Ideological Polarization and Organizational Form Evolution: A Study of Islamic–Secular Rivalry and High Schools in Turkey, 1971–1998

Ideological Polarization and Organizational Form Evolution: A Study of Islamic–Secular Rivalry... Abstract How do rival organizational forms that are associated with different ideological values affect each other’s growth? And how do ideological polarization and state regulation impact the growth of these rival organizational forms? We adopt a community ecology framework as the stepping-stone to answer these questions in the context of a study of the dynamics of the Turkish high school educational system from 1971 to 1998—a period characterized by important regulatory changes amid growing ideological polarization between religious Islamic and secular factions. We explore how micro-level ideological Islamic–secular polarization and macro-level state regulation as to female participation may impact the meso-level interaction among high school forms. We demonstrate that rival organizational forms affect each other’s growth positively, on average, but that this mutualism is reduced by ideological polarization and reinforced by the increase in female participation in Islamic high school education. Introduction Interdependencies between different populations of organizations have long been understudied (Ingram and Simons 2000). However, a growing number of community ecology studies focus on sets of functional and/or spatial relationships between different organizational forms (Freeman and Audia 2006). As organizations affect the communities they belong to and are also affected by those communities, such work has the potential to “spread the relevance of results beyond organizationally defined problem areas” (Freeman and Audia 2006, 145). In this study, we apply such community analysis to explore how the reciprocal interplay of political and ideological struggles in societies impacts the emergence and prevalence of different organizational forms (Barnett and Woywode 2004; Boone and Özcan 2014; Schneiberg, King, and Smith 2008; Simons and Ingram 2004). As formal organizations are carriers of ideologies, their emergence and diffusion are major outcomes and drivers of social and political change (Barnett and Woywode 2004). Scholars agree that such work is therefore likely to advance organization theory “especially in light of recent criticisms that the field has lost sight of its mission to identify the role of organizations in society” (Simons and Ingram 2004, 34; see also Hinings and Greenwood [2002]). Only a limited number of studies have directly addressed how ideology affects the interdependencies between organizational forms in a community context. Two competing hypotheses, both supported by empirical evidence, have been proposed (Simons and Ingram 2004). The first hypothesis suggests that ideological affinity between organizations is an important source of mutualism: organizations that share the same set of ideological values positively co-evolve together. The second hypothesis claims that ideological similarity goes hand in hand with overlapping resource utilization: this increases competition between forms that share the same ideological values (Barnett and Woywode 2004). Simons and Ingram (2004) tried to reconcile these seemingly contradictory positions by invoking a key contingency—that is, the populations’ location in resource space, an important concept in ecology theory (van Witteloostuijn and Boone 2006). Ideological similarity is argued to be a source of mutualism when organizational forms do not share key resources, but a source of competition when they occupy similar positions in resource space. Current empirical evidence appears to be consistent with this unifying proposition (Simons and Ingram 2004). There is, however, an important blind spot in the literature on the ecology of ideology because ideological dissimilarity and how it affects interdependencies between organizational forms that compete for resources have seldom been studied (Simons and Ingram 2004). Given that ideological struggles between rival ideologies are main drivers of social and political change, this is unfortunate (Barnett and Woywode 2004; Olzak and West 1991). We relate to this by studying how an ideological struggle in society affects the prevalence of organizational forms carrying these rival ideological values, and whether such an ideological struggle is a source of competition or mutualism among these rival organizational forms. This is our first contribution. The interaction among rival organizational forms does not appear in a vacuum. Especially given the value-laden tensions associated with rival ideologies, the institutional and local environmental conditions are extremely important (Boone and Özcan 2014; Marquis and Huang 2009; Wade, Swaminathan, and Saxon 1998; Wholey and Sanchez 1991). We therefore analyze the interplay between polarization of local ideological preferences, central government regulation, and the dynamics of rival organizational forms. We explore how ideological polarization and state regulation may impact the ecological interaction among rival value-laden organizational forms. By doing so, we explicitly recognize that the type of interdependence between rival organizational forms is likely to be time and place bound (Aldrich 2009; Boone and Özcan 2014). Most studies in this tradition, however, assume that the ideological contextual climate and critical resource base are constant over time. By directly measuring shifts in polarization and resource niches, we show that the mutualistic interdependence between rival ideological forms is suppressed by increasing polarization (as reflected in election outcomes) and reinforced by resource niche expansion (through state regulation). This is our second contribution. We develop specific hypotheses that are tested in the Turkish high school community context in the 1971–1998 period. We analyze how religious Islamic–secular ideological polarization in Turkish society affects the growth at the district level of two school forms carrying oppositional ideologies: religious high schools that provide Islamic education (so-called Imam Hatip schools) and cosmopolitan high schools that provide Western-based secular education in foreign languages (so-called Anatolian schools). We investigate the conditions that affect the nature of interdependencies between organizational forms that carry this pair of rival ideologies. So, we seek to provide a richer understanding of the mechanisms and implications of ideological contention for the organizational dynamics of the Turkish high school educational system. The empirical setting we study is particularly interesting for our purposes for two reasons. First, the Turkish Republic is characterized by an old ideological tension involving the role of religion, particularly of Islam, in public life. In the twentieth century, ideological pressure was gradually built up to increase the role of Islam in public life, and to develop a kind of Turkish–Islamic synthesis. However, this triggered opposition from adherents of “Western”-style ideologies, leading to deep ideological polarization in Turkish society. As education is an extremely important vehicle for the diffusion of ideologies and values in societies (Nielsen and Hannan 1977), the Turkish high school educational system was and still is a major battlefield (Güven 2005), where this ideological polarization materializes via the establishment of Islamic and secular high schools.1 These rival high school dynamics are driven by both central government regulation trying to control high school densities and diversity, and local demand for different forms of high school education as reflected by ideological preferences. Given that data are available at the Turkish district level, we have both substantial cross-sectional and longitudinal variation, which offers the opportunity to examine the interplay of local ideological environments with high school organizational ecologies in the context of local communities. Such local variation is essential (though seldom available; see Marquis and Huang [2009, 1223]) to evaluate the contingent nature of the consequences of nationwide public educational policies. In addition, our setting allows us to substantiate Freeman and Audia’s (2006) plea for studies that focus on explaining the composition of different organizational forms (in casu high schools) in particular locales (in casu districts). Second, during the period of our study, the Turkish context features three major high school types that prepare students for higher education that nicely map onto the ideological resource space: regular high schools in the ideological “center” (which is the dominant secular school type established by Ataturk in 1924) and two oppositional types at both extremes, where Imam Hatip schools target the Islamic “tail” and the cosmopolitan schools are squarely located in the secular “tail.” As these school types all compete for the same resources (i.e., students), the setting is ideal to explore the blind spot in the ecology of ideology literature referred to above (Simons and Ingram 2004). Moreover, the resource space shifted substantially during the period of our study because only as of 1976 were female students allowed to study at Islamic high schools. This fundamental regulatory change moved up the carrying capacity for Islamic high schools substantially, and altered the nature of competition between Islamic and cosmopolitan schools, as now the education of females was at stake. Before developing our hypotheses, we first introduce the details of our setting. Islamic and Cosmopolitan High Schools in Turkey Growing Islamic–secular polarization in Turkey in the twentieth century The history of modern Turkey starts with the establishment of the Republic in 1923, with Mustafa Kemal Ataturk as the founding president. As of 1923, Ataturk tried to transform the remains of the Ottoman Empire into a modern and Western-oriented state, based on secular principles, with reforms that unified education, abolished Sharia, and introduced a modern civil code inspired by the Swiss example. Because the Kemalist modernization project implied a major revolution in the context of a formerly conservative and traditional Islamic society, Islamic opposition was mobilized from the very beginning (Zurcher 1993). This opposition eventually gave rise to the emergence of so-called political Islam in the second half of the twentieth century (Güven 2005; Öniş 1997; Sakallıoğlu 1996). During the single-party regime of the Republican People’s Party (RPP), the bureaucracy and army took control of all aspects of public affairs (Weiker 1981). Because of increasing contention and repression, Turkey shifted to a multi-party system in 1945. In this era, the population became politically mobilized, demanding to be involved in deciding on the nation’s policies (Weiker 1981). The 1950 elections resulted in the victory of the Democratic Party (DP), which criticized RPP to be coercive and not sensitive to the needs and traditions of the people, particularly freedom of religious activity (Ayata 1996). Although the DP remained loyal to the basic principles of secularism, they emphasized religion as a “necessary social cement for the cohesion of society” (Ayata 1996, 43) in their political discourse. The DP introduced voluntary Islamic courses into primary schools, restored the Arabic instead of the Turkish call to prayer, and established Imam Hatip (i.e., preacher) schools and a Faculty of Divinity at Ankara University. During the 1950s, the state elite was very critical of the DP government’s policies (Heper and Keyman 2006). The DP government ended in 1960 with a military coup. The DP was closed down and banned from politics. The National Unity Committee, consisting of army leaders, prepared a liberal-democratic constitution, which was approved by a referendum. This 1960 constitution paved the way for a period of liberalization, which ended with the prolonged clash of many ideologies, especially in the second half of the 1960s and the 1970s. (Landau 1997). In 1970, the National Order Party (NOP), the first openly pro-Islamic party of the Turkish Republic, was founded. In comparison to the DP, the NOP introduced a more aggressive discourse in terms of the role of religion in public life and education, and rejected any type of secularism that could be perceived as being hostile toward the Islam (Landau 1997). The NOP was banned in 1971 because of violating the secularist principles written in the constitution. The reason for political/social unrest was not only religious-secular polarization, but also the emergence of several extreme groups, both left and right wing (Zurcher 1993). To end this state of chaos, the military took over, again, in 1971. However, this military intervention did not end the chaos. Successive governments failed to effectively deal with both economic and political problems. The ideological spectrum widened even more with the foundation of many political parties and social movements, representing all kinds of ideologies, ranging from extreme left to extreme right, including religious sects and Kurdish separatists (Özbudun 1996). This period cumulated in another military coup in 1980. Like the earlier coup, the aim of the 1980 military intervention was to defend the Kemalist principles. However, this coup was also different from the previous ones in several aspects. First of all, the leaders of the military coup put restrictions on individual, group, and organizational freedom (e.g., limiting the role of labor unions), because they perceived the 1961 constitution to be too liberal in light of the ideological polarization in Turkish society. Second, they realized that religion can be a unifying force against fragmentation and contestation caused by the rise of leftists, socialists, radical religious sects, ultra-nationalists, and Kurdish separatists (Ahmad 1991; Pak 2004; Salt 1995). Following the coup, the military government took several measures, including a reform of the education system, to blend religion and national sentiments in an attempt to create a more homogeneous society (Yavuz 1997). They opened up new Qur’anic schools and made religious courses compulsory (Yavuz 1997). This ideology initiated the development of the so-called “Turkish Islamic Synthesis” (Güven 2005). This ideology was supported by the center-right Motherland Party, which took over government after the military. This ideology began to be disseminated through the educational system and media to the masses (Yavuz 1997), which eventually led to the victory of the Islamic Welfare Party in the elections of 1995, and the appointment of the first Islamic Prime Minister (Necmettin Erbakan) of the Turkish Republic. At the other side of the ideological spectrum, the secular groups interpreted these developments as a revolt against Kemalism (Yavuz 1997), which resulted in an escalation in tension and antagonism between both factions (Güven 2005; Heper and Güney 2000). The secular counter-movement resonated well not only among the traditional defenders of the secular Kemalist heritage such as state bureaucrats, retired army officers, secular journalists, and Kemalist organizations, but also among adherents of a new wave of Westernization favoring pluralism, freedom of expression, and social equality (Güven 2005). These platforms challenged the “anti-democratic religious tradition and impinged on emerging ideas of expanding freedoms and a socially diverse and pluralistic form of modernization” (Güven 2005, 203). The increasing role of the Islamic religion in public life, and particularly in education, was regarded as a threat to these treasured secular values. This is witnessed by the establishment of several counter-movement organizations in this era, such as the Association of Kemalist Thought, the Association of Kemalist Women, and the Support of Modern Life Association. All these controversies ultimately led to increasing ideological polarization, dividing Turkey into two very distinct factions: a religious/Islamic and a secular/Kemalist one. The Turkish high school educational system This polarization and contention between ideologies reverberated throughout the educational sector (Güven 2005). Particularly the high school system was an important arena, as here the seeds are sown for the development of the future elite of the country. One of the major strategies of the Islamic movement was the creation of a counter-elite by strengthening the Islamic stream in the educational system (Güven 2005). Two main subgroups can be distinguished: general and vocational/technical high schools. The former educate students for further study in higher education, such as universities, whereas the latter are orientated at specific vocations. In the current study, we focus on the regular, Anatolian and Imam Hatip schools, because they (1) are the most prevalent ones, (2) prepare students for higher education and entry into the elite, and (3) were engaged in strong ideological rivalry during the period of our study.2 These three types of general high schools were most salient in the period under study; the other high school forms (i.e., science high schools, private high schools, and Anatolian Imam Hatip high schools) are less prevalent and/or originated more recently. Indeed, including these latter high school types in the models does not have any effect on the analyses (available upon request). Within our time window, those who wanted to have a (standard) high school education had to decide between attending one of these three types of high schools. Because of the ideological contention, the educational system experienced many changes in regulations and restrictions, particularly with respect to the position of Islam in education. The aim of Ataturk was to provide one unified and secular educational system for all Turkish citizens (through the regular high schools), which was codified in the Law of Unification of Educational Instruction in 1924. During the twentieth century, however, the number of Imam Hatip high schools expanded to such an extent that duality in education is, by now, a fact of life (Güven 2005). The roots of Imam Hatip or Islamic high schools go back to Medresetü-l Eimmeti vel Hutaba, an educational institution founded in 1913 aimed at educating imams and preachers. The madrasah were first abolished as a result of the Law of Unification in 1924, and later re-established in order to educate enlightened imams and preachers who were loyal to the Kemalist ideology. Due to lack of students, the schools were closed again in 1930. In 1949, Imam Hatip high schools were opened to offer 10-month courses, monitored by the Republican People’s Party. It was the Democratic Party that finally established full-fledged Imam Hatip high schools in 1951. Even though the “official aim”3 of Imam Hatip high schools still was the education of imams and preachers, their curricula contained just as much arts and science classes as did regular high schools. As a result, these high schools provided an educational alternative for the children of Islamic and conservative families, and therefore gained popularity all around the country. A first important regulation that positively affected the expansion of Imam Hatip high schools was the 1973 change in the law. Before the 1973 law change, the official purpose of Imam Hatip schools was “educating imams and preachers for religious services.” The 1973 law expanded the official mission of Islamic schools, including “the preparation of students for higher education.” A very fundamental second regulation was the one of 1976 that allowed Imam Hatip high schools to admit female students, even though women cannot be employed as preachers or imams. This increased the “niche” for Islamic high schools substantially, as now conservative Islamic families felt comfortable to send their girls to school (Bozan 2007; Çarkoğlu and Toprak 2006).4 This regulation represents the hallmark of the transformation of Imam Hatip high schools from Islamic vocational high schools to general educational institutions. The number of Imam Hatip high schools increased rapidly from 71 in 1971 to 249 by 1980. Many of these schools were first established by means of public donations, and staffed afterward by the Ministry of Education. Imam Hatip high schools could strengthen their position even more after the regulation in 1982, which allowed their graduates to apply to any faculty program they would like after taking the university entrance exam. Even though these regulatory changes were initiated by the government, they can be seen as tactical moves of the Islamic faction in society to strengthen their position in the high school arena. Secularists perceived the expansion of both the mission and number of Imam Hatip high schools as a threat because a growing number of young people were educated to become not an imam or a preacher, but instead “potential agents of Islamic traditionalism and a potential source of votes for traditional-fundamentalist politicians” (Sezer 1993, 23–24). The graduates of Imam Hatip high schools step by step entered into elite functions in many sectors of Turkish society (Dorian 1997). For example, Recep Tayyip Erdoğan, the current president, graduated from an Imam Hatip high school before he obtained his university degree. The growing duality between secular and Islamic education reinforced the ideological polarization in society. All this brought the issue of Islamic schools onto the political agenda again. Specifically, in 1997, the military strongly opposed the Islamic policy of the government, arguing that “the threat against secularism was now more serious than formerly” (Esposito 2000, 6). In 1997, the military put the government under pressure to restore the restrictions for the graduates of Imam Hatip high schools with respect to the university entrance exam.5 Cosmopolitan, or Anatolian, high schools are positioned at the other extreme side of the ideological spectrum, as their educational curriculum embodies secularist values and includes education in foreign languages. The first cosmopolitan high schools were established as alternatives to foreign (e.g., French, German, or Italian) high schools that provided high-quality and foreign language education in Turkey. The state was suspicious that students of foreign schools would be influenced too much by foreign values, and would not sufficiently be taught about their own history and culture. Therefore, the government established Turkish high schools in 1955, which provide education in foreign languages and which accept their students based on the Nationwide High School Entrance Exam. These Anatolian high schools can be seen as equivalent to the gymnasium in Germany or the Netherlands, in terms of their student acceptance criteria, curriculum, and success. Anatolian high schools claim that, by giving education in foreign languages, they open up students’ minds, helping them communicate and be in touch with the rest of the world much more easily. All course material consisted of books imported from Western countries. The students who were accepted to these schools were the most competitive ones, and the level of education was more challenging than in other types of high schools in the period we analyze. Only one compulsory religious course is taught in cosmopolitan high schools. This course (of one hour per week) is called “Religion and Morality,” and is not designed to introduce students to a religious worldview. Figures 1A and B show the density of Islamic, cosmopolitan, and regular high schools over time. Figure 1. View largeDownload slide A. The density of regular high schools, 1971–1998. Figure 1. View largeDownload slide A. The density of regular high schools, 1971–1998. Figure 1. View largeDownload slide B. The density of Islamic and cosmopolitan high schools, 1971–1998. Figure 1. View largeDownload slide B. The density of Islamic and cosmopolitan high schools, 1971–1998. Islamic versus cosmopolitan high schools are located at the opposite sides (or tails) of the ideological spectrum, with regular schools positioned in the middle (or center). Both rival forms have strongly dissimilar ideologies, but share a similar resource base by targeting the country’s potential student population. The Turkish high school system constitutes a zero-sum game, ceteris paribus; enrollment of a student in one school form goes at the expense of other forms. Islamic and cosmopolitan high schools might not directly attract (potential) students from each other, but both aim to enlarge their student base by competing for center students who would otherwise opt for regular schools. Predicting the Growth of Islamic and Cosmopolitan High Schools Formal organizations play a prominent role in affecting social and political change by advocating new ideological agendas or by defending the status quo (Barnett and Woywode 2004; McCarthy and Zald 1977; Olzak and West 1991). Organizations are carriers of ideologies that are often designed to diffuse and reinforce preferences regarding social outcomes, and the way these outcomes can be obtained (Simons and Ingram 2004). Particularly, schooling is an organized activity consisting of organizations that are highly infused with ideologies, and the schooling system is of major strategic importance to political elites committed to changing the relationship of the individual to the state or polity (Nielsen and Hannan 1977). The demand and supply of specific types of schooling is, therefore, strongly driven by ideologies that prevail among different interrelated actors in society, such as parents choosing education for their children and political elites designing educational policies, and private actors trying to influence political elites via all sorts of social movement organizations, also to obtain schooling that fits with their ideology (Güven 2005). Indeed, in the Turkish context, local communities have opportunities to push for the establishment of new high schools in their communities. Formally, the final decision about high school founding is the authority of the Ministry of Education. However, whether a new high school will be founded and which type of high school is going to be established can be influenced by the local community and the local governor, implying that often the decision of the Ministry of Education depends on the needs of the district. For instance, if the local community donates the necessary buildings and equipment for a new high school, then the government tends to give permission for the founding of the high school, subsequently assigning teaching staff. In that case, the donating individual(s) have the right to make the decision as to which type of high school is going to be established. Hence, the founding process of high schools in Turkey involves both a top-down and a bottom-up process, implying that both local conditions and local collective actions are important drivers of new high school formation. We argue that especially contention between rival ideologies will spur the growth of organizational forms that carry these rival ideologies—in our case, high schools. Polarization increases the saliency of individual identities, in-group solidarity, and between-group competition and conflict (Esteban and Ray 1994; Olzak and West 1991; Sherif and Sherif 1953; Simmel 1923). Polarization between rival ideologies is a very potent mobilizing force, spurring actors to protect their ideological position by organized activity (Meyer and Staggenborg 1996; Olzak and West 1991). Polarization increases the saliency of ideological positions by raising the awareness of ideological markers and identities (Olzak and West 1991). Increasing polarization and contention between opposed ideological groups also creates the threat that the existing benefits may be taken away, and/or new harms might be imposed by the opposing group or the state (Almeida 2003; Goldstone and Tilly 2001; Jasper 1997; Tilly 1978; Van Dyke and Soule 2002). The growing importance of political Islam triggered secular countermovements that mobilized the secular public opinion against the anti-secular rhetoric of the former (Güven 2005; Heper and Güney 2000). During this period, especially Imam Hatip high schools became highly politicized, becoming an instrument for political actors to signal the community’s standing regarding the religiosity scale. As a result, the need for Imam Hatip high schools and the parents’ willingness to send their children to these schools became an issue of taking a side with either position on the ideological spectrum. That is why the type of high school founded in a district is a signal to the local community, which resulted in competition between religious and center-right parties in order to get Islamic votes (Yavuz 1997). On top of this, during this period of contestation, many civil society associations and organizations were formed around Imam Hatip high schools and Islamist political parties, examples being İlim Yayma Cemiyeti, Ensar Vakfi, and ÖNDER. These organizations “provided scholarships to Imam Hatip students, fund dormitories, organize conferences and lobby on behalf of the Imam Hatip community” (Sarfati 2015, 685). These organizations not only support Imam Hatip high schools, but also “act as intermediaries between the [Imam Hatip] schools and Islamist political formations by creating shared identities among students, graduates, conservative business circles and activists and by [fostering] a feeling of trust within religiously conservative communities that the politicians will recognize” (Sarfati 2015, 685). Given their curricula and the atmosphere in these schools, Anatolian high schools represent the other end of the secular-religious dimension (Divarcı Çakmaklı, Boone, and van Witteloostuijn 2017). Therefore, for secular families these schools became the best alternative. During this period, graduates of Anatolian high schools also established associations in support of Anatolian high schools. In addition to that, many other organizations emerged to form a countermovement vis-à-vis the Islamic movement, such as the Association of Kemalist Thought (Atatürkçü Düşünce Derneği) and the Association for the Support of Contemporary Living (Çağdaş Yaşamı Destekleme Derneği). In the decision-making process regarding high school foundings, the movement and countermovement organizations act as mobilizing forces, by targeting the state in an attempt to influence public policy (Johnson, Agnone, and McCarthy 2010). As a result of increasing polarization, both polarized groups felt threatened (Almeida 2003). If the recipients of this threat are well organized, the level of collective action and resistance to the threat is expected to be higher (Almeida 2003; Gould 1991; Jenkins 1983; Walton and Seddon 1994). Such a cycle of movement and countermovement creates a political opportunity structure favorable to those organizational forms that carry the associated rival ideological values, as long as the underlying tensions are not “resolved” by the state (Meyer and Staggenborg 1996). Hypothesis 1: Ideological Islamic–secular polarization increases the growth of both Islamic and cosmopolitan high schools. Several scholars have contributed to the so-called ecology of ideology by analyzing the ideological drivers of inter-organizational dynamics (Barnett and Woywode 2004; Boone and Özcan 2014; Minkoff 1994, 1995; Simons and Ingram 2004; Vermeulen 2013; Wade, Swaminathan, and Saxon 1998). Some claim and find that ideological similarity between organizational forms is associated with strong competition (Barnett and Woywode 2004; Vermeulen 2013), while others report that similar ideologies will help each rival form in forwarding their own ideology (Boone and Özcan 2014; Simons and Ingram 2004). Simons and Ingram (2004) argue that these contradicting findings are due to the fact that some studies confound ideological similarity with high resource overlap between organizational forms. When resource overlap is low and ideological similarity is high, interdependency between forms is mutualistic. Then, forms sharing similar ideologies enjoy the benefit of a strong esprit de corps (Barnett and Woywode 2004; Vermeulen 2013), without fighting for the same resources: “organizations can generate ideological mutualism by sharing experience and information, providing patronage, money, encouragement, and legitimacy, and by cooperating in efforts at political influence” (Ingram and Simons 2000, 34–35). This claim is corroborated in studies regarding the founding and mortality rates of different types of cooperatives in Israel during the twentieth century (Ingram and Simons 2000; Simons and Ingram 2004). The density of credit cooperatives reduces the failure rate of worker cooperatives (Ingram and Simons 2000), and increases the founding rate of kibbutzim and moshavim, both agricultural cooperatives (Simons and Ingram 2004). The density of kibbutzim, in turn, suppresses the mortality rate of worker cooperatives (Ingram and Simons 2000). All these organizational forms share a similar “socialist” ideology, without competing in the same niche for resources. In a similar vein, Minkoff (1995, 1997) found that the development of the civil rights organizations was an important driver of the subsequent growth of the number of women’s organizations between 1955 and 1985. However, when resource overlap is large, the potential benefit of an esprit de corps cannot compensate for the cost of strong competition. Simons and Ingram (2004) find that kibbutzim and moshavim affect each other competitively: the density of the former has a negative effect on the founding rate of the latter and vice versa, notwithstanding the fact that both forms share a common ideology. This is due to both forms’ reliance on similar resources for survival, such as land, distributors, and participants (Simons and Ingram 2004). In their study of the Viennese newspaper industry for the period 1918–1938, Barnett and Woywode (2004) also report that competition among newspapers is strongest when newspapers share similar ideologies. Competitive relationships have also been reported for populations that have dissimilar ideologies, but do not compete for the same resources directly. Due to competition in the ideological arena, organizations might try to coerce organizations that carry rival ideologies by withholding resources (Ingram and Simons 2000). Banks, the ideal-typical representatives of the capitalist ideology, have had a strong influence on the cooperative, socialist form of economic activity in Israel. Simons and Ingram (1997) report that kibbutzim that relied on banks for financial input were more likely to adopt capitalist principles, such as employing hired labor. Bank density indeed increased the failure rate of worker cooperatives that were perceived to be close to the core of the socialist ideology (Ingram and Simons 2000). Similarly, the number of corporations was associated with a decrease in the founding rate of kibbutzim and moshavim cooperatives (Simons and Ingram 2004). Scholars have speculated that when organizations’ ideologies/identities are dissimilar and the resource overlap is high, there will be conflicting forces impacting their relationship. Although “diametric opponents are hardly friends … they at least enhance each other’s ideological identity by their sharp contrast” (Barnett and Woywode 2004, 1461). This might strengthen a population’s identity, cohesion, and appeal, partially offsetting strong competition in the resource arena. Simons and Ingram (2004) suggest that ideological rivalry among populations might promote collective action and the populations’ capacity to build institutions. Boone and Özcan (2014) claim that this relationship is moderated by local-level variables. In their study of county-level founding processes of cooperatives in the US ethanol industry, they show that local conditions spur and facilitate collective action among farmers to establish cooperatives in response to the local diffusion of corporations. It still remains unclear, however, what outcomes these conflicting forces will generate in different circumstances. Following the arguments of Simons and Ingram (2004), Olzak and West (1991), Boone and Özcan (2014), and Barnett and Woywode (2004), we predict as a baseline that the interdependence between populations (groups) that carry rival ideologies will be mutualistic. Rivalry between dissimilar ideologies makes the ideologies extremely sharp and salient, which mobilizes the adherents of these ideologies to diffuse their values and preferences. This is consistent with Meyer and Staggenborg’s (1996, 1638) proposition that “when movements effectively create or exploit events, they are likely to encourage countermovement mobilization at the same time that they advance their own causes.” The salience argument contends that ideological rivalry “galvanizes” the opposing groups, and so spurs the formation of formal organizations that carry these ideological values (Olzak and West 1991, 459). Such organization building mobilizes even more collective action (Tilly 1978), such that the establishment of one organizational form will reinforce the dynamics of the rival one, leading to mutualistic interdependence. Hypothesis 2a: The growth of Islamic high schools increases with the number of cosmopolitan high schools. Hypothesis 2b: The growth of cosmopolitan high schools increases with the number of Islamic high schools. However, such a mutualistic relationship is unlikely to hold indefinitely as the focus of the ideological and resource rivalry eventually becomes the “destruction of the foe” (Barnett and Woywode 2004, 1461). Vermeulen (2013), for instance, analyzed to what extent the context affects the form of competition and/or collaboration among different ideologies. His analysis of Turkish immigrant organizations in Germany and the Netherlands reveals that an open, supportive environment spurs the activities of different ideological groups and also encourages intra- and inter-ideological cooperation. Following this logic, we focus on two context moderators—one repressing mutualism (i.e., increases in ideological polarization), and the other reinforcing mutualism (i.e., increases in the carrying capacity). First, we expect that progressing ideological polarization will eventually repress the positive impact of the growth of one organizational form on the growth of the rival form. Although ideological polarization is predicted to spur mobilization initially (hypothesis 1), further polarization may well intensify competition between organizational forms associated with these extreme positions. In fact, the ultimate aim of such mobilization and competition is to repress the rival ideology to become the dominant one. The political opportunity structure has an important impact on the ideology that will turn out to be the “winner” (Meyer and Staggenborg 1996). For instance, in Vienna from 1918 to 1938, the right-wing newspapers eventually won the contest, destroying the clerical center as “the environment of political opportunity shifted in such a way as to give apparent support for the claims of the extreme right” (Barnett and Woywode 2004, 1489). We expect that especially the mutualistic effect of the cosmopolitan schools on their Imam Hatip counterparts will be repressed by Islamic–secular polarization, because the rapid expansion of the Islamic high schools in the 1970s triggered a cosmopolitan countermovement that gained momentum in a political environment that gradually became more favorable to secular and Western values in the second half of the 1980s. We hence expect a negative interaction effect of Islamic–secular polarization and the density of cosmopolitan schools on the growth of Islamic high schools. As interaction effects are symmetric, such a negative interaction implies that we expect that the positive effect of ideological polarization on the growth of Imam Hatip schools decreases, or may even become negative, with the density of cosmopolitan high schools. Hypothesis 3: The positive impact of the number of cosmopolitan high schools on the growth of Islamic schools decreases (or becomes negative) when Islamic–secular polarization increases. Second, what happens when carrying capacity shifts? As discussed in detail by Soule and King (2008), social movement organizations adopt more specialized tactical repertoires when competition increases. The law of 1976, which allows female students in Imam Hatip schools, can be seen as a tactical move of Islamic schools, channeled through governmental intervention, to keep up competition with cosmopolitan schools. This law increased the high school population’s carrying capacity, especially in the Islamic high school niche. Many Islamic parents who formerly were not willing to send their daughters to regular or other high schools felt now at ease to enroll their girls in Islamic schools. Competition between cosmopolitan and Islamic schools entered a new arena—that is, competition for female education. This not only shifted the opportunity structure for both organizational forms upward (Meyer and Staggenborg 1996), but also revitalized the mobilization of both rival ideologies. This expansion of the resource niche and renewed mobilization are expected to extend the sustainability of the mutualistic interdependence between cosmopolitan and Islamic high schools. Hypothesis 4a: The growth of Islamic high schools increases with the number of cosmopolitan high schools especially when the proportion of female students in Islamic high schools to all female students increases. Hypothesis 4b: The growth of cosmopolitan high schools increases with the number of Islamic high schools especially when the proportion of female students in Islamic high schools to all female students increases. Data, Measures, and Methods Our dataset covers the period from 1971 to 1998, for 80 districts in Turkey. Even though full-fledged Islamic high schools were already founded in the 1950s, they only became part of the general high school educational field in the beginning of the 1970s when their official mission was expanded to include the preparation of students for higher education. Therefore, we start our analysis as of 1971, the year of the second military coup in Turkey that marks this new era. We end our analysis in 1998 because a new regulation in 1998 fundamentally transformed the high school system, implying that the data before and after 1998 are not comparable. The high school data are obtained from the National Education Statistics that are published by the Prime Ministry State Institute of Statistics. The core of our dataset consists of the number of Islamic (i.e., Imam Hatip), cosmopolitan (i.e., Anatolian), and regular high schools in Turkey that existed between 1971 and 1998. Following Ruef (2004) and Zhou and van Witteloostuijn (2010), our dependent variable is the first difference of the density (i.e., the number) of Islamic and cosmopolitan high schools at the district level in time t+1 and time t, respectively.6 As schools were seldom closed in our time period, this first difference variable is almost tantamount to the number of foundings. For modeling purposes, we therefore can treat this variable as a count, replacing the limited number of negative first difference values by zeros.7 Our main independent variables consist of the density of Islamic and cosmopolitan high schools at the district level. We lagged densities in order to avoid simultaneity. To test the niche expansion H4a and 4b, we employ the (lagged) proportion of female students in the Islamic high schools to all female students. For the measurement of polarization, we take the polarization index proposed by Esteban and Ray (1994). This index is based on the Gini coefficient, which accounts for both the size of the subgroups and distance: P=k∑i=1N∑j=1Nπi1+απj|yi−yj| in which k is the scaling factor, π denotes the size of the groups (i.e., the proportion of individuals belonging to group i), |yi−yj| gives the distance between groups i and j on a salient continuum (in this study, the religious-secular dimension), and α refers to the degree of polarization sensitivity that takes a value between 0 and 1.6 (Esteban and Ray 1994). When α is zero, P equals the Gini index. The lower α, the more P measures inequality (stressing alienation due to distance); the higher α, the more sensitive P is to polarization between large homogeneous subgroups (Stark, Hyll, and Behrens 2010). This measure captures the intuition that polarization is high when there is high homogeneity within each group, a high degree of heterogeneity between groups (i.e., distance), and a small number of groups of significant size (Esteban and Ray 1994). We set k = 1, and α = 0.8. By doing so, we follow most researchers by giving a balanced weight to the size of and the distance between the subgroups. We compute P using the outcomes of parliamentary elections at the district level to measure religious-secular polarization in the local community. The proportion of each political party’s votes denotes the size of different groups that vote for each political party in each district. To obtain an indicator of distance, the second author (who has Turkish nationality) classified political parties on a secular/religious scale ranging from 1 (extremely secular) to 10 (extremely religious) (available upon request). In order to score each political party’s religiosity on this scale, the second author analyzed each political party’s official program (for this period) to determine its emphasis on religion, and consulted several articles/books on the religious orientation of political parties during this period (Cizre 2008; Cizre-Sakallıoğlu and Çınar 2003; Göle 1997; Keyman 2007; Öniş 2007; Sakallıoğlu 1996; Somer 2007; Yavuz 1997). This classification was double-checked by two Turkish political scientists. The data on the parliamentary elections are obtained from the Turkish Prime Ministry State Institute of Statistics. The years without elections were filled in with the election outcomes from the beginning of the term. Our main control variables are the density of regular high schools at the district level, the (logarithm of) human population at the district level, human population growth at the district level, national GDP in real prices, and the percentage of government expenditures devoted to education at the country level, all lagged to avoid simultaneity. Regular high school density is included to capture the interaction with the high schools in the ideological center. The other variables are introduced to control for the high school population’s carrying capacity. We also add the percentage of votes that religious parties received at the district level (taking into account the votes received by the parties with a score of 10). We include dummies for the different official regions in Turkey, assigning 1 if the district is in that region and 0 otherwise. We add a dummy variable to control for the years following a territorial change of a few districts (1 for the years after the change, and 0 otherwise). Finally, to control for the important regulations of 1976 (that allowed female students to enroll in the Islamic high schools) and 1997 (that restored the university restrictions on the graduates of the Islamic high schools), we enter time dummies (0 before and 1 after the regulation). Since our dependent variable is a count measure, a Poisson process provides a natural baseline for modeling the growth of high schools (Hannan and Freeman 1987, 1989). However, Poisson regression should not be applied in the case of overdispersion (McCullagh and Nelder 1989). Overdispersion does not affect the coefficient estimates, but results in underestimated standard errors. Negative binomial regression adequately deals with this issue by adding a stochastic error component, following a gamma distribution, to the model (Cattani, Pennings, and Wezel 2003). The structure of the data is a pooled cross-section and time series (districts * years). Pooled data generally exhibit autocorrelation. Following Baron, Hannan, and Burton (2001), we use Liang and Zeger’s (1986) method of generalized estimating equations (GEE), which generalizes quasi-likelihood estimation to the panel data context, and which is a very flexible way to deal with clustered data. We assume an exchangeable autocorrelation structure. Because the observations within the districts cannot be assumed to be independent, we report Huber-White robust standard errors. A drawback of GEE is that summary goodness-of-fit statistics are problematic because the residuals of these models are correlated (Ballinger 2004; Zorn 2001). We therefore report Wald chi-square statistics, which test the null hypothesis that all regression coefficients are equal to zero. As Ballinger (2004, 146) warned, this statistic is not suited to compare the goodness of fit of alternative models. All models are estimated using the XTGEE routine of version 10.0 of STATA. Findings Descriptive statistics and bivariate correlations between the variables are reported in tables 1 and 2. Tables 3, 4, and 5 provide GEE negative binomial regression estimates of the growth of Islamic, cosmopolitan, and regular high schools, respectively. We add analyses with regular high school growth as the dependent variable to shed further light on the underlying cross-form interactions (see below). Table 1. Descriptives of Major Variables under Study Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 View Large Table 1. Descriptives of Major Variables under Study Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 View Large Table 2. Correlations of Major Variables under Studya 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − an = 1943; * = p < 0.05 (two-tailed). View Large Table 2. Correlations of Major Variables under Studya 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − an = 1943; * = p < 0.05 (two-tailed). View Large Table 3. Negative Binomial GEE Regression Estimates of the Growth in the Number of Islamic High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < . 01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 3. Negative Binomial GEE Regression Estimates of the Growth in the Number of Islamic High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < . 01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 4. Negative Binomial GEE Regression Estimates of the Growth in the Number of Cosmopolitan High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** an = 1943; + = p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 4. Negative Binomial GEE Regression Estimates of the Growth in the Number of Cosmopolitan High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** an = 1943; + = p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 5. Negative Binomial GEE Regression Estimates of the Growth in the Number of Regular High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 5. Negative Binomial GEE Regression Estimates of the Growth in the Number of Regular High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Both types of high schools grow slower in the East and Southeast of Turkey (Marmara being the omitted region). In central Anatolia, Islamic high schools grow faster, and in the Black Sea region cosmopolitan high schools grow slower. District population strongly spurs the growth of both Islamic and cosmopolitan high schools, and population growth hampers their growth. National GDP is negatively related to the growth of Imam Hatip high schools, and positively to the growth of cosmopolitan high schools. This finding is related to the rise of a devout bourgeoisie middle class, whose welfare started to increase over time, and hence did not want to send their children to Imam Hatip high schools anymore. The percentage of people in the district voting on Islamic parties has the expected effect: positive on Islamic and negative on cosmopolitan high schools. As expected, the 1976 regulation strongly enhanced the growth of Islamic high schools, whereas the 1997 regulation reduced its growth in favor of cosmopolitan high schools. Model 2 in table 3 shows a strong positive effect of ideological polarization on the growth of Islamic high schools. This estimate does not change when we add the densities of the three high school types (model 3 in table 3). For cosmopolitan high schools, we find a marginally significant negative effect of ideological polarization that becomes insignificant, however, when we control for the high school type densities (models 2 and 3 in table 4). These findings provide partial support for H1: ideological polarization spurs the growth of Islamic high schools, but not that of cosmopolitan high schools. The estimates of model 3 in both tables 3 and 4 provide support for H2a and 2b: the nature of interdependence between both extreme high school forms is mutualistic. The positive impact of cosmopolitan on Islamic high schools (B = 0.13) is much larger than the other way around (B = 0.04). Model 4 in table 3 reveals that the mutualistic effect of cosmopolitan on Islamic high schools is attenuated by the extent of ideological polarization, as is clear from the negative and significant interaction effect of cosmopolitan high school density and ideological polarization, providing support for H3. Figure 2 visualizes this interaction. This moderation effect of ideological polarization does not appear with respect to the impact of Islamic high school density on cosmopolitan high school growth (model 4 in table 4). Figure 2. View largeDownload slide The interaction effect of polarization and cosmopolitan school density on Islamic school growth Figure 2. View largeDownload slide The interaction effect of polarization and cosmopolitan school density on Islamic school growth In model 5 in both tables 3 and 4, we test whether (relative) niche expansion (i.e., proportion of female students in Islamic high schools) positively moderates the mutualistic interdependence between both forms. Both interaction coefficients are significant: (relative) niche expansion indeed positively reinforces this mutualistic interdependence, which provides evidence for H4a and 4b. Figures 3A and B show both interactions. Figure 3. View largeDownload slide A. The interaction effect of the proportion of female students in Islamic schools to all female students and cosmopolitan school density on Islamic school growth. B. The interaction effect of the proportion of female students in Islamic schools and Islamic school density on cosmopolitan school growth Figure 3. View largeDownload slide A. The interaction effect of the proportion of female students in Islamic schools to all female students and cosmopolitan school density on Islamic school growth. B. The interaction effect of the proportion of female students in Islamic schools and Islamic school density on cosmopolitan school growth Finally, to further explore the center–tail cross-form interaction, and hence to reveal the underlying interactive processes, we estimate the same models to analyze the growth of the regular high schools. District population increases the growth of regular high schools, while the growth of district population has a negative effect on the growth of regular high schools. The 1976 and 1997 regulations have a negative impact, too. Models 2 and 3 in table 5 reveal that neither ideological polarization nor the densities of Islamic and cosmopolitan high schools have a significant impact on the growth of regular high schools. Important for our analyses is the result in model 4 in table 5, which reveals a negative and significant interaction effect of the density of Islamic high schools and ideological polarization. The interaction of cosmopolitan high school density and ideological polarization is also negative and significant (model 5 in table 5): Islamic and cosmopolitan high school densities suppress the expansion of the regular high schools when ideological polarization increases, suggesting strong competition between all forms. Moreover, models 6 and 7 in table 5 show that the (relative) niche expansion variable does not moderate the impact of Islamic or cosmopolitan high school density on the growth of regular high schools. These findings suggest that the expansion of both Islamic and cosmopolitan schools affects the growth of regular high schools, and both school forms compete for the students at the center. Finally, we ran a series of robustness analyses, all available upon request, to explore the extent to which our findings might be sensitive to adding additional control variables and different calculations of the polarization index. First, we added (1) dummy variables for the military coup years (see the section on the history of the Republic), (2) the density of science high schools, (3) the density of private high schools, such as Robert College, and (4) a dummy for districts that have a population above 250,000 (as a proxy for urbanization). Note that the density of the other high school types is low in our time window. Second, we ran analyses with our polarization index calculated using α = 0.4 and α = 1.2, employing linear interpolation to fill in the years without elections. None of these robustness analyses changed the results reported above, increasing our confidence in the results found. Discussion Our main findings can be summarized as follows. Economic forces (i.e., national GDP) positively affect the expansion of cosmopolitan high schools: this high school form appears to follow general trends of economic development. State regulation and the size of the local Islamic faction (i.e., the number of votes on Islamic parties), in contrast, strongly affect the expansion of Islamic high schools. Islamic–secular polarization at the district level has a strong mobilizing impact on the growth of these Islamic high schools over and above the size of the local Islamic faction. Additionally, the interdependence between both organizational forms is mutualistic, on average. Both forms stimulated each other’s growth rate as a result of a sequence of movements and countermovements, starting with the initial rise of Islamic high schools in the late 1970s. The mutualistic effect of the cosmopolitan high schools on the growth of Islamic high schools declines, however, with further increases with Islamic–secular polarization. Mutualism between rival ideology-laden organizational forms is eventually repressed by ideological polarization, ceteris paribus. The expansion of the resource space of Islamic high schools by allowing enrollment of female students, in contrast, reinforces the mutualistic relationships between both rival forms. Moreover, both Islamic and cosmopolitan schools feed on the center, as the former grow at the expense of the latter when ideological polarization increases. This pattern of findings reveals that Islamic–secular polarization shifts the density distribution of high schools with different ideologies such that the number of rival ideology forms grows at the expense of the center. A major underlying driver of this pattern of findings is the subtle balance between two major conflicting forces that operate simultaneously: the mobilizing force of ideological polarization and the repressing force of competition among rival ideologies. The first force makes ideologies and values salient, and creates strong subgroup identities that facilitate collective action (e.g., social movements and/or government regulation). Such collective action subsequently spurs defensive countermovements of the opposing ideology. The second force results from the desire of adherents of specific ideologies to become dominant. The purpose of ideological rivalry is to repress the rival ideology. The interplay between the context’s opportunity structure and endogenous organizational dynamics determines which force dominates at a particular moment in time. As a result of this complex interplay of conflicting forces, we cannot predict ex ante the specific outcome of ideological rivalry (especially in settings where the contention remains unsettled). However, being able to ex post explain and understand the mechanisms behind these processes is scientifically valuable in its own right (Davis and Marquis 2005). Our community ecology approach reveals that formal organizations play an important role in the rivalry among ideologies, as such organizations serve as the vehicles of the diffusion of values in society. This is literally the case for the high schools that we studied. By establishing formal organizations, ideological rivalry is institutionalized in a relatively permanent way, underscoring Tilly’s (1978) observation that organizations are important in prevailing and promoting ideological positions. Our paper contributes to the literature in several ways. First, we add to community ecology and the ecology of ideology. Although several ecologists have stressed the importance of the study of interdependencies between different organizational forms, the number of such studies is still limited. Especially studies that combine relations between (classes of) organizations (so-called functional complementarities) with spatial differentiation are rare. In their review of the field, Freeman and Audia (2006, 165) propose that “simultaneous consideration of spatial and functional differentiation and complementarity can provide more nuanced views of the mechanism underlying density effects.” This is exactly what we sought to do in the present study by modeling the interplay between the growth of different high school forms against the background of local ideological differences (and state regulation). Related to this, we contribute to the ecology of ideology by analyzing the dynamics of organizational forms with dissimilar (opposing) ideologies that compete for similar resources, which has been identified as a major gap in the literature (see also Boone and Özcan [2014]). Although Simons and Ingram (2004) and Barnett and Woywode (2004) already speculated about the importance of the conflicting forces central to our argument, ours is one of the first studies that tries to unravel the conditions driving their relative impact. Most studies in this tradition implicitly assume that the ideological contextual climate and critical resource base are constant over time. By directly measuring shifts in ideological polarization and resource niches, we show that the mutualistic interdependence between rival ideological forms is suppressed by ideological polarization (as reflected in election outcomes) and reinforced by resource niche expansion (through state regulation). Second, by introducing the moderating effects of ideological polarization and state regulation, we contribute to “new organizational synthesis” research that “goes beyond studying movements as movement organizations, aiming instead to trace how movements shape organizational fields, and vice versa” (Schneiberg, King, and Smith 2008, 636). We do not study movements directly, but we do analyze the dynamic interplay between ideological struggles and the expansion of organizational forms carrying specific ideologies. Although it is generally accepted that organizations play a central role in facilitating many kinds of collective actions, much less is known about whether, when, and how social tension and conflict (such as polarization) affect the life chances of organizations (Olzak and West 1991). We show that ideological polarization is a major mobilizing force, having long-lasting effects on the specific development of organizational fields of major importance, such as high school education. Our study reveals that the dynamics of contention, with its cycles of opportunity, mobilization, and countermobilization, produces historical and path-dependent trajectories of different kinds of organizations (Marquis and Lounsbury 2007; Schneiberg, King, and Smith 2008). Third, our study analyzes the impact of culture and policy by incorporating both polarization and regulation, revealing that “policy makers should take underlying cultural factors into account when anticipating the intended effects of policy … Though policy may reflect majority will, there may be significant pockets of resistance or social movements that condition the intended effect … Thus, identifying the heterogeneous cultural beliefs and their organization may be important when implementing policy” (Marquis and Huang 2009, 1242). Turkish educational high school policies have generated many unexpected consequences as a result of these heterogeneous and polarized beliefs. Islamic high schools were regulated following the military coup of 1971 for the purpose of controlling religious Islamic education in Turkey. However, their increasing popularity spurred their numbers as well as the number of rival ideological high school forms. Female emancipation policies enlarged the ideological battlefield to competition in education. All this did not settle the issue of the role of Islam in society, but is likely to have, in contrast, nurtured ideological polarization in Turkey, which the government in the 1970s wanted to avoid in the first place. Although we focused on the Turkish high school system, we think our community ecology framework can be applied to understand organizational dynamics in many other settings as long as ideological polarization is salient. For example, growing polarization can also be observed in the United States among Democrats and Republicans with very clear-cut distinctions in terms of several issues concerning healthcare, civil rights, immigrants, and so on (Jacobson 2016). This, in turn, is associated with the establishment of new organizations, such as Breitbart and BuzzFeed, representing either side of the polarized debate. Another example is the emergence of Islamic banking as an insurgent logic against conventional banking in Turkey and other Islamic countries, fueled by ideological competition between secular and traditional Islamic factions in society (Boone and Özcan 2016). Our study is not without limitations, which provide opportunities for further research. First, the relationships between ideological polarization, state regulation, and expansion of high school forms are reciprocal in nature. Regulation is not exogenous, but affected by the distribution of ideological values. Moreover, different school forms diffuse different ideologies among a society’s population, which subsequently affects ideological polarization. As a first step, we decided to treat state regulation and ideological polarization as context variables that affect the growth of high school forms in a variety of ways. In future research, it would be interesting to build models that focus on Islamic–secular polarization as the phenomenon to explain, emphasizing this two-way channel of reciprocal causation. Second, we only focused on the three major high school forms in Turkey. More recently, interesting new high school types emerged that are probably also related to the continuing ideological struggle in Turkey. For instance, hybrid high school types were established, such as the Anatolian Imam Hatip high schools. These schools blend the quality of non-religious Anatolian high schools (which we labeled cosmopolitan high schools) with the intensive Islamic education of Imam Hatip high schools. Most courses in these schools are also taught in foreign languages, as in cosmopolitan-secular high schools. At the other side of the ideological spectrum, more and more private cosmopolitan high schools emerged, where students have to pay for top-quality education. This suggests that the frontline of the ideological struggle is gradually shifting to high-quality education of the future elite in Turkish society. Notes 1 Since Turkey is a secular country, officially all school types are secular. In this paper, the secular-religious distinction does not refer to the formal status, but to the curriculum of, and atmosphere in, the schools (Erkan and Akçayöz 2003; Pak 2004). 2 Note that all high school types analyzed in this study are public schools funded by the state. None of them required a tuition fee. 3 Official aim refers to what is written in the mission statement of Imam Hatip high schools. 4 We acknowledge that not all conservative families, especially Kurdish ones, were comfortable with sending their children to Imam Hatip schools—not only their girls, but also their boys. To this day, there are still informal madrasah in the Kurdish area, which are run by local religious leaders, serving as an alternative to Imam Hatip schools. The Kurdish conservative families historically have approached the Imam Hatip schools with suspicion, mostly because of their mistrust in their secular curriculum. Unfortunately, we do not have any other data on the number of students attending madrasah instead of high schools (as these madrasah are informal, this data is not available). Moreover, because of its politicized nature, the government can decide to establish Imam Hatip high schools in Kurdish areas as a signal, irrespective of the number of students that go to Imam Hatip high schools. So, we only focus on the number of high schools established, and not on the number of students that go to these high schools. 5 These restrictions relate to the coefficients that are applied as weights to the high school grades of students in order to determine whether a student is allowed to enter the university. For instance, if an Imam Hatip high school graduate wanted to enter the Theology faculty, then her high school grades were multiplied by 0.5; however, if she wanted to enter any other faculty at the university, then her high school grades were multiplied by 0.2. Obviously, this restriction was a disadvantage for Imam Hatip high school graduates to enter university in comparison to other high school graduates. 6 We employ the number of schools instead of the number of students because, in the short run, mobilization is related to the emergence of different school types while affecting the number of students in the long run. 7 Unfortunately, no information is available on the limited number of school closing events. About the Authors Christophe Boone is Professor of Organization Theory and Behavior at the Faculty of Applied Economic Sciences at the University of Antwerp (Belgium). His research interests focus on the dynamics of organizational populations in local communities, the antecedents and consequences of team and organizational diversity, CEO values and cognition, and the neuroeconomics of decision-making. Anıl Divarcı Çakmaklı is an Assistant Professor in the Department of Management at Istanbul Şehir University, Turkey. She holds master’s degrees in Economics from Istanbul Technical University and in Philosophy and Economics from Erasmus University Rotterdam, and a PhD in Management from the University of Antwerp. Her research focuses on organization theory and strategy. Arjen Van Witteloostuijn is Dean of the School of Business and Economics of the Vrije Universiteit Amsterdam in the Netherlands. He is associate editor of the Journal of International Business Studies and reviewing editor of Cross Cultural and Strategic Management, and sits on the editorial board of the British Journal of Management and Industrial and Corporate Change. He has published about 170 articles in international journals, including the American Journal of Sociology, American Sociological Review, Academy of Management Journal, Academy of Management Review, American Journal of Political Science, Journal of Public Administration Research and Theory, Organization Science, PLoSONE, and Strategic Management Journal. References Ahmad , Feroz . 1991 . “ Politics and Islam in Modern Turkey .” Middle Eastern Studies 27 ( 1 ): 3 – 21 . Google Scholar CrossRef Search ADS Aldrich , Howard . 2009 . “ Lost in Space, Out of Time: Why and How We Should Study Organizations Comparatively .” Research in the Sociology of Organizations 26 : 21 – 44 . Google Scholar CrossRef Search ADS Almeida , P. D. 2003 . “ Opportunity Organizations and Threat-Induced Contention: Protest Waves in Authoritarian Settings .” American Journal of Sociology 109 ( 2 ): 345 – 400 . Google Scholar CrossRef Search ADS Ayata , Sencer . 1996 . “ Patronage, Party, and State: The Politicization of Islam in Turkey .” Middle East Journal 50 : 40 – 56 . Ballinger , Garry A. 2004 . “ Using Generalized Estimating Equations for Longitudinal Data Analysis .” Organizational Research Methods 7 : 127 – 50 . Google Scholar CrossRef Search ADS Baron , James N. , Michael T. Hannan , and M. Diane Burton . 2001 . “ Labor Pains: Organizational Change and Employee Turnover in Young, High-Tech Firms .” American Journal of Sociology 106 : 960 – 1012 . Google Scholar CrossRef Search ADS Barnett , William P. , and Michael Woywode . 2004 . “ From Red Vienna to the Anschluss: Ideological Competition among Viennese Newspapers during the Rise of National Socialism .” American Journal of Sociology 109 : 1452 – 99 . Google Scholar CrossRef Search ADS Boone , Christophe , and Serden Özcan . 2014 . “ Why Do Cooperatives Emerge in a World Dominated by Corporations? The Diffusion of Cooperatives in the US Bio-Ethanol Industry, 1978–2013 .” Academy of Management Journal 57 : 990 – 1012 . Google Scholar CrossRef Search ADS Boone , Christophe , and Serden Özcan . 2016 . “ Ideological Purity vs. Hybridization Trade-Off: When Do Islamic Banks Hire Managers from Conventional Banking? ” Organization Science 27 : 1380 – 96 . Google Scholar CrossRef Search ADS Bozan , Irfan . 2007 . Devlet ile Toplum Arasında Bir Okul: İmam Hatip Liseleri, Bir Kurum: Diyanet İşleri Başkanlığı. Istanbul : TESEV Publishing . Çarkoğlu , Ali , and Binnaz Toprak . 2006 . Degisen Turkiye’de Din, Toplum ve Siyaset . Istanbul : TESEV Publishing . Cattani , Gino , Johannes M. Pennings , and Filippo C. Wezel . 2003 . “ Spatial and Temporal Heterogeneity in Founding Patterns .” Organization Science 14 : 670 – 85 . Google Scholar CrossRef Search ADS Cizre , Ümit . 2008 . Secular and Islamic Politics in Turkey: The Making of the Justice and Development Party. London and New York: Routledge . Cizre-Sakallıoğlu , Ümit , and Menderes Çınar . 2003 . “ Turkey 2002: Kemalism, Islamism, and Politics in the Light of the February 28 Process .” South Atlantic Quarterly 102 ( 2 ): 309 – 32 . Google Scholar CrossRef Search ADS Davis , Gerald F. , and Christopher Marquis . 2005 . “ Prospects for Organization Theory in the Early Twenty-First Century: Institutional Fields and Mechanisms .” Organization Science 16 : 332 – 43 . Google Scholar CrossRef Search ADS Divarcı Çakmaklı , Anıl , Christophe Boone , and Arjen van Witteloostuijn . 2017 . “ When Does Globalization Lead to Local Adaptation? The Emergence of Hybrid Islamic Schools in Turkey, 1985–2007 .” American Journal of Sociology 122 ( 6 ): 1822 – 68 . Google Scholar CrossRef Search ADS Dorian , Jones . 1997 . Higher Education in Turkey: Times Higher Education Supplement. Issue 1277, p. 10. Erkan , Rüstem , and Harika Akçayöz . 2003 . “ Siyasi İslam Tartışmaları Açısından İmam–Hatip Lisesi Öğrencilerinin Demokratik Tutum ve Davranışlarının İncelenmesi .” Cumhuriyet Üniversitesi Sosyal Bilimler Dergisi 27 ( 2 ): 174 – 201 . Esposito , John L. 2000 . “Islam and Secularism in the Twenty-First Century.” In Islam and Secularism in the Middle East , edited by John Esposito and Azzam Tamini . London : Hurst and Company , 1 – 12 . Esteban , Joan , and Debraj Ray . 1994 . “ On the Measurement of Polarization .” Econometrica: Journal of the Econometric Society 62 : 819 – 51 . Google Scholar CrossRef Search ADS Freeman , John , and Pino G. Audia . 2006 . “ Community Ecology and the Sociology of Organizations .” Annual Review of Sociology 32 : 145 – 69 . Google Scholar CrossRef Search ADS Goldstone , Jack , and Charles Tilly . 2001 . “Threat (and Opportunity): Popular Action and State Response in the Dynamic of Contentious Action.” In Silence and Voice in the Study of Contentious Politics , edited by R. Aminzade , D. McAdam , E. Perry , W. Sewell , S. Tarrow , and C. Tilly , 179 – 94 . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Göle , Nilüfer . 1997 . “ Secularism and Islamism in Turkey: The Making of Elites and Counter-Elites .” Middle East Journal 51 : 46 – 58 . Gould , Roger . 1991 . “ Multiple Networks and Mobilization in the Paris Commune, 1871 .” American Sociological Review 56 : 716 – 29 . Google Scholar CrossRef Search ADS Güven , Ismail . 2005 . “ The Impact of Political Islam on Education: The Revitalization of Islamic Education in the Turkish Educational Setting .” International Journal of Educational Development 25 : 193 – 208 . Google Scholar CrossRef Search ADS Hannan , Michael T. , and John Freeman . 1987 . “ The Ecology of Organizational Founding: American Labor Unions, 1836–1985 .” American Journal of Sociology 92 : 910 – 43 . Google Scholar CrossRef Search ADS Hannan , Michael T. , and John Freeman . 1989 . Organizational Ecology. Cambridge, MA : Harvard University Press . Heper , Metin , and Aylin Güney . 2000 . “ The Military and the Consolidation of Democracy: The Recent Turkish Experience .” Journal of Armed Forces and Society 26 : 635 – 58 . Google Scholar CrossRef Search ADS Heper , Metin , and Fuat Keyman . 2006 . “ Double-Faced State: Political Patronage and the Consolidation of Democracy in Turkey .” Middle Eastern Studies 34 ( 4 ): 259 – 77 . Google Scholar CrossRef Search ADS Hinings , Chris R. , and Roy Greenwood . 2002 . “ Disconnects and Consequences in Organization Theory? ” Administrative Science Quarterly 47 : 411 – 21 . Google Scholar CrossRef Search ADS Ingram , Paul , and Tal Simons . 2000 . “ State Formation, Ideological Competition, and the Ecology of Israeli Workers’ Cooperatives, 1920–1992 .” Administrative Science Quarterly 45 : 25 – 53 . Google Scholar CrossRef Search ADS Jacobson , G. C. 2016 . “ Polarization, Gridlock, and Presidential Campaign Politics in 2016 .” ANNALS of the American Academy of Political and Social Science 667 ( 1 ): 226 – 46 . Google Scholar CrossRef Search ADS Jasper , James . 1997 . The Art of Moral Protest. Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Jenkins , J. Craig . 1983 . “ Resource Mobilization Theory and the Study of Social Movements .” Annual Review of Sociology 9 : 527 – 53 . Google Scholar CrossRef Search ADS Johnson , E. W. , J. Agnone , and J. D. McCarthy . 2010 . “ Movement Organizations, Synergistic Tactics and Environmental Public Policy .” Social Forces 88 ( 5 ): 2267 – 92 . Google Scholar CrossRef Search ADS Keyman , E. Fuat . 2007 . “ Modernity, Secularism and Islam: The Case of Turkey .” Theory, Culture & Society 24 ( 2 ): 215 – 34 . Google Scholar CrossRef Search ADS Landau , Jacob M. 1997 . “Turkey between Secularism and Islamism.” Jerusalem Letter/Viewpoints 352. Liang , Kung-Yee , and Scott L . Zeger . 1986 . “ Longitudinal Data Analysis Using Generalized Linear Models .” Biometrika 73 : 13 – 22 . Google Scholar CrossRef Search ADS Marquis , Christopher , and Michael Lounsbury . 2007 . “ Vive la résistance: Competing Logics and the Consolidation of U.S. Community Banking .” Academy of Management Journal 50 : 799 – 820 . Google Scholar CrossRef Search ADS Marquis , Christopher , and Zhi Huang . 2009 . “ The Contingent Nature of Public Policy and the Growth of U.S. Commercial Banking .” Academy of Management Journal 52 : 1222 – 46 . Google Scholar CrossRef Search ADS McCarthy , John D. , and Mayer N . Zald . 1977 . “ Resource Mobilization and Social Movements: A Partial Theory .” American Journal of Sociology 82 : 1212 – 41 . Google Scholar CrossRef Search ADS McCullagh , Peter , and John A. Nelder . 1989 . Generalized Linear Models , 2nd ed . Boca Raton, FL : CRC Press . Google Scholar CrossRef Search ADS Meyer , David S. , and Suzanne Staggenborg . 1996 . “ Movements, Countermovements, and the Structure of Political Opportunity .” American Journal of Sociology 101 : 1628 – 60 . Google Scholar CrossRef Search ADS Minkoff , Debra C. 1994 . “ From Service Provision to Institutional Advocacy: The Shifting Legitimacy of Organizational Forms .” Social Forces 72 ( 4 ): 943 – 69 . Google Scholar CrossRef Search ADS Minkoff , Debra C. 1995 . Organizing for Equality: The Evolution of Women’s and Racial-Ethnic Organizations in America, 1955–1985 . New Brunswick, NJ : Rutgers University Press . Minkoff , Debra C. 1997 . “ The Sequencing of Social Movements .” American Sociological Review 62 : 779 – 99 . Google Scholar CrossRef Search ADS Nielsen , François , and Michael T . Hannan . 1977 . “ The Expansion of National Educational Systems: Tests of a Population Ecology Model .” American Sociological Review 42 : 479 – 90 . Google Scholar CrossRef Search ADS Olzak , Susan , and Elizabeth West . 1991 . “ Ethnic Conflict and the Rise and Fall of Ethnic Newspapers .” American Sociological Review 56 : 458 – 74 . Google Scholar CrossRef Search ADS Öniş , Ziya . 1997 . “ The Political Economy of Islamic Resurgence in Turkey: The Rise of the Welfare Party in Perspective .” Third World Quarterly 18 : 743 – 66 . Google Scholar CrossRef Search ADS Öniş , Ziya . 2007 . “ Conservative Globalists versus Defensive Nationalists: Political Parties and Paradoxes of Europeanization in Turkey .” Journal of Southern Europe and the Balkans Online 9 ( 3 ): 247 – 61 . Google Scholar CrossRef Search ADS Özbudun , Ergun . 1996 . “ Turkey: How Far from Consolidation? ” Journal of Democracy 7 ( 3 ): 123 – 38 . Google Scholar CrossRef Search ADS Pak , Soon-Yong . 2004 . “ Cultural Politics and Vocational Religious Education: The Case of Turkey .” Comparative Education 40 : 321 – 41 . Google Scholar CrossRef Search ADS Ruef , Martin . 2004 . “ The Demise of an Organizational Form: Emancipation and Plantation Agriculture in the American South, 1860–1880 .” American Journal of Sociology 109 : 1365 – 1410 . Google Scholar CrossRef Search ADS Sakallıoğlu , Ümit C. 1996 . “ Islam-State Interaction in Turkey .” International Journal of Middle East Studies 28 : 231 – 51 . Google Scholar CrossRef Search ADS Salt , Jeremy . 1995 . “ Nationalism and the Rise of Muslim Sentiment in Turkey .” Middle Eastern Studies 31 ( 1 ): 13 – 27 . Google Scholar CrossRef Search ADS Sarfati , Yusuf . 2015 . “ Islamic Schools in Modern Turkey: Faith, Politics, and Education .” British Journal of Middle Eastern Studies 42 ( 4 ): 684 – 86 . Google Scholar CrossRef Search ADS Schneiberg , Marc , Marissa King , and Thomas Smith . 2008 . “ Social Movements and Organizational Form: Cooperative Alternatives to Corporations in the American Insurance, Dairy and Grain Industries .” American Sociological Review 73 : 635 – 67 . Google Scholar CrossRef Search ADS Sezer , Duygu B. 1993 . State and Society in Turkey: Continuity and Change? Santa Monica, CA : Rand Publications . Sherif , Muzaver , and Carolyn W. Sherif . 1953 . Groups in Harmony and Tension. New York : Harper & Row . Simmel , Georg . 1923 . Conflict and the Web of Group Affiliation. Glencoe, IL : Free Press . Simons , Tal , and Paul Ingram . 1997 . “ Organization and Ideology: Kibbutzim and Hired Labor, 1951–1965 .” Administrative Science Quarterly 42 : 784 – 813 . Google Scholar CrossRef Search ADS Simons , Tal , and Paul Ingram . 2004 . “ An Ecology of Ideology: Theory and Evidence from Four Populations .” Industrial and Corporate Change 13 : 33 – 59 . Google Scholar CrossRef Search ADS Somer , Murat . 2007 . “ Moderate Islam and Secularist Opposition in Turkey: Implications for the World, Muslims and Secular Democracy .” Third World Quarterly 28 ( 7 ): 1271 – 89 . Google Scholar CrossRef Search ADS Soule , Sarah , A and Brayden G. , King . 2008 . “ Competition and resource partitioning in three social movement industries .” American Journal of Sociology 113 : 1568 – 610 . Google Scholar CrossRef Search ADS Stark , Oded , Walter Hyll , and Doris A. Behrens . 2010 . “ Gauging the Potential for Social Unrest .” Public Choice 143 : 229 – 36 . Google Scholar CrossRef Search ADS Tilly , Charles . 1978 . From Mobilization to Revolution. Reading, MA : Addison Wesley . Van Dyke , Nella , and Sarah Soule . 2002 . “ Structural Social Change and the Mobilizing Effect of Threat: Explaining Levels of Patriot and Militia Organizing in the United States .” Social Problems 49 ( 4 ): 497 – 520 . Google Scholar CrossRef Search ADS van Witteloostuijn , Arjen , and Christophe Boone . 2006 . “A Resource-Based Theory of Market Structure and Organizational Form.” Academy of Management Review 31 : 409 – 26 . Google Scholar CrossRef Search ADS Vermeulen , Floris . 2013 . “ Mutualism, Resource Competition and Opposing Movements among Turkish Organizations in Amsterdam and Berlin, 1965–2000 .” British Journal of Sociology 64 ( 3 ): 453 – 77 . Google Scholar CrossRef Search ADS Wade , Jim B. , Anand Swaminathan , and Michael-Scott Saxon . 1998 . “ Normative and Resource Flow Consequences of Local Regulations in the American Brewing Industry, 1845–1918 .” Administrative Science Quarterly 43 : 905 – 35 . Google Scholar CrossRef Search ADS Walton , John , and David Seddon . 1994 . “Food Riots Past and Present.” In Free Markets and Food Riots: The Politics of Global Adjustment , edited by J. Walton and D. Seddon , 23 – 54 . Oxford : Blackwell . Google Scholar CrossRef Search ADS Weiker , Walter F. 1981 . The Modernization of Turkey: From Ataturk to the Present Day . London: Holmes and Meier Pub . Wholey , Douglas R. , and Susan M . Sanchez . 1991 . “ The Effects of Regulatory Tools on Organizational Populations .” Academy of Management Review 16 : 743 – 67 . Google Scholar CrossRef Search ADS Yavuz , M. Hakan . 1997 . “ Political Islam and the Welfare (Refah) Party in Turkey .” Comparative Politics 30 : 63 – 82 . Google Scholar CrossRef Search ADS Zhou , Chaohong , and Arjen van Witteloostuijn . 2010 . “ Institutional Constraints and Ecological Processes: Evolution of Foreign-Invested Enterprises in the Chinese Construction Industry, 1993–2006 .” Journal of International Business Studies 41 : 539 – 56 . Google Scholar CrossRef Search ADS Zorn , Christopher J W 2001 . “ Generalized Estimating Equation Models for Correlated Data: A Review with Applications .” American Journal of Political Science 45 : 470 – 90 . Google Scholar CrossRef Search ADS Zurcher , Erik J. 1993 . Turkey: A Modern History . London : I. B. Tauris . Author notes We thank the participants of the 2010 Organizational Ecology conference in Helsinki, of the ACED conference on team and community diversity (Antwerp, 2010), and of the seminars at Copenhagen Business School (November 2010) and Sabanci University (January 2011), especially Özgecan Koçak and Serden Özcan, for stimulating us to develop the ideas presented in the paper and for their useful comments. This project is financed by the Flemish Science Foundation (FWO) in the realm of the Odysseus program of the Flemish government. © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Forces Oxford University Press

Ideological Polarization and Organizational Form Evolution: A Study of Islamic–Secular Rivalry and High Schools in Turkey, 1971–1998

Loading next page...
 
/lp/ou_press/ideological-polarization-and-organizational-form-evolution-a-study-of-uc8ZDKcpz9
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0037-7732
eISSN
1534-7605
D.O.I.
10.1093/sf/sox103
Publisher site
See Article on Publisher Site

Abstract

Abstract How do rival organizational forms that are associated with different ideological values affect each other’s growth? And how do ideological polarization and state regulation impact the growth of these rival organizational forms? We adopt a community ecology framework as the stepping-stone to answer these questions in the context of a study of the dynamics of the Turkish high school educational system from 1971 to 1998—a period characterized by important regulatory changes amid growing ideological polarization between religious Islamic and secular factions. We explore how micro-level ideological Islamic–secular polarization and macro-level state regulation as to female participation may impact the meso-level interaction among high school forms. We demonstrate that rival organizational forms affect each other’s growth positively, on average, but that this mutualism is reduced by ideological polarization and reinforced by the increase in female participation in Islamic high school education. Introduction Interdependencies between different populations of organizations have long been understudied (Ingram and Simons 2000). However, a growing number of community ecology studies focus on sets of functional and/or spatial relationships between different organizational forms (Freeman and Audia 2006). As organizations affect the communities they belong to and are also affected by those communities, such work has the potential to “spread the relevance of results beyond organizationally defined problem areas” (Freeman and Audia 2006, 145). In this study, we apply such community analysis to explore how the reciprocal interplay of political and ideological struggles in societies impacts the emergence and prevalence of different organizational forms (Barnett and Woywode 2004; Boone and Özcan 2014; Schneiberg, King, and Smith 2008; Simons and Ingram 2004). As formal organizations are carriers of ideologies, their emergence and diffusion are major outcomes and drivers of social and political change (Barnett and Woywode 2004). Scholars agree that such work is therefore likely to advance organization theory “especially in light of recent criticisms that the field has lost sight of its mission to identify the role of organizations in society” (Simons and Ingram 2004, 34; see also Hinings and Greenwood [2002]). Only a limited number of studies have directly addressed how ideology affects the interdependencies between organizational forms in a community context. Two competing hypotheses, both supported by empirical evidence, have been proposed (Simons and Ingram 2004). The first hypothesis suggests that ideological affinity between organizations is an important source of mutualism: organizations that share the same set of ideological values positively co-evolve together. The second hypothesis claims that ideological similarity goes hand in hand with overlapping resource utilization: this increases competition between forms that share the same ideological values (Barnett and Woywode 2004). Simons and Ingram (2004) tried to reconcile these seemingly contradictory positions by invoking a key contingency—that is, the populations’ location in resource space, an important concept in ecology theory (van Witteloostuijn and Boone 2006). Ideological similarity is argued to be a source of mutualism when organizational forms do not share key resources, but a source of competition when they occupy similar positions in resource space. Current empirical evidence appears to be consistent with this unifying proposition (Simons and Ingram 2004). There is, however, an important blind spot in the literature on the ecology of ideology because ideological dissimilarity and how it affects interdependencies between organizational forms that compete for resources have seldom been studied (Simons and Ingram 2004). Given that ideological struggles between rival ideologies are main drivers of social and political change, this is unfortunate (Barnett and Woywode 2004; Olzak and West 1991). We relate to this by studying how an ideological struggle in society affects the prevalence of organizational forms carrying these rival ideological values, and whether such an ideological struggle is a source of competition or mutualism among these rival organizational forms. This is our first contribution. The interaction among rival organizational forms does not appear in a vacuum. Especially given the value-laden tensions associated with rival ideologies, the institutional and local environmental conditions are extremely important (Boone and Özcan 2014; Marquis and Huang 2009; Wade, Swaminathan, and Saxon 1998; Wholey and Sanchez 1991). We therefore analyze the interplay between polarization of local ideological preferences, central government regulation, and the dynamics of rival organizational forms. We explore how ideological polarization and state regulation may impact the ecological interaction among rival value-laden organizational forms. By doing so, we explicitly recognize that the type of interdependence between rival organizational forms is likely to be time and place bound (Aldrich 2009; Boone and Özcan 2014). Most studies in this tradition, however, assume that the ideological contextual climate and critical resource base are constant over time. By directly measuring shifts in polarization and resource niches, we show that the mutualistic interdependence between rival ideological forms is suppressed by increasing polarization (as reflected in election outcomes) and reinforced by resource niche expansion (through state regulation). This is our second contribution. We develop specific hypotheses that are tested in the Turkish high school community context in the 1971–1998 period. We analyze how religious Islamic–secular ideological polarization in Turkish society affects the growth at the district level of two school forms carrying oppositional ideologies: religious high schools that provide Islamic education (so-called Imam Hatip schools) and cosmopolitan high schools that provide Western-based secular education in foreign languages (so-called Anatolian schools). We investigate the conditions that affect the nature of interdependencies between organizational forms that carry this pair of rival ideologies. So, we seek to provide a richer understanding of the mechanisms and implications of ideological contention for the organizational dynamics of the Turkish high school educational system. The empirical setting we study is particularly interesting for our purposes for two reasons. First, the Turkish Republic is characterized by an old ideological tension involving the role of religion, particularly of Islam, in public life. In the twentieth century, ideological pressure was gradually built up to increase the role of Islam in public life, and to develop a kind of Turkish–Islamic synthesis. However, this triggered opposition from adherents of “Western”-style ideologies, leading to deep ideological polarization in Turkish society. As education is an extremely important vehicle for the diffusion of ideologies and values in societies (Nielsen and Hannan 1977), the Turkish high school educational system was and still is a major battlefield (Güven 2005), where this ideological polarization materializes via the establishment of Islamic and secular high schools.1 These rival high school dynamics are driven by both central government regulation trying to control high school densities and diversity, and local demand for different forms of high school education as reflected by ideological preferences. Given that data are available at the Turkish district level, we have both substantial cross-sectional and longitudinal variation, which offers the opportunity to examine the interplay of local ideological environments with high school organizational ecologies in the context of local communities. Such local variation is essential (though seldom available; see Marquis and Huang [2009, 1223]) to evaluate the contingent nature of the consequences of nationwide public educational policies. In addition, our setting allows us to substantiate Freeman and Audia’s (2006) plea for studies that focus on explaining the composition of different organizational forms (in casu high schools) in particular locales (in casu districts). Second, during the period of our study, the Turkish context features three major high school types that prepare students for higher education that nicely map onto the ideological resource space: regular high schools in the ideological “center” (which is the dominant secular school type established by Ataturk in 1924) and two oppositional types at both extremes, where Imam Hatip schools target the Islamic “tail” and the cosmopolitan schools are squarely located in the secular “tail.” As these school types all compete for the same resources (i.e., students), the setting is ideal to explore the blind spot in the ecology of ideology literature referred to above (Simons and Ingram 2004). Moreover, the resource space shifted substantially during the period of our study because only as of 1976 were female students allowed to study at Islamic high schools. This fundamental regulatory change moved up the carrying capacity for Islamic high schools substantially, and altered the nature of competition between Islamic and cosmopolitan schools, as now the education of females was at stake. Before developing our hypotheses, we first introduce the details of our setting. Islamic and Cosmopolitan High Schools in Turkey Growing Islamic–secular polarization in Turkey in the twentieth century The history of modern Turkey starts with the establishment of the Republic in 1923, with Mustafa Kemal Ataturk as the founding president. As of 1923, Ataturk tried to transform the remains of the Ottoman Empire into a modern and Western-oriented state, based on secular principles, with reforms that unified education, abolished Sharia, and introduced a modern civil code inspired by the Swiss example. Because the Kemalist modernization project implied a major revolution in the context of a formerly conservative and traditional Islamic society, Islamic opposition was mobilized from the very beginning (Zurcher 1993). This opposition eventually gave rise to the emergence of so-called political Islam in the second half of the twentieth century (Güven 2005; Öniş 1997; Sakallıoğlu 1996). During the single-party regime of the Republican People’s Party (RPP), the bureaucracy and army took control of all aspects of public affairs (Weiker 1981). Because of increasing contention and repression, Turkey shifted to a multi-party system in 1945. In this era, the population became politically mobilized, demanding to be involved in deciding on the nation’s policies (Weiker 1981). The 1950 elections resulted in the victory of the Democratic Party (DP), which criticized RPP to be coercive and not sensitive to the needs and traditions of the people, particularly freedom of religious activity (Ayata 1996). Although the DP remained loyal to the basic principles of secularism, they emphasized religion as a “necessary social cement for the cohesion of society” (Ayata 1996, 43) in their political discourse. The DP introduced voluntary Islamic courses into primary schools, restored the Arabic instead of the Turkish call to prayer, and established Imam Hatip (i.e., preacher) schools and a Faculty of Divinity at Ankara University. During the 1950s, the state elite was very critical of the DP government’s policies (Heper and Keyman 2006). The DP government ended in 1960 with a military coup. The DP was closed down and banned from politics. The National Unity Committee, consisting of army leaders, prepared a liberal-democratic constitution, which was approved by a referendum. This 1960 constitution paved the way for a period of liberalization, which ended with the prolonged clash of many ideologies, especially in the second half of the 1960s and the 1970s. (Landau 1997). In 1970, the National Order Party (NOP), the first openly pro-Islamic party of the Turkish Republic, was founded. In comparison to the DP, the NOP introduced a more aggressive discourse in terms of the role of religion in public life and education, and rejected any type of secularism that could be perceived as being hostile toward the Islam (Landau 1997). The NOP was banned in 1971 because of violating the secularist principles written in the constitution. The reason for political/social unrest was not only religious-secular polarization, but also the emergence of several extreme groups, both left and right wing (Zurcher 1993). To end this state of chaos, the military took over, again, in 1971. However, this military intervention did not end the chaos. Successive governments failed to effectively deal with both economic and political problems. The ideological spectrum widened even more with the foundation of many political parties and social movements, representing all kinds of ideologies, ranging from extreme left to extreme right, including religious sects and Kurdish separatists (Özbudun 1996). This period cumulated in another military coup in 1980. Like the earlier coup, the aim of the 1980 military intervention was to defend the Kemalist principles. However, this coup was also different from the previous ones in several aspects. First of all, the leaders of the military coup put restrictions on individual, group, and organizational freedom (e.g., limiting the role of labor unions), because they perceived the 1961 constitution to be too liberal in light of the ideological polarization in Turkish society. Second, they realized that religion can be a unifying force against fragmentation and contestation caused by the rise of leftists, socialists, radical religious sects, ultra-nationalists, and Kurdish separatists (Ahmad 1991; Pak 2004; Salt 1995). Following the coup, the military government took several measures, including a reform of the education system, to blend religion and national sentiments in an attempt to create a more homogeneous society (Yavuz 1997). They opened up new Qur’anic schools and made religious courses compulsory (Yavuz 1997). This ideology initiated the development of the so-called “Turkish Islamic Synthesis” (Güven 2005). This ideology was supported by the center-right Motherland Party, which took over government after the military. This ideology began to be disseminated through the educational system and media to the masses (Yavuz 1997), which eventually led to the victory of the Islamic Welfare Party in the elections of 1995, and the appointment of the first Islamic Prime Minister (Necmettin Erbakan) of the Turkish Republic. At the other side of the ideological spectrum, the secular groups interpreted these developments as a revolt against Kemalism (Yavuz 1997), which resulted in an escalation in tension and antagonism between both factions (Güven 2005; Heper and Güney 2000). The secular counter-movement resonated well not only among the traditional defenders of the secular Kemalist heritage such as state bureaucrats, retired army officers, secular journalists, and Kemalist organizations, but also among adherents of a new wave of Westernization favoring pluralism, freedom of expression, and social equality (Güven 2005). These platforms challenged the “anti-democratic religious tradition and impinged on emerging ideas of expanding freedoms and a socially diverse and pluralistic form of modernization” (Güven 2005, 203). The increasing role of the Islamic religion in public life, and particularly in education, was regarded as a threat to these treasured secular values. This is witnessed by the establishment of several counter-movement organizations in this era, such as the Association of Kemalist Thought, the Association of Kemalist Women, and the Support of Modern Life Association. All these controversies ultimately led to increasing ideological polarization, dividing Turkey into two very distinct factions: a religious/Islamic and a secular/Kemalist one. The Turkish high school educational system This polarization and contention between ideologies reverberated throughout the educational sector (Güven 2005). Particularly the high school system was an important arena, as here the seeds are sown for the development of the future elite of the country. One of the major strategies of the Islamic movement was the creation of a counter-elite by strengthening the Islamic stream in the educational system (Güven 2005). Two main subgroups can be distinguished: general and vocational/technical high schools. The former educate students for further study in higher education, such as universities, whereas the latter are orientated at specific vocations. In the current study, we focus on the regular, Anatolian and Imam Hatip schools, because they (1) are the most prevalent ones, (2) prepare students for higher education and entry into the elite, and (3) were engaged in strong ideological rivalry during the period of our study.2 These three types of general high schools were most salient in the period under study; the other high school forms (i.e., science high schools, private high schools, and Anatolian Imam Hatip high schools) are less prevalent and/or originated more recently. Indeed, including these latter high school types in the models does not have any effect on the analyses (available upon request). Within our time window, those who wanted to have a (standard) high school education had to decide between attending one of these three types of high schools. Because of the ideological contention, the educational system experienced many changes in regulations and restrictions, particularly with respect to the position of Islam in education. The aim of Ataturk was to provide one unified and secular educational system for all Turkish citizens (through the regular high schools), which was codified in the Law of Unification of Educational Instruction in 1924. During the twentieth century, however, the number of Imam Hatip high schools expanded to such an extent that duality in education is, by now, a fact of life (Güven 2005). The roots of Imam Hatip or Islamic high schools go back to Medresetü-l Eimmeti vel Hutaba, an educational institution founded in 1913 aimed at educating imams and preachers. The madrasah were first abolished as a result of the Law of Unification in 1924, and later re-established in order to educate enlightened imams and preachers who were loyal to the Kemalist ideology. Due to lack of students, the schools were closed again in 1930. In 1949, Imam Hatip high schools were opened to offer 10-month courses, monitored by the Republican People’s Party. It was the Democratic Party that finally established full-fledged Imam Hatip high schools in 1951. Even though the “official aim”3 of Imam Hatip high schools still was the education of imams and preachers, their curricula contained just as much arts and science classes as did regular high schools. As a result, these high schools provided an educational alternative for the children of Islamic and conservative families, and therefore gained popularity all around the country. A first important regulation that positively affected the expansion of Imam Hatip high schools was the 1973 change in the law. Before the 1973 law change, the official purpose of Imam Hatip schools was “educating imams and preachers for religious services.” The 1973 law expanded the official mission of Islamic schools, including “the preparation of students for higher education.” A very fundamental second regulation was the one of 1976 that allowed Imam Hatip high schools to admit female students, even though women cannot be employed as preachers or imams. This increased the “niche” for Islamic high schools substantially, as now conservative Islamic families felt comfortable to send their girls to school (Bozan 2007; Çarkoğlu and Toprak 2006).4 This regulation represents the hallmark of the transformation of Imam Hatip high schools from Islamic vocational high schools to general educational institutions. The number of Imam Hatip high schools increased rapidly from 71 in 1971 to 249 by 1980. Many of these schools were first established by means of public donations, and staffed afterward by the Ministry of Education. Imam Hatip high schools could strengthen their position even more after the regulation in 1982, which allowed their graduates to apply to any faculty program they would like after taking the university entrance exam. Even though these regulatory changes were initiated by the government, they can be seen as tactical moves of the Islamic faction in society to strengthen their position in the high school arena. Secularists perceived the expansion of both the mission and number of Imam Hatip high schools as a threat because a growing number of young people were educated to become not an imam or a preacher, but instead “potential agents of Islamic traditionalism and a potential source of votes for traditional-fundamentalist politicians” (Sezer 1993, 23–24). The graduates of Imam Hatip high schools step by step entered into elite functions in many sectors of Turkish society (Dorian 1997). For example, Recep Tayyip Erdoğan, the current president, graduated from an Imam Hatip high school before he obtained his university degree. The growing duality between secular and Islamic education reinforced the ideological polarization in society. All this brought the issue of Islamic schools onto the political agenda again. Specifically, in 1997, the military strongly opposed the Islamic policy of the government, arguing that “the threat against secularism was now more serious than formerly” (Esposito 2000, 6). In 1997, the military put the government under pressure to restore the restrictions for the graduates of Imam Hatip high schools with respect to the university entrance exam.5 Cosmopolitan, or Anatolian, high schools are positioned at the other extreme side of the ideological spectrum, as their educational curriculum embodies secularist values and includes education in foreign languages. The first cosmopolitan high schools were established as alternatives to foreign (e.g., French, German, or Italian) high schools that provided high-quality and foreign language education in Turkey. The state was suspicious that students of foreign schools would be influenced too much by foreign values, and would not sufficiently be taught about their own history and culture. Therefore, the government established Turkish high schools in 1955, which provide education in foreign languages and which accept their students based on the Nationwide High School Entrance Exam. These Anatolian high schools can be seen as equivalent to the gymnasium in Germany or the Netherlands, in terms of their student acceptance criteria, curriculum, and success. Anatolian high schools claim that, by giving education in foreign languages, they open up students’ minds, helping them communicate and be in touch with the rest of the world much more easily. All course material consisted of books imported from Western countries. The students who were accepted to these schools were the most competitive ones, and the level of education was more challenging than in other types of high schools in the period we analyze. Only one compulsory religious course is taught in cosmopolitan high schools. This course (of one hour per week) is called “Religion and Morality,” and is not designed to introduce students to a religious worldview. Figures 1A and B show the density of Islamic, cosmopolitan, and regular high schools over time. Figure 1. View largeDownload slide A. The density of regular high schools, 1971–1998. Figure 1. View largeDownload slide A. The density of regular high schools, 1971–1998. Figure 1. View largeDownload slide B. The density of Islamic and cosmopolitan high schools, 1971–1998. Figure 1. View largeDownload slide B. The density of Islamic and cosmopolitan high schools, 1971–1998. Islamic versus cosmopolitan high schools are located at the opposite sides (or tails) of the ideological spectrum, with regular schools positioned in the middle (or center). Both rival forms have strongly dissimilar ideologies, but share a similar resource base by targeting the country’s potential student population. The Turkish high school system constitutes a zero-sum game, ceteris paribus; enrollment of a student in one school form goes at the expense of other forms. Islamic and cosmopolitan high schools might not directly attract (potential) students from each other, but both aim to enlarge their student base by competing for center students who would otherwise opt for regular schools. Predicting the Growth of Islamic and Cosmopolitan High Schools Formal organizations play a prominent role in affecting social and political change by advocating new ideological agendas or by defending the status quo (Barnett and Woywode 2004; McCarthy and Zald 1977; Olzak and West 1991). Organizations are carriers of ideologies that are often designed to diffuse and reinforce preferences regarding social outcomes, and the way these outcomes can be obtained (Simons and Ingram 2004). Particularly, schooling is an organized activity consisting of organizations that are highly infused with ideologies, and the schooling system is of major strategic importance to political elites committed to changing the relationship of the individual to the state or polity (Nielsen and Hannan 1977). The demand and supply of specific types of schooling is, therefore, strongly driven by ideologies that prevail among different interrelated actors in society, such as parents choosing education for their children and political elites designing educational policies, and private actors trying to influence political elites via all sorts of social movement organizations, also to obtain schooling that fits with their ideology (Güven 2005). Indeed, in the Turkish context, local communities have opportunities to push for the establishment of new high schools in their communities. Formally, the final decision about high school founding is the authority of the Ministry of Education. However, whether a new high school will be founded and which type of high school is going to be established can be influenced by the local community and the local governor, implying that often the decision of the Ministry of Education depends on the needs of the district. For instance, if the local community donates the necessary buildings and equipment for a new high school, then the government tends to give permission for the founding of the high school, subsequently assigning teaching staff. In that case, the donating individual(s) have the right to make the decision as to which type of high school is going to be established. Hence, the founding process of high schools in Turkey involves both a top-down and a bottom-up process, implying that both local conditions and local collective actions are important drivers of new high school formation. We argue that especially contention between rival ideologies will spur the growth of organizational forms that carry these rival ideologies—in our case, high schools. Polarization increases the saliency of individual identities, in-group solidarity, and between-group competition and conflict (Esteban and Ray 1994; Olzak and West 1991; Sherif and Sherif 1953; Simmel 1923). Polarization between rival ideologies is a very potent mobilizing force, spurring actors to protect their ideological position by organized activity (Meyer and Staggenborg 1996; Olzak and West 1991). Polarization increases the saliency of ideological positions by raising the awareness of ideological markers and identities (Olzak and West 1991). Increasing polarization and contention between opposed ideological groups also creates the threat that the existing benefits may be taken away, and/or new harms might be imposed by the opposing group or the state (Almeida 2003; Goldstone and Tilly 2001; Jasper 1997; Tilly 1978; Van Dyke and Soule 2002). The growing importance of political Islam triggered secular countermovements that mobilized the secular public opinion against the anti-secular rhetoric of the former (Güven 2005; Heper and Güney 2000). During this period, especially Imam Hatip high schools became highly politicized, becoming an instrument for political actors to signal the community’s standing regarding the religiosity scale. As a result, the need for Imam Hatip high schools and the parents’ willingness to send their children to these schools became an issue of taking a side with either position on the ideological spectrum. That is why the type of high school founded in a district is a signal to the local community, which resulted in competition between religious and center-right parties in order to get Islamic votes (Yavuz 1997). On top of this, during this period of contestation, many civil society associations and organizations were formed around Imam Hatip high schools and Islamist political parties, examples being İlim Yayma Cemiyeti, Ensar Vakfi, and ÖNDER. These organizations “provided scholarships to Imam Hatip students, fund dormitories, organize conferences and lobby on behalf of the Imam Hatip community” (Sarfati 2015, 685). These organizations not only support Imam Hatip high schools, but also “act as intermediaries between the [Imam Hatip] schools and Islamist political formations by creating shared identities among students, graduates, conservative business circles and activists and by [fostering] a feeling of trust within religiously conservative communities that the politicians will recognize” (Sarfati 2015, 685). Given their curricula and the atmosphere in these schools, Anatolian high schools represent the other end of the secular-religious dimension (Divarcı Çakmaklı, Boone, and van Witteloostuijn 2017). Therefore, for secular families these schools became the best alternative. During this period, graduates of Anatolian high schools also established associations in support of Anatolian high schools. In addition to that, many other organizations emerged to form a countermovement vis-à-vis the Islamic movement, such as the Association of Kemalist Thought (Atatürkçü Düşünce Derneği) and the Association for the Support of Contemporary Living (Çağdaş Yaşamı Destekleme Derneği). In the decision-making process regarding high school foundings, the movement and countermovement organizations act as mobilizing forces, by targeting the state in an attempt to influence public policy (Johnson, Agnone, and McCarthy 2010). As a result of increasing polarization, both polarized groups felt threatened (Almeida 2003). If the recipients of this threat are well organized, the level of collective action and resistance to the threat is expected to be higher (Almeida 2003; Gould 1991; Jenkins 1983; Walton and Seddon 1994). Such a cycle of movement and countermovement creates a political opportunity structure favorable to those organizational forms that carry the associated rival ideological values, as long as the underlying tensions are not “resolved” by the state (Meyer and Staggenborg 1996). Hypothesis 1: Ideological Islamic–secular polarization increases the growth of both Islamic and cosmopolitan high schools. Several scholars have contributed to the so-called ecology of ideology by analyzing the ideological drivers of inter-organizational dynamics (Barnett and Woywode 2004; Boone and Özcan 2014; Minkoff 1994, 1995; Simons and Ingram 2004; Vermeulen 2013; Wade, Swaminathan, and Saxon 1998). Some claim and find that ideological similarity between organizational forms is associated with strong competition (Barnett and Woywode 2004; Vermeulen 2013), while others report that similar ideologies will help each rival form in forwarding their own ideology (Boone and Özcan 2014; Simons and Ingram 2004). Simons and Ingram (2004) argue that these contradicting findings are due to the fact that some studies confound ideological similarity with high resource overlap between organizational forms. When resource overlap is low and ideological similarity is high, interdependency between forms is mutualistic. Then, forms sharing similar ideologies enjoy the benefit of a strong esprit de corps (Barnett and Woywode 2004; Vermeulen 2013), without fighting for the same resources: “organizations can generate ideological mutualism by sharing experience and information, providing patronage, money, encouragement, and legitimacy, and by cooperating in efforts at political influence” (Ingram and Simons 2000, 34–35). This claim is corroborated in studies regarding the founding and mortality rates of different types of cooperatives in Israel during the twentieth century (Ingram and Simons 2000; Simons and Ingram 2004). The density of credit cooperatives reduces the failure rate of worker cooperatives (Ingram and Simons 2000), and increases the founding rate of kibbutzim and moshavim, both agricultural cooperatives (Simons and Ingram 2004). The density of kibbutzim, in turn, suppresses the mortality rate of worker cooperatives (Ingram and Simons 2000). All these organizational forms share a similar “socialist” ideology, without competing in the same niche for resources. In a similar vein, Minkoff (1995, 1997) found that the development of the civil rights organizations was an important driver of the subsequent growth of the number of women’s organizations between 1955 and 1985. However, when resource overlap is large, the potential benefit of an esprit de corps cannot compensate for the cost of strong competition. Simons and Ingram (2004) find that kibbutzim and moshavim affect each other competitively: the density of the former has a negative effect on the founding rate of the latter and vice versa, notwithstanding the fact that both forms share a common ideology. This is due to both forms’ reliance on similar resources for survival, such as land, distributors, and participants (Simons and Ingram 2004). In their study of the Viennese newspaper industry for the period 1918–1938, Barnett and Woywode (2004) also report that competition among newspapers is strongest when newspapers share similar ideologies. Competitive relationships have also been reported for populations that have dissimilar ideologies, but do not compete for the same resources directly. Due to competition in the ideological arena, organizations might try to coerce organizations that carry rival ideologies by withholding resources (Ingram and Simons 2000). Banks, the ideal-typical representatives of the capitalist ideology, have had a strong influence on the cooperative, socialist form of economic activity in Israel. Simons and Ingram (1997) report that kibbutzim that relied on banks for financial input were more likely to adopt capitalist principles, such as employing hired labor. Bank density indeed increased the failure rate of worker cooperatives that were perceived to be close to the core of the socialist ideology (Ingram and Simons 2000). Similarly, the number of corporations was associated with a decrease in the founding rate of kibbutzim and moshavim cooperatives (Simons and Ingram 2004). Scholars have speculated that when organizations’ ideologies/identities are dissimilar and the resource overlap is high, there will be conflicting forces impacting their relationship. Although “diametric opponents are hardly friends … they at least enhance each other’s ideological identity by their sharp contrast” (Barnett and Woywode 2004, 1461). This might strengthen a population’s identity, cohesion, and appeal, partially offsetting strong competition in the resource arena. Simons and Ingram (2004) suggest that ideological rivalry among populations might promote collective action and the populations’ capacity to build institutions. Boone and Özcan (2014) claim that this relationship is moderated by local-level variables. In their study of county-level founding processes of cooperatives in the US ethanol industry, they show that local conditions spur and facilitate collective action among farmers to establish cooperatives in response to the local diffusion of corporations. It still remains unclear, however, what outcomes these conflicting forces will generate in different circumstances. Following the arguments of Simons and Ingram (2004), Olzak and West (1991), Boone and Özcan (2014), and Barnett and Woywode (2004), we predict as a baseline that the interdependence between populations (groups) that carry rival ideologies will be mutualistic. Rivalry between dissimilar ideologies makes the ideologies extremely sharp and salient, which mobilizes the adherents of these ideologies to diffuse their values and preferences. This is consistent with Meyer and Staggenborg’s (1996, 1638) proposition that “when movements effectively create or exploit events, they are likely to encourage countermovement mobilization at the same time that they advance their own causes.” The salience argument contends that ideological rivalry “galvanizes” the opposing groups, and so spurs the formation of formal organizations that carry these ideological values (Olzak and West 1991, 459). Such organization building mobilizes even more collective action (Tilly 1978), such that the establishment of one organizational form will reinforce the dynamics of the rival one, leading to mutualistic interdependence. Hypothesis 2a: The growth of Islamic high schools increases with the number of cosmopolitan high schools. Hypothesis 2b: The growth of cosmopolitan high schools increases with the number of Islamic high schools. However, such a mutualistic relationship is unlikely to hold indefinitely as the focus of the ideological and resource rivalry eventually becomes the “destruction of the foe” (Barnett and Woywode 2004, 1461). Vermeulen (2013), for instance, analyzed to what extent the context affects the form of competition and/or collaboration among different ideologies. His analysis of Turkish immigrant organizations in Germany and the Netherlands reveals that an open, supportive environment spurs the activities of different ideological groups and also encourages intra- and inter-ideological cooperation. Following this logic, we focus on two context moderators—one repressing mutualism (i.e., increases in ideological polarization), and the other reinforcing mutualism (i.e., increases in the carrying capacity). First, we expect that progressing ideological polarization will eventually repress the positive impact of the growth of one organizational form on the growth of the rival form. Although ideological polarization is predicted to spur mobilization initially (hypothesis 1), further polarization may well intensify competition between organizational forms associated with these extreme positions. In fact, the ultimate aim of such mobilization and competition is to repress the rival ideology to become the dominant one. The political opportunity structure has an important impact on the ideology that will turn out to be the “winner” (Meyer and Staggenborg 1996). For instance, in Vienna from 1918 to 1938, the right-wing newspapers eventually won the contest, destroying the clerical center as “the environment of political opportunity shifted in such a way as to give apparent support for the claims of the extreme right” (Barnett and Woywode 2004, 1489). We expect that especially the mutualistic effect of the cosmopolitan schools on their Imam Hatip counterparts will be repressed by Islamic–secular polarization, because the rapid expansion of the Islamic high schools in the 1970s triggered a cosmopolitan countermovement that gained momentum in a political environment that gradually became more favorable to secular and Western values in the second half of the 1980s. We hence expect a negative interaction effect of Islamic–secular polarization and the density of cosmopolitan schools on the growth of Islamic high schools. As interaction effects are symmetric, such a negative interaction implies that we expect that the positive effect of ideological polarization on the growth of Imam Hatip schools decreases, or may even become negative, with the density of cosmopolitan high schools. Hypothesis 3: The positive impact of the number of cosmopolitan high schools on the growth of Islamic schools decreases (or becomes negative) when Islamic–secular polarization increases. Second, what happens when carrying capacity shifts? As discussed in detail by Soule and King (2008), social movement organizations adopt more specialized tactical repertoires when competition increases. The law of 1976, which allows female students in Imam Hatip schools, can be seen as a tactical move of Islamic schools, channeled through governmental intervention, to keep up competition with cosmopolitan schools. This law increased the high school population’s carrying capacity, especially in the Islamic high school niche. Many Islamic parents who formerly were not willing to send their daughters to regular or other high schools felt now at ease to enroll their girls in Islamic schools. Competition between cosmopolitan and Islamic schools entered a new arena—that is, competition for female education. This not only shifted the opportunity structure for both organizational forms upward (Meyer and Staggenborg 1996), but also revitalized the mobilization of both rival ideologies. This expansion of the resource niche and renewed mobilization are expected to extend the sustainability of the mutualistic interdependence between cosmopolitan and Islamic high schools. Hypothesis 4a: The growth of Islamic high schools increases with the number of cosmopolitan high schools especially when the proportion of female students in Islamic high schools to all female students increases. Hypothesis 4b: The growth of cosmopolitan high schools increases with the number of Islamic high schools especially when the proportion of female students in Islamic high schools to all female students increases. Data, Measures, and Methods Our dataset covers the period from 1971 to 1998, for 80 districts in Turkey. Even though full-fledged Islamic high schools were already founded in the 1950s, they only became part of the general high school educational field in the beginning of the 1970s when their official mission was expanded to include the preparation of students for higher education. Therefore, we start our analysis as of 1971, the year of the second military coup in Turkey that marks this new era. We end our analysis in 1998 because a new regulation in 1998 fundamentally transformed the high school system, implying that the data before and after 1998 are not comparable. The high school data are obtained from the National Education Statistics that are published by the Prime Ministry State Institute of Statistics. The core of our dataset consists of the number of Islamic (i.e., Imam Hatip), cosmopolitan (i.e., Anatolian), and regular high schools in Turkey that existed between 1971 and 1998. Following Ruef (2004) and Zhou and van Witteloostuijn (2010), our dependent variable is the first difference of the density (i.e., the number) of Islamic and cosmopolitan high schools at the district level in time t+1 and time t, respectively.6 As schools were seldom closed in our time period, this first difference variable is almost tantamount to the number of foundings. For modeling purposes, we therefore can treat this variable as a count, replacing the limited number of negative first difference values by zeros.7 Our main independent variables consist of the density of Islamic and cosmopolitan high schools at the district level. We lagged densities in order to avoid simultaneity. To test the niche expansion H4a and 4b, we employ the (lagged) proportion of female students in the Islamic high schools to all female students. For the measurement of polarization, we take the polarization index proposed by Esteban and Ray (1994). This index is based on the Gini coefficient, which accounts for both the size of the subgroups and distance: P=k∑i=1N∑j=1Nπi1+απj|yi−yj| in which k is the scaling factor, π denotes the size of the groups (i.e., the proportion of individuals belonging to group i), |yi−yj| gives the distance between groups i and j on a salient continuum (in this study, the religious-secular dimension), and α refers to the degree of polarization sensitivity that takes a value between 0 and 1.6 (Esteban and Ray 1994). When α is zero, P equals the Gini index. The lower α, the more P measures inequality (stressing alienation due to distance); the higher α, the more sensitive P is to polarization between large homogeneous subgroups (Stark, Hyll, and Behrens 2010). This measure captures the intuition that polarization is high when there is high homogeneity within each group, a high degree of heterogeneity between groups (i.e., distance), and a small number of groups of significant size (Esteban and Ray 1994). We set k = 1, and α = 0.8. By doing so, we follow most researchers by giving a balanced weight to the size of and the distance between the subgroups. We compute P using the outcomes of parliamentary elections at the district level to measure religious-secular polarization in the local community. The proportion of each political party’s votes denotes the size of different groups that vote for each political party in each district. To obtain an indicator of distance, the second author (who has Turkish nationality) classified political parties on a secular/religious scale ranging from 1 (extremely secular) to 10 (extremely religious) (available upon request). In order to score each political party’s religiosity on this scale, the second author analyzed each political party’s official program (for this period) to determine its emphasis on religion, and consulted several articles/books on the religious orientation of political parties during this period (Cizre 2008; Cizre-Sakallıoğlu and Çınar 2003; Göle 1997; Keyman 2007; Öniş 2007; Sakallıoğlu 1996; Somer 2007; Yavuz 1997). This classification was double-checked by two Turkish political scientists. The data on the parliamentary elections are obtained from the Turkish Prime Ministry State Institute of Statistics. The years without elections were filled in with the election outcomes from the beginning of the term. Our main control variables are the density of regular high schools at the district level, the (logarithm of) human population at the district level, human population growth at the district level, national GDP in real prices, and the percentage of government expenditures devoted to education at the country level, all lagged to avoid simultaneity. Regular high school density is included to capture the interaction with the high schools in the ideological center. The other variables are introduced to control for the high school population’s carrying capacity. We also add the percentage of votes that religious parties received at the district level (taking into account the votes received by the parties with a score of 10). We include dummies for the different official regions in Turkey, assigning 1 if the district is in that region and 0 otherwise. We add a dummy variable to control for the years following a territorial change of a few districts (1 for the years after the change, and 0 otherwise). Finally, to control for the important regulations of 1976 (that allowed female students to enroll in the Islamic high schools) and 1997 (that restored the university restrictions on the graduates of the Islamic high schools), we enter time dummies (0 before and 1 after the regulation). Since our dependent variable is a count measure, a Poisson process provides a natural baseline for modeling the growth of high schools (Hannan and Freeman 1987, 1989). However, Poisson regression should not be applied in the case of overdispersion (McCullagh and Nelder 1989). Overdispersion does not affect the coefficient estimates, but results in underestimated standard errors. Negative binomial regression adequately deals with this issue by adding a stochastic error component, following a gamma distribution, to the model (Cattani, Pennings, and Wezel 2003). The structure of the data is a pooled cross-section and time series (districts * years). Pooled data generally exhibit autocorrelation. Following Baron, Hannan, and Burton (2001), we use Liang and Zeger’s (1986) method of generalized estimating equations (GEE), which generalizes quasi-likelihood estimation to the panel data context, and which is a very flexible way to deal with clustered data. We assume an exchangeable autocorrelation structure. Because the observations within the districts cannot be assumed to be independent, we report Huber-White robust standard errors. A drawback of GEE is that summary goodness-of-fit statistics are problematic because the residuals of these models are correlated (Ballinger 2004; Zorn 2001). We therefore report Wald chi-square statistics, which test the null hypothesis that all regression coefficients are equal to zero. As Ballinger (2004, 146) warned, this statistic is not suited to compare the goodness of fit of alternative models. All models are estimated using the XTGEE routine of version 10.0 of STATA. Findings Descriptive statistics and bivariate correlations between the variables are reported in tables 1 and 2. Tables 3, 4, and 5 provide GEE negative binomial regression estimates of the growth of Islamic, cosmopolitan, and regular high schools, respectively. We add analyses with regular high school growth as the dependent variable to shed further light on the underlying cross-form interactions (see below). Table 1. Descriptives of Major Variables under Study Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 View Large Table 1. Descriptives of Major Variables under Study Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 Mean S.D. Min Max Growth of Islamic schools 0.22 0.66 0 9 Growth of cosmopolitan schools 0.20 0.72 0 12 Growth of regular schools 0.71 1.31 0 21 Ln (District population) (t-1) 13.15 0.71 10.13 16.01 District population growth (t-1) 0.02 0.21 −0.90 0.17 Educational expenditure (t-1) 11.28 2.34 7.23 15.10 GDP (t-1) 218.44 72.42 110.65 363.25 Votes Islamic parties 11.12 10.29 0 51.60 Regulation 1976 0.83 0.38 0 1 Regulation 1997 0.08 0.27 0 1 Ideological polarization 0.95 0.18 0.41 1.43 Density Islamic schools (t-1) 3.91 3.44 0 21 Density cosmopolitan schools (t-1) 0.83 1.73 0 23 Density regular schools (t-1) 15.77 15.99 0 171 Proportion females in Islamic schools (t-1) 0.06 0.08 0 1 View Large Table 2. Correlations of Major Variables under Studya 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − an = 1943; * = p < 0.05 (two-tailed). View Large Table 2. Correlations of Major Variables under Studya 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Growth of Islamic schools − 2. Growth of cosmopolitan schools 0.002 − 3. Growth of regular schools 0.08* 0.03 − 4. Ln (District population) (t-1) 0.15* 0.28* 0.38* − 5.District population growth (t-1) −0.004 −0.004 −0.01 0.02 − 6. Educational expenditure (t-1) −0.05* −0.21* 0.14* −0.05* −0.001 − 7. GDP (t-1) −0.08* 0.37* −0.10* 0.06* 0.001 −0.32* − 8. Votes Islamic parties −0.005 0.17* −0.06* −0.01 0.07* 0.10* 0.49* − 9. Regulation 1976 0.08* 0.12* −0.10* 0.09* −0.005* −0.40* 0.59* 0.16* − 10. Regulation 1997 −0.09* 0.44* −0.11* 0.01 −0.01 −0.46* 0.55* 0.31* 0.13* − 11. Ideological polarization 0.17* −0.17* 0.07* 0.08* 0.004 0.20* −0.37* −0.17* 0.01 −0.24* − 12. Density Islamic schools (t-1) 0.01 0.37* 0.15* 0.59* −0.02 −0.26* 0.43* 0.11* 0.40* 0.17* −0.19* − 13. Density cosmopolitan schools (t-1) 0.07* 0.59* 0.23* 0.53* 0.01 −0.14* 0.48* 0.25* 0.19* 0.36* −0.17* 0.54* − 14. Density regular schools (t-1) 0.15* 0.40* 0.38* 0.76* 0.001* −0.09* 0.27* 0.09* 0.23* 0.08* −0.02 0.67* 0.78* − 15. Proportion females in Islamic schools (t-1) −0.07* 0.23* −0.06* 0.01 0.02 −0.20* 0.68* 0.49* 0.32* 0.43* −0.40* 0.42* 0.30* 0.13* − an = 1943; * = p < 0.05 (two-tailed). View Large Table 3. Negative Binomial GEE Regression Estimates of the Growth in the Number of Islamic High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < . 01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 3. Negative Binomial GEE Regression Estimates of the Growth in the Number of Islamic High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −3.48*** −4.65*** −7.61*** −9.40*** −9.38*** (0.71) (0.73) (1.44) (1.63) (1.90) District territory change 0.63*** 0.71*** 0.31+ 0.11 0.04 (0.15) (0.14) (0.19) (0.30) (0.31) Mediterranean region 0.09 0.18 0.45+ 0.07 0.16 (0.19) (0.16) (0.23) (0.23) (0.23) Eastern Anatolia −0.75*** −0.57*** −0.42* −0.64*** −0.53** (0.18) (0.15) (0.20) (0.17) (0.20) Aegean region 0.24+ 0.26+ 0.46** 0.17 0.25 (0.13) (0.14) (0.18) (0.18) (0.19) Southeast Anatolia −0.81*** −0.68** −0.65+ −0.79** −0.81*** (0.25) (0.24) (0.34) (0.28) (0.25) Central Anatolia 0.09 0.19* 0.64*** 0.28+ 0.46* (0.09) (0.10) (0.16) (0.15) (0.20) Black Sea region 0.12 0.17 0.48* 0.17 0.24 (0.13) (0.13) (0.19) (0.16) (0.17) Ln (District population) (t-1) 0.37*** 0.34*** 0.52*** 0.65*** 0.66*** (0.05) (0.05) (0.12) (0.13) (0.14) District population growth (t-1) −0.50 −0.74 −1.39* −1.53* −1.36* (0.52) (0.66) (0.56) (0.62) (0.66) Educational expenditure (t-1) −0.22*** −0.22*** −0.25*** −0.23*** −0.24*** (0.02) (0.02) (0.03) (0.03) (0.03) GDP (t-1) −0.01*** −0.01*** −0.005** −0.005** −0.003 (0.001) (0.001) (0.002) (0.002) (0.002) Votes Islamic parties 0.06*** 0.05*** 0.04*** 0.03*** 0.03*** (0.01) (0.005) (0.01) (0.01) (0.01) Regulation 1976 1.13*** 0.84*** 0.97*** 0.97*** 0.94*** (0.18) (0.20) (0.17) (0.18) (0.18) Regulation 1997 −2.91*** −3.09*** −3.18*** −3.08*** −5.20*** (0.76) (0.74) (0.70) (0.73) (1.42) Proportion females in Islamic schools (t-1) −0.88 0.76 1.24 1.67 −2.61 (1.22) (1.10) (1.45) (1.39) (1.95) Ideological polarization 1.36*** 1.76*** 1.88*** 1.58*** (0.30) (0.34) (0.34) (0.34) Density Islamic schools (t-1) −0.15*** −0.16*** −0.15*** (0.03) (0.03) (0.02) Density cosmopolitan schools (t-1) 0.13*** 0.90*** −0.22** (0.03) (0.24) (0.08) Density regular schools (t-1) −0.0005 0.01 0.002 (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) * Ideological polarization −1.02*** (0.29) Density cosmopolitan schools (t-1) * Proportion females in Isl. schools (t-1) 3.17*** (0.69) Wald Chi2 979.98*** 1674.57*** 1079.91*** 1210.77*** 1280.96*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < . 01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 4. Negative Binomial GEE Regression Estimates of the Growth in the Number of Cosmopolitan High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** an = 1943; + = p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 4. Negative Binomial GEE Regression Estimates of the Growth in the Number of Cosmopolitan High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** Model 1 Model 2 Model 3 Model 4 Model 5 Constant −13.38*** −12.86*** −14.71*** −13.09*** −15.28*** (0.61) (0.69) (1.64) (2.57) (1.58) District territory change −0.17*** −0.16*** 0.20+ 0.18+ 0.18 (0.05) (0.05) (0.11) (0.10) (0.13) Mediterranean region −0.02 −0.07 −0.08 −0.03 −0.11 (0.04) (0.05) (0.10) (0.11) (0.09) Eastern Anatolia −0.57*** −0.58*** −0.55*** −0.51** −0.73*** (0.11) (0.12) (0.17) (0.17) (0.12) Aegean region −0.21* −0.22** −0.22* −0.16 −0.22** (0.08) (0.09) (0.09) (0.10) (0.08) Southeast Anatolia −0.35* −0.29* −0.41* −0.28 −0.56*** (0.16) (0.12) (0.19) (0.18) (0.16) Central Anatolia −0.05 −0.09 −0.28* −0.25* −0.53*** (0.09) (0.09) (0.12) (0.10) (0.15) Black Sea region −0.15* −0.15* −0.26* −0.25* −0.35*** (0.07) (0.08) (0.11) (0.11) (0.09) Ln (District population) (t-1) 0.56*** 0.56*** 0.67*** 0.61*** 0.71*** (0.05) (0.05) (0.11) (0.14) (0.11) District population growth (t-1) −2.38*** −1.96* −2.19** −2.12* −2.15* (0.66) (0.79) (0.82) (0.87) (0.84) Educational expenditure (t-1) −0.02 −0.01 −0.003 −0.01 −0.002 (0.02) (0.01) (0.02) (0.02) (0.02) GDP (t-1) 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) (0.002) Votes Islamic parties −0.01 −0.01 −0.02* −0.02** −0.01* (0.005) (0.01) (0.01) (0.01) (0.004) Regulation 1976 0.88 1.01 0.58 0.39 0.71 (0.64) (0.67) (0.55) (0.55) (0.56) Regulation 1997 0.13 0.18 0.79** 0.77** 0.66** (0.18) (0.18) (0.30) (0.30) (0.25) Prop. females in Islamic schools (t-1) 0.16 0.02 −0.65 −0.48 −1.02 (0.32) (0.35) (0.63) (0.60) (0.63) Ideological polarization −0.59+ −0.04 −0.85 −0.14 (0.35) (0.31) (0.51) (0.32) Density Islamic schools (t-1) 0.04** −0.09 0.01 (0.02) (0.08) (0.01) Density cosmopolitan schools (t-1) −0.13*** −0.12*** −0.13*** (0.02) (0.02) (0.03) Density regular schools (t-1) 0.01*** 0.01*** 0.01*** (0.002) (0.003) (0.002) Density Islamic schools (t-1) * Ideological polarization 0.17 (0.11) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.10+ (0.05) Wald Chi2 4472.35*** 4698.58*** 6281.23*** 5492.63*** 6447.41*** an = 1943; + = p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 5. Negative Binomial GEE Regression Estimates of the Growth in the Number of Regular High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Table 5. Negative Binomial GEE Regression Estimates of the Growth in the Number of Regular High Schools, 1971–1998a Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Constant −7.18*** −7.06*** −7.23*** −8.34*** −7.59*** −7.34*** −7.35*** (0.29) (0.29) (0.41) (0.49) (0.47) (0.43) (0.50) District territory change −0.25*** −0.24*** −0.16** −0.23*** −0.17** −0.19** −0.17** (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) Mediterranean region 0.18*** 0.18*** 0.14*** 0.15*** 0.09+ 0.14*** 0.12*** (0.04) (0.04) (0.04) (0.05) (0.05) (0.04) (0.05) Eastern Anatolia 0.11 0.10 0.06 0.10 0.02 0.06 0.04 (0.12) (0.12) (0.13) (0.12) (0.12) (0.13) (0.13) Aegean region 0.17*** 0.17*** 0.15** 0.13** 0.09+ 0.14** 0.13* (0.04) (0.04) (0.05) (0.05) (0.05) (0.05) (0.05) Southeast Anatolia 0.03 0.03 −0.02 −0.01 −0.06 −0.03 −0.04 (0.06) (0.06) (0.06) (0.06) (0.06) (0.06) (0.07) Central Anatolia 0.26*** 0.26*** 0.19*** 0.23*** 0.13*** 0.19*** 0.17*** (0.03) (0.03) (0.05) (0.04) (0.04) (0.04) (0.05) Black Sea region 0.06 0.05 0.01 0.04 −0.04 0.01 −0.01 (0.06) (0.06) (0.06) (0.06) (0.05) (0.06) (0.06) Ln (District population) (t-1) 0.44*** 0.44*** 0.46*** 0.51*** 0.49*** 0.47*** 0.47*** (0.02) (0.02) (0.04) (0.04) (0.04) (0.04) (0.04) District population growth (t-1) −1.56*** −1.58*** −1.66*** −1.17*** −1.72*** −1.69*** −1.68*** (0.26) (0.27) (0.25) (0.36) (0.28) (0.25) (0.25) Educational expenditure (t-1) 0.05*** 0.05*** 0.04*** 0.04*** 0.04*** 0.04*** 0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) GDP (t-1) −0.0001 −0.0005 −0.001 −0.001 −0.001 −0.001 −0.001 (0.001) (0.0005) (0.0005) (0.001) (0.001) (0.001) (0.001) Votes Islamic parties −0.0003 −0.0002 0.001 0.002 0.001 0.001 0.001 (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Regulation 1976 −0.17*** −0.14** −0.12* −0.09+ −0.12* −0.11* −0.11* (0.05) (0.05) (0.05) (0.05) (0.06) (0.05) (0.06) Regulation 1997 −0.48* −0.44* −0.41+ −0.48* −0.46* −0.45* −0.44+ (0.22) (0.22) (0.22) (0.21) (0.22) (0.22) (0.22) Prop. females in Islamic schools (t-1) −0.30 −0.32 −0.66 −1.42** −0.62 −1.69* −0.88 (0.51) (0.49) (0.50) (0.55) (0.50) (0.85) (0.63) Ideological polarization −0.16 −0.18 0.25+ −0.03 −0.17 −0.19 (0.11) (0.14) (0.14) (0.15) (0.14) (0.14) Density Islamic schools (t-1) 0.01 0.12*** 0.01 0.0001 0.01 (0.01) (0.03) (0.01) (0.01) (0.01) Density cosmopolitan schools (t-1) −0.004 0.01 0.15** −0.01 −0.01 (0.01) (0.01) (0.06) (0.01) (0.01) Density regular schools (t-1) −0.002 −0.004** −0.001 −0.002 −0.002 (0.001) (0.002) (0.002) (0.001) (0.002) Density Islamic schools (t-1) * Ideological polarization −0.11*** (0.03) Density cosmopolitan schools (t-1) * Ideological polarization −0.18** (0.06) Density Islamic schools (t-1) * Proportion females in Islamic schools (t-1) 0.14 (0.08) Density cosmopolitan schools (t-1) * Proportion females in Islamic schools (t-1) 0.11 (0.12) Wald Chi2 2718.57*** 2849.94*** 3262.69*** 3127.64*** 4097.77*** 3252.33*** 3323.78*** an = 1943; += p < 0.10 * = p < 0.05 ** = p < 0.01 *** = p < 0.001 (two-tailed); robust standard errors in parentheses. View Large Both types of high schools grow slower in the East and Southeast of Turkey (Marmara being the omitted region). In central Anatolia, Islamic high schools grow faster, and in the Black Sea region cosmopolitan high schools grow slower. District population strongly spurs the growth of both Islamic and cosmopolitan high schools, and population growth hampers their growth. National GDP is negatively related to the growth of Imam Hatip high schools, and positively to the growth of cosmopolitan high schools. This finding is related to the rise of a devout bourgeoisie middle class, whose welfare started to increase over time, and hence did not want to send their children to Imam Hatip high schools anymore. The percentage of people in the district voting on Islamic parties has the expected effect: positive on Islamic and negative on cosmopolitan high schools. As expected, the 1976 regulation strongly enhanced the growth of Islamic high schools, whereas the 1997 regulation reduced its growth in favor of cosmopolitan high schools. Model 2 in table 3 shows a strong positive effect of ideological polarization on the growth of Islamic high schools. This estimate does not change when we add the densities of the three high school types (model 3 in table 3). For cosmopolitan high schools, we find a marginally significant negative effect of ideological polarization that becomes insignificant, however, when we control for the high school type densities (models 2 and 3 in table 4). These findings provide partial support for H1: ideological polarization spurs the growth of Islamic high schools, but not that of cosmopolitan high schools. The estimates of model 3 in both tables 3 and 4 provide support for H2a and 2b: the nature of interdependence between both extreme high school forms is mutualistic. The positive impact of cosmopolitan on Islamic high schools (B = 0.13) is much larger than the other way around (B = 0.04). Model 4 in table 3 reveals that the mutualistic effect of cosmopolitan on Islamic high schools is attenuated by the extent of ideological polarization, as is clear from the negative and significant interaction effect of cosmopolitan high school density and ideological polarization, providing support for H3. Figure 2 visualizes this interaction. This moderation effect of ideological polarization does not appear with respect to the impact of Islamic high school density on cosmopolitan high school growth (model 4 in table 4). Figure 2. View largeDownload slide The interaction effect of polarization and cosmopolitan school density on Islamic school growth Figure 2. View largeDownload slide The interaction effect of polarization and cosmopolitan school density on Islamic school growth In model 5 in both tables 3 and 4, we test whether (relative) niche expansion (i.e., proportion of female students in Islamic high schools) positively moderates the mutualistic interdependence between both forms. Both interaction coefficients are significant: (relative) niche expansion indeed positively reinforces this mutualistic interdependence, which provides evidence for H4a and 4b. Figures 3A and B show both interactions. Figure 3. View largeDownload slide A. The interaction effect of the proportion of female students in Islamic schools to all female students and cosmopolitan school density on Islamic school growth. B. The interaction effect of the proportion of female students in Islamic schools and Islamic school density on cosmopolitan school growth Figure 3. View largeDownload slide A. The interaction effect of the proportion of female students in Islamic schools to all female students and cosmopolitan school density on Islamic school growth. B. The interaction effect of the proportion of female students in Islamic schools and Islamic school density on cosmopolitan school growth Finally, to further explore the center–tail cross-form interaction, and hence to reveal the underlying interactive processes, we estimate the same models to analyze the growth of the regular high schools. District population increases the growth of regular high schools, while the growth of district population has a negative effect on the growth of regular high schools. The 1976 and 1997 regulations have a negative impact, too. Models 2 and 3 in table 5 reveal that neither ideological polarization nor the densities of Islamic and cosmopolitan high schools have a significant impact on the growth of regular high schools. Important for our analyses is the result in model 4 in table 5, which reveals a negative and significant interaction effect of the density of Islamic high schools and ideological polarization. The interaction of cosmopolitan high school density and ideological polarization is also negative and significant (model 5 in table 5): Islamic and cosmopolitan high school densities suppress the expansion of the regular high schools when ideological polarization increases, suggesting strong competition between all forms. Moreover, models 6 and 7 in table 5 show that the (relative) niche expansion variable does not moderate the impact of Islamic or cosmopolitan high school density on the growth of regular high schools. These findings suggest that the expansion of both Islamic and cosmopolitan schools affects the growth of regular high schools, and both school forms compete for the students at the center. Finally, we ran a series of robustness analyses, all available upon request, to explore the extent to which our findings might be sensitive to adding additional control variables and different calculations of the polarization index. First, we added (1) dummy variables for the military coup years (see the section on the history of the Republic), (2) the density of science high schools, (3) the density of private high schools, such as Robert College, and (4) a dummy for districts that have a population above 250,000 (as a proxy for urbanization). Note that the density of the other high school types is low in our time window. Second, we ran analyses with our polarization index calculated using α = 0.4 and α = 1.2, employing linear interpolation to fill in the years without elections. None of these robustness analyses changed the results reported above, increasing our confidence in the results found. Discussion Our main findings can be summarized as follows. Economic forces (i.e., national GDP) positively affect the expansion of cosmopolitan high schools: this high school form appears to follow general trends of economic development. State regulation and the size of the local Islamic faction (i.e., the number of votes on Islamic parties), in contrast, strongly affect the expansion of Islamic high schools. Islamic–secular polarization at the district level has a strong mobilizing impact on the growth of these Islamic high schools over and above the size of the local Islamic faction. Additionally, the interdependence between both organizational forms is mutualistic, on average. Both forms stimulated each other’s growth rate as a result of a sequence of movements and countermovements, starting with the initial rise of Islamic high schools in the late 1970s. The mutualistic effect of the cosmopolitan high schools on the growth of Islamic high schools declines, however, with further increases with Islamic–secular polarization. Mutualism between rival ideology-laden organizational forms is eventually repressed by ideological polarization, ceteris paribus. The expansion of the resource space of Islamic high schools by allowing enrollment of female students, in contrast, reinforces the mutualistic relationships between both rival forms. Moreover, both Islamic and cosmopolitan schools feed on the center, as the former grow at the expense of the latter when ideological polarization increases. This pattern of findings reveals that Islamic–secular polarization shifts the density distribution of high schools with different ideologies such that the number of rival ideology forms grows at the expense of the center. A major underlying driver of this pattern of findings is the subtle balance between two major conflicting forces that operate simultaneously: the mobilizing force of ideological polarization and the repressing force of competition among rival ideologies. The first force makes ideologies and values salient, and creates strong subgroup identities that facilitate collective action (e.g., social movements and/or government regulation). Such collective action subsequently spurs defensive countermovements of the opposing ideology. The second force results from the desire of adherents of specific ideologies to become dominant. The purpose of ideological rivalry is to repress the rival ideology. The interplay between the context’s opportunity structure and endogenous organizational dynamics determines which force dominates at a particular moment in time. As a result of this complex interplay of conflicting forces, we cannot predict ex ante the specific outcome of ideological rivalry (especially in settings where the contention remains unsettled). However, being able to ex post explain and understand the mechanisms behind these processes is scientifically valuable in its own right (Davis and Marquis 2005). Our community ecology approach reveals that formal organizations play an important role in the rivalry among ideologies, as such organizations serve as the vehicles of the diffusion of values in society. This is literally the case for the high schools that we studied. By establishing formal organizations, ideological rivalry is institutionalized in a relatively permanent way, underscoring Tilly’s (1978) observation that organizations are important in prevailing and promoting ideological positions. Our paper contributes to the literature in several ways. First, we add to community ecology and the ecology of ideology. Although several ecologists have stressed the importance of the study of interdependencies between different organizational forms, the number of such studies is still limited. Especially studies that combine relations between (classes of) organizations (so-called functional complementarities) with spatial differentiation are rare. In their review of the field, Freeman and Audia (2006, 165) propose that “simultaneous consideration of spatial and functional differentiation and complementarity can provide more nuanced views of the mechanism underlying density effects.” This is exactly what we sought to do in the present study by modeling the interplay between the growth of different high school forms against the background of local ideological differences (and state regulation). Related to this, we contribute to the ecology of ideology by analyzing the dynamics of organizational forms with dissimilar (opposing) ideologies that compete for similar resources, which has been identified as a major gap in the literature (see also Boone and Özcan [2014]). Although Simons and Ingram (2004) and Barnett and Woywode (2004) already speculated about the importance of the conflicting forces central to our argument, ours is one of the first studies that tries to unravel the conditions driving their relative impact. Most studies in this tradition implicitly assume that the ideological contextual climate and critical resource base are constant over time. By directly measuring shifts in ideological polarization and resource niches, we show that the mutualistic interdependence between rival ideological forms is suppressed by ideological polarization (as reflected in election outcomes) and reinforced by resource niche expansion (through state regulation). Second, by introducing the moderating effects of ideological polarization and state regulation, we contribute to “new organizational synthesis” research that “goes beyond studying movements as movement organizations, aiming instead to trace how movements shape organizational fields, and vice versa” (Schneiberg, King, and Smith 2008, 636). We do not study movements directly, but we do analyze the dynamic interplay between ideological struggles and the expansion of organizational forms carrying specific ideologies. Although it is generally accepted that organizations play a central role in facilitating many kinds of collective actions, much less is known about whether, when, and how social tension and conflict (such as polarization) affect the life chances of organizations (Olzak and West 1991). We show that ideological polarization is a major mobilizing force, having long-lasting effects on the specific development of organizational fields of major importance, such as high school education. Our study reveals that the dynamics of contention, with its cycles of opportunity, mobilization, and countermobilization, produces historical and path-dependent trajectories of different kinds of organizations (Marquis and Lounsbury 2007; Schneiberg, King, and Smith 2008). Third, our study analyzes the impact of culture and policy by incorporating both polarization and regulation, revealing that “policy makers should take underlying cultural factors into account when anticipating the intended effects of policy … Though policy may reflect majority will, there may be significant pockets of resistance or social movements that condition the intended effect … Thus, identifying the heterogeneous cultural beliefs and their organization may be important when implementing policy” (Marquis and Huang 2009, 1242). Turkish educational high school policies have generated many unexpected consequences as a result of these heterogeneous and polarized beliefs. Islamic high schools were regulated following the military coup of 1971 for the purpose of controlling religious Islamic education in Turkey. However, their increasing popularity spurred their numbers as well as the number of rival ideological high school forms. Female emancipation policies enlarged the ideological battlefield to competition in education. All this did not settle the issue of the role of Islam in society, but is likely to have, in contrast, nurtured ideological polarization in Turkey, which the government in the 1970s wanted to avoid in the first place. Although we focused on the Turkish high school system, we think our community ecology framework can be applied to understand organizational dynamics in many other settings as long as ideological polarization is salient. For example, growing polarization can also be observed in the United States among Democrats and Republicans with very clear-cut distinctions in terms of several issues concerning healthcare, civil rights, immigrants, and so on (Jacobson 2016). This, in turn, is associated with the establishment of new organizations, such as Breitbart and BuzzFeed, representing either side of the polarized debate. Another example is the emergence of Islamic banking as an insurgent logic against conventional banking in Turkey and other Islamic countries, fueled by ideological competition between secular and traditional Islamic factions in society (Boone and Özcan 2016). Our study is not without limitations, which provide opportunities for further research. First, the relationships between ideological polarization, state regulation, and expansion of high school forms are reciprocal in nature. Regulation is not exogenous, but affected by the distribution of ideological values. Moreover, different school forms diffuse different ideologies among a society’s population, which subsequently affects ideological polarization. As a first step, we decided to treat state regulation and ideological polarization as context variables that affect the growth of high school forms in a variety of ways. In future research, it would be interesting to build models that focus on Islamic–secular polarization as the phenomenon to explain, emphasizing this two-way channel of reciprocal causation. Second, we only focused on the three major high school forms in Turkey. More recently, interesting new high school types emerged that are probably also related to the continuing ideological struggle in Turkey. For instance, hybrid high school types were established, such as the Anatolian Imam Hatip high schools. These schools blend the quality of non-religious Anatolian high schools (which we labeled cosmopolitan high schools) with the intensive Islamic education of Imam Hatip high schools. Most courses in these schools are also taught in foreign languages, as in cosmopolitan-secular high schools. At the other side of the ideological spectrum, more and more private cosmopolitan high schools emerged, where students have to pay for top-quality education. This suggests that the frontline of the ideological struggle is gradually shifting to high-quality education of the future elite in Turkish society. Notes 1 Since Turkey is a secular country, officially all school types are secular. In this paper, the secular-religious distinction does not refer to the formal status, but to the curriculum of, and atmosphere in, the schools (Erkan and Akçayöz 2003; Pak 2004). 2 Note that all high school types analyzed in this study are public schools funded by the state. None of them required a tuition fee. 3 Official aim refers to what is written in the mission statement of Imam Hatip high schools. 4 We acknowledge that not all conservative families, especially Kurdish ones, were comfortable with sending their children to Imam Hatip schools—not only their girls, but also their boys. To this day, there are still informal madrasah in the Kurdish area, which are run by local religious leaders, serving as an alternative to Imam Hatip schools. The Kurdish conservative families historically have approached the Imam Hatip schools with suspicion, mostly because of their mistrust in their secular curriculum. Unfortunately, we do not have any other data on the number of students attending madrasah instead of high schools (as these madrasah are informal, this data is not available). Moreover, because of its politicized nature, the government can decide to establish Imam Hatip high schools in Kurdish areas as a signal, irrespective of the number of students that go to Imam Hatip high schools. So, we only focus on the number of high schools established, and not on the number of students that go to these high schools. 5 These restrictions relate to the coefficients that are applied as weights to the high school grades of students in order to determine whether a student is allowed to enter the university. For instance, if an Imam Hatip high school graduate wanted to enter the Theology faculty, then her high school grades were multiplied by 0.5; however, if she wanted to enter any other faculty at the university, then her high school grades were multiplied by 0.2. Obviously, this restriction was a disadvantage for Imam Hatip high school graduates to enter university in comparison to other high school graduates. 6 We employ the number of schools instead of the number of students because, in the short run, mobilization is related to the emergence of different school types while affecting the number of students in the long run. 7 Unfortunately, no information is available on the limited number of school closing events. About the Authors Christophe Boone is Professor of Organization Theory and Behavior at the Faculty of Applied Economic Sciences at the University of Antwerp (Belgium). His research interests focus on the dynamics of organizational populations in local communities, the antecedents and consequences of team and organizational diversity, CEO values and cognition, and the neuroeconomics of decision-making. Anıl Divarcı Çakmaklı is an Assistant Professor in the Department of Management at Istanbul Şehir University, Turkey. She holds master’s degrees in Economics from Istanbul Technical University and in Philosophy and Economics from Erasmus University Rotterdam, and a PhD in Management from the University of Antwerp. Her research focuses on organization theory and strategy. Arjen Van Witteloostuijn is Dean of the School of Business and Economics of the Vrije Universiteit Amsterdam in the Netherlands. He is associate editor of the Journal of International Business Studies and reviewing editor of Cross Cultural and Strategic Management, and sits on the editorial board of the British Journal of Management and Industrial and Corporate Change. He has published about 170 articles in international journals, including the American Journal of Sociology, American Sociological Review, Academy of Management Journal, Academy of Management Review, American Journal of Political Science, Journal of Public Administration Research and Theory, Organization Science, PLoSONE, and Strategic Management Journal. References Ahmad , Feroz . 1991 . “ Politics and Islam in Modern Turkey .” Middle Eastern Studies 27 ( 1 ): 3 – 21 . Google Scholar CrossRef Search ADS Aldrich , Howard . 2009 . “ Lost in Space, Out of Time: Why and How We Should Study Organizations Comparatively .” Research in the Sociology of Organizations 26 : 21 – 44 . Google Scholar CrossRef Search ADS Almeida , P. D. 2003 . “ Opportunity Organizations and Threat-Induced Contention: Protest Waves in Authoritarian Settings .” American Journal of Sociology 109 ( 2 ): 345 – 400 . Google Scholar CrossRef Search ADS Ayata , Sencer . 1996 . “ Patronage, Party, and State: The Politicization of Islam in Turkey .” Middle East Journal 50 : 40 – 56 . Ballinger , Garry A. 2004 . “ Using Generalized Estimating Equations for Longitudinal Data Analysis .” Organizational Research Methods 7 : 127 – 50 . Google Scholar CrossRef Search ADS Baron , James N. , Michael T. Hannan , and M. Diane Burton . 2001 . “ Labor Pains: Organizational Change and Employee Turnover in Young, High-Tech Firms .” American Journal of Sociology 106 : 960 – 1012 . Google Scholar CrossRef Search ADS Barnett , William P. , and Michael Woywode . 2004 . “ From Red Vienna to the Anschluss: Ideological Competition among Viennese Newspapers during the Rise of National Socialism .” American Journal of Sociology 109 : 1452 – 99 . Google Scholar CrossRef Search ADS Boone , Christophe , and Serden Özcan . 2014 . “ Why Do Cooperatives Emerge in a World Dominated by Corporations? The Diffusion of Cooperatives in the US Bio-Ethanol Industry, 1978–2013 .” Academy of Management Journal 57 : 990 – 1012 . Google Scholar CrossRef Search ADS Boone , Christophe , and Serden Özcan . 2016 . “ Ideological Purity vs. Hybridization Trade-Off: When Do Islamic Banks Hire Managers from Conventional Banking? ” Organization Science 27 : 1380 – 96 . Google Scholar CrossRef Search ADS Bozan , Irfan . 2007 . Devlet ile Toplum Arasında Bir Okul: İmam Hatip Liseleri, Bir Kurum: Diyanet İşleri Başkanlığı. Istanbul : TESEV Publishing . Çarkoğlu , Ali , and Binnaz Toprak . 2006 . Degisen Turkiye’de Din, Toplum ve Siyaset . Istanbul : TESEV Publishing . Cattani , Gino , Johannes M. Pennings , and Filippo C. Wezel . 2003 . “ Spatial and Temporal Heterogeneity in Founding Patterns .” Organization Science 14 : 670 – 85 . Google Scholar CrossRef Search ADS Cizre , Ümit . 2008 . Secular and Islamic Politics in Turkey: The Making of the Justice and Development Party. London and New York: Routledge . Cizre-Sakallıoğlu , Ümit , and Menderes Çınar . 2003 . “ Turkey 2002: Kemalism, Islamism, and Politics in the Light of the February 28 Process .” South Atlantic Quarterly 102 ( 2 ): 309 – 32 . Google Scholar CrossRef Search ADS Davis , Gerald F. , and Christopher Marquis . 2005 . “ Prospects for Organization Theory in the Early Twenty-First Century: Institutional Fields and Mechanisms .” Organization Science 16 : 332 – 43 . Google Scholar CrossRef Search ADS Divarcı Çakmaklı , Anıl , Christophe Boone , and Arjen van Witteloostuijn . 2017 . “ When Does Globalization Lead to Local Adaptation? The Emergence of Hybrid Islamic Schools in Turkey, 1985–2007 .” American Journal of Sociology 122 ( 6 ): 1822 – 68 . Google Scholar CrossRef Search ADS Dorian , Jones . 1997 . Higher Education in Turkey: Times Higher Education Supplement. Issue 1277, p. 10. Erkan , Rüstem , and Harika Akçayöz . 2003 . “ Siyasi İslam Tartışmaları Açısından İmam–Hatip Lisesi Öğrencilerinin Demokratik Tutum ve Davranışlarının İncelenmesi .” Cumhuriyet Üniversitesi Sosyal Bilimler Dergisi 27 ( 2 ): 174 – 201 . Esposito , John L. 2000 . “Islam and Secularism in the Twenty-First Century.” In Islam and Secularism in the Middle East , edited by John Esposito and Azzam Tamini . London : Hurst and Company , 1 – 12 . Esteban , Joan , and Debraj Ray . 1994 . “ On the Measurement of Polarization .” Econometrica: Journal of the Econometric Society 62 : 819 – 51 . Google Scholar CrossRef Search ADS Freeman , John , and Pino G. Audia . 2006 . “ Community Ecology and the Sociology of Organizations .” Annual Review of Sociology 32 : 145 – 69 . Google Scholar CrossRef Search ADS Goldstone , Jack , and Charles Tilly . 2001 . “Threat (and Opportunity): Popular Action and State Response in the Dynamic of Contentious Action.” In Silence and Voice in the Study of Contentious Politics , edited by R. Aminzade , D. McAdam , E. Perry , W. Sewell , S. Tarrow , and C. Tilly , 179 – 94 . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Göle , Nilüfer . 1997 . “ Secularism and Islamism in Turkey: The Making of Elites and Counter-Elites .” Middle East Journal 51 : 46 – 58 . Gould , Roger . 1991 . “ Multiple Networks and Mobilization in the Paris Commune, 1871 .” American Sociological Review 56 : 716 – 29 . Google Scholar CrossRef Search ADS Güven , Ismail . 2005 . “ The Impact of Political Islam on Education: The Revitalization of Islamic Education in the Turkish Educational Setting .” International Journal of Educational Development 25 : 193 – 208 . Google Scholar CrossRef Search ADS Hannan , Michael T. , and John Freeman . 1987 . “ The Ecology of Organizational Founding: American Labor Unions, 1836–1985 .” American Journal of Sociology 92 : 910 – 43 . Google Scholar CrossRef Search ADS Hannan , Michael T. , and John Freeman . 1989 . Organizational Ecology. Cambridge, MA : Harvard University Press . Heper , Metin , and Aylin Güney . 2000 . “ The Military and the Consolidation of Democracy: The Recent Turkish Experience .” Journal of Armed Forces and Society 26 : 635 – 58 . Google Scholar CrossRef Search ADS Heper , Metin , and Fuat Keyman . 2006 . “ Double-Faced State: Political Patronage and the Consolidation of Democracy in Turkey .” Middle Eastern Studies 34 ( 4 ): 259 – 77 . Google Scholar CrossRef Search ADS Hinings , Chris R. , and Roy Greenwood . 2002 . “ Disconnects and Consequences in Organization Theory? ” Administrative Science Quarterly 47 : 411 – 21 . Google Scholar CrossRef Search ADS Ingram , Paul , and Tal Simons . 2000 . “ State Formation, Ideological Competition, and the Ecology of Israeli Workers’ Cooperatives, 1920–1992 .” Administrative Science Quarterly 45 : 25 – 53 . Google Scholar CrossRef Search ADS Jacobson , G. C. 2016 . “ Polarization, Gridlock, and Presidential Campaign Politics in 2016 .” ANNALS of the American Academy of Political and Social Science 667 ( 1 ): 226 – 46 . Google Scholar CrossRef Search ADS Jasper , James . 1997 . The Art of Moral Protest. Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Jenkins , J. Craig . 1983 . “ Resource Mobilization Theory and the Study of Social Movements .” Annual Review of Sociology 9 : 527 – 53 . Google Scholar CrossRef Search ADS Johnson , E. W. , J. Agnone , and J. D. McCarthy . 2010 . “ Movement Organizations, Synergistic Tactics and Environmental Public Policy .” Social Forces 88 ( 5 ): 2267 – 92 . Google Scholar CrossRef Search ADS Keyman , E. Fuat . 2007 . “ Modernity, Secularism and Islam: The Case of Turkey .” Theory, Culture & Society 24 ( 2 ): 215 – 34 . Google Scholar CrossRef Search ADS Landau , Jacob M. 1997 . “Turkey between Secularism and Islamism.” Jerusalem Letter/Viewpoints 352. Liang , Kung-Yee , and Scott L . Zeger . 1986 . “ Longitudinal Data Analysis Using Generalized Linear Models .” Biometrika 73 : 13 – 22 . Google Scholar CrossRef Search ADS Marquis , Christopher , and Michael Lounsbury . 2007 . “ Vive la résistance: Competing Logics and the Consolidation of U.S. Community Banking .” Academy of Management Journal 50 : 799 – 820 . Google Scholar CrossRef Search ADS Marquis , Christopher , and Zhi Huang . 2009 . “ The Contingent Nature of Public Policy and the Growth of U.S. Commercial Banking .” Academy of Management Journal 52 : 1222 – 46 . Google Scholar CrossRef Search ADS McCarthy , John D. , and Mayer N . Zald . 1977 . “ Resource Mobilization and Social Movements: A Partial Theory .” American Journal of Sociology 82 : 1212 – 41 . Google Scholar CrossRef Search ADS McCullagh , Peter , and John A. Nelder . 1989 . Generalized Linear Models , 2nd ed . Boca Raton, FL : CRC Press . Google Scholar CrossRef Search ADS Meyer , David S. , and Suzanne Staggenborg . 1996 . “ Movements, Countermovements, and the Structure of Political Opportunity .” American Journal of Sociology 101 : 1628 – 60 . Google Scholar CrossRef Search ADS Minkoff , Debra C. 1994 . “ From Service Provision to Institutional Advocacy: The Shifting Legitimacy of Organizational Forms .” Social Forces 72 ( 4 ): 943 – 69 . Google Scholar CrossRef Search ADS Minkoff , Debra C. 1995 . Organizing for Equality: The Evolution of Women’s and Racial-Ethnic Organizations in America, 1955–1985 . New Brunswick, NJ : Rutgers University Press . Minkoff , Debra C. 1997 . “ The Sequencing of Social Movements .” American Sociological Review 62 : 779 – 99 . Google Scholar CrossRef Search ADS Nielsen , François , and Michael T . Hannan . 1977 . “ The Expansion of National Educational Systems: Tests of a Population Ecology Model .” American Sociological Review 42 : 479 – 90 . Google Scholar CrossRef Search ADS Olzak , Susan , and Elizabeth West . 1991 . “ Ethnic Conflict and the Rise and Fall of Ethnic Newspapers .” American Sociological Review 56 : 458 – 74 . Google Scholar CrossRef Search ADS Öniş , Ziya . 1997 . “ The Political Economy of Islamic Resurgence in Turkey: The Rise of the Welfare Party in Perspective .” Third World Quarterly 18 : 743 – 66 . Google Scholar CrossRef Search ADS Öniş , Ziya . 2007 . “ Conservative Globalists versus Defensive Nationalists: Political Parties and Paradoxes of Europeanization in Turkey .” Journal of Southern Europe and the Balkans Online 9 ( 3 ): 247 – 61 . Google Scholar CrossRef Search ADS Özbudun , Ergun . 1996 . “ Turkey: How Far from Consolidation? ” Journal of Democracy 7 ( 3 ): 123 – 38 . Google Scholar CrossRef Search ADS Pak , Soon-Yong . 2004 . “ Cultural Politics and Vocational Religious Education: The Case of Turkey .” Comparative Education 40 : 321 – 41 . Google Scholar CrossRef Search ADS Ruef , Martin . 2004 . “ The Demise of an Organizational Form: Emancipation and Plantation Agriculture in the American South, 1860–1880 .” American Journal of Sociology 109 : 1365 – 1410 . Google Scholar CrossRef Search ADS Sakallıoğlu , Ümit C. 1996 . “ Islam-State Interaction in Turkey .” International Journal of Middle East Studies 28 : 231 – 51 . Google Scholar CrossRef Search ADS Salt , Jeremy . 1995 . “ Nationalism and the Rise of Muslim Sentiment in Turkey .” Middle Eastern Studies 31 ( 1 ): 13 – 27 . Google Scholar CrossRef Search ADS Sarfati , Yusuf . 2015 . “ Islamic Schools in Modern Turkey: Faith, Politics, and Education .” British Journal of Middle Eastern Studies 42 ( 4 ): 684 – 86 . Google Scholar CrossRef Search ADS Schneiberg , Marc , Marissa King , and Thomas Smith . 2008 . “ Social Movements and Organizational Form: Cooperative Alternatives to Corporations in the American Insurance, Dairy and Grain Industries .” American Sociological Review 73 : 635 – 67 . Google Scholar CrossRef Search ADS Sezer , Duygu B. 1993 . State and Society in Turkey: Continuity and Change? Santa Monica, CA : Rand Publications . Sherif , Muzaver , and Carolyn W. Sherif . 1953 . Groups in Harmony and Tension. New York : Harper & Row . Simmel , Georg . 1923 . Conflict and the Web of Group Affiliation. Glencoe, IL : Free Press . Simons , Tal , and Paul Ingram . 1997 . “ Organization and Ideology: Kibbutzim and Hired Labor, 1951–1965 .” Administrative Science Quarterly 42 : 784 – 813 . Google Scholar CrossRef Search ADS Simons , Tal , and Paul Ingram . 2004 . “ An Ecology of Ideology: Theory and Evidence from Four Populations .” Industrial and Corporate Change 13 : 33 – 59 . Google Scholar CrossRef Search ADS Somer , Murat . 2007 . “ Moderate Islam and Secularist Opposition in Turkey: Implications for the World, Muslims and Secular Democracy .” Third World Quarterly 28 ( 7 ): 1271 – 89 . Google Scholar CrossRef Search ADS Soule , Sarah , A and Brayden G. , King . 2008 . “ Competition and resource partitioning in three social movement industries .” American Journal of Sociology 113 : 1568 – 610 . Google Scholar CrossRef Search ADS Stark , Oded , Walter Hyll , and Doris A. Behrens . 2010 . “ Gauging the Potential for Social Unrest .” Public Choice 143 : 229 – 36 . Google Scholar CrossRef Search ADS Tilly , Charles . 1978 . From Mobilization to Revolution. Reading, MA : Addison Wesley . Van Dyke , Nella , and Sarah Soule . 2002 . “ Structural Social Change and the Mobilizing Effect of Threat: Explaining Levels of Patriot and Militia Organizing in the United States .” Social Problems 49 ( 4 ): 497 – 520 . Google Scholar CrossRef Search ADS van Witteloostuijn , Arjen , and Christophe Boone . 2006 . “A Resource-Based Theory of Market Structure and Organizational Form.” Academy of Management Review 31 : 409 – 26 . Google Scholar CrossRef Search ADS Vermeulen , Floris . 2013 . “ Mutualism, Resource Competition and Opposing Movements among Turkish Organizations in Amsterdam and Berlin, 1965–2000 .” British Journal of Sociology 64 ( 3 ): 453 – 77 . Google Scholar CrossRef Search ADS Wade , Jim B. , Anand Swaminathan , and Michael-Scott Saxon . 1998 . “ Normative and Resource Flow Consequences of Local Regulations in the American Brewing Industry, 1845–1918 .” Administrative Science Quarterly 43 : 905 – 35 . Google Scholar CrossRef Search ADS Walton , John , and David Seddon . 1994 . “Food Riots Past and Present.” In Free Markets and Food Riots: The Politics of Global Adjustment , edited by J. Walton and D. Seddon , 23 – 54 . Oxford : Blackwell . Google Scholar CrossRef Search ADS Weiker , Walter F. 1981 . The Modernization of Turkey: From Ataturk to the Present Day . London: Holmes and Meier Pub . Wholey , Douglas R. , and Susan M . Sanchez . 1991 . “ The Effects of Regulatory Tools on Organizational Populations .” Academy of Management Review 16 : 743 – 67 . Google Scholar CrossRef Search ADS Yavuz , M. Hakan . 1997 . “ Political Islam and the Welfare (Refah) Party in Turkey .” Comparative Politics 30 : 63 – 82 . Google Scholar CrossRef Search ADS Zhou , Chaohong , and Arjen van Witteloostuijn . 2010 . “ Institutional Constraints and Ecological Processes: Evolution of Foreign-Invested Enterprises in the Chinese Construction Industry, 1993–2006 .” Journal of International Business Studies 41 : 539 – 56 . Google Scholar CrossRef Search ADS Zorn , Christopher J W 2001 . “ Generalized Estimating Equation Models for Correlated Data: A Review with Applications .” American Journal of Political Science 45 : 470 – 90 . Google Scholar CrossRef Search ADS Zurcher , Erik J. 1993 . Turkey: A Modern History . London : I. B. Tauris . Author notes We thank the participants of the 2010 Organizational Ecology conference in Helsinki, of the ACED conference on team and community diversity (Antwerp, 2010), and of the seminars at Copenhagen Business School (November 2010) and Sabanci University (January 2011), especially Özgecan Koçak and Serden Özcan, for stimulating us to develop the ideas presented in the paper and for their useful comments. This project is financed by the Flemish Science Foundation (FWO) in the realm of the Odysseus program of the Flemish government. © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Social ForcesOxford University Press

Published: Feb 2, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off