56 (.50) 23 .54 (.50) 72 .56 (.50) 65 13
It should be noted that the assumption of equal variances in the population has been violated, Levene's F (5, 244) = 9.938, p < .000.
14
MEDIA BIAS IN DUTCH WAR COVERAGE 24 Motivation Israel .50 (.51) 21 .51 (.51) 18 .52 (.51) 23 .44 (.50) 18 .51 (.51) 24 .51 (.51) 21 .50 (.51) 68 .49 (.50) 57 Motivation Hamas .36 (.49) 15 .34 (.48) 12 .34 (.48) 15 .44 (.50) 18 .19 (.40) 9 .34 (.48) 14 .29 (.46) 39 .38 (.49) 44 Articles that mentioned situational context for both Israel and Palestine² .33 (.48) 14 .26 (.44) 9 .27 (.45) 12 .37 (.49) 15 .19 (.40) 9 .29 (.46) 12 .26 (.44) 35 .31 (.46) 36
¹ This variable overlaps with the previous two variables; ² id..
Table 6 – Results of a two-way ANOVA (N = 250) for presence of historical context of Palestine
Sum of Squares df Mean Square F p η2
Newspaper .633 2 .316 1.478 .230 .012
War 2014 .139 1 .139 .650 .421 .003
Newspaper * War 2014 1.416 2 .708 3.306 .038 .026
Error 52.261 244 .214
Total 80.000 250
Table 7 – Results of a two-way ANOVA (N = 250) percentage of total number articles that mentioned historical context for both Israel and Palestine in one article
Sum of Squares df Mean Square F p η2
Newspaper .146 2 .073 .438 .646 .004
War 2014 .380 1 .380 2.283 .132 .009
Newspaper * War 2014 1.139 2 .570 3.421 .034 .027
Error 40.623 244 .166
Total 54.000 250
In addition, considering the second research question a similar analysis was
performed looking at the accessed voices. Table 8 shows that in 2008 81% (N total = 133) of
the articles contributed space to accessed voices compared to 87% (N total = 117) in 2014. The
MEDIA BIAS IN DUTCH WAR COVERAGE 25
accessed voices pro Palestine or pro Israel or for both parties per article was not significant.15 Nonetheless, in terms of percentages, Volkskrant paid less attention to voices pro Palestine over time (2008 = 67%; 2014 = 57%), which is contrary to hypothesis 4; however, NRC paid most attention to accessed voices pro Palestine over time (2008 = 66%; 2014 = 78%). Finally, as hypothesis 5 predicted, De Telegraaf paid more attention to voices pro Israel in 2014 (76%) compared to 2008 (72%).
Table 8 – Accessed voices per newspaper for both wars over time
Volkskrant Volkskrant NRC NRC Telegraaf Telegraaf Total Total
2008 2014 2008 2014 2008 2014 2008 2014 M (SD) N= 42 M (SD) N=35 M (SD) N=44 M (SD) N=41 M (SD) N=47 M (SD) N=41 M (SD) N=133 M (SD) N=117 Accessed voices .76 (.43) 32 .86 (.36) 30 .84 (.37) 37 .83 (.38) 34 .83 (.38) 39 .93 (.26) 38 .81 (.39) 108 .87 (.34) 102 Critical of Israel or pro Palestine .67 (.48) 28 .57 (.52) 20 .66 (.48) 29 .78 (.42) 32 .66 (.48) 31 .66 (.48) 27 .66 (.48) 88 .68 (.47) 79 Critical of Hamas or pro Israel .62 (.49) 26 .54 (.51) 19 .64 (.49) 28 .51 (.51) 21 .72 (.45) 34 .76 (.44) 31 .66 (.48) 88 .61 (.49) 71
Articles that are both critical of Israel or pro Palestine and critical of Hamas or pro Israel¹ .52 (.50) 22 .26 (.44) 9 .45 (.50) 20 .46 (.51) 19 .55 (.50) 26 .49 (.50) 20 .51 (.50) 68 .41 (.50) 48
¹This variable overlaps with the previous two variables.
Third, the interaction between type of newspaper and over time changes on the attention given to victim and perpetrator representation was also measured. Table 9 shows that in 2008 overall 65% (N total = 133) of the cases identified victims, compared to 64% (N total = 117) in 2014. Furthermore, in 2008 47% (Ntotal = 133) cases identified casualties,
compared to 61% (Ntotal = 117) in 2014. There are no significant interactions between type of
15
MEDIA BIAS IN DUTCH WAR COVERAGE 26
newspaper and over time changes on the attention given to victim and perpetrator
identification.16 Nevertheless, the results in table 9 do suggest a tendency of newspapers to report differently regarding victim and perpetrator representation. To start with, in line with hypothesis 4, Volkrant paid more attention to Palestinian victims 2014 (71%) compared to 2008 (60%). De Telegraaf paid roughly equal attention to Israel victims in 2014 (41%) compared to 2008 (40%), oppositely to hypothesis 5. Furthermore, in line with hypothesis 3, NRC most often represented victims from both sides in one article, also over time.
Concerning casualty representation, results in table 9 do suggest a tendency to support hypothesis 4 which predicted Volkskrant would pay increased attention to Palestinian
casualties; note NRC and De Telegraaf also paid increased attention to Palestinian casualties. NRC paid most attention to Israeli casualties in 2014 (37%) compared to De Telegraaf (34%) and the Volkskrant (26%).
Regarding perpetrator representation, results in table 9 do suggest a tendency to support hypothesis 5. De Telegraaf represented Israel less often as perpetrator in 2014 (33%) compared to 2008 (47%). There is no indication for support of hypothesis 4, since Volkskrant represented Hamas more often as perpetrator in 2014 (69%) compared to 2008 (52%).
Volkskrant tended to represent both parties as perpetrator (2008= 38%; 2014 = 54%) relatively more frequently compared to NRC (2008 = 33%; 2014= 46%) and De Telegraaf (2008 = 32%; 2014 = 30%).
Table 9 – Representation of victims and perpetrators
Volkskrant Volkskrant NRC NRC Telegraaf Telegraaf Total Total
2008 2014 2008 2014 2008 2014 2008 2014 M (SD) N= 42 M (SD) N=35 M (SD) N=44 M (SD) N=41 M (SD) N=47 M (SD) N=41 M (SD) N=133 M (SD) N=117 Victims .74 (.45) 31 .80 (.406) 28 .55 (.50) 24 .59 (.50) 24 .66 (.48) 31 .56 (.50) 23 .65 (.48) 86 .64 (.48) 75 16
MEDIA BIAS IN DUTCH WAR COVERAGE 27 Israeli victims .36 (.49) 15 .46 (.51) 16 .36 (.49) 16 .44 (.50) 18 .40 (.50) 19 .41 (.50) 17 .38 (.49) 50 .44 (.50) 41 Palestinian victims .60 (.50) 25 .71 (.46) 25 .52 (.51) 23 .54 (.51) 22 .51 (.51) 24 .49 (.51) 20 .54 (.50) 72 .57 (.50) 67 Articles that identifies both Israeli and Palestinian victims¹ .24 (.43) 10 .37 (.49) 12 .34 (.48) 15 .44 (.50) 18 .30 (.46) 14 .39 (.49) 16 .29 (.46) 39 .40 (.49) 46 Casualties .49 (.51) 21 .60 (.50) 21 .45 (.50) 20 .51 (.51) 21 .47 (.50) 22 .46 (.51) 19 .47 (.50) 63 .52 (.50) 61 Israeli Casualties .20 (.40) 8 .26 (.44) 9 .30 (.46) 13 .37 (.49) 15 .30 (.46) 14 .34 (.48) 14 .27 (.44) 35 .32 (.47) 37 Palestinian Casualties .46 (.51) 19 .51 (.51) 18 .43 (.50) 19 .51 (.51) 21 .38 (.49) 18 .44 (.50) 18 .42 (.50) 56 .49 (.50) 57 Articles that identifies both Israeli and Palestinian casualties² .12 (.33) 5 .20 (.41) 7 .27 (.45) 12 .37 (.49) 15 .21 (.41) 10 .34 (.48) 14 .20 (.40) 27 .31 (.46) 36 Israeli perpetrator .66 (.50) 28 .63 (.49) 22 .44 (.50) 19 .54 (.51) 22 .47 (.50) 22 .33 (.47) 14 .52 (.50) 69 .49 (.50) 58 Palestinian perpetrator .52 (.51) 22 .69 (.47) 24 .35 (.48) 15 .46 (.51) 19 .49 (.51) 23 .50 (.51) 21 .48 (.50) 64 .54 (.50) 63 Articles that identifies both Israeli and Palestinian perpetrator³ .38 (.49) 16 .54 (.51) 19 .33 (.47) 14 .46 (.51) 19 .32 (.47) 15 .30 (.46) 12 .34 (.48) 45 .43 (.50) 50
¹This variable overlaps with the previous two variables; ² id.; ³ id..
Finally, table 10 shows Israeli historical context had a significant main effect of time F(1, 244) = 3.994, p = .047, η² = .01617, indicating that newspapers less often mentioned Israeli historical context in 2014 compared to 2008. Thus over time, all three newspaper paid less attention to Israeli historical context.
Table 10 – Results of a two-way ANOVA (N = 250) percentage of total number articles that mentioned historical context for Israel
17
It should be noted that the assumption of equal variances in the population has been violated, Levene's F (5, 244) = 7.006, p < .000.
MEDIA BIAS IN DUTCH WAR COVERAGE 28
Sum of Squares df Mean Square F p η2
Newspaper .252 2 .126 .603 .548 .005 War 2014 .833 1 .833 3.994 .047 .016 Newspaper * War 2014 .823 2 .412 1.974 .141 .016 Error 50.896 244 .209 Total 76.000 250 Discussion
The aim of this research was to provide a better insight in the concept of biased reporting in newspapers on foreign wars. Research question one asked if a newspaper’s political orientation influences balanced war reporting. In general, the results indicate that a newspaper’s political orientation influences balanced war reporting to a certain extent. The
results showed only minor, albeit sometimes significant, differences between the analyzed newspapers. The overall findings suggest that Volkskrant paid more attention to the
Palestinian side. Oppositely, De Telegraaf paid more attention to the Israeli side. Finally, the data did not show that NRC reporting was clearly more balanced than Volkskrant and/ or De Telegraaf. These results fit the assumption that a country with low press partisanship reports relatively unbiased (Brants, 2002, in: Bardoel et al. 2002; Vliegenthart & Mena Montes, 2014) compared a country with high press partisanship (Thomas, 2011; Neureiter, 2017). These results also underline that newspapers are not likely to take unconditional sides in a conflict not involving its country compared to a conflict involving its country (Walgrave & Verhulst, 2005; Maslog et al., 2006; Lindner, 2009; Fahmy & Eakin, 2014).
Research question two asked to what extent balanced war reporting changed during the last decade. In all three newspapers, balanced war reporting did not substantially change comparing 2008 to 2014. In the time frame studied, it was shown that apart from small differences, overall war coverage did not strongly reflected political polarization.
MEDIA BIAS IN DUTCH WAR COVERAGE 29
Although Dutch newspapers did not report strongly biased, there are differences between newspapers across the whole period. Differences between newspapers are most strongly with regard to Palestinian historical context, and the representation of victims and perpetrators. Overall, Volkskrant paid more attention to Palestinian historical context across the whole period. These results reflect a tendency on the political left, and amongst younger generations to sympathize with the Palestinian refugee case relatively more compared to earlier times (Smith & Doherty, 2016; Hirsh, 2017). This outcome thus reflects political polarization of public opinion a very general sense.
Furthermore, Volkskrant paid more attention to the representation of victims and perpetrators. Socialism is characterized by standing up for the underdog. During war, civilians are victims of combatting parties (Schulzke & Carroll, 2015), and perpetrators are responsible for people’s grievances. Strongly differentiating between victims and perpetrators thus corresponds with a leftish orientation. Furthermore, the so-called ‘underdog’ narrative, which casts the Palestinians as helpless and the Israelis as aggressive, is often used
concerning the Israel-Palestine/Arab conflict (Neireiter, 2017).
In general, the results suggest Dutch news over time seems to reflect reality (Bischop, 2003), instead of polarization. All newspapers paid frequently more attention to casualties and victims. Palestinian victims in particular received attention. That is not surprising, since Palestinians’ primary living circumstances deteriorated because of the conflict. According to
proportional reporting (Bischop, 2003; Schwalbe & Dougherty, 2015) more attention should be paid to Palestinian victims compared to Israeli victims. However, it is up for debate to what extent proportionality is a reflection of reality. It is beyond the scope of this study to argue what is proportional and what is disproportional.
Past studies found statements and interviews in favor of Israel dominated the news in high press partisanship countries (Philo & Berry, 2004; Thomas, 2011). No such bias is found
MEDIA BIAS IN DUTCH WAR COVERAGE 30
in The Netherlands. Bearing in mind increased international critique against Israel (Ban Ki-moon, 2008), this outcome indicates all Dutch newspapers succeeded to reflect reality to a certain extent.
Past studies on the Israeli-Palestinian/Arab conflict found news tended to provide inadequate background information on foreign wars (Philo & Berry, 2004; Thomas, 2011; Fahmy & Eakin, 2014). Previous research showed people would sympathize more often with Palestinians if they had more historical background information (Philo & Berry, 2004). This study shows all Dutch newspapers provided context in only fifty percent of the articles. Dutch newspapers thus fail to report (balanced) in terms of context.
The positioning of NRC as a newspaper in the center of the political spectrum is debatable. NRC can also be categorized as a quality right newspaper (Walgrave & Verhulst, 2005), a liberal neutral newspaper (Jonkman & Verhoeven, 2013) or a quality newspaper (Hijmans et al., 2011). Nevertheless, earlier research on war coverage between Volkskrant, NRC and De Telegraaf found that NRC followed real-world developments in its coverage (Ruigrok, 2008). Since I define balanced reporting as a reflection of reality, Ruigrok’s findings (2008) underline NRC as a newspaper in the center of the political spectrum. Moreover, this study showed NRC’s growing critical view regarding Israel. NRC thus
reflects current international critique against Israel (Ban Ki-moon, 2008). A critical attitude towards politics is an aspect of quality journalism (Van Der Wurff & Schönbach, 2011). Also the results of this study underline the positioning of NRC as a newspaper in the center of the political spectrum.
The Israel-Palestine/Arab conflict, being highly emotional, is prone to media bias (Thomas, 2011). One the one hand, one could argue that Israel-Palestine/Arab conflict could serve as an example because it is polarized. On the other hand, because of its highly
MEDIA BIAS IN DUTCH WAR COVERAGE 31
regarding foreign war coverage. One way or the other, the conflict receives much media coverage even in countries not involved in the analyzed conflict.
Furthermore, the results cannot be compared to high press partisanship countries’ media coverage about conflicts not involving its government. Also, since The Netherlands was not involved in the conflict, this research does not predict assumptions about media bias (in The Netherlands) in countries directly involved in a conflict. Dutch media could report relatively more biased concerning a conflict directly involving The Netherlands. Future research could repeat this study analyzing a conflict The Netherlands is directly involved in.
This study analyzed relative media bias. Future studies could analyze to what extent proportionality reflects reality, in order for proportional reporting to serve as a measure for absolute media bias. A measure for proportional reporting could prevent unequal media representation of war and thus bias (Gaisbauer, 2012).
Biased war coverage could have broad implications since it forms public
understanding about conflicts (Reese & Buckalew, 1995). The results of this study are mixed. Overall the results suggest that Dutch media reports relatively balanced because a
newspaper’s political orientation influences balanced war reporting only to a certain extent. Additionally, war coverage did not substantially change during the last decade. The outcome of this research contributes to the understanding of war coverage over time in a single country not involved in the conflict that was analyzed.
References
Abdullah, D., & Chehata, H. (2011). European Public Perceptions of the Israel-Palestine Conflict. London: Middle East Monitor, European Muslim Research Centre, AlJazeera Center For Studies.
MEDIA BIAS IN DUTCH WAR COVERAGE 32
Amnesty International. (2009). Israel/Gaza: Operation “Cast Lead”: 22 Days of Death and Destruction (Rep.). Retrieved form Amnesty International
file:///Users/Danielle/Downloads/mde150152009en%20(1).pdf
Azrout, R., Moeller, J. & de Vreese, C. H. (2010), European Election Media Study Data and Documentation, 01/07/2015.
Bardoel, J., Vos C., Vree van F., Wijfjes, H. (2002). Journalistieke cultuur in Nederland. Amsterdam: Amsterdam University Press.
Baum, M. A., & Zhukov, Y. M. (2015). Filtering revolution: Reporting bias in international newspaper coverage of the Libyan civil war. Journal of Peace Research, 52(3), 384-400.
BBC. (2014). UN's Navi Pillay warns of Israel Gaza 'war crimes'. Retrieved from BBC http://www.bbc.com/news/world-middle-east-28437626
Bishop, E .(2003) Balanced vs. equal (Editorial). St. Louis Journalism Review 33(259): 3. Boykoff, M. T., & Boykoff, J. M. (2004). Balance as bias: global warming and the US
prestige press. Global environmental change, 14(2), 125-136.
Brandenburg, H. (2006). Party strategy and media bias: A quantitative analysis of the 2005 UK election campaign. Journal of elections, public opinion and parties, 16(2), 157-178.
Brants, K. (2002). Opgejaagd door Cerberus. De moeizame mediatisering van de politieke communicatie.
Choi, J. H., Watt, J. H., & Lynch, M. (2006). Perceptions of news credibility about the war in Iraq: Why war opponents perceived the Internet as the most credible medium. Journal
of Computer‐ Mediated Communication, 12(1), 209-229.
D’ agostino, R. B. (1971). A second look at analysis of variance on dichotomous data.
MEDIA BIAS IN DUTCH WAR COVERAGE 33
D’Alessio, D, & Allen, M. (2000). Media bias in presidential elections: a meta analysis.
Journal of Communication, 50(4), 133–156.
Entman, R. M. (1991). Symposium framing US coverage of international news: Contrasts in narratives of the KAL and Iran air incidents. Journal of communication, 41(4), 6-27. Fahmy, S. (2010). Contrasting visual frames of our times: A framing analysis of English-and
Arabic-language press coverage of war and terrorism. International Communication Gazette, 72(8), 695-717.
Fahmy, S., & Eakin, B. (2014). High drama on the high seas: Peace versus war journalism framing of an Israeli/Palestinian-related incident. International communication gazette, 76(1), 86-105.
Fahmy, S., & Neumann, R. (2012). Shooting war or peace photographs? An examination of newswires’ coverage of the conflict in Gaza (2008-2009). American Behavioral
Scientist, 56(2).
Fico, F & Drager, M. (2001). News stories about conflict generally balanced. Newspaper Research Journal 22(1): 2–11.
Gaisbauer, F. (2012). Representations of victimization and responsibility during the Second Intifada and the Gaza War in German quality newspapers. Conflict &
Communication, 11(2).
Goodwin, M. (2011). Right Response Understanding and Countering Populist Extremism in Europe. London: Retrieved from The Royal Institute of International Affairs:
https://www.chathamhouse.org/publications/papers/view/178301.
Hallin, D. C., & Mancini, P. (2004). Comparing media systems: Three models of media and politics. Cambridge university press.
Hijmans, E., Schafraad, P., Buijs, K., & d'Haenens, L. (2011). Wie schrijft ons nieuws?. Tijdschrift voor Communicatiewetenschap, 39(2), 77.
MEDIA BIAS IN DUTCH WAR COVERAGE 34
Hirsh, D. (2017). Contemporary left antisemitism. Routledge.
Human Rights Council. (2015). Report of the detailed findings of the independent
commission of inquiry established pursuant to Human Rights Council resolution S 21/1.
Jonkman, J., & Verhoeven, P. (2013). From risk to safety: Implicit frames of third-party airport risk in Dutch quality newspapers between 1992 and 2009. Safety science, 58, 1-10.
Kahmann, B. (2017). ‘The most ardent pro-Israel party’: pro-Israel attitudes and anti
antisemitism among populist radical-right parties in Europe. Patterns of Prejudice, 51(5), 396-411.
Kuypers, J. A. (2006). Bush's war: Media bias and justifications for war in a terrorist age. Rowman & Littlefield Publishers.
Lee, S. T., & Maslog, C. C. (2005). War or peace journalism? Asian newspaper coverage of conflicts. Journal of Communication, 55(2), 311-329.
Leitz, L. (2011). Oppositional Identities: The Military Peace Movement's Challenge to Pro Iraq War Frames. Social Problems, 58(2), 235-256.
Lindner, A. M. (2009). Among the troops: Seeing the Iraq War through three journalistic vantage points. Social Problems, 56(1), 21-48.
Lucaites, J. L., & Hariman, R. (2001). Visual rhetoric, photojournalism, and democratic public culture. Rhetoric Review, 20(1/2), 37-42.
Maslog, C. C., Lee, S. T., & Kim, H. S. (2006). Framing analysis of a conflict: How newspapers in five Asian countries covered the Iraq War. Asian Journal of Communication, 16(1), 19-39.
MEDIA BIAS IN DUTCH WAR COVERAGE 35
Murray, C., Parry, K., Robinson, P., & Goddard, P. (2008). Reporting dissent in wartime: British press, the anti-war movement and the 2003 Iraq War. European Journal of Communication, 23(1), 7-27.
Neijens, P. C. & Voorveld, H. A. M. (2015). Cross-platform advertising: Current practices and issues for the future. Journal of Advertising Research, 55(4), 1-7.
Neumann, R., & Fahmy, S. (2012). Analyzing the spell of war: A war/peace framing analysis of the 2009 visual coverage of the Sri Lankan civil war in western newswires. Mass Communication and Society, 15(2), 169-200.
Neureiter, M. (2017). Sources of media bias in coverage of the Israeli–Palestinian conflict: the 2010 Gaza flotilla raid in German, British, and US newspapers. Israel Affairs, 23(1), 66-86.
Pal achter Israël! Nederland en de Zesdaagse Oorlog. (2017). Andere Tijden. The Netherlands: VPRO.
Pew Research Center. (2014). Where News Audiences Fit on the Political Spectrum. Retrieved from Pew Research Center http://www.journalism.org/interactives/media polarization/
Pew Research Center. (2017). Political Polarization, 1994-2017. Retrieved from Pew
Research Center: http://www.peoplepress.org/interactives/political-polarization-1994 2017/
Philo, G., & Berry, Mike. (2004). Bad news from Israel. London: Pluto Press.
Reese, S. D., & Buckalew, B. (1995). The militarism of local television: The routine framing of the Persian Gulf War. Critical Studies in Media Communication, 12(1), 40-59. Riffe, D., Aust, C. F., Jones, T. C., Shoemake, B., & Sundar, S. (1994). The shrinking foreign
