Refugees,xenophobia,and domestic conflict

Abstract

What factors influence attitudes towards refugees? Do negative attitudes towards refugees also influence attitudes towards conflict in the host countries? Previous studies suggest that an influx of refugees, and locals’ reaction to them, may destabilize receiving countries and lead to conflict. In particular, actual or perceived negative effects of refugees’ presence, such as increased economic competition with the locals, disruption of ethnic balance in the host country, and arrival of people with ties to rebel groups may lead to an increased likelihood of civil conflict in countries that receive refugees. These effects can lead to instability by changing the locals’ incentives and opportunities of engaging in violence. Indeed, some studies find a positive correlation at the cross-national level between influx of refugees and conflict in receiving countries. We contribute to this literature by experimentally manipulating information about the externalities of hosting refugees. We conducted a survey-experiment in the summer of 2014 in Turkey, a country that hosts the largest number of Syrian refugees. We examine how different messages about the possible effects of hosting refugees – increased economic burden, disruption of ethnic balance, and ties with rebels, as well as a positive message of saving innocent women and children – affect locals’ perceptions of the refugees and their attitudes towards the Turkish-Kurdish peace process. We find that some messages cause locals, especially majority non-Kurds, to hold more negative views of the refugees, and in some cases to view them as a threat. Generally speaking, this information does not affect support for the peace process within Turkey. Rather, fundamental factors, such as partisanship, and previous exposure to conflict are better predictors of attitudes towards peace.

Keywords

civil wars conflict refugees survey experiment Syria Turkey

Introduction

The number of refugees and displaced persons worldwide has surged to its highest levels since World War II due to several ongoing conflicts, particularly the Syrian Civil War (UNHCR, 2015: 2–3). Syrian refugees have fled primarily to neighboring countries, with the largest group of them residing in Turkey (UNHCR, 2015: 15). We study the attitudes and perceptions of the local population in Turkey towards Syrian refugees, and whether exposure to refugees affects the locals’ willingness to resolve conflicts in their own country to avoid escalation.

Several factors are hypothesized to increase negative attitudes towards refugees: (1) competition over economic resources and social services (Scheve & Slaughter, 2001; Burns & Gimpel, 2000); (2) ‘disruption’ of the host country’s ethnic balance and increase in intergroup tensions (Loescher & Milner, 2004); and (3) concerns that refugees may bring weapons, have connections to rebels, and experience in fighting that may fuel conflicts in the receiving country (Weiner, 1992–93). Importantly, locals’ perception of these factors, and not necessarily their objective impact, is argued to increase tensions in host countries (Salehyan & Gleditsch, 2006: 342–344). Several studies find a positive correlation between the arrival of refugees and the spread of conflict, but also acknowledge that such links may be confounded by, or conditioned on other factors, such as the ability to control the country’s borders and the incumbent government’s perceived legitimacy (Weiner, 1992–93; Whitaker, 2003). Other studies that use subnational data on refugees and conflict do not find that the arrival of refugees leads to more conflict in places where the refugees settle (Fisk, 2014; Shaver & Zhou, 2015).

We contribute to this debate by exploring how messages about refugees influence the perceptions of refugees looking at the case of Syrian refugees in Turkey. We conducted a survey experiment in June and July of 2014 among 1,257 local residents in central, eastern, and southeastern Turkey. We randomly assigned respondents to different primes that reflect refugee-related messages (negative and positive). These primes are not only grounded in the literature on the effects of hosting refugees, but also reflect the Turkish media discourse about Syrian refugees. Thus, we also bring a note of realism into our study by manipulating information about refugees that is similar to elite discourse in Turkey. We examine the effect of these primes on attitudes towards refugees, and on support for the Turkish-Kurdish peace process.

The effect of information on attitudes towards refugees and towards domestic conflict is important for several reasons. First, perceptions of refugees’ impact on the host country are one of the channels through which refugees may spread instability (Salehyan & Gleditsch, 2006: 342). We explicitly focus on messages that manipulate perceptions of the potential impacts of refugees on the host society. We do not examine whether the ‘actual’ impact of refugees affects the likelihood of domestic conflict in Turkey. Second, attitudes towards conflict offer a lower bar to test for the negative effect of information about refugees on a host country. This is because actual violence may not accurately reflect latent attitudes as it is also conditioned on other factors, such as opportunity and resources. Additionally, while attitudes and behavior are not perfectly correlated, they are positively associated (van Stekelenburg & Klandermans, 2013), and attitude change can be used to predict future violence (Hirose, Imai & Lyall, 2016).

We have five key findings. (1) Messages about negative effects of refugees lead to negative views about the refugees and their effects on Turkey, but primarily among non-Kurdish (majority Turkish) respondents. This result is especially important because it shows that negative messages about mostly Muslim refugees can negatively impact attitudes towards them even in a middle-income, Muslim-majority country such as Turkey. (2) There is some evidence of co-ethnic bias – Kurdish respondents exhibit more positive views towards Kurdish refugees than towards other refugees, and more than non-Kurdish respondents. Sunni refugees are not as disliked as the other groups following our negative treatments. (3) Additionally, even though some of our primes negatively affect attitudes towards refugees (especially the message that refugees may be bringing arms), they do not lead the respondents to be less supportive of the Turkish-Kurdish peace process. This calls into question previous studies that argue that negative perceptions of refugees, and generally a higher salience of conflict, may increase support for violence (Salehyan & Gleditsch, 2006; Kuran, 1998). (4) Locals who report a high level of exposure to refugees in their daily life express more negative views of, and higher threat perceptions of refugees. However, exposure to refugees does not change the effects of our treatments. (5) Finally, partisan identification matters: supporters of the governing party (AKP) are less likely to view the refugees as a threat.

While information may induce negative attitudes towards refugees among some respondents, it does not change attitudes towards domestic conflict. Instead, these are shaped to a great extent by geography and partisanship. Our study makes three important contributions. (1) We study a major host country (Turkey) that has received the largest number of Syrian refugees compared to other host countries.¹ (2) While many of the media and policy debates on Syrian refugees focus on their effects on Europe, the vast majority of refugees are hosted in surrounding countries, such as Turkey, Lebanon, and Jordan (Arar, Hintz & Norman, 2016). This pattern is not unique to the Syrian refugee crisis, as refugees from other civil wars also tend to flee primarily to neighboring countries (Moore & Shellman, 2007). (3) Finally, by examining the individual-level perceptions that are said to affect the locals’ response to refugees, we show that although anti-refugee appeals may have some traction, they do not increase support for conflict in the host country.

We proceed first by providing background on Syrian refugees in Turkey, and on the main policy issue that coincided with their arrival – the Turkish-Kurdish peace process. We then review previous studies on the effect of refugees on their host countries, followed by a brief discussion of our survey design and sampling (a detailed description is in the Online appendix), and present our findings. We conclude by outlining the implications of our study for understanding the impact of refugees on public attitudes of the locals.

Background: Syrian refugees and Turkey

In response to the arrival of the first Syrian refugees in early 2011, Turkey opened several refugee camps to provide food, healthcare, and education. At the time of our survey, 22 camps were operational in ten provinces (see map in Figure A.1 in the Online appendix). Despite these efforts, only 219,688 out of over 900,000 refugees in Turkey resided in camps during our survey. The remaining refugees lived among the local population, primarily in provinces with camps (UNHCR, 2014).

Relations between Syrian refugees and the local residents have been relatively peaceful.² Nonetheless, locals showed signs of discontent: some blamed the refugees for the increase in housing prices (Sak, 2014), the rise in unemployment, competition with local businesses (Çetingüleç, 2014), and even for social ills such as thefts, murders, smuggling, and prostitution (Erdoğan, 2015).

Media coverage reflected those negative sentiments, but also contained positive depictions of refugees. We searched the archive of Hürriyet – a leading newspaper website – between April 2011 and June 2014 and found 112 articles directly related to Syrian refugees in Turkey.³ The most frequent negative theme is the possible connection between refugees and rebels (Hürriyet, 2012). Other articles highlight the economic burden of hosting the refugees, especially in border provinces (Hürriyet, 2014), and the refugees’ ethnic identity as a contributor to sectarian tensions. For instance, Alawite refugees are reluctant to reside in majority Sunni refugee camps (Yezdani, 2013). Finally, a number of articles highlight the positive humanitarian aspects of hosting refugees (Erkuş, 2014). This theme echoes messages of the AKP government and the open-door policy and assistance to Syrian ‘brothers’ fleeing from the violence of a cruel regime.⁴ These different aspects of the refugee crisis form the basis of the treatments in our survey experiment.

Turkish-Kurdish peace process

At the time of our survey, one of the main issues at the forefront of Turkish domestic politics was the Turkish-Kurdish peace process – negotiations between the government and the Kurdistan Workers’ Party (PKK) to end a 30-year-long conflict in exchange for political reforms.⁵ The peace process began at the end of 2012 when the government announced negotiations with the jailed PKK leader Öcalan (Haberturk, 2012). Although the exact details of these talks are not fully known, the broad outlines of a potential deal included a ceasefire and withdrawal of the PKK, democratization, and PKK’s disarmament and normalization (Yeğen, 2015).

Turkish public opinion in 2014 was divided with respect to these negotiations (Yılmaz, 2014). While 57% supported the peace process, about 38% opposed it. Ethnicity and partisanship were the two strongest determinants of attitudes towards the peace process, with the Kurds and the AKP supporters being more in favor of this initiative than others.

Attitudes towards refugees and the peace process

Locals may harbor negative attitudes towards refugees for several reasons. For instance, they might be concerned that refugees alter the demographic balance in the host country (Loescher & Milner, 2004). This is especially true if refugees are ethnic kin to one of the aggrieved groups (for instance, the arrival of Kurdish refugees from Syria in Turkey). Locals may also mobilize into a ‘sons of the soil’ movement if they feel that refugees threaten their long-standing status in their native regions (Weiner, 1992–93). For example, around the time of our survey, there were reports of tensions in Hatay between Turkish Alawites and Arab Sunni refugees from Syria,⁶ and concerns over the arrival of Sunni Arab refugees to the historically Kurdish areas (Çağaptay & Menekşe, 2014).⁷

Besides fomenting negative attitudes, information about refugees may affect positions about conflict and cooperation in the host country. First, perceptions that refugees have connections to rebel groups may contribute to the locals’ sense of insecurity, not only making them more negative towards refugees, but also fueling existing conflicts, or spurring new disputes (Schweitzer et al., 2005; Lischer, 2005; Faulkner et al., 2004). For example, Turkey’s assistance to Sunni Arab rebels and an ‘open-door’ policy that brought many Sunni Arab refugees, some of whom were possibly connected to rebels (Hakan, 2012), may harden Kurdish negotiation positions. In this context, the Turkish-Kurdish leader Demirtaş suggested that ‘developments in Syria might come to a stage to block the Kurdish peace process in Turkey’ due to the government allowing support to flow from Turkey to radical Islamists who are attacking Kurds in Syria (Yetkin, 2013). Moreover, Forsberg (2014) argues that when an ethnic group engages in an ethnic conflict in one country, it may also encourage kin in neighboring countries to rebel. In this context, the fact that the Syrian Kurds play such a central role in the Syrian Civil War may also encourage the Kurds in Turkey to harden their negotiation positions. Since refugees often settle along ethnic lines (Rüegger & Bohnet, 2012), exposure to Kurdish refugees coming from Syria may further lower their willingness to compromise with the Turkish government. In addition, the perception that refugees exacerbate economic competition may lead to conflicts between the locals and the refugees over jobs, housing, and public goods (Loescher & Milner, 2004; Salehyan & Gleditsch, 2006). Indeed, rents in Gaziantep, an industrial city of 1.4 million people (around 300,000 of whom are Syrians), doubled with the arrival of refugees. This has led to fights between locals and refugees, and anti-Syrian rallies in the city (Kütahyalı, 2014). Moreover, refugees’ arrival and the government’s integration policy may lead to conflict, or at least lower the locals’ willingness to compromise with each other, by augmenting the salience of ethnic-based grievances among the locals (Kuran, 1998). For instance, the Kurds may accuse the government of a ‘double standard’ for allowing Syrian refugees to receive education in Arabic while not permitting education in Kurdish in Turkish public schools (Kirişçi, 2014: 25). Kurds in Turkey have long demanded education in the Kurdish language, with both Kurdish politicians and the PKK suggesting this should be part of the reforms agreed upon through the peace process negotiations (Yeğen, 2015: 13). Conversely, government assistance to refugees and the fact that some of them are Kurdish and are more likely to settle among Turkish Kurds (Rüegger & Bohnet, 2012) may predispose the latter to compromise with the government.

Attitudes towards refugees and their effect on conflict are likely not uniform, with different parts of the local population taking different views. For instance, negative attitudes towards newcomers are often more prevalent among members of the ethnic majority who fear that their status may be threatened (Valentino, Suhay & Brader, 2008; Koopmans & Olzak, 2004). Common religion may also attenuate negative attitudes towards refugees. However, as Lazarev & Sharma (2017) show in the context of Syrian refugees in Turkey, economic concerns may trump the effects of shared identity. Residing in a place previously affected by political violence increases altruism (Gilligan, Pasquale & Samii, 2014), and may positively predispose respondents towards refugees, although in some cases members of such communities become more hawkish towards the outgroup (Getmansky & Zeitzoff, 2014). Partisan identification may also shape the locals’ position towards refugees. Indeed, AKP supporters in Turkey are more sympathetic towards Syrians (EDAM Center for Economics and Foreign Policy Studies, 2014). Lastly, political elites in host countries can play a decisive role in framing public opinion (Zaller, 1992; Hainmueller & Hopkins, 2014). Ordinary citizens are often not well informed about refugees, so elites may try to inflame (or downplay) the perceived threat from newcomers as a way to increase their own political support (Gadarian & Albertson, 2014).

In sum, primes about an influx of refugees may exacerbate negative attitudes toward refugees, particularly when emphasizing the negative externalities of hosting them (i.e. economic competition, disruption of ethnic balance, connection to militants). Whether negative perceptions of the refugees translate into lower support for the Turkish-Kurdish peace process is less clear.

Figure 1.

Map of the sampled districts

Negative primes may increase the salience of ethnicity, further exacerbating ethnic tensions, and reduce support for compromise between Turks and Kurds. Yet, ethnicity is also likely to moderate the effect of our primes about refugees given that Kurds and Turks have different economic status and reside in places with different exposure to past violence.

Research design and data

Sampling

We employed a stratified sampling procedure to produce variation on three key factors associated with refugees and Turkish politics: refugee presence, past support for the incumbent party (AKP), and a history of Turkish-Kurdish violence. More detailed discussion of our sampling procedure is in the Online appendix (section A.3). Figure 1 depicts our sample’s geographic distribution. Table A.2 in the Online appendix lists the provinces, districts, and the number of respondents in each district.

We surveyed 1,257 respondents, and our response rate was 34%.⁸ Our sample includes non-Kurds (58%) and Kurds (42%). Similar to previous studies of public opinion in Turkey (Yılmaz, 2014; Kalaycıoğlu, 2009; Kalaycıoğlu & Çarkoğlu, 2007; Çarkoğlu & Toprak, 2007), we classify someone as Kurdish if they list their first language as Kurdish.⁹ Sample demographics are shown in Table I.

The Kurdish and non-Kurdish samples are well balanced on gender, age groups, AKP support, and residing in an urban district. A total of 43% of Kurds in the sample live in provinces that were previously under state-of-emergency rule (OHAL) due to the Turkish-Kurdish conflict, whereas only 13% of non-Kurds lived in these provinces (see section A.3 in the Online appendix for the list of these provinces). As we would expect from a marginalized minority, Kurds on average have lower household income, and are less likely to own some household goods. They are also slightly less educated than the non-Kurds. Kurds also appear to be more religious, as

Table I.

Sample demographics

	Non-Kurds					Kurds
Variable	N	Mean	Std. dev.	Min.	Max.	N	Mean	Std. dev.	Min.	Max.
Female	716	0.50	0.50	0	1	526	0.48	0.50	0	1
Age group	716	2.43	1.10	1	4	526	2.44	1.09	1	4
Married	715	0.70	0.46	0	1	524	0.76	0.43	0	1
AKP supporter	681	0.50	0.50	0	1	479	0.51	0.50	0	1
OHAL	716	0.13	0.33	0	1	526	0.43	0.50	0	1
Urban district	716	0.72	0.45	0	1	526	0.75	0.44	0	1
Ramadan survey	716	0.12	0.32	0	1	526	0.09	0.29	0	1
Education	716	3.99	1.42	1	8	524	3.45	1.65	1	8
Household income	697	5.50	2.87	1	16	487	4.21	2.12	1	16
Components of Wealth
Smart phone	716	0.33	0.47	0	1	526	0.17	0.38	0	1
Car	716	0.31	0.46	0	1	526	0.20	0.40	0	1
Computer	716	0.51	0.50	0	1	526	0.25	0.44	0	1
Washing machine	716	0.93	0.26	0	1	526	0.89	0.31	0	1
Dish washer	716	0.65	0.48	0	1	526	0.45	0.50	0	1
Components of Religious
Cover hair	702	5.00	1.96	1	7	518	6.19	1.51	1	7
Alcohol not OK	709	4.35	2.30	1	7	524	4.72	2.64	1	7
Pray	682	3.88	2.03	1	7	512	5.03	1.56	1	7
Components of self-reported Refugee exposure
Public transport	708	0.67	0.37	0	1	503	0.50	0.42	0	1
Street	707	0.64	0.37	0	1	502	0.69	0.36	0	1
Business	706	0.28	0.40	0	1	499	0.23	0.38	0	1
Social life	702	0.47	0.42	0	1	495	0.36	0.40	0	1
Market	702	0.62	0.38	0	1	503	0.56	0.40	0	1

Wealth, Religious, and Refugee exposure are factor variables created using the components listed below each item. The differences in the number of respondents are due to missing values.

demonstrated in questions about alcohol use, frequency of prayer, and whether women in the family cover their hair. Finally, Kurds report slightly lower levels of refugee exposure in their daily lives except when it comes to seeing them on the streets.¹⁰

Treatments

We randomly assigned subjects to one of five experimental conditions (Table II). In the Control condition, subjects did not receive any information about refugees. In the other four treatments, the enumerator read a brief statement heightening the refugees’ salience in Turkey, and then subjects received treatments (see section B.1 in the Online appendix for the exact wording of our treatments). The Economic cost, Ethnic balance, and Militant ties treatments were all meant to heighten a key negative aspect of hosting refugees. The Women and children

Table II.

Experimental conditions

Treatment	View on refugees	Description
Control	---
Economic cost	Negative	Syrian refugees impose large costs on government resources, and increase unemployment among Turkish citizens.
Ethnic balance	Negative	Syrian refugees upset Turkey’s ethnic balance.
Militant ties	Negative	Syrian refugees have ties to militant groups that make Turkey less safe.
Women and children	Positive	Turkey’s refugee policy has saved many innocent women and children.

treatment was meant to balance out the negative tone of the treatments with a more positive tone, and to reflect AKP’s justification of its open-door policy towards the refugees. By including the positive prime, we can also see whether it is the negative aspects of

Figure 2.

Responses to the main dependent variables questions – Kurdish/non-Kurdish respondents comparison

hosting refugees, or simply mentioning refugees, that affects public opinion. We present randomization checks in section C in the Online appendix.

Dependent variables

Following the treatment, the enumerators asked subjects about their views on our main dependent variables: sympathy towards different groups of refugees,¹¹ whether respondents view the refugees as an economic threat and as a threat to safety, and support for the Turkish-Kurdish peace process. The exact wording of specific items is in section B.2 of the Online appendix.

Figure 2 shows the distribution of our dependent variables. We present the responses of Kurds and non-Kurds separately because group identification is an important cleavage in Turkey, and likely affects the way respondents view refugees (Valentino, Suhay & Brader, 2008; Koopmans & Olzak, 2004). Moreover, Kurds are the biggest minority group in Turkey, and the peace process concerns their welfare directly, especially in regions that we surveyed. There are noticeable differences between Kurds and non-Kurds. The four upper panels of Figure 2 show that: (1) Kurdish respondents have warmer attitudes than non-Kurds towards all groups of refugees; (2) Kurdish respondents exhibit the warmest attitudes towards Kurdish refugees, and then towards Sunnis. Non-Kurds express slightly warmer attitudes towards Sunni refugees, and treat the other groups equally. This suggests that common religion may be associated with a more sympathetic attitude towards refugees,¹² but not as much as common ethnic identity. The two lower panels show that both Kurds and non-Kurds are concerned that refugees are making Turkey less safe, and that they pose an economic threat to Turkey. These views are slightly more prevalent among non-Kurds than Kurds. Finally, the figure also suggests – similarly to other studies (Yılmaz, 2014) – that Kurds are more supportive of the peace negotiations than non-Kurds.

Findings

Empirical Strategy

We are interested in how information about different externalities of hosting refugees (our treatments) affects the dependent variables while controlling for fundamental attributes – partisanship, refugee exposure, living in areas that experienced political violence, and ethnicity. Our empirical specification for individual i in province j is the following OLS model:

Y_{i j} = α_{1} \times T_{i} + α_{2} \times K u r d_{i} + α_{3} \times K u r d_{i} \times T_{i} + β_{1} \times A K P_{i} + β_{2} \times O H A L_{j} + β_{3} \times R e f u g e e E x p o s u r e_{i} + γ \times X_{i} + μ_{j} + ∊_{i}

where Y_ij is the individual response to our dependent variables, T_i is a vector of binary treatment indicators, Kurd_i is a binary indicator of whether respondent i’s first language is Kurdish, AKP_i is an AKP supporter binary indicator, OHAL_j is a binary variable equal 1 if individual i lives in an OHAL province.¹³ Refugee exposure_i is the individual (self-reported) level of refugee exposure,¹⁴ and X_i is a vector of individual controls.¹⁵ The inclusion of these parameters allows us to investigate how individual-level ethnicity and party identity, as well as residing in an area of conflict, affect views of the refugees and their impact on Turkey. We control for other province-level covariates by including a province-level dummy μ_j. ∊_i is a normally distributed error. We allow for differential treatment effects for non-Kurds (α₁) and Kurds (α₁ + α₃) based on our theoretical underpinnings that messages about refugees are likely to influence minorities and majorities differently. We are also interested in the effect of the key observational variables (β).

Results

We address three principal research questions: (1) Do the different primes about Syrian refugees change attitudes towards refugees relative to the Control? (2) How do they influence attitudes towards the Turkish-Kurdish peace process? (3) How do fundamental attributes – partisanship, refugee exposure, living in a place with past experience of violence, and ethnicity – influence attitudes towards refugees and domestic conflict? We present bootstrapped coefficient plots of the treatments and observational variables of interest in Figures 3–4. The full regression results are provided in Tables E.7–E.8 in the Online appendix. All the dependent and independent variables (that are not dummy variables) are rescaled to lie between 0 and 1. In the Online appendix, we show that our main results hold when we estimate our models using weighted regression.¹⁶ In addition, our results hold when we estimate the models without control variables (Tables F.11–F.12, and section F.2 in the Online appendix), split the sample into non-Kurds vs. Kurds (Tables F.22–F.25 in the Online appendix), and use an ordered probit model instead of OLS (Tables F.20–F.21 in the Online appendix). We discuss additional robustness checks below.

Attitudes towards refugees

Figure 3 demonstrates that the Militant ties and the Women and children treatments make non-Kurds less likely to express positive attitudes towards refugees than non-Kurds in the control group. In particular, with respect to Arab, Alawite, and Kurdish refugees, these treatments lower the response of non-Kurds by 9 to 12 percentage points compared to the response of non-Kurds in the control group (with a 95% confidence interval of [1, 20] using bootstrap standard errors).

Whereas the negative effect of Militant ties treatment is intuitive, the result with respect to Women and children is surprising. One might expect that reminding the respondents that many refugees are innocent women and children would increase positive attitudes towards them. We believe that this negative effect is driven by the fact that over 75% of the refugee population in 2014 was women and children (UNHCR, 2014), and the perception that women and children are unproductive populations that would continue to require support

Figure 3.

Positive attitudes towards different groups of Syrian refugees

in the future.¹⁷ Orhan & Gündoğar (2015) provide support for this interpretation, and find that seeing Syrian children begging on the streets in Turkey increases the negative perception of refugees among the locals. An additional factor that may contribute to the negative effect of the Women and children prime are reports around the time of the survey that local men were marrying Syrian refugee brides as second or third wives. Erdoğan (2015) reports that instances of polygamy and child marriages involving Syrian refugees generate discontent among locals towards young Syrian females. The increase in divorce rates has also been

Figure 4.

Refugee primes, threat perceptions, and support for the Turkish-Kurdish peace process

blamed on marriages between Turkish men and Syrian women (Orhan & Gündoğar, 2015). Thus, a seemingly positive treatment about saving Syrian women and children may have also reminded respondents of specific demographic-related, negative externalities that come with hosting young women and children refugees.

Our treatments do not influence the attitudes of Kurds towards the refugees, with the exception of the Economic cost treatment, which leads Kurdish respondents to express more positive attitudes towards Kurdish refugees (not towards other refugees), compared to Kurds that do not receive such treatment (significant only at 90% level; p-value is 0.06). We interpret this as weak evidence of co-ethnic bias in treatment of refugees, especially when Kurdish respondents are informed that the Turkish government is spending resources on refugees who also include members of their own ethnic group, the Syrian Kurds.¹⁸

The Militant ties treatment has no significant effect on Kurds’ attitudes towards refugees. This may be partially because Kurds in Turkey have sympathies for the Syrian Kurdish Democratic Union Party (PYD) and its armed wing, Popular Defense Units (YPG), most of whose recruits are also Kurdish (Paasche, 2015). Therefore, the Kurds in our sample may have interpreted the refugees having ties to armed groups not necessarily as a negative development. However, this account fails to explain why the Kurds do not show negative attitudes towards other groups of refugees, such as Arab refugees, when they receive the Militant ties treatment.

Finally, attitudes towards Sunni refugees are not influenced by our treatments. The only exception is the Women and children treatment that makes non-Kurds more negative towards Sunnis, but this result is significant only at 90% level. Following all other treatments, respondents become neither negative nor positive towards Sunni refugees compared to the control group. This is consistent with Lazarev & Sharma (2017), who also find that shared religion is not a strong determinant of pro-refugee attitudes.

Among the non-experimental variables, living in provinces previously under OHAL (provinces that experienced political violence due to the Turkish-Kurdish conflict) is strongly associated with more positive attitudes towards all groups of refugees. Self-reported daily exposure to refugees (Refugee exposure) is not associated with negative attitudes towards refugees (in the case of Arab and Alawite refugees, it is associated with more positive attitudes towards these groups, and these results are significant at 90% and 99% levels, respectively). This is contrary to the argument that contact with refugees is likely to foment xenophobic sentiments among the locals. Finally, partisanship (AKP) is not a significant determinant of respondents’ attitudes towards refugees. AKP supporters are somewhat positive towards Sunnis, but this result is marginally significant (90%).

In sum, our results suggest that: (1) primes related to possible military ties of the refugees, as well as the arrival of women and children, negatively affect attitudes towards the refugees; and (2) these primes affect non-Kurds that receive these treatments relative to non-Kurds in the control group, and do not affect Kurds (ethnic minority). Finally, (3) residing in a province with a history of violence (OHAL) is associated with warmer feelings towards refugees. As a robustness check, we show in the Online appendix that the differences between Kurds and non-Kurds are not because the Kurds are less exposed to Syrian refugees in their daily life.

Refugees as a threat

In the upper panel of Figure 4, we observe a similar pattern of negative attitudes towards refugees among the non-Kurdish respondents. Following the Women and children treatment, non-Kurdish respondents are 9 percentage points more likely to view the refugees as an economic threat, compared to non-Kurdish respondents in the control group (95% confidence interval of [0.4, 17]). The same treatment causes the non-Kurds to be 8 percentage points more likely to agree that refugees are making Turkey less safe compared to non-Kurds in the control group (95% confidence interval of [0.1, 17]). Non-Kurds are also 9 percentage points more likely to say that refugees are a safety threat following the Militant ties treatment compared to non-Kurds in the control group (95% confidence interval [0.4, 17]). Kurdish respondents are not more likely to view refugees as an economic or a safety threat following the Militant ties treatment. Comparable to our explanation for the Militant ties treatment not affecting Kurds’ attitudes towards Kurdish refugees, this result can be explained by the fact that Kurds in Turkey sympathize with, or at least are less threatened by the PYD and its armed wing, the YPG. Therefore, Kurds in our sample may have interpreted the refugees having ties to armed groups not necessarily as a threat.

The non-experimental variables also exert a strong effect on attitudes towards refugees. AKP supporters are significantly less likely to view refugees as a threat, in line with the AKP platform. Respondents from OHAL provinces are not particularly threatened by the refugees. There is even some indication that they have a lower perception of economic threat, though it is significant only at 90% level. Conversely, higher self-reported exposure to refugees (Refugee exposure) is associated with a heightened threat perception. In our sample, 43% of the Kurds live in former OHAL provinces,¹⁹ and the primes do not increase Kurds’ negative feelings towards the refugees. Taken together, these two points strongly suggest that Kurds are more welcoming and less threatened by the refugees than non-Kurds. In the robustness checks section, we demonstrate that this result is not due to their lower exposure to refugees.

Refugee primes and attitudes towards peace

In the lower panel of Figure 4, we look at the effect of the experimental primes and the key non-experimental variables on support for the peace process. We chose to focus on the peace process because our goal is to examine whether refugees affect attitudes towards domestic conflict; the PKK insurgency, and the efforts to end the insurgency through peaceful means were the main issue of domestic conflict in Turkey at the time of our survey.

The only treatment that has a significant effect for Kurds is the Economic cost treatment on support for the peace process. Specifically, there is an increase of 12 percentage points in support for the peace process among Kurds in the Economic cost group compared to Kurds in the control group, with 95% confidence interval of [5, 20] using bootstrap standard errors. Apart from this, our treatments do not affect respondents’ view on the peace process.

For the non-experimental variables, AKP supporters are more supportive of the peace process, reflecting a strong partisan effect on political attitudes. Other non-experimental variables are not statistically significant, except for respondents in former OHAL provinces who are more supportive of the peace process but this effect is significant only at the 90% level. This finding is not surprising because this region suffered the most from the Turkish-Kurdish conflict, and the majority of the respondents from former OHAL provinces are Kurdish (71%). Overall, these results show that partisanship has the strongest effect on the respondents’ positions on domestic conflict while the treatments do not have significant effects except the Economic cost treatment on the Kurds’ support for the peace process. One explanation of this finding is that Kurds, hearing that the government is spending resources on helping the refugees who also include their ethnic kin (Syrian Kurds), become more in favor of the peace process, which aims at expanding the political rights of the Kurds in Turkey.

Robustness checks

So far we find that Kurds and non-Kurds respond differently to our treatments: Kurds do not express negative attitudes towards refugees, do not view refugees as a threat, and become more supportive of the peace process. There are two alternative explanations for this finding: (1) Kurds are less likely to interact with refugees in their daily lives compared to non-Kurds. Indeed, Kurds are more likely to reside in the OHAL region, where the reported exposure to refugees is lower; and (2) Kurds as members of a minority group may provide favorable answers to sensitive political questions.

To explore the first possibility – that Kurds’ answers are driven by their low exposure to refugees – we repeat our tests using data on provinces with camps where refugee exposure is higher.²⁰ We explore this in Tables F.13 and F.14, and section F.3 in the Online appendix. Our main results hold. Even if we focus only on provinces with camps, where exposure to refugees is higher, Kurds still do not express negative attitudes towards refugees, view them as a threat or become less supportive of the peace process.

It is also possible that Kurds do not express their ‘true’ views, but instead, as an ethnic minority, provide socially desirable answers. To explore this possibility, we also asked the respondents to what extent they identify with Turkey. This question is sensitive in the Turkish context (Şirin, 2013: 81), and to the extent that social desirability is an issue, we expect it to be evident in this question. Specifically, if the Kurdish respondents express ‘politically correct’ views, then the variable Kurd should have no association with the level of identification with Turkey. However, the Kurds in our sample are less likely to identify strongly with Turkey than the non-Kurds, which suggests that they do not necessarily provide socially desirable answers. Results are in Table F.15 in the Online appendix.

One of our strongest findings is that respondents from provinces with a history of violence (OHAL) have warmer attitudes towards the refugees. To examine whether we observe the same relationship with a more nuanced measure of political violence at the province-level, we estimate our models using the province-level PKK fatalities data in Tezcür (2015).²¹ Similar to the results with the OHAL dummy, the results show that respondents from provinces with larger number of PKK fatalities have warmer attitudes towards refugees, and are more favorable to the peace process.

An additional concern may be that response rates vary systematically, and that refusal to participate in the survey is higher in the former OHAL region. In fact, refusal rates are slightly lower in OHAL provinces (p < 0.01). This strongly suggests that our results are not driven by respondents from areas with past conflict refusing to participate.

Finally, a critique may be that our experimental findings would be different if we conducted the survey today, with nearly three million refugees currently residing in Turkey (as of June 2017). Two of our results provide some clues suggesting that the experimental results would not be so different. First, we explicitly studied areas in Turkey’s southeast, some of which were already experiencing significant refugee presence. Importantly, as we show in the Online appendix (Tables F.16–F.17), exposure to refugees does not moderate the effects of our treatments, so those with high levels of daily exposure to refugees do not respond differently to our treatments relative to those who have lower refugee exposure. This suggests that additional influx of refugees should not change our experimental findings.

Summary and implications

We examine how messages about Syrian refugees affect attitudes of local residents in Turkey towards refugees and towards the Turkish-Kurdish peace process. We show that priming commonly hypothesized negative effects of refugees (especially Militant ties), as well as a positive prime (Women and children), increases negative attitudes towards refugees and threat perceptions among members of majority group (non-Kurds). Yet these primes do not lower support for the Turkish-Kurdish peace process. Conversely, for Kurds, the Economic cost treatment in effect makes them more supportive of the peace process.

The non-experimental variables also provide several interesting findings. Supporters of the ruling AKP party are not more positive towards the refugees compared to supporters of other parties, but they are systematically less likely to see refugees as a threat. There is strong evidence that residing in a province with a history of political violence (OHAL) is associated with more positive feelings towards refugees and greater support for the peace process. This is consistent with previous findings that community-level experience with political violence is associated with an increase in altruistic attitudes.

This is not to say that refugees do not have negative externalities, and that emphasizing these externalities does not heighten negative attitudes towards refugees.²² Yet we show that in the context of one of the largest refugee crises, and against an ethnically tense backdrop (Turkey-PKK conflict), locals’ political attitudes are not as malleable as many hypothesize. While information about refugees may affect perception of refugees among some respondents, it does not affect support for the peace process. Rather, pre-existing factors (partisanship and place of residence) shape attitudes towards peace.

It is also important to highlight that the finding that primes change attitudes, but not policy preferences, is in line with other survey experiments that find similar effects of information (Kuziemko et al., 2015). Relatedly, Grigorieff, Roth & Ubfal (2016) demonstrate that while information about immigrants affects the locals’ attitudes towards them, it has a limited effect on immigration policy preferences of the local respondents. In our case, we provide information about possible externalities of hosting refugees from a civil war in a neighboring country, thereby increasing the salience of the negative effects of civil conflict. We show that some of this information moves attitudes towards refugees and even threat perception, but has no effect on policy preferences related to solving a domestic conflict.

One concern might be that the recent ethnic tensions between Kurdish groups and the Turkish government potentially call into question our findings that refugees do not necessarily heighten local tensions. For instance, a Kurd from Turkey, who received training in the Islamic State camps in Syria, perpetrated a suicide attack in the border town of Suruç in Turkey that killed 32 individuals and triggered the collapse of the peace process (BBC, 2015). Hence, this attack might be considered as evidence for the transnational rebel networks mechanism identified in the literature. However, the recent round of violence in Turkey seems to have more to do with electoral dynamics (İdiz, 2015), or foreign policy related to the Syrian civil war (Çağaptay, 2015), rather than the influx of refugees. Moreover, an equally plausible, alternative interpretation of the Suruç attack is that Turkey’s open-door policy was the cause of both the large Syrian refugee flows and the higher chances of armed conflict in Turkey.

It is also important to acknowledge that some features of Turkey may set it apart from other recipients of Syrian refugees, especially European countries. In particular, Turkey is a predominantly Sunni-Muslim country hosting mostly Sunni refugees, and the effect of refugee-related messages on the locals may differ if there are religious differences between the locals and the refugees. Yet it is also the case that most refugees flee to neighboring countries that tend to be similar to their country of origin.

Finally, our findings have important policy implications. Many NGOs and governments seek to reduce the locals’ anxiety and promote positive views of the refugees.²³ However, we show that emphasizing positive messages featuring women and children refugees does not necessarily make the locals more supportive of refugees, and can actually backfire, depending on how these messages are interpreted by the locals. Another implication that emerges from our findings is that information that refugees may have weapons and ties with militant groups has a particularly strong effect on attitudes and threat perception among the locals. Thus, addressing these concerns and reassuring the local population that refugees are not a threat should be an important priority.

Footnotes

Replication data

The dataset and do-file for the empirical analysis, along with the Online appendix, can be found at .

Acknowledgements

We are grateful for the feedback from participants of the Conflict Consortium Virtual Workshop on November 19, 2014, and from participants of the Peace Science Society (International) 2014 Annual Meeting and the Empirical Studies in Political Analysis Workshop in 2015. Finally, we would like to thank David Carter, Jacob N Shapiro, and Empirical Studies of Conflict (ESOC) workshop participants for their helpful comments on an earlier experimental design.

Funding

This project was supported by Boğaziçi University Research Fund (BAP) Grant Number 8681. We also would like to thank Amaney Jamal and the Mamdouha S Bobst Center for Peace and Justice, as well as Helen Milner and the Niehaus Center for Globalization and Governance for their generous funding.

Notes

References

Arar

Rawan

Hintz

Lisel

Norman

Kelsey P

(2016) The real refugee crisis is in the Middle East, not Europe. Washington Post 14 May (https://www.washingtonpost.com/news/monkey-cage/wp/2016/05/14/the-real-refugee-crisis-is-in-the-middle-east-not-europe/).

Australian Associated Press (2014) Almost 1 million Syrian refugees in Turkey, 22 April (http://www.sbs.com.au/news/article/2014/04/22/almost-1-million-syrian-refugees-turkey).

BBC (2013) Blasts kill dozens in Turkish town Reyhanlı on Syria border, 11 May (http://www.bbc.com/news/world-middle-east-22494128).

BBC (2015) Suruc massacre: ‘Turkish student’ was suicide bomber, 22 July (http://www.bbc.com/news/world-europe-33619043).

Burns

Peter

Gimpel

James G

(2000) Economic insecurity, prejudicial stereotypes, and public opinion on immigration policy. Political Science Quarterly 115(2): 201–225.

Çarkoğlu

Ali

Toprak

Binnaz

(2007) Religion, Society and Politics in a Changing Turkey. İstanbul: Turkish Economic and Social Studies Foundation (TESEV).

Çağaptay

Soner

(2015) Turkey is in serious trouble. Atlantic, 5 October (http://www.theatlantic.com/international/archive/2015/10/turkey-isis-russia-pkk/408988/).

Çağaptay

Soner

Menekşe

Bilge

(2014) The Impact of Syria’s Refugees on Southern Turkey. Policy Focus no. 130. Washington Institute for Near East Policy (https://www.washingtoninstitute.org/uploads/Documents/pubs/PolicyFocus130_Cagaptay_Revised3s.pdf).

Çetingüleç

Mehmet

(2014) Syrian refugees aggravate Turkey’s unemployment problem. Al-Monitor, 9 July (http://www.al-monitor.com/pulse/originals/2014/07/cetingulec-syrian-refugees-turkey-unemployment-illegal-work.html).

10.

Daloğlu

Tülin

(2014) Turkey admits Reyhanli was attacked by al-Qaeda. Al-Monitor 4 April (http://www.al-monitor.com/pulse/originals/2014/04/reyhanli-qaeda-bombing-attack-admits.html).

11.

Deutsche Welle (2013) After attacks, distrust lingers in Turkey’s Reyhanlı, 25 August (http://www.dw.com/en/after-attacks-distrust-lingers-in-turkeys-reyhanli/a-17043402).

12.

Dressler

Markus

(2008) Religio-secular metamorphoses: The re-making of Turkish Alevism. Journal of the American Academy of Religion 76(2): 3280–3311.

13.

Economist (2015) Refugee realpolitik, 24 October. (https://www.economist.com/news/europe/21676780-angela-merkels-response-migrant-crisis-less-emotional-and-more-risky-it).

14.

EDAM Center for Economics & Foreign Policy Studies (2014) Reaction Mounting Against Syrian Refugees in Turkey. Public Opinion Surveys of Turkish Foreign Policy 2014/1 (http://edam.org.tr/en/reaction-mounting-against-syrian-refugees-in-turkey/).

15.

Erdoğan

Murat

(2015) Türkiye’deki Suriyeliler: Toplumsal Kabul ve Uyum Araştırması [Syrians in Turkey: Social Acceptance and Integration Research]. Ankara: Hacettepe Üniversitesi Göç ve Siyaset Araştırmaları Merkezi (HUGO) Yayınları (https://mmuraterdogan.files.wordpress.com/2015/02/tc3bcrkiyedeki-suriyeliler-syrians-in-turkey-rapor-tr-en-190220151.pdf).

16.

Erkuş

Sevil

(2014) 900,000 Syrians take refuge in Turkey. Hurriyet Daily News 10 March (http://www.hurriyetdailynews.com/900000-syrians-take-refuge-in-turkey-unicef.aspx?pageID=238&nid=63374&NewsCatID=341).

17.

Evin

Mehveş

(2015) Merciless and hopeless reality for Syrian refugees. Al-Monitor 2 February (http://www.al-monitor.com/pulse/en/politics/2015/02/turkey-syria-refugee-crisis.html).

18.

Faulkner

Jason

Schaller

Mark

Park

Justin H

Duncan

Lesley A

(2004) Evolved decease-avoidance mechanisms and contemporary xenophobic attitudes. Group Processes & Intergroup Relations 7(4): 333–353.

19.

Fisk

Kerstin

(2014) Refugee geography and the diffusion of armed conflict in Africa. Civil Wars 16(3): 255–275.

20.

Forsberg

Erika

(2014) Transnational transmitters: Ethnic kinship ties and conflict contagion 1946–2009. International Interactions 40(2): 143–165.

21.

Gadarian

Shana Kushner

Albertson

Bethany

(2014) Anxiety, immigration, and the search for information. Political Psychology 35(2): 133–164.

22.

Getmansky

Anna

Zeitzoff

Thomas

(2014) Terrorism and voting: The effect of rocket threat on voting in Israeli elections. American Political Science Review 108(3): 588–604.

23.

Gilligan

Michael J

Pasquale

Benjamin J

Samii

Cyrus

(2014) Civil war and social cohesion: Lab-in-the-field evidence from Nepal. American Journal of Political Science 58(3): 604–619.

24.

Grigorieff

Alexis

Roth

Christopher

Ubfal

Diego

(2016) Does information change attitudes towards immigrants? Evidence from survey experiments. Unpublished manuscript (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2889665).

25.

Haberturk (2012) İmralı’yla görüşüyoruz [We are in talks with İmralı], 29 December (http://www.haberturk.com/gundem/haber/807198-imraliyla-gorusuyoruz).

26.

Hainmueller

Jens

Hopkins

Daniel J

(2014) Public attitudes toward immigration. Annual Review of Political Science 17(1): 225–249.

27.

Hakan

Ahmet

(2012) Looking for Al-Qaida members in Antakya. Hurriyet Daily News 22 September (http://www.hurriyetdailynews.com/looking-for-al-qaida-members-in-antakya.aspx?pageID=238&nID=30709&NewsCatID=396).

28.

Hirose

Kentaro

Imai

Kosuke

Lyall

Jason

(2016) Civilian attitudes and insurgent tactics in civil war. Unpublished manuscript (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2446168).

29.

Hürriyet (2012) 71 Syrian soldiers flee to Turkey, 9 November (http://www.hurriyetdailynews.com/71-syrian-soldiers-flee-to-turkey-report.aspx?pageID=238&nid=34307&NewsCatID=352).

30.

Hürriyet (2014) Syrian refugee inflow doubles house prices in Turkish border cities, 5 March (http://www.hurriyetdailynews.com/syrian-refugee-inflow-doubles-house-prices-in-turkish-border-cities.aspx?pageID=238&nID=63204&NewsCatID=345).

31.

İdiz

Semih

(2015) Erdogan continues to stir the pot in Turkey. Al-Monitor 8 September (http://www.al-monitor.com/pulse/originals/2015/09/turkey-erdogan-continues-stir-cauldron-elections-pkk.html).

32.

Kalaycıoğlu

Ersin

(2009) Public choice and foreign affairs: Democracy and international relations in Turkey. New Perspectives on Turkey 40(1): 59–83.

33.

Kalaycıoğlu

Ersin

Çarkoğlu

Ali

(2007) Turkish Democracy Today: Elections, Protest and Stability in an Islamic Society. New York: IB Tauris.

34.

Kennedy

Craig

Glenn

John

Balme

Natalie La

Isernia

Pierangelo

Everts

Philip

Eichenberg

Richard

(2011a) Transatlantic Trends Survey 2008, Codebook. Ann Arbor, MI: Inter-University Consortium for Political and Social Research.

35.

Kennedy

Craig

Zsolt

Nyiri

Balme

Natalie La

Isernia

Pierangelo

Everts

Philip

Eichenberg

Richard

(2011b) Transatlantic Trends Survey 2009, Codebook. Ann Arbor, MI: Inter-University Consortium for Political and Social Research.

36.

Kirişçi

Kemal

(2014) Syrian Refugees and Turkey’s Challenges: Going Beyond Hospitality. Washington, DC: Brookings Institution (https://www.brookings.edu/wp-content/uploads/2016/06/Syrian-Refugees-and-Turkeys-Challenges-May-14-2014.pdf).

37.

Kirişçi

Kemal

Winrow

Gareth M

(1997) The Kurdish Question and Turkey: An Example of a Trans-State Ethnic Conflict. London: Frank Cass.

38.

Koopmans

Ruud

Olzak

Susan

(2004) Discursive opportunities and the evolution of the right-wing violence in Germany. American Journal of Sociology 110(4): 975–997.

39.

Kuran

Timur

(1998) Ethnic dissimilation and its international diffusion. In: Lake

David A

Rothchild

Donald

(eds) The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation. Princeton, NJ: Princeton University Press, 35–60.

40.

Kütahyalı

Rasim Ozan

(2014) Syrian refugees under attack in Turkey. Al-Monitor 14 August (http://www.al-monitor.com/pulse/ar/originals/2014/08/kutahyali-syrian-refugees-under-attack-turkey-gaziantep.amp.html).

41.

Kuziemko

Ilyana

Norton

Michael I

Saez

Emmanuel

Stantcheva

Stefanie

(2015) How elastic are preferences for redistribution? Evidence from randomized survey experiments. American Economic Review 105(4): 1478–1508.

42.

Lazarev

Egor

Sharma

Kunaal

(2017) Brother or burden: An experiment on reducing prejudice toward Syrian refugees in Turkey. Political Science Research and Methods 5(2): 201–219.

43.

Lischer

Sarah Kenyon

(2005) Dangerous Sanctuaries: Refugee Camps, Civil War, and the Dilemmas of Humanitarian Aid. Ithaca, NY: Cornell University Press.

44.

Loescher

Gil

Milner

James

(2004) Protracted refugee situations and state and regional insecurity. Conflict, Security & Development 4(1): 3–20.

45.

Marcus

Aliza

(2007) Blood and Belief: The PKK and the Kurdish Fight for Independence. New York: New York University Press.

46.

McDowall

David

(1996) A Modern History of the Kurds. London: IB Tauris.

47.

Moore

Will H

Shellman

Stephen M

(2007) Whither will they go? A global study of refugees’ destinations, 1965–1995. International Studies Quarterly 51(4): 811–834.

48.

Nisan

Mordechai

(2002) Minorities in the Middle East: A History of Struggle and Self-Expression. Jefferson, NC: McFarland.

49.

Orhan

Oytun

Gündoğar

Sabiha Senyücel

(2015) Effects of the Syrian Refugees on Turkey. Ankara: Center for Middle Eastern Strategic Studies (ORSAM) (http://www.orsam.org.tr/files/Raporlar/rapor195/195eng.pdf).

50.

Paasche

Till F

(2015) Syrian and Iraqi Kurds: Conflict and cooperation. Middle East Policy 22(1): 77–88.

51.

Rüegger

Seraina

Bohnet

Heidrun

(2012) No undifferentiated mass: Introducing a quantitative data set on the ethnic composition of refugee movements. Working paper (http://file.prio.no/files/projects/RueggerBohnet_RefugeeDataPaper.pdf).

52.

Sak

Güven

(2014) Of fake housing bubbles and Syrian refugees. Hurriyet Daily News, 6 September (http://www.hurriyetdailynews.com/of-fake-housing-bubbles-and-syrian-refugees.aspx?pageID=449&nID=71347&NewsCatID=403).

53.

Salehyan

Idean

Gleditsch

Kristian Skrede

(2006) Refugees and the spread of civil war. International Organization 60(2): 335–366.

54.

Scheve

Kenneth F

Slaughter

Matthew J

(2001) Labor market competition and individual preferences over immigration policy. Review of Economics and Statistics 83(1): 133–145.

55.

Schweitzer

Robert

Perkoulidis

Shelley

Krome

Sandra

Ludlow

Christopher

Ryan

Melanie

(2005) Attitudes towards refugees: The dark side of prejudice in Australia. Australian Journal of Psychology 57(3): 170–179.

56.

Shaver

Andrew

Zhou

Yang-Yang

(2015) Questioning refugee camps as sources of conflict. Working paper (https://scholar-dev.princeton.edu/sites/default/files/ISADraft_0.pdf).

57.

Şirin

Çiğdem V

(2013) Analyzing the determinants of group identity among Alevis in Turkey: A national survey study. Turkish Studies 14(1): 74–91.

58.

Tezcür

Güneş M

(2015) Electoral behavior in civil wars: The Kurdish conflict in Turkey. Civil Wars 17(1): 70–88.

59.

UNHCR (2014) UNHCR Turkey Syrian Refugee Daily Sitrep, June 6 (http://data.unhcr.org/syrianrefugees/download.php?id=6183).

60.

UNHCR (2015) Global Trends: Forced Displacement in 2015 (http://www.unhcr.org/uk/statistics/unhcrstats/576408cd7/unhcr-global-trends-2015.html).

61.

Valentino

Nicholas A

Suhay

Elizabeth

Brader

Ted

(2008) What triggers public opposition to immigration? Anxiety, group cues, and immigration threat. American Journal of Political Science 52(4): 959–978.

62.

Van Stekelenburg

Jacquelien

Klandermans

Bert

(2013) The social psychology of protest. Current Sociology Review 61(5–6): 886–905.

63.

Weiner

Myron

(1992–93) Security, stability, and international migration. International Security 17(3): 91–126.

64.

Whitaker

Beth Elise

(2003) Refugees and the spread of conflict: Contrasting cases in Central Africa. Journal of Asian and African Studies 38(2–3): 211–231.

65.

Yeğen

Mesut

(2015) The Kurdish peace process in Turkey: Genesis, evolution and prospects. Global Turkey in Europe Working paper 11 (http://www.iai.it/sites/default/files/gte_wp_11.pdf).

66.

Yezdani

İpek

(2013) Alawites escape Jihadists in Syria, take shelter in Turkey. Hurriyet Daily News 9 September (http://www.hurriyetdailynews.com/alawites-escape-jihadists-in-syria-take-shelter-in-turkey.aspx?pageID=238&nID=54073&NewsCatID=352).

67.

Yetkin

Murat

(2013) Turkey’s Syria policy problem. Hurriyet Daily News 21 September (http://www.hurriyetdailynews.com/turkeys-syria-policy-problem-.aspx?%20pageID=238%20&nid=54869&NewsCatID=409).

68.

Yılmaz

Hakan

(2014) Türkiye’de Kimlikler, Kürt Sorunu ve Çözüm Süreci: Algılar ve Tutumlar [Identities, Kurdish question and the solution process in Turkey: Perceptions and attitudes]. Working paper (http://hakanyilmaz.info/yahoo_site_admin/assets/docs/OSI_Yurttaslik_2014_Cozum_Sureci_Sunum_Basin_v08.259182726.pdf).

69.

Zaller

John R

(1992) The Nature and Origins of Mass Opinion. New York: Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB