Ideological Extremism Among Syrian Refugees Is Negatively Related to Intentions to Migrate to the West

The conflict in Syria created a dire humanitarian situation, as nations around the world struggled with how best to deal with the more than 6.6 million Syrian refugees who fled their homes to escape aggression. Resistance to granting refugee status to individuals often originates in the belief that the influx of refugees endangers national security because of the presumably extremist religious and political beliefs that refugees hold. The present research surveyed Syrian refugees residing in Jordan, Lebanon, Turkey, and Iraq (N = 1,000). The results revealed that the majority of surveyed refugees did not intend to migrate to the West and would rather return to their home country. More importantly, refugees most interested in moving to Western countries were the least likely to subscribe to Islamist extremism or to harbor negative sentiment toward the West. Theoretical and practical implications for addressing the current refugee crisis are discussed.

likelihood of terrorism in their country (Wike et al., 2016) and wished that their country would take in no refugees (Esipova & Ray, 2017). Also widespread are the concerns that refugees threaten the cultural cohesion of the host communities (Landmann et al., 2019).
The popularity of such narratives says more about the psychology of the people who endorse them than about the refugees themselves, given that little evidence supports the validity of such narratives. Admittedly, an analysis of refugee-related violence found an association between streams of refugees and terrorism within the hosting nations, but violence against refugees ( Gineste & Savun, 2019) and the aid workers assisting them (Choi & Salehyan, 2013) outpaced violence perpetrated by refugees. Research has yet to address why these few refugees engaged in political violence. It could be that they held extremist views prior to migration (which could have determined or been inconsequential to their decisions to migrate), or their extreme views could have been instilled in them after their arrival in the host nation. These issues cannot be addressed only by surveying individuals who are already residing in host communities.
Most research on refugees' migration focuses on the conditions in the refugees' country of origin that motivated a decision to leave. It demonstrates that refugee streams are more likely when violence is present in refugees' countries of origin (e.g., Adhikari, 2013;Balcilar & Nugent, 2019;Moore & Shellman, 2004). This is unsurprising as, by definition, refugees are escaping threats to their safety in their home country. Presently, however, Syrian refugees are traveling through relatively safe neighboring countries. The question, therefore, is what instills in them the desire to migrate further? One argument is that in such a situation, refugees are similar to other migrant populations, and they choose to reside in locations that best afford the fulfillment of various economic (Constant & Massey, 2003) or relational (Constant & Zimmermann, 2012) needs. This research has largely neglected the role that ideological and political beliefs have in this process. Thus, although this work has advanced our understanding of the dynamics of migration, it is possible that it has examined an incomplete set of motivational factors, including some that could possibly supersede the importance of those that have already been studied.
Research on the importance of homophily suggests that political beliefs should play an important role in these decisions. People generally prefer to distance themselves from those who do not share their norms and values (Howells, 1966). Indeed, this tendency toward homophily is strengthened among people with strong attachments to the norms and values of the in-group, and it has been shown to result in spatial and social segregation of peoples with differing values (Schelling, 1971;Semyonov et al., 2007). Moreover, people avoid intergroup contact when they wish to protect important values from cultural dilution (Waterman & Kosmin, 1988) and when they predict that contact with strangers will result in anxiety (Landmann et al., 2019). Most relevant to migration decisions, findings of past research show that the perceived ideological fit between one's personal political beliefs and the political beliefs of one's community impacts the desire to relocate within one's country (Motyl et al., 2014).
We built on this literature by examining whether the political attitudes held by Syrian refugees were systematically related to their intentions to migrate to the West or their intentions to migrate to another country. Rather than study the role of general political attitudes (i.e., conservative vs. liberal), we measured endorsement of extreme beliefs, namely, strong commitment to an absolutist version of one's religion and willingness to make large sacrifices for it, support for violence to advance ideological goals, and hostility toward the out-group (i.e., Western countries). Given that such attitudes might predispose individuals to conflicting intergroup relations (Atran & Ginges, 2012), understanding their connection to migration intentions might be also practically important.
We therefore suggest that refugees with extreme political and religious beliefs should be less willing to emigrate to Western countries, whose citizens do not share their worldviews. Such individuals likely expect difficulties adapting to a new cultural context and a negative reception from the locals should they decide

Statement of Relevance
The conflict in Syria has resulted in the displacement of more than 6.6 million refugees. Denizens of Western countries have been reluctant to accept Syrian refugees, in part because of the perception that these individuals hold extremist religious or political views. We surveyed 1,000 Syrian refugees living in Jordan, Lebanon, Turkey, and Iraq. The results show that the refugees most motivated to migrate to Western countries are those that are least likely to endorse extreme tenets of their religion or harbor anti-Western sentiment. On the other hand, those refugees who are more ideologically extreme do not want to migrate to Western countries and choose to move back to their home country instead.
to migrate. They should be more motivated to return to their home country instead. In contrast, refugees who express attitudes and beliefs that are sympathetic to the population in which they wish to embed themselves (i.e., Western culture) should be more willing to migrate to the West. Whereas these predictions are in line with past findings on intergroup contact, to our knowledge they have not been examined in the context of longterm, possibly even lifelong, decisions about relocating to a different country.

Method
We first explored these ideas in Jordan in 2016. We recruited Syrian refugees from outside refugee camps, given that around 90% of Syrian refugees did not live in camps at this time (UNHCR, 2019). To replicate these findings, we expanded our sample to Turkey, Lebanon, and Iraq in 2016 and 2017. In those countries, we also recruited refugees who resided in refugee camps, although this effort was less successful in Turkey. We also revised the surveys used in Jordan to better capture our constructs of interest. In what follows, we describe our samples, measures, and main results. The details of the materials and additional analyses are presented in the Supplemental Material available online.
Data sets, analysis scripts, and materials for all studies have been made publicly available via OSF (https:// osf.io/86tbg/). Because of potential concerns with violating the privacy and anonymity of our participants, demographic information has been removed from data sets. However, whether the demographic variables were included in the analyses or not, the results for the focal variables did not differ. In the Supplemental Material, we present the correlations between the focal variables and the demographics.

Jordan
Participants. In Jordan, we recruited 250 Syrian refugees, a sample size that gave us at least .80 power to detect a small to medium correlation (r = .20). Moreover, a sample of this size should allow for stable correlation estimates (Schönbrodt & Perugini, 2013). To increase their confidence that their responses would be private and would in no way be linked to them personally, we asked participants to consent verbally. They were also informed that the research was conducted independently of local governments and that participants' responses would not be shared with those governments and hence would not impact their migration decisions. An independent research organization (with no governmental or university affiliation) was responsible for the recruitment of participants. The organization had extensive past experience conducting research with refugees in Jordan in partnership with local and international organizations (e.g., UNHCR). Participants were recruited in two ways. First, potential participants were identified by publicly available registry information and contacted by trained Arabic-speaking research associates. Second, potential participants were identified by a local nongovernmental organization (NGO) from their lists of individuals who have received their services, and they were recruited by the same NGO to participate in the study. Gender and ethnicity-matched interviewers collected data using computer-assisted personal interviewing. Specifically, one interviewer read each question and the corresponding answer options aloud to one participant at a time and then filled in the responses. The organization in collaboration with researchers was also responsible for translating the survey from English to Arabic and then from Arabic back into English.
Six non-Muslim participants were excluded from analyses because one measure of ideological extremism was directly related to religious belief (i.e., Islamist ideology), and it was not relevant to non-Muslim participants. More importantly, the non-Muslim group was too small to allow for meaningful comparisons on other measures. The remaining sample consisted of 244 individuals (122 women, 122 men; mean age = 35.80 years, SD = 12.47). Because of missing values on some variables (defined as more than 50% of items left unanswered), the sample sizes differed in each analysis. Because the results obtained using listwise deletion were not different from those using pairwise deletion, we present the latter analyses to avoid reducing power. The demographic details of the samples are presented in Table S1 in the Supplemental Material.

Measures.
Migration intentions. Two items measured participants' willingness to relocate to a Western country: "How much would you like to move to a Western country?" and "Do you intend to move to a Western country?" Responses were given on a scale ranging from 1, not at all, to 7, very much so (r = .93, p < .001). The same two items were asked regarding migration back to Syria (r = .53, p < .001).

Ideological extremism.
Because there is no one agreed-on measure of ideological extremism, we used several scales to test the generalizability of results. Those scales included measures of (a) Islamist ideology, (b) willingness to sacrifice for one's religion, and (c) willingness to sacrifice for a political cause.
Islamist ideology was operationalized as commitment to the fundamental tenets of Islam (e.g., "I think it is important to establish an Islamic state in my country"), including support for religiously motivated violence (e.g., "Armed Jihad is a personal obligation of all Muslims today"). It was measured using a 15-item scale created in conjunction with Muslim clerics in Singapore and has been used in past research ( Jasko, Webber, Kruglanski, et al., 2020;Webber et al., 2018). Two reverse-coded items were excluded because they did not correlate with the scale ("There are many different ways of interpreting Islam" and "Under no circumstances does Islam support the killing of civilians"). We averaged the remaining 13 items (α = .87).
The extent to which refugees were willing to engage in violent and nonviolent forms of sacrifice for religion was measured with six items (e.g., "I would be willing to give away all my belongings to defend my religion," α = .87). We used the same set of items to measure sacrifice for a political cause, namely in service of defending the rights of immigrants in Western countries (α = .94).
Sociodemographic variables. We also measured the following sociodemographic variables: age, gender, education, socioeconomic status in Jordan and in Syria, and whether participants had family living in the United States, European Union, Syria, and Jordan. More than 90% of participants had family in Syria and Jordan.

Lebanon, Turkey, and Iraq
Participants. In order to replicate the findings of the first study in a larger sample, we surveyed 250 Syrian refugees in each of the three countries: Lebanon, Turkey, and Iraq. In these countries, we collaborated with a different company, but the procedure was similar.
As in Jordan study, 28 non-Muslim participants were excluded from analyses (23 in Lebanon, three in Turkey, and two in Iraq). The remaining sample consisted of 722 individuals (347 women [48.1%], 375 men, mean age = 36.57 years, SD = 13.17), of which 246 participants lived in the refugee camps. Because of differences in accessibility, the group of refugees living in the camps was unevenly distributed across countries (119 in Lebanon, 29 in Turkey, and 98 in Iraq).
Measures. Some measures were slightly refined and revised from those used with the Jordan sample (see the Supplemental Material for the full list of questions). Migration intentions were measured with the same items as in the Jordan study.
Ideological-extremism scales. The measure of Islamist ideology was expanded to capture endorsement of both extreme ideology and extreme violence. Thus, 18 items were used to measure both aspects of Islamist extremism.
One item was added to the scales measuring willingness to sacrifice for religion and willingness to sacrifice for political rights. In Jordan, the items measuring the latter variable asked participants about immigrants' rights. In the present study, those items were rephrased to ask about refugees' rights, to be more relevant to participants (i.e., "I would be willing to give my life if it prevented the oppression of refugees in Western countries").
Finally, two other scales of ideological extremism were added to capture political attitudes and beliefs regarding the West. A five-item scale measured the willingness to support violence toward the West (Lebanon: α = .91, Turkey: α = .86, Iraq: α = .99). Sample items include, "I would support Syrian people fighting against the West" and "I would support my friends and family if they were fighting against the West." Another five-item scale measured general negative attitudes toward the West (e.g., "Western cultures are generally immoral," "Western countries often violate other people's human rights around the world"; Lebanon: α = .92, Turkey: α = .79, Iraq: α = .97).
Sociodemographic variables. In addition to the variables used in Jordan, we also controlled for whether an interview was conducted inside or outside of a refugee camp.

Results
The Jordan sample did not include refugees living inside refugee camps, but the remaining three samples did. Controlling for this factor did not change the results for Lebanon and Turkey, but it did change the results in Iraq, where we obtained large differences between refugees living inside and outside of camps. Therefore, for Iraq, we report correlations separately for participants surveyed inside and outside of camps. The correlations split by camp for Lebanon are presented in the Supplemental Material. The number of refugees living inside camps in Turkey was too small to allow for meaningful comparisons.

Descriptive statistics
Means and standard deviations for all variables of interest are provided in Table 1. Distributions for the migration-intention variables and Islamist ideology are depicted in Figure 1. Because of space limitations, the distributions for the remaining variables are presented in the Supplemental Material.
Migration intentions. In Jordan, Lebanon, and Turkey, participants expressed little interest in migrating to the West, as demonstrated by mean levels of intentions that were below the midpoint of the scale. As depicted in Figure 1, the mode of responses was a 1 in all three of these samples. This means that the most frequent pattern of responding was that participants answered both questions about Western migration (i.e., "would like" and "intend" to move to a Western country) using the lowest option on the scale, not at all. For intentions to return to Syria, participants in all three locations expressed a much stronger motivation. Opposite to the mode for intentions to migrate West, the mode of the responses in Jordan, Lebanon, and Turkey was a 7, indicating that many participants responded to both questions with the highest scale option of very much so. In Iraq, differences between refugees inside camps and outside of camps emerged. Specifically, refugees living outside of camps declared stronger intentions to migrate to the West, but they also wanted to go back to Syria.
Ideological extremism. In Jordan, Lebanon, and Turkey, mean responses on most of the ideologicalextremism variables were around or below the midpoint of the scales, indicating an overall low level of ideological extremism among refugees. Two patterns are worth highlighting. First, higher means were observed on scales assessing religious views (i.e., Islamist ideology and willingness to sacrifice for the religion) relative to the more political and Western-focused scales (i.e., willingness to sacrifice for political rights, fighting against the West, negative attitudes toward the West). Second, we examined the responses on scales measuring religious extremism at the individual-items level. This is informative, because some of the questions directly mentioned violence (e.g., "I would be willing to attack police or security forces to defend my religion"), whereas others did not (e.g., "I would be willing to give away all my belongings to defend my religion"). The overall level of agreement with the items that referred to violence was lower than agreement with the items that did not mention violence (see Fig. 2 for the sacrifice-for-one's-religion variable). Overall, nonviolent sacrifice for religion and commitment to fundamentalist religious ideals seemed to resonate the most with participants, whereas support for violence in general and aggressive tendencies toward the West in particular evoked low agreement.
In Iraq, there were large differences between refugees living inside and outside camps. Refugees living inside of camps reported low levels of Islamist ideology, less negative attitudes toward the West, and lower support for fighting against the West. In contrast, refugees living in Iraq outside of camps held the most negative views toward the West out of all samples.

Bivariate analyses: ideological extremism and migration intentions
To test our main hypotheses, we next examined correlations between migration intentions (i.e., willingness to migrate to the West and willingness to go back to Syria) and different measures of ideological extremism (see Table 2). Given that migration intentions and some of the extremism variables did not follow a normal distribution, we used Spearman's correlation. The results for Jordan, Lebanon, and Turkey remained the same when analyzed using Pearson's correlation. Here, we briefly summarize the relations between migration intentions and all measures of extremism to provide information about the consistency of findings across measures. Because of space limitations, Figure 1 presents scatterplots only in regard to Islamist ideology. The remaining plots are presented in full in the Supplemental Material.
We first examined intention to migrate to Western countries. In Jordan, Lebanon, and Turkey, endorsement of Islamist ideology was negatively correlated  with the intention to migrate to the West, as depicted in Figure 1. More generally, willingness to migrate to the West was significantly and negatively related to all measures of political and religious extremism in Jordan and Lebanon. In Turkey, it was negatively related to all but one measure of extremism. Relative to the other locations, the correlations found in Iraq were generally weaker, and the pattern of correlations was less consistent across measures. This applied to refugees living outside and inside camps.
Examination of intentions to return to Syria revealed a different pattern. Some measures of ideological extremism were positively correlated with intentions to return, and others were unrelated to willingness to go back to Syria. In Jordan, the significant correlation was found with willingness to sacrifice for political rights, whereas in Lebanon, Islamist extremism was related to return intentions. In Turkey, four out of five extremism variables were positively related to willingness to go back to Syria. Again, the pattern of results in Iraq was inconsistent.
Together, these findings reveal a pattern (albeit less so in Iraq): Refugees who are more ideologically extreme are less interested in relocating to Western countries and, if anything, are more interested in returning to their homeland.

Additional analyses
Ideological extremism and negative expectations toward the West. To help illuminate why ideologically extreme individuals would be hesitant to migrate to Western countries, we explored the correlations between ideological extremism and refugees' expectations of the West. Specifically, we anticipated that more ideologically extreme refugees would expect that they would not be accepted in Western countries and that they would receive negative treatment. This should be reflected in expectations that their various psychological needs (both higher-order needs for respect and more basic concerns such as safety and employment) would not be met living in the West.
Our surveys contained items tapping these perceptions. Specifically, a series of items (13 in Jordan and 12 in the other locations) measured the extent to which refugees felt their basic needs (e.g., physical security, access to food, health care, opportunities for employment) would be frustrated (or satisfied) should they move to a Western country. Another series of 10 items measured the extent to which participants felt their higher-order needs would not be satisfied upon moving to a Western country-for example, items related to humiliation, respect (reverse coded), and acceptance (reverse coded). We averaged across those two sets of items to form one index of negative expectations toward the West. Analyses conducted separately for basic needs and higher-order needs are available in the Supplemental Material, and the differences between correlations were small or nonexistent.
The analysis of correlations between expected frustration of needs and ideological extremism supported the predictions. In Jordan, Lebanon, and Turkey we found that all measures of ideological extremism were consistently and significantly correlated with negative expectations toward the West ( Jordan: .37 ≥ r S ≥ .23; Lebanon: .77 ≥ r S ≥ .43; Turkey: .59 ≥ r S ≥ .30). In other words, refugees who were more ideologically extreme expected to experience greater frustration of both their lower-and higher-order needs after moving to the West. To provide insight into these results, we depict correlations between Islamist ideology and negative expectations in Figure 3. All correlations can be found in the Supplemental Material. For Iraq, the pattern of results again differed from those for the other three locations, and it differed between refugees living inside and outside of camps. In short, the results in three locations supported the idea that more extreme beliefs are related to more negative expectations toward the treatment the refugees might encounter in the West.
Desirability concerns. Because of the sensitive character of our study and the sample, the potential effects of desirability concerns should be considered. Specifically, refugees who were motivated to migrate to the West may have been more likely to provide less extreme responses to questions about their ideology and attitudes toward the West if they believed that their true, and possibly more extreme, responses could hurt their migration chances. We were aware of these concerns and designed consent procedures to reduce this likelihood. Still, in this section, we present several additional analyses, some of which relied on additional variables included in our study but not pertinent to the main analyses, conducted to examine whether possible desirability concerns impacted the results. All questions included in our surveys are listed in the Supplemental Material.
Sacrifice for an ideological cause versus sacrifice for the family's safety. First, if refugees adjusted their endorsement of violence downward because they were afraid that it could impact their chances of getting asylum, they should have done so consistently across different measures of aggressive intentions. Instead, we found a rather nuanced pattern of results. Our survey included three scales that measured willingness to support and engage in extreme actions (e.g., willingness to attack police or security forces). Two of those scales were related to a political and religious cause (i.e., sacrifice for religion and sacrifice for political rights) and were described earlier. The third scale was related to the goal of "keeping one's family safe," which we assumed would be higher among refugees and was included to verify that participants responded to questions in a discriminating way. Thus, participants responded to the same set of "extreme" items as for religious and political causes but with regard to the issue of the safety of their family. Responses to all three scales were positively correlated with each other across the samples (the exception was Iraq, where there were nonsignificant relationships between the scales), and indeed, the willingness to sacrifice for one's family's safety was the strongest of all three. They were, nonetheless, differentially related to migration intentions when included in one model. Although willingness to sacrifice for religion was negatively related to migration intentions to the West (the result that was reported earlier in the main analyses), willingness to sacrifice for the safety of one's family through extreme measures was positively and significantly related to migration intentions in Lebanon, Turkey, and Iraq (.22 ≥ β ≥ .15) and was unrelated to migration intentions in Jordan. In other words, refugees who were more motivated to seek safety for their families and were willing to do whatever it takes to achieve this goal (including intending to engage in violent and sacrificial actions) were at the same time more interested in migrating to the West. Although this is not a surprising result in itself, it suggests that participants shared their true opinions even when those opinions expressed support for extreme actions.
Violent versus nonviolent extremism. Second, we separately analyzed the responses to questions that explicitly mentioned violence (vs. those that did not). As reported earlier, the mean level of endorsement of those questions was different. Importantly, however, responses to violent and nonviolent (but still extreme) items correlated strongly with each other and correlated to a similar extent with migration intentions. This suggests that even when social-desirability concerns were less likely to influence the responses to some questions (as indicated by higher levels of agreement with those items), those questions nonetheless show the same pattern of results as questions that would be more likely to be influenced by social desirability. These analyses are available in the Supplemental Material.

Ideological extremism and need for cognitive closure.
Finally, the validity of participants' responses would be reinforced if they correlated in a theoretically meaningful way with other variables unlikely to be influenced by desirability concerns. To test this possibility, we examined the correlation between our extremism items and the need for cognitive closure (Kruglanski & Webster, 1996). The need for cognitive closure captures individuals' desire for certainty and predictability. Past research has found positive associations between need for cognitive closure and ideological extremism (Webber et al., 2018) and intergroup hostility (cf. Roets et al., 2015). Thus, if refugees truthfully answered the questions measuring ideological extremism, we should have obtained correlations between those measures and the need for cognitive closure. Refugees had no reason to adjust their responses on the latter scale given that its content and placement (as the first scale in the survey) gave no reason for suspicion. As expected, this scale was positively and significantly correlated with a majority of extremism measures, including Islamist ideology, willingness to sacrifice for religion, and negative attitudes toward the West across all samples (see the Supplemental Material for details). These findings, which replicate relations previously found between need for cognitive closure and extremism, suggest that our participants' responses were authentic rather than a reflection of social desirability.
In short, even though we cannot completely rule out the possibility that participants in our studies who were more motivated to go to the West somewhat modified their answers to questions about political or religious extremism, we believe that the additional analyses presented above attest to the validity of our findings.

Results in Iraq.
Given the inconsistent pattern of results obtained in the Iraqi sample, in May 2017, we used the same survey to collect data from another 250 refugees living in Iraq. The goal was to verify whether the results from the first sample could be attributed to sampling variability or whether there was something unique and replicable about Syrian refugees in Iraq. After we excluded one non-Muslim participant, the sample consisted of 99 refugees living inside camps and 150 refugees living outside camps. Again, there were large differences between participants living inside and outside of camps. Notably, this time refugees living outside of camps were less extreme and less motivated to go to the West than in the sample reported earlier. The bivariate results across samples and measures were also inconsistent. However, the types of inconsistencies found in the first sample were not replicated in the second sample, so we refrain from making conclusions regarding the correlations between migration intentions and extremism in this group. One point deserves mention: After collecting data, we realized that participants in Iraq differed from refugees in other countries in that they all identified as Kurds. We suspect this might have played a role in their responses, given the history of this minority group both in Syria and in Iraq and its ambivalent relationship to Western countries. We included a detailed comparison of both waves of our Iraqi data collection in the Supplemental Material.

Discussion
The findings of our studies help clarify the migration decisions of refugees who, after managing to escape the ravages of conflict in their home countries, are temporarily residing elsewhere. Refugees in this situation might consider staying where they are, returning home, or migrating to yet another country. We studied these decisions with refugees from the Syrian conflict. In contrast to past studies that focused largely on safety concerns or material needs of asylum seekers as predictors of their migration intentions, our study focused on refugees' social and political attitudes because there is virtually no empirical evidence on this topic.
The first result worth highlighting is that the majority of participants did not intend to migrate to the West and would rather return to their home country. These findings replicate surveys conducted by UNHCR (2020) with Syrian refugees living in Egypt, Lebanon, Jordan, and Iraq. In a similar vein, the majority of participants in our samples did not support political violence, particularly when the questions about its use referred to the Western context. This finding is also consistent with other research on political violence ( Jasko, Webber, & Kruglanski, 2021), which shows that majorities reject ideologically motivated violence.
Arguably the most important finding is that the refugees most interested in moving to the West are less likely to endorse tenets of Islamist extremism, do not wish to engage in costly sacrifice or violence for their beliefs, and harbor no negative sentiment toward Western countries. Additional analyses suggested that this might occur because refugees who hold such beliefs recognize that their strong commitment to their cultural and political values would lead to their being treated as deviants in a Western society. In contrast, the ideological concerns of ideologically extreme refugees may appear to be best served by returning to their homelands, where the issues they care about are a salient priority. Indeed, those refugees in our samples who were more committed to an extreme version of their ideology seemed to be more motivated to go back home to Syria, although this relationship was less consistent across measures of extremism. This could be due to the immense uncertainty that surrounds the possibility of return and what the country would look like when they returned.
Taken together, these results have important implications. On a theoretical level, we hope to extend the literature on migration decisions by incorporating ideological factors, which have not been tested in this context before. If migration indicates a willingness to break ties with one's country of origin and to immerse oneself in a different culture, sociopolitical attitudes should be related to those decisions. Our data are consistent with this contention. Past research has predominantly focused on social beliefs of immigrants and refugees who had already resettled in their host countries (e.g., research on acculturation patterns), so more research on the underpinnings of migration intentions is needed. Our findings suggest that treating political and social attitudes as predictors of migration intentions is a promising new direction for such research.
From a practical perspective, our findings paint a picture of refugees who want to travel to the West as holding less extreme political and religious views and as being more positive toward Western countries than those refugees who do not want to leave the region. Importantly, not only are they less extreme in relative terms, but also examination of the distribution of responses show that their absolute levels of extremism are low. Still, these results run counter to the views of some populations in the West that are wary of accepting large numbers of refugees into their countries because of the hostile views that they attribute to the strangers, a view that is historically rather common (Krogstad, 2019).
The present results build on the limited existing literature to suggest that concerns regarding the security threat posed by incoming Syrian refugees may be overblown. Whereas the refugees themselves may pose little security threat, the process of migration is still a concern that needs to be addressed. Perhaps because of the fears held by people in countries hosting refugees, an influx of refugees is related to increased violence by extremists against refugees (Gineste & Savun, 2019). Moreover, organizations with nefarious intentions may exploit (and have exploited) migration streams into Western countries as a means of importing violence. For instance, although the perpetrators of the Paris attacks in 2015 were European nationals, some of them posed as refugees and used migration streams as a means of traveling to and from Middle Eastern countries (Wihtol de Wenden, 2015). And finally, our data speak only to political views prior to migration. It is possible that after migrating to Western countries, refugees' political views may become more extreme. This scenario is most likely if refugees are met with hostility and experience alienation in the host countries. One potentially fruitful avenue is to use the findings of the present study as an intervention to reduce intergroup hostility: Providing members of host nations with these results may increase positive attitudes toward integration of refugees. Past research, although scarce, suggests that correcting inaccurate metaperception about the out-group might be an effective way of reducing intergroup conflict (e.g., for metahumanization, see Kteily et al., 2016; for correcting beliefs about group polarization, see Lees & Cikara, 2020).
Although the present research contains important findings, it is not without limitations. In addition to the desirability concerns that we addressed earlier, we also acknowledge that the findings in Iraq were inconsistent with the findings in other locations. These inconsistencies were at least partially driven by stark differences between individuals residing inside and outside of refugee camps. Moreover, as noted earlier, the refugees in Iraq self-identified as Kurds. It is possible that this unique context changed the way our participants interpreted the surveys, their underlying attitudes toward migration and their current location, and their attitudes toward the West.
Second, we measured only intentions to migrate, and we cannot be sure that those intentions will translate into actual decisions. For instance, past research has shown that although partisans' residential preferences differ, they do not necessarily act on their preferences (e.g., Mummolo & Nall, 2017), and one reason for this is that alternative concerns common to all ideological groups override the impact of ideological fit. We suspect that competing goals also play a role in refugees' migration decisions. Assessment of the relative impact of different needs in migration would be a fruitful direction for future research.
Finally, we present only correlational data, and therefore the precise causal process underlying these relationships is unclear. We were also agnostic as to the relationship between different extreme attitudes. Whereas it is possible to hypothesize causal paths between the various measures of extremism that we included (e.g., endorsement of ideological extremism might lead to an increased willingness to sacrifice for those goals), our goal was simpler: to test the robustness of the main effect over different operationalizations of extremism present in the literature. We hope that the consistent pattern we found will facilitate future research that will disentangle the underlying processes.
These limitations notwithstanding, the pros of this research outweigh the potential cons. Unfortunately, the research on Syrian refugees (and refugees in general) mimics the media coverage-people speak about them rather than allowing them to speak for themselves (Georgiou & Zaborowski, 2017). Thus, the literature on refugees predominantly involves surveys carried out with members of the host country to inquire into their perspective on refugees or analyses of country-level data to predict migration streams (Neumann, 2004). Research conducted with the refugees themselves tends to survey individuals who have already migrated to the West (Ellis et al., 2008) and/or primarily focuses on the traumas that this move entailed (Lustig et al., 2004;Steel et al., 2009). To our knowledge, the present research is the first to survey actual refugees' migration intentions and to anchor these intentions in individuals' social and political attitudes. Making a decision about migration is one of the most impactful moments in people's lives, particularly when one decides to relocate to a culturally distant country. We hope that the present findings contribute to a better understanding of refugees' intentions and in so doing further the scientific study of the fast-growing migration phenomenon.

Transparency
Action Editor: Mark Brandt