Xenophobia and Social Contact in University Students

Due to increasing human mobility in today’s world, relations among groups are becoming more and more important. As people with different cultures come in close contact they begin to influence inter-group attitudes (Barni et al., 2020; Turner et al., 2020). Xenophobia, which can be described as negative attitudes towards migrants, is one of the most serious problems between groups caused by migration (Peterie & Neil, 2020). This study aims to investigate the relationship between social contact and xenophobic attitudes towards Syrian migrants among university students in Turkey. Data were collected from 142 university students through the Xenophobia Scale and the Revised Social Contact Scale. The data obtained were analyzed by multiple linear regression technique. The findings show that while the quality of social contact significantly predicts the xenophobic attitudes of university students, the quantity of social contact does not. The quality of social contact explains 43% of university students’ xenophobic attitudes. It is understood that the quality of social contact is highly effective in reducing xenophobic attitudes towards migrants.


INTRODUCTION
facilitating and hindering the acculturation of international students in Australia revealed that the support network made up of mixed-nationals and especially host locals reduces the acculturation stress and facilitates psychological and sociocultural adaptation. Perceived negative stereotypes loosen the ties with the host culture and also strengthen the ties with ethnic culture. While social resources associated with both cultures were found to be beneficial for acculturation, each independently contributed to the psychological well-being of the participants. Communication with the host locals played a particularly important role in the development of these resources. This shows the effect of host community members on the problems experienced by migrants. One problem that migrants often face is xenophobia. Xenophobic attitudes towards migrants are gradually increasing across the world and educational systems are also influenced by this rise (Organising Bureau of European School Student Unions [OBESSU], 2015). Local university students view international students from different nations a threat to their achievement and group identity. International students' cultural differences, lifestyles, in-class and on-campus behaviors may become the target of xenophobic attitudes by local students (Harrison & Peacock, 2010). It can be argued that because of xenophobic attitudes international students fall behind on their education, fail to achieve linguistic competence and make no progress on their literacy skills. When faced with xenophobic attitudes, migrant students become distanced from educational environment, causing their integration with the society to suffer. The development of linguistic and literacy skills in particular helps migrant students' social inclusion in educational environment. Migrant students have the opportunities in educational institutions to develop basic life skills required for social inclusion and social networking (OBESSU, 2015).

LITERATURE REVIEW
Xenophobia is defined as intense antipathy, hatred and fear of individuals who are perceived as foreigners (Tafira, 2011). Xenophobia, which is a combination of the Greek words "xenos", meaning stranger or foreigner, and "phobos", meaning fear, is used by many organizations such as IOM, ILO, OHCHR and UNHCR to refer to hatred of foreigners (McKinley et al., 2001). Zenophobia has often been associated with migrants (Yakushko, 2009). Those perceived as foreigners are understood to be migrants, refugees, asylum seekers, displaced individuals and non-citizens (Adjai, 2010). Canetti-Nisim et al. (2006) explain that hostile attitudes towards foreigners are defined as xenophobia, as individuals within the group consider those outside the group as a threat to their culture, socioeconomic status, identity and values. In other words, hostile attitudes caused by feelings of fear, hatred and humiliation towards individuals not belonging to the group are considered as xenophobia.
Among the causes of xenophobia, factors such as economic factors, regional migration movements, possible perceived threat against the culture, political imbalances, religious doctrines and terrorism come to the fore (Omoluabi, 2008). On the other hand, the factors influencing xenophobic attitudes are basically discussed as internal and external factors. While genetic structure and personality traits are assessed as internal factors, education and intergroup relationships are assessed as external factors (Wagner, 2017). In the context of intergroup relations, it is stated that dominant groups in the society can approach other groups in a prejudiced way and exhibit discriminatory attitudes (Case et al., 2006). For example, a study examining the relationship and social interaction of university students with foreign students in the UK put forth that local students perceive students with different nationalities as a threat to their academic achievement and group identity, and exhibit xenophobic attitudes towards them (Harrison & Peacock, 2010). Similar results were obtained in another study conducted by Genkova and Grimmelsmann (2020) in Germany. Accordingly, this study determined that there is a positive association between the national identity of the host community members and their xenophobic attitudes and that xenophobia increases the social distance between migrants and foreigners, and members of the host society. Thus, individuals who do not develop an inclusive attitude towards other groups in society socioculturally have more xenophobic tendencies (Litvinova & Tarasov, 2012). As Alrababa'h et al., (2020) stated, prejudice and xenophobia against the outgroup underlie the opposition to migration movements in society.
Possible xenophobic attitudes towards migrants may affect them even more negatively considering they are already in a disadvantaged position. Disadvantaged groups in society struggle to have the same rights as dominant social groups that shape social values as well as possessing financial resources or important social symbols (Sidanius & Pratto, 1999). In this struggle, prejudiced attitudes of individuals in the group not only lead to discrimination but they also play an important role in the breakout of conflict and violence between groups (Çuhadar-Gürkaynak, 2012). The solution to this problem is thought to be dependent on increasing social contact and developing tolerance among groups.
Social contact refers to real face-to-face interaction between members of different groups (Christ & Kauff, 2019). It has been promoted for years to prevent conflict and violence by improving inter-group relationships (Dovidio et al., 2003). Social contact, which is defined as the exact opposite of social distance, reduces bias and discrimination between groups (Barni et al., 2020;Coban, 2020;Christ & Kauff, 2019;Çuhadar-Gürkaynak, 2012;de Coninck et al., 2020). On the other hand, it increases mutual trust, solidarity and forgiveness (Coban, 2020;Christ & Kauff, 2019). While social contact reduces prejudice and discrimination, Beller's (2020) study revealed that xenophobia generalizes prejudice and increases discrimination. The aforementioned study determined that xenophobia also affects well-being and health negatively and increases violence and aggression. Considering that xenophobia is defined as hostile attitudes towards migrants, a negative relationship is expected to exist between social contact and xenophobia. In other words, it is predicted that individuals' xenophobia levels can be lowered by intensifying social contact.
Reducing xenophobia through social contact also improves migrant students' education experiences. On the other hand, education can help increase social contact and reduce local students' xenophobic attitudes. When migrant students develop their linguistic skills through schooling, they find it easier to establish social contact with local students. Linguistic obstacles lead to limitations, shortage of sufficient information on the new culture and various problems in daily life (Björn, 2013;Kim, 2016). Literacy skills have a considerable impact on the development of migrant students' linguistic competence. Literacy refers to the ability to read, write and speak competently in the host society's language in order to communicate, make decisions and solve problems in the family, at workplace and in public (Wilson, 2002). As their literacy skills improves, migrant students can cope with problems effectively, stand up for their rights, become integrated into the society and develop interpersonal communication skills (Larrotta & Chung, 2020). Improved interpersonal communication skills naturally intensify social contact and thus reducing local students' xenophobic attitudes.
Xenophobic attitudes of individuals in the host society tend to be influenced by the convergence of different groups as a consequence of migration. Whether group members consider their group as superior or not and the quantity and quality of social contact with other groups are presumed to have a bearing on individuals' xenophobic attitudes. While a search of the literature revealed a limited number of studies on the relationship between social contact and xenophobic attitudes (e.g. Brown et al., 2001;DiGiusto & Jolly, 2009;Jolly & DiGiusto 2014;Ommunds et al., 2013), only one research study (Padır, 2019) was found to have been conducted in the context of Turkey. Therefore, this study aims to investigate the relationship between xenophobia expressed as negative attitudes of members of the host society towards migrants and social contact, considering the sizeable migrant population in Turkey. The study also attempts to produce meaningful findings regarding the steps to be taken to engender positive relations between the host society and the migrant community. In addition, the Social Contact Scale used to collect data for this study was revised and its validity and reliability analysis was performed.

Research Design
This study is designed to explain the relationship between university students' xenophobic attitudes and social contact. In this regard, it is an explanatory correlational research, one of the types of quantitative methods. Explanatory correlational research aims to determine the degree and the direction of the relationship between two or more variables (Fraenkel et al., 2011).

Study Group
The study group was determined by convenience sampling method. In this method, in accordance with the objective of the research, accessible groups are selected from which comprehensive data can be obtained (Fraenkel et al., 2011). In this context, the researcher collected face-to-face data from Turkish university students who had Syrian migrant students in their classes through group practices. The study group of the research consists of 142 university students, 97 (68%) females and 45 (32%) males. The age range of the participants is 18-25 and the average age is 20.33. The participants are students at a faculty of education and 47% of them are in first year, 13% in second year, and 40% are in third year of their study. The reason the study group was composed of university students was that the university students represent a unique demographic structure that can provide inferences about future trends. It is believed that this group, which will affect different areas of the society in their future professional lives, will also be determinant in attitudes towards migrants in the future.

Xenophobia Scale (XS)
The Xenophobia Scale is a five-point Likert-type scale developed by Bozdağ and Kocatürk (2017) to measure individuals' attitudes towards migrants. The scale developed for this purpose consists of 18 items. Validity and reliability studies for the scale were conducted with data collected from 537 participants between the ages of 18-25. Explanatory factor analysis (AFA) revealed that the scale is made up of three factors. These are hate, fear and humiliation. It was observed that the first factor (hate) explained 44.47% of the variance and that the total variance explained was 55.47%. Confirmatory factor analysis (CFA) was performed to determine whether the three-factor structure obtained by the explanatory factor analysis was confirmed. The obtained values showed that the structure of the three-factor scale has acceptable and valid results. Cronbach's alpha value of the scale was found to be .87. In this study too, Cronbach's alpha value of the scale was calculated as .87.
Each item in the scale is scored as 1 point for the "Strongly Disagree" option, 2 points for the "Disagree" op-tion, 3 points for the "Neutral" option, 4 points for the "Agree" option, and 5 points for the "Strongly agree" option. Two of the items (7 th and 11 th items) require reverse scoring. The lowest score that can be obtained from the scale is 18 while the highest score is 90, and the higher is the score, the higher is the level of xenophobia and vice versa.

Social Contact Scale (SCS)
The Social Contact Scale was developed by Islam and Hewstone (1993) to measure the social contact between Muslims and Hindus living in India. While adapting to the Turkish language by Akbaş (2010), the expressions in the scale were translated as Alevis-Sunnis in accordance with the purpose of the researcher's study. In this research, the expression "Syrian migrants" was used instead of "Alevis-Sunnis".
The Social Contact Scale consists of two subscales to measure the quantitative and qualitative aspects of social contact. The Quantity of Social Contact subscale aims to measure the frequency of social contact between two groups whereas the Quality of Social Contact subscale contains questions regarding the quality of social contact between two groups. Each subscale contains five questions. The sixth item in the scale is calculated by reverse coding. In the Quantity of Social Contact subscale, each item is evaluated with a Likerttype five-point rating ranging from "Never" (1) to "Always" (5). A high score indicates more frequent social contact with the other group. In the Quality of Social Contact subscale, on the other hand, each item is evaluated according to the question with a Likert-type five-point rating (For example, "Definitely not equal" (1) and "Definitely equal" (5)). A high score obtained from the subscale indicates a more quality social contact. The total variance explained by the Social Contact Scale is 62.2%, and the Cronbach Alpha values are calculated as .83 for each subscale. In this study, since the scale was revised, validity and reliability analyses were performed. The results regarding the validity and reliability analysis of the Revised Social Contact Scale (SCS-R) are provided below.
Validity and reliability analysis results for SCS-R. CFA was conducted for the validity of the SCS-R. The values suggested by Schermelleh-Engel et al. (2003) (2<X 2 /df ≤ 3, .05 < RMSEA ≤ .08, .05 < SRMR ≤.10, .85 ≤ AGFI <.90, .90 ≤GFI <.95, .95 ≤CFI <.97, .90 ≤NFI <.95, .95 ≤NNFI <.97) were taken as reference for the model fit. As a result of the CFA in which the two-factor structure was tested, it was determined that the model (model 1) did not have acceptable fit. Therefore, modification indices were checked and the model was retested after error covariance was inserted between the items 2 and 3 as well as 2 and 5 as they measure the same construct and belong to the same factor. The resulting analysis was found to have acceptable fit. The results for the models are given in Table 1 and Table 2.
In Table 2, it can be seen that the standardized factor loads obtained for SCS-R are between .29 and .83 for the quantity of social contact subscale and between .31 and .87 for the quality of social contact subscale. The R 2 values were examined to understand to what extent the items explain the variances in the relevant factor. The R 2 values range from .08 to .69 for the quantity of social contact subscale and from .10 to .80 for the quality of social contact subscale. The z values of the items were determined between 3.13 and 8.59 for the quantity of social contact, and 3.57 and 3.67 for the quality of social contact (p <.01). The correlation coefficient between latent variables (quantity of social contact and quality of social contact) was specified as .42. This shows that quantity of social contact and quality of social contact are moderately positively correlate with each other.
Convergent validity of SCS-R. The Average Variance Extracted (AVE) and Composite Reliability (CR) values for the convergent validity of the SCS-R were examined. To ensure convergent validity, AVE value should be larger than .50, CR value larger than .70 and AVE value lower than CR value (Fornell & Larcker, 1981). According to the results, AVE values were determined as .41 for the quantity of social contact subscale and .57 for the quality of social contact subscale while CR values were determined as .75 for the quantity of social contact subscale and .86 for the quality of social contact subscale. Since the AVE value for the quantity of social contact subscale was .41 (<.50), the results showed that convergent validity was partially met.
Divergent validity of SCS-R. For divergent validity, AVE value must be bigger than Shared Variance (SV) value (Hair et al. 2014). On the other hand, Fornell and Larcker (1981) state that the square roots of AVE value should be bigger than the correlation coefficient between the latent variables for the divergent validity. The results show that AVE values of two subscales were bigger than the SV value (.18). In addition, the square roots of the AVE values of the subscales (.64 for the quantity of social contact subscale, and .75 for the quality of social contact subscale) were bigger than the correlation coefficient between the latent variables (.42). Therefore, the divergent validity of the scale was ensured.
Reliability of SCS-R. Cronbach's Alpha coefficient was calculated for the reliability of SCS-R. The Cronbach's Alpha coefficients was found to be .75 for the quantity of social contact subscale. It was calculated .84 for the quality of social contact subscale.

Procedure and Data Analysis
During the data collection process, the participants were informed about the rationale of the research and the significance of providing sincere responses was explained to them. The data collection process took approximately two weeks to complete. After the data collection process was completed, the data were classified and transferred to SPSS for analysis. During data classification, missing values (10) were removed from the data set. Then, statistical analysis was performed using SPSS 25 and AMOS 24. Multiple linear regression analysis was performed to determine the relationship between xenophobia and social contact, and confirmatory factor analysis for the validity analysis of SCS-R. The key assumptions are primarily examined for the analysis. Accordingly, sample size, univariate and multivariate outliers, normality, linearity, homoscedasticity, multicollinearity and independence of errors assumptions were checked (Field, 2013;Hair et al., 2014;Tabachnick & Fidell, 2013). Eight outliers were specified and removed from the data set. The analysis resumed with the remaining 142 data items. The sample size was considered to be sufficient based on the criterion of (n ≥ 50 + 8m [number of independent variables in m]) (Tabachnick & Fidell, 2013). It was observed that the scatter plots of the residuals met the assumptions of normality, linearity and homoscedasticity. For multicollinearity assumption, it was verified that the correlation coefficient between variables is less than .80 (Field, 2013), VIF (Variance Inflation Factor) is less than 10 and TV (Tolerance Value) is greater than .10 (Hair et al., 2014). The assumption of independence of errors was met by calculating the Durbin-Watson value as 1.82 (Field, 2013). Mean and standard deviation values of variables, correlation coefficients between variables, TV, VIF, skewness and kurtosis values are provided in Table 3.

RESULTS
When the mean and standard deviation values of the measurement tools used in the study are examined it is seen that ( Table 3) the mean score for xenophobia is 44.90 (12.65), for the quantity of social contact 8.54 (3.36) and for the quality of social contact 16.23 (5.00). There is a significant weak negative correlation between university students' xenophobic attitudes and the quantity of social contact (r = -.19, p <.05) while a significant moderate negative correlation between university students' xenophobic attitudes and the quality of social contact (r = -.66, p <.01). In addition, the items of the Likert scales used in the study have been given in a table containing the percentages of the responses (Appendix A).
Multiple linear regression analysis was conducted to determine the contribution of the quantity of social contact and the quality of social contact in explaining xenophobic attitudes of university students. The results are given in Table 4 below.
As seen in Table 4, the quantity of social contact does not significantly predict xenophobic attitudes of university students (β = 0.01, t (139) = 0.19, p > .05). However, the quality of social contact (β = -0.66, t (139) = -9.87, p < .01) is found to explain university students' xenophobic attitudes by 43% (R 2 = .43, F (2, 139) = 53.10, p = .000). The quality of social contact of university students significantly negatively predict their xenophobic attitudes. In other words, as university students' quality of social contact with Syrian migrant intensifies, their xenophobic attitudes diminish.

DISCUSSION
In this research, the Social Contact Scale, which was developed by Islam and Hewstone (1993) and adapted to the Turkish language by Akbaş (2010), was revised and its validity and reliability analyses were conducted to find out the social contact of university students with Syrian migrants. The results confirmed that the two-factor structure of the SCS-R has acceptable fit. Furthermore, convergent and divergent validity of SCS-R was ensured. Cronbach's Alpha coefficient was calculated for reliability analysis of SCS-R. Cronbach's Alpha coefficient, which was calculated as .75 for the quantity of social contact and .84 for the quality of social contact, revealed that the scale is sufficiently reliable. Therefore, it can be argued that SCS-R meets the requirements to be used to determine the social contacts of individuals with Syrian migrants.
The analysis results of the relationship between university students' xenophobic attitudes and the quantity and the quality of social contact revealed a weak negative correlation between xenophobic attitudes and the quantity of social contact, and a moderate negative correlation between xenophobic attitudes and the quality of social contact. It was found that among the variables only the quality of social contact significantly predicts the xenophobic attitudes of university students. As a result, xenophobic attitudes in university students were found to diminish as the intensity of the quality of social contact grew.
Lack of social contact among social groups causes prejudice, differentiation and increased social distance (Ata et al.,  Barlow et al., 2012). This, in turn, drives discrimination (Beller, 2020). However, with the establishment of social contact, faulty generalizations underlying the prejudices are noticed and similarities as well as differences are discovered (Kotzur et al., 2019). Thus, it is realized that the other group is not as bad and homogeneous as it is thought (Çuhadar-Gürkaynak, 2012). This study too confirmed that establishing social contact (the quantity of social contact) with migrants can reduce xenophobic attitudes, albeit at a low level. Consistent with this finding, there are certain studies that show there is a negative relationship between the size of the migrant population and xenophobic attitudes of host society (e.g. DiGiusto & Jolly, 2009;Ha, 2010;Jolly & DiGiusto, 2014;Wagner et al., 2006). According to Ommundsen et al.'s (2013) study, as the quantity of social contact increases, fear-based xenophobic attitudes decrease. The more important issue is the quality of social contact (Ahmed, 2017;Dirksmeier, 2014). When a quality social contact is established, there is a significant decline in xenophobic attitudes. A study that investigated the xenophobic attitudes towards Syrian refugees in Turkey found that the quality of social contact lowers xenophobic attitudes. However, according to the same study, unlike the current research, the quantity of social contact was found to raise xenophobic attitudes (Padir, 2019). According to Abrams (2010), the quality social contact enables the development of positive attitudes among groups. The results of a study by Brown et al. (2001) concluded that the quality of social contact is more effective than the frequency of contact in relations among groups. In another study conducted with university students, it was argued that anxiety about being socially desirable increases xenophobic attitudes (Haque, 2015). In other words, individuals may display negative attitudes towards individuals from other groups in order to be accepted by their own groups and to become popular. Jolly and DiGiusto (2014) also concluded that individuals display more xenophobic attitudes towards those outside their group. Foreign university students are excluded and maltreated by other students. Cliques are formed to exclude foreign students (Singh, 2013). On the other hand, exclusionary attitudes have been found to lose intensity among those who have friends from the minority group (McLaren, 2003). In a conducted by Şen (2014), it was observed that the quality of social contact also lowers discriminative treatment of ingroup members. Therefore, it can be concluded that the results of earlier research are consistent with the results of this study in general. The quality of social contact is considered as an important factor in reducing negative attitudes among groups. The reduction in xenophobic attitudes with the heightened quality of social contact also positively influences educational experiences of migrant students at university. Linguistic competence and literacy skills of migrant students can increase social contact and thus contributing to the lessening of xenophobic attitudes. According to Larrotta and Chung (2020), developing linguistic skills of migrant students and to this end, offering letter writing opportunities aid interpersonal communication competence. As their linguistic skills improve, migrants interact more frequently with members of the host society, establish more contacts among them and consequently encounter fewer sociocultural problems (Feliciano, 2001;Yoshida, 2015). Migrant students gain access to more opportunities and resources when they learn the language of the host society. This enables them to fully integrate into social life (Bacquet, 2020). The impact of linguistic competence and literacy skills in reducing xenophobic attitudes of local students towards migrant students should not be ignored.
Another factor influential in intensifying quality social contact in order to lessen xenophobic attitudes towards migrant students is multicultural education. Multicultural education helps students develop positive attitudes towards individuals from different cultural groups, reduces discriminatory attitudes and raises tolerance among groups (Köşker & Erdoğan, 2020). In an educational system based on a single culture that denies existent diversity in social structure, individuals may develop negative and marginalizing attitudes towards different cultural groups (Parekh, 2000). In multicultural education system, however, reducing xenophobic attitudes becomes possible by understanding people from different nations, respecting opposing views and empathizing with people with different lifestyles (Boehnke et al., 1998). Therefore, it is essential that educational institutions adopt a policy based on multiculturalism and accept differences as diversity.

CONCLUSION
It is vital that precautions are taken to lower the marginalization of and prejudice against migrants, refugees or asylum seekers arriving in Turkey for various reasons. There have been warnings against possible rise in concrete negative incidents where foreigners such as migrants, refugees or asylum seekers are labelled as the enemy and an entity that precautions need to be taken against (Ünal, 2014). At this point, the findings of the current study suggest increasing the quality of social contact between Syrian migrant university students in Turkey and Turkish university students. This should certainly not be limited to university students, and quality social contact should be maintained between individuals who are perceived as foreigners in different parts of society and members of the host society. Many studies in the literature provide evidence that social contact is effective in precluding discrimination and prejudice (e.g. Barni et al., 2020;Coban, 2020;Christ & Kauff, 2019;de Coninck et al., 2020). This study also shows that quality social contact is highly effective in preventing the negative impact of xenophobia, which is a more radical exclusionary attitude than discrimination and prejudice. The four basic conditions stated by Allport (1954) for social contact between groups to be effective were confirmed in many studies (e.g. Pettigrew, 1998;Pettigrew & Tropp, 2006). Accordingly, the fact that the two groups have equal status, have common goals, cooperate in line with these goals, and support institutional and structural positive relationships ensure effective social contact. On the other hand, it is stated that meeting these four conditions strengthens the effects of social contact, however, even if these conditions are not present, social contact yields effective results (Everett and Onu, 2013;Pettigrew & Tropp, 2006). In addition to direct contact, indirect contact (e.g. extended, imaginary) is also reported to produce effective results (Christ & Kauff, 2019;de Coninck et al., 2020). Therefore, it is important to increase the direct and indirect social contact of university students with migrant students. In this context, various group works can be organized in universities in order to increase the interaction between migrant students and students from the host community. Through these group works, cooperation can be improved in line with common goals. Teaching faculty members at universities can make an effort to create a positive interaction environment between migrant students in their classes and students from the host community. Host community students' awareness about the lifestyles and cultures of migrant students can be raised through print and visual media channels.
There are certain limitations of the current research. It can be considered as a limitation that the participants included only university students and that the research is a cross-sectional study. Further studies can be conducted by collecting data from different segments of society in order to explore xenophobic attitudes in a broader context. It is important to research attitudes through longitudinal studies in order to corroborate findings and reveal more valid results. Using convenience sampling method in the study can be considered as a limitation in terms of the generalizability of the study results. The variables predicting xenophobic attitudes in the current study are limited to the quantity of social contact and the quality of social contact. Studies in which more variables are used to predict xenophobic attitudes of individuals in the host society are recommended. These variables can be at the individual level or in conjunction with those at the environmental level. Finally, mixed method studies that integrate quantitative and qualitative data analyses are suggested. Quantitative data analysis may be employed to find out what variables predict xenophobic attitudes of individuals in the host society and qualitative data analysis to determine what kind of xenophobic attitudes migrants face. Various model studies can be conducted based on the multi-factor variables that affect xenophobic attitudes. Thus, a holistic perspective on the psychosocial effects of xenophobia can be achieved.