Assimilation and Perceptions of War: A Micro-Level Analysis of Ex-Yugoslavs in Sweden

(1)

Essay III

(2)

(3)

Assimilation and Perceptions of War: A Micro-Level Analysis of Ex-Yugoslavs in Sweden

Introduction

Scholars widely recognize that efforts to secure a lasting peace are often undermined by a lack of popular legitimacy and sensitivity to local opinion (Lundy and McGovern 2008; Mac Ginty 2008; Reno 2008; Richmond 2008, 109). Survey research has significantly contributed to our understanding of the factors that shape local perceptions after war (Brounéus 2008; Canetti- Nisim et al. 2009; Gibson 2004; Halperin and Bar-Tal 2011; Kostić 2007;

Pham et al. 2007). However, the mass displacement violent conflicts gener- ate reminds us that substantial portions of national communities affected by conflict do not live within the boundaries of the state. Nevertheless, system- atic data collected in places such as Bosnia or Rwanda appear to be largely truncated at the border. As a result of this ‘methodological nationalism’

(Wimmer and Glick Schiller 2002, 2003), the attitudes of migrants from war-torn countries remain largely uninvestigated.

This paper contributes to addressing this problem by examining a key assumption in migration and conflict research: that more assimilation in settlement countries leads migrants to hold more peaceful attitudes about war in origin countries, while less assimilation leads migrants to hold more conflictive attitudes about war in origin countries.¹ According to previous literature, cross-border refugees may turn their backs on homeland conflicts by inter- marrying with locals and adopting their culture, while those remaining in camps maintain ethnic purity and participate in homeland struggles (Lischer 2005; Malkki 1995; Zolberg, Suhrke, and Aguayo 1989). When tensions between refugees and locals rise, the pressure to return may reverse assimilation and increase refugee militarization (Adelman 1998). In liberal settle-

1 While forced and voluntary migration are traditionally distinguished by the ‘push’ and ‘pull’

factors that influence migration decisions (Kunz 1973), scholars today recognize migration motives are mixed within both groups as well as individuals. On the crises and subsequent

‘migration orders’ that structure forced and voluntary migration flows, see Van Hear (1998).

In much of the literature, terms such as ‘refugees’, ‘immigrants’, and ‘diasporas’ tend to be treated synonymously to refer to international migrants and their descendants who maintain emotional and social ties with a homeland (Brubaker 2005). I will use these terms in this broad and multigenerational sense.

(4)

ment countries, some migrants may adopt progressive values, enabling them to contribute to peace-building and positive social transformation in origin countries (Levitt 1998; Shain 1999; Shain and Barth 2003). However, others respond to the loss of cultural heritage and status by maintaining ethnic distance from the majority and pursuing ‘long-distance’ or ‘vicarious’ nationalism (Anderson 1998; Smith 1986). Migrants from war-torn countries may organize their identities around traumatic myths and war memories that in- voke sectarian, categorical and uncompromising views (Lyons 2006). More- over, these processes may be informed by nostalgia rather than recent experience, resulting in perceptions of conflict that are frozen at the time of war and displacement (Hockenos 2003). The tendency is thus to view unassimi- lated migrants as more politically extreme.²

These claims are based largely on descriptive and theoretical accounts.

Moreover, case studies tend to be selected on the dependent variable and thus suffer from selection bias (e.g Adamson 2013; Angoustures and Pascal 1996; Fair 2005; Gunaratna 2003; Hockenos 2003). At the same time, within-case data are usually collected through observations of, and interviews with, members of diaspora organizations who undoubtedly selected themselves into such roles. Generalizations are typically based on the results of large-N statistical studies linking migrant communities to war; the most widely cited being Collier and Hoeffler (2004), which found an association between the risk of renewed homeland conflict and the size of the migrant population living in the USA. Yet, this finding was discredited by subsequent research (Collier, Hoeffler, and Soderbom 2008). Others partly attrib- ute geographical conflict clustering, interstate disputes, and the duration of conflict to cross-border refugee flows (Salehyan 2007, 2008; Salehyan and Gleditsch 2006). However, we also know that most refugee crises do not result in refugee militarization (Adelman 1998; Salehyan and Gleditsch 2006, 361; Stedman and Tanner 2003). In sum, neither qualitative nor quan- titative analyses provide a careful treatment of the concept of assimilation and its empirical effects on migrants’ attitudes towards conflict. This would require fine-grained data on the individual-level; an empirical lacuna that has impeded the progress of migration and conflict research more generally (Salehyan 2007, 136).

This paper offers both a theoretical and an empirical contribution to migration and conflict research. My theoretical contribution involves a more nuanced understanding of immigrant incorporation grounded in the assimilation and acculturation literature and connects this with social-psychological research on intergroup conflict and bias. There are two aspects to my argu-

2 Moreover, the xenophobic reaction to migrant extremism in settlement countries potentially

‘feeds the immigrant community divide between those who feel themselves assimilated and those who retreat into a long-distance identity’ (Tarrow 2005, 55).

(5)

ment. The first relates to socioeconomic assimilation and the second to acculturation.

First, I argue that by increasing access to coping resources socioeconomic assimilation may reduce the need to defend group identity in response to the conflict crisis. Conflictive beliefs, which for example glorify in-groups and delegitimize out-groups, provide one set of psychological resources for coping with conflict. However, psychological research has shown that individuals with more personal and social resources are less likely to develop counterproductive defensive coping strategies such as hostility towards out- groups (Canetti-Nisim et al. 2009; Hobfoll, Canetti-Nisim, and Johnson 2006).

Second, I argue that cultivating a bicultural identity in settlement coun- tries may reduce the need to rely on conflictive beliefs to cope with conflict crisis because it fosters greater identity complexity. According to intergroup bias research, social identity complexity reduces the significance of intergroup comparisons and the importance of any particular in-group for satisfying psychological needs, undermining the motivational basis for intolerance towards out-groups generally (Brewer and Pierce 2005). Roccas and Brewer (2002) illustrate the concept of identity complexity using the example of immigrant biculturalism – simultaneous identification with both one’s heritage culture and the dominant culture in the settlement country. According to them, persistent bicultural identification represents the most complex acculturation strategy available to migrants. However, as yet no one has examined the impact of acculturation on migrants’ attitudes toward homeland conflict.

Testing my arguments requires entirely new data on the individual-level. I examine the results of a new large-scale survey of ex-Yugoslavs in Sweden conducted in 2010. Respondents were located by searching online public listings. Because the listings are geo-referenced, the survey captures the geographical distribution of Bosniaks, Croats, and Serbs and represents the whole of Sweden. Using regression analysis, I relate respondents’ attitudes towards the Bosnian War (1992-95) to their level of socioeconomic assimilation and acculturation strategy in Sweden, controlling for individual background characteristics such as migration cohort and urban background.

The findings suggest that both greater socioeconomic assimilation and biculturalism among ex-Yugoslavs in Sweden are associated with more peaceful attitudes related to war in former Yugoslavia. Each therefore may contribute independently to reducing psychological dependence on conflictive beliefs to cope with the conflict crisis. On the other hand, cultural assimilation appears less relevant since so few (5 respondents) adopt this acculturation strategy.

In addition, the findings also uplift another source of identity complexity:

cultural heritage itself, the maintenance of which is often equated with ethno- nationalism. Many respondents nurture a separate but inclusive Yugoslav (rather than ethnic) heritage. Moreover, those that combine inclusive herit-

(6)

age with new attachments in the settlement country hold the most complex identity and display the most peaceful attitudes. Together these findings suggest that incorporation in settlement countries may encourage peaceful attitudes not by encouraging migrants to adopt the dominant culture’s values but by supporting identity complexity and socioeconomic security. By im- plication, the recent pendulum swing away from multiculturalism in Europe and elsewhere may inhibit the formation of complex identities and thus more peaceful attitudes among refugees and diaspora communities with links to war-torn origin countries.

The paper proceeds as follows: in the next section I outline my arguments relating migrants’ conflictive beliefs to their socioeconomic assimilation and acculturation strategy in settlement countries. Next, I discuss the context of migration from former Yugoslavia to Sweden, including the war in Bosnia and Swedish migration and integration policy. Thereafter, the data and methods are presented and discussed. This is followed by the empirical analysis and subsequently the final discussion.

How Immigrant Incorporation Shapes Beliefs about Homeland War

This section explains how incorporation in settlement countries influences migrants’ perceptions of war in origin countries. First, it describes how under stressful and threatening wartime and postwar conditions social categorization processes give rise to conflictive beliefs that in turn help meet basic psychological needs. Next, it describes how socioeconomic assimilation in settlement countries reduces migrants’ needs for conflictive beliefs in order to deal with the conflict crisis by increasing access to coping resources. Last- ly, it describes how cultivating a bicultural identity in settlement countries is more likely to reduce the significance of intergroup comparisons related to homeland conflict than pursuing cultural assimilation by fostering greater identity complexity.

In many cases, wars represent large-scale crises in which ideological shifts, traumatic events, and political machinations culminate in massive violence between social groups, generating widespread fear and uncertainty.

In the process, public support for divisive rather than compatible images of group identity spread through cascades of social reactions, culminating in the

‘dissimilation’ or ‘polarization’ of society (Kuran 1998, 626; Somer 2001).

Under these conditions, adaptation and survival require forming new shared beliefs that enable individuals to cope psychologically and contribute suc- cessfully to their group’s costly efforts to win the war (Bar-Tal 1998). These conflictive beliefs justify the war’s aims and make security the central con- cern, glorify the in-group and cast it as the victim, delegitimize opponents,

(7)

promote patriotism and unity, and maintain the hope of peace, albeit in the abstract. The stable reproduction of these beliefs is supported by the development of a social-psychological infrastructure, involving for example public education, the media, official statements and commemorations (Bar-Tal 2007). Over time, these beliefs become part of a society’s ethos and shape its culture, language, and symbols, contributing to the conflict’s intractability.

Conflictive beliefs may be further reinforced by stressful post-war conditions if conflict resolution and peace-building processes do not address the conflicts root causes and improve living standards (Mac Ginty 2006).

Social categorization processes contribute to the formation of conflictive beliefs under these stressful and threatening wartime and postwar conditions.

The organization of the self and others in terms of group memberships is a universal process driven by the fundamental survival strategy of group living (Brewer 1999). In-groups define communities of mutual trust, obligation, interdependence and cooperation, but they also delimit out-groups. In-group belonging entails conforming to the in-group image as well as accentuating intergroup differences. Social categorization thus satisfies the opposing yet basic human needs for assimilation with and differentiation from others (Brewer 1991). In-group favoritism is pervasive and not limited to conflict situations (Tajfel and Turner 1986). However, under threatening conditions in-group biases are transformed into more salient and durable forms of out- group discrimination, derogation and intolerance (Hewstone, Rubin, and Willis 2002, 579).

The conflictive societal beliefs that emerge from this process are essen- tially a set of shared coping resources employed in part to meet basic psychological needs. These needs include: making sense of a chaotic situation, reinforcing a positive self-image, and reducing anxieties about death that inevitably arise in the context of war. On a fundamental level, periods of violent social conflict disrupt the plausibility structures that individuals rely upon to make sense of the world around them. According to Berger and Luckman (1966, 121):

All social reality is precarious. All societies are constructions in the face of chaos. The constant possibility of anomie terror is actualized whenever the legitimations that obscure the precariousness are threatened or collapse. The dread that accompanies the death of a king, especially if it occurs with sud- den violence, expresses this terror. Over and beyond emotions of sympathy or pragmatic political concerns, the death of a king under such circumstances brings the terror of chaos to conscious proximity.

First, by for example casting the in-group as the victim and delegitimizing out-groups, conflictive beliefs render the social world and one’s place within it intelligible again. Second, conflictive beliefs help reinforce self-esteem when it is threatened by conflict. Group memberships, and the value and emotional significance attached to them, form part of the self-concept (Tajfel

(8)

1981, 255). Threats to the in-group during conflict thus threaten self-esteem, while in-group favoritism and out-group derogation reinforce self-esteem.

Third, war also raises the possibility of one’s own death, frustrating on a psychological level the basic human need for self-preservation (Becker 1973). However, fear of death may be reduced, and psychological equanimi- ty restored, by investing in and maintaining psychological structures like conflictive beliefs that restore a sense of meaning, order and permanence to life (Greenberg, Pyszczynski, and Solomon 1986).

Socioeconomic Assimilation

I argue that socioeconomic assimilation in settlement countries may reduce migrants’ reliance on conflictive beliefs for coping with the conflict crisis by increasing access to personal and social resources that help meet psychological needs. The early assimilation literature largely portrayed immigrant incorporation as a single process. As a result, classical works such as Gordon (1964) focused largely on sociocultural processes, relegating to footnotes key issues such as economic success and mobility in the labor market; the assumption being that cultural assimilation would eventually give way to structural integration and economic mobility. However, today immigration research conceives of socioeconomic assimilation and acculturation as separate processes associated with independent outcomes (Berry 1997; Gans 2007; Portes and Zhou 1993). Socioeconomic assimilation is typically defined as socioeconomic parity and measured by factors such as educational attainment, employment, and income: outcomes that may be attained independently of acculturation (Portes and Zhou 1993).³

By increasing access to resources for coping, socioeconomic assimilation may reduce reliance on conflictive beliefs. Coping positively with traumatic events such as those that occur during war and displacement requires developing an alternative set of beliefs about what caused the crisis, perceptions of control over oneself and life more generally, as well as a positive self- image. Taylor (1983) refers to these aspects as meaning, mastery and self- enhancement. Socioeconomic status has been shown to encourage beliefs that enable positive coping. According to Kerr (1969), modern society may be divided in terms of productivity into the inner and outer society. The inner society, or “multidimensional modern core”, consists of skilled laborers, while the outer society consists of the unskilled, unemployed, students, and the elderly. Belonging to the inner circle provides individuals with greater opportunity to influence their own lives, enhances feelings of productivity

3 This is captured by the concept of ‘segmented assimilation’, which describes how immigrants may experience divergent economic outcomes regardless of their social and cultural assimilation due to incorporation into higher or lower socioeconomic classes (Portes and Zhou 1993).

(9)

and effectiveness, and places them in a meaningful life context, giving them purpose (Tornstam 2005, 93). According to Jahoda (1982, 6), employment provides a structure for each day, week and year, creates a broader social horizon beyond connections to family and friends, cajoles individuals to participate in collective efforts, and defines their social status. Conversely, the outer circle has a range of negative consequences for individual well- being. For example, unemployment is associated with less satisfaction with one’s own life (Warr 1987), lower self-esteem and loss of self-respect (Branthwaite and Garcia 1985), deteriorating psychological health (Bartley 1994), poorer physical health, and higher mortality rates (Warr 1987).

Empirical research on psychological stress and exclusionary attitudes supports this view, showing that individuals with access to greater resources for coping are less likely to develop counterproductive coping behaviors such as intolerance and hostility towards out-groups under threatening conditions (Canetti-Nisim et al. 2009; Hobfoll, Canetti-Nisim, and Johnson 2006).

According to Antonovsky (1979), economic resources help meet the demands posed by traumatic experiences, enabling individuals to perceive them not as debilitating stresses but rather as challenges worthy of their time and investment. By empowering migrants with access to a range of personal and social resources, socioeconomic assimilation may therefore reduce reliance on conflictive beliefs for coping with the conflict crisis, instead setting the stage for positive ‘posttraumatic growth’ (Tedeschi and Calhoun 2004).

Acculturation

According to previous research, cultural assimilation in liberal settlement countries may lead to moderation among migrants through their absorption of liberal cultural traits (Brinkerhoff 2008, 4-5; Fukuyama 2006; Huntington 2001; Leiken 2005; Shain 1999; Shain and Barth 2003). By contrast, I argue that biculturalism in settlement countries may reduce the significance of intergroup comparisons related to homeland conflict by fostering greater identity complexity. For our purposes, it is thus important to distinguish between different acculturation strategies migrants adopt vis-à-vis the dominant culture in the settlement country (Berry, 1997). Acculturation describes the social and cultural changes that accompany migration and intergroup contact as individuals who developed in one context adapt to new surround- ings (see Table 1). Biculturalism – also referred to as integration by Berry – denotes simultaneous identification with both the heritage and dominant culture, while assimilative acculturation signifies identification with the dominant culture without maintaining one’s cultural heritage. Striving to preserve one’s cultural heritage while maintaining distance from the dominant culture of the settlement country may be referred to as the strategy of separation, while alienation from both one’s cultural heritage and the domi- nant culture of the settlement country may be described as marginalization.

(10)

Table 1. Acculturation Strategies

Maintains cultural heritage?

Yes No

Identifies with dominant culture of settlement country?

Yes Biculturalism

(Integration) Assimilation

No Separation Marginalization Based on Berry (1997, 10).

As mentioned, under threatening conditions social categorization processes give rise to conflictive beliefs that in turn help meet individuals’ basic psychological needs. Once in place, social identities become difficult to change.

However, research suggests that as social categorizations become more complex intergroup hostility may decrease. Reviewing and synthesizing previous literature, Roccas and Brewer (2002) propose that identity com- plexity should be associated with greater out-group tolerance generally.

Complex group memberships reduce the evaluative significance of intergroup comparisons and diminish the importance of any particular group membership for satisfying self-definition and belonging needs, which un- dermines the motivational basis for intergroup discrimination. This view is in line with the general results of acculturation research, which suggest that individuals with a positive and secure sense of their own culture will hold more positive attitudes towards other groups in general (Phinney, Ferguson, and Tate 1997). Brewer and Pierce (2005) offer additional support for this claim. Using a telephone survey in the United States, they find that greater complexity of respondents’ national, religious, occupational, political and recreational social identities is systematically related to more positive attitudes towards ethnic out-groups and diversity.

Biculturalism represents a more complex form of social identity than cultural assimilation, separation and marginalization. According to the social identity complexity model, groups vary in the amount of overlap and em- beddedness they share. These configurations of overlap may be arranged from the least to the most complex. Roccas and Brewer (2002) draw upon the acculturation literature to illustrate this point (e.g. Berry 1997). Immi- grant and dominant groups in settlement countries represent distinct cultures and overlapping but non-convergent social categories. Each places competing demands on loyalties and resources and represents different, and at times competing, sets of norms and values. To resolve potential identity conflicts, immigrants adopt different acculturation strategies. Ranging from the least to

(11)

most complex,⁴ these include hyphenated identity, assimilation or separation, alternating biculturalism, and integrated biculturalism (Roccas and Brewer 2002, 92-93). The key point is that a persistent and enduring sense of bicultural identification among immigrants in settlement countries represents the greatest amount of social identity complexity, which as mentioned above may be associated with more tolerance towards out-groups generally.

While the literature on intergroup bias points to bicultural acculturation as a prime example of identity complexity, to date no one has attempted to relate acculturation strategies to attitudes towards intergroup conflict. Accord- ingly, I argue that migrants from war-affected countries that foster a bicultural identity in settlement countries should attach less importance to identity boundaries involved in homeland conflicts and thus express more peaceful attitudes in relation to homeland war.

War in Former Yugoslavia and Migration to Sweden

The breakup of the Socialist Federal Republic of Yugoslavia⁵ eventually coalesced in Bosnia and Herzegovina⁶ (1992-95), pitting Bosniaks, Croats, and Serbs against one another in Europe’s most devastating conflict since the Second World War. The violence involved civilian victimization on a massive scale, resulting in the rapid ethnic unmixing of Bosnian society. More than 100,000 individuals lost their lives as a result of the conflict, at least half of which were civilians (Tabeau and Bijak 2005). Tuathail and Dahlman (2004) describe the staggering scale of displacement caused by ethnic cleansing. Over 2 million of Bosnia’s pre-war population of 4.4 million fled their homes during the war, at least 1 million of which sought refuge outside the country. Half returned in the years that followed the signing of the Day- ton Accords in 1995. The other half settled abroad. The return of international refugees to Bosnia peeked in 1997 and subsequently declined. Today, approximately 38 percent of Bosnia’s citizens lives beyond its borders (Valenta and Ramet 2011, 1).

Those escaping the crisis in former Yugoslavia tended to follow pre- established patterns of Yugoslav gastarbeiter migration to Europe (Van Hear 1998, 29-30). As in Germany and Austria, Swedish industries attracted many labor migrants from Yugoslavia during the 1950s and 1960s. Sweden thus became a prime destination for asylum seekers during the Bosnian War.

4 Moving from least to most complex: an in-group may be defined by the overlap of multiple categories (‘intersection’, e.g. ‘I am a female lawyer’); one in-group membership may domi- nate all others (‘dominance’); group memberships may be isolated from one another and context specific (‘compartmentalization’); or non-convergent in-groups may be simultaneously recognized and embraced (‘merger’) (Roccas and Brewer 2002).

5 Henceforth referred to as Yugoslavia.

6 Henceforth referred to as Bosnia.

(12)

Most recipient countries offered only temporary protection to asylum appli- cants, expecting the crisis to be short-lived. However, Sweden took a different approach, taking a blanket decision in June of 1993 to grant 42,000 asylum seekers permanent residency (Frykman 2012, 2). Despite its small population size, Sweden accepted a total of 58,000 refugees between 1992-95, making it the third largest recipient of war refugees from Bosnia beyond the borders of former Yugoslavia. Policy differences among reception countries had a major impact on refugee settlement patterns. Today, 96 percent of refugees from Bosnia remain settled in Sweden. This may be contrasted with the situation in Germany, where only 6 percent remain (Valenta and Ramet 2011, 4). As a result of this unique situation, Sweden now hosts one of the world’s largest immigrant communities from former Yugoslavia; the second largest foreign-born population in Sweden after those born in Finland.

The refugees arrived during a period of economic crisis in Sweden and as a result faired poorly on the labor market. Just after the war, 90 percent of refugees from Bosnia received social assistance in Sweden. The situation has, however, greatly improved over the years. Currently, the number receiv- ing social assistance is less than 10 percent (Hedberg and Malmberg 2010, 45). Despite initial disadvantages, today refugees from Bosnia fair relatively well on the labor market in Sweden in comparison to other asylum groups (Frykman 2012, 4).

According to the National Human Development Report 2007, which doc- uments post-war socioeconomic conditions in Bosnia, incentives for refugees to return remain low (UNDP 2007). Prior to the war, Yugoslavia was a relatively egalitarian society that enjoyed high living standards. The war both destroyed the economy and reshaped the demography of the country.

Fearing discrimination and difficult economic conditions, many do not wish to return to their pre-war residences. Despite major international efforts to promote post-war reconstruction and reconciliation, the situation remains precarious for returnees. Post-war economic growth is both jobless and une- qually shared. Around a fifth of the population is below the poverty line and up to a third are poor in relative terms. Housing shortages and poor labor market conditions mean that returnees compete with locals for housing and jobs. Minority returns to majority areas therefore constitute a distinctly so- cially excluded group. Returnees suffer disproportionately from poverty and unemployment and face greater difficulty accessing public services and par- ticipating in political life.

In sum, prior Yugoslav gastarbeiter migration to Sweden in the 1950s and 1960s made it a main destination country for refugees from Bosnia in the 1990s. Moreover, the unique policy of blanket asylum for these refugees made their permanent settlement possible. Difficult conditions in Bosnia today contribute to making settlement in Sweden attractive. The majority of refugees thus remain in Sweden, creating a rich context in which to examine how incorporation affects migrants’ perceptions of homeland war.

(13)

Data and method

This paper examines whether differences in immigrant incorporation in settlement countries explain differences in attitudes towards conflict in origin countries. Using regression analysis, I relate ten different measures of attitudes related to war in former Yugoslavia to two separate measures of incor- poration of ex-Yugoslavs in Sweden: socioeconomic assimilation and accul- turation strategy. Socioeconomic assimilation is measured using an additive index, while acculturation strategy is measured by a categorical variable that represents the bicultural, assimilative and separation strategies.⁷ In all regressions I control for a rich set of individual background characteristics.⁸

Sampling and Data Collection

During the spring of 2010, together with Roland Kostić, Uppsala University, I conducted a large-scale survey in Sweden designed to equally represent the views of members of each major ethnic group involved in the Bosnian War.⁹ Previously a unit of the Socialist Federal Republic of Yugoslavia, Bosnia today is comprised of three constituent nations: Bosniaks, Croats and Serbs.

In Sweden, census data do not include ethnicity, making stratified sampling on the basis of census information impossible. In the sample from Sweden, three ethnic strata are defined according to Bosniak, Croat and Serb family names.¹⁰ Individuals were located by searching online public listings. Be- cause the listings are geo-referenced, the survey captures the geographical distribution of each group in Sweden. Most respondents live in the major cities of Malmö, Gothenburg, and Stockholm. Individuals to contact were selected proportionally by location (city/village). Conducting the survey by mail ensured the anonymity and informed consent of respondents. It was translated into the respondent’s native language and included 71 questions as well as a cover letter providing basic information about the project. Of the 2580 individuals contacted, 28 percent (714 individuals) responded.¹¹

7 I was unable to capture the strategy of marginalization. However, given that it is expected to be the least common strategy employed by migrants its absence should not present a problem for the analysis (Berry 1997, 10).

8 The software package used for the statistical analysis is Stata 11. The dataset, codebook, and do-files for the empirical analysis are available from the author.

9 In cooperation with IPSOS, Kostić conducted a survey in Bosnia in 2005 and again in the spring of 2010 – simultaneously with the Swedish survey – as part of a long-term project to track societal reconciliation over time (Kostić 2007). By employing the same questions gaug- ing conflict perceptions, the present paper extends these efforts beyond the borders of Bosnia.

10 A door-to-door survey in Sweden is also infeasible. There are ‘ethnic neighborhoods’ in Sweden. However, migrants from former Yugoslavia largely do not live in ethnic enclaves but instead intermixed with other members of society throughout the country.

11 Kostić and I conducted an initial pilot study to test out our survey guideline in the spring of 2008 (N=16). Then, in the fall of 2008 we conducted an initial round of data collection by mail (N=239). Finding that the procedure worked very smoothly, we then rolled out the rest of

(14)

It is important to consider whether the unit non-response rate will affect the representativeness of the results. Due to steadily declining survey response rates in many countries over the past decades, a large body of research has emerged investigating the importance of non-response (Berinsky 2008). Numerous studies have shown that although respondents and non- respondents often differ demographically, correcting for non-response does not appear to substantially affect estimates of political attitudes. Broad re- views find that even surveys with relatively low response rates are highly representative of public opinion (Holbrook, Krosnick, and Pfent 2007).

The reason unit non-response does not necessarily result in more response error is that the factors that drive non-response may be uncorrelated with the variables of interest in the survey. In a series of experiments, the Pew Re- search Center (2012) finds that respondents are more likely to engage in political and social action than non-respondents. Yet such engagement is not correlated with political preferences, partisanship, ideology or opinions on a variety of other social issues. It is therefore a reasonable assumption that the survey is representative of the attitudes of migrants from former Yugoslavia, although it may over-represent those more likely to take action with regard to those attitudes. However, the political and social engagement of migrants is not the subject of the current paper.

In sum, though the data are not flawless the analysis I provide offers a significant contribution to our understanding of what shapes migrants’ perceptions of homeland war. Given that previous research relies largely on indirect evidence rather than individual-level data, the data alone are an important contribution to existing knowledge.

Capturing the Nature of Societal Beliefs After War

Addressing the societal beliefs that emerge in the context of war is crucial to ensure lasting peace. Developing shared understandings of the past, recog- nizing the past experiences of other groups, and forgiveness lay the ground- work for a common future in which social relations are characterized by trust, positive attitudes, and the mutual consideration of needs (Ericson 2001; Galtung 2001). In essence, the conflictive ethos must give way to an ethos of peace (Bar-Tal 1998). This requires changes in beliefs about societal goals, the in-group, out-groups, intergroup relations, and the nature of peace (Bar-Tal 2000).

To capture prospects for peaceful coexistence, Hayner (2001) suggests examining how the past is integrated and spoken about between former enemies, if contradictory versions of the past have been reconciled, and if rela-

the survey during the spring of 2010 (N=475). I find no significant differences in attitudes between the two samples collected by mail and thus include data from both rounds in the analysis (N=714).

(15)

tionships are based on the present rather than the past. I use ten indicators to evaluate the extent to which respondents’ attitudes are embedded in a peace ethos. Answers to these questions represent respondents’ attitudes on: The extent to which the participation of their own group during the last war in Bosnia should be understood as purely defensive; whether ethnic cleansing is ever justified; whether their own group suffered more than others throughout history; whether their own group has a stronger culture than others in the region; whether their own group deserves to play the leading role in the region; whether one should forgive those who persecuted their compatriots;

the ideal arrangement of multicultural society in Bosnia, including the role of ethnicity, cultural mixing and intermarriage; the importance they place on their own ethnic belonging; ethnic distances; and belief in the possibility of peaceful coexistence among Bosniaks, Croats and Serbs in Bosnia unsuper- vised by external actors. Table 2 presents descriptive statistics for these variables.

(16)

Table 2. Conflict-Related Attitudes: Descriptive Statistics

Survey questions (10) Answer categories

1. What is your relation to your own national

belonging? I Accept but don’t think about it much. 0.54

National belonging is not important in my life. 0.13 National belonging is an important part of my life. 0.33 2. In Bosnia-Herzegovina live people of different

nationalities. According to you, which would be the ideal societal arrangement?

A society in which ethnic belonging is irrelevant. People

socialize freely, and mixed marriages are very common. 0.64 Groups don't mix, each protects its own culture, and there are

no mixed marriages. 0.23

Groups protect their own culture, but there are some mixed

marriages. 0.13

3. Do you think Bosniaks, Croats and Serbs in Bosnia can live peacefully together without international supervision, or has war made peaceful coexistence impossible?

They can live peacefully together without supervision. 0.41 War has made peaceful coexistence impossible. 0.26

Don’t know. 0.34

4a. Are you willing to accept a member of the

Bosniak ethnic group to: Be your neighbor? 0.96

Be your colleague at work? 0.94

Be your friend? 0.90

Become close kin by marrying into your family? 0.68 4b. Are you willing to accept a member of the

Croat ethnic group to: Be your neighbor? 0.97

Become close kin by marrying into your family? 0.77 4c. Are you willing to accept a member of the Serb

ethnic group to: Be your neighbor? 0.95

Become close kin by marrying into your family? 0.69

5. My people fought only defensive wars. Completely Disagree. 0.09

Partly Disagree. 0.11

Partly Agree. 0.22

Completely Agree. 0.48

Don’t Know. 0.10

6. My people have suffered more than other groups

in the region throughout history. Completely Disagree. 0.11

Partly Agree. 0.32

Don’t Know. 0.12

7. My people are not perfect but have a stronger cultural tradition than other groups in the Balkan region.

Completely Disagree. 0.09

Partly Agree. 0.31

Don’t Know. 0.10

8. My people deserve and should take the leading

role in the Balkan region. Completely Disagree. 0.27

Partly Agree. 0.23

Don’t Know. 0.15

9. According to you, are there situations where it might be justified to force members of other ethnic groups to leave their homes?

It is never justified. 0.94

Justified if they did it to us; Justified in the case of politically disloyal people; Justified for the sake of higher national interest; There was no ethnic cleansing.

0.06

10. Should one forgive those who persecuted your

compatriots? Forgive, but never forget; Forgive and forget. 0.66

Never forgive. 0.24

Don’t know. 0.10

Number of respondents: 714

4a-c measure the level of closeness one is willing to accept in their contact with each group. Positive responses were added together to form one index (0-12: Mean=10.4; SD=2.3).

(17)

Measuring Immigrant Incorporation

Socioeconomic assimilation is typically defined as socioeconomic parity and measured by factors such as educational attainment, employment, and income: outcomes that may be attained independently of acculturation. Im- portant gateways to these outcomes are language acquisition and secure legal status (Alba and Nee 1997, 854-855). To capture socioeconomic assimila- tion, I create an index of six indicators: monthly income at or above the na- tional mean; completion of upper secondary or university education in Swe- den; qualified employment; no reliance on social benefits during the last five years; high Swedish language competency; and Swedish citizenship (see Table 4).

To measure acculturation strategies, I use two questions placed on the first page of the survey providing two ways of capturing respondents’ self- ascription of identity. Cross-tabulating these two questions provides a simple way of representing acculturation strategies (see Table 3). The first asks simply ‘To which people do you belong?’ The second question asks ‘Which national anthem best expresses your feelings towards your home country?’

(See Table 4.) Although the Socialist Federal Republic of Yugoslavia no longer exists, the legacy of Yugoslav nationalism clearly persists in the minds of many respondents. As a result, the acculturation strategies of separation and integration come in two forms: one in which respondents’ maintain their ethnic heritage (Bosniak, Croat, Serb or ‘other’ ex-Yugoslav national group), and one in which respondents maintain their Yugoslav heritage. The breakdown of responses according to acculturation strategy is presented in Table 3. Table 4 presents descriptive statistics for both socioeconomic assimilation and acculturation strategy. In addition, in Table 5 the relationship between socioeconomic assimilation and acculturation strategy is presented descriptively. It is clear that those adopting bicultural and assimilative acculturation strategies are more assimilated socioeconomi- cally.

Table 3. Acculturation Strategies among ex-Yugoslavs in Sweden Maintains cultural heritage?

Yes Yes No

Identifies with dominant culture of settlement country?

Yes

Biculturalism ethnic heritage

10%

Biculturalism Yugoslav heritage

3%

Assimilation

1%

No

Separation ethnic heritage

54%

Separation Yugoslav heritage

31%

Marginalization

0%

(18)

Table 4: Socioeconomic Assimilation and Acculturation Socioeconomic Assimilation additive index Dummy=1 if: 1. Total monthly income as of last month Income > 20,000 Swedish kronor1 0.53 2. Highest level of education completed in Sweden Upper-secondary school; University2 0.61 3. Primary occupation in Sweden Qualified (e.g. nurse); Highly qualified (e.g. dentist)3 0.66 4. Reliance on social benefits at any time during the last five years No0.94 5. Swedish language competency Excellent4 0.54 6. Legal statusSwedish Citizen5 0.90 Mean: 4.18. St. Dev.: 1.57. Min: 0. Max: 6 Acculturation strategy (5 categories) To which people do you belong?Which national anthem best expresses your feelings towards your home country? 6 1. SeparationBosniak/Croatian/Serbian/other ex-Yugo. groupBosnian/Croatian/Serbian/other ex-Yugoslav rep.0.54 2. SeparationYugoslavYugoslav 0.31 3. IntegrationBosniak/Croatian/Serbian/other ex-Yugo. groupSwedish 0.11 4. IntegrationYugoslavSwedish 0.03 5. AssimilationSwedish Swedish 0.01 Sample size: 714 Questions 1-6 are coded as dichotomies (0-1), the cut-off point being a strict test of socioeconomic assimilation. These dummies are added together to form an index. 1 Splits the sample near the mean into high/low income. 2 Dummy=0 if Primary school; No education in Sweden. s3 Dummy=0 if Unqualified; Unemployed; Housewife; Student; Retired. 4 Splits the sample near the mean into high/low Swedish language competency. This additive index of four questions assesses an individual’s ability to ‘understand news reports on radio/tv’; ‘communicate with authorities over phone’; ‘read books’; and 'complete an application for employment' (options: Poor; Average; Good; Excellent). The cut-off point (16) is the highest possible score. 5 Dummy=0 if Temporary Resident; Permanent Resident. 6 Options: Nacionalna Himna Bosne i Hercegovine (Bosnian); Lijepa Nasa (Croatian); Boze Pravde (Serbian); Hej Slaveni (Ex-Yugoslav); Du Gamla, Du Fria (Swedish); other ex-Yugoslav republic.

(19)

Table 5. Socioeconomic Assimilation by Acculturation Strategy

Obs Mean Std. Dev. Min Max Separation (ethnic heritage) 368 4.068 1.555 0 6 Separation (Yugoslav heritage) 210 4.152 1.594 0 6 Integration (ethnic heritage) 69 4.884 1.312 2 6 Integration (Yugoslav heritage) 23 5.478 0.846 3 6

Assimilation 5 5.200 1.304 3 6

Individual Background Variables

As mentioned in the theoretical section, individuals with greater personal and social resources are less likely to develop counterproductive coping behaviors. I argue that socioeconomic assimilation may increase access to such resources, thus decreasing the need to rely on defensive psychological resources such as conflictive beliefs. However, migrants may differ in terms of their personal and social resources prior to migration. Moreover, labor and asylum migrants may differ in part because labor migrants are more likely to be selected for economic success (see discussion of selection effects below).

At the same time, refugees from war-torn countries are believed to harbor more animosity towards former enemies as a result of their experiences of violence and displacement (Lischer 2005; Lyons 2006). To account for these issues, I control for belonging to pre-war, wartime, or post-war migration cohorts as well as having an urban background (versus rural background) prior to migration. In addition, I control for two additional basic demograph- ic variables: gender and age upon arrival in Sweden.

Sample characteristics for these variables are described in Table 6. Com- paring my sample with census data provided by Statistics Sweden, demo- graphic characteristics of the respondents do not appear to differ dramatical- ly from those of the population of migrants from former Yugoslavia living in Sweden; however, men and those with more education are somewhat overrepresented.¹²

12 In my sample, 42.5 percent of respondents are women. The mean age is 44.6 years. 12.2 percent completed at most primary school, 51.05 percent at most secondary school, and 36.75 percent up to one year or more of university studies. According to Statistics Sweden, 50 percent of migrants from former Yugoslavia living in Sweden are women, the mean age is 46 years, 28 percent completed at most primary school, 49 percent at most secondary school, and 21 percent up to one year or more of university studies. See the sections ‘befolkningsstatistik’

and ‘utbildning och forskning’ on Statistics Sweden’s website (www.scb.se).

(20)

Table 6. Control Variables: Descriptive Statistics

Age upon arrival in Sweden

01-15 0.25

16-30 0.40

31-45 0.27

46-60+ 0.08

Period of arrival in Sweden

Came prior to war 0.28

Came during war 0.57

Came after war 0.15

Lived in an urban settlement before moving to

Sweden 0.70

Female 0.42

Sample size 714

Lastly, one might ask whether the results will be affected by: 1) the selection of migrants (either into migration to Sweden or return migration to Bosnia);

or 2) country specific characteristics (such as Swedish policies towards refugees). After all, migrant perceptions of homeland war are most likely the result of a combination of both structural factors and individual characteristics.

The selection of immigrants in or out of Sweden should not be expected to influence the relationship between incorporation in Sweden and attitudes towards conflict on the individual level. It is true that labor migrants tend to be positively ‘self-selected’ for labor market success prior to migration, while the least economically successful tend to return home (Borjas 1991).

However, while labor migrants as a whole may be positively selected for economic success in Sweden, this cannot explain individual level differences in their socioeconomic incorporation. In addition, there is no reason to as- sume migrants are selected based upon their conflict perceptions.

Both types of selection are much less relevant when it comes to asylum migration since the choice to stay or leave either the origin or settlement country is greatly reduced (Chiswick 1999). This is obvious when consider- ing return migration trends. According to Statistics Sweden (2011, 21), after 10 years nearly 60 percent of labor migrants to Sweden have either returned to their country of origin or migrated to a third country. By contrast, only 4 percent of asylum migrants leave Sweden. Nevertheless, controlling for migration cohort should help capture potential differences between labor and asylum migrants.

Turning to the impact of country specific characteristics, as discussed above Sweden along with Germany and Austria was a major destination country for migrants from former Yugoslavia during the two most recent waves of outward migration: the gastarbeiter migration during the 1950s and 1960s and asylum migration during the 1990s. However, a key policy

(21)

feature distinguished Sweden from other asylum countries in Europe during the 1990s: the fact that asylum seekers were granted permanent residency in toto in 1993. Since most refugees settled permanently in Sweden after the war, the stock of migrants from former Yugoslavia living in Sweden is argu- ably more representative of migrants from former Yugoslavia generally.

Still, it is important to recognize that each country context will involve special characteristics that are unaccounted for in the analysis here. Moreo- ver, both across countries and over time integration policies may vary (from the more multicultural to the more assimilationist). For example, Sweden followed a more multicultural approach up until the 1990s when attempts were made to reconnect welfare provision to the adoption of the national language and culture (Favell 2010, 5). Nevertheless, while national context inevitably shapes immigrant incorporation it does not present a problem for the present analysis. By examining individual-level differences among immigrants in one country (Sweden), country-level characteristics are held constant.

Results: Estimated Effects of Incorporation on Attitudes towards Conflict

I argue both socioeconomic assimilation and bicultural acculturation should lead to more peaceful attitudes among migrants. Table 7 presents the relationship between an individual’s attitudes related to the Bosnian War and their levels of incorporation in Sweden. I regress ten indicators of attitudes towards conflict on socioeconomic assimilation and acculturation strategy.

Higher values denote more peaceful attitudes. I also control for the variables discussed in the previous section (For information on bivariate regressions and regressions without controls, see the Appendix).