Master’s Thesis
Framing immigrants in Dutch newspapers
Name: Michiel van Aarnhem
Student ID: 10245669
Course: Master’s Programme Communications Science: Master’s Thesis Graduate School: Graduate School of Communication
Supervisor: Linda Bos
"Migrants and refugees are not pawns on the chessboard of humanity."
1 Abstract
This thesis is about framing migration in Dutch newspapers. The framing of migration is a widely researched topic in the field of political communication. The findings of this thesis complement our understanding of the process of framing by differentiating not only between frames, but also between different types of immigrants. This research resulted in three important findings, based on a content analysis of 360 articles from five Dutch newspapers, NRC, De Volkskrant, De Telegraaf, Trouw and Algemeen Dagblad, during a three months’ time period. The main findings are: (1) Asylum seekers and refugees are dominantly present in today’s Dutch newspapers, although they only account for 5 percent of the total amount of immigrants. (2) De Telegraaf makes more use of negative framing in comparison with other Dutch newspapers. (3) Labour migrants are significantly more framed as positive for the economy in comparison with asylum seekers, refugees and family migration.
Introduction
Migration is an interdisciplinary field of study, embedded in anthropology, sociology, political economy, international relations and political communication. Apart from broad interest displayed by scholars of various subfields, migration is also perceived as salient by people within societies (Castles et, al., 2014, p. 57). This is particularly so in the case of receiving societies, because the settlement of groups of migrants can “fundamentally change the social, cultural, economic and political fabric of societies’’ (Castles et, al., 2014, p. 1). The Netherlands is an example of such a receiving society. By 2010, at least 11 percent of the population in the Netherlands was foreign-born. If one includes the children of immigrants, the total percentage would raise to 20, thus one in five (Hollifield, et, al., 2014, p. 256-257).
There are many examples of studies investigating the way these immigrants changed and shaped the Dutch society. Within the field of political communication, many studies are conducted concerning the framing of migration or immigrants in a media context (Buonfino, 2004; Van Gorp, 2005; Helbling, 2014; Roggebrand & Vliegenthart, 2007; Nickels, 2007). These studies are important, because they contribute to our understanding of the consequences of framing migration in influencing public opinion, public debate and policy outcomes. This study complements this research, by focusing on the way newspapers in the Netherlands frame immigrants. In doing so, this study tries to find the salient frames used in the Netherlands in relation to different types of immigrants. This is an important first step to understand and measure the effect of framing migration in the Netherlands.
2 The studies of Roggebrand & Vliegenthart (2007), and Helbling (2014), have come up with interesting and useful definitions of different types of frames. The studies by Buonfino (2004), Van Gorp (2005) and Nickels (2007), have found interesting effects of framing asylum seekers and immigration in Europe, Belgium and Luxembourg. However, they only looked at migration in general or at specific types of immigrants. Buonfino (2004), shows the evolution of policy discourse and investigates the politicization of migration. Buonfino concludes that migration becomes securitized within a dynamic interplay between public opinion, mass media and governments (Buonfino, 2004). Van Gorp (2005), conducted a content analysis on Belgium newspapers. Van Gorp finds that on the one hand ‘asylum seekers are innocent victims’ while on the other hand, ‘asylum seekers are intruders’ (Van Gorp, 2005). Nickels (2007), examines the asylum discourse in Luxembourg. Nickels identifies four frames that media use to refer to asylum seekers and refugees (Nickels, 2007). These studies show how media framing affects the migration discourse in these specific countries. However, because Buonfino (2004) focuses on migration in general, his study does not differentiate between the various forms of immigrants that exist. The studies by Van Gorp (2005) and Nickels (2007) only look at one specific type of migration, asylum seekers and refugees. None of these studies differentiated types of immigrants and none of these studies included all types of immigrants that concerned their country specific context. Not differentiating on immigrant type can be considered as a gap in the literature about the effects of framing migration.
This master’s thesis fills that gap. It makes use of studies that define different types of frames, in order to investigate how these types of frames are used in Dutch newspapers. However, this study goes beyond the studies that conceptualize different frames (Roggebrand & Vliegenthart, 2007), analyze public discourses in different countries on migration in general, asylum seekers or refugees (Buonfino, 2004; Van Gorp, 2005; Nickels, 2007), or investigate how general actors perceive immigration (Helbling, 2014). This study does this by looking at different types of migration and immigrants, instead of analyzing the effect of framing migration in general. We know from previous research that frames have effect on public opinion, public social debate and policy outcomes in different countries with different contexts (Buonfino, 2004; Van Gorp, 2005; Nickels, 2007). However, before we can investigate the effect of frames in the Netherlands, it is important to find out which frames are salient within the specific context of the Netherlands.
Thus, this thesis is first to look at the use of framing in relation to all different types of immigrants in the Dutch context, making this research unique in the field of framing migration
3 studies. The types of frames used in this study are; the humanitarian, positive economic, negative economic, security, nationalistic cultural and multicultural frames. These frames are further elaborated on in the theory section of this study. Next to these different types of frames, there are also different types of migration or immigrants distinguished, these are; labour, asylum seekers, refugees, family, student and post-colonial migration. In order to answer the research question of this study, a content analysis on news articles of the five most read Dutch newspapers will be conducted in a three months’ time period from 1 September, 2014 until 1 December, 2014. An advantage of this thesis is that it included the five biggest newspapers in the Netherlands. Making this study extensive and its conclusions not easily ignored.
In short, this study focusses on the way migration is framed in Dutch newspapers, thereby specifically looking at differentiation between types of immigrants. The research question of this thesis is: In what way do Dutch newspapers frame different types of immigrants?
Theory
The first part of this section is about framing. The theory of Entman is used to define framing. Then the effect of framing will be elaborated on and theory about frame building is discussed. In the last part about framing, the different immigration frames will be identified. Theory of Helbling (2014), Roggebrand and Vliegenthart (2007) and Huysmans (2006) is used to come up with six different types of frames. The second part of the theory section is about the different types of immigration. Theory by Hollifield (2014) and Castles (2014) is used in order to make categorizations of immigrant types. Together with demographic and historical information, six different types of immigrants are identified which are relevant within the Dutch context.
The effect of framing
A frame is a scheme or way of interpretation. According to Reese frames are; “organized principles that are socially shared and persistent over time, that work symbolically to meaningfully structure the social world.’’ (Reese et, al., 2001: p. 11). One of the most cited definitions of framing comes from Entman, who defines framing as: “Framing is selecting some aspects of a perceived reality, to enhance their salience in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation’’ (Entman, 1993: p. 53). In this way, some facts are becoming more important than others. This is emphasized by Gamson, who argues that: ‘‘facts take on their meaning by being embedded
4 in a frame or story line that organizes them and gives them coherence, selecting certain ones to emphasize while ignoring others’’ (Gamson, 1989, p. 157).
One of the processes of framing is the application of frames by the media, called frame-building (Helbling, 2014, p. 22). On this frame-frame-building phase lies the focus of this thesis. Frame-building is about how the media makes use of specific frames. Furthermore it looks at how specific topics, such as migration, are framed. While agenda setting theory could provide answers in the direction of how the media comments on a specific topic, framing analysis would make it able to show how these media comments shape our perceptions of this topic. In other words, journalists are able to provide contextual cues for interpretation. In this way, news media has the potential power to manipulate audience perceptions on a certain topic (Kuypers, 2002, p. 9). Understanding these frames becomes important, especially “when one considers that a powerful feature of frames is that they define problems, causes and solutions, although not necessarily in that order’’ (Kuypers & Cooper, 2005, p. 3).
In other words, the frames journalist use, partly determine whether a topic becomes a problem, what causes this problem and possible ways to solve it. In this way, journalists and news media can frame the topic of migration and, in doing so, migration can become considered as a ‘problem’ for the Dutch society. News media could for example argue that this problem is caused by our liberal immigration procedures and generous welfare state provisions. A possible solution, provided by ‘experts’ presented in the same media, could then be to apply stricter nationalization procedures and make it more difficult for aliens to gain welfare state provisions. Or to put it in Kuypers & Cooper’s words: “when journalists frame, they construct a particular point of view that encourages the facts of a given situation to be interpreted in a specific way. Thus journalists can, knowingly or unknowingly, guide the interpretation of readers toward a particular point of view’’ (Kuypers & Cooper, 2005, p. 3). Based on this theory, it is to be expected that journalists of the different newspapers make use of frames that are in line with their political point of view. The selected newspapers cover a broad public. De Volkskrant is a quality newspaper that represents the left and social part of the population. De NRC is a quality, right leaning newspaper which is more liberal. De Telegraaf and Algemeen Dagblad are right of the centre and represent the more populist and less educated part of the population. Finally, Trouw is more leftish and coming from a protestant tradition, should therefore pay relatively more attention to asylum seekers and refugees (Roggebrand & Vliegenthart, 2007, p. 528-529). Helbling (2014), develops certain expectations based on the left-right placement of parties. Helbling (2014), argues that the more left a party is, the more likely it is that this party makes
5 more use of the multicultural, humanitarian and positive economic frames (Helbling, 2014, p. 26). This thesis uses these expectations and applies them on newspapers instead of political parties. Thus, news articles from the leftish newspapers, De Volkskrant and Trouw, and liberal NRC, are more likely to use the humanitarian and positive economic frame in comparison to the right newspapers, De Telegraaf and Algemeen Dagblad. While in turn, De Telegraaf and Algemeen Dagblad are more likely to use the negative economic and security frames in comparison with De Volkskrant, NRC and Trouw.
The media plays a significant role in the way Dutch citizens and politicians perceive immigration. Two theories about the influence of the media on citizens and politicians are shortly mentioned here. First the view of Norris (2000), who argues that the media act as an intermediary between citizens and politicians. Second the view presented by Meyer (2002), who argues that the media is an active political player who decides what is newsworthy and how to present this. Because the media has the power to decide, Meyer argues that the media must be perceived as an ‘active’ political player. This argument is in line with the view of Kuypers and Cooper, mentioned earlier in this section. Results from other studies show that there exists a significant relation between framing migration in the media, which in turn affects the political arena (Roggebrand & Vliegenthart, 2007). Both the claim made by Meyer and Roggebrand & Vliegenthart are convincingly in favor of a direct link between media framing and influencing the public opinion and political debate. The view of Meyer (2002), and Roggebrand & Vliegenthart (2007), that the media players a significant political role in framing migration, thereby influencing attitudes and opinion within the Dutch society, makes the study on the use of framing even more relevant.
Complementary to research about the influence of news media is research that investigated how immigration is framed (Buonfino, 2004; Van Gorp, 2005; Nickels, 2007). The research by Buonfino makes use of different types of frames applied in different countries in Europe, on the topic of migration in general (Buonfino, 2004). While asylum seekers and refugees are specifically investigated in the other articles about the influence of framing by news media (Van Gorp, 2005; Nickels, 2007). A more recent study by Helbling (2014), combines the findings of these studies and comes up with more generic frames. These generic frames can be used for analyzing framing in the field of migration in different contexts. But the study of Helbling (2014), focuses on the role political actors play in framing political issues, thereby leaving out the application of frames by the media. However, because of the generic character of the frames produced in this study, they can also be used for research on the framing
6 of migration by the media. And because these frames are developed to be applicable within different contexts, they can also be applied for the analysis of Dutch newspapers articles. The next paragraph will categorize the different types of immigrant frames on the basis of the study by Helbling and others.
Categorizing types of frames
One general study of framing by Semetko & Valkenburg (2000), identified five frames in the news, by using a deductive approach. The five frames identified are; the conflict frame, human interest frame, economic consequences frame, morality frame and the responsibility frame (Semetko & Valkenburg, 2000, p. 95-96). These five frames are general and need to be further specified in order to be useful for the analysis of news articles about migration. This step is done by Helbling (2014), Roggebrand & Vliegenthart (2007), and Huysmans (2006).
The humanitarian frame as described by Helbling (2014), comes from the human interest frame of Semetko & Valkenburg (2000). But the frame by Helbling is more specific and therefore useful for the analysis of news articles about migration. The humanitarian frame of Helbling (2014), refers to migration in terms of ‘fairness, equality, human rights, freedom, rule of law and democracy’.
Helbling (2014), further developed the economic consequences frame into two new frames called the ‘positive economic frame’ and the ‘negative economic frame’. News media can, knowingly or unknowingly, use these frames in order to positively or negatively frame migration in relation to the economy.
Roggebrand and Vliegenthart (2007), together with Helbling (2014), created the ‘nationalistic’ and ‘multicultural’ frames, which are especially relevant in the context of migration. The nationalistic cultural frame consists of items related to ‘foreign infiltration, national identity, loss of tradition, avoiding flows of refugees and national sovereignty’ (Helbling, 2014, p. 25). The multicultural frame is often referred to as pro typically Dutch and refers to ‘respect of cultural diversity, tolerance between immigration groups and integration’ (Roggebrand & Vliegenthart, 2007, p. 530; Helbling, 2014, p. 25).
Finally, Huysmans (2006), writes about the securitization or migration. His main argument is that migration has become a ‘security issue’ and has been identified or constructed as a ‘threat’. He writes: “Framing refugees as a humanitarian question introduces different relations to refugees than framing it as a security question, for example. While the former allows for compassion or for relating to the refugee as a rights holder, the latter sustains fear of refugees
7 and policies of territorial and administrative exclusion.’’ (Huysman, 2006, preface xii). In this way, framing asylum and migration can lead to adoption of certain security policies, such as identity and border controls. Furthermore Huysmans writes about a connection between “immigration and asylum with terrorism, transnational crime and border control.’’ (Huysmans, 2006: p, 68). Huysmans even makes the link between framing migration in security terms and “policing migration, possibly to the disadvantage of demographic or economic policies regulation migration.’’ (Huysmans, 2006, p. 31). Huysmans is considered expert on the securitization of migration, therefore his work is used in order to determine the security frame.
On the basis of the above-mentioned literature and theory, six types of frames are selected which are relevant to study the framing of migration in Dutch newspapers. The types of frames identified are; the humanitarian frame, the positive economic frame, the negative economic frame, the security frame, the national cultural frame and the multicultural frame.
Humanitarian frame
Referring to migrants or migration in terms of moral issues can be seen as humanitarian. The humanitarian discourse tries to make moral universal claims. Using a human rights language, referring to different types of legal rights and humanitarian rights of migrants, can be viewed as using the humanitarian frame. Examples of this type of frame are; discrimination of groups, human rights, freedom of movement, Geneva convention, rule of law, fairness, equality and democracy (Helbling, 2014, p. 25).
Positive economic frame
This variable is selected when economic issues around migration refer to migration/immigrants as economically beneficial. When referring to migrants or migration as economically beneficial, most arguments are made in favor of cheap labour for the labour market. Examples of positive frames used are; attracting high skilled workers, productivity, fair international competition, cheap labour force and doing the jobs Dutch people don’t want to do anymore (Semetko & Valkenburg, 2000; Helbling, 2014, p. 25).
Negative economic frame
When referring to migrants or migration as economically costly. Most arguments refer to the welfare state provisions these immigrants receive, but all kinds of arguments used
8 in order to underscore the costs of migration are coded in this variable. Examples of negative frames used are; unfair competition, using our welfare system, lazy people, benefits recipients and people who steal our jobs (Semetko & Valkenburg, 2000; Helbling, 2014, p. 25).
Security frame
The security frame refers to security rhetoric which is used for political purposes to reframe migration as a security threat. This variable measures all kinds of security rhetoric used in order to link migrants or migration to security related issues or words. Examples of such frames are; linking migrants or migration with ‘terrorism, criminality, security, organized crime, Islamisation, and other threats’ (Huysmans, 2006).
National cultural frame
The nationalistic type of frame can be seen as negative towards immigration. These frames are mostly used in a context of preserving national boundaries and the culturally homogeneous society. Examples of such frames are when referred to migrants or migration as; ‘foreign infiltration, national identity, loss of traditions, avoid Islamisation, avoid flows of refugees and national sovereignty’ (Helbling, 2014, p. 25; Roggebrand & Vliegenthart, 2007).
Multicultural frame
The multicultural type of frame can be seen as positive towards immigration. These frames “favor cultural openness, exchange and the peaceful coexistence of various cultural and religious groups within a society.’’ (Helbling, 2014, p. 24). Examples of such frames are; ‘integration through tolerance, advantages of cultural diversity, tolerance between religious groups and European identity’ (Helbling, 2014; Roggebrand & Vliegenthart, 2007, p. 530).
Immigration in the Netherlands
The 21st century can be defined as the age of migration (Castles, et. Al., 2014). In Europe, many countries face difficulties in their attempt to control immigration flows. Italy, for example, faces a humanitarian crisis due to an explosion of refugees from northern Africa. Spain on the other hand, struggles with the impact immigration has on the economy, especially
9 in the labour market which faces high unemployment rates. In the Netherlands, both humanitarian and economic issues dominate the debate of migration (Hollifield et ,al., 2014).
Kymlicka (1995), distinguishes two models for integrating immigrants into receiving societies. Based on these two models, different immigrant groups have different rights. Immigrants may receive polyethnic rights, which means that the group must integrate into the culture of the majority, but that they can maintain certain aspects of their ethnicity (Kymlicka, 1995, pp. 78-79). The other model entails societal cultures, which enables groups to acquire stronger rights, such as self-governance and special representation rights for national minorities (Kymlicka, 1995, pp. 27-33). This second model is considered multicultural, which enables immigrant groups to keep their cultural values. This model asks for a tolerant and open government with respect to immigrant rights. The Netherlands is considered to have a multicultural tradition. However, immigrants do not have the same rights as national minorities, because they come to the Netherlands voluntarily and therefore have a responsibility of integrating into the cultural values of the receiving state (Kymlicka, 1995). This multicultural model is criticized by Joppke. According to Joppke, there is a shift away from multiculturalism in the Netherlands: “the supposedly difference-friendly, multicultural Netherlands is currently urging migrants to accept ‘Dutch norms and values’ in the context of a policy of civic integration that is only an inch away from the cultural assimilation that had once been attributed to the French’’ (Joppke, 2007, p. 2). Together with the rise of the populist anti-immigration party of Geert Wilders, the Netherlands is becoming more hostile towards immigrants. This is especially the case for those groups of immigrants that are perceived as costly for society, such as asylum seekers and refugees and to a lesser extent also for labour and family migration. The multicultural theory by Kymlicka together with the critique of Joppke, increase the expectation that different groups of immigrants are perceived different within the Netherlands. It is to be expected that asylum seekers and refugees receive relatively more media attention than labour and family migration, because they are perceived as more costly for society.
Political studies in the field of migration are used to identify the different types of immigrants relevant for the Netherlands. Most of these studies underscore the specific context of the Netherlands. These studies look into immigrant groups who came to the Netherlands throughout the 20th century, which in turn produces more context specific results than only identifying broad immigrant categorizations, based on international migratory movements worldwide. Demographics together with historical information is used to identify the six most relevant immigrant groups present in the Netherlands.
10 Categorizing types of immigrants
In the international migration outlook of 2013, the OECD identified six different categories of permanent immigration. First family migration, second migration related to free movement, third migration related to work, fourth migration related to humanitarian issues, fifth migrants who accompany family of workers and finally migration due to other reasons (OECD, 2013). These six different categories are very broad and need to be further specified in order to be applicable in the Netherlands and useful for the research purposes of this thesis. What are the main motives for immigrants to come to the Netherlands as off 2011? The OECD identified the four main motives for immigrants to come to the Netherlands; labour (40%), family (32%), study (15%) and asylum seeking/refugees (5%). (OECD, 2013, p. 278). These four motives point in the direction of the six different types of immigrants identified, however one needs to include historical information about the Netherlands in order to be complete and context specific.
Historically, one could identify at least six dominant types of immigrants that occur in the Netherlands. These are; labour migrants, asylum seekers, refugees, family migrants, student migrants and post-colonial migration (Castles et, al., 2014, p. 57-68; Hollifield, et, al., 2014, p. 256-272). In the first quarter of the 20th century, the Netherlands became a country of immigration. Only after WOII, there was a period of net emigration, but this soon ended at the start of the 1960’s, because of the guest worker programs. During these decades, the Dutch government promoted foreigners to come to the Netherlands for work. This cheap labour force was perceived good for the growing economy and again turned the Netherlands into a net immigration country. These guest worker programs were bound to be temporarily, however most Turkish and Moroccan guest workers stayed. Because the Netherlands had relatively liberal family reunification and formation policies, these guest workers were also able to bring their family to the Netherlands (Hollifield, et, al., 2014, p. 260-263). In this way, the substantial labour migration of the ‘60’s and ‘70’s resulted in high numbers of family migration. Another common type of migration towards the Netherlands has been labeled post-colonial. The three biggest colonies of the Netherlands, Indonesia, Suriname and the Antilles, resulted in high numbers of post-colonial migration, especially after WOII (Hollifield, et, al., 2014, p. 261-262). An additional type of immigrants are asylum seekers and refugees who come to the Netherlands in large numbers since the Vietnam war. In the ‘70’s the Dutch government invited refugees from Vietnam and Chile. After the fall of the Iron Curtain in 1989, and the Yugoslavia crisis of 1992-1994, many more refugees came from eastern Europe to the Netherlands (Meijer, Internet
11 Archives, visited: 6, January 2014). More recently, asylum seekers form the Horn of Africa and Syria came to the Netherlands in order to apply for asylum. Furthermore, the Dutch government tries to attract high skilled labour, therefore students from all over the world attend our universities.
Using this historical information combined with the broad OECD categorizations and the four main motives for immigrants to come to the Netherlands, six types of immigrants are defined. These types of immigrants are; labour migration, asylum seekers, refugees, family migration, study migration and post-colonial migration. (Castles, et, al., 2014; Hollifield, et, al., 2014; OECD, 2013). Regardless of country of origin, social class, or other specific characteristics, the types of immigrants used for this study, should fall under one of these six categories.
Labour migration
This type of migration refers to migrants that come to the Netherlands for reasons of work. Historically, the biggest group of labour migrants came from Turkey and Morocco during the seventies and eighties. Although these people where temporary migrant workers, they stayed permanently, either legal or illegal. More recently, people from eastern Europe are the biggest group of labour migrants. People from within the EU, such as workers from Poland, Hungary and Romania are migrating to the Netherlands. Labour migrants account for approximately 40% of the total amount of immigrants in the Netherlands (Hollifield, et, al., 2014; OECD, 2013, p. 278).
Asylum seeker
The terms asylum-seeker and refugee are often confused; an asylum-seeker is someone who says he or she is a refugee, but whose claim has not yet been definitively evaluated (UNHCR, 2015). Asylum seekers are a group who awaits the decision of the IND, whether or not they can stay in the Netherlands. The IND is short for immigration and naturalization service, which is a governmental organ that decides whether or not an alien can stay. If not, they sometimes are sent back to their country of origin, or put in detention. Together with refugees, asylum seekers account for approximately 5% of the total immigrant population (Hollifield, et, al., 2014; OECD, 2013, p. 278).
12 This group of immigrants comes mostly from African countries and want to stay in the Netherlands, mainly because it is not safe in their home country. The difference between a asylum seeker and a refugee is that a refugee is already accepted as a person who is in need of protection. Although they are the smallest group, with only 5% of the total amount of immigrants, they are subject of strong debate (Hollifield, et, al., 2014; OECD, 2013, p. 278).
Family reunification
The Geneva convention and the Amsterdam treaty provide a legal framework for family migration. Immigrants that legally live in the Netherlands have the right of family reunification. This means that the temporary guest workers from Turkey and Morocco, after legalized, where able to legally reunify with their wives and children. Historically, many immigrants came to the Netherlands in this way, mainly from Turkey and Morocco. Family migration accounts for approximately 32% of the total immigrant population in the Netherlands (Hollifield, et, al., 2014; OECD, 2013, p. 278).
Students
This type of immigrants is highly skilled and educated and therefore provides opportunities for the Dutch society. This group of migrants comes from all over the world and perceive certain benefits, such as tax cuts from the Dutch government in order to keep them in the Netherlands after their studies. Students account for approximately 15% of the total immigrant population in the Netherlands (Hollifield, et, al., 2014; OECD, 2013, p. 278).
Post-colonial migration
This type of migration refers to a migrant or group of migrants that came to the Netherlands while originally born in a ‘former’ Dutch colony, such as Suriname, Antilles or Indonesia (Hollifield, et, al., 2014, pp. 261-261) .
One last important theory to mention is the theory of news values or news criteria. This theory is about how much attention news stories are given by news media, for example newspapers. Newsworthiness is considered to be based on specific news values such as; frequency, negativity, unexpectedness, unambiguity, personalization, meaningfulness, references to elite
13 nations and persons, conflict and continuity (Galtung and Ruge, 1965; Hardcup & O’neill, 2001). Because asylum seekers and refugees are mostly perceived as unexpected visitors, coming from dangerous countries torn apart by wars and conflict (UNHCR Statistical Yearbook, 2012). And because asylum seekers and refugees have often meaningful and personal stories to tell, it is more likely that asylum they will gain attention of newspapers. Asylum seekers and refugees are more often concerned with issues of negativity, frequency, unexpectedness, personalization, meaningfulness, conflict and continuity in comparison with labour and family migrants (UNHCR Statistical Yearbook, 2012). Based on the theory of news values, it is more likely that newspapers write about asylum seekers and refugees, because they include more aspects of news criteria (Galtung and Ruge, 1965; Hardcup & O’neill, 2001).
Research questions
The research question of this thesis is: In what way do Dutch newspapers frame different types of immigrants? In order to answer this research question properly, three sub-questions are prepared.
The first sub-question is based on the expected that asylum seekers and refugees receive relatively more attention from newspapers than labour and family migration. This expectation is based on the multicultural theory of Kymlicka (1995), and the critical remarks made by Jopkke (2007). Complementary to this, the expectation can be addressed to the theory of news values by Galtung and Ruge, (1965), and Hardcup & O’neill, (2001). This leads to the first sub-question, (Q1): Do newspapers preferably write about asylum seekers and refugees in comparison with labour and family migration?
The second sub-question comes from the expectation that different newspapers use different types of frames in line with their political point of view. This expectation is based on the view of Kuypers and Cooper (2005) about journalistic framing, and the distinction made between right and left newspapers by Roggebrand and Vliegenthart, (2007) and Helbling (2014). This leads to the second sub-question, (Q2): Do different newspapers make use of different types of frames when framing migration?
The third sub-question comes from the expectation that due to the differences between immigrant types, it becomes likely that newspapers use different types of frames when writing about migration. This expectation is based on the categorization of immigration types by Castles (2014), Hollifield et, al., (2014) and the OECD (2013). This leads to the third sub-question,
14 (Q3): Which frames are used differently when writing about asylum seekers, refugees, labour migrants, family migration, student migration and post-colonial migration?
Methods
This section further elaborates on the quantitative content analysis that is conducted. The first part is about sampling. This part elaborates on data selection and sample size determination. The second part is about operationalization, this part describes all variables and operationalizes them. Finally all central variables of the study are theoretically substantiated.
In order to answer the research questions, I conducted a content analysis of Dutch immigration news published between 1 September, 2014 until 1 December 2014. The content analysis focused on articles from the five most read Dutch newspapers: NRC Handelsblad, De Volkskrant, Trouw, De Telegraaf and Algemeen Dagblad (Bakker & Scholten, 2003). The total amount of new articles coded is N=360.
Sample
For the content analysis, I looked for news articles containing at least one of the six types of immigrants identified in the theory section. Furthermore, news articles needed to come from one of the selected newspapers; De Volkskrant, NRC Handelsblad, Trouw, Algemeen Dagblad and De Telegraaf. Thus, the first criterion entails that the article has to refer to at least one type of migration; labour, refugees, asylum seekers, family, student or post-colonial. Two more criteria are considered when the first criterion is met. These criteria are; first the article has to deal with migration as ‘main topic’ and second ‘at least half’ of the article has to deal with the topic of migration.
The LexisNexis database is used in order to find relevant news articles. Based on the criteria mentioned above, the search categories are determined, these are; migration, labour-migration, family-reunification, asylum-seekers, refugees, student migration and colonial migration. Secondly, these categories need to be addressed in either ‘the head’ or ‘the lead’ of the article, in order to meet the second and third criteria described above. The option HLEAD is chosen because this serves as a pre-selection method, ensuring migration is mentioned in either ‘head’ or ‘lead’. When searching in LexisNexis, the ‘Power Search’ option is enabled. The search term is: HLEAD (Migra! OR arbeidsmigra! OR gezinshereniging OR asielzoeker! OR vluchteling! OR familiemigra! OR studentmigra! OR kolonialemigra!). The selected period is from 1 September, 2014 until 1 December, 2014 and the selected newspapers are: De
15 Volkskrant, NRC Handelsblad, Trouw, Algemeen Dagblad and De Telegraaf. I selected A three months period because all relevant immigrant groups, news items and topics are expected to be covered in three months’ time. During these three months there were no elections or other special events that would, on beforehand, suggest to have implications for the reliability of the data. Elections could influence the way newspapers write about certain topics, for example migration. It could for instance be the case that anti-immigration parties will receive more attention in months previous to elections, thereby influencing the coverage of immigration related news and the outcome of this study.
The performed search resulted in 552 articles. These articles were read and only selected when meeting all the criteria mentioned. Finally, a total of 360 articles were selected and coded on the basis of a codebook that is included in appendix 1.
Operationalization
The first part of each coded article contains the general information of the article. These variables are; source of the article, date of publication, page number, title and number of words. The second part of the coding focused on the different types of migration, the six types identified are; labour migration, asylum seekers, refugees, family migration, student migration and post-colonial migration. Labour migration (LAB); selected when referred in the article to a migrant or group of migrants in the context of work. Asylum seekers (ASY); selected when referred to a person that claims to be a refugee, but whose claim has not yet been definitively evaluated. Refugee (REF); selected when referred to someone with a rightful claim of asylum based on the fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion. Family reunification (FAM); selected when referred to a migrant or group of migrants who migrated in order to be with their family. Student migrant (STU); selected when referred in the article to a migrant or group of migrants who migrated for reasons of studying. Post-colonial migrant (COL); selected when referred in the article to a migrant or group of migrants which come from a former colony and migrated because of this. These types of migration are translated into an analytical framework consisting of different indicators or questions. Based on these indicators the types of migration are identified, examples of these indicator questions are presented in the codebook (see appendix 1).
Furthermore, the coding identified the types of frames used in a selected article. Based on the theory section of this paper, the six frames are identified are; the humanitarian frame,
16 positive economic frame, negative economic frame, security frame, nationalistic cultural frame and multicultural frame. These frames are identified on the basis of three to five ‘yes’ and ‘no’ questions which together form a frame (See appendix 2). Because these questions are dichotomous, a Mokken Scale analysis is done in order to determine the reliability of the identified frames. “The main purpose of the Mokken model is to validate an ordinal measure of a latent variable: for items that satisfy the criteria of the Mokken model, the sum of the responses across items can be used to rank respondents on the latent trait. The Mokken scale generally allows keeping more important items.’’ (Hardouin, et, al., 2011, p. 31). The items are the questions that together form the frame, thus by doing a Mokken Scale analysis, the most important questions are determined. Mokken (1971), considered a scale weak if .3 ≤ H < .4, moderate if .4 ≤ H < .5, and strong if H > .5 (Van der Ark, 2007, p. 4). Thus when placed on a scale, the items together need to meet the reliability criteria of LoevH = ≥ 0.30. In order to meet this criterion, questions 4,5,8,9,13,14,15,19,20,23 and 24 were excluded from the analysis. Because these questions received almost no positive response, they conflict with the assumption of monotonicity, which means “that the number of respondents having R−j = r may be too small for an accurate estimation of P (Xj ≥ x|R−j = r).’’ (Van der Ark, 2007, p.4-5). In other words, the number of articles with answer ‘yes’ for the questions 4,5,8,9,13,14,15,19,20,23 and 24 were simply too less, ‘yes’ < 5, to include them in the analysis without making concessions for the reliability of the selected frames. Table 2.1 shows the results of the Mokken Scale analysis performed in Stata. The Mokken Scale analysis shows which questions are included per frame in order to be reliable. The relevant questions that were used to operationalize each frame are presented here, the other questions are presented in appendix 2.
Humanitarian frame: Q1: The article refers to immigrants or migration in respect with ‘human rights’ or the ‘rule of law’ (M=0.33, SD=0.47). Q2: The article refers to immigrants or migration in terms of ‘fairness’ (M=0.10, SD=0.30). Q3: The article refers to immigrants or migration in terms of ‘equality’ (M=0.04, SD=0.19).
Positive economic frame: Q6: Migration is viewed as ‘economically beneficial’ (M=0.04, SD=0.19). Q7: The article refers to immigrants or migration as attracting ‘a cheap labour force’ (M=0.03, SD=0.16). Q10: The article refers to immigrants or groups of migrants in terms of ‘doing the jobs Dutch people don’t want to do anymore’ (M=0.01, SD=0.12).
17 Negative economic frame: Q11: Migration is viewed as ‘economically costly’ (M=0.17, SD=0.38). Q12: The article refers to immigrants or migration as ‘receiving too many welfare state provisions’(M=0.05, SD=0.22).
Security frame: Q16: Migration or immigrants are linked to ‘security threats’ (M=0.17, SD=0.38). Q17: The article refers to immigrants or migration related to ‘terrorism’ (M=0.02, SD=0.13). Q18: The article refers to immigrants or migration related to ‘criminality’ (M=0.10, SD=0.30).
National cultural frame: Q21: The article links migration or immigrants to a ‘loss of traditions’ (M=0.02, SD=0.13). Q22: The article refers to immigrants or migration negatively in terms of ‘flows of immigrants’ (M=0.14, SD=0.35).
Multicultural frame: Q25: The article refers to immigrants or migration in terms of ‘tolerance’ (M=0.01, SD=0.11). Q26: The article refers to immigrants or migration in terms of ‘being part of the Dutch culture’ or ‘enhancing the same values’ (M=0.03, SD=0.16).
Table 2.1 – Mokken Scale results.
Consisting of Questions
Scale significance Loevinger H
Frames: Humanitarian Q1, Q2, Q3 0.017*** 0.53** Positive Economic Q6, Q7, Q10 0.017*** 0.55** Negative Economic Q11, Q12 0.017*** 0.79** Security Q16, Q17, Q18 0.008*** 0.75** Nationalistic cultural Q21, Q22 0.017*** 0.42* Multicultural Q25, Q26 0.017*** 0.49*
* Scale is considered moderate ** Scale is considered strong *** Scale is significant
18 The results of the Mokken scale analysis are used to build reliable frames, based on two or three items per frame. In total, there are 345 items answered positively. This does not mean that almost all 360 articles contained one positive ‘yes’ answer to one item of a frame, because in theory one article could contain 15 positive answers to 15 items. However in most cases only one ‘yes’ item occurred, out of two or three items of a frame. In 251 cases, only one item per frame was answered positively. However, in many articles, one positive item of more than one frame occurred. For example, many articles contained one positive answer to Q1, about the rule of law, in combination with one positive answer to Q17, perceiving migration as a ‘threat’. The humanitarian frame is dominant in news articles, a total of 128 items were answered positive, N=128. The security frame was also commonly used, a total of 67 items were answered positive, N=67. For the negative economic frame, 65 items were answered positive, N=65. For the national cultural frame, 55 items were answered positive, N=55. The positive economic frame, with 19 positive answers, N=19 and the multicultural frame, with 9 positive answers, N=9 are almost never found in Dutch newspaper articles. The questions, scales, mean and standard deviation of the six frames, and of all the specific items, are presented in table 2.2.
Table 2.2 Items selected and frames
Mean and standard deviation
M SD Minimum Maximum
Humanitarian frame 0,15 0,23 .00 1.00
Q1, Human rights/rule of law 0.33 0.47 .00 1.00
Q2, Fairness 0.10 0.30 .00 1.00
Q3, Equality 0.04 0.19 .00 1.00
Positive economic frame 0.03 0.12 .00 1.00
Q6, Economically beneficial Q7, Cheap labour force Q10, Doing not wanted jobs
Negative economic frame
Q11, Economically costly Q12, Welfare state provisions
Security frame 0.04 0.03 0.01 0.11 0.17 0.05 0.10 0.19 0.16 0.12 0.25 0.38 0.22 0.22 .00 .00 .00 .00 .00 .00 .00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
19 Q16, Linked to threats
Q17, Related to terrorism Q18, Related to criminality
National cultural frame
Q21, Loss of traditions Q22, Flows of immigrants
Multicultural frame
Q25, Tolerance
Q26, Part of Dutch culture
0.17 0.02 0.10 0.08 0.02 0.14 0.02 0.01 0.03 0.38 0.13 0.30 0.20 0.13 0.35 0.11 0.11 0.16 .00 .00 .00 .00 .00 .00 .00 .00 .00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Table 2.2 shows the results of the combined frames and their different items. The mean shows how much framing the news articles contained. The humanitarian frame (M=0.15, SD=0.23), the Negative economic frame (M=0.11, SD=0.25) and the security frame (M=0.10, SD=0.22), are most commonly used in Dutch newspapers. In comparison with the national cultural frame (M=0.08, SD=0.20), positive economic frame (M=0.03, SD=0.12) and the multicultural frame (M=0.02, SD=0.11), which are used scarcely. There are four individual items that are specifically interesting to mention, because they are much more commonly used than the other items. Within the humanitarian frame, but also in general, most of the news articles referred to migration or migrants in relation to human rights or the rule of law (M=0.33, SD=0.47). Within the negative economic frame, most of the articles referred to migrants or migration as economically costly (M=0.17, SD=0.38). Within the security frame, migrants or migration is mostly perceived as a threat (M=0.17, SD=0.38). And finally, migrants or migration is often referred to in terms of flows of immigrants (M=0.14, SD=0.35).
The total number of articles coded is N=360, which are divided amongst the newspapers with De Telegraaf with the lowest amount of news articles coded N=59, and the NRC Handelsblad with the highest amount N=83.
Results
This section presents the results of the conducted content analysis. Each relevant analysis that contributes to answering the sub-questions are mentioned in this section. First, in order to answer sub-question Q1, various descriptive results will be shown. Second, in order to answer sub-questions Q2 and Q3, the results of two paired samples t-tests and two different one-way ANOVA analyses will be analyzed. Finally sub-questions Q2 and Q3 will be further explored
20 with the results of the conducted multiple regression analyses. This is done in order to show significant relations, when controlling for all other independent variables in the regression model. The regression is conducted with the frames as dependent variables and newspapers, immigrant types, amount of words and page number as independent variables.
The first sub-question is (Q1): Do newspapers preferably write about asylum seekers and refugees in comparison with labour and family migration? The six different types of immigrants expected to be found in news articles, were specified on the basis of historical and theoretical arguments. However, both student migration and post-colonial migration were not mentioned once in the 360 articles about migration or immigrants. In order to find significant differences between immigrant type occurrence in newspapers, a paired samples t-test is conducted. The results are shown in table 3.1.
Table 3.1: Paired samples t-test on differences between immigrant type in newspapers.
Paires of immigrant type Mean Difference
SD SE t (359) p
Pair 1: Labour migrants – Asylum seekers
-0.31 0.64 0.03 -9.290 0.000***
Pair 2: Labour migrants – Refugees
-0.35 0.65 0.03 -10.280 0.000***
Pair 3: Labour migrants – Family migrants
0.06 0.37 0.02 2.856 0.005**
Pair 4: Asylum seekers- Refugees
-0.04 0.93 0.05 -0.795 0.427
Pair 5: Asylum seekers- Family migrants
0.37 0.56 0.03 12.447 0.000***
Pair 6: Refugees – Family migrants
0.41 0.57 0.03 13.573 0.000***
Paired samples t-test statistic results (p < .05*. p < .01**. p < .001***)
Table 3.1 shows the significant differences between immigrant types occurrence in newspapers. Most significant differences are found between labour migrants and asylum seekers and refugees, p, <.001. And family migrants and asylum seekers and refugees, p, <.001. Less
21 significant are the differences between labour migrants and family migrants, p, <.01. And not significant are the differences between asylum seekers and refugees. The mean differences are most prominently found between labour migrants and asylum seekers, (Mdifference = -0.31, p <.001) and refugees, (Mdifference = -0.35, p <.001). When comparing asylum seekers with family migrants, this resulted in (Mdifference = 0.37, p <.001). When comparing refugees with family migrants, this resulted in (Mdifference = 0.41, p <.001). These results show that newspapers write significantly more about asylum seekers and refugees, in comparison with labour and family migrants.
In order to get a more complete picture about the immigrant types newspapers prefer to write about, some descriptive results are shown in table 3.2. These results show for each newspaper separately, how many articles they wrote about labour migration, asylum seekers, refugees or family migration.
Table 3.2 Type of newspaper * immigrants Type of immigrant Labour migration Asylum seeker Refugees Family migration Total Type of newspaper NRC Handelsblad 16 21 42 4 83 De Volkskrant 5 28 38 4 75 De Telegraaf 3 31 23 2 59 Trouw 9 33 35 5 82 Algemeen Dagblad 2 35 24 0 61 Total 35 148 162 15 360
Table 3.2 shows the different types of immigrants that occur in different newspapers. There is a significant relation between type of newspaper and type of immigrant, Pearson’s chi-square (12) = 17,12, p <.005. However, the assumptions that cells need to be greater than 5, is violated in 8 cells, which makes the test less reliable. In general, refugees are most commonly referred to in news articles, N=162. Asylum seekers are also referred to in many news articles, N=148. However, labour migration N=35 and family migration N=15, are scarcely used in these five Dutch newspapers.
22 of the specific Dutch newspapers. NRC Handelsblad writes more about labour migration, N=16 in comparison with the other newspapers, N= ≤9. With regard to asylum seekers, the biggest difference is between Algemeen Dagblad, N=35 and NRC Handelsblad, N=21. NRC Handelsblad, N=42, De Volkskrant, N=28 and Trouw, N=35 write more about refugees in comparison with Algemeen Dagblad, N=24 and De Telegraaf, N=23. Finally, when looking at family migration, Trouw pays attention to this type of migrants in five news articles, N=5, while Algemeen Dagblad does not write about family migration at all, N=0.
Thus, table 3.1 shows that Dutch newspapers give significantly more attention to a relatively small group of immigrants, asylum seekers and refugees. Table 3.2 provides insight in how these differences are divided amongst the five biggest Dutch newspapers. This leads to the question whether or not these immigrant groups are also framed differently within these five Dutch newspapers?
The second sub-question is (Q2): Do different newspapers make use of different types of frames when framing migration? The descriptive results of table 2.2 show that the humanitarian frame, the negative economic frame and the security frame are mostly used in Dutch newspapers. The positive economic frame and the multicultural frame are hardly used in the news articles of the different newspapers. Within the most common used frames, three specific items are notable because they occur most often; (Q1), referring to immigrants in relation to the rule of law, (Q16), viewing immigrants as economically costly and (Q11), perceiving immigrants as a threat. However, these results do not show whether or not the frames differ significantly in news articles about immigration. Therefore a paired samples t-test is conducted. The results are presented in table 3.3.
Table 3.3: Paired samples t-test on differences between frames used in news articles.
Paires of frames Mean SD SE t (359) p
Pair 1: Humanitarian frame – Positive economic frame
0.13 0.27 0.01 8.875 0.000***
Pair 2: Humanitarian frame – Negative economic frame
0.04 0.33 0.02 2.485 0.013*
Pair 3: Humanitarian frame – Security frame
23 Pair 4: Humanitarian frame –
National cultural frame
0.07 0.30 0.02 4.659 0.000***
Pair 5: Humanitarian frame – Multicultural frame
0.14 0.26 0.01 9.872 0.000***
Pair 6: Positive economic frame – Negative economic frame
-0.09 0.29 0.02 -5.591 0.000***
Pair 7: Positive economic frame – Security frame
-0.07 0.25 0.01 -5.422 0.000***
Pair 8: Positive economic frame – National cultural frame
-0.05 0.23 0.01 -4.541 0.000***
Pair 9: Positive economic frame – Multicultural frame
0.01 0.16 0.01 0.943 0.346
Pair 10: Negative economic frame- Security frame
0.01 0.32 0.02 0.867 0.387
Pair 11: Negative economic frame- National cultural frame
0.03 0.30 0.02 1.952 0.052
Pair 12: Negative economic frame- Multicultural frame
0.09 0.28 0.01 6.307 0.000***
Pair 13: Security frame- National cultural frame
0.02 0.24 0.01 1.071 0.285
Pair 14: Security frame- Multicultural frame
0.08 0.24 0.01 6.083 0.000***
Pair 15: National cultural frame- Multicultural frame
0.06 0.22 0.01 5.395 0.000***
Paired samples t-test statistic results (p < .05*. p < .01**. p < .001***)
The results of table 3.3 show significant differences between the use of the humanitarian frame and all other frames in news articles. However, most significant differences are found for the humanitarian frame and the positive economic, national cultural and multicultural frames, p, <.001. For the security frame the significance is, p, <.01, and for the negative economic frame the significance is p, <.05. The mean differences are most prominently found for the humanitarian frame and the positive economic frame, (Mdifference = 0.13, p <.001), and the multicultural frame (Mdifference = 0.14, p <.001). This means that newspapers make
24 significantly more use of the humanitarian frame in comparison with the positive economic and multicultural frame. Furthermore, the results of table 3.3 show significant differences between the positive economic frame and all other frames p, <.001, except for the multicultural frame which is not found to be significant. The mean differences are most prominently found for the positive economic frame and the security frame, (Mdifference = -0.07, p <.001), and the negative economic frame (Mdifference = -0.09, p <.001). This means that newspapers significantly make more use of the security frame and the negative economic frame in comparison with the positive economic frame.
Finally, table 3.3 show significant differences between the multicultural frame and the security, national cultural and the negative economic frame, p, <.001. However, the M differences are only small with for security (Mdifference = 0.08, p <.001), for national cultural (Mdifference = 0.06, p <.001) and for negative economic (Mdifference = 0.09, p <.001). Concluding, the results of tables 3.3 show significant differences in the use of the humanitarian, positive economic and multicultural frames in newspapers. The biggest effects were measured for the humanitarian frame in comparison with other frames. The most significant effects were found for the positive economic frame in comparison with other frames. However, these results do not show the differences between the use of framing in newspapers when they are specifically compared with each other. Therefore a one-way ANOVA needs to be conducted.
One-way ANOVA on groups of newspapers and frames
A one-way ANOVA is conducted to compare groups of newspapers and the frames they use. The groups of newspapers differ significantly between the use of the humanitarian frame (F = 2.775, df = 4, p = 0.027), positive economic frame (F = 3.157, df = 4, p = 0.014) and security frame (F = 3.413, df = 4, p = 0.009). The other frames are not found to be significant and are therefore left out of the analysis, because this means that they are more or less equally divided amongst the different newspapers. The test of homogeneity of variances was significant for all frames with a Levene Statistics for the humanitarian frame, F (4, 355) = 6.53, p= 0.000, for the positive economic frame F (4, 355) = 13,56, p =.000 and for the security frame, F (4, 355) = 11.72, p = .000. In order to further explore the differences between the groups, a post-hoc Bonferroni is conducted. The Bonferroni correction is a method used to counteract the problem of multiple comparisons. It is one of the most strict methods to control the familywise error rate. The results of this test are presented in table 3.4. For the mean and standard deviation of each newspaper separately, see appendix 4.
25
Table 3.4 Bonferroni Post Hoc results on frames and newspapers Newspaper group (i) Newspaper group (j) (M difference) SE p Type of frame Humanitarian NRC Telegraaf 0.13 0.04 0.017*
Positive Economic Volkskrant Telegraaf 0.07 0.02 0.019*
Security Telegraaf Trouw 0.10 0.04 0.03*
Post-hoc Bonferroni test statistic results (p < .05*. p < .01**. p < .001***)
The results presented in table 3.4 show a significant difference between the NRC (M=0.20, SD =0.26) and Telegraaf (M=0.07, SD=0.18), in their use of the humanitarian frame. The (Mdifference = 0.13, p <.05), means that the NRC makes significantly more use of the
humanitarian frame in their news articles in comparison with Telegraaf. When comparing De Volkskrant (M=0.07, SD=0.22) with de Telegraaf (M=0.00, SD=0.00), it makes significantly more use of the positive economic frame (Mdifference = 0.07, p <.05). Finally De Telegraaf
(M=0.16, SD=0.27), uses significantly more security frames when comparing with Trouw (M=0.06, SD=0.16), with as a result (Mdifference = 0.10, p <.05).
The fact that NRC Handelsblad make significantly more use of the humanitarian frame in comparison with De Telegraaf, is in line with expectations and can be explained because of differences between liberal and left-right oriented newspapers. However, the level of significance is only p, <.05. The results concerning the use of positive economic framing in newspapers is also in line with theoretical expectations. The positive economic frame is significantly more used in De Volkskrant when comparing with De Telegraaf. This difference can be explained based on expectations that left newspapers write more positively about the economic consequences of migration than right newspapers. However, the significance level of all findings of framing in newspapers was considered only small p, <.05.
The findings presented in Table 3.4 lead to the conclusion that newspaper only marginally differentiate in their use of framing when writing about migration or immigrants. Table 3.3 shows that the most significant differences found are between the humanitarian, the positive economic and the multicultural framing in newspapers articles. Finally, the results of the Bonferroni post-hoc test shows more negative framing of De Telegraaf in comparison with
26 the framing of NRC, Trouw and De Volkskrant for the humanitarian, positive economic and security frames.
The third sub-question is (Q3): Which frames are used differently when writing about asylum seekers, refugees, labour migrants, family migration, student migration and post-colonial migration? When looking into the different frames used in news articles, three frames dominate; the humanitarian, negative economic and security frame. When looking into the different immigrant types newspapers write about, two types dominate; asylum seekers and refugees. However, is there also a significant difference between the six different frames used, when writing about different types of immigrants? In order to answer this question properly, a one-way ANOVA needs to be conducted.
One-way ANOVA on groups of immigrants and frames
A one-way ANOVA is conducted in order to find differences between immigrant types and the use of framing. Earlier results show that only four types of immigrants occurred in the news articles, therefore only these four types are included in this one-way ANOVA. There are significant differences between the type of immigrant and the use of the humanitarian frame (F = 2.749, df = 3, p = 0.043), the positive economic frame (F = 3,647, df = 3, p = 0.000) and the multicultural frame (F = 2.775, df = 3, p = 0.013). The other frames were not found to be significantly different, which means that there are more or less no differences between immigrant types and the negative economic, security and national cultural frames, therefore these are left out of the analysis. The test of homogeneity of variances was significant for all frames with a Levene Statistics for the humanitarian frame, F (3, 356) = 3.71, p = 0.012 for the multicultural frame, F (3, 356) = 13.91, p = .000 and for the positive economic frame F (3, 356) = 57,70, p =.000. The Bonferroni Post Hoc test is done in order to further explore the differences between the use of framing and immigrant type, results are presented in table 3.5. For the mean and standard deviation of each immigrant type separately, see appendix 4.
Table 3.5 Bonferroni Post Hoc results on frames and immigrant type Immigrant group (i) Immigrant group (j) (M difference) SE p Type of frame
27 Positive
Economic
Labour Asylum seekers Refugees Family 0.17 0.15 0.13 0.02 0.02 0.04 0.000*** 0.000*** 0.002**
Multicultural Labour Asylum seekers Refugees 0.06 0.05 0.02 0.02 0.008** 0.029*
Post-hoc Bonferroni test statistic results (p < .05*. p < .01**. p < .001***)
The results of table 3.5 shows significant differences in the way labour migration is framed in comparison with the three other types of migration. When news articles write about labour migration (M=0.09, SD=0.17), the humanitarian frame is significantly less used in comparison with family migration (M=0.29, SD=0.31), resulting in (Mdifference = -0.20, p <.05). Because of
the small groups, labour N=35, and family N=15, the question whether or not the results presented here are reliable, could be raised. However the results are significant on a level p, <.05 and this finding is in line with expectations based on the characteristics of labour and family migration. The expectation based on the theory predicts family migration to be linked with specific rules of law, because this type of migration considers specific human rights to family life (Hollifield, et, al., 2014). It is therefore not surprisingly that the humanitarian frame is more commonly found in news articles about family migration.
With regard to the multicultural frame, this type of framing is positively related with labour migration (M=0.07, SD=0.21), in comparison with asylum seekers (M=0.01, SD=0.06) and refugees (M=0.02, SD=0.10). For labour in comparison with asylum seekers, the result is (Mdifference = 0.06, p <.01). For labour in comparison with refugees, the result is (Mdifference =
0.05, p <.05). This means that labour migration is often more portrait as multicultural in comparison with asylum seekers and refugees. However, these results are questionable because there is no underlying theoretical expectation and the number of labour migrants N=35, and amount of multicultural frames used N=9, is small.
Finally, there is a significant difference in the use of the positive economic frame between labour migrants (M=0.17, SD=0.27) and asylum seekers (M=0.00, SD=0.00) resulting in (Mdifference = 0.17, p <.001). Also between labour migrants and refugees (M=0.02, SD=0.10)
resulting in (Mdifference = 0.15, p <.001). And finally between labour migrants and family
28 positive economic frame is significantly more used when writing about labour migration, in comparison with all other immigration types. These results are significant p, ≤.001, and all point in the direction of more positive framing of labour migrants in news articles. There are no significant effects found of labour migrants in relation to negative economic framing.
Concluding, the findings presented in table 3.5 show significant differences between immigrant types and the use of framing in newspaper articles. The most significant differences are found between labour migrants and all other types of migration. For the humanitarian frame, the significant difference is found between more use of the humanitarian frame for family migration in comparison with labour migration. For the multicultural frame, the significant results show more use of the multicultural frame for labour migration in comparison with asylum seekers and refugees. Finally the most significant relation found, is between making more use of positive economic framing in newspaper articles about labour migrants, in comparison with newspaper articles about all other types of immigrants.
Multiple regression analysis
A multiple regression analysis is performed to find significant effects on the different frames used, when controlling for all other independent variables. The regression is done exploratory, because the dependent variables, the frames, are only three or four point item scales and therefore not completely suitable for a multiple regression analysis. The model is conducted for all six types of frames, the significant results of the different frames are presented in this section.
Model 1, presented in table 3.6, show the results of the multiple regression analysis with as dependent variable the humanitarian frame and as independent variables; asylum seekers, refugees, labour migrants, NRC, De Volkskrant, De Telegraaf, Algemeen Dagblad, amount of words and page number. The model is significant, F(9,359) = 5.499, p <. 001. The variables family migrants and Trouw are left out of the analysis in order to serve as a reference category. The humanitarian frame can be predicted on the basis of the independent variables. However, the strength of the predicted model is weak; only 10 percent of the variation of the humanitarian frame can be predicted on the basis of the independent variables.
Table 3.6 Humanitarian frame
Model 1 b SE b b* Sig.
