Type of locality and the public support of European Integration throughout the migration crisis

(1)

University of Leiden

Public Administration MSc

Sándor Kozák

s2573067

Dr. Dimiter Toshkov

2020.

Type of locality and the public support of European

Integration throughout the migration crisis

(2)

Introduction

Public opinion is considered as a standard method of quantifying preferences of a large population. This can not only help to applied social sciences, but it is also essential in democrat politics (Lavrakas, 2008). Public opinion recordings on the perception of the European Union (EU) dates back to the 1950s, and since then it’s relevance has grown from an informative report to a tool of policy formulation and governance in the EU institutions (Priestley, 2012). This connects public opinion to European integration, that is often viewed and conceptualized as the constant enlargement of the EU, while it is actually about shifting power between and through multi-level decision-making (Zürn, 2012). To eliminate the existing democratic deficit of the international level, these processes are further opened to the public (Papadopoulos, 2010). In the post-Maastricht period, the public has gained enough relevance to be included in these processes. Thus, measuring the support of the tendencies that are shaping the EU has become fundamental.

Public support for European integration seems to be on the rise in the past decades. With the appearance of consistent surveys such as the European Social Survey or the Eurobarometer, it is possible to highlight both long-term and short-term effects between variables. In terms of theory there are multiple explanations for differing support (Hobolt & de Vries, 2016). These can range from post-materialist, ideology-based all the way to economic measurements. These explanatories do not exist in a vacuum, multiple of them can clarify the same effect, which leads to the implication that while empirical research can feature these interests as a shaping factor, it cannot prove them to be the universal reasoning (Kanthak & Spies, 2018; Ejrnæs & Jensen, 2019). It is known that the two main drivers that shape public opinion towards European Integration are utilitarian explanations such as societal classes and type of labour, and identity based, that is the cultural and traditional background along political values (Hobolt & de Vries, 2016). The best way to explain and see how these different explanations behave is to look at specific time periods where a sharp change happened on the supranational level, which also affects the everyday life of the citizens. Therefore, we know several theoretical assumptions that lead to greater explanations, but due to the lack of certainty it is paramount to constantly come up with new ideas to prove these variables.

This thesis works on two levels: it strives to introduce a new possible statistical association between the type of locality and public support on European integration. Furthermore, it wants to demonstrate the differences between the rural and non-rural populations’ support. Therefore, the thesis has two research questions. The first is directly related to the public support studies that are measuring the factors behind public opinion: “Can the type of community be a strong explanatory factor in public support?”. This requires additional statistical analysis, which will be detailed in the followings. The second research question is „how did the European migration crisis affect the rural and urban populations’ public opinion with regards to European integration?”. Choosing the migration crisis

(4)

as an event makes the research more nuanced but also helps to highlight how the crisis acted as a major effect. The research question uses it as a changing point to find out how a quite identity-based issue changes these interests behind public support (Hobolt & Wratil, 2015).

The characteristics of a certain type of locality is assumed due to cultural, geographical and economic differences that shape the identity of the citizens (Marks, 1999). The concept of rural areas, for instance, can be viewed as a geographical environment that attracts capital from different economic backgrounds compared to big urban areas (Dijkstra & Poelman, 2014). Others, look at it as a type of community that resembles a unique lifestyle, which has its own cultural heritage (Öğdül, 2010). On the other side of the spectrum, large urban areas are often described as the metropolitan focal point of multiculturalism, and international interest, especially in the European Union (Antalovsky et al., 2005). Furthermore, there is a middle ground, the middle-sized cities or small towns, which also introduce unique territorial identities. Regardless the way of defining it, the type of communities shapes the identity and the behaviour of the people. In this thesis, the main type of locality that is examined and challenged against the others, is the rural area. There, the perception of post-materialistic values, the connection to the local community, the assumed conservative traditions all shape someone’s perception towards European integration (Atkin, 2003). This thesis presents all characteristics of “rurality” (as it was defined by Cloke (1977)), and try to measure how differently the people living in such communities perceive international decision-making compared to the “urban” areas (in the future, urban areas also include middle and small sized cities).

Due to the wars and conflicts in Middle Eastern countries and deprivation in Sub-Saharan Africa, already in 2014, a large number of people started traveling to the outer borders of the EU. Then, more than 600 thousand refugee statuses have been requested citing the Dublin III Regulation of 2013 (No 343/2003). Around 180 thousand refugees were accepted (EUROSTAT, 2015). The number of asylum applications grew to an all-time high in 2015 when this number exceeded 1.2 million cases (EUROSTAT, 2020). The migration crisis first started an institutional problem in 2015 with handling the mass influx of refugees and migrants at the EU borders in the south and with the appearance of the Western Balkan route. In most countries, temporary border control was reintroduced (Colombeau, 2019). The crisis became more political when the quota system has been introduced which is an EU wide distribution of asylum seekers. The proposal met with vehement opposition by the leaders of the Visegrád group. The relocation has started however in 2016 the quota system was abandoned (Gotev, 2017). The framing of the issue led to different behaviours in each country: the national medias created an effect which exacerbated the issue (Trilling, 2019). Anti-migrant policies, decreasing public support and racist sentiments started flowing, which later led to the elective success to right-wing populist parties in several countries (Ratković, 2017).

(5)

There is no better time period to measure rurality than the immediate and further time frame of the migration crisis. While general immigration is a quite accepted or at least an unusual trend in today’s globalized European Union, the migration towards rural areas is perceived as a less common habit, which usually accommodates large level of hostility from the locals. There is no developed infrastructure for it and migrants have to face the discrimination of citizens that want to keep the homogeneity of the area (Fabrizio et al., 2019). In certain countries the national media framed an issue in such a way, which created clear resentment in the rural people especially (Spike, 2017). On the other hand, urban areas, specifically big cities’ perception ranges in economic, social and political areas. Economically immigration is considered as a highly positive pattern, while politically and socially it can create tensions (World Economic Forum, 2017. pp. 30). Nevertheless, in large European cities, the cosmopolitan values are prominent, therefore it could be assumed that the support in the EU was not affected on the same level citizens from big cities (Maxwell, 2019). Thus creates relevance in of itself: the research could be a valuable tool for not only public opinion studies, but it enriches the knowledge regarding rurality and support in the EU during the migration crisis. Overall, it can be implied that the rural population perceived the migration crisis in a different way which still affects their support.

The structure of the thesis can be explained in 5 major steps. As the research questions imply, the thesis is based on several theoretical assumptions that have been forming the studies of European Integration. First, it is necessary to highlight what we do know about these processes. Thus, to understand why being rural or a certain crisis could affect someone’s views on such things as the authoritybei allocation of a supranational system, it is necessary to conceptualize certain terms, and highlight different models. The best way to achieve this is through a literature review, where we highlight all what we currently know about research on public opinion, about all the variables and effects. Then, the hypotheses will be detailed. The migration crisis has effects on both the short-term and long-term future of the citizens, thus it is needed to be analysed both accordingly. Subsequently, the rurality as a factor requires the additions of other variables that might be strong confounders and might explain the causal link well. Since the thesis uses two explanatory research question, the hypotheses will also be divided based on the dimension it tries to explain.

Then, the operationalization will be portrayed, which includes data collection a variable description. The thesis uses Eurobarometer surveys as datasets and considers the time frame of 2012-2018. Thus, it is necessary to highlight the main strength of Eurobarometer research and also the areas which will be used in the analysis. In the next chapter, the results will be shown which are divided into two main categories. First, basic descriptive analyses will be shown to explain whether there is a discrepancy among the the rural and non-rural support and if so, what is the direction of it. Two dependent variables will be used in this case and both of them will be put to the test. Then, the association will be analysed. Later, a linear regression analysis shows the main drivers of public support. Lastly, a

(6)

conclusion will be written, in which the answers to the hypotheses and the general patterns will be discussed.

Theoretical Foundation

The goal of the thesis is to add another variable into the group of the already existing ones. The implication is that being rural can be an important factor either alone or as a supplementary variable. Explanations such as educational level, age, inflation or identity are somewhat rooted in the research of public support and they are backed up from a number of directions (Gabel, 1998). The thesis argues that living in a rural environment shapes people’s views towards European integration just as much as the soon-to-be presented factors. Unfortunately, this implication cannot be proven easily. Using the migration crisis as an event between the time-frame, helps to measure these causal relationships easier and also aids to find out how robust effect being rural or urban has. While it is certainly possible to find differing support in rural areas, the causal link might be caused by other, previously mentioned variables. In order to find out the reasons why rural residents vote differently, all approaches could be applicable (Hobolt & de Vries, 2016). In order to determine whether or not locality is a main explanatory factor or just another usable cofounder, other variables will be added to the analysis the control the link between the dependent and the independent variable.

The groundwork of the thesis are a mix of conceptualization frame and theoretical foundations. Thus, in order to make it more comprehensible, first I will lay out the evolution of approaches to European Integration and how the post-Maastricht era led to the consideration of public support as an important circumstance (Eichenberg & Dalton, 2007). While at first it could be viewed as a marginal matter in the big scheme of things, there are specific findings on how public view can affect supranational legitimacy or even policy-making (Toshkov, 2011). If we consider public support as a base of European Integration, it is also important to explain how a crisis can affect that. During the strike of the Schengen or migration crisis, the public support towards the European Integration was lowered while the support of the common migration policy remained high (Börzel & Risse, 2017). Using the crisis as an event can portray variations of support in subgroups and present basic assumptions on which effects are shaping their opinion. In the case of rurality a general hypothesis on low public support could be the threat of letting the European Union exert power to shape the cultural distribution of their local community (McLaren, 2002). Or is it more based on measuring the costs and benefits based on the performance of the national and the supranational level (Sánchez-Cuenca, 2000)?

Later, I will summarize what public opinion really is and why there is an observable disparity between certain class of people’s support towards the European Union (Hobolt & de Vries, 2016). Gabel (1998) highlights five major theories that could explain public support. While the thesis won’t be able to explain which theory explains the rural population’s behaviour, it regardless helps to legitimize the

(7)

variable’s usage. Most specifically, it will be necessary to look at empirical and theoretical studies that are basing public opinion on utilitarian or identity explanations (Hooghe & Marks, 2004).

The concept of European Integration and the postfunctionalist application

Ernst B. Haas (1958) viewed regional integration as loyalty towards a new “center”. When we consider the EU (then European Community) as a new center, the concept of European integration is about equalizing the field for everyone to achieve collective benefits for member-states, albeit it requires not only immense formal authority but also a deep insight on how it affects nation-states and citizens alike (Genschel & Jachtenfuchs, 2016). To achieve these means, the supranational system has to gain several decision-making and regulatory power from the national level (Hooghe & Marks, 2015). There are multiple theoretical approaches that touche upon sources and effects of European Integration (Hooghe & Marks, 2008). Neofunctionalism believed that the whole reason for integration is that at the basic national level, the required solutions cannot be achieved to problems due to them being on different levels. That makes the system inefficient and it cries for a jurisdictional reform that would somewhat rewrite the territorial integrity of the state. This approach was mainly concentrating on economic considerations and argued that international organizations produce functional benefits to nation states. (Hooghe & Marks, 2009. pp. 3-4) While neofunctionalism believed that the integration is initiated from the supranational level, the intergovernmentalist theory believes that the reallocation of authority is based on mutually profitable co-ordination between states and business groups (Moravcsik, 1993). Scharpf (1998) highlighted considerable differences between the two approaches since he argues that supranational decision making and European law is a type of negative integration, while intergovernmental policy-making is positive. What is common between the two approaches is that both imagined decision-making behind closed doors and there was a “permissive consensus” between the elite and the citizens in the integration’s direction and processes (Genschel & Jachtenfuchs, 2016. pp. 52-54). That means that this informal agreement between the elite and the public let the European integration process to be unconstrained and unchecked by the mass (Eichenberg & Dalton, 1993).

The turning point occurred in 1992, after the adoption of the Maastricht Treaty which redefined the main objectives of key EU institutions. At this point, the European Union became more than just an economic matter in the eyes of the public (Eichenberg & Dalton, 2007. pp. 139). The appearance of majority voting in the European Council, the strengthened European Parliament (EP) which began to be directly elected and the extended competences led the EU to a more publicized and subsequently, politicized world (Genschel & Jachtenfuchs, 2016). The effects were immediate starting with the Danish rejection of the treaty in 1992 to the results of the first EP elections and the appearance of Euroscepticism: the need for public legitimacy towards European Integration became crucial (Hobolt, 2012). In 1991 the net public support declined to a then lowest point (Eichenberg & Dalton, 2007. pp. 42). This “constraining dissensus” led to the re-evaluation of the theories regarding to regional integration, which created postfunctionalism. In this theory, public opinion is regarded as an important

(8)

factor of European integration that can cause further cooperation but also can be caused by existed integration. They argue that the increased politicization of the decision-making process led to a more involved public. However politicization can create not only integration but also constraints (Börzel & Risse, 2017. pp. 5-6). The media became an influencer which is powerful enough to shape views towards the EU that has the possibility to lead to the involvement of national governments in the process of context creation (Hobolt & Wratil, 2015).

Public support is measured through surveys, experiments, and interviews. The opinion towards European integration in surveys is the most prominent where questions related to trust, identity, and membership are current (Hobolt & de Vries, 2016). For a long time, it was argued that public opinion cannot affect EU policy-making. However, empirical research argues otherwise. In Toshkov (2011) it is stated that the number of important adopted policies clearly resonated with the then-current public support. Even though this correlation disappears in 1995 the high number of policies could have created oversaturation in the eyes of the public which reverses the effect. Public legitimacy is needed especially during crises, which can be seen in Hobol & Wratil (2015), where it is highlighted that in the eurocrisis, public support was essential. In order to analyse public opinion of European integration, it is cardinal to see through postfunctionalist lenses. It leads us to realize that in a system where the European Union has an effect on identity, redistribution and sovereignty we have to research how different subgroups respond to the policy-making of the EU. This is extensively imperative if we include the important role of the media and national parties in this picture. But what are the factors that can affect public opinion? What are the sources of beliefs and how can we approach these with efficiency?

The effect of crises on public support

An international crisis can unconditionally shape the authority allocation of a supranational system. During these times, reaching consensus on letting the international level work as a coping mechanism is crucial (Herszenhorn et al., 2018). On the level of the European Union, the financial crisis was dealt with poorly, the supranational system did not have the proper coping mechanisms, and Member States were left uncertain. Later, the Eurozone crisis challenged the whole supranational system and the initial failure of the Economic and Monetary Union (EMU) created a critical juncture point in the history of the European economic integration (Verdun, 2015). The Member States were highly threatened by the crisis, and due to the strong linkages, interdependence and uncertainty. bailing out was too costly. Hence, they eventually transferred authority to supranational EU institutions such as the Commission and other newly created systems and granted more power to existing institutions (Schimmelfennig, 2018.; Dehousse, 2016). In the case of the Schengen crisis, that the thesis is mainly concentrating on, different patterns were apparent. During both crises, the events were highly politicized on the domestic level, making the citizens aware of what is truly happening. The neofunctionalist respond to this, was depoliticizing the issue, by creating negative integration processes, meaning non-majoritarian institutions (Börzel & Risse, 2017). This did not work in the case of the migration crisis.

(9)

Both Frontex and the European Asylum Support Office lacked the necessary competences to deal with the issue. The exist-cost was low, and the countries seemed to be able to deal with the issue in unilateral means. (Schimmelfennig, 2018). Deliberative intergovernmentalism examines how this shift affected integration especially during a crisis: the Council strives countries to create a consensus that requires individual commitment and arrangements behind the curtains. Due to the linkage of “low and high” politics, the uncertainty of legitimization in the domestic level forces national actors to be in consensus (Fabbrini & Puetter, 2016). Thus, not only the framing of the issue was different, but the “political and economic loss” differed: in the migration crisis everyone had major positions to lose, but the offered solutions were somewhat even more politically lethal to several Central and Eastern European countries (Börzel & Risse, 2017). This created even higher politicization, which lowers the chance of integration significantly (Schimmelfennig et al., 2015. pp. 754).

The reason why I included the eurocrisis into to the explanation is because significant results came in terms of public opinion towards integration through analysing that event. An international crisis creates interest towards the EU due to the increased amount of information that arrives to the public (Hobolt & Wratil, 2015). This can be positive or negative, but it certainly affects public support. What is striking is that during the eurocrisis, public support remained stable: in Hobolt and Wratil (2015) it is seen that the support for the euro (which is not equivalent to European integration, but it is useful to mention) was still high even during the crisis and argues that identity based discources were replaced with utilitarian explanatories. That of course does not mean that one explained all types of public support. Nevertheless, it seemed like in the individual, economic based interest outweighed identity based ones. The switch argued to be founded on the fact that politicization of monetary policies creates more interest in the former than the latter (Hobolt & Wratil, 2015. pp. 252), meaning that identity is still considered as predictor important to measure. It is crucial because others found not contradictory, but different results. Kuhn and Stoeckel (2014) discovered that several variables both utilitarian and non-utilitarian ones are strong variables, but only a few of them are actually significant. Exclusive national identity is actually one of these. On the other hand among significant factors, economic ones can be found as well, such as being unemployed and being in a wealthy, economically healthy Member State. According to Lefkofridi and Schmitter (2015) the eurocrisis created this “savior” image for the European Union which also affected the public support. It created new opportunities of evaluation for both political actors and the public.

There is a lack of studies that analyse how public support was affected by the migration crisis. According to surveys inclusive national identity is growing (Kuhn, 2019). Positive support on EU has been stagnating at a low level since 2007 (European Commission, 2013) and the trust in the European Union has been has been gradually decreasing (Schout & Holderied, 2018). In this thesis, I argue that the Schengen crisis did affect public support and the rural attribute will be easier to highlight if we use both pre- and post-crisis data. Since the Schengen crisis was politicized thorough the national level even

(10)

socio-economic arguments were spilled with national identity and cultural reasoning (Zingg, 2016. pp. 17).

Literature Review

Explanations of public support

Explanations of public support differ in a number of ways, but generally there is a basic division that can be highlighted with Easton (1975) who conceptualized two major type of support towards political institutions. While “diffuse” is more based on long-term ideologies and values, “specific” is more of a short-term utilitarian support based on rational thinking. Gabel (1998) elaborated on this existing distinction and analysed five major theories that can explain different approaches towards the EU. As he stated, the biggest issue with these explanations is that they all strive to answer to the same question. None of them give perfect and unilateral answer, meanwhile all of them can be detected in different significance and with different angles (Kanthak & Spies, 2018). The following models and theories all explain public support on the individual level.

The first theory is based on Inglehart (1970), in which it was argued that the high level of education that appeared in the 20th century led to move post-materialist based thinking, which helps to understand the reasons of European integration, thus people with such mind-set are more likely to support European integration. Cognitive mobilization argues that education overwrites every other existing variables and factors like nationality, class or income. This is a well-regarded and often cited theory, however it has limited empirical foundation (Jacquier, 2012). The second theory is another assumption from Inglehart and it based on political values that are the socioeconomic values that citizens achieved in their pre-adult years. This can cover national identities, ideologies and traditions (Eichenberg & Dalton, 1993. pp. 509-511; Gabel, 1998. pp. 336-337). While political values also assume that post-materialists will support integration more, in reality it is the exact opposite (Anderson & Reichert, 1995. pp. 241).

The third theory is class partisanship that comes from tradition voting behaviour studies. It is related to unconditional party support: what the party says is what the citizens believe (Gabel, 1998). While party support is meaningful in the case of public support towards the EU, it is usually a side variable not the main concern (Toshkov, 2011; de Vries & Steenbergen, 2013). This concept of party support has been revamped over the years and it is mostly analysed through politicization effects and how national actors shape and literate EU policies and events (Sánchez-Cuenca, 2000; McLaren, 2002; Hobolt, 2012). The fourth explanation comes from the support of governments. That is based on the nature of regional integration that is pooling authority from the national level (Hooghe & Marks, 2015). While this approach in itself seems to have small substantive impact (Gabel, 1998. pp. 351), using this thinking led really effective findings. Sánchez-Cuenca (2000) combined this and utilitarian views and laid out remarkable frameworks for other scholars. It is useful to state that according to the distributional

(11)

theory, the greater the influence of supranational institutions the bigger the pro-integration. Also, if an individual is not interested in national politics, but direct and indirect benefits appear from the EU, they become pro-integration. Continuing this thinking, based how badly the national government is performing, the judgement towards the European Union will be more positive (Sánchez-Cuenca, 2000; McLaren, 2002).

Lastly, the utilitarian model argues that people with different socio-economic cultures gain or lose differently from European integration and this is the backbone of the way of them approaching it. This is logical, knowing that the European Union started as an economic based international organization. The economic interest can be approached through multiple angles. Anderson and Reichert (1995) argues that there are direct and indirect economic effects that can change public opinion. The former is in connection with services such as payment or subsidy that is directly sent to the citizen. Thus, it is assumed that people with high-skilled jobs that directly benefit from the European Union will support it. On the other hand, indirect effects are things that are not tangible by the public, but it clearly affects the everyday life such as trade policies (Anderson & Reichert, 1995). Kuhn and Stoeckel (2014) grasp similar distinction, by differentiating support for the EU and support of the EU governance. The former is connected to market liberalization which is highly beneficial for competitive citizens. On the other hand, the latter is the “supranational oversight on markets” that can affect people differently in welfare states or in newly joined states (Kuhn & Stoeckel, 2014. pp. 625). These division style can be connected to the previously highlighted negative and positive integration differentiation (Scharpf, 1998). This individual rationalist cost-benefit analysis also differ in policies. While special policy support means the support of authority allocation to the international level, which is more related to the fourth model, it clearly has a utilitarian side as well. If the economic effects are reaching all the way to the citizen, the effects will be highly different than in economic policies that only indirectly affect them (Eichenberg & Dalton, 1993).

Before continuing, it is imperative to include a newer distinction of theories, which highlight a significant and new approach. Hobolt and de Vries (2016) distinguishes three type of causes of public support. While the utilitarian covers the same, already mentioned characteristics, the identity cause is a combination of political values and support for government. In here, they explain explanatory variables that are “pooling soverignty that potentially erodes national self-detemination by blurring boundaries between national communities” (Hobolt & de Vries, 2016. pp. 420). In the following of this thesis this term will cover most of ideology or cultural based factors. The third category is “cue-taking and benchmarking” that has been an emerging term in the field. This approach covers the effect of the media and national political parties in the citizen’s decision-making. It argues that the European Union is too complex to deal with individuals, thus the abovementioned actors create proxies or cues that citizens rely on (Hobolt & de Vries, 2016. pp. 422-423). This is the model that explains considerable differences between public opinion in different Member States. These can be positive or negative, based on the cue

(12)

availability and the existing cue competition in the country. Furthermore, it can also be based on predisposition that, once again, goes back to political values in Gabel (1998).

Main empirical findings

As it was mentioned utilitarian reasoning and factors of public support have been ruling the research field over the decades. However, utilitarian explanations can be accessed through different angles. For instance, macroeconomic effects that were created by the national elite to deal with issues such as GDP growth within and outside the national level, inflation or unemployment seemed to be great explanatories (Eichenberg & Dalton, 1993, 2007; Gabel & Palmer, 1995; Anderson & Reichert, 1995). This obviously comes from the fact that at first the EU functioned as an economic enterprise, thus the growth of intra-EU trade and the assessment of the gross national products seemed to be the most reasonable ways of explaining public support. Eichenberg and Dalton (1993) found that citizens evaluate the then EC based on effective price controls. Several studies included the effect of inflation as a significant explanation (Gabel & Whitten, 1997; Eichenberg & Dalton, 2007), while unemployment seems to be a moderate but important variable (Anderson & Kaltenhaler, 1996), especially if we use regional levelled unemployment rate (Gabel & Whitten, 1997. pp. 90).

It is time to examine microeconomic variables: Jacquier (2012) summarized the main deterrents of public support. He argues that macroeconomic predictors are less important nowadays, and the main economic factor is the personal potential gaining from trade liberalization. Thus, we turn to basic socio-economic factors that can decide how you will respond to the free-market self of the EU. The general argument is connected to the distribution logic in Hooghe (2003). While the elite looks at European integration as a functionality, meaning that if a policy better fits in the supranational level, then it should be dealt there. On the other hand, while citizens understand this, they are looking at liberalization as an economic insecurity. They are afraid of the Europeanization of policies that are affecting their financial flow directly (Hooghe, 2003). Thus, a quite obvious explanation is that actors who are threatened by liberalization will be against integration (Sánchez-Cuenca, 2000). Thus, skill and industry are significant variables. Farmers, clerks, blue-collar workers are assumed to view public opinion based on their own economic interest which has been elaborated in de Vries and Steenbergen (2013), where it was clear that physical workers are less supported in unification.

Several economic-based explanations and deterrents were highlighted, however, it is necessary to include the national controlling variables before moving to other findings. Due to the multi-levelness of regional integration, it is customary, that citizens tend to blame the national politicians for macroeconomic declines. National traditions are exceptionally important in public support (Eichenberg & Dalton, 1993), since national institutions are based upon different historical trajectories that highly affect the discourse (Hooghe & Marks, 2009. pp. 14). For instance, countries that are rich in labour, it is predicted that the elite will be more Eurosceptic while in reverse it is the exact opposite (Hooghe &

(13)

Marks, 2004. pp. 415.). This is also highly connected to national evaluations: if the national economy does not provide possibilities to these labour classes to thrive, than they will look at the EU as a “alternative”, while in reverse situations, the poor fears to be affected by EU policies directly. Others argue that in fact, high-skilled citizens will oppose integration in welfare states (Brinegar & Jolly, 2005), which somewhat contradicts this line of thinking. That is because in this distinction there are high-skill endowment and low-skill endowment countries. That is a similar, yet different differentiation than in Hooghe & Marks (2004). It was also already mentioned, how the performance of national governments can alter support towards the supranational institutions (Sánchez-Cuenca, 2000). This was essentially important during the Eurocrisis (Kuhn & Stoeckel, 2014). Furthermore, as it was mentioned through the cueing benchmark approach that the national media and politics can work as a proxy of information which citizen can use as an aggregated source of opinion. Based on that, it is not surprising that a number of scholars found dynamics that explain these hypotheses. National parties can use European integration as an important polity field in debates during national elections (Ladrech, 2009). However, if there is no significant conflict framing by the parties, then the variance is much higher then otherwise (de Vries & Steenbergen, 2013).

This introduces the other highly analysed field that is national identity. Hooghe & Marks (2009) argue that identity is one of the most important postfunctionalist term, that since can be discussed in public forums and mass organizations, creates a collective group loyalty. As it was already highlighted it can be viewed two eyes. It either means the exclusive nationalism or the authority pooling of the supranational institutions that undermine nationalism (Jacquier, 2012). The basic argument of the national identity as the first interpretation is that if the citizen identifies more strongly or exclusively with the national level, it will lower the support towards the EU integration (Carey, 2002). While there is little consensus on how it is possible to measure it, there are some ways of conceptualizing it into real life. McLaren (2002) for instance used the perceived cultural threat of supranational immigration policies, to explain the importance of national identity. While in the view of the utilitarian explanations, immigration is a realistic threat, in connection with the collapse of social benefits, people with strong national identity fear from symbolic threats such as religious practices. This approach of national identity comes from instrument Euroscepticism that concentrates on the membership of the country (Hakhverdian et al., 2013). The other approach (pooling authority) can be analysed by asking citizens how they view the loss of sovereignty of the state (Hobolt, 2012; Jacquier, 2012; Hobolt & de Vries, 2016). Possible measurements to find out identity-based support is the principle of membership and the desired direction and speed of integration (Hooghe & Marks, 2004; Hakhverdian et al., 2013). According to empirical findings, in countries where states do not lose significant powers citizens are more supportive towards integration (Sánchez-Cuenca, 2000). This explanation can be best used when you specifically measure different policy support (Eichenberg & Dalton, 2007). To summarize, national identity is an important explanation that has to be integrated into research, even though it works more

(14)

as an additional explanation. While some studies believe that it is much more of a factor than economic values (see Carey, 2002), it is usually the opposite. Lastly, it is important to mention the concept of dual identity, which means that people who are having a strong but inclusive national identity, could still support integration (Klandermans et al., 2004).

Two more important variables still have to be discussed that are largely connected to cognitive mobilization. It was assumed that a large level of education will involve more political knowledge, thus more interest in regional integration (Inglehart, 1970). While education in fact does seem to be a really strong explanatory predictor, lately it has been showing that it is not a significant. While it is through that education affects how citizens view Euroscepticism (Hakhverdian et al., 2013, Dijkstra et al., 2019), in the case of public support this is not that universal, and it is even thought that it has no major effect at all (Brinegar & Jolly, 2005). To be able to use it is usually connected with socio-economic variables, where it has more significant success (Gabel & Whitten, 1997; Kuhn & Stoeckel, 2014). Regardless of these criticisms, education is one of the most essential variables to include in an empirical public support research, thus it was necessary to mention. It is also important to highlight political knowledge, which in some ways can be a more nuanced way of view of how politically educated citizens behave. Through multiple studies, de Vries measured political knowledge and assumed that it is a highly ambivalent variable. It seems to differ based on the length of being a member and even then, especially in the post-2004 countries it is insignificant (de Vries, 2013; de Vries & Steenbergen, 2013).

Can the type of community be an explanatory factor?

The thesis is based on the general concept of the political cleavage theory raised by Lipset and Rokkan (1967). In there they believe that voters can be categorized into certain distinct groups that are sometimes opposite of each other. These groups represent their interests. Rural/urban cleavage is one of the four original ones. Thus it is argued that being rural has always been a unique characteristic. The locality factor can actually be dated back to attachment questions. The creation of a European identity required the uncovering of attachment styles of citizens. While in Marks (1999) it is argued that national identity kept being the leading identity, regional and territorial attachments also appeared. Due to the geographical polarization between locality types, this sense of belonging differs (Maxwell, 2019). In the case of rural communities for instance, local identity is assumed to be strong and directly connected to working possibilities as well (Kaneff, 2002).

The last question that still needs to be conceptualized is why being a rural resident can be a factor? What being rural really means and what if there is no difference between rural and other citizens? First, the thesis does not particularly expect that being rural, meaning living in an area with less population density is an individual predictor that could explain public support on its own. Age or gender is not enough of an explanation either even though there are differences between elder citizen’s support and so on. What is important to find out that with the addition of socio-economic variables does it still

(15)

show an explanatory power? Naturally, being rural or urban in itself could be a socio-economic variable given the fact that in several cases regional integration affects regions and people differently. For instance, farmers and people from small villages and towns are much likely to be directly affected by the Common Agricultural Policy (CAP) than urban residents. Also, the political value argument is based on socio-tropic indicators that are coming from people’s pre-adult years: that can also mean that people from different types of communities are looking differently on European integration (Gabel, 1998). After all, regional actors are likely to benefit from the multi-levelness of this system (Klandermans et al., 2004). Regardless, the reason why identity predictors were highlighted is because if there is a general difference between the urban public support and the rural population’s support, there also might be an identity or cultural explanation that overwrites even socio-economic differences.

Being rural has not really been used as a variable in public support studies. In some cases like in Gabel & Whitten (1997) it is highlighted as a characteristic that could be a predictor. It actually performs well, but it is not regarded with significance (pp. 89). What is the closest to considering the rural population is the analysis on farmers and their opinion towards European integration. In Gabel & Palmer (1995) farmers are one of the professions that are analysed and compared to other socio-economic classes, they are moderately supportive towards regional integration. Data from the 1990 show a general negative support from the farmers that can directly be explained through the then-current CAP reforms (Anderson & Reichert, 1995. pp. 243). Similar opinions came out in de Vries and Steenbergen (2013), where they are less supportive than the average. As it can be seen, farmers are usually considered as factor for utilitarian evaluations, however when we look at identity-based studies, it can even be said that due to their localness, they are keener to accept cumulative identities: that means they are identifying with the nationality and also with being European (Hooghe & Marks, 2004). Klandermans and his colleagues (2004) argue that farmers’ attitude is quite similar in every countries but in general they are lacking the collective political activity. Of course the thesis is well aware that not everyone is a farmer in the rural level. But it shows how rural people were dealt with in public opinion studies before.

One of the possible reasons why public opinion research usually does not deal with special residence attributes can come from the fact that being rural means highly different things in every country. It is based on national tradition but also regional development. The differences certainly even increase if we add economical divisions to the regions (Fratesi & Rodríguez-Pose, 2016). In Dijkstra et al. (2019), it is argued that geographical drivers are great explanatories in connection with Eurosceptic party support. Rural people could behave differently based on labour economy or even distance from the capital. While this is certainly a factor that makes work harder, with acknowledging, it gives a possibility for an even more nuanced research. Cloke (1977) created an index where the rural areas distinguished through demographic density. This is a proper way of categorizing, but the density cannot be generalized on multiple countries. It requires other sources such as land use, agricultural activities and employed population (Braga et al., 2018). Another good index to see is Index Mundi’s statistics

(16)

which show the rural share of the population in Europe. The data has been collected through the United Nations Population Divison’s World Urbanization Prospects and estimated by the World Bank (Index Mundi, 2018). Here it is visible that the rural share of the population ranges highly from only 2% in Belgium to Slovakia where it is 46%. Unfortunately, as later it will be assessed, the Eurobarometer suveys lack the ability to highlight such nuanced differences.

As it was mentioned extensively in the Introduction, the migration crisis, and the effect on the rural perception on the EU has more expected theoretical relevancies. While migrants can create highly new possibilities for the declining rural population in multiple countries (Bunic, 2016), due to the crisis being largely politicized by the national level, it might come off as an even bigger cultural symbolic threat here (McLaren, 2002). In some countries such as Hungary, it is shown that politicization led the rural population to believe that migration leads to great threats that directly affect the individuals. There has been more than one occasion where villages straight-up locked up their lands from the mere possibility of settlement of foreigners (Spike, 2017).

Hypotheses

In the literature review, I have highlighted not only the main characteristics of public opinion studies but also how this thesis imagines the locality “factor”. Now, it is necessary to create implications or possible propositions that seem to be true given the logic that was shown. Since the research works on two-level, several hypotheses need to be portrayed. In the Introduction, I have highlighted that the thesis strives to show two things. Therefore, multiple hypotheses can be highlighted.

Hypotheses on the effect of the migration crisis on public support

Regarding the first hypothesis, we need to look at how the rural population’s public support differs from the non-rural and how the migration crisis affected this. If only one time period is taken, the numbers would be quite self-explanatory. However, to prove it in the long run, it is necessary to look at multiple years. This way we can map the general change in public support, but also the general line of support among the rural population. The best way to do this is to place an important event in the time period. That is the reason why the research question notably mentions the migration crisis as a factor. It also needs to be mentioned that the migration crisis had direct and indirect effects on the population. In the short-term, it was an agenda that broke through the traditional boundaries of international issue-framing. It went into every household and due to the previously mentioned way of behaving during it, it felt like a national problem. It is necessary to come with two different types of interpretations when it comes to the difference between the total and the rural support. Thus, the first hypothesis will assume that in the short-term, both the rural population’s and the total population’s public support decreased during the crisis event. On the other hand, it is also argued that the difference between the two supports grew overtime. The implication is that after the migration crisis the rural population perceived the public support highly differently than the total. This can be due to fear of losing the territorial integrity of their

(17)

or region, or it is straight-up cue-taking from the media. They might look at it in a much more accepting way as well (Centraal Bureau voor de Statistiek, 2018; Scmitt & Lebas-Joly, 2019). To portray these, general cross-tab analysis is enough and it can depict both the country and the EU-level.

H1: Both the rural population’s and the urban population’s public support shifted negatively during the migration crisis.

H2: Throughout the migration crisis, the direction of the rural population’s support and the urban population’s support diverged in the majority of the Member States

Another additional hypothesis that is worth exploring is whether or not, there can be any regional similarities in between countries with similar levels of trust or between neighbouring Member States. Due to the cue-taking behaviour and the importance of domestic politics, the thesis argues that there are only minor regional similarities between the shifts and changes in public opinion towards European Integration.

H3: There are temporary, regional similarities between the same population’s opinions across countries.

The type of community as a predictor

The second research question somewhat drifts away from the migration crisis and tries to concentrate on laying groundwork to another variable for future analyses. Although a crisis is a proper period to highlight associations between traits since the data is more heterogeneous and analysing the relationships is easier. The first hypothesis refers to the general assumption that there is some type of association between the type of community and public support towards European Integration. This hypothesis will be analysed on the country level through a Pearson Chi-Square analysis, while on the EU level with the regression analysis.

H1: There is a statistically significant relationship between the type of community and public support

The other hypothesis that is highlighted here is somewhat an extension of the previous. Even if there is an association between the two indicators, it is still worth including classical predictors in the analysis. With their controlling power, it is possible to see whether the association is strong enough and how relevant it is. Thus, with the last hypothesis the thesis wants to find out whether the differences are actually caused by the rural traits or whether by other more classical explanations. Both utilitarian and identity-based variables will be included as control variables. The other geographic indicators will be added as well. The job is two-stepped. First, it is necessary to see whether or not there is even any significance in the type of community and public support. We can also view it as a secondary explanatory variable that clings onto the primer control variables. The next step is to see how strong it is compared to the others.

(18)

H2: The type of community is a strong causal factor on public support, even with the inclusion of utilitarian and identity-based control variables.

In the following, the thesis will discuss the data source and collection, the methodology that will be used and most importantly the control variables that will be taken into account. Then, the results will be highlighted, which subsequently leads to the discussion of whether the hypotheses were valid or not.

Operationalization

Data collection

To highlight different tendencies and factors over a longer period of time, a dataset, that is measuring public opinion and general utilitarian and identity-based attributes is needed. For the current research, the data are collected from various Eurobarometer (EB) surveys. Standard Eurobarometers were established in the 1970s and since then they are conducted at least once a year, but usually twice. It is collected using the Computer Assisted Personal Interviewing (CAPI) technique that is considered as a progressive interviewing method that has a low cost and high level of data quality (de Leeuw, 2008). The goal of these surveys are to make an in-depth analysis source where trends and tendencies can be witnessed (European Commission, n.d.). A great reason to use these datasets come from this exact fact: they are highly consistent, and they can be used easily for longitudinal research (Hakhverdian et al., 2013). Generally speaking, standard EB surveys represent themes related to the legitimacy of the EU, the main problems and issues the Member States and citizens face and the general attitude towards the international system. It is usable to highlight whether a financial crisis opened the gap between countries, or how the image of the EU has been declining over the decades (European Commission, 2013). The surveys are not finalized at the same time, the datasets are usually collected over a long period, making several sub-reports that are based on one particular asset (GESIS, 2020). Main, reoccurring topics are the “European citizenship”, the “EU budget”, “media use”, “views on priorities” and so on. In this research “public opinion in the European Union” topic-based data will be used, which usually represents standard questions, thus it is a great source to use in situations such as this (see e.g EB 91.5, 2019).

Before continuing the explanation of the data collection, it is impossible to avoid the discussion that, Eurobarometer surveys have their own deficiencies as well. In Nissen (2014) it was argued that while EB surveys are great sources of public opinion, it is not a panel survey meaning that its’ sample selection is different each year and in many countries the style of selection and data collection differs. Furthermore, the EB is not created to highlight the changes in the countries’ demographic characteristics: meaning, minorities are not highlighted and even with weighting, the distribution of the rural and urban population is incorrect (Nissen, 2014). Some scholars argue, that EB surveys are manipulative, the questionnaires are selective, and it “constructs” an opinion that might have never existed before (Höpner & Jurczyk, 2015. pp. 4-5). In Schout & Holderied (2018) it is argued that EB surveys are politically inadequate and they propose the idea of making Eurobarometer independent from

(19)

the European Commission. While the thesis acknowledged these matters, in the field of European integration, Eurobarometer is the standard type of database to use. There is no better and more consistent way of getting these cross-national tendencies (Hobolt & de Vries, 2016. pp. 416-417). Regardless, these deficiencies need to be highlighted in order to become as accountable as possible.

Researches primarily use EB surveys to create basic cross-tabulations, descriptive analysis and to show relationships between variables. This thesis will do all of the above mentioned. As the hypotheses and the literature review have mentioned, the thesis would like to see whether being rural can be its’ own factor in public opinion on European Integration. To achieve that, we look at tendencies over a long period of time and use the migration crisis as an event. That even helps to witness how it differs from previous crises: while in the financial crises integration and common acting has improved and the public opinion strengthened, here these were not endured (Debomy, 2016. pp. 34). It is fairly important to state what we consider as the date of the crisis event since some scholars argue that already in 2011, a migration crisis had started when Tunisian refugees had entered the European coast. Still, the general consensus is that the Schengen crisis started in the summer of 2015, when the number of refugees entering the EU has increased, and later that year some Member States reintroduced inner border control (Colombeau, 2019). Thus, to be able to use it as an event when changes are much more perceivable, the thesis uses surveys before and after 2015. Considering the data collection period for each Eurobarometer helps to point out the surveys that have been taken place in the heat of the crisis. Colombeau (2019) argues that the most critical period of the crisis ended in 2016. Thus, when the thesis reveals the perceived positions of respondents during the migration crisis, it will use the second survey in 2015 (E.B 84.3) and the first one in 2016 (EB 85.2).

The data is collected from 2012 to 2018. In the appendix, Table 1 shows the codes and the recording date of the surveys. Aside from the last, in each year two surveys were conducted, thus the quantitative analysis used 13 survey databases (from Eurobarometer 77-89). The research will be Large-N based analysis in which we include all Member States. The number of respondents differed in each dataset, but it ranged between 31 and 33 thousand. The total rural population ranged between 30 and 35%. The data included all the 27 (and the UK) Member States, and other prospective EU candidates such as Serbia or Turkey and partners like Iceland. In the final analysis these were taken out, and the UK was only included in aggregated data. Since, the number of respondents are more or less around 1000 in each country, the numbers are not representing well the population of the EU. Thus EB surveys offer different weighting methods, mostly post-stratification weights, and population sized weightings. The former’s goal is to reduce sampling errors and non-response bias with the help of control data and design weights (European Social Survey, 2014). On the other hand, the latter is weighting the proportion of the countries based on their population size (GESIS, n.d.). In this research the latter was used, most notably the code W23 (or W24, based on the given dataset) which represents the 27 European Union Member States and Croatia and weights the dataset according to their population-based on EUROSTAT

(20)

measures (EB 87.3, 2017). The EUROSTAT does make research on the percentages of the rural population, however, the datasets are not meeting with those numbers (EUROSTAT, 2018). This has already been mentioned as a deficiency of the EB surveys. In the analysis, the size of the rural population in different states is somewhat irrelevant, in the sense that the portrayed percentages show how different responses appear in one certain type of locality and later that will be compared to each other.

Before presenting the used variables, it is necessary to state some essential information about the datasets. Aside from the aforementioned UK data, the EB tends to still measure West-German and East-German data separately. In this thesis, Germany will not be “unified”, based on a simple implication: are the citizens from East-German perceive the EU drastically differently than from the West? Furthermore, it needs to be highlighted that missing and non-answered units are not included in the data. In the latter four datasets (86.2, 87.3, 88.3, 89.1) the tables had to be altered since in these surveys, the option of “don’t know” was also included as a full unit, thus they excluded by me. The research creates a continuity of socio-political topics using the reoccurring indicators. In some cases, these objective is eased thanks to trend files that are aggregated or cumulative datasets based on trends. Such a trend file could be the Mannheim EB Trend file of 1970-2002 (Schmitt et al., 2008). Unfortunately, in the case of this research, this help does not exist yet, thus data had to be individually analysed and cumulated. The data was analysed through IBM SPSS Statistics 24 and the graphs were created through Microsoft Excel.

Variables

In the following, the variables will be presented, which are more or less the same throughout the whole research. The first two hypotheses are quite self-given. It needs one dependent and one independent variable. This however later changes, when other factors and attributes are added to the mix. There, several confounding variables will be highlighted, that are set to uncover other causal effects that might affect our dependent variable (Toshkov, 2015. pp. 245). This is especially important in such a Large-N research, due to the large level of randomness in the analysis. The casual mechanisms were detailed extensively in the literature review. What is essential is to state that the thesis assumes and implies that the independent variable might actually end up being the confounding one. In the latter hypotheses in which the causal interaction gets more complex, the casual strategy will be based on conditioning. That is, the calculation of the association between the dependent variable and the independent one, after the conditions of the confounders. Thus, at the end, a mean casual effect will be gained (Toshkov, 2015. pp. 224).

Dependent variables

In previous studies that were analysing the public opinion towards European Integration, the most used indicators were questions where the EU membership’s perception was questioned (Hobolt & deVries, 2016). Even in studies where that was not the main concern, such as in Luedtke (2005) it was

(21)

mentioned as an alternative control variable. As Gabel (1998) states, other indicators can be used, and the reason why this one was so prominent is that this was a question they asked each time. This has changed nowadays, this question is mainly asked to countries that are prospective Member States (89.1). Thus, the thesis switched to two alternatives. The coding differs each year, thus they will not be mentioned.

EU IMAGE POSITIVE/NEGATIVE - In general, does the EU conjure up for you a very positive, fairly positive, neutral, fairly negative or very negative image?

Very positive (1) Fairly positive (2) Neutral (3) Fairly negative (4) Very negative (5) DK (6)

TRUST IN INSTITUTIONS: EUROPEAN UNION - I would like to ask you a question about how much trust you have in certain media and institutions. For each of the following media and institutions, please tell me if you tend to trust it or tend not to trust it. (…)

European Union Tend to trust (1) Tend not to trust (2) DK (3)

The former dependent variable is more of a secondary indicator that is only used in the first few hypotheses. As it was mentioned before, the “DK” answers will be taken out from the equation. The explanation behind this indicator is that the perception of the European Union can be strictly connected whether citizens perceive positively or negatively the current actions of the EU. It is necessary to state of course that these considerations often revolve around certain policies or statements not related to European Integration. Regardless, due to the crisis, the image of the EU is fairly important.

The latter variable is the main dependent variable, and this is the one that replaces the “EU membership” indicator. It is only a part of a question, often along with other institutions like political parties, media or national parliaments. In Hobolt & de Vries (2016. pp. 416), apart from the perception of the membership, the desired speed of integration and trust in the institutions were mentioned as dynamics of public support. Given this analogy, this thesis assumes that the “trust” in the European Union is a good measuring system that can explain integration. Some critics could say that this variable is two-edged since the interviewed could perceive it as a reference to one given institution such as the European Commission or the European Bank. I argue, that the way the question is asked implies the

(22)

topic well enough. It also stands against the more classical criticism towards the EB, that is the conceptualization of previously non-existent arguments (Höpner & Jurczyk, 2015). Measuring perception of EU integration based on membership is adequate, however, a large number of people don’t believe in the current European integration albeit they want to stay in the EU. This is a sort of soft Euroscepticism that is harder to be grasped with the traditional indicator (Taggart & Szczerbiak, 2002).

Independent variable

In the thesis, it is assumed that the geographical attributes of the living environment of a citizen affects its’ perception of the highlighted dependent variables. Thus, we use the “Type of Community” (d25 codenamed) numeric and nominal variable. The portraying of the question is taken from the EB 89.1 (2018) survey questionnaire.

TYPE OF COMMUNITY: d25 - Would you say you live in a…?

Rural area or village (1) Small or middle-sized town (2) Large town (3)

DK (4 or 8 depend on the survey)

This is a great variable to easily highlight the rural population in the dataset. However, it lacks the adequacy that was described as necessary part in the literature review. As it is visible in Dijkstra et al. (2019), it is essential to correctly state what really the countryside entails. How improved the region is, how close the certain town is from a larger city or to an agglomeration, and so on. An obvious criticism of the EB surveys is that there is no additional information about these characteristics. Thus, even though it should be more than necessary, it is impossible to include in the research. The current coding also limits the operationalization of the rural population. While it can be argued that small towns could also very much be included in this category, the current coding blocks this possibility. Middle-sized towns are the concern here, especially because this is where different Middle-sized countries’ perception shifts. For instance, according to the European Spatial Planning Observation Network’s (ESPON) survey data, in Germany towns are considered small up until 50 thousand inhabitants and middle till 100 thousand. This number is exactly halved in France, in the case of middle-sized towns, it is doubled in Poland and Spain, and there are countries such as Hungary where there is not even an upper limit (ESPON, 2006. pp. 34) Thus, the safest decision was to only include villages in the consideration. Unfortunately without any proper distinctions laid out to the interviewers and the citizens, such clear cuts cannot be expected.

Confounding variables

The confounding variables are the main and necessary utilitarian and identity-based indicators that have been used thoroughly in the past decades in the field. Each of these will be used in a linear

(23)

regression analysis, to highlight whether the expected relationship between the type of community and the trust in the European Union is real or actually enforced by them. Gabel and Palmer (1996) and Kanthak and Spies (2018) are the inspiration in deciding which topics should be chosen. In the following, I will briefly highlight each variable, and which question could be connected to it.

Age: It is one of the most basic indicator that has to be included into each and every research. It is a quite established theory that age creates political a social conflict between the elder and the youth (Foner, 1974). This is especially important in the case of the European Union, where generational differences seems to be a generally significant factor on how you perceive the integration (Down & Wilson, 2013). To determine this indicator, we use variable D11 (“How old are you?”).

Gender: The gender of the respondents are just as essential of an indicator as age is. It is something that needs to be addressed, thus the question coded as D10 (“Gender”) will provide the data for this variable.

Employment: Occupation is the main and most important utilitarian indicator that needs to be included in the analysis. Given the fact that education seems not to be a thing that is questioned directly, it is also, in a way a representative of the cognitive mobilization theory as well (Inglehart, 1970). It introduces the argument that, mainly revolves around whether skill and the international embededness of the given job is a factor or not in public opinion. This topic is represented with the question codenamed to D15a (“What is your current occupation?”). This question includes a vast variety of answers, from which some of them (farmer, skilled manual worker, employed professional) could be used as variables. For the regression analysis, the “occupant of respondents” variable will be recorded into two different variables: low skilled workers and high skilled workers. The former includes farmers, fishermen, skilled and unskilled manual workers, low-level entrepreneurs, and so on. The latter involves employed position with high skill requirements, middle and higher-level executive positions, professionals and employment that mainly revolves around office jobs. These distinctions are created for the thesis, and it can be challenged in case of disagreement. There could be valuable debates about whether office jobs is considered as a quote on quote “high skilled work”. Same can be said about skilled manual labour. The way the thesis distinguishes the two is based on the expected market value and required educational level (Inglehart, 1970; Jacquier, 2012). Lastly, unemployment responses are separated into a dummy variable.

National identity and Attachment to the European Union: Lastly, national identity will be used as a control variable. This is related to the identity discussion argued in Carey (2002). National identity is hard to really measure, and Eurobarometer surveys tend to avoid asking direct questions about people’s political ideologies and nationalistic tendencies. However, in certain questions, it can be deducted how close someone is to their country. The thesis assumes that the question codenamed QD1a_2 (“Please tell me how attached you feel to…YOUR COUNTRY”) is adequate enough to highlight these theories. The answers can be ranged from very, fairly attached and not very or not at all attached. In the