The Role of School Libraries in Reducing Learning Disadvantages in Migrant Children

The educational achievement of children from non-Western migrant families in the Netherlands and other Western countries lags behind that of natives, especially when it comes to language proficiency and reading ability. This literature review pinpoints what is known and what is as yet unknown about reducing learning disadvantages through school libraries to point to directions for future research. A considerable body of research has shown that school libraries are positively related to learning outcomes in children, as well as to their reading behavior and attitude toward reading, factors that correlate positively with reading and language skills. However, on the basis of existing research, it is hard to draw firm conclusions about the effect of school libraries on students from migrant families in particular. This article indicates that future research should explicitly focus on the impact of school libraries’ reading promotion efforts on the reading behavior, attitude toward reading, and reading and language skills of migrant students, leading to more effective educational policies.

Good language skills and reading proficiency are important for participating successfully in modern society (Johnsson-Smaragdi, 2006; National Endowment for the Arts [NEA], 2007). However, it is evident that not every child sufficiently masters these skills. If we consider, for instance, the situation in a highly developed Western country like the Netherlands, approximately a quarter of the students finish primary school with insufficient technical reading proficiency, which in turn negatively affects their reading comprehension skills (Vernooy, 2009). In fact, there are categories of children who in particular face difficulties in developing good language and reading skills. An urgent issue-typically observed in many Western countries with non-Western migrant groupsthat provides a case in point is the ethnic inequality in school performance (Gijsberts & Iedema, 2012;Schnepf, 2007).
Since the 1960s, migrants have come to the Netherlands in large numbers (Herweijer, 2009). Statistics show that in 2012, the Netherlands counted almost 1.6 million Western migrants and more than 1.9 million non-Western migrants, 1 making up 9% and 12%, respectively, of the total population. Turkish and Moroccan migrants constitute the largest non-Western migrant groups, followed by Surinamese migrants and people of Dutch Antillean origin (Nicolaas, Loozen, & Annema, 2012). In line with the migration flows, multiethnic schools and schools with a high percentage of migrant students (especially in the largest cities) have become more common in the Netherlands over the last decades (Herweijer, 2008). In 2013, about 17% of the primary school-aged children (4-12 years) were of non-Western origin (Centraal Bureau voor de Statistiek [CBS], 2013).
The situation of migrant students differs in several respects from that of their indigenous counterparts. First of all, migrant children are often partly or even entirely raised in another language. Accordingly, many migrant children have to acquire both the minority language of their cultural community and the Dutch language (Scheele, 2010). Another important difference is that the socioeconomic position of migrant families is usually weaker, as expressed by their lower education level, higher unemployment rate, and lower income compared with native Dutch families (Herweijer, 2009). In addition, it has been suggested that parents from migrant groups in the Netherlands are generally less often involved in their children's education (Herweijer & Vogels, 2004) and that the socialization in migrant families more often creates characteristics and expectations among children that run counter to what is valued in formal schooling (Elbers, 2002). Overall, the situation of many migrant 580369S GOXXX10.1177/2158244015580369SAGE OpenKleijnen et al.

research-article2015
1 University of Amsterdam, The Netherlands 2 Utrecht University, The Netherlands families puts ethnic minority students at higher risk for poorer school performance (Gijsberts & Herweijer, 2009).
Research has clearly demonstrated that the educational achievement of children from non-Western migrant groups in the Netherlands falls behind, especially when it comes to language and reading ability (Gijsberts & Iedema, 2012). As reported by Herweijer (2009), language deficiencies among migrant students are already manifest at the start of primary school. 2 Herweijer demonstrates that in kindergarten, children from non-Western migrant groups not only lag behind native Dutch students whose parents have a secondary or higher education background but also-albeit to a lesser extent-relative to native children from families with a loweducation background. The language disadvantages of migrant students are much more pronounced than their disadvantages in arithmetic, which makes fewer demands on their language skills (Herweijer, 2009). Although the language proficiency of non-Western migrants at the start of primary school has improved over the years, there is still a considerable gap with their native peers (Gijsberts & Iedema, 2012).
The initial learning disadvantage among migrant students continues throughout their school career (Herweijer, 2009;Meelissen et al., 2012). Differences between the various migrant groups and native Dutch students are still evident in the last year of primary school, particularly in language proficiency (Gijsberts & Herweijer, 2009). At the end of primary school, Turkish Dutch, Moroccan Dutch, and Antillean Dutch students have a language disadvantage of approximately 2 years on average (Herweijer, 2009). On a more positive note, studies comparing different cohorts have indicated that the language disadvantages of migrant children in sixth grade have diminished over time. Still, it is a relatively slow process and upon finishing primary school, non-Western migrant students continue to be educationally disadvantaged (Driessen & Merry, 2013;Gijsberts & Iedema, 2012).
Given the crucial role of good language and reading skills, it is important to combat the learning deficiencies in migrant children. Reducing educational disadvantages is suggested to go hand in hand with reading promotion efforts (Hermans, 2002). In the Netherlands, most primary schools and public libraries cooperate when it comes to reading promotion, and school libraries supported or run by public libraries have become increasingly common (Kasperkovitz, van Tits, & von der Fuhr, 2009;Oberon, 2009Oberon, , 2011. The ultimate objective of the collaboration between schools and libraries is enhancing the students' language development (Huysmans, Kleijnen, Broekhof, & van Dalen, 2013). Considering the investments involved and the importance of reducing educational disadvantages, it is relevant to have insight into the effectiveness of reading promotion efforts by school libraries in the Netherlands.
Although ample studies have addressed the impact of school libraries, there are also gaps in the literature that need to be filled. By identifying and eliminating these gaps, more effective policies targeted at reducing ethnic inequality in school performance can be conceived. Therefore, the aim of this contribution is to point to directions for future research into the effectiveness of school libraries in reducing learning deficiencies in migrant children.
For this purpose, a literature study has been conducted casting the net rather widely. Besides paying attention to research that explicitly addresses the influence of school libraries on the educational achievement of (migrant) students, other literature that indirectly sheds light on this subject was included. For instance, because school libraries are aimed at reading promotion, studies on the relationship between reading behavior and school performance are discussed as well. The study is also concerned with what is known about the effectiveness of reading promotion efforts by parents, schools, and public libraries, given that school libraries function amid these other pedagogical agents. Moreover, this contribution also describes findings from research on children in general, because little research explicitly focuses on students with a migrant background.
In this literature review, the contribution of school libraries in reducing ethnic inequality in school performance is described and analyzed primarily for the Dutch situation. Realizing that conditions for ethnic groups vary from group to group and from country to country, which undermines the generalizability of findings, it was deemed wise to make an in-depth study of one country rather than assume that findings from various nations and continents can be integrated into a coherent whole with general validity. Nevertheless, the evidence presented here can within reasonable boundaries be generalized to other Western countries with the same or similar ethnic groups. Also, research and policy implications provided in the "Discussion" section will be more widely applicable.
The main research question of this article is as follows: Research Question 1: What is known and what is as yet unknown about reducing learning disadvantages among children in general, and children from a migrant background in particular, through school libraries in the context of other pedagogical agents (i.e., parents, schools, and public libraries)?
This main question can be divided into four sub-questions that concern both children in general and migrant children: Research Question 1a: What is the correlation between leisure reading and school performance? Research Question 1b: What is known about the effectiveness of reading promotion efforts by parents, schools, and public libraries on reading attitude, reading behavior, and school performance? Research Question 1c: What is known about the effectiveness of reading promotion efforts by school libraries on reading attitude, reading behavior, and school performance?
Research Question 1d: What directions could future research take to help policy enhance the contribution of school libraries in reducing learning deficiencies?

Method
To answer these questions, we searched for literature on the relationship between leisure reading and school performance as well as for literature that addressed the effects of reading promotion efforts by parents, schools, public libraries, and school libraries on students' reading attitude, reading behavior, and school performance. Our initial focus was on literature about migrant students in the Dutch context, yet we also searched for studies that addressed children in general and/or studies conducted outside the Netherlands. The findings primarily describe the situation in the Netherlands, although within reasonable assumptions they can be regarded as indicative for migrant groups in comparable developed Western countries as well.
Literature was searched for through a library discovery tool accessing various electronic databases (most importantly, ScienceDirect, CrossRef, ERIC, LexisNexis Academic, Emerald, PsychINFO, and JSTOR), and through search engines (Google Scholar and Google), using varying combinations of keywords in both Dutch and English. 3 We also located potentially relevant papers by scanning the reference sections of papers already identified and by searching through the publications of organizations that address topics related to the focus of our review (e.g., Organisation for Economic Co-Operation and Development [OECD], Stichting Lezen (Dutch Reading Foundation), and National Literacy Trust). This search strategy identified a wide range of sources such as articles in peer-reviewed journals, chapters in edited books, dissertations, and research reports. From this gross list, we selected those sources that, judged from the abstracts, could explicitly or implicitly contribute to answering our research questions.
It should be noted that the studies discussed here vary considerably in the research methods used, having consequences for the value that can be attached to their results. For example, reading and language skills are mostly assessed through tests, whereas some studies conducted in this area use questionnaires to gain insight into the respondents' skills. The self-reported data can be seen as less valid than those obtained by administering tests (Poll, 2012). Furthermore, as reading attitude is not directly observable (Alwin & Krosnick, 1991), this construct is typically assessed using questionnaires. Reading behavior is also often measured via selfreports. Yet it is not certain that the children's answers completely cover their actual reading behavior or attitude, as bias may occur caused by factors such as social desirability, unclear instructions, and vague wording of the questions and response categories (Pouwer, van der Ploeg, & Bramsen, 1998). Another instrument used for assessing reading behavior is a Print Exposure Checklist or a Title Recognition Test that measures familiarity with book titles and authors or magazines (Cunningham & Stanovich, 1998;Mol & Bus, 2011b). This instrument provides an index of relative differences in reading volume and it is considered more objective than self-report questionnaires (Cunningham & Stanovich, 1998).
Apart from the type of instrument used, other methodological characteristics such as the sample and the operationalization of the measured constructs may have affected the research outcomes. In the remainder of this article, the methods used in former studies are mentioned as much as possible so the reader can take some note of the quality of the discussed research.

Reading Behavior and School Performance
Given the relevance of good reading proficiency and language ability, it is important to reduce disadvantages in this area among migrant children. To overcome, for instance, a vocabulary disadvantage of 2,000 words in the first school years, non-Western children in the Netherlands should learn 3,000 Dutch words (i.e., 1,000 to not fall further behindbecause their native Dutch peers will learn new words as well-and 2,000 words to bridge the gap). If this is supposed to happen during 1 school year, these students should learn 75 new words a week. However, with intentional vocabulary learning at school, on average just 25 new words are learned on a weekly basis, which is by no means sufficient for children of non-Western origin to catch up (Broekhof, 2011).
To reduce the disadvantages among these children, their language and reading skills have to be stimulated in another way. This is possible through incidental learning, which is a process of learning without the intention of doing so (Ahmad, 2011;Broekhof, 2011). In terms of language acquisition, one learns new words while listening to a story or reading an interesting text (Broekhof, 2011). In fact, most theorists are convinced that the bulk of vocabulary growth during childhood occurs indirectly, through language exposure, rather than directly, through teaching. Moreover, as most speech is lexically impoverished compared with written language, many researchers have argued that reading volume, rather than oral language, is the main contributor to individual differences in children's vocabularies (Cunningham & Stanovich, 1998). Leisure reading thus seems to be an important complement to instruction at school when it comes to reducing learning disadvantages.
Leisure reading, also often referred to as reading for pleasure, independent reading, recreational reading, and voluntary reading (Clark & Rumbold, 2006;Moyer, 2011), has been defined in the literature in various ways, with the common component being that it is part of a non-school (and non-work) recreational activity (Moyer, 2011). Leisure reading typically involves materials that reflect one's own choice (Clark & Rumbold, 2006). Nowadays, reading is not limited to printed texts, as electronic devices such as personal computers, mobile phones, e-readers, and tablets offer a wide variety of digital reading opportunities, taking the form of, for example, e-books, emails, articles on websites, text and WhatsApp/Ping messages, and postings on social networking sites such as Facebook and Twitter (Clark & Douglas, 2011;Gillebaard & Jager, 2011).
It has been suggested that the nature of reading is changing due to the use of digital devices (Coiro, 2003;OECD, 2011b;van der Weel, 2011), as digital texts can differ from printed texts in several respects, such as the nonlinearity of page structures, the potential dynamic nature (OECD, 2011b), the amount of text visible to the reader at any one time (Gil-Flores, Torres-Gordillo, & Perera-Rodíguez, 2012), the (possible) interactivity, and the inclusion of multiple media forms (Coiro, 2003). Culture critics Carr (2011) and Wolf (2007;as cited in N. Bakker, 2013) feared that the rather shallow and fragmented form of reading that is thought to come with non-linear reading and multitasking (i.e., switching between the different functionalities of the device) diminishes the capacity for concentration, interpretation, and contemplation. Although empirical research that either confirms or rejects such claims is scarce, it cannot be argued that every electronic device and digital text comes with a similar degree of (potential) shallow and fragmented reading (N. Bakker, 2013). For instance, blogs or messages on Facebook are much more concise and prone to stimulating non-linear reading and multitasking compared with an e-book. Moreover, texts on websites are more susceptible to language errors and logical inconsistency due to lack of editorial assistance (Huysmans, 2013). Thus, different types of digital texts do not seem to offer equal reading experiences.
This review focuses on the rather immersive and sustained forms of reading as a recreational activity, because this is the kind of reading that reading promotion is aimed at (N. Bakker, 2013). This type of leisure reading includes, for instance, reading story books and comic books (paper and digital forms), but not text messages, emails, and postings on social networking sites. Although leisure reading is usually considered to be limited to a non-school activity, we also discuss some studies that address in-school free reading (i.e., time set aside for reading for pleasure at school during which students can read whatever they wish-within reasonable limits-with little or no accountability in the form of grades or book reports; Krashen, 2004a), because this form can also be considered a similar kind of recreational reading. The remaining part of this section discusses the state of the art regarding leisure reading in the Netherlands as well as research on the relationship between leisure reading and school performance.
Leisure reading in the Netherlands. A study on the reading behavior of 7-to 15-year-olds in the Netherlands showed that, in 2012, 68% of children aged 7 years reported reading a book nearly every day, whereas this was true for 35% of 12-year-olds and 20% of 15-year-olds. Fiction books were most often read by the children, followed by comic books and non-fiction books (Huysmans, 2013). The Progress in International Reading Literacy Study (PIRLS), a large-scale international comparative study using self-reported data, demonstrated that between 2006 and 2011, the frequency of reading for pleasure slightly increased among fourth graders in the Netherlands. In 2011, 36% of the students indicated spending 30 min to 1 hr on leisure time reading a day and 22% of the students reported reading more than 1 hr per day. Migrant children more often reported reading to acquire information than native Dutch children. The two groups did not differ in time spent reading for pleasure (Meelissen et al., 2012).
With regard to digital reading among children in the Netherlands, findings from the 2006 PIRLS showed that fourth graders indicated spending more time on a typical day reading articles and stories in books or magazines than online (Mullis, Martin, Kennedy, & Foy, 2007). Huysmans (2013) showed that in 2012 only 3% of children aged 7 to 15 years reported reading book apps (e-books) during leisure time, whereas story books were read by 85% of the children. With respect to ethnical background, a survey-based study by Hirzalla, de Haan, and Ünlüsoy (2011) indicated that migrant adolescents, more than their native Dutch counterparts, were involved in online discussions, searching for information online, and emailing, types of reading we chose not to consider in this article (see above). Ethnic differences in e-book reading have as yet not been examined in the Netherlands.
Relationship between leisure reading and school performance. There is a vast amount of research demonstrating the importance of reading behavior for reading proficiency and language ability (Broekhof, 2011;Krashen, 2004b). In 2007, the American institute NEA published To Read or Not to Read, a report which-on the basis of the most accurate national (American) data available-showed that reading for fun strongly correlates with academic achievement. Youth who reported reading for pleasure on a daily or weekly basis had higher reading scores than less frequent readers. Similar results were found in the 2006 PIRLS. Internationally, students who indicated reading for fun (almost) every day had an average reading score of 516 points, compared with 503 points for students reading for fun once or twice a week, and 484 points for those reading only twice a month or less (Mullis et al., 2007). For Dutch children in particular, these reading achievement scores were 566, 550, and 530, respectively (Netten & Verhoeven, 2007). Other research has demonstrated that Dutch children (between 7 and 15 years of age) who indicated being very good at reading reported more reading behavior than children who perceived themselves as less advanced readers (Huysmans, 2013).
With respect to vocabulary, it has been estimated that children who read approximately 15 min a day read 1 million words a year and can thereby enlarge their vocabulary with 1,000 words (Broekhof, 2011). In line with this, Sullivan and Brown (2013), who studied a nationally representative sample of around 6,000 people born in Britain, found that (selfreported) leisure reading of books at both ages 10 and 16 was related to vocabulary at age 16. Reading frequency was even a stronger predictor of vocabulary than was reading proficiency at age 10. According to Krashen (2013), this finding supports the claim that literacy development can be improved anytime through free reading.
Researchers who have reviewed many studies on reading for pleasure also emphasize a positive link between reading and educational development. A publication that provides a case in point is The Power of Reading (Krashen, 2004b). On the basis of evidence from studies conducted all over the world, Krashen concluded that free voluntary reading (i.e., reading because you want to, both in-and out-of-school) positively affects the development of reading comprehension, vocabulary, spelling, and grammar. The findings of Mol and Bus (2011b) support this conclusion. They meta-analyzed 99 studies that addressed leisure reading of preschoolers and kindergartners, children attending Grades 1 to 12, and college and university students. In sum, print exposure, as measured by checklists, appeared to be an important correlate of scores on reading and language tests. Children reading books in their leisure time appeared to have better reading comprehension skills, larger vocabularies, and better spelling and technical reading skills than peers with less reading experience.
As most school subjects depend to varying degrees on reading and language ability, one would expect mastering these skills to also positively affect other areas of school (Logan & Johnston, 2009). This is supported by the findings of Kortlever and Lemmens (2012) who investigated the correlation between self-reported reading frequency outside school and sixth-grade students' scores on a widely used standardized test in Dutch primary schools, the Cito test. This test establishes several learning outcomes: language, math, study skills, and general knowledge of the world. Kortlever and Lemmens found that leisure reading positively related to the overall test score as well as to the scores on the separate Cito-test elements. The relation between reading behavior and the Cito-test scores was fully mediated by the language test score, indicating that reading frequency correlates with better language skills, which in turn explained the better scores on the other Cito-test elements.
Research has suggested that the positive effects of reading also hold true for migrant children (Broekhof, 2011;Krashen, 2004b;Lao & Krashen, 2000). Krashen (2004b) referred to studies conducted by Elley showing that reading for pleasure has a strong effect on second-language learners. In one of these studies (Elley & Mangubhai, 1983), he studied Fiji Islands' primary school students who were learning English by following daily classes of 30 min. A sample of fourth-and fifth-grade students was randomly assigned to one of two experimental groups or a matched control group that followed the normal structured English language program. The first experimental group engaged in free reading instead of the standard program, whereas the second experimental group participated in a shared book reading experience method in which the teacher shared a book several times with the class by, for instance, reading it aloud and discussing it with the students. The time devoted to the English language was the same for all groups. After 20 months, both experimental groups-who performed quite similar-outperformed the traditionally taught group on tests of reading comprehension, writing, and grammar. Eight years later in Singapore, studying approximately 3,000 children (aged 6-9 years), Elley again found students who engaged in a free reading program (i.e., a combination of shared book reading experience, language experience, and free reading) performing far superior not only on reading comprehension, writing, and grammar but also on vocabulary, listening comprehension, and oral language than their traditionally taught peers.
Research focusing on low-ability readers can also shed more light on the effects of leisure reading for migrant children, given their learning disadvantages. For example, Cunningham and Stanovich (1998) found a positive link between reading behavior and language ability both for more able readers and children with limited skills. On the basis of their meta-analysis, Mol and Bus (2011b) demonstrated that leisure reading is especially important for low-ability readers. Basic reading skills of students with a lower reading ability were more strongly related to print exposure than those of higher ability readers. These results indicate that the effects of leisure reading also, or even more strongly, hold for students with a migrant background.
In some studies addressing the relationship between reading and school achievement, the measurement of leisure reading is limited to reading (fiction) books, whereas in other studies, other types of printed media are taken into account as well. The different kinds of reading materials can contribute to children's vocabulary development, which holds true especially for learning rare (low-frequent) words like those used in school language and subjects like geography and history (Broekhof, 2011). Research has shown that reading a wide variety of materials, as assessed through a questionnaire, is strongly related to reading achievement scores (Gille, Loijens, Noijons, & Zwitser, 2010).
Using data from student questionnaires and tests, research has suggested that books, in particularly fiction books, have the strongest impact on children's linguistic skills (Gille et al., 2010;OECD, 2011a). In line with this, the 2006 PIRLS showed that reading informational texts has a less clear-cut relationship with reading achievement than reading novels and short stories (Mullis et al., 2007). A publication reporting on empirical studies also indicated that reading fiction has a greater impact on vocabulary than reading of non-fiction does (Hakemulder, 2011). Cunningham, Stanovich, and West (1994;as cited in Kortlever & Lemmens, 2012) have shown that reading magazines also correlates with academic achievement, although not as strongly as book reading, whereas Kortlever and Lemmens (2012) did not find a relationship between reading magazines and learning outcomes. Nevertheless, Cunningham and Stanovich (1998) cited statistical studies on several reading materials, demonstrating that print media (i.e., adult books, newspapers, abstracts of scientific articles, children's books, comic books, and popular magazines) contain a far greater average number of rare words than TV shows or (higher educated) adult speech.
Other research has been conducted addressing the relation between reading digital texts and reading proficiency. For instance, Clark and Douglas (2011), who analyzed survey data and reading attainment scores of a large sample of youth (aged 7-16 years) in the United Kingdom, found that youth who reported reading more traditional paper-based materials (e.g., fiction, non-fiction, and poems) were more likely to read above the expected level for their age than those who read, for instance, websites, text messages, and emails (Clark, 2012;Clark & Douglas, 2011). Respondents who reported reading e-books were likely to read above the expected level as well (Clark, 2012). Jeong (2010), in his study of a small sample of Korean sixth graders, found that paper books enabled better reading comprehension than e-books, although he stated that previous studies on reading comprehension comparing e-books and printed books found that, overall, comparable attainment can be achieved.
It should be noted here that the source used for reading e-books seems to be important and can therefore have consequences for research outcomes when comparing e-books and paper books. For example, on the basis of self-reported survey data of about 1,300 Dutch people (from 13 years of age onward), N. Bakker (2013) showed that the experience of reading a book from an e-reader, a unifunctional device, was rather similar to reading a book on paper. On the contrary, respondents who read e-books from a tablet or laptop, multifunctional devices, reported lower levels of concentration, comprehension, relaxation, and pleasure than when reading a paper book. Experimental research is needed to further investigate differences in reading experience between various media used for digital reading.

Reading Promotion Through Parents, Schools, and Public Libraries
As leisure reading appears to be key to school performance, reducing educational disadvantages among children seems to go hand in hand with reading promotion activities (Hermans, 2002). Reading promotion is typically aimed at increasing reading frequency, and improving reading and language skills as well as reading attitude (Stalpers, 2005). These factors are found to be related in a reciprocal manner: Children who report more positive attitudes toward reading will tend to read more, which in turn translates to a higher reading and language ability (Broekhof, 2011;Cubiss, 2012;Meelissen et al., 2012;Mol & Bus, 2011a).
It is assumed that childhood is a crucial period for stimulating reading, as reading socialization experiences seem most effective during the so-called formative years (between ages 5 and 20; Kraaykamp, 2003). Over the last decades, national (annual) recurring reading promotion activities and projects aimed at children have been initiated in the Netherlands, as well as various unique and local reading promotion interventions. However, research on the effectiveness of Dutch reading promotion activities is rather scarce (Bonset & Hoogeveen, 2009;Piek, 1995). Studies that did examine the results of such interventions pointed to positive effects, such as increased reading motivation and reading frequency, and broadening of reading interests (e.g., de Haan & Kok, 1990;Hermans & Jans, 2012;Stokmans, 2007).
According to sociologist Kraaykamp (2002Kraaykamp ( , 2003, however, it is far from clear if reading socialization activities really are effective. Kraaykamp (2003) stated that "cultural knowledge and skills are not acquired by unique and nonrecurrent introductions to literature or culture" (p. 236). He has suggested that to be effective, reading promotion efforts should take place through intensive recurrent contact in a relevant social context. The conditions for reading socialization occur primarily in three institutions: at home, in schools, and in libraries (Kraaykamp, 2002(Kraaykamp, , 2003. In the following sections, these three socializing agencies will be discussed.

Parents.
A large body of literature points to the importance of parental socialization for children's reading behavior, reading attitude, reading ability, and overall educational success (R. Bakker, 2011;Broekhof, 2011;De Graaf, De Graaf, & Kraaykamp, 2000;Kraaykamp, 2002Kraaykamp, , 2003Mol & Bus, 2011a, 2011bNotten, 2011;van Steensel, 2006;Verboord, 2003). Parents may influence their children's reading development either by setting an example or by actively stimulating children's reading habits, referred to as imitation and instruction (Bandura & Walters, 1963;Kraaykamp, 2002;Leseman & de Jong, 1998). Applied to reading promotion, imitation, which takes place more or less unconsciously, refers, for example, to parental reading behavior and the availability of reading materials at home. In instruction, however, parents deliberately encourage their child's development by, for example, reading books to children, recommending books, discussing books, and giving books as a gift (De Graaf et al., 2000;Kraaykamp, 2002Kraaykamp, , 2003. In migrant families, there seems to be a less favorable reading climate compared with native Dutch families. Studies have indicated that migrant children are less likely to be read to, have fewer reading materials at home, and their parents are less inclined to set an example by reading themselves (de Vries, 2007;Hermans, 2002;Scheele, 2010;van Steensel, 2006). Literature also points to differences in reading climate between different migrant groups. For example, Turkish Dutch parents more often read to their children than parents of Moroccan origin. In families with a Turkish background, maintenance of the Turkish language is generally highly valued and also facilitated by the widespread availability of Turkish reading materials, including books for young children. Accordingly, children in these families are often read to in their first language. Moroccan Dutch families, however, more often have to rely on Dutch reading materials, because the primary language of the majority of these families is Berber, a non-scripted language (R. Bakker, 2011;Scheele, 2010).
Schools. Children spend a considerable amount of time at school, an institution where they acquire knowledge and skills. Like parental socialization, education is also suggested to be influential in children's cultural development, as school introduces many children to reading and culture (Kraaykamp, 2003). Using nationally representative survey data, Dutch research on the effect of secondary school socialization in the long run has shown that cultural education and an extensive humanities oriented set of finals correlate with a stronger preference for literature later in life (Kraaykamp, 2003). On the basis of surveys administered to a sample of teachers and their former students, Verboord (2003Verboord ( , 2005 found that a more student-centered literary education (i.e., with lessons adjusted to the students' preferences) results in a higher book reading frequency at an adult age. Furthermore, Stalpers (2005Stalpers ( , 2007 has demonstrated on the basis of data from student questionnaires that teachers' behaviors such as reading themselves, and discussing and recommending books, are positively related to the reading attitudes of secondary school students. With regard to primary school, it is suggested that different accents between school curricula may have different consequences for children's reading development, with more effective instruction, in terms of method and scheduled time, leading to a higher reading ability (P. F. de Jong & Leseman, 2001). Reading for pleasure at school also seems of importance. For example, on the basis of a review of studies, Krashen (2004b) indicated that students engaging in a free reading program often had a higher reading achievement than students in traditionally taught classes, and the longer the reading program lasted, the greater the gains. On average and almost without exception, students who read for fun within the context of a reading program lasting longer than 12 months performed better than their counterparts in classes that lacked free reading.
In the Netherlands, elementary schools are free to shape their curricula. Therefore, schools differ in the methods they use as well as in the time spent on reading and language education (Meelissen et al., 2012). Although instructional time is considered a crucial resource with regard to students' opportunity to learn, it is difficult to capture its effect on student performance, because various factors can influence the productivity of the instruction hours (Mullis, Martin, Foy, & Drucker, 2012). The 2011 PIRLS sheds light on the instructional time spent on language and reading in Dutch schools (fourth grade). Teachers who participated in this study indicated spending on average 8.4 hr a week on language education and/or other language activities. This is more than was reported in 2006 and 2001 (7.8 and 7.7 hr, respectively). On average, teachers spend the most hours on reading (31%), followed by grammar and spelling (27%). In addition, the teachers often give their students the opportunity to read for pleasure in class, and students are frequently read to by the teacher or by fellow students (Meelissen et al., 2012).
On the basis of interviews and surveys, Oberon (2009), however, concluded that many Dutch primary schools do not exhibit a systematical and structural approach regarding reading promotion. Teachers themselves often engage on their own initiative in reading activities with their students. However, many schools do participate in annual national reading promotion campaigns and activities (Oberon, 2009;Stichting Lezen, 2012).
Public libraries. Studies on reading socialization point to the importance of the availability of reading facilities (see Kraaykamp, 2003). Having easy access to a wide variety of reading materials is positively linked to children's reading behavior, reading proficiency, and motivation to read (Gille et al., 2010;Guthrie & Humenick, 2004;Krashen, 2004b). In the Netherlands, such circumstances are created by public libraries (Kraaykamp, 2003), and a library membership is usually free for children under 16 years of age and 16-and 17-year-olds are often charged a reduced fee (Huysmans & Hillebrink, 2008;Huysmans & Röst, 2009). Dutch research by Kraaykamp (2002Kraaykamp ( , 2003 has indicated that the library is effective in promoting interest in reading and that people who were in their youth library members for a long time exhibit a stronger preference for literary books and suspense novels later in life. Since the 1990s, library use in the Netherlands has shown a decline in terms of memberships, visits by members, lending, and number of books borrowed (Huysmans & Hillebrink, 2008). Between 2000 and 2013, library membership fell by more than 10% to 3.86 million members. Yet, this trend does not apply to young people between 0 and 17 years of age. In 2013, 2.24 million youths were library members, compared with 1.98 million in 2005 and 2.05 million in 2000. This is partly a result of public libraries' increased efforts directed to young people, such as the implementation of reading promotion programs. Consistent with the trend among adults, though, the number of items borrowed by young people has decreased between 2000 and 2013 from 58.8 million to 38.0 million. The number of materials they borrowed in the non-fiction section declined more steeply than in the fiction segment (CBS, 2014). Research conducted in 2009 among a representative sample of primary school children (aged 7-12 years) revealed that 73% of 7-to 9-year-olds and 70% of 10-to 12-year-olds who reported having a library membership (77% and 84% of the total group, respectively) had borrowed at least one book in the month before the survey (Siebelhoff, Caarels, & Shen Cheung, 2010).
Research by Huysmans and Hillebrink (2008) has shown a difference in self-reported library use between native Dutch citizens and ethnic minority groups. Contrary to the overall downward trend, library membership and recent borrowing of books increased among citizens with a Turkish and Moroccan background between 1995 and 2003. Ethnic minorities also show a different pattern in library use by age. Unlike the native Dutch population, library use is high among migrant teenagers and young adults (aged 15-24 years), especially among those of Turkish and Moroccan origin, whereas the library is rarely visited by older adults with a migration background. Ethnic differences in library usage among children under 15 years could not be assessed in this particular study. Some insight is given by W. de Jong et al. (2010) who drew on a large-scale survey study conducted in 2007. They found that non-Western children between 4 and 12 years of age more often reported visiting the public library at least once a month than did their native Dutch counterparts (55% vs. 46%). However, library membership was more common in indigenous children (85%) than in those of non-Western origin (77%).
Nowadays, the reading promotion activities of Dutch libraries go beyond lending books. Libraries all over the country organize a range of activities to enthuse children, making the library an institution that promotes reading in a broad sense (http://www.lezen.nl/; Kraaykamp, 2003;Oberon, 2009). Cooperation between schools and public libraries is also very common in the Netherlands when it comes to reading promotion: More than 80% of the libraries who participated in a 2009 study reported working together with primary schools in all their branches (Kasperkovitz et al., 2009). For instance, students visit the local public library with their teacher, libraries organize shows for students, and libraries inform teachers about national reading promotion interventions (Oberon, 2009).
Moreover, many primary schools in the Netherlands have some sort of school library (Bonset & Hoogeveen, 2009;Heesters, van Berkel, van der Schoot, & Hemker, 2007;Oberon, 2009). A study conducted in 2005 showed that, according to questionnaires administered to a sample of teachers, the library's collection was often in the school's own possession. To a lesser degree, the collection was provided by the public library (Heesters et al., 2007). Currently, school libraries supported or even run by public libraries are becoming increasingly common (Oberon, 2009(Oberon, , 2011.

Effectiveness of School Libraries
Since the 1960s, many studies have addressed the effectiveness of school libraries (Roberson, Schweinle, & Applin, 2003). According to Krashen (2004b;Krashen, Lee, & McQuillan, 2012), it has been well established that access to books, as provided by school libraries, results in more reading. Given the positive relationship between reading behavior and proficiency in reading and language, school libraries seem to have the potential to improve the school performance of students.
Indeed, a substantial body of foreign research has established a positive link between school libraries and improved learning outcomes (see Ryan, 2004;Scholastic, 2008). Some of the most significant studies include the work of Lance and associates (Roberson et al., 2003). Their work encompasses a number of statewide studies conducted in the United States and involves both primary and secondary schools (Ryan, 2004). The first of these studies was carried out in Colorado (Lance, Welborn, & Hamilton-Pennell, 1993) and revealed that the size of the school library (in terms of its collection and staff) was positively related to scores on reading tests. Between 5% and 15% of the differences in reading scores among students of various elementary and secondary grades were explained by library size, while controlling for confounding factors.
The results from subsequent studies in Colorado and other American states consistently indicated that access to school libraries enhances student achievement, irrespective of how achievement was measured (i.e., standardized reading scores, literacy, or learning more generally). In addition, reading enjoyment was strongly related to the presence of a teacherlibrarian (Blackett & Klinger, 2006). Based on varying instruments (e.g., student surveys, teacher-librarian surveys, tests, and focus groups of teachers and students), research carried out in other countries, such as England (Clark, 2010), Scotland (Boelens, 2010;Williams & Wavell, 2001), and Australia (Hay, 2003;Softlink, 2012), confirmed that school libraries contribute to school attainment, as well as to reading enjoyment (Clark, 2010).
Foreign research also sheds light on the attributes of school libraries that are positively linked to student achievement. First of all, the presence of adequate, qualified library staff is found to be of vital importance. An effective school librarian has an instructional role, guides students toward a love of reading, and interacts and collaborates with classroom teachers. Moreover, studies have also demonstrated the importance of funding, library usage, and flexible library access as well as the size and quality of the collections and access to technology (Gavigan, Pribesh, & Dickinson, 2010;Lance, Rodney, & Hamilton-Pennell, 2005;Michie & Chaney, 2009;Roberson et al., 2003;Scholastic, 2008).
The rather limited available data from the Netherlands on school libraries also point to positive outcomes. A pilot study of the Library at School involving seven libraries and 31 primary schools showed that youth membership grew with 65% and book loans with 115% (Oberon, 2011). A growth in book loans was also found in the town of Almere (Oberon, 2011). The research and statistics department of the public library in Vlissingen noticed that a school library was related to more self-reported reading activity among children (Oberon, 2011).
A more extensive research was conducted by Geurtsen (2008), who studied the effectiveness of a specific library run for and by students of three primary schools in Hoorn. He found that these children reported more leisure time reading than children in a control group. Controlled for confounding factors, the difference was over 10%. The attitude toward books was also more positive among the experimental group. The difference amounted to nearly 13%. Due to data limitations, it was not possible to adequately test the impact of the library on reading ability (Geurtsen, 2008;Geurtsen & Huysmans, 2008).
Most recently,  studied the effects of the Dutch policy program the Library at School on primary school students' leisure reading and attitude toward reading books in the first year of the nationwide implementation of the program. Based on survey data collected in the school year 2011-2012 from a sample of 4,682 students and 284 teachers from 68 schools, multilevel regression analyses showed that effects of the library at school could not yet be discerned in this starting phase, although slightly positive univariate effects were found. It should be noted that the number of participating schools was limited in this early effect evaluation; hence, statistical power was low on that level.
Although there are numerous studies on the effectiveness of school libraries, little is known about the impact of a school library on subgroups, particularly on groups of disadvantaged and at-risk children (Lonsdale, 2003), such as ethnic minorities. Several studies addressing the impact of school library characteristics on student achievement have attempted to statistically adjust for school and student characteristics, including the students' racial or ethnic background. Linkages between characteristics of school libraries and improved test scores appeared to persist after making such adjustments. For example, flexible scheduling and library staffing were found to exert a positive influence on achievement, regardless of the students' race/ethnicity (Lance et al., 2005;Michie & Chaney, 2009). These findings seem to imply that success factors of school libraries apply to students of various racial/ethnic backgrounds (at least in the United States). To shed more light on the possible influence of school libraries on migrant students, we will next discuss other relevant literature that indirectly address this subject.
If we, for example, follow the line of reasoning of the theory of social and cultural reproduction (cf. Bourdieu, 1992), we would expect only a marginal effect of reading interventions taking place outside the home on migrant students (Broeder, Stokmans, & van Wijk, 2011). After all, the impact of parental socialization (i.e., primary socialization) is granted more importance here than the school context (i.e., secondary socialization; Verboord, 2003). Given their less favorable reading climate at home, interventions would-in this line of reasoning-barely have an enduring effect on migrant children, as the home environment is not supportive to these interventions, especially if one endorses that primary socialization can hardly be compensated by secondary socialization (Broeder et al., 2011).
In line with this, research has shown that parents are of utmost importance. For instance, Stalpers (2005Stalpers ( , 2007, who surveyed students attending Dutch secondary schools, found that the reading climate created by parents (e.g., number of books they read themselves and the frequency they talked about books) had a direct effect on both the students' reading frequency and reading attitude, whereas the reading climate of teachers was only, and to a smaller extent, positively related to the students' reading attitude. In the study of , it was found that parental reading socialization was also a stronger predictor of primary school students' reading behavior and attitude than were reading promotional efforts in schools and classes. With regard to school performance, Veenstra (1999), who analyzed data on around 7,000 students from 150 Dutch secondary schools, pointed to the limited impact of schools. Differences in test scores were for the most part explained by characteristics of the student and his or her family.
However, there are studies suggesting that, besides parents, public libraries and schools are of importance (as was shown in the former section) and that these institutions can even compensate for an unfavorable home climate. Kraaykamp (2002), for example, who analyzed nationally representative survey data on adults, concluded in his study that parents, schools, and libraries are all effective in promoting reading, with no institution standing out. They seem to work complementarily rather than competitively. Verboord (2003) also showed that activities of both parents and schools strongly affect later reading frequency, with primary socialization even seeming to be of slightly less importance than secondary socialization. Leseman and de Jong (1998), who studied a multiethnic sample of 89 families in the Netherlands, found that by the end of first grade, Turkish Dutch children attained far lower vocabulary scores than native Dutch children (the Surinamese Dutch children scored in between), whereas such marked differences were not found with regard to word decoding (i.e., the technical ability to decipher written words). According to the authors, this finding points to the equalizing effects of reading instruction in primary school, given that (at the time of the study) the acquisition of decoding was often strongly emphasized in the early years' curriculum, whereas vocabulary development was often not an explicit part of the first years of primary school.
Furthermore, Rivkin, Hanushek, and Kain (2005) argued that high-quality instruction by teachers throughout primary school could compensate considerably for disadvantages associated with a low socioeconomic background, on the basis of a rich data set providing information on students in the State of Texas. Access to books and libraries has also been found to counter negative effects of poverty on reading achievement scores, as was demonstrated by several largescale studies conducted in various countries (see Krashen, 2011;Krashen et al., 2012). In line with this, Allington et al. (2010), who used student questionnaires and tests to study the effect of a reading intervention on leisure reading and reading achievement, found the strongest effects for students from the poorest families.
With regard to ethnic background, Broeder et al. (2011), who analyzed differences in reading climate between native Dutch and migrant families, argued that reading interventions for youth are probably just as likely to be effective for migrant students as for their native peers, differences in educational level and gender taken into account. According to Broekhof (2011), it is plausible that benefits from language stimulation activities at school are particularly obtained by children from non-Dutch-speaking families (if not compensated by parents talking much in their mother tongue with their children on a challenging level) and where reading to the children is not a common activity. de Haan and Kok (1990), who studied the effects of reading promotion activities in the school context in the city of Utrecht on students' reading motivation and frequency, indeed found stronger effects for children from families with a low developed reading culture. Given the generally less developed reading climate in migrant families, this result would especially apply to children from ethnic minorities.

Discussion
Ethnic inequality in school performance is an urgent policy and research issue. Primary school students with a non-Western migrant background-who make up a considerable proportion of the total school population in the Netherlands and in other Western countries-often face educational disadvantages, especially in language and reading proficiency (Gijsberts & Iedema, 2012;Schnepf, 2007). This is a cause of concern considering the importance of good language and reading skills for participating successfully in society (Johnsson-Smaragdi, 2006;NEA, 2007).
Reducing educational disadvantages is suggested to go hand in hand with reading promotion activities (Hermans, 2002). These are typically aimed at increasing reading frequency, and improving reading and language skills as well as the attitude toward reading (Stalpers, 2005), factors that are related in a reciprocal manner (Broekhof, 2011;Cubiss, 2012;Meelissen et al., 2012;Mol & Bus, 2011a). Schools and libraries are, besides parents, important agents when it comes to reading promotion (e.g., Kraaykamp, 2002Kraaykamp, , 2003.
Nowadays, school libraries in primary schools supported or run by public libraries and aiming at reading promotion are becoming increasingly common in the Netherlands (Oberon, 2009(Oberon, , 2011. Given the investments involved and the importance of reducing educational disadvantages, it is relevant to have insight into the effectiveness of reading promotion efforts by school libraries. By identifying what is known and what is as yet unknown about reducing learning disadvantages through school libraries among children in general, and children from a migrant background in particular, this literature review aimed at pointing to directions for future research. According to the existing literature, it has been established that access to books, as provided by school libraries, results in higher levels of reading (Krashen, 2004b;Krashen et al., 2012). Given the positive relationship between reading behavior and proficiency in reading and language, school libraries seem to have the potential to improve these skills (e.g., Broekhof, 2011;Krashen, 2004b;Mol & Bus, 2011b). In line with this, a considerable amount of research conducted outside the Netherlands has shown that school libraries are related to better learning outcomes for students in general (Clark, 2010;Lance et al., 1993;Ryan, 2004;Scholastic, 2008). Studies conducted abroad stress the importance of adequate school library staff in this context, and attributes such as funding, flexible library access, and adequate collections are also found to be important success factors of school libraries (e.g., Lance et al., 2005;Scholastic, 2008). The limited amount of Dutch research on school libraries has also pointed to positive outcomes, such as more leisure reading among children and a more positive attitude toward books as well as a growth in youth memberships and book loans (Geurtsen, 2008;Oberon, 2011).
However, on the basis of previous research, it is hard to draw firm conclusions about the effects of school libraries' efforts on migrant students in particular. To the best of our knowledge, there are no studies that explicitly focused on the impact of school libraries on migrant students, although research conducted in the United States has suggested that success factors of school libraries apply to students of various racial/ethnic backgrounds. Moreover, literature that indirectly sheds light on the effect of school libraries on children with a migrant background is not unambiguous. On one hand, studies have indicated that the home environmentwhich is usually not that favorable among migrant families in the Netherlands-is of utmost importance, suggesting only a limited impact of (interventions taking place at) other socializing institutions (e.g., Broeder et al., 2011;Stalpers, 2005Stalpers, , 2007Veenstra, 1999). On the other hand, there are studies suggesting that, besides parents, public libraries and schools do play an important role, and that these institutions can even compensate for a reading-unfriendly home climate (e.g., de Haan & Kok, 1990;Kraaykamp, 2002Kraaykamp, , 2003Krashen, 2011;Krashen, 2004b;Leseman & de Jong, 1998;Verboord, 2003). Differences in methods could have contributed to these mixed results, although it should be noted that studies pointing in the same direction varied in methods as well.
The gaps in existing research on the effectiveness of school libraries should be addressed in the future to guide governmental policy in the Netherlands and other Western countries with a considerable proportion of non-Western migrant families. First of all, future research should explicitly focus on children with a migrant background. Although there are numerous studies on the effectiveness of school libraries on children in general, little is known about the impact of a school library on subgroups, particularly on disadvantaged and at-risk children (Lonsdale, 2003), such as ethnic minorities.
Given the hitherto ambiguous research results, it is as yet too early to formulate firm policy recommendations about the contribution of reading promotion efforts in lessening disadvantages for ethnic minority children, and in particular the role of school libraries therein. More research is needed to see if and how so-called "Matthew effects" can be prevented to occur (children from advantaged backgrounds could profit more from the school library's services than their less advantaged counterparts, with a widening rather than a diminishing gap as the result; cf. Cunningham & Stanovich, 1998;Neuman & Celano, 2012).
Future research among migrant students should therefore focus on the possibly differential impact of reading promotion efforts by school libraries on reading behavior, reading attitude, and different areas of reading and language skills, such as vocabulary, reading comprehension, and spelling skills. Furthermore, future research could address whether a possible positive effect of school libraries also indirectly affects the performance of migrant students on other school subjects such as math (cf. Kortlever & Lemmens, 2012). Studies could compare outcomes for migrant children and their native peers, and comparisons could be made between migrant students from various ethnic backgrounds. By having insight into the role of school libraries for migrant students in particular, more effective policies can be conceived. For instance, if school libraries are found to be effective for migrants, supporting school libraries could be incorporated in governmental policy targeted at reducing ethnic inequality in school performance, and collaborations between local governments, schools, and public libraries could be stimulated and formally established.
Furthermore, most studies on the effectiveness of school libraries have been conducted elsewhere, and it is far from clear if reading promotion efforts, such as offered by school libraries, actually are effective in the Netherlands (Bonset & Hoogeveen, 2009;Kraaykamp, 2002Kraaykamp, , 2003Piek, 1995). To mount a strong case for recognizing a positive link between school libraries and student outcomes, and thus to justify policy along these lines, it is important to know whether the findings of foreign studies are also true for the Dutch context (Lonsdale, 2003). In addition, as it has been suggested that to be effective, reading promotion efforts should take place through intensive recurrent contact in a relevant social context (Kraaykamp, 2002(Kraaykamp, , 2003; it seems also desirable to assess the long-term effects of school libraries. Several studies have indicated that the impact of reading promotion activities was related to, for instance, the reading culture at home and poverty (Allington et al., de Haan & Kok, 1990;Krashen et al., 2012). Therefore, future research in the Netherlands and elsewhere could take into account that the impact of school libraries may depend on these and other family characteristics (i.e., there may be interaction effects), but possibly also on characteristics of the child itself (e.g., gender) and the class he or she is in (e.g., reading promotion efforts of the teacher). Moreover, future studies could address what criteria school libraries should meet to be effective for migrant students in, for instance, the Netherlands. By knowing the success factors of school libraries and the groups of children who profit (more) from school libraries, more effective policies aimed at reducing learning disadvantages can be designed, and financial resources can be used more efficiently.
Nowadays, reading also includes digital reading, and Dutch school libraries increasingly pay structural attention to what can be referred to as digital/new literacies or 21st-century literacies/skills. This involves skills and strategies required for the effective use of new technologies (e.g., Coiro, Knobel, Lankshear, & Leu, 2008;Leu et al., 2007;Schmar-Dobler, 2003), such as searching for adequate information on the Internet (Walraven, Paas, & Schouwenaars, 2013). From a policy perspective, it is of the utmost importance that this field will be addressed now that reading behavior may be gradually shifting from print to screens globally.
In conclusion, there seems to be a need for longitudinal research on the possible effects of reading promotion efforts by school libraries (e.g., guidance of a qualified librarian and offering children a varied collection) on migrant students. This is true not only of the Netherlands but also of all countries witnessing ethnic inequality in educational performance. Future research should not only focus on the impact of school libraries' reading promotion efforts on the students' reading and language ability but also on the other factors that reading promotion is directed at and which are found to correlate positively with reading and language skills, that is, the students' reading behavior and attitude toward reading. Attention should also be paid to the possible contribution of school libraries to the students' digital literacies.
Addressing these issues is relevant from a scientific perspective, as gaps in the research literature can be filled by investigating whether a library facility in a primary school leads to an improvement in reading among migrant students. Besides being scientifically relevant, addressing these issues is of great societal relevance, as this can give indications whether or not a school library is an effective tool in reducing disadvantages among migrant students, and should be preferred (or not) above other policy instruments in the Netherlands and elsewhere. These indications can contribute to the empirical foundation of government policy with regard to reading promotion and public libraries. Identifying reading promotion projects that perform well in disadvantageous contexts helps policy makers design effective policy recommendations to overcome inequalities in learning opportunities (OECD, 2007).

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.