Islamic Primary Schools in the Netherlands

ABSTRACT During the last 20 years of the 20th century, Islamic primary schools were founded in the Netherlands thanks to its constitutional “freedom of education” (which allows state-funded religious schools), its voucher system (each school receives the same amount of money per pupil), and school choice by parents. This essay gives some background information about the Dutch system of religious schools and the history of Dutch Islamic schools.1 I address four aspects of Islamic schools: (a) contradictions around the quality of education in Islamic schools; (b) attitudes and values of pupils and parents in Islamic schools, deviating from the broader Dutch society; (c) serious administrative problems around establishing and running Islamic schools, due to the nonexistence of Islamic Dutch elites and teachers; and (d) negative relations between the current Islam religion and educational performance in modern societies.


Introduction
Founding and maintaining religious schools in the Netherlands is relatively quite simple, since the great political compromise of 1917 (the socalled school pacification). Since that year, the equal funding for all nonpublic schools but also equal quality standards and curricula for public and nonpublic schools have been a part of the constitution. This was a constitutional right not only for one dominant religion, but for all religions in the Netherlands: Roman Catholic and several Orthodox-Protestant churches (and even religiously neutral schools). This constitutional right to found schools is, however, older: the liberal constitution of 1848 said that "teaching was free," provided that the teachers had ability and state supervision. It also stated that sufficient public education should be provided by the national government. The Dutch constitutional "freedom of education" is thus foremost the right by religious and nonreligious foundations and associations (not churches) to found schools within the state parameters of educational quality and supervision. The parental right to choose a school for their children is a consequence of this founder Note. Only schools with pupils in the last grade of primary school (age 11/12).

The Dutch system of religious schools
Because there is sufficient literature on the Dutch system of religious schools (Dijkstra et al., 2004;Glenn & de Groof, 2004), I summarize here only its most important characteristics, which are relevant for Islamic schools. Although the Dutch system is an extreme case, many European countries have some form of state-funded religious schools, combined with a kind of voucher system and school choice by parents (Dronkers, 2004;Smyth, Lyons, & Darmody, 2013). As a consequence, Islamic state-funded schools might also be established in other European countries because all have substantial Islamic minorities.
Religious schools have to be founded and owned by associations or foundations, if they want to receive financial grants from the state. As a consequence, religious schools are not owned nor run by churches or mosques; they have independent legal identities. For instance, in the late 20th century a very orthodox Catholic bishop could not remove the adjective Catholic from more liberal Catholic schools in his diocese, because the courts ruled they were owned and run by independent institutions. The legal position of religious schools in the Netherlands resembles that of charter schools in the United States, although the former have a much higher stability than the later, and also less degree of freedom in their operation (national curriculum, inspectorate, nation-wide tests and exams, salaries, ban of fees). Accordingly, differences between public and religious schools in the Netherlands are more cultural than material, more school-climate differences than factual operation differences.
Religious state-funded schools have the same national curriculum as public schools, but they can add religious teaching to the slate and have a bit more freedom in applying the standards of that national curriculum inside their schools. The quality criteria of the national educational inspectorate apply to all state-funded schools, public, religious, or not. National curriculum is the basis for the nation-wide test at the end of primary school and the national exams at the end of secondary school. Participation in this test at the end of primary education is obligatory and essential for obtaining access to the higher tracks in secondary education (grammar school and higher secondary school). The same holds for the national exam at the end of secondary education. Without a successful exam, entrance into university or vocational college is virtually impossible. Nationwide publications of quality of all primary and secondary schools by the inspectorate and daily papers have become common since the early 21st century. These publications of school quality both of the Inspectorate and daily papers use as their core the comparable school results of testing at the end of primary school and the national exams at the end of secondary schools.
There are national standards for teacher training and qualification. Payment and application of these standards is mandatory for all public and religious school in order to get state funding. If these standards are not met, closure of the school will follow, as I discuss as follows in the case of Islamic secondary schools.
One of the interesting unintended consequences of this Dutch version of "freedom of education" is that elite private schools, operating without state funding but with parental fees, disappeared in the mid-20th century. Given the high standard of both public and religious schools and their low costs (from parental perspectives), the demand for elite private schools dwindled away. The few private schools, which still existed at the start of the 21st century, are institutions to help school dropouts with well-to-do parents to obtain a (higher) secondary school diploma, but do not have high prestige.
Mandatory schooling ends at the age of 18, although from 16 years old onwards schooling can be combined with work (apprenticeship schooling) and a minimum diploma is required before one can leave the educational system. Grading is based on performance in class: social promotion to higher grades hardly exists (only in the lowest vocational track of vocational education). Stepping down to a lower track in secondary education is an option after performance failure, although higher educated parents will prefer repetition of the same grade within the same track (avoiding downward mobility).
Although the Dutch do not use the term often, they have a voucher system: Each nonpublic school gets the same grants per pupils as the public schools get per pupil. As a consequence, if pupil numbers of a school drop, the amount of state money for that school decreases as well (although with some delay). Substantial sponsoring of schools by churches, firms, or other agencies will be deducted from the state grant, although the strict rules were somewhat relaxed in 2014. Parents using state-funded schools are not obliged to pay a school fee and the nonobligatory parental school fee should not be used for core activities of the school (teaching; building). Schools (both public and nonpublic) get additional funding for pupils with less educated parents in the same way. Because most parents at Islamic schools have very low education, Islamic schools are an important beneficiary of this additional money.
School choice by parents within municipalities is more or less free, depending on whether schools want to increase their number of pupils. Nonpublic schools cannot be obliged by the public authorities to increase their number of pupils above a funding threshold. The housing for primary schools is a task of the municipality and can be a reason for the delay of opening new religious schools or their further expansion. Housing of the growing and diminishing schools is often a hindrance for a quick adjustment of school to the changing demand by parents and demographic population changes. But municipalities have never been successful in thwarting infinitely school founders with enough involved parents, nor delaying enlargement of the housing of popular schools. A common strategy of schools with increasing numbers of pupils is to establish a new location of their school elsewhere in the municipality under the same school board. Municipalities tried often to restrict parental school choice, because it increases school segregation. They try to form catchment areas of several schools, which parents can choose. Most attempts have failed in the long run due to the constitutional impossibility to force these schemes on all schools and parents.
In the early 20th century most nonpublic schools were Catholic, Protestant-Christian, or neutral nonpublic. Their school boards were closely related to the Protestant and Catholic political parties, which were members of nearly all Dutch governments since 1919. The number of Catholic and Protestant-Christian schools and their pupils increased until the 1960s and became stable, despite the widespread secularization of Dutch society. By the beginning of the 21st century more than 50% of the adult population belonged to no church or religious association. Yet the variation of religious schools increased, mainly by the foundation of smaller orthodox Protestant schools like Reformed Orthodox, Reformed Liberated, and Evangelical. These smaller orthodox Protestant schools lacked political backing by influential national Christian-Democrat parties, but still grew during the second half of the 20th century. Simultaneously, the number of anthroposophical schools increased strongly, though anthroposophy is not a religion. This reflects the increasing influence of nonreligious views on education.

The history of Dutch Islamic schools
During the last 20 years of 20th century, two types of non-Christian religious schools were founded: Hindu and Islamic. 2 Hinduism came to the Netherlands via their former colony Suriname (North Latin-American) where they were imported from British-India after the abolishment of slavery in the 19th century (as in the British parts of the Caribbean). Before or shortly after the independence of Suriname, substantial numbers of Hindus migrated to the Netherlands, thanks to their Dutch passports. Among the Surinamese migrants were Muslims, who also originated from British-India (34,000). 3 But the number of these Surinamese migrants was dwarfed by the large numbers of Islamic migrants coming since the 1960s as guest workers from Turkey (285,000) and Morocco (296,000), of which the majority settled themselves with their families in Western Europe. Later smaller groups of migrants came from Afghanistan (31,000), Iraq (27,000), and Somalia (20,000).
The first attempts to found Islamic primary schools came in 1980, and in 1988 the first two Islamic schools opened their doors. In 2014 there are 46 Islamic primary schools, three of which do not yet have pupils in their last grade. There were two Islamic secondary schools founded as well (in Rotterdam and Amsterdam), but both were closed by the Dutch educational authorities due to their insufficient quality and administrative problems. After this failure the Christian School Foundation in Rotterdam decided in 2014 to establish an Islamic secondary school under its legal umbrella and auspices. 4 All Islamic primary schools have both male and female pupils, a bit less of the latter. Co-education is standard in the Netherlands and only-boys or only-girls schools do not exist. Also classes in Islamic school are mixed, but some lessons (swimming, gymnastics) will not be mixed.
There is no easy explanation for the successful foundation of Islamic primary schools and the failure of secondary schools. Perhaps founding a secondary school is more difficult, because they must contain a number of hierarchal tracks (from grammar school to vocational training) with more specialized teaching, more administrative burdens, and more pupils. Another explanation might be that the two secondary schools were founded with primarily orthodox Islamic religious aims, while the primary schools had more often also an emancipation aim. When the focus of the school founding was on combating the educational disadvantage of Muslim children, the Dutch local authorities were rather more accommodating than when the focus was on the religious character of the school (Driessen & Merry, 2006).

Do Islamic schools meet minimum school quality?
The political compromise of 1917 was also an effort to create equal conditions and equal quality of all schools, public and nonpublic. One element was the creation of the same final examination of secondary education and equal rules for the transition from the common primary school into one of the tracks of secondary school (from vocational schooling to grammar school). The score on a final test at the end of primary school and the teacher's recommendation about the most fitting track are the essentials for admission to the higher tracks. The aggregates of these final scores (measures of language and math) and the teacher recommendations per school are public and published by national and local media. Table 1 shows these aggregated scores for all schools in 2014 on a nationally used standardized test. This national test has an average score of 535. Its minimum is 501 (equivalent for lowest level of vocational schooling) and its maximum is 550 (equivalent for grammar school). Islamic schools have on average the lowest score: 531, but that is above the official minimum score. The Dutch inspectorate applies 527-529 as lower band for schools with more than 60% low educated parents. Schools with lower scores and without a prospect of quick improvement will be called publically a weak or very weak school by Dutch inspectorate. These schools will lose pupils (in most cases by fewer new pupils) and thus funding and run the risk of closure by Dutch authorities. Table 1 also shows the parental socioeconomic status, based on the postal code of the homes of the pupils. Pupils of Islamic schools have also a very low parental socioeconomic status, although their average is not the lowest. I computed the added value of all schools, based on their final test score average, the average parental socioeconomic status and the percentage parents with very low education. The estimation of parental socioeconomic status was based on the zip code of the pupils of a school in combination with a populationcomposition scale of all zip codes of the Netherlands, made by the Dutch government. Fifty is the highest possible score of a zip code on this population-composition scale (% more twice of average income, % minimum income; % tertiary education, % unemployment, % nonwestern migrants), and minus 50, the lowest possible score. The percentage of parents with low education are provided by the schools: high percentages give them more funding. This computation shows that Islamic schools have a higher final test score then one might expect given the social background of the parents of the pupils, around 1.6 points. Only Hindu schools have a higher added-value of 1.7 points. A difference of one point on this final test can give a better chance to enter a higher track in secondary education (Korthals, 2015). A comparable positive added value of Islamic schools is also found by Merry and Driessen (2014). So, there is a paradox: pupils at Islamic primary schools have very low final test scores, but given their parental background these low final scores are higher than expected.
According to the education inspectorate, 4% of all primary schools were weak or very weak in September 2011 and 3% in September 2012 (Inspectie, 2013, p. 63). Twenty percent of the Islamic schools were weak or very weak in 2011, against 7% in 2012. It is important to note that the inspectorate takes the parental social-economic background into account when they formulate their verdict, but given their very low parental socioeconomic background, Islamic schools run a higher risk that their average final test score and other quality indicators fall below the lowest admitted by the inspectorate. The substantial change in the percentage of weak and very weak primary schools illustrates clearly that schools make a big effort to improve as soon as possible after a negative inspectorate verdict (Koning & van der Wiel, 2013). Weak schools are often helped by their national organizations such as the Islamic School Board Organization (ISBO), or by special pedagogical teams.
What explains Islamic schools' combination of low final test scores, but positive added value and strong improvements? There is no systematic research, but several possible explanations. First, according to the reports of the inspectorate, the didactical and pedagogical approach of Islamic schools is conservative compared with other Dutch primary schools: they use more often front-class teaching and homework. The effectiveness of modern didactics is highly debatable, but most scholars agree that pupils with few parental cultural resources are helped more by structured teaching and clear curriculum requirements. Second, the same inspectorate reports suggest that Islamic schools tend to invest more teaching and learning time to the basic skills (Dutch language; math; geography; history) and avoid spending much time on noncore activities (music, discussion, swimming). In this way, they increase the amount of time actually spent on learning the knowledge and skills measured in final tests (Slavin, 2003), a point I will return to below. Third, Islamic primary schools tend to have low ethnic diversity, typically serving only two or three ethnic different groups (Turks, Moroccans, Indian-Suriname); in strong contrast, urban public schools have considerable ethnic diversity (Veerman, van de Werfhorst, & Dronkers, 2013). Although there is no agreement whether ethnic diversity harms scholastic achievements, there is agreement that high levels of ethnic diversity challenge school operations and might hamper quality. Fourth, most Islamic schools are situated in an urban context and have pupils from poor neighborhoods. The Islamic religious activities of their schools and the active Islamic community of their parents might act as an extra protection against the temptations of that urban context and neighborhoods (comparable to the Catholic school effect: Bryk, Lee, & Holland, 1993;Coleman & Hoffer, 1987;Coleman, Hoffer, & Kilgore, 1982).

Attitudes and values of Islamic schools
The Netherlands has been a multireligious country since the 17th century with a small Protestant majority and a large Catholic minority, which live alongside each other. The Protestant majority broke in the 19th century into different streams ranging from liberal to very orthodox. As a consequence, at the end of the 19th century Dutch society was "pillarized," with politicodenominational segregation of Dutch society in Catholic, Protestant, Social-Democrat, and Neutral segments or "pillars." The Netherlands was "vertically" divided by religions and ideologies. These pillars all had their own social institutions: their own newspapers, broadcasting organizations, political parties, trade unions and farmers' associations, banks, schools, hospitals, universities, scouting organizations, and sports clubs. Political compromises were hammered out by the elites of these "pillars." The great political compromise of 1917 (the so-called school pacification), which allowed religious schools funded by the state and equal quality-standards and curricula for public and nonpublic schools is an example of such a compromise between the elites of these "pillars" (Lijphart, 1968). Only in the 1970s and 1980s did this "pillarization" erode. Now pillarization of Dutch society has disappeared, but remnants can still be seen in the 21st century: religious schools funded by the state being one of these remnants. The breakdown of pillarization did not decrease religious schools, although many of the Catholic and Protestant schools are only religious in name, not in daily practice or by the religion of the teachers and even of the school boards. Yet there is no increasing preference among the growing numbers of nonreligious parents for public schools (see further about the resilience of religious school in a nonreligious society Dijkstra et al., 2004).
Moreover, some orthodox Protestant communities continue to behave as small "pillars." Members of the Reformed Churches (Liberated) have their own primary and secondary schools, their own national newspaper, and some other organizations, such as a labor union. Members of several Orthodox Reformed Churches have also founded their own schools, newspaper, and political party. Muslim immigrants in the Netherlands are also using the legal possibilities created for the pillarized structure of society, by setting up their own schools.
As a consequence of these religious differences and the following pillarization, there was no national consensus or standards about the values and norms, which should be taught in schools. These noncognitive educational goals are left to the schools within the former "pillars." Merry and Driessen (2014, pp. 15-17) provide information about citizenship of primary school pupils from a nationwide study. Their four components of citizenship are measured knowledge, reflection, skills, and attitudes; they also refer to four central societal tasks: acting democratically, acting in a socially responsible manner, dealing with conflicts, and dealing with differences (Ten Dam, Geijsel, & Reumerman, 2011). The Knowledge scale (27 items) varies between 0 and 1. The scales Reflection (28 items), Skills (15 items), and Attitude (24 items) vary between 1 and 4. Table 2 provides insight into the citizenship competences of pupils of Islamic schools and compares these with the scores of pupils of comparable schools (socio-economic composition) and the average school.
With regards to three dimensions (reflection, skills, and attitudes) pupils at Islamic schools score considerably higher than pupils at comparable schools, and still higher than pupils at the average school. Only with regard to knowledge competence do pupils at Islamic schools score nearly the same as pupils at comparable schools, but significantly lower than pupils at the average school. These findings directly challenge the assumption that pupils at Islamic schools are less likely to cultivate the relevant civic virtues for Dutch society at large. To be sure, some schools manage to cultivate civic competences better than others, but this is not related with their religious or private background (also Avram and Dronkers, 2011). An older article (Driessen, 1997, p. 56) contains information about the parental background of pupils of Islamic schools. The differences between Islamic schools and non-Islamic schools are generally not very large. The largest is that pupils at Islamic schools attend Koran classes more often than pupils at comparable schools. Koran classes are part of the regular curriculum at Islamic schools, in the same way that Bible classes are a part of the regular curriculum at Protestant schools. Not all pupils in Islamic school attend Koran classes because schools might have trouble finding qualified Koran teachers or some pupils get exemption from these classes. Another significant difference is that pupils at Islamic schools are far more frequently given homework. This might be a part of a specific strategy at these schools to make more of an "authoritarian" effort to reduce the educational disadvantage of immigrant children than "regular," more liberal primary schools tend to make. A third difference is the importance attached to religion as an important aspect of upbringing. Parents of pupils at Islamic schools attach more importance to religion than parents in other schools (remember that parents at most Catholic and Protestant schools are far more secularized). Also parents of pupils at Islamic have more often not the Dutch nationality, although obtaining Dutch nationality by immigrants from less developed countries is far more common than by immigrants from within the European Union (Dronkers & Vink, 2012).

Administration problems
Nonpublic schools are administrated by foundations or associations with their own legal status, independently from churches, religious organizations, employers' organizations, and so forth. The ground for this rule is the separation between church and state, and thus the impossibility to grant state money to religious schools. As a consequence, boards of nonpublic schools need members who can run their schools, who can negotiate with local and national authorities, and so forth. Catholic and Orthodox-Protestant schools could find enough well-educated and wellconnected members for their school boards. The large Catholic minority had enough upper class believers to recruit efficient school-board members in the 19th century and the same held for the smaller Orthodox-Protestant groups in the 20th century. But what was true for these indigenous religious groups is not necessary true for the Islamic and Hindu groups. The later two religions were endogenous religions in the Netherlands until 1960s (if we ignore the Dutch colonies) and given their migration-history (unskilled labor migration into a society which was not very open to non-European newcomers) lacked well-educated and well-connected believers for their own organizations, like boards of their religious schools. As a consequence of this lack of well-educated and well-connected believers, there were many serious administrative problems in Islamic schools due to mismanagement of the school boards: misuse of educational money for other purposes, fraud, mismanagement of nominating teachers, serious conflicts within boards, and so forth. The quality of the (financial) administration is scrutinized by the education inspectorate. If that quality is too low without prospects of improvement the school will be closed down (formally they do not receive the state grant anymore, and are thus insolvent). However, this lack of well-educated and well-connected believers also means that Islamic schools miss a common network with the Dutch authorities and society at large. This weak network between Islamic school board members and the Dutch political and administrative authorities means that the Dutch usual way to solve administrative problems with schools (strike a compromise between board and authorities) cannot be applied. This lack of well-educated and well-connected members of Islamic school boards and thus their failure to run their school properly according to the national norms of the education inspectorate are important explanations for the closing of the only two Islamic secondary schools in the Netherlands.
Most teachers at Islamic schools are not Muslims, and the qualification of these teachers is more or less equal to that of non-Islamic schools. As stated previously, there are national standards for teacher training and qualification and application is mandatory for state funding. There are too few qualified Islamic teachers available, so schools are forced to nominate non-Islamic teachers acceptable to the education inspectorate. This lack of believing teachers is not unique for Islamic schools. Catholic and Protestant schools have the same problem attracting believing teachers, due to the high level of secularization in the Netherlands (more than 50% in no religion; more among the educated). The secularized parents of pupils attending Catholic and Protestant schools do not care much whether the teachers believe or do not believe (Dijkstra et al., 2004), but the parents of Islamic schools value religion as important in upbringing of their children (see Table 3). The need to use non-Muslim teachers by Islamic schools adds additional tensions within these schools, more than in other religious schools.

Islamic religion
Although individual socioeconomic differences between migrant pupils and their families are the most important explanations of difference in their educational performance, there remain-despite all controls for these backgrounds-substantial differences between migrant children originating from different origins. These origin differences (van Tubergen, Maas, & Flap, 2004), can be summarized as follows (Dronkers & de Heus, 2013): "Migrants' pupils from Islamic origin countries (Turkey, Morocco, Pakistan) have lower scores than comparable migrants' pupils from Christian origin countries (Yugoslavia, Poland, Russia), who have lower scores than comparable migrants' pupils from non-Islamic Asia (India, Vietnam, Korea, China)." This is an outcome that is repeatedly found in research. Further analyses show that this outcome is related with Islamic religion, not with an origin from a country with a majority of Islamic believers. Dronkers and Fleischmann (2010) show that individual religion is related to lower educational attainment of second-generation Islamic men in Europe, not their origin from a dominantly Islamic country. Without any doubt Islamic migrants perceive and face discrimination in Europe (André, Dronkers, & Fleischmann, 2009), but they feel no more discriminated against than migrants who adhere other non-Christian religions (Jews; Eastern religions). Thus discrimination alone cannot be a valid explanation of the lower educational performance of Islamic migrants, because otherwise the migrant pupils originating from Asian non-Islamic countries should also have low educational performance. Migration to Western Europe from Islamic countries like Turkey and Morocco started in the 1960s with guest-workers, temporary labor migrants recruited for unskilled work in dwindling industrial sectors (textile, coal, shipbuilding;Icduygu, 2009). They followed earlier waves of labor migration to Western Europe from Italy, Spain, and Yugoslavia in the 1950s, which also came for unskilled work. The possible negative selectivity of these guest-workers programs (firms and immigration organizations sought young, uneducated workers) seems not to be unique for immigrants from Turkey or Morocco; rather negative selectivity seems also true of guest workers from Italy, Spain, and Yugoslavia (Dronkers & de Heus, 2010). It is possible that some values and norms related with Islam explain the low educational performance of pupils from Islamic countries. These countries score very low at the Gender Empowerment Measurement (GEM). The GEM evaluates women's participation and decision-making ability in political and economic forums (Klasen, 2006). Ranging from 0 to 100, it combines variables such as women's share of parliamentary seats and ministerial positions; as well as managerial, senior official, and legislative jobs; their share of technical and professional jobs; and gender income differences. The very low GEM of Islamic countries explains not only the low educational performance of female migrant pupils from Islamic origin countries, but also the low performance of male migrant pupils from Islamic origin countries (Dronkers & Kornder, 2015). 5 Thus the unequal gender norms in the Islamic countries (masculine virility and feminine purity) may offer a valid explanation for the low educational performance of both male and female migrant pupils from countries with Islam as the dominant religion. Religion need not be a "black box" of cultural phenomena, its various aspects can be analyzed (gender equality, economic values, authority) and their importance in adherents' behaviour estimated. It can imply that values and norms, which are related to a religion, can be adjusted to new circumstances and challenges. An example of such an adjustment is Catholicism, which adjusted to capitalism without losing its critical stance toward gluttony.

Conclusion
Islamic primary schools have gained a foot in the Dutch educational system and their numbers are increasing. A partial explanation of this growth is their higher quality compared with public schools with the same socioeconomic composition. Another reason for their growth is that Islamic parents value a religious upbringing of their children and see Islamic schools as an extra protection against the temptations of their urban neighborhoods. A more conservative pedagogic approach, more time-on-task for core subjects, lower ethnic school diversity and the active Islamic community around the primary school could explain the relative higher quality of Islamic primary schools.
However, the running of secondary Islamic schools has failed because of low quality and serious administrative problems. An important reason for this failure is the nonexistence of a well-educated and well-connected Islamic elite for the foundations and associations that establish and run the religious schools according the national norms of the education Inspectorate.
A final challenge for Islamic schools are the gender values and norms related with the Islamic religion in the Mediterranean region, where most Islam migrants in continental Europe originate. These gender values and norms hamper educational performance of both male and female children in their migration societies because they reference a society which no longer exists for them and thus blocks adjustment to modern society. In modern society, high education performance by male and female students is of tantamount importance, and very different and unequal gender roles (masculine virility; feminine purity) seem not helpful for this performance. Islamic schools in Europe might be instrumental for that adjustment of Islam to modernity. An indication of their possible success as such an instrument is the positive added value of Islamic primary schools in the Netherlands, for boys and girls.
A next step in the study of Islamic schools in Europe is to see whether these Dutch observations apply to other countries where Islamic schools are now established (Belgium for instance) or where quasi-Islamic schools exist (England).