SOCIAL DISTANCE IN IRAQ AND LEBANON By Jihan A. Mohammed A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Sociology-Master of Arts 2017 ABSTRACT SOCIAL DISTANCE IN IRAQ AND LEBANON By Jihan A. Mohammed Research has found that tensions, conflicts, and wars worsen the views groups hold towards each other and, plausibly, increase social distance (Parks 1924; Bogardus 1925; Owen et al. 1981; Siber 1997; Parrillo and Donoghue 2005; Oswald 2005; Strabac 2016). Since the twentieth century the Middle East is caught up in interlocking pattern of crises, conflicts, wars, and terrorism. Almost every country in the region have serious problems in social and political stability. The focus of this study is to investigate levels of social distance in conflict areas like Iraq and Lebanon where thousands of people have been killed and displaced because of civils wars and counter-terrorism. We use Arab Democracy Barometer Survey data, Wave II, a nationally representative data on ten Middle Eastern countries collected between 2010-2011, to investigate the determinants of social distance. We are specifically interested to see the relative importance of: practicing religious rituals i.e. praying, fasting, attending Friday prayer/Sunday service; sectarian/denominational identification; and voting behavior as determinants of social distance. Our findings indicate that Iraqis have higher levels of social distance than the Lebanese. Further, controlling for age, gender, education, and geographical area (urban vs rural) we found a negative association between practicing religious rituals and social distance. Finally, we found a negative association between in-group identification, in-group favoritism, and the outcome variable. This thesis is dedicated to my mother, my father, and my beloved husband. Thank you for your endless support. iii TABLE OF CONTENTS LIST OF TABLES v INTRODUCTION 1 BACKGROUND 2 THE IRAQI AND LEBANESE CONTEXT Iraq: population groups by religion and ethnicity Civil wars and conflicts Lebanon: population groups by religious denominations Civil wars and conflicts 6 6 6 8 8 THEORETICAL NOTIONS AND HYPOTHESES 10 DATA AND METHODS Dependent Variable Independent Variables Control Variables 13 13 13 15 ANALYSES 16 FINDINGS 17 DISCUSSION 23 BIBLIOGRAPHY 26 iv LIST OF TABLES Table 1: Descriptive Data of Variables of Interest (Mean) 18 Table 2: OLS Regressions 19 Table 3: OLS Regressions/Lebanon 20 Table 4: OLS Regressions/Iraq 21 v INTRODUCTION Research has found that tensions, conflicts, and wars worsen the views groups hold towards each other and, plausibly, increase social distance (Park, 1923; Bogardus, 1925; Owen et al., 1981; Siber, 1998; Parrillo and Donoghue, 2005; Oswald, 2005; Strabac, 2016). For over a century the Middle East has been caught up in interlocking pattern of crises, conflicts, wars, and terrorism. Almost every country in the region has serious problems in social change and political stability. The focus of this study is to investigate levels of social distance in conflict areas like Iraq and Lebanon where thousands of people have been killed and displaced because of civils wars and counter-terrorism. We use Arab Democracy Barometer Survey data, Wave II, a nationally representative data on ten Middle Eastern countries collected between 2010-2011, to investigate the determinants of social distance. We are specifically interested to see the relative importance of: practicing religious rituals i.e. praying, fasting, attending Friday prayer/Sunday service; sectarian/denominational identification; and voting behavior as determinants of social distance. Given the fact that Iraq is still in the middle of chaos and Lebanon is in the process of slow recovery, we expect to find higher social distance in Iraq than in Lebanon. Further, we argue that all major religions promote benevolence and caring towards out-groups. Therefore, controlling for age, gender, education, and geographical area (urban vs rural) we expect to find a negative association between practicing religious rituals and social distance. Finally, the idea that conflicts heighten in-group identification (Sherif and Sherif, 1979; Bobo, 1996) and in-group favoritism (Bratton and Kimenyi, 2008), we expect the relationship between each of these variables and social distance to be positive. 1 BACKGROUND In 1924 the pioneering figure in social distance, Emory S. Bogardus, developed a Social Distance Scale to measure prejudice among ethnic and racial groups in America. In his book published in 1922, A History of Social Thought, Bogardus expressed his concern with what he referred to as “the race problem,” which he acknowledged to be one of the major social dilemmas confronting America (Owen et al., 1981, p. 80). During that period a large number of immigrants were coming into the country. In the one hand, the “older” Americans or the so called the “first wave” rejected immigrants and tried to stop them from entering the country. In the other hand, government officials signed agreements and enacted laws to slow down and eventually stop immigration. For example, in 1907-1908, president Theodore Roosevelt made “Gentlemen’s Agreements” with Japan to stop Japanese from entering the country and in 1913 California passed “Alien Land Law” which prohibited Chinese, Japanese, Korean, and Asian Indians from purchasing and leasing land. These agreements and many other enacted laws enhanced segregation and heightened prejudice and racism towards immigrants from different backgrounds (Wark and Galliher, 2007). Bogardus’ concern with ethnic and racial issues pushed him to develop a quantitative indicator of social distance. He defined social distance as” the degrees and grades of understanding and feeling that persons experience regarding each other. It explains the nature of a great deal of their interaction. It charts the character of social relations” (Bogardus, 1925, p. 299). For Bogardus, the social distance is a function of the social contact of the person with another person or with a person of another group in different social settings i.e., social contact within the family; within social and fraternal groups; within neighborhoods; within occupational groups; and within political or national groups (Bogardus, 1925). He argued that the weaker the social contact the higher, presumably, the social distance and prejudice. The original scale consisted of the following seven statements, all of which express the willingness to admit the person or the groups 2 considered: 1- to close kinship by marriage; 2- to one’s club as personal chums; 3- to one’s street as neighbors; 4- to employment in one’s occupation; 5- citizenship in one’s country; 6- as visitors only to one’s country; and 7- to exclude from one’s country (Bogardus 1925). Miller (1991) states that the scale may be used “to estimate the amount of potential and real conflict existing between any cultural groups, anywhere in industrial, political, racial, religious, and other phases of life” (p. 482). Between 1920 and 1977 Bogardus’ scale was used five times to trace the evolution of ethnicracial relations in America and it has been used commonly by other researchers to measure social distance in different contexts and with a variety of groups (see Wark and Galliher 2007). As a matter of fact, the General Social Survey (GSS), until today, have kept using two of Bogardus’ questions “would you have as next neighbors?” and “would you marry into group?” In this study, we will analyze the social distance in Iraq and Lebanon using the question about desirability of having different group members as neighbors using Arab Barometer, Wave II, collected between 2011-2012. Previous social distance studies (Triandis and Triandis, 1960; Parrillo and Donoghue 2013) have found that geographical area (urban vs rural) and education are significant indicators of social distance. Bogardus noted that in rural areas there was much more spatial distance between people as compared to urban areas, hence, levels of social distance in rural areas were much higher. Other studies (Borgadus, 1933; Melikian and Prothro, 1952; Hunt, 1956; Kirsch, 1957) have found that race, religion, and nationality are common determinants of social distance. Generally speaking, people tend to have low social distance with those of the same race, ethnicity, and religion because people feel more comfortable being around those similar to them. For example, in America, as well as in Europe, White people are more inclined to maintain social distance with Blacks and 3 Latinos than with Whites themselves (Owen et al., 1981, Duckitt, 1922)1. Having said that, in multi-ethnic and multi-religious states, intergroup hostilities play a significant role in determining social distance (Starr, 1978; Meade and Singh, 1973; Šiber, 1998). Meade and Singh (1973) found that the India/Pakistan war in 1971 increased social distance between Hindus and Muslims and Hindus and Pakistanis. Abanes et al. (2014) found that the long-running armed conflict between Christians and Muslims in the Philippines resulted in the salience of religious identification, which in turn, resulted in high levels of social distance between the two groups. In China, however, Guo and Tynen (2015) found that among the Han, the Hui, and the Tibetans, the latter have the highest levels of social distance. Authors argue that it could be explained by Tibetans’ “experience of state-inflicted violence and social and cultural exclusion as a historically marginalized minority on the periphery” (p. 706). To sum, in conflict areas, religious/denominational identification is inversely related to social distance (Johnson 1977; Šiber 1998; Poppe and Hagendoorn, 2004; Abanes et al. 2014; Bloom et al. 2015). Further, studies (Horowitz, 1985; Terkildsen, 1993; Lijphart, 1999; Snyder, 2000; Mendelberg’s, 2001; Greene,2004; Dickson and Scheve, 2006; Ferree, 2006; Bratton and Kimenyi, 2008) have found that conflicts increase the likelihood to vote for a candidate who belongs to voter’s family/tribe, ethnicity, race, sect. In social identity theory this voting behavior is known as in-group favoritism and is found to be significantly associated with social distance. Turner and Tajfel (1979) defines in-group favoritism as “any tendency to favor the in-group over the out-group, in behavior, attitudes, preferences or perception” (p. 187). Given the history of conflicts in Iraq and Lebanon we are interested to examine if in-group favoritism is associated with social distance. Although regional and civil conflicts are found to heighten ethnic, religious, and 1 For a critique of the patterns of social distance in America see (Weaver, 2008). 4 sectarian identities (Dawisha, 2013; Kirmanj, 2013), no study so far has been done to examine social distance in Iraq and only one study was done in Lebanon in 1952 by Prothro and Melikian. Therefore, it is difficult to know enough about the evolution of social distance in these two countries and/or predict an outcome. Nevertheless, the two countries are seriously effected by the cumulative impact of a long series of regional and civil conflicts, hence, we expect that in-group favoritism is inversely related to social distance. 5 THE IRAQI AND LEBANESE CONTEXT The reason we choose Iraq and Lebanon is because their history and social structure are very similar. In the early twentieth century, Iraq was a British colony and Lebanon was a French colony. While an Iraqi monarchy was established in 1921 under the role of Britain, a French Mandate of Lebanon was established around the same time. Further, Iraq gained independence from Britain in 1932 and Lebanon from France in 1943. It is beyond the purpose of this paper to discuss the socio-historical background of conflicts and civil wars in each of the two countries. Nevertheless, below we briefly touch upon demographic composition and key points of conflicts. Iraq: population groups by religion and ethnicity According to the Central Intelligence Agency website (CIA), the Iraqi population is about 38 million. It is composed of various ethno-religious groups which are heavily concentrated in specific geographical areas. Ethnically, Iraq is divided between Arabs and Kurds. Arabs make up to 77% of the population and are concentrated in middle and southern areas of the country, whereas, the Kurds make about 20% and are heavily concentrated in north Iraq. Religiously, Muslims make up 97% and about 3% of the population belong to other small religious minorities concentrated in north west in the city of Mosul i.e., Chaldo-Assyrians, Armenians, Yezidis, Turkemn, Mandaean, and Jews (Kirmanj, 2013). Finally, the country is divided between the two major sects of Islam; Sunni and Shia. In Iraq, 75% of the Muslim population are Shia and 25% are Sunni. The Sunnis are concentrated in the north and middle eastern areas of the country, whereas, the Shias are concentrated in middle and south western areas. Civil wars and conflicts Since the establishment of Iraq in 1932, the three major population groups, the Kurds, Arab Sunnis and Shias, have been at odds with each other because of power, oil, territory, and identity. Historians and political scientists (Dodge, 2003; O’Leary, 2003; Cole, 2004; Fontan, 2009; Zeidel, 6 2010; Natali, 2016) provide detailed analyses on the origins and the evolutions of conflicts and civil wars in Iraq. They all agree that Iraq is still struggling to bring together all Iraqis under one umbrella. Scholars also agree that the reason Iraq has been through many wars and conflicts is because of the frailty of the state to consolidate the Iraqi national identity due to external interventions in Iraq’s internal issues (Cole, 2004; Dawisha, 2013; Kirmanj, 2013). As an outcome, ethnic and sectarian identities have become salient. Arab Sunnis and Shia fought each other throughout the history of Iraq over “who is the true Iraqi?” Arab Shias are accused of being Iranian sympathizers, hence, their Iraqiness is perceived suspiciously (Dawisha, 2013; Kirmanj, 2013; Haddad, 2014). The oppression of Arab Shias by Arab Sunnis led to many civil wars i.e., the 1991 Arab Shias revolution against the Iraqi government. The revolution failed and thousands of Shias were killed. However, after the American invasion in 2003, Arab Shias dominated the political order and they have been governing the country since then. Shortly after 2003, Arab Sunni insurgent groups started to rise up against the Shia government (Al-Marashi, 2005; Chaplin, 2006; Haddad, 2014; Kuznetsov, 2015). These conflicts widened the social, cultural, and political distance in Iraq (Dawisha, 2013), particularly, after the American invasion in 2003. Between 2006 and 2007, Arab Sunnis and Shias’ sectarian war has led to the killing of 23,000 innocent civilians (Kirmanj, 2013), and internally displaced over 3 million people (Dawisha, 2013). Additionally, according to the United Nations Office for the Coordination of Humanitarian Affairs (OCHA), 3 million more civilians have been internally displaced since 2014 because of ISIS’s control of some parts of Iraq. Furthermore, Arab Sunnis and the Kurds have been in conflict because the Kurds demand secession from Iraq to establish their own Kurdish State (Kirmanj, 2013). Persecutions against the Kurds reached its worst levels in 1970s and 1980s when the Iraqi regime mass executed thousands of civilians and used chemical weapons to eliminate the entire population of the city of Halabja and other areas near the Iranian borders. At the present time intergroup relations in Iraq 7 are, yet, very tense. Lebanon: population groups by religious denominations According to Central Intelligence Agency website (CIA), as of July, 2016, the population od Lebanon is more than 6 million. About 27% of the Lebanese population are Shias and about 27% are Sunnis (Viret, 2015). However, Christians are divided into different small denominations. About 21% of Lebanese are Christian Maronite, 8% are Greek Orthodox, 6% are Greek Catholic, and other 6% are Druze (Viret, 2015). Other smaller Christian denominations compose 5% of the population (Viret, 2015). Civil wars and conflicts The three main group populations in Lebanon are Christians, Arab Sunnis and Shias (Moaddel, 2012). The conflict among these groups has been over the political power and equal representation in the government (Weiss, 2009). Until 1920, Lebanon’s territory included Mount Lebanon only. At the time, the Christian group, Maronites, claimed political supremacy over other groups due to their wealth and numerical majority. Christians composed 80% of Mount Lebanon’s population, estimated of 400,000. After the formation of the Greater Lebanon in 1920 and the independence from France in 1943, the Lebanese territories expanded, with it the proportion of Muslims increased. Further, the influx of refugees from Palestine to Lebanon in 1948, and the Syrian military occupation from 1975 to 2005, heightened religious, sectarian, and political tensions over equal representation in the government. Tensions broke up into a 15 years (19751990) civil war that resulted in more than 100 thousand fatalities and, according to Internal Displacement Monitoring Center (IDMC), another 800 thousand people were displaced. Moreover, tensions between Sunnis and Shias, on the one hand, and between Shias and Israel on the other hand, brought more violence and instability to the country. By the end of the 2000s Lebanon started to recover despite the many challenges i.e., terrorism and the impact of the civil 8 war in the neighbor country, Syria. 9 THEORETICAL NOTIONS AND HYPOTHESES This study derives its explanation from the theories of Realistic Group Conflict and Social Identity. The Realistic Group Conflict Theory (RGCT) focuses on explaining social distance in contexts of conflicts and intergroup hostility. The core concept of RGCT is that intergroup relations are significantly effected by factors: a. economic dispute or competition over scarce resources; b. clash of interests and goals; c. the presence of real or imagined threats to the safety of the group; and d. political advantage and military considerations (Sherif and Sherif ,1979; Bobo 1983; Kinder and Sears, 1981). Any of these factors may cause an escalation of intergroup hostility and social distance (Sherif et al., 1961; Sherif and Sherif, 1979; Bobo, 1996). Relevant to our study are competition over land, resources, and unequal distribution of power. These factors create feelings of “frustration and perceptions of relative deprivation” (Jackson, 1993, p. 404) which lead to hate and aggression towards the out-group. As we discussed above, intergroup conflicts in Iraq are over power, oil, and land. Arab Sunnis and Shia and the Kurds fight to gain equal power and recognition in Baghdad. Further, these groups have a historical dispute over rich oil areas. The conflict over oil is not because of its scarcity but because of its uneven distribution. RGCT states that groups get involved in conflicts when they attempt to access scarce resources or when resources are abundant but are unevenly distributed. The Kurdish-Arab dispute over the rich oil city of Kirkuk is a relevant example. The Kurds control some parts of the city and the Arabs controls some other parts. The dispute has many times lead to high tensions and clashes. In the case of Lebanon, the reinforcement of political power along confessional lines, as well as, the Syrian and Israeli occupation of the Lebanese territories, heightened civil and sectarian conflicts. Given this discussion, we would expect to find relatively high levels of social distance towards out-groups in both countries. However, the idea that levels of social distance is determined by the intensity of competition and conflicts (RGCT), we assume that, unlike Iraq, Lebanon is recovering 10 from war and intergroup relations are in the process of slow normalization. Hence, our first proposition is: H1: Lebanese are less likely to maintain social distance towards out-groups than the Iraqis. Religiosity is one of the most complex constructs in social science. One limitation of the previous studies on social distance, some of which are reviewed above, is that they looked at the effect of only one dimension of religiosity on social distance. They examined the effect of religious/denominational identification and generalized that religiosity increases social distance. We second the argument that religiosity is multi-dimensional (Scheepers et al., 2002; Vaos, 2007; Smidt et al., 2009; Saroglou, 2011; Bloom et al., 2015) and each dimension may have a different impact on an individual’s social, psychological, and political attitudes (Brinkerhoff and Mackie, 1986; Scheepers et al., 2002; Shen et al., 2013; Bloom et al, 2015). Therefore, in this study we will measure two dimensions. The first one is the ritualistic dimension (Glock and Stark, 1965) which consists of “the performance of any rites or sacred acts designed for communion with the Divine” (Kupke, 1971, p. 8-9) i.e., fasting, public church worship, prayers, and meditation. We argue that all major religions i.e. Islam and Christianity, promote benevolence, caring, solidarity, and compassion towards out-groups (Bloom et al, 2015). We also argue that the rituals that we will analyze in this study emphasize these beliefs. The rituals are praying, fasting, watching or listening to religious programs, attending religious lessons in mosques or churches, attending Friday prayer/Sunday service, and reading religious books. In fact, studies (Allport and Ross, 1967; Brinkerhoff and Mackie, 1986; Steensland et al., 2000) that analyzed the relationship between the frequency of attending the church and social distance have found that high frequency of church attendance is associated with low social distance. The second dimension is religious/denominational identification. Social Identity Theory (SIT) posits that a sense of the self is derived from identification with a group i.e., ethnic, racial or religious and in case of threat or 11 conflict, in-group identification increases (Tajfel, 1981). We expect our findings to be consistent with the literature reviewed above and the idea that sectarian/denominational identification is inversely related with social distance. Having said that, it is important to mention that the question on religious/denominational identification was not asked in Lebanon, therefore, the first dimension will be measured in both countries, but the second dimension will be measured in Iraq only. Given the discussion above on religiosity, our second and third propositions are: H2: The level of social distance is associated with the frequency of performing religious rituals. By that we mean, the higher the frequency of praying, fasting, watching or listening to religious programs, attending religious lessons in mosques or churches, attending Friday prayer/Sunday service, and/or reading religious books, the lower the level of social distance. H3: People who self-identify with a sectarian/denominational group have higher social distance than people who do not. Finally, RGCT theory posits that intergroup conflicts strengthen intragroup relations and reinforce in-group favoritism which is found to be associated with social distance (Sherif and Sherif 1979). Essentially, during conflicts the group becomes more important to the self. Attitudes i.e., voting for a candidate who belongs to the voter’s group, is found to be of vast importance to the individual. Having said that and due to the fact that the questions on in-group favoritism were not asked in Lebanon, we will investigate the relative importance of in-group favoritism as determinant of social distance in Iraq only. We expect our findings to be consistent with the literature and assume in-group favoritism in Iraq is positively associated with the outcome variable. Hence, our fourth proposition is: H4: The higher an individual’s in-group favoritism in Iraq, the higher the level of social distance towards out-groups. 12 DATA AND METHODS We will use Arab Democracy Barometer survey data, Wave II, collected between 20102011. This data seeks to measure and trace over time the social and political attitudes in more than ten Middle Eastern countries. In this data, over 1,231 Iraqis and 1,387 Lebanese are sampled. The use of this data is advantageous because it is one of the few nationally representative data ever collected in the region. After dropping the missing data, the sample size is dropped to 2,311 and the distribution of both countries are: Iraqis 44% and Lebanese 56%. Dependent Variable Our outcome variable is social distance. We use the following four items to create a social distance scale. The items assess the desirability of having different group members as neighbors. In the original data people were asked “members of which of the following groups would you not like to have as neighbors? 1- followers of other religions; 2- people of a different race and color; 3- expatriate workers and immigrants; and 4- displaced people and refugees.” A two-point scale is utilized to measure all four items (1= I do not want them to be my neighbors; and 2= I do not object). Scale items were reverse coded so that 1 indicates low social distance and 2 indicates high social distance. Then, we created a new scale out of the four items. Scale internal consistency reliability was determined by calculating Cronbach's coefficient alpha (0.71). Independent Variables The core independent variables are country, performing religious rituals, sectarian/denominational identification, and in-group favoritism. The country measure is categorical in the original data. Out of the ten countries surveyed, we dropped cases of 8 countries and kept Iraq and Lebanon. Then, we recoded country as a dummy so that (1=Lebanon) and (0=Iraq). As for performing religious rituals, we operationalize it as the frequency of performing religious activities and rituals such as praying, fasting, going to church/mosque, etc. Our measure 13 of religious rituals is a scaled measure that is composed of seven items. People were asked: “Do you pray daily? Do you fast during Ramadan/fast the 40 days of Lent for Christians? Do you watch or listen to religious programs on the radio or television? Do you attend religious lessons in mosques or churches? Do you attend Friday prayer/Sunday service? Do you listen to or read the Quran/the Bible? And do you read religious books?” In the original data a four-point scale is used to measure all the seven items: 1= always; 2= most of the times; 3=sometimes; and 4= rarely2. We reverse coded all the items so that low scores indicate low frequency of performance and high scores indicate high frequency of performance. Then, we created a scale and internal consistency reliability was determined by calculating Cronbach's coefficient alpha (0.82). Further, the indicator of sectarian/denominational identification is drawn upon the item “if I asked about your religion, would you prefer the answer be: 1- Orthodox; 2- Catholic; 3- Protestant; 4-Christian; 5-Sunni Muslim; 6-Shia Muslim; and 7-Muslim?” First, we dropped Christian denominations because of the small number of cases. Then, to have a better analysis of between-group variation in terms of self- identification with a sectarian group, we recoded the variable as a dummy so that anyone who self-identified with her/his sectarian group (Sunni Muslim and Shia Muslim=1) and those who did not self-identify with their sectarian group (otherwise=0). This method is previously used by Moaddel et al. (2012), Spierungs (2014) and many others. Finally, we operationalize in-group favoritism as the preference to cast a vote during elections for a candidate who belongs to voter’s family, tribe, ethnicity, denomination. We used the following three questions to create an in-group favoritism scale: 1- “In general, to what extent is it important to you that the candidate is from your family/tribe (relatives) in deciding who to vote for in elections whether the parliamentary, municipal or local elections?” 2- “In general, to 2 In this study only Muslims and Christians are included. Other religions were dropped because of the limited number of cases. 14 what extent is it important to you that that the candidate belongs to your ethnicity in deciding who to vote for in elections, whether parliamentary, municipal or local elections?” And 3- “In general, to what extent is it important to you that that the candidate belongs to your sect in deciding who to vote for in elections, whether parliamentary, municipal or local elections?” In the original data a four-point scale is used to measure all three questions: 1=to a great extent; 2= to a medium extent; 3=to a limited extent; and 4= not important. We reverse coded the scales so that low scores indicate lower in-group favoritism and high scores indicate higher in-group favoritism. Scale internal consistency reliability was determined by calculating Cronbach's coefficient alpha (0.76). Control Variables The analysis includes geographical area; we recoded it as a dummy (1=rural; 0=urban). Age is measured in years ranging from 18 to 75 and is treated as continuous. Educational attainment is, as well, treated as continuous (1=illiterate/literate; 2=elementary; 3=preparatory/basic; 4= secondary school; 5=mid-level diploma/professional or a technical degree; 6= Bachelors degree; and 7= higher education). Finally, gender is dummy coded (1= female; 0=male). 15 ANALYSES We used STATA package, version 14.1 to run the analyses. First, we conducted descriptive statistical analyses (see Table 1) in order to describe the population of Iraq and Lebanon with respect to the variables of interest. Then, we ran an OLS regression model to examine demographic factors and how they influence the outcome variable. After that, we ran another OLS regression to investigate the effect of country on social distance. Finally, we ran separate OLS models for each country to investigate how social and demographic variables, as well as, the frequency of performing rituals influence social distance. For Iraqis, we ran additional models and added in measures of sectarian/denominational identification and in-group favoritism. 16 FINDINGS Descriptive data in (Table 1) show that 26% of the Iraqi sample and 2% of the Lebanese sample are from rural areas. Average age is about 36 years old in Iraq and about 39 years old in Lebanon. Regarding education, Lebanese scored a slightly higher mean than Iraqis. Average education of the Iraqi sample is elementary and middle school, whereas, for Lebanese, average education is middle and high school. Further, 44% of the Iraqi sample and 42% of the Lebanese sample are women. Regarding the frequency of performing religious rituals, Iraqis reported higher frequencies of praying daily, fasting during Ramadhan/fasting the 40 days of Lent for Christians, and watching or listening to religious programs on the radio or television. Nevertheless, Lebanese reported higher frequencies of attending religious lessons in mosque or church, attending Friday prayer/Sunday service, listening or reading the Quran/Bible, and reading religious books. Further, 18% of Iraqis said they do not like to have followers of other religions as neighbors, 16% of them said they do not like to have people of a different race/color as neighbors, 40% said they do not like to have expatriate workers and immigrants as neighbors, and 18% said they do not like to have displaced people and refugees as neighbors. The percentages from the Lebanese sample are much lower (05%, 09%, 17%, and 16%). Moreover, 05% of Iraqis said that it is very important to vote for a candidate who is their relative, 14% said it is very important to vote for a candidate who belongs to their sectarian group, and 21% said is it is very important to vote for a candidate who belongs to their ethnic group. Finally, almost 80% of the Iraqis self-identified with a sectarian group (Sunni/Shia). The (Table 2) presents the results of the first two models that we ran to determine the impact of control variables (Model 1) and country (Model 2) on social distance. Model 1 includes predictors of age, education, gender, and geographical area of which only the geographical area show to significantly affect social distance. In other words, people in rural areas are more inclined to maintain social distance with other groups than people living in urban areas. 17 Table 1: Descriptive Data of Variables of Interest (Means) Variables Controls Iraq Lebanon Geographical area (Rural) 0.26 0.02 Age 35.91 38.61 Education 3.46 3.75 Female 0.44 .42 Praying daily 0.79 0.59 Fast during Ramadan/Fast the 40 days of Lent for Christians 0.76 0.57 Watch or listen to religious programs on the radio or television 0.28 0.22 Attend religious lessons in mosques or churches 0.05 0.16 Attend Friday prayer/Sunday services 0.25 0.40 Listen to or read the Quran/the Bible 0.20 0.36 Read religious books 0.08 0.21 Sectarian/Denominational Identification 0.79 __ Performing Religious Rituals In-Group Favoritism __ Voting for relatives 0.05 __ Voting for one’s sect 0.14 __ Voting for one’s ethnicity 0.21 __ Disfavoring to have followers of other religions as neighbors 0.18 0.05 Disfavoring to have people of a different race/color as neighbors 0.16 0.09 Disfavoring to have expatriate workers and immigrants as neighbors 0.40 0.17 Disfavoring to having displaced people and refugees as neighbors 0.16 Social Distance 0.18 Notes: Geographical area, age education, gender, sectarian/denominational identification are mean values: the remainder are proportions. N=1,303 for Lebanon; N=973 for Iraq * p<0.05; ** p<0.01; *** p<0.000 18 To test hypotheses #1, we added in the measure of country in Model 2. The results for age, education, and gender are identical to Model 1, except for the significant impact of geographical area. When the variance of country is accounted for, the relationship between geographical area and social distance is no longer statistically significant. As recommended by Baron and Kenny (1996), we performed tests for mediation and results indicate that we have a perfect situation of mediation, whereby the effect of geographical area is mediated by country variable. We ran three regression equations and results show that geographical area significantly affects country, geographical area significantly affects social distance, and country significantly affects social distance. Hence, we see that in Model 2, geographical area is no longer statistically significant. Further, Model 2 indicates that Lebanese have lower social distance than Iraqis. Table 2: OLS Regressions Variables Age Education Female Rural Model 1 B/SE Model 2 B/SE -0.000 -0.000 (000) (0.00) 0.004 0.007 (0.04) (0.00) 0.016 0.015 (0.01) (0.01) 0.087*** 0.032 (1.76) (0.01) Lebanon -0.106*** (0.02) Intercept 1.151*** 1.196*** (0.02) 2 R Mean VIF N (0.02) 0.01 1.05 2,311 0.04 1.10 2,311 * p<0.05; ** p<0.01; *** p<0.000 Given the fact that there was a variance in social distance across countries we ran separate group models (by country) to investigate the outcome variable. Ifatunji (2016) argues that to 19 measure the variance across ethnic, racial, religious, or national groups running separate group models is, conceptually and statistically more efficient than running one model in which groups are treated as dummies. Following (Tables 3), presents the results of the models ran to test hypothesis #2 for Lebanon, and (Table 4) presents the results of the models ran to test hypotheses #2; #3; and #4 for Iraq. Table 3: OLS Regressions/Lebanon Variables Model 3 B/SE Model 4 B/SE Age -0.000 -0.000 (0.00) (0.00) Education Female Rural Frequency of Rituals 0.000 0.000 (0.00) (0.00) 0.002 0.002 (0.14) (0.14) -0.051 -0.052 (0.04) (0.04) __ -0.000 (0.00) Intercept 1.112 (0.03)*** R2 Mean VIF N 0.004 1.06 1.097 (25.31)*** 0.005 1.05 1,303 * p<0.05; ** p<0.01; *** p<0.000 First, in the case of Lebanon (Table 3, Model 3 and 4), no variable shows any statistical significance. Second, in the case of Iraq (Table 4), we first looked to determine which control variables are associated with social distance (see Model 5). Then, we tested hypothesis #2 and added the frequency of performing rituals to Model 6. Further, to test hypothesis #3 we added religious/denominational identification to Model 7. Finally, to test hypothesis #4 we added in the measure of in-group favoritism to Model 8. Results indicate that the social and demographic variables of age, education, and gender do not have a statistically significant relationship with social distance. However, the relationship between geographical area and social distance becomes 20 statistically not significant when we account for the variance of religious/denominational identification (see Model 7). Before accounting for the variance of that predictor (Models 5 and 6), results indicate that people living in rural areas of Iraq are more inclined to maintain social distance with other groups than people living in urban areas. As recommended by Baron and Kenny (1996), we performed tests for mediation and results indicate that we have a second perfect situation of mediation, whereby the effect of geographical area is mediated by religious/denominational identification predictor. Table 4: OLS Regressions/Iraq Variables Age Model 5 B/SE Model 6 B/SE 0.000 Female Rural Frequency of Rituals Sectarian/Denominational Identification Voting for one’s own Group Model 8 B/SE 0.000 0.000 0.000 (0.00) (0.00) (0.00) 0.001 0.001 0.007 0.006 (0.00) (0.00) (0.00) (0.00) (000) Education Model 7 B/SE 0.000 0.015 -0.015 -0.001 (0.01) (0.01) (0.01) (0.01) 0.050* 0.049* 0.034 0.033 (0.02) (0.02) (0.02) (0.02) __ -0.021 -0.032 -0.031 (0.01) (0.01) (0.01) __ __ __ __ 0.133*** 0.107*** (0.02) (0.02) __ 0.061*** (0.01) Intercept 2 R Mean VIF N 1.151*** 1.196*** 1.133*** 1.192*** (0.02) (0.02) (0.02) (0.02) 0.007 1.05 973 0.009 1.06 973 0.04 1.06 973 0.07 1.07 973 * p<0.05; ** p<0.01; *** p<0.000 21 We ran three regression equations and results show that geographical area significantly affects religious/denominational identification, geographical area significantly affects social distance, and religious/denominational identification significantly affects social distance. Hence, we see that in Model 7, geographical area is no longer statistically significant. Model 8 predicts the impact of all the predictors on social distance. Interestingly, the ritualistic dimension of religiosity has no statistically significant relationship with social distance, whereas, religious denominational identification dimension has significance. As predicted, people who identified with their sectarian group are more inclined to maintain social distance with other groups than people who did not identify with their sectarian group. Finally, results in Model 8 indicate that the stronger the ingroup favoritism, the higher the social distance. 22 DISCUSSION In this study we investigated the determinants of social distance in Iraq and Lebanon. We found substantial differences across the two countries. As predicted, we found that Iraqis are more inclined to maintain social distance with out-groups than Lebanese. Then, when we ran separate country-specific models, we found no statistically significant relationship in the Lebanese sample between the variables we controlled for and social distance. However, in the Iraqi sample we found that people who self-identify with their sectarian group are more likely to maintain social distance with out-groups than people who do not. Moreover, we found that the higher the likelihood to cast a vote for a candidate from voter’s tribe, ethnicity, sect the higher the social distance. Additionally, we found that in the Iraqi sample the relationship between geographical area and social distance is moderated by the variable religious/denominational identification. We explain the difference in levels of social distance across the two countries by arguing that despite the fact these two countries have many things in common i.e., the heterogeneous social structure and the ethno-sectarian conflicts, they also have many differences. For example, ethno-sectarian violence in Lebanon has not been as bloody as in Iraq (Vibert, 2010). Also, in spite of persistent external and internal threats the Lebanese government has been more successful in maintaining the social and political stability in the last two decades. Furthermore, an important finding of this study concerns the impact of religiosity on social distance. We found that the ritualistic dimension of religiosity has no statistically significant impact on social distance in both samples. This finding is inconsistent with the literature that argues that the relationship between the frequency of performing religious rituals (such as church attendance) and social distance is positive. Nevertheless, as predicted, we found that in the Iraqi sample the religious/denominational identification dimension has a statistically significant impact on social distance. These findings support our argument and the idea that different dimensions of religiosity may have different impacts on individual’s social, 23 psychological, and political attitudes (Brinkerhoff and Mackie, 1986; Bloom et al, 2015). We argue that the insignificant impact of the first dimension on social distance could be due to a model misspecification problem. The ritualistic scale that we created out of seven different rituals may have lost or hidden some information. To diagnose that, we re-ran Models 4 and 8 with each ritual separately3. In the Lebanese sample none of seven rituals were associated with the outcome variable, whereas, in the Iraqi sample three rituals (daily prayers, fasting, and watching religious programs) showed a negative impact on the outcome variable and the other four rituals showed no statistically significant impact. We infer that each religious ritual may have a different impact on social distance. Findings of Brinkerhoff and Mackie (1986) support our argument. Authors divided rituals into formal (i.e., church attendance and contributing money to church) and informal (i.e., praying and reading the Bible) and found out that informal rituals positively affect social distance, whereas, informal rituals add very little to the model. Regarding the second dimension of religiosity, SIT and RGCT contribute in explaining that religion is a potential source of social identity and that the perceived threat to one’s religious/denominational group results in heightening in-group identification (Tajfel, 1981; Sherif and Sherif, 1979; Bobo, 1996). We discussed above that the long history of conflicts between Arab Sunnis and Shia in Iraq resulted in the salience of the Sunni and Shia identities (Dawisha, 2013; Haddad, 2014). Hence, it is logical to predict that in Iraq sectarian/denominational identification increases social distance and our findings do support that. Finally, we acknowledge that this study has notable limitations that could be addressed in the future. First, for the Lebanese sample we measured the impact of only one dimension of religiosity on the outcome variable and did not measure the impact of the second dimension nor the impact of in-group favoritism. Therefore, some of our findings are limited to 3 Models can be provided if needed. 24 the case of Iraq due to the fact that the survey data we used did not ask all the questions in both countries. Another potential limitation is that the previous social distance studies used at least two questions from Bogardus’ scale (would you have as next neighbors? and would you marry into group?) In this study we used the first question only because the second question was not asked in Lebanon. 25 BIBLIOGRAPHY 26 BIBLIOGRAPHY Abanes, M. S., Scheepers, P. L., & Sterkens, C. (2014). Ethno-religious groups, identification, trust and social distance in the ethno-religiously stratified Philippines. Research in Social Stratification and Mobility, 37, 61-75. Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. Journal of personality and social psychology, 5(4), 432. Al-Marashi, I. (2005). Iraqi Constitutional Debate, Gloria Center, Vol. 9, no. 3, Article 8, pp. 131. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182 Bloom, P. B. N., Arikan, G., & Courtemanche, M. (2015). Religious social identity, religious belief, and anti-immigration sentiment. American Political Science Review, 109(02), 203-221. Bobo, L., & Hutchings, V. L. (1996). Perceptions of racial group competition: Extending Blumer's theory of group position to a multiracial social context. American Sociological Review, 951-972. Bogardus, E. S. (1925). Measuring social distance. Journal of applied sociology, 9(2), 299-308. Bogardus, E. S. (1933). A social distance scale. Sociology & Social Research Bratton, M., & Kimenyi, M. S. (2008). Voting in Kenya: Putting ethnicity in perspective. Journal of Eastern African Studies, 2(2), 272-289. Brinkerhoff, M. B., & Mackie, M. M. (1986). The applicability of social distance for religious research: An exploration. Review of Religious Research, 151-167. Chaplin, E. (2006). Iraq's New Constitution: Recipe for Stability or Chaos?. Cambridge Review of International Affairs, 19(2), 271-284. Cole, J. (2004, March11). The Three-State https://www.thenation.com/article/three-state-solution/. Solution? The-Nation. Retrieved: Dawisha, A. (2013). Iraq: A political history. Princeton University Press. Dickson, Eric, and Kenneth Scheve. 2006. “Social Identity, Political Speech, and Electoral Competition.” Journal of Theoretical Politics 18 (1): 5–39. Dodge, Toby. Inventing Iraq: The failure of nation building and a history denied. Columbia University Press, 2003. 27 Duckitt, J. H. (1992). The social psychology of prejudice. Praeger Publishers/Greenwood Publishing Group. Ferree, K. E. (2006). Explaining South Africa's racial census. Journal of Politics, 68(4), 803-815. Fontan, V. (2009). Voices from post-saddam iraq: Living with terrorism, insurgency, and new forms of tyranny. Westport, Conn: Praeger Security International. Greene, S. (2004). Social identity theory and party identification. Social Science Quarterly, 85(1), 136-153 Glock, C. Y., & Stark, R. (1965). Religion and society in tension. Chicago: Rand McNally. Guo, H., & Tynen, S. (2015). Social distance, ethnicity, and religion in Northwest China: a survey of inter-and intra-group attitudes among Han, Hui, and Tibetan students. Eurasian Geography and Economics, 56(6), 679-712. Haddad, F. (2014). Sectarianism in Iraq: Antagonistic visions of unity. Oxford University Press. Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. Hunt, Chester L. (1956)."Social distance in the Philippines." Sociology & Social Research. Johnson, D. P. (1977). Religious commitment, social distance, and authoritarianism. Review of Religious Research, 99-113. Jackson, JW. (1993). Realistic group conflict theory: A review and evaluation of the theoretical and empirical literature. Psychological Record, Vol. 43, issue 3. Ifatunji, Mosi Adesina. (2016). Race, Ethnicity and Quantitative Methodology. Summer Program in Quantitative Methods of Social Research, Inter-university Consortium for Social and Political Research, Institute for Social Research, University of Michigan at Ann Arbor. Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life. Journal of personality and social psychology, 40(3), 414. Kirmanj, S. (2013). Identity and nation in Iraq. Lynne Rienner Publishers. Kirsch, A. D. (1957). Social distance in voting behavior related to N variables. Purdue University. Division of Educational Reference. KUPKE, H. G. (1971). religiosity and social distance: A community study. Kuznetsov, A. A. (2015). " The Islamic State" in the Context of Political Changes in the Middle East. Vestnik MGIMO, (2), 184-192. 28 Lijphart, Arend. 1999. Patterns of Democracy. New Haven: Yale University Press. Natali, D. (2016, June 23). The Myth of a Tripartite Iraq. Al-monitor. Retrieved from: http://www.al-monitor.com/pulse/originals/2016/06/iraq-myth-tripartite-sunni-shia-kurdstability-isis.html Melikian, L., & Prothro, E. T. (1952). Social distance and social change in the Near East. Sociology & Social Research. Moaddel, M., Kors, J., & Garde, J. (2012). Sectarianism and Counter-Sectarianism in Lebanon. Population Studies Center Research Reports, (12). Meade, R. D., & Singh, L. (1973). Changes in social distance during warfare: A study of the India/Pakistan War of 1971. The Journal of Social Psychology, 90(2), 325-326. Mendelberg, Tali. 2001. The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality. Princeton: Princeton University Press. O’Leary, B. (2003, September). Multi-national federalism, federacy, power-sharing and kurds of Iraq. In Keynote address to ‘Multi-nationalism, Power-Sharing and kurds in a New Iraq’, Cafritz Foundation Conference Center, George Washington University (Vol. 12). Oswald, D. L. (2005). Understanding Anti-Arab Reactions Post-9/11: The Role of Threats, Social Categories, and Personal Ideologies1. Journal of Applied Social Psychology, 35(9), 1775-1799. Owen, C. A., Eisner, H. C., & McFaul, T. R. (1981). A half-century of social distance research: National replication of the Bogardus studies. Sociology and Social Research, 66, 30–32 Prothro, T., & Melikian, L. (1952). Social distance and social change in the Near East. Sociology and Social Research, 37(1), 3-11. Park, R. E. (1923). The concept of social distance. Journal of Applied Sociology, 8, 339–344. Parrillo, V. N., & Donoghue, C. (2005). Updating the Bogardus social distance studies: a new national survey. The Social Science Journal, 42(2), 257-271. Parrillo, V. N., & Donoghue, C. (2013, September). The national social distance study: Ten years later. In Sociological Forum (Vol. 28, No. 3, pp. 597-614). Poppe, E., & Hagendoorn, L. (2004). Social distance of Russian minorities from titular population in former Soviet Republics. In M. Gijsberts, L. Hagendoorn, & P. Scheepers (Eds.), Nationalism and exclusion of migrants: Cross-national comparisons (pp. 143–156). England: Ashgate Publishing Limited. Saroglou, V. (2011). Believing, bonding, behaving, and belonging: The big four religious dimensions and cultural variation. Journal of Cross-Cultural Psychology, 42(8), 1320-1340. 29 Scheepers, P., Gijsberts, M., & Hello, E. (2002). Religiosity and prejudice against ethnic minorities in Europe: Cross-national tests on a controversial relationship. Review of Religious Research, 242265. Sherif, M. (1961). The Robbers Cave experiment: Intergroup conflict and cooperation.Wesleyan University Press. Sherif, M., & Sherif, C. (1979). Research on intergroup relations. In W. Austin, & S. Worchel (Eds.), The Social Psychology of Intergroup Relations. Monterey, CA: Brooks/Cole Shen, M. J., Yelderman, L. A., Haggard, M. C., & Rowatt, W. C. (2013). Disentangling the belief in God and cognitive rigidity/flexibility components of religiosity to predict racial and valueviolating prejudice: A Post-Critical Belief Scale analysis. Personality and Individual Differences, 54(3), 389-395. Šiber, I. (1998). War and the changes in social distance toward the ethnic minorities in Croatia. Politička misao, 34(5), 3-26. Smidt, Corwin E., Lyman A. Kellstedt, and James L. Guth. 2009. “The Role of Religion in American Politics: Explanatory Theories and Associated Analytical and Measurement Issues.” In Corwin Smidt, Lyman A. Kellstedt, and James L. Guth (eds.), Oxford Handbook on Religion and American Politics. Oxford: Oxford University Press, 3–42. Spierings, N. (2014). The Influence of Islamic Orientations on Democratic Support and Tolerance in five Arab Countries. Politics and Religion, 7(04), 706-733. Steensland, B., Robinson, L. D., Wilcox, W. B., Park, J. Z., Regnerus, M. D., & Woodberry, R. D. (2000). The measure of American religion: Toward improving the state of the art. Social Forces, 79(1), 291-318. Strabac, Z. (2016). Social Distance and Ethnic Hierarchies in Croatia, in, The aftermath of war: experiences and social attitudes in the western Balkans. Edited by Simkus, A. Routledge Starr, P. D. (1978). Continuity and Change in Social Distance: Studies from the Arab East-A Research Report. Social Forces, 1221-1227. Snyder, Jack. 2000. From Voting to Violence: Democratization and Nationalist Conflict. New York: W.W. Norton and Company. Wark, C., & Galliher, J. F. (2007). Emory Bogardus and the origins of the social distance scale. The American Sociologist, 38(4), 383-395 Tajfel, H. (1981). Human groups and social categories: Studies in social psychology. CUP Archive. Triandis, H. C., & Triandis, L. M. (1960). Race, social class, religion, and nationality as determinants of social distance. The Journal of Abnormal and Social Psychology, 61(1), 110 30 Terkildsen, Nayda. 1993. “When White Voters Evaluate Black Candidates: The Processing Implications of Candidate Skin Color, Prejudice, and Self-Monitoring.” American Journal of Political Science 37 (4): 1032–53. Turner, J. C., Brown, R. J., & Tajfel, H. (1979). Social comparison and group interest in ingroup favouritism. European journal of social psychology, 9(2), 187-204. Vibert, P. (April 15, 2010). Lebanon and Iraq. Foreign Policy Blogs). Retrieved from: http://foreignpolicyblogs.com/2010/04/15/lebanon-and-iraq/. Viret, C. (2015). Schools’ Role in Shaping Youth Collective Memory of Conflict and Experiences of Sectarianism in Lebanon. Voas, D. (2007). Surveys of behaviour, beliefs and affiliation: Micro-quantitative. The SAGE handbook of the sociology of religion, 144-166. Weaver, C. N. (2008). Social distance as a measure of prejudice among ethnic groups in the United States. Journal of Applied Social Psychology, 38(3), 779-795. Weiss, M. (2009). The historiography of sectarianism in Lebanon. History Compass, 7(1), 141154. Zeidel, R. (2010). On Servility and Survival: Sunnis Opposition to Saddam and the Origins of the Current Sunni Leadership in Iraq. In Zeidel, R., Baram, A., & Rohde, A. (Eds.). Iraq between Occupations: Perspectives from 1920 to the Present. Springer. 31