2.2.1 Ecological and Developmental Framework

We take a perspective on reading that emphasizes ecological influences on child development, while integrating insights from neurocognitive, sociological, and economic perspectives. An ecological model of human development (e.g., Reference BronfenbrennerBronfenbrenner, 1994; Reference Bronfenbrenner, Morris, Lerner and DamonBronfenbrenner & Morris, 2006) helps us understand the development of individual children as influenced by the variety of social contexts in which they are growing, changing, and learning. Specifically, an ecological perspective on reading development acknowledges the central role of the “microsystem” – proximal processes (e.g., teacher’s instructional practices, parent’s interaction styles) in children’s immediate contexts (e.g., classrooms, homes) – in their development of skilled reading. At the same time, this perspective views these proximal processes and immediate context as shaped by multiple broader exosystems – family social networks, neighborhood contexts, school district practices – and macrosystems – overarching patterns characteristic of a given culture or, in the case of our analysis, a national context.

Applying an ecological perspective to reading, in particular, we focus on what Reference Aaron, Joshi, Gooden and BentumAaron et al. (2008) called the “ecological component” of reading in their Componential Model of Reading, which also includes a cognitive and a psychological component. Like these authors, we see this ecological component as influencing and interacting with the development of cognitive skills (i.e., word reading and linguistic comprehension) and psychological attributes (e.g., motivation, interest) required for successful reading (Reference Kieffer and VukovicKieffer & Vukovic, 2012). Neurocognitive research provides further insight into the nature of cognitive development required for reading. We share the perspective of many scholars that reading is a parasitic system on spoken language; that is, that reading written language is a culturally invented and “unnatural” activity, which draws on “natural” language systems that evolved to serve other purposes for Homo sapiens. Given that assumption, it is not surprising that acquiring language is a nearly universal human accomplishment, but that learning to read complex texts with comprehension is far from universal, even in countries like the United States and Canada with compulsory and free education through adolescence. Basic literacy rates are very high in both countries (e.g., 96 percent of fourth-grade students in each country reached the “low international benchmark” on the 2016 Progress in International Reading Literacy Study [PIRLS]; Reference Mullis, Martin, Foy and HooperMullis et al., 2017), but a large gap exists between these two countries in terms of the level of more advanced literary skills required in them for participation in the modern, global economy and democratic processes.

2.2.2 Ecology of Education in the United States and Canada

To apply these perspectives to a cross-national study of reading inequalities in the United States and Canada, we also start by acknowledging some key similarities and differences in the challenges associated with reading in these two countries. First, given the position of English as the common language in the United States and as one of two official languages in Canada, learning to read in these two contexts requires learning the same alphabetic writing system, with a deep orthography (i.e., a spelling system characterized by many inconsistencies and complexities in grapheme–phoneme correspondences; Reference Seymour, Aro and ErskineSeymour, Aro, & Erskine, 2003). Students learning to read in French in Canada are also learning an alphabetic writing system with a moderately deep orthography (e.g., Reference Seymour, Aro and ErskineSeymour et al., 2003). So, the cognitive challenges in learning to read are quite similar for Anglophone students and roughly comparable for Francophone students in Canada as well. (That said, French readers may face particular challenges due to the language’s higher frequency of historically preserved morphological patterns that produce speech-to-spelling inconsistencies; Reference Deacon, Desrochers, Levesque, Verhoeven and PerfettiDeacon, Desrochers, & Levesque, 2017). Second, the United States and Canada, as OECD countries with long histories of stable government, have relatively similar levels of economic and institutional development. As a result, both countries offer free and compulsory education through secondary school, with relatively large levels of average investment in public education (e.g., NCES, 2016; UNESCO, 2016 ). Despite these similarities, the two countries differ in their ecologies of education in major ways. Importantly, these include major differences in the social services available, including the availability in Canada of universal, nationalized healthcare and more robust economic support in Canada than that found in the United States (e.g., Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al., 2015).

Third, as part of the legacy of settler colonization, the United States and Canada have both been defined by immigration since their founding. They have both varied over time in their treatment of new waves of immigrants – from encouragement through grudging acceptance to xenophobic rejection (Environics Institute for Survey Research, 2019; Reference Waters and PineauNational Academies of Sciences, Engineering, and Medicine, 2015). As these broader forces inevitably affect schools, questions about the education of immigrants and the children of immigrants have been politically charged and answered differently over time. Foremost among these have been questions about schools’ responsibilities as regards to meeting the language- and content-learning needs of language-minority learners, that is, students who speak a language other than the society’s dominant language(s) at home. Of course, the other legacy of settler colonialism in both countries since their inception has been policies and practices that sought to eradicate Indigenous peoples and their ways of being. These have had the cumulative effect both of wiping out Indigenous languages and underfunding Indigenous education efforts, leading to significant education and opportunity gaps for Indigenous students in both countries (NCES, 2019; Truth and Reconciliation Commission of Canada, 2015).

In both countries, education for language-minority and immigrant learners has been defined by major tensions. In the United States, on one hand, these students benefit from civil rights protections, including the right to equal education opportunities guaranteed by the United States federal court decisions (Lau v. Nichols, 1974; Castañeda v. Pickard, 1981; Plyler v. Doe, 1982). On the other hand, US schooling has been characterized by persistent negative attitudes toward bilingual education and immigrant students that produce major inequalities in resources and opportunities to learn (e.g., Reference Gandara, Rumberger, Maxwell-Jolly and CallahanGandara et al., 2003). Canada has a more decentralized educational system, whereby each province has jurisdiction over education. Under this system, there has been relatively little federal attention to teaching language-minority learners in their native language, but the country also has official policies that embrace bilingualism (at least English/French bilingualism) and promote multiculturalism. There has also been a recent resurgence in Indigenous-language revitalization in Canada.

The immigration policies of the two countries have differed in major ways historically and continue to do so, with Canada accepting most immigrants using a points system that favors highly educated and skilled workers while the United States system prioritizes family reunification rather than education and skills (e.g., Reference BloemraadBloemraad, 2006). It is worth noting that the geographic isolation of Canada from the rest of the world – surrounded by oceans to the east, west, and north, and sharing a border with only one other country – has also facilitated highly selective immigration criteria and policies. Given the two countries’ demographic and political histories (and futures), describing the reading development of language-minority learners in the United States and Canada is central to understanding how diversity and structural inequalities in both countries play out in schools.

The two countries also have different histories and public rhetoric concerning multilingualism and multiculturalism. In Canada, French and English have had equal official status in government and all of the services it controls since the Official Languages Act of 1969; in the United States, there is no official language by law, but English remains the overwhelmingly predominant language in government and schooling, with only minor protections for other languages, most of which have been introduced as a result of civil rights actions rather than democratic processes (e.g., Reference CrawfordCrawford, 2004). Moreover, Canadian law (i.e., the 1982 Canadian Charter of Rights and Freedoms; the 1985 Canadian Multiculturalism Act) affirms an official policy of recognizing and promoting multiculturalism (Reference Ambrose and MuddeAmbrose & Mudde, 2015). This proactive encouragement of multiculturalism goes beyond the prohibitions against discrimination that characterize the corresponding US laws (e.g., the 14th Amendment; the 1964 Civil Rights Act). (At the same time, Canada’s Indian Act still engenders discriminatory practices and limits Indigenous sovereignty [Reference JosephJoseph, 2018], which contradicts the spirit of Canada’s commitment to multiculturalism.) Finally, the two countries differ in the fundamental scale of their societies and educational systems. In the United States, roughly 50 million public K-12 students are taught by over 3 million teachers (NCES, 2015), whereas in Canada, roughly 5 million students are taught by approximately 700,000 teachers (Stats Canada, 2016). By point of comparison, New York City schools serve more students than those of every province and territory in Canada, except for the two largest, Quebec and Ontario.

2.2.3 Sociocultural Predictors of Reading Development

Understanding the ecological differences between the United States and Canada leads us to construct hypotheses about the two sociocultural predictors that are our focus: SES and linguistic diversity. Regarding SES, the stronger social safety net available to children living in poverty in Canada compared to that in the United States would suggest smaller effects of SES in the former country than in the latter, given the importance of health and well-being to reading development (e.g., Reference Shonkoff, Garner and SiegelShonkoff et al., 2012; Reference Snow, Burns and GriffinSnow et al., 1998). Regarding linguistic diversity, Canada’s more selective immigration system is likely to result in language-minority learners with more resources and consequently better reading outcomes compared to their counterparts in the United States. That said, if these differences are associated primarily with differences in socioeconomic resources, we would expect smaller differences between language-minority learners in the two countries once we have taken SES into account.

Given our ecological perspective, we resist attributing differences by SES and home-language background to individual students’ characteristics or to those of their families or teachers. Rather, we interpret these differences as the culmination of complex interactions among broad structural inequalities rooted in cultural and historical patterns of oppression and marginalization. These patterns play out in a variety of proximal processes in classrooms that constrain opportunities to learn and develop for some students, while maximizing these opportunities for others.

As we have seen, written text (from orthography through discourse patterns) is “unnatural” and culturally specific. Given this, it is not surprising that reading development will be linked to children’s access to and opportunities to learn the oral language skills for the languages that are privileged in written school texts. For example, when language-minority learners enter kindergarten with limited oral proficiency in the language (or languages) of instruction, we can expect that their trajectories toward proficient reading will differ from those of their peers who enter with strong proficiency in the language of instruction. (That said, language-minority learners’ trajectories often converge with those of language-majority students over time.) Moreover, language-minority students disproportionately come from low-income backgrounds, particularly in the United States (and to a much lesser extent in Canada), so disentangling the effects of poverty and home language on reading development is important to understanding these students’ academic trajectories.

2.2.4 Evidence of Disparities in Reading Achievement

To synthesize evidence on the magnitudes of disparities in reading achievement by SES and home language, we leveraged existing published analyses and conducted new analyses of large-scale, nationally representative datasets. Specifically, we leveraged analyses by an international team of scholars (Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al., 2011, Reference Bradbury, Corak, Waldfogel and Washbrook2015; Reference Washbrook, Waldfogel, Bradbury, Corak and GhanghroWashbrook et al., 2012) of the US Early Childhood Longitudinal Study – Birth, 2001 cohort (ECLS-B) and Kindergarten, 1998 cohort (ECLS-K:98) as well as the Canadian National Longitudinal Study of Children and Youth (NLSCY) and analyses (Kieffer, 2008, 2011, 2012) of the ECLS-K:98 by one of this chapter’s authors. We conducted new analyses using PIRLS 2006 and 2016 for fourth-grade reading in both countries. When possible, we report results from the most recent PIRLS administration, but rely on earlier datasets when they are the only administrations to include important variables of interest (e.g., parental education). We also conducted new analyses using the OECD Programme for International Student Assessment (PISA) data from 2018 as well as reports from PISA 2009 (which included useful analyses of home-language controlling for SES). Finally, we conducted new analyses with the US National Assessment for Educational Progress (NAEP) data from 2019 for SES and from 2015 for home language (the most recent year when the appropriate variable was available). All new analyses were conducted with online data explorers that appropriately account for weighting, matrix sampling, and other features of the datasets. All of the datasets are nationally representative, so include the full range of racial–ethnic and linguistic diversity in each country. In particular, the participating language-minority learners spoke a variety of languages, though the ability to conduct analyses separately by individual language group was limited.

In reporting and interpreting results, we present effect sizes, that is, standardized differences expressed in standard deviation units. In this way, we focus on the question of the size of the differences found, rather than simply whether differences exist. We interpret the effect sizes using Reference CohenCohen’s (1992) conventions, where 0.2 is small, 0.5 medium, and 0.8 large. This emphasis on magnitudes over statistical significance tests is appropriate because trivially small differences can be statistically significant in large datasets. In addition, effect sizes allow us to make rough comparisons across datasets for which statistical significance tests are not appropriate.

2.3.1 Role of SES in Language Skills at School Entry

Differences in students’ language skills when they enter school provide key insights into the structural inequalities that will play out later, given the well-established role of early reading-related language skills in predicting later reading development (e.g., Reference Snow, Burns and GriffinSnow et al., 1998). Fortunately, this is also a topic on which high-quality and highly comparable data has been collected in both the United States and Canada, particularly recent work by an international team of scholars (Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al., 2011, Reference Bradbury, Corak, Waldfogel and Washbrook2015; Reference Washbrook, Waldfogel, Bradbury, Corak and GhanghroWashbrook et al., 2012). Specifically, they analyzed data from the Canadian NLSCY and from the US ECLS-K and ECLS-B. Here, we focus on their comparisons of the oral vocabulary skills of four- and five-year-olds in the two countries.

Their results indicated substantial differences in early vocabulary by SES in both countries, with much larger differences in the United States relative to Canada, as shown in Table 2.1. The difference in vocabulary scores between students with at least one college-educated parent and those with parents with high-school education or less was 1.22 standard deviations (SDs) in the United States and 0.66 SDs in Canada. Similarly, when estimated using family-income quintile, differences were large in both countries, but more so for the United States than Canada. Given the potential differences in structural inequalities caused by other demographic characteristics in the two countries, we also provide estimates from Bradbury et al., which have been adjusted for race/ethnicity and parental immigrant status in Table 2.1. These adjusted estimates remained large, but were more similar across the two countries. This suggests that these other structural inequalities, which are historically and fundamentally intertwined with economic inequality, explain some of the differences observed between the United States and Canada.

2.3.2 Role of SES in Learning to Read

After four or five years of schooling, differences in reading achievement by SES continue to be moderate to large in both countries, as shown in Table 2.2. The available data did not allow direct comparisons between Canada and the United States, but nonetheless supported the general claim that differences in both countries are substantial. Consistent with results for early childhood, differences by parental education were moderate to large. For instance, for Canada, reading gaps by parental education in fourth grade on PIRLS 2006 were moderately sized (0.53 SDs), though smaller than the average for all countries participating in PIRLS 2006 (0.66 SDs) and slightly smaller than the corresponding gaps in early childhood vocabulary reported above. (We used PIRLS 2006 rather than PIRLS 2016 here because this was the last year in which parental education was measured.) Comparable PIRLS data were not available for the United States, which did not participate in all aspects of PIRLS data collection, but Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al. (2011) estimated these same gaps using ECLS-K data to be large in both third grade (0.99 SDs) and fifth grade (1.03 SDs).

Differences in reading by other SES indicators were also moderate to large. The PIRLS 2016 data for Canada include a home educational resource index, which included parental education and occupation, number of books in the home, and other study supports (i.e., having one’s own room to study; having an internet connection). On this measure, achievement differences between Canadian children with different levels were moderate (0.48 SDs) and smaller than the average for thirty-one OECD countries which participated in PIRLS 2016 (0.66 SDs). Again, such analyses were not possible for the United States. Instead, if we look at differences by income on the Grade 4 NAEP 2019, reading differences by free- or reduced-price lunch status were large (Table 2.2). Similarly, on the Grade 3 and Grade 5 waves of ECLS-K, differences by an SES composite (i.e., a combination of parental education, income, and occupation and the school’s percentage of students on free and reduced-price lunch) were also moderate to large in size for US children (Reference KiefferKieffer, 2012). Although these estimates are not strictly comparable across the two countries, they support the general conclusion that reading-achievement differences by SES in middle childhood are substantial in both countries.

Longitudinal studies corroborate these cross-sectional findings of substantial differences but also indicate some narrowing of SES gaps in each country between kindergarten and Grade 3. For the United States, Reference KiefferKieffer (2012) found that SES gaps in overall reading achievement narrowed by about 30 percent between kindergarten and Grade 3 (see also Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al., 2015). For Canada, a large epidemiological (but not nationally representative) study conducted in one city indicated a dramatic narrowing of gaps by neighborhood SES on a word-reading measure (Reference D’Angiulli, Siegel and HertzmanD’Anguilli, Siegel, & Hertzman, 2004). The latter study was not comparable to Reference KiefferKieffer (2012) and Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al. (2015), due not only to differences in sampling and SES indicators but also to fundamental differences in the reading skills measured; the greater narrowing of gaps found by D’Anguilli et al. may be specific to word-reading skills, which would be consistent with constrained-skills theory (Reference ParisParis, 2005). Nonetheless, these longitudinal findings suggest some degree of narrowing of SES gaps between early and middle childhood – though the cross-sectional results make it clear that large gaps remain.

2.3.3 Role of SES in Adolescent Reading Development

Similar to earlier developmental periods, evidence suggests moderate to large differences between levels of adolescent reading comprehension in both countries. On the most recent PISA in 2018, both countries demonstrated substantial differences between students in the top and bottom quartiles on an SES composite. As shown in Table 2.3, these large differences were somewhat greater in the United States (0.92 SDs) than in Canada (0.68 SDs), with the latter substantially smaller than the OECD average (0.90 SDs).

Other national datasets specific to the United States, including NAEP 2019, also suggest relatively large differences by SES indicators (Table 2.3). These gaps are reduced somewhat when adjusting for race/ethnicity, but remain substantial. Moreover, longitudinal studies using data from ECLS-K indicate that reading-achievement gaps widened between Grade 3 and Grade 8 (Reference KiefferKieffer, 2012; Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al., 2015), in contrast to these studies’ findings that the same gaps narrowed between kindergarten and Grade 3.

2.3.4 Summary of SES Differences

Looking across early childhood, childhood, and adolescence, differences in reading by SES are moderate to large in magnitude, regardless of the SES indicator used. In early childhood, these differences are notably larger in the United States than in Canada. Childhood differences are also moderate to large in magnitude in both countries, but cannot be compared directly. Canada’s differences are smaller than international averages in childhood. By adolescence, differences remain large in both countries, with larger differences again in the United States compared to Canada. In adolescence, Canada’s SES differences are smaller than the international average, whereas the United States’s differences are similar to this average.

2.4.1 Linguistic Diversity at School Entry

Comparisons of national datasets indicate large differences in vocabulary (measured in the dominant language of English or French) between language-minority and language-majority students when they enter school in both the United States and Canada, as shown in Table 2.4. These results come from the same cross-national research team mentioned above (Reference Washbrook, Waldfogel, Bradbury, Corak and GhanghroWashbrook et al., 2012). Here, language-minority students were defined as those who spoke a language other than the dominant language(s) at home and were either immigrants or native-born children of immigrants. Language-majority students were those who had native-born parents and spoke English (or French in the case of Francophone Canadians) primarily at home. As shown in Table 2.4, the differences in vocabulary scores were similar across the two countries. In addition, adjusting these estimates for income, parental education, and other SES indicators (see note, Table 2.4) did not lead to a substantial reduction in these differences, indicating that large gaps in early language are demonstrated between language-minority and language-majority students, even when comparing those from similar SES backgrounds. This is consistent with the findings of several US studies conducted with smaller samples of Spanish-speaking language-minority learners, which suggest that they enter school with underdeveloped English vocabularies, relative to national norms on standardized assessments (e.g., Reference Mancilla-Martinez and LesauxMancilla-Martinez & Lesaux, 2011a, Reference Mancilla‐Martinez and Lesaux2011b; Reference Manis, Lindsey and BaileyManis, Lindsey, & Bailey, 2004).

2.4.2 Linguistic Diversity and Learning to Read

In comparison to results for early childhood vocabulary, differences in reading achievement in Grade 4 between language-minority and language-majority learners appear to be much smaller, particularly in Canada, as shown in Table 2.5. On the 2016 PIRLS, differences in reading comprehension between language-majority and language-minority students were moderately sized in the United States (0.34 SDs) and small in Canada (0.04 SDs). Specific to this survey, the United States gap was similar to differences in other OECD countries (0.32), while the Canadian gap was notably smaller. It is worth noting that the definition of language minority was not identical to the one used so far. Here, language-minority students included those who reported “sometimes” speaking the dominant language(s) at home, while language-majority students include those who reported “always” speaking the dominant language(s) at home; only a trivial number of students (3 percent in Canada, 1 percent in the United States) chose the other answer choice of “never” speaking the dominant language(s) at home, consistent with prior evidence that the vast majority of language-minority learners speak a combination of their home language and the dominant language, particularly after they enter school (e.g., Reference Mancilla-Martinez and KiefferMancilla-Martinez & Kieffer, 2010; Reference Mancilla-Martinez and LesauxMancilla-Martinez & Lesaux, 2011a). Although this definition is not strictly comparable to the one used by Reference Washbrook, Waldfogel, Bradbury, Corak and GhanghroWashbrook et al. (2012), they likely capture roughly similar populations of students.

Given differences in immigration policies and patterns that in turn yield differences in the SES of immigrant families in the two countries, we also provide estimates adjusted for a comparable index of SES available from PIRLS 2016 (i.e., the presence of home study supports). The adjusted estimate was very close to 0 for both Canada and the United States, indicating that SES explained nearly all of the apparent differences between language-minority and language-majority students in both countries. This represented a much greater reduction for the United States, indicating greater confounding of SES and home-language backgrounds.

Similarly, United States-specific evidence from national datasets suggests moderately sized differences between language-minority and language-majority learners that decline to small or trivial magnitudes when adjusting for SES. Specifically, analyses of ECLS-K indicated a moderately sized reading achievement gap (0.48 SDs) in Grade 3 that became very close to 0 after adjusting for an SES composite and school concentration of poverty (see also Reference KiefferKieffer 2008, Reference Kieffer2011). Results for the Grade 5 wave of ECLS-K were similar. NAEP 2015 results indicated a moderately sized gap (0.35 SDs) that became small when adjusting for students’ eligibility for free or reduced-cost lunch (0.15 SDs). Across these datasets, US language-minority learners demonstrated reading levels that were substantially below national averages, but quite similar to those of their language-majority counterparts from similar SES backgrounds.

2.4.3 Linguistic Diversity and Adolescent Reading Development

In contrast to evidence for early childhood vocabulary and childhood reading, the PISA 2009 and 2018 data indicate relatively small US reading achievement differences by home language in adolescence (0.33 and 0.27 SDs, respectively); see Table 2.6. These differences were also somewhat smaller than OECD averages. Moreover, for the 2009 PISA, when adjusted for SES (using the PISA composite), this difference reversed in direction, such that language-minority learners outperformed language-majority learners to a small degree (-.10 SDs).Footnote ¹ For Canada, PISA data indicated a small gap (0.13 in 2009, 0.24 in 2018) by home language that was notably smaller than the OECD averages. Adjusting for SES, the 2009 gap was reduced somewhat (to 0.08 SDs). Here, language-minority children included those who spoke a language other than the dominant language(s) “most of the time” at home.

Similarly, United States-specific evidence from NAEP 2015 suggests a moderately sized reading-achievement gap in Grade 4 (0.34 SDs). As with PISA, the NAEP gap declined to a trivial size (0.02 SDs) when adjusting for SES (indexed by student eligibility for free or reduced-cost lunch and student-reported parental education). In additional converging evidence from ECLS-K, the United States reading-achievement gap by home language in Grade 8 was smaller than in earlier grades and reversed in direction to favor language-minority learners when adjusting for SES (Reference KiefferKieffer, 2011).

2.4.4 Summary of Home Language Differences

Looking across time, we see that large vocabulary differences between language-minority and language-majority students in both countries at kindergarten entry turn into moderate differences in reading in the United States and small differences in Canada by middle childhood. By adolescence, these differences become small-to-moderate in the United States and remain small in Canada. Moreover, while SES differences do not explain much of these initial vocabulary differences in kindergarten, they do explain a substantial proportion of later reading differences in middle childhood and adolescence. Comparing students from similar SES backgrounds in later grades, we find only small-to-trivial differences between language-majority and language-minority students, with some analyses indicating better performance for language-minority learners. Given the greater confounding of SES and language background in the United States, more of the observed differences are explained by the low SES, on average, for language-minority learners in the United States, compared to Canada.

2.5.1 Promoting Equity across SES Differences

Our conclusions regarding SES echo those of Reference Bradbury, Corak, Waldfogel and WashbrookBradbury et al. (2015) and provide further support for their suggestion that structural inequalities in both the United States and Canada prevent far too many students from low-SES backgrounds from reaching their potential to become skilled readers. As Reference Reardon, Duncan and MurnaneReardon’s (2011) analyses of national US datasets from the 1940s through the 2000s suggest, these inequalities are greater now than they have ever been since the government began collecting such data. Canada appears to have had some success relative to the United States, but does not escape notable SES differences. Indeed, achievement differences in both countries are large across early childhood, childhood, and adolescence. (For an international perspective, see Nag, Chapter 15 in this volume.)

From an ecological perspective, this relative success on the part of Canada in producing more equal outcomes might reflect influences at local levels (e.g., differences in childcare experiences) as well as influences at broader levels that prioritize social supports (e.g., maternity leave, universal healthcare). Bradbury et al. note these public investments in families as key levers that promote more equity in Canada and that should be considered in public policy discussion about education in the United States. Noting the larger differences by SES in the United States than in Canada at school entry, they argue for the particular importance of investment in early childhood development – including evidence-based preschool and parenting-support programs as well as income-support programs to reduce poverty and its associated stresses among families with young children. Historically, the United States has made large investments in K-12 education, while leaving working families on their own to pay for the education and development of children before age five – precisely the time at which children’s brain development is most influenced by environmental affordances (e.g., Reference Yoshikawa, Raver and MorrisYoshikawa, Raver, & Morris, 2016).

At the same time, our results suggest that both countries could also do more to promote equity among elementary and secondary students. There is a strong and growing global research base on which to recommend specific instructional approaches for improving reading outcomes and how they can be tailored and implemented to meet the needs of students from low-SES backgrounds. From systematic phonics instruction to rich oral-language instruction, many curricula and instructional approaches have been found to be particularly beneficial to students from low-SES backgrounds (see Dickinson et al., Chapter 17, and Schwartz, Chapter 19, in this volume). Although instructional questions remain, we certainly know enough to improve classroom reading instruction in important ways.

2.5.2 Understanding Linguistic Diversity in Relation to SES

Our conclusions regarding language-minority learners underscore the long-term nature of the process of learning to become a proficient reader in a second language. The fact that language-minority students arrive at school with underdeveloped second-language vocabulary but many go on to catch up with their language-majority peers by adolescence aligns with a wide consensus that language learning is a long-term developmental process (National Research Council, 2017). Despite this widespread consensus, educational policies and practices continue to emphasize the short-term underperformance of language-minority learners in ways that may actually hurt their long-term success (e.g., Reference Hopkins, Thompson, Linquanti, Hakuta and AugustHopkins et al., 2013; Reference Kieffer and ThompsonKieffer & Thompson, 2018). In particular, many US policies and practices focus attention on the subset of language-minority learners who are currently and temporarily classified as English learners. This group is defined by their current lack of proficiency in speaking, listening, reading, and writing English. Thus, differences between this selected subgroup and native English speakers is a “gap that can’t go away” (Reference Saunders and MarcellettiSaunders & Marcelletti, 2013, p. 139). Recent analyses of NAEP data over time (Reference Kieffer and ThompsonKieffer & Thompson, 2018) indicate that when one analyzes the entire population of language-minority or multilingual learners, one finds they have made substantial progress in reading and math that is not apparent when analyzing only current English learners.

At the same time, the data provide some support for the hypothesis that the process of learning the dominant language and catching up with language-majority peers is quicker in Canada than in the United States. Canadian gaps by home language are similar to US gaps in early childhood vocabulary, but much smaller by Grade 4. In addition to the social supports described in Section 2.5.1, these differences might reflect how Canadian provinces tend to have intentional educational policies in place to support immigrant students in learning the language of instruction (OECD, 2006). Such policies, in turn, might reflect macrosystem values as reflected in Canada’s commitment to the protection and preservation of multiculturalism through federal law (i.e., Canadian Charter of Rights and Freedoms, 1982; Multiculturalism Act, 1985). That said, it is important to note that Canada is not immune to anti-immigrant and xenophobic sentiments; a recent report on Canadian public opinion revealed that 64 percent of respondents felt that Canada accepts too many immigrants from racial minority groups, and 50 percent felt too many immigrants coming into Canada were not adopting Canadian values (Environics Institute for Survey Research, 2019). Nonetheless, it does appear that Canada’s commitment to multiculturalism creates a more conducive learning environment for immigrant students compared to countries without explicit educational integration policies (OECD, 2006); if so, this may be an example of policy making a difference in spite of countervailing popular sentiment. However, before one takes too rosy a view of Canada’s relative success with language-minority learners, it is worth reiterating that its selective immigration policies favor highly educated and skilled parents.

In addition, our conclusions underscore how much the achievement of US language-minority learners may be constrained by structural inequalities associated with their disproportionately low SES. Although converging evidence suggests that US language-minority learners catch up with their peers from similarly low-SES backgrounds, this is a low bar, given the poor performance of those peers. One potential implication is that US educators may be overemphasizing the need for interventions specific to language-minority and English learners, rather than integrating such efforts into broader approaches to reduce inequalities by socioeconomic status.

This pattern of confounding SES and home language background does not appear to the same extent for Canada. This is likely due to the fact that very different immigration policies and practices are in place in Canada than those in the United States, but also perhaps to a variety of Canadian social policies that reduce the structural inequalities that would otherwise constrain the reading development of both language-minority learners and students from low-SES backgrounds.

2.5.3 What We Can’t Learn from Current Large-Scale Datasets

Although the available data support some interesting conclusions, there remain many questions that cannot be easily answered from the existing large-scale datasets. First, although national and international assessments provide valuable insights into reading development, they pay little or no attention to writing development in either country. Given the importance of writing in college success and access to meaningful employment (e.g., Reference Levy and MurnaneLevy & Murnane, 2004), more attention to writing in large-scale research is warranted. Similarly, oral language is rarely assessed and when it is, it is assessed in superficial ways, despite its key role in literacy development (e.g., Reference Snow, Burns and GriffinSnow et al., 1998). In addition, returning to the ecological approach we have described, there are many variables within the exosystems (e.g., neighborhood contexts, school practices) and microsystems (e.g., instructional practices) that cannot be captured effectively by national and international assessments.

Second, the differences between the United States and Canada in their systems and investment in collecting educational achievement data mean that we have more complete answers for many questions about achievement differences in the United States than we have for Canada. Unlike the United States, Canada maintains few national data-collection systems – in education or otherwise – leaving to individual provinces and territories the responsibility of identifying, collecting, and tracking relevant metrics, including demographic identifiers such as race/ethnicity or Indigenous status. In the context of education, individual provinces/territories implement regular testing programs that focus primarily on language arts and numeracy in both elementary and high school, with some provinces also assessing science and social studies in high school. However, there is little consistency across provinces/territories in what content is assessed or when, resulting in functional barriers to making comparisons across demographic groups even if such data were available. Canada also lacks longitudinal datasets, which limits the ability to monitor changes and trends over time.

Third and more problematic in terms of identifying and correcting for structural inequalities as a result of Canada’s colonial history, provinces/territories have not devised a way to monitor achievement gaps between Indigenous and non-Indigenous Canadians, nor are provinces required to disaggregate data by demographic characteristics. Canada’s legacy of the genocide of Indigenous peoples is well documented (Truth and Reconciliation Commission of Canada, 2015), but Canada also has a history of discrimination against immigrants of Asian and African descent (Reference Ambrose and MuddeAmbrose & Mudde, 2015). The inability to comparably disaggregate data across provinces/territories obscures the successes of marginalized communities and failures of the education system for historically disenfranchised groups over time. This is especially concerning for Indigenous communities, who have experienced repeated neglect from the Canadian government. A recent call to action by Indigenous leaders explicitly summons the federal government to regularly publish educational attainments of Indigenous peoples compared with non-Indigenous peoples (Truth and Reconciliation Commission of Canada, 2015). There is also a need to move beyond standard testing programs that focus only on academic attainment to include outcomes that are important for Indigenous peoples, such as the social, physical, and spiritual wellbeing of individuals and communities (Reference Friesen and KrauthFriesen & Krauth, 2012). The most important message for Canadian policymakers is to work with Indigenous communities to ensure that the right data are collected, policies written, and programs implemented.

2.5.4 Structural Inequalities and Reading Development Revisited

Even now, decades into the twenty-first century, the promises of public education to be the great equalizer remain unfulfilled. In the United States and Canada, structural inequalities in opportunities to learn the complex reading skills required for the global economy persist. So, too, do the legacies – and continuing practices – of racism, classism, xenophobia, and oppression of Indigenous peoples, thereby perpetuating structural inequalities inside and outside of schools. To begin to address unequal outcomes in a real way, we must begin by considering how we make sense of the differences in achievement that we see (see also Navas, Chapter 3, Araújo and da Costa, Chapter 5, and Morgan et al., Chapter 10 in this volume). Too often, these are interpreted as evidence of how students from low-income or language-minority backgrounds are failing to reach the standards set by the educational system; too rarely are these differences interpreted as evidence of how the system is failing them.

4.2.1 Sociolinguistic Background

Sociolinguistic studies provide evidence that the development of literacy has important consequences for individuals and societal communities. It guides personal and communal communication and serves social, economic, and political actions in society (cf. Reference StubbsStubbs, 2013). In the Caribbean context, as societies make the transition from colonization to autonomy, careful language planning is needed, as well as an appreciation of social and cultural resources which does not exclude the possibility of their utilization. The Caribbean islands are among the Small Island Developing States (SIDS), which were first recognized as a distinct group of developing countries at the United Nations Conference on Environment and Development in June 1992. They include countries that tend to share similar sustainable development challenges, including historical underachievement in terms of school success.

In many cases, a Creole variety is the native language of a location’s inhabitants, or the predominant language heard at home. Creole languages derive from pidgins – language varieties that emerge when groups of people, mostly as a result of population movement, develop a new language by simplifying their own language(s) and adding words and phrases from other contact languages. In case a new generation of learners acquires the emerging variety as their home language, the pidgin is then redefined as a Creole (see Reference SiegelSiegel, 2005). Creoles such as Papiamentu in the Dutch Caribbean and Kwéyòl in St. Lucia are autonomous in that the colonial language is not the lexifier language, whereas Creoles such as Jamaican and Trinidad Creoles can be considered more continuous. Creoles can be seen as contact languages in situations where groups of people have to communicate with each other without the availability of a common language (cf. Reference Migge, Léglise, Bartens, Migge, Léglise and BartensMigge, Leglise, & Bartens, 2010a). The most widely spoken Creole is Haitian Creole with more than 10 million speakers. It is estimated that more than 75 million people speak a Creole as their native language, of which the majority develop literacy in a second language, often the colonial language (Reference SiegelSiegel, 2005).

The focus of language planning in Caribbean societies during the past decades has been on the processes and individual variation of literacy and school success in L1 Creole and L2 colonial language(s). The low educational achievements of these societies can, to a large extent, be explained as resulting from a mismatch between the language of the population (mostly a Creole language) and the school language (mostly the ex-colonial language). Only gradually has investment been provided for the codification and further development of Creole languages. Reports of large numbers of class repeats and school dropouts have urged for a broad scientific evaluation of these countries’ educational systems, with a special focus on the role of Creole language, not only as a school subject but also as a possible tool to enhance the development of literacy and school success in the second language (see Reference Migge, Léglise, Bartens, Migge, Léglise and BartensMigge et al., 2010a). Programs of action have been proposed to assist the SIDS in their sustainable development efforts, with a strong focus on the development of human resources via education. Such programs vary widely and appear to be linked with the nature of the individual Caribbean country’s remaining colonial ties, the nature of linguistic diversity, the developmental state of the indigenous language(s), and the language attitudes among the speakers. In most societies passing through a transitional period from colonial to postcolonial time, the tension between the position of indigenous languages and the colonial language brings about an upgrading of the mother tongues and even a takeover of the position of the colonial language (cf. Reference SiegelSiegel, 2005). Decolonization involves not only the appreciation of the mother tongue over a colonial language, but also the strengthening of autonomy and political power. However, in many cases the drive to displace colonial languages tends to be softened because of economic or political forces, usually resulting in a slowing down of the pace of enactment of new laws. According to Reference ChurchillChurchill (1986), the strengthening of indigenous languages generally proceeds stage-wise. It starts out with recognition of their existence as a vehicle for communication and education in the local community. Then it is followed by an extension, a consolidation, and an adaptation phase, leading to multilingual coexistence.

Depending on the situation in a society, language planning may take different forms (see Reference AgerAger, 2001; Reference Appel, Verhoeven, Arends, Muysken and SmithAppel & Verhoeven, 1994). Usually, the first task in a postcolonial state is to determine which language(s) should fulfill the role of national language. In the vast majority of cases, the colonial language is abandoned in favor of one of the indigenous languages. In addition to processes of language selection, the position of minority languages must be dealt with. Accordingly, decisions need to be made about the developmental support of languages used in a society. A common procedure for language support is that of codification, which can be seen as a prerequisite for the standardization of a language. Codification includes at least an explicit statement of the model for standard language usage via dictionaries, grammars, spelling, punctuation, and pronunciation guides, and reference materials. It is important to note that any language variety is modifiable in such a way that orthographies can be developed and lexical elaboration provided. As Reference SiegelSiegel (2010) has pointed out, the development of a standard form of a Creole differs from that of other language varieties in that the Creole needs to be made autonomous from its lexifier so that it is perceived as a separate language. Moreover, the standardized Creole needs to be accessible to the majority of its speakers.

4.2.2 Educational Background

Language policy primarily manifests itself in two domains: the mass media (including social media) and education. As to the mass media, it is a matter of policy to what extent the (post)colonial and indigenous language will perform public functions in the community. Measures in this domain concern the use of written language in institutions, periodicals, and libraries. As to education, different perspectives can be taken on the legal opportunities for both indigenous and (post)colonial literacy education (Reference SiegelSiegel, 2005). There can be a monolingual approach focusing on the indigenous language, or the colonial language, only. There can also be a transitional model in which the indigenous language is used as a vehicle to better learn to become literate in the second language. Children who have learned the writing code in one language do not need to start from scratch in the other. Basic literacy skills can be easily transferred from one language to the other (Reference Chen, Dronjic and Helms-ParkChen, Dronjic, & Helms-Park, 2015). Finally, there can be a model focusing on full biliteracy. In such a case, literacy instruction may start from the indigenous language, the colonial language, or simultaneously from both languages. For a review of the effects of alternative biliteracy instruction models, see Reference Verhoeven, Kamil, Pearson, Moje and AfflerbachVerhoeven (2010, Reference Verhoeven, Cain, Compton and Parilla2017). Although evaluations of multilingual programs in the Caribbean are only beginning to emerge, it can be concluded that the implementation of Creole literacy in the school curriculum brings about an awareness that the native language can be seen as an educational resource for higher motivation and better learning outcomes among students (Reference PereiraPereira, 2018; Reference SiegelSiegel, 2010).

With respect to literacy development in a Caribbean context, an important question is whether learning to read should be taught first in the child’s indigenous language (L1) and then in the second language (L2); simultaneously; or the other way around. It can be argued that the acquisition of literacy will be facilitated when the instruction links up with the child’s linguistic background (Reference CumminsCummins, 2001). Moreover, the child’s motivation and self-concept may increase if the instruction is given in their native language (Reference SpolskySpolsky, 2000). Importantly, there can be linguistic transfer in bilingual reading instruction in that children who have learned to read in one language do not have to start from scratch in another: What they do have to learn is a new written language code. With respect to the acquisition of academic language skills, Reference CumminsCummins (1991: 85) has hypothesized the role of interdependence as follows: “To the extent that instruction in a certain language is effective in promoting proficiency in that language, transfer of this proficiency to another language will occur, provided there is adequate exposure to that other language (either in the school or environment) and adequate motivation to learn that language.” The interdependence hypothesis predicts that optimal input in one language leads not only to better skills in that language but also to a deeper conceptual and linguistic proficiency that can clearly facilitate the transfer of various cognitive and academic language skills across languages. A high level of proficiency in the first language may thus enhance the development of the second language. However, if skills in the first language are not well developed and education in the early years is in the second language, the further development of the second language may stagnate. Importantly, it is argued that certain aspects of bilingual development may also facilitate the more general cognitive and metalinguistic functioning of children (Reference BialystokBialystok, 2001; Reference VerhoevenVerhoeven, 2007). However, the exact conditions under which language transfer typically occurs and just how particular patterns of bilingual development relate to children’s levels of literacy in L1 and L2 are as yet unclear (Reference Durgunoglu and GoldenbergDurgunoglu & Goldenberg, 2011; Reference Verhoeven, Kamil, Pearson, Moje and AfflerbachVerhoeven, 2010, Reference Verhoeven, Cain, Compton and Parilla2017).

The sociocultural background of children in the Caribbean region may also have a large impact on their language and literacy learning. A first relevant aspect is language exposure. Language input in and outside the home environment may have an impact on children’s L1 and L2 development (Reference Lesaux, Geva, August and ShanahanLesaux & Geva, 2006; Nag et al., 2018). Likewise, the home literacy environment may have a differential effect on children’s (bi)literacy development (Reference Dickinson and NeumanDickinson & Neuman, 2006; Dickinson et al., Chapter 17 in this volume,). Another relevant characteristic is the sociocultural orientation of the child and their parents (see Nag, Chapter 15 in this volume). Parental attitudes toward L1 maintenance and L2 learning may have an impact on the child’s attitudes and motivation (Reference AgerAger, 2001; Reference PereiraPereira, 2018). Apart from family-related characteristics, school-related sociolinguistic factors may also have an impact on children’s language and literacy learning in L1 and L2, such as the teacher’s educational level, teaching experience, language proficiency, and sociocultural orientation (Reference Lemberger and Reyes-CarrasquilloLemberger & Reyes-Carrasquillo, 2011; Nag et al., 2016; Dickinson et al., Chapter 17 in this volume).

Finally, initiatives must be taken to develop and implement (bi)literacy programs in schools. For such programs, two sets of curriculum materials should be made available, besides literary books and reference works in the two languages. For the Creole language, it is extremely important that the lexicon is continuously elaborated with new entries in order to enable the development of curricula for basic, secondary, and higher education. For the colonial language, it is critical that curriculum materials are an ethnically appropriate mirror of the local realities. Another important issue is the availability of qualified teachers. Teachers must be well trained in literacies in the language(s) of instruction in order to be able to offer the planned literacy programs. With reference to the teaching of Creole besides French in Guadeloupe, Reference Bolus, Migge, Léglise and BartensBolus (2010) noted that there is a high load on teachers given the fact that the availability of curricular materials in the Creole languages is limited and they are in short supply, so that teachers have to develop their own curricular materials. Therefore, there is an urgent need for teacher training to ground teachers’ knowledge on biliteracy development and apply this knowledge so as to make their literacy teaching more effective.

4.2.3 Literacy Levels in Caribbean Societies

UNESCO (2013) provided global literacy rates of the adult population (fifteen years and older) in Caribbean countries. Overall, the literacy rate in 2011 was estimated to be 92 percent as compared to 90 percent in 2000 and 86 percent in 1990. A breakdown by gender showed that 55 percent of illiterate people were women. Over the years, the literacy rate has grown faster among females, narrowing the gap with men. There is significant variation between the different territories. In Antigua and Barbuda, the Cayman Islands, Trinidad and Tobago, Aruba, and Costa Rica, the illiteracy rate is less than 5 percent; in Puerto Rico and the Dominican Republic, less than 10 percent; in Jamaica, Honduras, and Guyana less than 15 percent; and in Haiti about 50 percent. Trinidad and Tobago is one of the few countries in the region that has participated in the International Reading Literacy Study (PIRLS). Over the years 2006, 2011, and 2016, PIRLS results show a steady improvement in the global literacy benchmark, which may be due to the successful implementation of nationwide programs that were designed to increase literacy levels among students (Reference CharlesCharles, 2013; Reference George and Quamina-AiyejinaGeorge & Quamina-Aiyejina, 2003).

UNESCO data also show that throughout Caribbean territories there is large variation in school trajectories and in characteristics related to the child, the home environment, and schools in the region. With respect to school trajectories, literacy is often taught in the official (ex-colonial) language only. In some cases, literacy instruction in the Creole language is also being implemented (see Migge et al., 2010). In quite a few countries (e.g., Guadeloupe, Haiti, Jamaica), literacy is being taught in the Creole language from scratch. In a quasi-experimental study, Reference Carpenter, Devonish, Migge, Léglise and BartensCarpenter and Devonish (2010) showed that the implementation of a bilingual literacy program in the early primary grades with Jamaican Creole and English as languages of instruction yielded high literacy outcomes in Jamaican Creole with no cost for the development of English literacy (Phillipson, 1990). In another quasi-experimental study, Reference Simmons-McDonald, Migge, Léglise and BartensSimmons-McDonald (2010) found that the introduction of French Creole as the language of instruction for literacy in primary schools in St. Lucia helped poor learners to develop literacy not only in French Creole but also in English as a second language. It is important to note that the pilot studies on bilingual literacy instruction that have been conducted so far have proved to be successful in developing literacy even in those children that show great variation in entrance levels of cognitive and linguistic abilities and cultural orientation, as well as in terms of the socioeconomic background of the family, the literacy practices in the home, and in the school support parents give to the child. Of course, it should be emphasized that experiments with bilingual literacy education can only be successful where there is full institutional support. Reference Migge, Leglise and BartensMigge et al. (2010b) evaluated three projects on the implementation of local languages in primary schools in French Guiana. They concluded that only when long-term support from the authorities is provided, including the development and provision of curricula in the local languages and teacher training in bilingual literacy programs, can sustained levels of biliteracy be attained.

4.3.1 Historical Context

The Dutch Caribbean comprises a group of islands, also named the Dutch Antilles, which were colonized by the Dutch in 1634. It is made up of the so-called Leeward Islands: Aruba, Bonaire, and Curaçao, with the Portuguese-based Creole Papiamentu as the generally used vernacular in a population of about 300,000; and the Windward Islands: Sint Maarten, Saba and St. Eustatius, with English as the common language of communication in a population of about 43,000. In this section, we focus on the development of literacy on the Leeward Islands, where Papiamentu is the language of children’s primary socialization process (cf. Reference Severing, Weijer, Faraclas, Severing and WeijerSevering & Weijer, 2008). It is interesting to note that until the beginning of the twentieth century, education was offered by missionary posts, with Creole Papiamentu as the language of instruction. However, in 1936 Papiamentu was banned from schools after a renewed economic interest in the islands on the part of the Dutch authorities in the Netherlands, despite the fact that Papiamentu was spoken by the vast majority of the population. Since then, the use of Dutch in schools was prescribed, whereas Papiamentu was considered to be an obstacle to learning. Formal education could best be characterized as L2 submersion, with children who spoke Papiamentu as their main language being engaged in Dutch instruction. Not until 1986 was Papiamentu assigned a place in the primary school curriculum on Curaçao and Bonaire. Thanks to great efforts in language codification and curriculum development, the role of Papiamentu within schools greatly expanded, while the majority of the instruction nevertheless took place in Dutch. From 2008, Dutch, Papiamentu, and English were recognized as the three official languages on the islands.

According to estimates, Papiamentu is the mother tongue of almost 80 percent of the population on the Leeward Islands. Most children first come into contact with Dutch at school. As Dutch is not spoken or hardly spoken in the child’s environment, it can be considered a foreign language. The language situation can best be characterized with the term “diglossia” (cf. Reference Appel, Verhoeven, Arends, Muysken and SmithAppel & Verhoeven, 1994; Reference ManaganManagan, 2016). On the one hand, there is the Creole Papiamentu, which was traditionally held in relatively low esteem but functions as the general lingua franca. On the other hand, there is Dutch, which has a relatively high social status and is generally learned as a foreign language in school. However, as a result of increasing political independence, Papiamentu’s status has increased dramatically in recent years, with the language being highly valued and increasingly used in formal situations. For example, most newspapers are published in Papiamentu and parliamentary debates also take place in that language (cf. Reference Severing, Weijer, Faraclas, Severing and WeijerSevering & Weijer, 2008). Moreover, serious attempts have been made to make Papiamentu suitable as a language for instruction by elaborating its lexicon with new entries, developing school curricula for language, literacy, and content matters, and by grounding teacher training in a multilingual perspective (Reference Severing, Weijer, Faraclas, Severing, Weijer and EchteldSevering & Weijer, 2010). In a recent study, Reference PereiraPereira (2018) investigated how 108 teachers in Aruba valued the role of Papiamentu in education. The great majority of teachers promoted educational reform that gives an important place to Papiamentu in the curriculum. They also pointed to the necessity of specific in-service training of teachers to prepare them for multilingual education. Twelve teachers with experience in Scol Multilingual – an innovative bilingual-education experiment on the island – were also interviewed; they indicated improved learning attitudes and better academic outcomes on the part of their pupils. In another survey, Pereira interviewed 1,141 parents of children at primary education level. By far, most parents were extremely supportive of educational reform from monolingual Dutch toward bilingual Papiamentu–Dutch education.

4.3.2 Early Language and Literacy Development

In a recent survey, the exposure to Papiamentu and Dutch of 183 children aged nought to four years on the island of Bonaire was examined (Reference Odenthal and BouwmanOdenthal & Bouwman, 2016). The children’s parents were interviewed as regarded their language of primary socialization, frequency of storybook reading, preferred television channels, and use of preschool facilities. It was found that the great majority of native parents from the Dutch Caribbean used Papiamentu as the primary home language, with only limited exposure to Dutch. Among a small minority of parents originating from other South American territories and from the Netherlands, the main languages in the home were reported to be Spanish and Dutch respectively, with limited contact with Papiamentu. With respect to storybook reading, parents reported reading to their child more often in Papiamentu than in Dutch. However, in one third of the cases, no storybook reading was reported by parents. Preschool facilities were reported to be widely used. In most cases the caretakers spoke Papiamentu or Papiamentu and Dutch, and the use of Dutch only was reported to be rare.

A quasi-experimental study examined the preliteracy development in Papiamentu and Dutch of eighty children aged four to six and living on the island of Curaçao (Reference NarainNarain, 1995; Reference Narain and VerhoevenNarain & Verhoeven, 1994). Children’s abilities in receptive and productive vocabulary, sentence reproduction, and narrative comprehension were assessed in Papiamentu and Dutch, and their phonological awareness in Dutch was also assessed. Children’s performance on these tests were measured at three moments: at the beginning of kindergarten, and after one and two years of education in kindergarten. At the end of kindergarten, the children’s phonological awareness was also measured. In order to explore the relationship between language proficiency and background features, sociocultural background characteristics (preschool duration, sociocultural orientation of the child and the parent, linguistic exposure, language used by the teacher) were also measured.

The results on the development of language abilities over time are presented in Figure 4.1. In all four linguistic skills examined in Papiamentu and Dutch, there was significant progress over time. Comparison of the two languages shows that participating children on Curaçao are significantly stronger in Papiamentu than in Dutch throughout the entire kindergarten school period and at all linguistic levels. Figure 4.1 also shows that the scores on productive vocabulary, sentence reproduction, and narrative comprehension were relatively low throughout the kindergarten period.

Figure 4.1 Means for the tests for receptive vocabulary, productive vocabulary, sentence reproduction, and narrative comprehension in Papiamentu and Dutch at the beginning (K1) and after one year (K2) and two years (K3) of kindergarten

It is important to note that evidence of cross-language transfer was also found. Over the two years of kindergarten, the longitudinal relations between Papiamentu-language abilities on the one hand and Dutch-language abilities on the other turned out to be extremely strong. Interestingly, there was also evidence of significant relations from L1 Papiamentu to L2 Dutch over time, as assumed in the framework of linguistic transfer (e.g., Reference Chen, Dronjic and Helms-ParkChen et al., 2015).

Two other important findings came out of the study. One is that sociocultural variables substantially predicted the children’s first- and second-language proficiency. Duration of preschool and the child’s cultural orientation combine to determine almost 40 percent of the variance in Papiamentu-language scores whereas the child’s linguistic exposure and parents’ cultural orientation explained 80 percent of the variance in Dutch-language scores. At the end of kindergarten, the language used by the teacher explained about 25 percent of the variance in Papiamentu-language scores. Simultaneously, language exposure at school and at home and the child’s cultural orientation together determine 77 percent of the variance in Dutch-language scores.

The other finding was that the degree of bilingual competence in children predicted their phonological awareness, which can be seen as an important precursor of their literacy development. The scores on phonological-awareness tasks were highest for children who scored above average in both languages, intermediate for children who scored above average in one language, and lowest for children who scored below average in the two languages. Apparently, a high level of bilingualism leads to good insight into the structure of language as such. This is in line with earlier research and theorizing (cf. Reference BialystokBialystok, 2001; Reference Cummins and RiveraCummins, 1984).

4.3.3 Bilingualism and Learning to Read

With respect to the development of basic literacy skills, Reference Prins-Winkel and MullerPrins-Winkel (1983) stated that Dutch-only language education on the island of Curaçao primarily leads to a high level of functional illiteracy. Children often prove to be able to perform reasonably well in the subject of Dutch as a foreign language, but have great difficulty using the language in everyday situations which require its written use. Recent evidence for this claim comes from a study by Reference Odenthal and BouwmanOdenthal and Bouwman (2016), who studied the language and literacy levels of 940 children in Bonaire in the age range of nine to eighteen. In a questionnaire, children reported that Papiamentu was spoken at home in almost 90 percent of families, in half of the cases in combination with Dutch. For half of the children, Papiamentu was the dominant language in school, while for 40 percent the dominant school language is Dutch. Outside school, children reported reading more books in Papiamentu than in Dutch. In Grades 2–5, children’s word-decoding fluency (the number of isolated words read in one minute) and reading comprehension were assessed. It was found that children on Bonaire developed L2 word-decoding skills across the grades. However, they stayed behind their monolingual Dutch peers in the Netherlands. The children’s reading-comprehension levels across the grades were extremely low compared to those of their Dutch peers. Finally, the literacy levels in Papiamentu and Dutch of 1,067 students at the end of secondary school were examined. The students from Bonaire remained behind in Dutch literacy compared with their peers in the Netherlands, whereas their level in Papiamentu literacy was higher than their Dutch. However, it should be noted that both in the Netherlands and in Bonaire there are three different educational tracks in the secondary school system (lower-vocational, higher-vocational, and pre-university level), with fewer students from Bonaire in the two higher tracks.

In several other studies attempts were made to examine the literacy development of children in both Papiamentu and Dutch. Reference SeveringSevering (1997) and Reference Severing and VerhoevenSevering and Verhoeven (1995) examined the development of word decoding, reading vocabulary, sentence comprehension, and text comprehension in written Papiamentu and Dutch from Grades 5 and 6 in primary school, with literacy being taught in Dutch and Papiamentu being taught as a subject for only half an hour a week. In Figure 4.2, the means of children’s word decoding (number of words read aloud correctly in one minute), reading vocabulary, sentence comprehension, and text comprehension at the beginning and end of Grade 5 and at the end of Grade 6 are presented.

Figure 4.2 Means for the tests for word decoding, reading vocabulary, sentence comprehension, and text comprehension in Papiamentu and Dutch at the beginning (B5) and the ends of Grades 5 (E5) and 6 (E6)

It was found that primary-school children on Curaçao learned to decode words significantly better in Dutch than in Papiamentu, but that their performance on reading vocabulary, sentence comprehension, and text comprehension was significantly better in Papiamentu than in Dutch. For word decoding, children were overall better in Dutch compared to Papiamentu, bearing in mind that reading instruction took place in Dutch. But the differences became smaller over time. For reading vocabulary, sentence comprehension, and reading comprehension, significant progress over the grades was also evidenced. The children scored higher overall in Papiamentu as compared to Dutch, and even more so for reading vocabulary. Next, the researchers examined to what extent the children’s reading abilities in Papiamentu and Dutch were related. In both Papiamentu and Dutch, word decoding, reading vocabulary, and reading comprehension abilities were not only longitudinally related but also cross-linguistically related to each other. Interestingly, the development of Papiamentu word-decoding ability could to a large extent be predicted from the level of Dutch word-decoding ability. This finding shows that the language of instruction has a high impact on the children’s literacy development and that development of Papiamentu literacy skills takes place even without much formal instruction in school. However, it is interesting to note that for reading comprehension it was the other way around, with transfer going from Papiamentu (the strongest language) to Dutch (the weaker language) despite the lack of instruction in Papiamentu.

The study by Severing and Verhoeven also explored to what extent the variation in Papiamentu and Dutch reading abilities is associated with factors which can be related to the child or their family background. Papiamentu word decoding at the three testing times was negatively associated with grade repetition and positively with the child’s reading behavior, whereas Papiamentu reading comprehension was negatively associated with grade repetition and positively with cognitive skills and socioeconomic status. For Dutch, word decoding and reading comprehension across the grades were negatively associated with class repeats and positively with cognitive skills, attitude toward Dutch, exposure to Dutch, and socioeconomic status. Dutch word decoding was also related to the child’s reading behavior. These findings show that both child factors related to the child – such as cognitive skills, class repeating, and reading behavior – and background factors related to language exposure, language attitudes, and socioeconomic status contribute to the prediction of literacy in Papiamentu and Dutch.

In three recent studies, the variation of Papiamentu and Dutch reading in the Dutch Caribbean was further explored. Reference HenderickxHenderickx (2018) examined the effect of the order of literacy instruction on word-decoding and reading-comprehension abilities in Papiamentu and Dutch in 166 children from Grade 3 and 140 children from Grade 5. For half of the children in each cohort, the order of reading instruction was Papiamentu–Dutch (i.e., Papiamentu starting in Grade 1, Dutch starting by the end of Grade 2), and for the other half it was Dutch–Papiamentu. Overall, the word-decoding scores in the two grades were about the same in Papiamentu and Dutch. This may be because there is a great deal of overlap in the letters and literacy conventions used in the two languages. Importantly, the reading-comprehension scores in the two grades were found to be higher in Papiamentu. The instruction order Dutch–Papiamentu yielded higher scores on Dutch reading comprehension, with no impact on Papiamentu reading comprehension, in both grades. Apparently, a high reading proficiency in Dutch as a foreign language is largely dependent on the amount of reading instruction and reading exposure in that language.

Reference Mercelina, Segers, Severing and VerhoevenMercelina et al. (2023) examined the variation in early decoding development in Papiamentu (Creole L1) and Dutch (postcolonial L2) from kindergarten to second grade of 156 children while considering order of initial decoding instruction (L1 or L2), kindergarten precursors, and transfer effects. Results revealed that in the case of L1 decoding instruction, the development of letter knowledge and decoding was higher in Papiamentu and in the case of L2 decoding instruction it was higher in Dutch. When instruction started in L1, Papiamentu decoding development was statistically significantly predicted by phonological awareness and letter knowledge, and Dutch decoding development by letter knowledge only. When instruction started in L2, Dutch decoding development was statistically significantly predicted by phonological awareness, letter knowledge, short-term memory, and speech decoding, and Papiamentu decoding development by letter knowledge and vocabulary. Transfer effects were found from the language of decoding instruction to the other language with L1 precursors significantly predicting L2, but not the other way around. These results reveal the pivotal role and benefit of including the L1 in early bilingual decoding development in a postcolonial context.

Finally, Reference Van der Elst-Koeiman, Segers, Severing and VerhoevenVan der Elst-Koeiman et al. (2022) examined the Papiamentu (L1) and Dutch (L2) reading comprehension development of 293 fourth-grade students in postcolonial Dutch Caribbean schools. It was found that reading comprehension in Papiamentu and Dutch was predicted by initial language of alphabetization, on the one hand, and word decoding, receptive vocabulary, and grammatical ability on the other hand. There was also evidence of linguistic interdependencies for decoding and reading comprehension between languages. The results highlight the importance of language of instruction for Papiamentu and Dutch reading skills.

4.3.4 Literacy and School Success

A number of studies have examined the relationship between language skills and school success. A classic study is that of Reference Prins-WinkelPrins-Winkel (1973). In the context of monolingual Dutch-language education, she related children’s academic performance to their intelligence, their socioeconomic class, and their home language. General cognitive abilities and socioeconomic class proved to have a positive correlation with academic performance, a finding corroborated by other research in this field. Furthermore, the home language proved to be a predictor: The more Dutch was spoken at home, the better the academic performance at the Dutch-only school. On the other hand, Reference Kook and VedderKook and Vedder (1989) found that Papiamentu-speaking children at middle-class schools scored as well in spelling and vocabulary tests as the Dutch-language children at those schools. They therefore concluded that the child’s home language does not by itself have a decisive influence on academic performance. This is in line with findings in other parts of the world (cf. Nag et al., 2018). A few studies have revealed that the exclusive attention to Dutch in education did not lead to positive educational results. First, periodically gathered statistics proved how low the educational return was during the time of Dutch monolingual education. As a result of the high level of discontinuity between the language offered to children at home and at school, over 60 percent of the children were found to have repeated one or more years in primary education (Minister of Education and Culture, 1992). Reference OlthetenOltheten (1980) examined the determinants of nonpromotion in primary education. Sociocultural characteristics and the family’s socioeconomic position, regional differences, and the efficacy of the school proved to primarily predict children’s nonpromotion. Finally, Reference SeveringSevering (1997) examined the relation between word decoding and reading comprehension abilities in Papiamentu and Dutch in fifth and sixth grade on the one hand and educational success by the end of primary school on the other. Notwithstanding the fact that formal literacy instruction throughout the grades had been given to a large extent in Dutch and that educational success was measured in Dutch only, it was found that not only children’s literacy attainment in Dutch but also their literacy development in Papiamentu significantly predicted their success in secondary education.

5.2.1 Historical Perspective

Educational attainment grew enormously from 1960 to 1995 in most of the Western world, with the average number of years of schooling nearly doubling (Reference Heckman and JacobsHeckman & Jacobs, 2009). In 2015, in the European Union,Footnote ² 80 percent of youngsters under twenty-five completed upper-secondary education or the equivalent of twelve years of schooling (OECD, 2017). This average is for twenty-two EU countries; the available data still includes the United Kingdom but excludes Bulgaria, Croatia, Cyprus, Lithuania, Malta, and Romania. It is possible that if data were available for the current twenty-seven EU Member States this average would be slightly different. For the EU22, it reflects wide variation among countries in terms of progress over time and differences in recent education attainment rates. For example, while in Portugal the secondary-school completion rate grew from 51 percent in 2005 to 83 percent in 2012, in Hungary it increased only two percentage points, from 80 to 82 percent, during the same time interval. In 2015, in Spain 68 percent of young adults completed secondary school compared to 87 percent in Finland.

The expansion of schooling is a major factor influencing the spread of literacy worldwide and three major periods have been identified in the diffusion of literacy (UNESCO, 2005). Before 1800, reading was already widespread in northern Europe in countries like Denmark, Finland, England, and France, but was confined to the upper classes. In Germanic-language countries and German-speaking regions, more than 50 percent of males were already literate before 1790 (Reference Diebolt and HippeDiebolt & Hippe, 2017). In contrast, in southern European countries like Spain, Portugal, and Italy, illiteracy was the norm. Between 1800 and 1860, little progress in literacy rates was registered in these countries, while in northern Europe modest progress was achieved. Several European countries introduced compulsory education in the first half of the nineteenth century, but actual enrollment in primary schools was very low (Reference Hippe and FouquetHippe & Fouquet, 2015). After 1860, literacy levels improved in much of Europe because of the introduction of legislation regulating compulsory education, although literacy levels in Hungary, Italy, Spain, and the Balkan countries remained low even around 1900. Following low literacy levels in most of nineteenth-century Europe, schooling became more available to all but social elites after the mid-nineteenth century. This change was rooted in the need for record keeping and for skillful labor linked to the use of new technologies, such as the steam engine, that required workers who could maintain and fix machinery. However, the spread of literacy remained a challenge (Reference Hippe and FouquetHippe & Fouquet, 2015) and the situation in the following century was as follows:

Different factors may account for the cross-country and within-country regional differences in adult literacy rates from the 1800s to the twenty-first century. These include literacy traditions (Reference Elley, Verhoeven and SnowElley, 2001), geographical location, and linguistic background. For example, during the 1800s, geographical proximity appears to have played a role in the spread of literacy from Germany, Switzerland, and the Netherlands to neighboring regions in Belgium, France, and northern Italy (Reference Diebolt and HippeDiebolt & Hippe, 2017). Furthermore, regional discrepancies in literacy rates within a country were often associated with linguistic variations. This may have been the case in Brittany, where French was not widely spoken before the nineteenth century, and which registered much lower literacy rates than other French regions in the 1800s (Reference Diebolt and HippeDiebolt & Hippe, 2017).

The beginning of the twentieth century clearly marked the widespread attendance of all children at least at primary school in most of the industrialized world (Reference Weber, Verhoeven and SnowWeber, 2001), although in some countries the literacy rate of all citizens did not significantly increase until the second half of the century (UNESCO, 2005). During this latter period, the spread of schooling in most of Europe was clearly linked to the attainment of more sophisticated levels of literate ability and to greater economic returns. Put differently, while in earlier periods an individual could be considered literate if he or she could only sign his or her name (Reference Venezky, Barr, Kamil, Mosenthal and PearsonVenezky, 1991) schooling in the second half of the twentieth century made it possible for individuals to further develop reading and writing abilities, and this rise in education levels translated into better jobs and higher salaries (Reference Heckman and JacobsHeckman & Jacobs, 2009). The accumulation of knowledge through schooling, often referred to as human capital, paid off in economic terms. However, at the turn of the century and into the twenty-first century the picture is more complex, because studies show that more school attainment does not translate directly and unequivocally into higher literate abilities.

Research conducted in the last decades has shown that quality of education is a better indicator of life outcomes than quantity of education, as measured in years of schooling (Reference Heckman and JacobsHeckman & Jacobs, 2009). This suggests that “School attainment is not a very good proxy for knowledge” (p. 27) and that, as Reference Hanushek and WoessmannHanushek and Woessmann (2015) contend, “direct measures of cognitive skills offer a superior approach to understanding how human capital affects the economic fortunes of nations” (p. 28). In relation to literacy attainment, children who fail to develop the basic reading skills needed to comprehend what they read by the fourth grade are likely to face reduced educational opportunities (Reference Adams and HiebertAdams, 2009; Reference Chall and JacobsChall & Jacobs, 2003). This, in turn, will affect their chances of acquiring the skills that are essential for full participation in society and in the labor force (Reference Hanushek and WoessmannHanushek & Woessmann, 2015).

As we have discussed, schooling and associated literacy rates in Europe vary from country to country and this variation has been present since the 1800s. While the relation between years of schooling and literacy abilities is not well documented prior to the twentieth century, in the 1800s someone with the ability to sign his or her name would have been considered a literate individual (UNESCO, 2005). If the ability to sign one’s name was already considered a literate behavior in the 1800s, and a useful one in terms of signing a work commitment or property possession (Reference Venezky, Barr, Kamil, Mosenthal and PearsonVenezky, 1991), what reading-literacy skills characterize the European student population of the twenty-first century, and how do children develop them?

5.2.2 Reading Development in Different Orthographies

Several cross-linguistic and/or cross-national studies have looked at how children develop reading skills in different orthographies (Reference Seymour, Aro and ErskineSeymour, Aro, & Erskine, 2003; Reference Vaessen, Bertrand and TothVaessen et al., 2010; Reference Ziegler, Bertrand and TóthZiegler et al., 2010). In the case of alphabetic writing systems, orthographic depth explains different rates of reading development and children seem to acquire reading faster in transparent orthographies (Reference Pollatsek and TreimanPollatsek & Treiman, 2015). The facilitating effect transparent orthographies have in reading development has mostly been established in the first stages of learning to read, first and second grades, and confirmed by studies that ask children to read words and nonwords (Reference SeidenbergSeidenberg, 2013). In a comparative study of reading acquisition in twelve different languages, Reference Seymour, Aro and ErskineSeymour et al. (2003) found that English and Danish first-graders take longer than Finish, Spanish, or Italian ones to master word-decoding skills.

In a transparent or shallow orthography, such as Finnish, there is a one-to-one phoneme-to-grapheme mapping whereby one letter represents one sound and vice versa. Research indicates that there is a continuum, with English being one of the most opaque alphabetic languages – different pronunciations for the same spelling patterns – while French and Danish are positioned somewhere in the middle of the orthographic-depth continuum (Reference Schmalz, Marinus, Coltheart and CastlesSchmalz et al., 2015). For example, in French, spelling-to-sound relations are reasonably predictable, but sound-to-spelling relations are more ambiguous (Reference Schmalz, Marinus, Coltheart and CastlesSchmalz et al., 2015). Among Western European languages, English has the most inconsistent orthography, and this impacts the rate at which English-speaking children develop literacy (Reference Sproat, Cook and RyanSproat, 2016). A child learning to read in English is confronted with the task of learning inconsistent print-to-speech correspondences, such as “ea” having one sound in “bread” and another in “leak” (Reference Schmalz, Marinus, Coltheart and CastlesSchmalz et al., 2015). To a lesser extent, a Portuguese first-grader is also confronted with a phonological inconsistent orthography; for example, “ca” can correspond to /ka/ or /kɐ/ (Reference Ventura, Fernandes, Leite, Pereira and WongVentura et al., 2019). In contrast, a child learning to read in Spanish does not encounter many words that have graphemes that share the same spelling but can be pronounced differently in different words. This facilitates learning to an extent that a child’s ability to name letters and segment speech sounds in kindergarten are not such strong determiners of that child’s reading ability in second grade and beyond (Reference Caravolas, Lervåg and MousikouCaravolas et al., 2012).

Other studies specifically addressing the impact of predictors of reading ability, for example the ability to segment speech sounds or phonological awareness (PA), indicate that reading development is modulated by orthography transparency. PA is a stronger predictor in less transparent orthographies, such as French and Portuguese, than in more transparent ones like Dutch, Hungarian, and Finish (Reference Ziegler, Bertrand and TóthZiegler et al., 2010). Although researchers have questioned the relevance of PA as a reading predictor in different languages because most studies have been conducted in English (Reference ShareShare, 2008), current evidence suggests that PA is a predictor of reading ability across European alphabetic languages (Reference Vaessen, Bertrand and TothVaessen et al., 2010). Reference Perfetti, Verhoeven, Verhoeven and PerfettiPerfetti and Verhoeven (2017) concluded that this association of PA and reading development holds for fourteen European languages. Some studies indicate that accuracy in word recognition is a better predictor of reading ability in less consistent orthographies, whereas variations in speed in more consistent orthographies explain reading performance (Reference Landerl, Ramus and MollLanderl et al., 2013; Reference Pollatsek and TreimanPollatsek & Treiman, 2015; Reference Vaessen, Bertrand and TothVaessen et al., 2010). However, the cognitive processes involved in word decoding – accuracy and speed – are identical in orthographies that vary along the transparency continuum (Reference Hulme and SnowlingHulme & Snowling, 2013; Reference Vaessen, Bertrand and TothVaessen et al., 2010), and this has been observed also in Grades 3 to 6 of primary school (Reference Moll, Ramus and BartlingMoll et al., 2014).

As children move beyond the initial stages of learning to read, knowledge of other aspects of language beyond accurate and fluent word reading, namely grammar and vocabulary knowledge, are associated with reading comprehension (Reference Hulme and SnowlingHulme & Snowling, 2013; Reference SeidenbergSeidenberg, 2013; Reference Sénéchal and WasikSénéchal, 2012). During the learning-to-read phase (Reference ChallChall, 1996) fast word-identification skills serve as the foundation for text comprehension (Reference Perfetti, Gough, Ehri and TriemanPerfetti, 1992), but as the ability to decode words develops, other factors, such as vocabulary knowledge, support reading to learn (Reference Chall and JacobsChall & Jacobs, 2003). In fact, “research indicates that reading with comprehension depends on understanding at least 95% of the words of the text” (Reference Adams and HiebertAdams, 2009, p. 172). This understanding of the meaning of words develops, for example, when preschool-age children are exposed to book reading and parents explore the meaning of print with them. Later, the oral comprehension they developed early in life will assist them in comprehending what they will read by themselves (Reference Sénéchal and WasikSénéchal et al., 2012). For example, Reference Sénéchal, Ouellette, Rodney, Neuman and DickinsonSénéchal, Ouellette, and Rodney (2006) found that preschool Canadian children’s vocabulary knowledge, acquired from parental book reading, predicts reading comprehension in Grade 3. The influence of this home-literacy practice of shared reading on future reading achievement has been found in studies with English- and French-speaking children (Reference Sénéchal, Ouellette, Rodney, Neuman and DickinsonSénéchal, 2006), as well as in studies in other more transparent languages. For instance, in Greek (Reference Manolitsis, Georgiou and ParrilaManolitsis, Georgiou, & Parrila, 2011; Lithuanian (Reference Silinskas, Aunola, Lerkkanen and RaizieneSilinskas et al., 2021); German (Reference Lehrl, Ebert, Roßbach, Pfost, Artelt and WeinertLehrl, Ebert, & Roßbach, 2013; Reference Niklas and SchneiderNiklas & Schneider, 2017; Reference Rose, Lehrl, Ebert and WeinertRose et al., 2018), and Finnish (Reference Silinskas, Lerkkanen and TolvanenSilinskas et al., 2012; Reference Silinskas, Torppa, Lerkkanen and NurmiSilinskas et al., 2020). As Reference Kalb and OursKalb and Van Ours (2014) put it, reading to young children gives them a head start in life. In their study of the impact of book reading by Australian parents they found that children who are read to in the home frequently, three-to-five days a week and six-to-seven days a week, obtain the equivalent of between six-to-twelve months’ higher scores in literacy at eight/nine years of age. Smaller effects of reading to children were found for numeracy skills.

Sénéchal’s home-literacy model (Reference Sénéchal and Wasik2012) postulates that during the preschool years both meaning-based interactions during shared reading and code-based teaching by parents, like naming alphabet letters, contribute to later reading achievement, and this view is consistent with theories of reading development (Reference Mol and BusMol & Bus, 2011; Reference Perfetti, Landi, Oakhill, Snowling and HulmePerfetti, Landi, & Oakhill, 2005). More specifically, shared reading supports vocabulary knowledge (Reference Sénéchal and WasikSénéchal, 2012) and code-based teaching by parents during shared reading can also assist children in learning to read. Nonetheless, “achievement-based skills such as early reading, early math, and letter recognition skills appear to be more sensitive to Head Start intervention attendance than cognitive skills such as IQ, vocabulary, and attention which are less sensitive to classroom instruction” (Reference Shager, Schindler and MagnusonShager et al., 2013, p. 90).

In summary, country variations in reading development are associated with linguistic differences related to orthographic depth in the early years of reading instruction and individual variations in reading achievement are linked to different levels of language knowledge and to the specific contexts, at home and/or at school, where this knowledge is acquired.

5.2.3 Variation in Reading Achievement

Home and school socioeconomic and linguistic contexts are relevant in explaining differences in reading achievement that are already present during the primary-school years. In the United States, for example, the reading gap between students from high- and low-income families and between students of different ethnic and linguistic backgrounds has been well documented (Reference SeidenbergSeidenberg, 2013; Kieffer & Vuković, Chapter 2 in this volume). Similarly, in Europe data from comparative education surveys show that there is a relationship between socioeconomic and linguistic backgrounds and achievement.

Large-scale studies, such as the Progress in International Reading Literacy Study (PIRLS), offer unique information about fourth-graders’ comprehension of authentic texts – the goal of learning to read – and related predictors of reading ability. The contextual information collected in different participant countries via student, home, school, and teacher questionnaires in PIRLS makes it possible to look at reading achievement from an international comparative perspective. This information can be used to study both student-background factors and environmental ones and to explore variations across countries in terms of reading-related literacy universals and particulars (Reference Lenkeit, Chan, Hopfenbeck and BairdLenkeit, et al., 2015). In this sense, PIRLS allows us to characterize the reading-literacy skills of fourth-grade students in European countries and to understand their relationship with different sociolinguistic contexts.

Measures of socioeconomic status (SES) can include income, number of books and children’s books at home, parental education level or occupation level, or a combination of these and other variables indicative of available home resources such as possession of an internet connection and one’s own room (Reference Araújo and CostaAraújo & Costa, 2015; Reference Caro, Sandoval-Hernández and LüdtkeCaro, Sandoval-Hernández & Lüdtke, 2014). Features of supportive literacy environments, such as having books at home, have been conceptualized as a measure of cultural capital and have been found to explain reading-score differences among students (e.g., Reference ParkPark, 2008). Moreover, both a school’s intake of children from a certain range of socioeconomic status levels and the school’s average of students’ cultural capital account for part of the reading-score differences between schools (Reference Caro, Sandoval-Hernández and LüdtkeCaro et al., 2014; Reference Myrberg and RosénMyrberg & Rosén, 2006). This indicates that there is a school compositional effect whereby students in schools with a higher SES composition have higher reading achievement and this trend has been found in Germany, France, and Denmark (Reference Stancel-Piatak, Mirazchiyski and DesaStancel-Piatak, Mirazchiyski, & Desa, 2013). Similarly, Reference Myrberg and RosénMyrberg and Rosén (2006) found that in PIRLS 2001, students’ cultural capital accounted for a great part of the reading-score differences between Swedish independent and public schools. In Sweden the share of students with a migrant background is higher in public schools, and this explains their lower reading achievement scores when compared to private schools. This has been corroborated in Dutch studies that also use PIRLS data and show that students who speak Dutch as a second language have lower achievement levels (Reference Netten, Voeten, Droop and VerhoevenNetten et al., 2014; Reference Netten, Luyten, Droop and VerhoevenNetten et al., 2016).

Literacy achievement in PIRLS in European countries reveals that common underlying factors such as socioeconomic and linguistic backgrounds are associated with student achievement. Nonetheless, the strength of the association between reading achievement and SES and between achievement and family/student characteristics varies across countries. For instance, a positive correlation of parents’ socioeconomic status and children’s fourth-grade reading achievement level is present in all European countries participating in PIRLS 2006 and 2011 (Reference Araújo and CostaAraújo & Costa, 2015; Reference ParkPark, 2008). However, score-point differences between students from high and from low SES backgrounds are wider in some countries than in others and both home-literacy activities and students’ early literacy skills also seem to mediate results (Reference Costa and AraújoCosta & Araújo, 2017). In a PIRLS study that considered aspects of the home-literacy model proposed by Reference Sénéchal and WasikSénéchal (2012), Reference Myrberg and RosénMyrberg and Rosén (2009) found that parental book reading and storytelling during preschool made a positive contribution to reading achievement, and that book reading was mediated by cultural capital, as measured by the number of books at home. Additionally, this study showed that Swedish students’ early literacy skills had a positive impact on their subsequent reading attainment, without any mediating effect of cultural capital or number of books at home. Other studies with PIRLS data that include several European countries (e.g., Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013; Reference Stancel-Piatak, Mirazchiyski and DesaStancel-Piatak et al., 2013) also support the notion that early literacy skills, such as recognizing letters of the alphabet and being able to write some words, exert a positive influence on achievement. However, Reference Netten, Voeten, Droop and VerhoevenNetten et al. (2014) examined PIRLS data for the Netherlands and found a positive influence of early literacy activities, but not of early literacy skills, on achievement.

In short, students who enter elementary school knowing how to name letters of the alphabet and how to write some words score higher in PIRLS. This has been observed in most European countries, after controlling for SES, except in the Netherlands (Reference Netten, Voeten, Droop and VerhoevenNetten et al., 2014). Nevertheless, the studies reviewed suggest that variations in reading achievement in PIRLS may be related to sociocultural and linguistic factors, as well as to individual literacy abilities and skills. Furthermore, PIRLS studies suggest that individual reading habits and reading for pleasure are related to achievement in Denmark, Sweden, and France (Reference Costa and AraújoCosta & Araújo, 2017), after accounting for SES. Similar findings are reported in a Dutch study that assessed reading ability using the PIRLS variable that measures reading motivation (Reference Netten, Droop and VerhoevenNetten, Droop, & Verhoeven, 2011).

With respect to school factors, analyses of PIRLS data have shown that adequate resources for teaching reading are related with increased reading achievement (Mullis et al., 2017), as is a school’s emphasis on academic success (Reference Costa and AraújoCosta & Araújo, 2017; Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013). Considering instructional factors, PIRLS studies do not show a strong relation with reading scores. For instance, an emphasis on reading skills in first grade is only positively related to achievement in Grade 4 in one European country – Germany (Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013). However, this may be due to a restriction of range in reading curricula (see Perfetti & Verhoeven, Chapter 11 in this volume). Similarly, no relationships have been reported between reading achievement and types of reading instruction – whole group vs. small group or individualized instruction and time spent on reading (Reference Shiel and EiversShiel & Eivers, 2009). Reading literacy achievement in PIRLS 2011 compared with PIRLS 2006 has been associated with a decline in students’ ability to answer questions that require higher-order comprehension processes, such as understanding text structure and main idea (Reference Netten, Voeten, Droop and VerhoevenNetten et al., 2014). However, this finding has not been linked to teachers’ instructional strategies or, more specifically, to a different emphasis in reading processes. It remains difficult to ascertain cause-and-effect relations between the way teachers teach comprehension and reading achievement in PIRLS, but studies suggest that reading instruction influences reading ability (Reference Hulme and SnowlingHulme & Snowling, 2013). For example, comprehension instruction that focuses on practicing making inferences from text enhances reading attainment (Reference Elbro and Buch-IversonElbro & Buch-Iverson, 2013).

Turning next to secondary-school attainments, observed variations in reading achievement level reveal that similar student-background factors are associated with achievement and that schools can also make a difference. All rounds of the Program for International Student Assessment (PISA) show, for example, that the higher the SES the higher students score in reading, and that migrant students’ scores are lower than those of native students. In PISA, the reading achievement gap between immigrant and native students is observed even after controlling for SES (Reference Lenkeit, Caro and StrandLenkeit, Caro, & Strand, 2015). Nonetheless, students’ achievement levels reveal that European immigrant students perform better in reading-test items that mirror educational situations or contexts where reading serves the purpose of learning or acquiring information (Reference Costa and AraújoCosta & Araújo, 2012). Moreover, they perform better than native students in test items linked to occupational reading or reading that involves accomplishing a task, such as looking for a job in a newspaper or following directions in the workplace. Conversely, native students perform better in personal and public situations that imply reading for recreational purposes as well as attending public events (e.g., a concert). Last, at the European level immigrant students perform better in exposition and instruction types of text, which again are text types likely found in textbooks used in school (Costa & Araújo, 2012).

Additional student-background factors related to students’ reading achievement in PISA include parental book reading and students’ reading for enjoyment. Results show that the parent effect in reading is present because students whose parents report reading to them frequently during first grade score higher (OECD, 2012). Furthermore, they show that resilient students, as defined by those who score above what would be expected given their SES background, read frequently outside of school.

The school factors associated with reading achievement in PISA are like those found in PIRLS studies. Besides emphasis on academic success, PISA results highlight that school climate – discipline – and student–teacher relations are associated with higher achievement. Students in schools where there is a good disciplinary climate, where teachers support students and emphasize academic achievement, score higher (OECD, 2012, 2013a, 2016). School compositional effects are also present in PISA. For example, students attending schools with a socioeconomically advantaged intake tend to perform better than those attending schools with more disadvantaged peers (OECD, 2010b).

It is with the PISA surveys that the European Commission monitors progress in the EU goal of reducing the share of low achievers in reading, mathematics, and science. However, PISA 2015 results indicate that, on average, the percentage of low achievers in reading has increased slightly from 17.8 percent in 2012 to 22.5 percent in 2018 in the EU (European Commission, 2020). This indicates that almost one fifth of fifteen-year-olds in the EU do not attain a satisfactory level of reading ability that would enable them to interpret texts effectively, although there are wide variations among countries (e.g., the share of low achievers is 26 percent in Hungary, but only 8 percent in Estonia).

In the following section, we present research evidence on literacy levels in the EU based on PIRLS primary-school data and PISA secondary-school data. The analyses illustrate the association of SES and home-literacy practices with reading scores in both surveys and the relation between early literacy skills and reading achievement in PIRLS.

5.3.1 Literacy Levels in Children

The International Association for the Evaluation of Educational Achievement (IEA) runs PIRLS every five years to measure trends in the reading achievement of fourth-grade students. At this point in their schooling, students have moved from learning to read to reading to learn (Reference ChallChall, 1996), and PIRLS tests students’ ability to comprehend both literary and informational texts. Its assessment framework defines reading literacy as “the ability to understand and use those written language forms required by society and/or valued by the individual. Young readers can construct meaning from a variety of texts. They read to learn, to participate in communities of readers in school and everyday life, and for enjoyment” (Reference Mullis, Kennedy, Martin and SainsburyMullis et al., 2006, p. 103). Reading comprehension is tested in questions that ask students to do the following: (1) focus on and retrieve explicitly stated information; (2) make straightforward inferences; (3) interpret and integrate ideas and information; and (4) evaluate and critique content and textual elements (Reference Mullis, Martin, Foy and HooperMullis et al., 2017).

The first cycle of PIRLS was carried out in 2001 and in 2016 recorded the participation of fifty countries and eleven regional entities. Twenty-four EU Member States (EU-MS)Footnote ³ collected representative data on the reading literacy skills of their fourth-grade students in 2001. The PIRLS achievement scale is based on item response theory (IRT) and scores are scaled to have an international average of 500 and a standard deviation of 100 points.

Figure 5.1 presents the average reading achievement by country and SES of fourth-graders in the twenty-two EU-MS that participated in PIRLS 2016. The graph shows that in 2016, Ireland, Finland, and Poland were the top-performing countries in Europe. Conversely, Malta, French Belgium, and France were the countries with the lowest reading scores in Europe.

Figure 5.1 Average reading achievement by country and by SES in Grade 4 in twenty-two countries participating in the PIRLS 2016 EU-MS dataset

Notes: Countries are ranked in ascending order of the reading-score average. High parental education denotes completion of tertiary education or higher education, and low parental education denotes completion of up to upper-secondary education.

Source: PIRLS, 2016, authors’ calculations

When we look at the reading scores of low SES students, using low parental educational level as a proxy, we see that those students whose parents have up to a high-school education have lower achievement than students whose parents have gone on to complete higher education. Differences in students’ reading achievement vary from about sixteen to seventy-three points, favoring students with a high parental education level, although the associations differ by country. Larger differences in reading achievement are found in Bulgaria and Hungary, and the smallest in Malta, Spain, and Denmark (Figure 5.1).

Turning next to variables prior to school entry, we use items from the PIRLS home questionnaire which asks parents to indicate how well their children were able to name letters of the alphabet before starting school and how often they or someone in the household read to their children. Evidence from studies with PIRLS data indicate that both greater alphabet knowledge and more frequent book reading at home before the start of compulsory education are associated with students’ achievement in Grade 4 (Reference AliverniniAlivernini, 2013; Reference Araújo and CostaAraújo & Costa, 2012, Reference Araújo and Costa2015; Reference Myrberg and RosénMyrberg & Rosén, 2009). Furthermore, good knowledge of the alphabet and high book-reading frequency benefit children from both high- and low-SES backgrounds across countries. These data are consistent with trends reported in several small-scale studies: Children’s ability to name letters of the alphabet before formal reading instruction begins is one of the strongest predictors of children’s reading ability (Reference Bond and DykstraBond & Dykstra, 1967; Reference RileyRiley, 1996, Reference Piasta and WagnerPiasta & Wagner, 2010), because naming letters shares a reciprocal relation with phonological awareness (Reference AdamsAdams, 1990; Reference Verhoeven, van Leeuwe, Irausquin and SegersVerhoeven et al., 2016). Similarly, parental book reading promotes reading development because when preschool-age children are exposed to book reading, they develop an understanding of vocabulary that is not commonly used in daily oral interactions (Reference Kalb and OursKalb & Van Ours, 2014; Reference Sénéchal and WasikSénéchal, 2012).

Figure 5.2 presents the reading-score average according to alphabet knowledge and students’ SES. High alphabet knowledge is positively associated with student scores. For students from low SES, Austria shows the smallest difference (one point) and Lithuania shows the largest, difference between high alphabet knowledge and low alphabet knowledge (eighty-seven points). For students whose parents have tertiary education, Austria is again the country with the lowest difference in student scores (thirteen points) and Ireland is the country with highest difference (seventy-five points).

Figure 5.2 Average reading achievement according to alphabet knowledge and SES in Grade 4 in twenty-two countries participating in the PIRLS 2016 EU-MS dataset

Notes: Grey lines and symbols correspond to low parental education. Black corresponds to high parental education. Dotted lines correspond to low alphabet knowledge and solid to high alphabet knowledge.

Source: PIRLS, 2016, authors’ calculations

Overall, except for Austria, Spain, and Belgium (Flemish speaking), and France, the differences in reading scores related to alphabet knowledge are larger for the low-parental-education group. That is, children whose parents have a low educational level seem to reap more benefits from knowing the letters of the alphabet very well than children whose parents have higher levels of education. However, as Figure 5.2 shows, within the same SES background knowing the letters of the alphabet very well is consistently related to higher student achievement.

Figure 5.3 shows the association between frequency of book reading before the start of compulsory education and the reading achievement of fourth-grade students whose parents have high and low education levels (SES). It clearly shows that within an SES band, high frequency of home book reading is always related to higher student achievement, although there are country-wise variations in the range within each band. For students whose parents have low educational levels, the Slovak Republic shows the largest effect of preschool book reading (160 points); for students whose parents have high educational levels, Sweden has the largest effect (79 points). The graph also reveals that the contribution of home reading to the achievement of students from distinct SES backgrounds is different in different countries. In particular, in a group of ten countries – Bulgaria, the Czech Republic, Finland, France, Hungary, Ireland, Latvia, the Slovak Republic, Slovenia, and Belgium Flemish – the magnitude of the reading score difference between high and low home-reading exposure is larger for students with low SES. In contrast, in Austria, Denmark, Germany, Italy, Lithuania, Malta, the Netherlands, Poland, Portugal, Spain, Sweden, and Belgium Flemish, the magnitude of the reading-score difference between high parental book reading and low parental book reading is larger for students with high SES.

Figure 5.3 Average reading achievement according to parental book reading and SES in Grade 4 in twenty-two countries participating in the PIRLS 2016 EU-MS dataset

Notes: Grey lines and symbols correspond to low parental education. Dark Black corresponds to high parental education. Dotted lines correspond to low book-reading frequency and solid to high book-reading frequency.

Source: PIRLS 2016, authors’ calculations

5.3.2 Literacy Levels in Adolescents

PISA is a cross-sectional survey launched in 2000 by the Organisation for Economic Co-operation and Development (OECD). Its purpose is to assess how ready youngsters are to either enter the workforce or continue further studies, enabling countries to monitor their progress in meeting key learning objectives (OECD, 2013a). The OECD has been running this international large-scale assessment of fifteen-year-old students’ skills in reading, mathematics, and science every three years. Each assessment cycle focuses on a main domain or knowledge area and reading was the main domain in 2000, 2009, and 2018, whereas science was the focus in 2006 and in 2015 and mathematics the focus in 2003 and 2012. In PISA, students’ scores are computed according to IRT and standardized with an OECD mean of 500 and a standard deviation set at 100 in 2000.

In PISA 2018, “reading literacy is understanding, using, evaluating, reflecting on and engaging with texts in order to achieve one’s goals, to develop one’s knowledge and potential and to participate in society” (OECD, 2019, p. 28). In much the same way as PIRLS does, PISA examines the extent to which students are able to understand and integrate the information in informational and literary texts by including the following dimensions in its assessment framework: (1) Retrieve texts and access them; (2) interpret and integrate texts; and (3) reflect and evaluate texts. In addition to the achievement score, PISA also collects information on students’ sociodemographic and dispositional characteristics, students’ home environments, and teaching and learning contexts in schools (Reference Lenkeit, Chan, Hopfenbeck and BairdLenkeit, et al., 2015) through the application of student and school questionnaires.

In order to examine the reading achievement of fifteen-year-old European students we focus on the most recent PISA 2018 data, which had reading as a main domain. Figure 5.4 shows the average reading score by country and SES across the PISA 2018 participating EU-MS, for which data are publicly available. The graph shows that there is great variation in students’ reading scores, with the difference in scores among EU-MS reaching about 100 points (Estonia vs Bulgaria). Estonia, Finland, and Ireland are the top performers, while the EU-MS with the lowest performance in reading are Bulgaria, Romania, and Malta.

Figure 5.4 Average reading achievement by country and by SES of fifteen-year-old students in twenty-five countries participating in the PISA 2018 EU-MS dataset

Notes: Countries are ranked in ascending order of the reading score average. High parental education denotes completion of tertiary education or higher education and low parental education denotes completion of up to upper-secondary education.

Source: PISA 2018, authors’ calculations

When comparing the reading scores for high- and low-SES students, using parental educational level as a proxy, we see that students whose parents have tertiary education perform better in reading than the ones whose parents have not completed higher education. Differences in students’ reading achievement vary from about one, in Croatia, to fifty-three points in Poland, favoring the students with a higher parental education level. These findings are in line with those from the previous section showing that differences in reading achievement according to students’ SES are already present in primary education. Additionally, this is in line with evidence from OECD that uses PISA 2009 data. In particular, across OECD countries students with a higher socioeconomic status outperform disadvantaged students in reading on average by thirty-eight score points, or about one year’s worth of education (OECD, 2010b). This corroborates evidence that there is an association between parents’ educational attainment, used as a proxy of students’ SES (Reference Jerrim and MicklewrightJerrim & Micklewright, 2014), and students’ achievement (Reference Pokropek, Borgonovi and JakubowskiPokropek, Borgonovi, & Jakubowski, 2015).

With respect to the home environment, and in accord with Sénéchal’s home literacy model, the PISA home questionnaire implemented in some countries in 2009 and 2018 asked parents how often they read to their children during their first year in primary education (OECD, 2012). Figure 5.5 presents the association between frequency of book reading and the reading achievement of fifteen-year-old students by parental educational level in EU-MS. It is evident that, in general and irrespective of parental educational level, high frequency of home book reading is associated with higher student achievement. The highest contribution of home reading for the achievement of students is found in Luxembourg and Belgium and the lowest difference is found in Croatia. These results are in line with those found using PIRLS data (Reference Araújo and CostaAraújo & Costa, 2015), showing that a higher frequency of home book reading is related to higher student achievement in the same MS, although these results do not pertain to the same cohort of students.

Figure 5.5 Average reading achievement by country, parental book reading, and SES of fifteen-year-old students in twenty-five countries participating in the PISA 2018 EU-MS dataset

Notes: Grey lines and symbols correspond to low parental education. Black corresponds to high parental education. Dotted lines correspond to low book-reading frequency and solid to high-Book-reading frequency. The graph includes EU-MS for which PISA 2018 data is available for the variable “home book reading.”

Source: PISA 2018, authors’ calculations

A comparison of home literacy in PIRLS and PISA shows that reading to young children gives them a head start in reading achievement, as retrospectively reported by parents in PIRLS (e.g., how often children were read to before the start of compulsory education) and during the first year of primary education in PISA (e.g., how often children were read to at the start of compulsory education).

Overall, the analyses presented in this chapter show that in PIRLS students are better readers in fourth grade when they were read to during the preschool years (Reference Araújo and CostaAraújo & Costa, 2015). Similarly, PISA 2018 students who were read to when in first grade are better readers at age fifteen, and this is in line also with findings using the PISA 2009 dataset (OECD, 2011). This highlights the positive reading outcomes accrued from home reading at a young age, a significant finding also with Australian children (Reference Kalb and OursKalb & van Ours, 2014) and with Canadian children (Reference Sénéchal and WasikSénéchal, 2012). Clearly, ILSAs such as PIRLS and PISA provide rich data for cross-country comparative analyses and the contextual information they collect allows us to understand what factors are associated with variations in the reading literacy performance of students. Although they are not longitudinal studies, the findings related with book reading in European countries converge with evidence from longitudinal research in Australia and Canada (Reference Kalb and OursKalb & van Ours, 2014, Reference Sénéchal and WasikSénéchal, 2012). The PISA results, however, also suggest that different European countries present particular relations between reading achievement and home book reading. In Croatia, for example, student scores change very little according to frequency of parental book reading with SES level. In contrast, in Poland, Belgium, and Bulgaria, score differences are much wider according to SES level. It is possible that children’s exposure to shared book reading in preschool kindergarten classrooms also varies among countries and that frequent exposure reduces the effects of home variables, but PIRLS and PISA do not collect such data. Research supports this conjecture and there is evidence that this effective early instruction practice promotes fourth-grade language and reading abilities (Reference Dickinson and PorcheDickinson & Porche, 2011).

5.4.1 Role of Socioeconomic Status

The analyses we present show that there are variations in achievement across countries and that socioeconomic variations within and across countries are associated with differences in reading achievement. High-SES students surveyed in PIRLS and PISA show higher achievement than low-SES students in all EU Member States. In addition, our analyses show that home book reading and early literacy skills can contribute to increased reading achievement in PIRLS for both groups of students. In this sense, early literacy skills and home practices before the start of primary school can help improve the reading abilities of students with different SES.

PIRLS research also shows that a school’s SES composition explains variation in students’ reading achievement in EU-MS (Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013). The study by Reference Stancel-Piatak, Mirazchiyski and DesaStancel-Piatak et al. (2013) found that in three PIRLS-2011-participating European countries – Denmark, Germany, and France – a school’s average of students’ cultural capital, measured by the number of books at home, was one of the most relevant variables in explaining variation in reading scores. Reference Myrberg and RosénMyrberg and Rosén’s study (2006) found that in PIRLS 2001, students’ cultural capital accounted for a large part of reading score differences between independent and public schools. These findings suggest that it is not only individual/home SES that matters, but that to a lesser extent a school’s SES compositional effect also shapes reading achievement in PIRLS (Reference Araújo and CostaAraújo & Costa, 2012; Reference Costa and AraújoCosta & Araújo, 2017). PISA findings corroborate this with adolescents (OECD, 2013b).

5.4.2 Role of Home Literacy Experiences

Different analyses of PIRLS confirm that both early literacy skills and practices, such as home book reading, in conjunction with other characteristics of effective schools, have a positive influence on achievement (Reference Araújo and CostaAraújo & Costa, 2012, Reference Costa and AraújoCosta & Araújo, 2017). Moreover, PIRLS studies suggest that motivational factors, like students’ enjoyment of reading, are positively related to achievement (Reference Costa and AraújoCosta & Araújo, 2017; Reference Stancel-Piatak, Mirazchiyski and DesaStancel-Piatak et al., 2013).

The home literacy environment, namely home book reading during the preschool years and the first year of primary school, contributes to higher reading achievement in PIRLS and PISA. This is in accord with Reference Sénéchal and WasikSénéchal’s (2012) home literacy model. In addition, and in accord with Reference Sénéchal and WasikSénéchal (2012) and with findings from research that looks at predictors of reading ability, the ability of young children to recognize the letters of the alphabet contributes to reading achievement in PIRLS. In this survey, parents are asked how well their children recognized the letters of the alphabet before the start of compulsory education. Children could have learned them at school or at home or in both settings. In this sense, this predictor of reading attainment in fourth grade is not restricted to the home literacy environment and research indicates that alphabet knowledge can be successfully acquired in school (Reference Shager, Schindler and MagnusonShager et al., 2013).

Findings from other studies that use PIRLS and PISA data and the information collected in student and parental questionnaires reveal interesting links between achievement and students’ sociodemographic and dispositional characteristics (Reference Lenkeit, Caro and StrandLenkeit, Caro, & Strand, 2015). For example, in PIRLS reading for enjoyment outside of school contributes to higher achievement in Denmark, Sweden, and France, after controlling for SES (Reference Costa and AraújoCosta & Araújo, 2017). Similarly, in all EU countries, secondary students who enjoy reading perform significantly better than students who have not developed an interest in reading (OECD, 2010a). This suggests that literacy practices can partially compensate for students’ socioeconomically disadvantaged backgrounds.

5.4.3 Role of Educational Factors

While the relationship of SES and home literacy practices with achievement is similar in all EU-MS, the relation of school-related literacy practices with achievement is more particular in terms of variation across countries. Similarly, for other school-level characteristics, such as emphasis on academic success, there is more variation across countries.

For example, research with PIRLS data shows that a school’s emphasis on academic success is related to higher achievement in France, but not in Denmark or Germany, which suggests that school effects are country specific (Reference Stancel-Piatak, Mirazchiyski and DesaStancel-Piatak et al., 2013). Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al. (2013) used PIRLS 2011 data to study the characteristics of effective schools and their findings corroborate the positive influence of good discipline and of emphasizing academic success. However, a school’s early emphasis on reading skills is only positively related to achievement in one European country – Germany. Conversely, the index of early literacy skills, which includes the ability to name alphabet letters before the start of compulsory education, is positively associated with higher achievement in all European countries (Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013).

Other factors related to schools and teachers have been found to also influence the achievement levels of fifteen-year-old students in PISA. For instance, students attending schools with a better disciplinary climate and better teacher–student relations perform better in reading (OECD, 2010c). This is supported by school effectiveness research, which suggests that school climate is a factor that explains variance in students´ achievement in reading (OECD, 2013b).

Clearly, PIRLS and PISA-related research has called attention to the school factors that seem to make a difference or that have “an effect on student achievement over and above” student/home predictors (Reference Martin, Foy, Mullis, O’Dwyer, Martin and MullisMartin et al., 2013, p. 111). Nonetheless, evidence of a school´s added value linked to student–teacher relations and instructional practices is limited, especially in PIRLS studies. Moreover, existing PIRLS studies suggest that factors like a school’s academic climate and literacy practices and their relationship with students’ achievement are more country specific than those factors associated with student-background variables (Stancel-Piatak et al., 2013).

6.3.1 Analytic and Global Methods

Russian literacy instruction is dominated by the analytical-synthetic method, which is focused on sound units in a word (Reference Lvov, Goretsky and SosnovskayaLvov, Goretsky, & Sosnovskaya, 2000). Methods that are based on global reading do not have wide popularity nor official recognition. They were relatively popular in the early 2000s, mainly in the parental environment: such methods include those outlined by Reference DomanDoman (1998), Reference ReznichenkoReznichenko (2012), and Reference ZaitsevZaitsev (2000), as well as Chaplygin’s cubes (2019).Footnote ¹ The first two approaches are global in the full sense, with children being introduced to a word as a whole. Doman’s method is more often used in kindergarten or even preschool. By contrast, Reznichenko’s method is designed for six–eight-year-old learners, gradually guiding them from a global to an analytical approach to decoding and coding the Russian orthography. The contemporary use of such strongly global methods was “tested” in Russia in the 1920s (Reference Lvov, Goretsky and SosnovskayaLvov et al., 2000). The syllabic teaching method (Reference ZaitsevZaitsev, 2000; Chaplygin’s cubes) can be classified as “semi-global”; its historical roots go back to the late nineteenth and early twentieth centuries. In this method, letter–sound correspondences are not explicitly taught; children practice syllable lists instead. This makes it easier to switch to analytical reading.

There are also many variations of the analytical-synthetic method. Russian creators of ABC books have gone from invariant models of teaching to the practical implementation of different technologies. In the Soviet period, from the end of the 1930s to the beginning of the 1980s, the literacy-teaching methodology was quite invariant. Children began school at the age of seven, with primary education for three years. The teacher had one common primer, and the only significant change that occurred during this period was the introduction of separate reading and writing lessons, that is, the transition from co-teaching reading and writing to parallel learning (Reference Lvov, Goretsky and SosnovskayaLvov et al., 2000). The teaching of the Russian language at school was characterized by a transition from practical language experiences to language generalizations, with literacy teaching based on moving from letter to sound. The sound values of a letter were simply communicated, and after a few exercises in which children were asked to find the target sound in a word, they were introduced to reading syllables and then words. These practice exercises began immediately after the introduction of letters, which made it possible for young learners to form implicit conclusions based on experience. The use of a special order in which to introduce the letters helped preserve the rule of simplicity in the lesson and not overload the child. The longest period in Soviet literacy studies is associated with the idea of the gradual introduction of those letters that are difficult in terms of their lack of transparency in letter–sound correspondences. For example, the letter M has only two sound values in any position in the word: [m] and [m ’]. Paired consonants, based on voiceless and voiced features, have four sound values, and the sound values of some vowels depend on their position. Therefore, in the Soviet school primers, the sonorants (m, n, r, l) were introduced first, then the paired unvoiced consonants, followed by voiced ones, with the complex vowels studied last.

From the early 1980s onward, reforms in the primary education system popularized the idea of the need for earlier literacy instruction (Reference SilchenkovaSilchenkova, 2006). This acceleration was perceived as a challenge to teaching practitioners. A gradual transition to four-year schooling began for six-year-old children. During this time, the literacy curriculum also changed by accounting for letter frequency in the Russian language. According to the curriculum developers (Reference Lvov, Goretsky and SosnovskayaLvov et al., 2000), this created the opportunity to offer full, lexically rich texts for reading earlier in the learning program, in order to contribute to a more comprehensive development of children’s literacy skills. However, this frequency-based approach led to the early introduction of complex letters. This problem of complexity was solved by increasing the proportion of exercises for phonological analysis, and by using detailed sound schemes that highlighted sound properties as follows: vowels, consonants, hard sounds, soft sounds, stressed sounds, and unstressed sounds. In our opinion, Goretsky’s curriculum of 1982 strengthened the phonics component in the Russian literacy education by instructing first-graders to master sound–letter connections and to make abstract generalization from these connections in words. Another systemic change in the 2000s was a transition from a unified to a more variable education program, with Russian schools receiving new primers and alphabet books. The elementary school began to be seen as a “school of cultural-activity pedagogy of variative education,” “the basic stage of designing universal educational activities,” the main task of which was to learn how to learn (Reference AsmolovAsmolov, 2013, p. 9).

Proponents of developmental education’s polemic against proponents of the traditional pedagogy had already been going on for a long time. Almost in parallel with the appearance of Goretsky’s alphabet book was the appearance of the school primer by Elkonin and collaborators based on the idea of development-centered child education (Reference ElkoninElkonin, 1992, Reference Elkonin and Feldstein1997). Its structure and content are founded on the principle of a functional meaning of letters and their role in Russian writing, and on the position-oriented principle. The position-oriented principle, on which Russian orthography is based, is actively deduced by students together with the teacher before reading. The path used is from sound to letter. First, sounds from a word are distinguished, then consonant characteristics (hard or soft) are clarified, after which a decision is made as to a letter that should represent the softness of the consonant and the actual vowel. This phonemic analysis precedes word decoding, and children only begin reading the first syllables several months later. This way of teaching literacy is based on the principle of ascent from the abstract to the concrete (Reference ZuckermanZuckerman, 2011). Even the introduction of the child to the rules of Russian morphology remains a subject of pedagogical discussion. And literacy is one of the most controversial issues here. It has been assumed that a preschooler with their syncretic perception of speech is faced with the need to be aware of complex linguistic phenomena and should therefore be exposed to a rich range of complex phonology in a structured way (Reference Velichenkova and VelichenkovaVelichenkova, 2018b). However, it has also been argued that a “language model developed for scientific purposes and a didactic model that is a way of presenting a language in the educational process are not required to coincide” (Reference LeontievLeontiev, 2016, p. 140).

Currently, Elkonin’s system is implemented in two school primers: Reference Zhurova and EvdokimovaZhurova and Evdokimova’s (2014) and Reference Repkin, Vostorgova and LevinRepkin, Vostorgova, and Levin’s (2019), with the former the most widely used (Reference Glagoleva, Arkhipova and BoykinaGlagoleva, Arkhipova, & Boykina, 2019).Footnote ² A recent federal list of textbooks that schools may use includes six school primers (Order of the Ministry of Education; 2019). The most popular is Goretsky’s textbook, which is used by about 40–60 percent of teachers (Reference Glagoleva, Arkhipova and BoykinaGlagoleva et al., 2019). Although Goretsky’s book is perceived by the pedagogical community as a traditional textbook, all the features of the new school methodology are inherent in it, including an increased role for analytical sound exercises and for the teaching of relevant knowledge in the field of phonetics.

6.3.2 Role of the Home Environment

In the contemporary Russian education system, we have started to observe phenomena very similar to the trends in Russia at the turn of the nineteenth and twentieth centuries: individual home-schooling flourishes in the context of state-education differentiation (Reference ShtetsShtets, 2009). These phenomena are especially characteristic in the initial stage of the child’s education. At the same time, one gets the feeling that the vectors of official and family approaches to teaching reading are regarded as opposites (Velichenkova, 2018а; Reference Velichenkova, Gorilko and BorisenkoVelichenkova & Gorilko, 2019). In the case of home literacy support, there is usually a parent acting as a teacher, and the book that is used by this “naive methodologist” is of particular importance. An approach often adopted by families who are trying to get away from the school primer is to look for a preschool one. Preschool textbooks do not require examination and testing; the choice to purchase them is the parents’ private matter. While only six school primers are on sale, there are about sixty preschool ones available. Despite such a huge selection and the significance of preschool family education, a comparative analysis of the manuals and methods in Russia has not been carried out.

Even more surprisingly, almost half of preschool primers are logopedic, with roots in speech therapy. It seems likely that there are objective factors that make logopedic textbooks attractive for parents. We attempted a comparative analysis of seven modern logopedic ABC books (Reference BatyaevaBatyaeva, 2017; Reference IlyinaIlyina, 2015; Reference KosinovaKosinova, 2006; Reference Krylova, Pisareva and IpatovaKrylova, Pisareva, & Ipatova, 1989; Reference NischevaNischeva, 2004; Reference ZhukovaZhukova, 2016; Reference ZhukovaZhukova, 2017). These primers usually emphasize that they have been compiled based on the authors’ extensive experience, but there is no scientific justification of this experience in peer-reviewed publications. Most authors highlight the traditional nature of the literacy-teaching methodology used in the textbook, and that it is supplemented by speech-therapy technologies. Parents wanting to find an alternative to the modern school primer and relying on their own experience of school accept this message and buy a traditional primer. For example, in the most popular “logopedic ABC book” by Reference ZhukovaZhukova (2017), the order in which letters are introduced keeps to the traditions of the Soviet period: The letters that come first are those that are unequivocal in terms of sound meaning. Other logopedic primers use a different order. Back in the 1960s, the idea emerged of postponing the study of late-mastered sounds (e.g., r, l, sh, s) to assist the literacy of children with a speech pathology. Since, in the case of speech disorders, these sounds are not being pronounced correctly by the time the child enters school, the corresponding letters are postponed to a later part of the curriculum.

Logopedic ABC-book preschool primers do not usually include many exercises for sound analysis and linguistic schematization. After all, they are focused on family education, and subtleties of phonology are inaccessible to the parent. If we compare the volume of speech material of three types (syllables, words, sentences), then it turns out that the logopedic ABC books on average are more extensive than, for example, the school primer by Reference Goretsky, Kiryushkin and VinogradskayaGoretsky et al. (2012) in terms of the numbers of syllables and words for reading. However, in terms of text volume, they are still behind due to the greater level of attention paid by their authors to reading a syllable and a word, and to their use of adapted texts that are shorter and simpler. Thus, in contradiction to the approach of strengthening the role of analytical strategy in schools, training in logopedic ABC books is being built rather more synthetically and at the same time more intuitively.

6.4.1 Monitoring Literacy Education

Preschool literacy teaching is a common phenomenon that is not regulated by the state. Let us try to briefly describe the state system of preschool and school education organization and education quality data. Preschool education is not obligatory; parents can use the services of the state kindergarten when their child reaches three years of age. There are state standards for preschool and school education that describe the requirements for the structure of educational programs, their implementation conditions, and how results should be assessed. The preschool standard (Federal State Educational Standards, 2019) emphasizes preschool education optionality in the Russian Federation. The educational program formulates its targets, but not children’s specific educational achievements. The program is compiled by the educational institution itself based on a sample program which the Ministry of Education has approved and posted on its website (Inventory of Sample Main Basic Education Programs, 2019). In addition to the development of all components of speech, it must only include the formation of sound analytical-synthetic activity as a precondition for literacy teaching. This means that in Russian kindergartens, literacy is not taught in an explicit way. A kindergarten teacher is not confronted with such a task. However, the teacher can conduct classes with children according to the chosen standard textbook, which to some degree develops sound-analytic and synthetic activity, whereas in some cases letters, syllables, and words are presented for literacy development.

Primary education starts at around seven years of age and lasts for four years. There are two state standards for elementary schools: general (valid since 2010) and those for children with special health needs (Health Impact Assessment HIA) (valid since 2016). The programs’ content formulates concrete targets for children to master in which reading literacy occupies a significant place. The disciplines in the philological block include the Russian language, literary reading, and a foreign language. At the time of writing, there are seven textbook series for teaching the Russian language and five for literary reading. The school carries out the assessments to establish whether students have mastered the program. This is done by means of marks attained at an intermediate stage in combination with the results of the final test. There is a state system for the monitoring of education quality. The marks received by schoolchildren through any such monitoring do not affect their final marks; they only serve as a tool for identifying problem areas and the education level in schools as well as in regions.

6.4.2 Literacy Outcomes

Russia has been participating in the PIRLS (Progress in International Reading Literacy Study) since 2001. From sixteenth place in its ranking, Russia has since climbed close to the top, where it has remained since 2006, which may be due to greater parental involvement. Research has consistently shown that Russian parents are the most active participants in the formation of children’s reading literacy among all the participating countries (Reference Mullis, Martin, Kennedy and FoyMullis et al., 2007 [PIRLS-2006]).

In PISA, reading literacy results have been extremely low. Starting in 2000 at twenty-seventh place, Russia kept worsening its position until 2012 (forty-second place), and then in 2015 returned to twenty-seventh place. The discussion about the poor results of Russian children in this study is ongoing (Reference Zuckerman, Kovaleva and KuznetsovaZuckerman, Kovaleva, & Kuznetsova, 2011). Obviously, students’ initial reading competency is well formed, but the complex of competencies that are usually described as making up functional literacy are not sufficiently developed.

The All-Russian Test Work initiative is compulsory in elementary schools; its tests are held once a year on the main subjects (the Russian language in all classes). As to outcomes on the All-Russian Test Work over the past five years, 4.6 percent of the population showed literacy problems.Footnote ³ For writing, 7.8 percent of primary-school graduates could not cope with the test, as shown from their performances on the ability to copy a text, use of spelling and punctuation norms, and independent writing.

6.5.1 Assistance for Children with Disabilities

Recently, progress has been made in the assistance system for children with learning disabilities. The system has undergone significant changes since 2013, especially in schools. The Law on Education was adopted in 2013 and introduced the concepts of “children with limited health abilities” and “adapted educational programs” (Federal Law of December 29, 2012). A child with limited health abilities (LHA) is defined as “an individual with physical and/or psychological developmental deficiencies confirmed by a psychological, medical and pedagogical commission as hindering education if special arrangements are not in place” (Federal Law of December 29, 2012). Adapted educational programs were created to teach children with disabilities, taking into account the peculiarities of their psychophysical development, individual capabilities, and social adaptation. These programs were focused on children with vision, speech, hearing, and motor impairments, as well as delayed mental development. It is important to note that reading and writing impairments are not specifically distinguished from each other (Reference Volkova and ShakhovskoyVolkova & Shakhovskoy, 1998). Reading and writing disorders are considered only because of either a phonological processes deficiency or general underdevelopment of speech. The terms dyslexia and dysgraphia are simply absent from the school standard, although, as we will see later, they are widely used in professional support systems.

Children who have difficulty in mastering reading and writing are often diagnosed as having a speech disorder or a delay in mental development. Their parents are advised to apply for an adapted program for the child based on a more intensive standard program or an approximate adapted program, (partly) outside the classroom. This new assistance system differs from the one that existed before primarily by virtue of its strengthening of the mediating role of medical-psychological-pedagogical commissions. Unfortunately, there are many unsolved problems with this system. The most important one is personnel. There are no mechanisms to encourage schools to employ a speech therapist. Furthermore, parents tend to mistrust the input of this new institute of medical-psychological-pedagogical commissions.

6.5.2 Public Perception of Children with Reading Disabilities

In 2019, on the initiative of the Association of Parents and Children with Dyslexia and with the financial support of the Our Future Foundation and the Ministry of Education of the Russian Federation, an independent research agency (MAGRAM MR) carried out an unprecedented survey to determine the level of awareness and attitudes toward problems with writing and reading in society and among professionals. Our role was to expertly evaluate the results. The all-Russian population survey included 2,516 telephone interviews (the sample represented the population of the Russian Federation by sociodemographic characteristics). The survey of professionals was carried out using an online questionnaire. The support of the Ministry of Education of the Russian Federation made it possible to collect data on 32,406 school specialists.

The survey found that only 17 percent of the Russian population understands what dyslexia or dysgraphia are. Eighty-three percent do not know these terms: 61 percent of these respondents do not know anything about the problem; 22 percent have heard of the terms but cannot explain their meaning. Young participants (from eighteen to thirty-four years old) demonstrate a higher level of knowledge of the terms in question than representatives of the older age groups. According to almost half of participants (49 percent), one of the main causes of dyslexia/dysgraphia is parents’ neglect their children and failing to provide the necessary participation in their upbringing and education. Thus, the child’s health problem is “shifted” onto the parents. Only a quarter to a third of respondents said that the cause of the problem relates to the child’s neurobiological characteristics. A fourth of respondents indicated that the causes of the problem under consideration lie in the irregularly organized school educational process, including a high level of high teacher workload (28 percent), high-school requirements (25 percent), and insufficient attention from support specialists (24 percent). More than a third of respondents (38 percent) believed that dyslexia in children is quite widespread. However, only 2 percent noted that their family had a child with dyslexia, and less than half of these children received specialized assistance.

To explore professional support for children with dyslexia, school staff were interviewed: primary-school teachers (23,849) school-department managers (3,497), psychologists (2,851), speech therapists (1,549) and speech pathologists (660). Ninety-eight percent of the speech therapists, 87 percent of the speech pathologists, 80 percent of the psychologists, 73 percent of the headteachers, and 66 percent of the teachers were familiar with the problem. Among causes for children’s writing and reading disabilities, teachers mentioned the child’s individual characteristics (63 percent). Half of the survey participants noted parents’ insufficient attention to the problem – rejection of the problem (54 percent) and lack of proper participation in education and training (49 percent). Fifty-three percent also mentioned lack of timely access to specialists. Among the professionals surveyed, the prevalence of the problem was rated as high by most specialists. Content-wise, 67 percent of pedagogical community representatives indicated that remedial classes with underachieving children were held. But many of them (38 percent) stated that classes were held much less frequently than required. Seventy-three percent of the professional representatives interviewed pointed out the insufficient number of psychological and pedagogical support specialists working with children with dyslexia, and 38 percent of them indicated that a substantial replenishment of specialized professionals is necessary. It can thus be concluded that the gap in awareness between the general population and professionals concerning learning disabilities in children and assessing their prevalence is significant.

6.5.3 Research on Developmental Dyslexia

The terms dyslexia and dysgraphia are widely used in the clinical and pedagogical classification of speech disorders. In Russia it is customary to consider reading and writing disorders as separate from each other (Reference Akhutina and VelichenkovaAkhutina, 2018; Reference Grigorenko and ElliottGrigorenko & Elliott, 2012; Reference SadovnikovaSadovnikova, 2011; Reference TokarevaTokareva, 1971; Reference Volkova and ShakhovskoyVolkova & Shakhovskoy, 1998). And this relates to the greater difficulty of mastering the Russian coding than decoding. Dyslexia is generally defined as persistent difficulties in reading. The diagnostic inclusion criterion for dyslexia is the same as for dysgraphia. However, most authors simply indicate the absence of intellectual disability as an important symptom for diagnosis (Reference SadovnikovaSadovnikova, 2011; Reference Velichenkova and RusetskayaVelichenkova & Rusetskaya, 2015; Reference Volkova and ShakhovskoyVolkova & Shakhovskoy, 1998). Reference KornevKornev (1995, Reference Kornev2004, Reference Kornev and Ishimova2010) insisted on the need to apply the criterion of reading level and level of mental development discrepancy. Defining dysgraphia, specialists normally name diagnostic criteria for inclusion (Reference KornevKornev, 1995; Reference SadovnikovaSadovnikova, 2011; Reference Velichenkova and RusetskayaVelichenkova & Rusetskaya, 2015). The main criterion for inclusion is the presence of errors associated with nonmastery of the graphics and a high frequency of such errors. The term dysorthography has been actively used in scientific and pedagogical practice (Reference KornevKornev, 1995; Reference PrishchepovaPrishchepova, 2006; Reference Velichenkova and RusetskayaVelichenkova & Rusetskaya, 2015), alongside the term dysgraphia, referring to the lack of spelling understanding. The diagnostic criteria remain the same for these conditions except for the higher frequency and different type of spelling errors associated with dysgraphia; these indicate a failure to master the morphological, grammatical, and traditional principles of writing.

Standardized reading assessment methods are based on the accuracy and speed of single-word reading (Reference Korneev, Akhutina and MatveevaKorneev, Akhutina, & Matveeva, 2019; Reference Akhutina and VelichenkovaRybchinskaya, Korneev & Akhutina, 2018) and text reading as well as reading comprehension (Reference Akhutina and InshakovaAkhutina & Inshakova, 2008; Reference Rybchinskaya, Korneev, Akhutina, Krylov, Soloviev and KibrikRybchinskaya et al., 2018). In a study by Reference Rusetskaya and VelichenkovaRusetskaya (2018), students in Grades 1–4 were given the same texts to read. It was shown that on average, correct scores increased from 91 percent in the first grade to 97 percent in the fourth grade. Reading efficiency as measured by the words read correctly in one minute increased from 45 words in the first grade to 67 words in the second grade, 84 words in the third grade, and 101 words in the fourth grade. These data show that reading problems in Russian are mainly a matter of reading speed. A psychometric procedure for assessing reading comprehension was proposed by Reference Akhutina and InshakovaAkhutina and Inshakova (2008). They assessed the integrity and coherence of retelling and comprehension in 197 third-graders by addressing accuracy, semantic adequacy of retelling, its deployment, and lexical-grammatical design. It was found that 11 percent of the children showed reading-comprehension problems.

With respect to dysgraphia, we obtained data on the average number of dysgraphic errors in the dictations of sixty-six third-graders in two Moscow schools in 2018. In total, about 600 control dictations of 45–70 words written by the children over one year were analyzed. Dysgraphic errors (2.01 on average) included omissions of letters, mixing graphically similar letters (т-п, б-д, и-у), mixing by acoustic-articulation similarity, rearrangement, perseveration, anticipation of letters, and violation of soft-consonants designation in writing. Furthermore, the development of writing in dictation, copying from printed material, and handwritten text was examined in 216 children from Grades 1 to 4 (Reference InshakovaInshakova, 2013). It was found that children made the most errors in dictation and that the numbers of errors remained constant throughout the grades.

9.5.1 Role of Phonological Awareness

China is a country where many different dialects are spoken.Footnote ⁷ These varieties are generally classified into seven major groups: Mandarin, Wu, Xiang, Gan, Hakka, Min, and Yue (Reference Li and ThompsonLi & Thompson, 1981). Among these, Mandarin is the standard language of China and used by the largest percentage of the population. Unlike English, which contains more than 6,000 syllables, Mandarin has only about 400 basic syllables and a total of 1,300 syllables when different tones are considered (Reference DuanmuDuanmu, 2007). Mandarin consists of twenty-one initial consonants, two final consonants, three glides, and six vowels (Reference DuanmuDuanmu, 2007). Possible syllable combinations in Mandarin include V (爱, /ai4/), CV (大, /da4/), VC (安, /an1/), and CVC (晚, /wan3/) (Yang, 1988). Mandarin is a tonal language; when the tone of a syllable changes, it turns into a different morpheme. For example, 妈 /ma1/ means mother, whereas 马 /ma3/ means horse. Mandarin has four different tones – high level, rising, fall-rise and high falling. Tones are usually indicated by either tone marks or the numbers 1, 2, 3, and 4 for ease of writing. A small number of syllables (e.g.,了 /le/) can be marked with a neutral tone, also called a zeroth tone. These syllables are shorter and less meaningful than other syllables. Chinese has a large number of homophones. On average, each tone syllable corresponds to five morphemes. These morphemes share the same pronunciation but are represented by idiosyncratic characters in print (e.g., 言, 岩, 盐, 颜, 严, 研, 阎, /yan2/) (Reference DuanmuDuanmu, 2007).

Cantonese is a dialect mainly spoken in Southeast China, including the provinces of Guangdong and Guangxi, and Hong Kong and Macao. Cantonese has a more complex phonological structure than Mandarin (Reference Chen, Anderson and LiChen et al., 2004). It consists of nineteen initial consonants, eight final consonants, and eight vowels (Reference So and DoddSo & Dodd, 1995, Reference ZengZeng, 1994). There are six tones in Cantonese: high level, high rising, mid-level, low falling, low rising, and low level. There are also three “entering tones,” the high-, mid- and low-stopped tones, which appear with syllables ending with /p/, /t/, and /k/. These tones are much shorter in duration and are considered allotones of the three level tones (Reference Bauer and BenedictBauer & Benedict, 1997; Reference So and DoddSo & Dodd, 1995). Despite the considerable differences between Cantonese and Mandarin, the two dialects, as well as all other dialects in China, are represented by the same written text, based on Mandarin. In other words, all Chinese children learn to read in Mandarin, though for many it is not their native dialect. This situation creates unique challenges that are currently understudied and need to be addressed by future research.

Phonological awareness refers to the ability to reflect on and manipulate smaller units of sounds in spoken words (Reference GoswamiGoswami, 2000). Phonological awareness consists of three different levels – syllable, onset-rime, and phoneme in most languages (Reference Lin, Cheng and WangLin, Cheng, & Wang, 2018; Reference Yoon, Bolger, Kwon and PerfettiYoon et al., 2002). Tone awareness is a unique aspect of phonological awareness in Chinese (Reference Chen, Pasquarella, Verhoeven and PerfettiChen & Pasquarella, 2017; Reference Treiman, Zukowski, Brady and ShankweilerTreiman & Zukowski, 1991). There is a one-to-one association between syllables and morphemes in Chinese, and both correspond to characters. For this reason, syllable awareness is salient for children at an early age (Reference Pan, Song and SuPan et al., 2016; Reference Zhang, Lin, Wei, Anderson, Chen, Wang and YangZhang et al., 2014). In Mainland China, when beginning readers learn to read characters using Pinyin, a syllable is usually split into two basic components: an onset (声母 /sheng1mu3/), which is the initial consonant of the syllable, and a rime (韵母 /yu4mu3/), which includes the rest of the syllable (Reference TřískováTřísková, 2011). As a result, onset-rime awareness is also important for reading characters. Finally, because Chinese is a tonal language, Chinese speakers must possess tone awareness to distinguish morphemes with the same syllable but different tones (少 /shao3/ -less, and 勺/shao2/-spoon).

An extensive body of research has found phonological awareness to be critical for success in word reading in alphabetic languages (Reference Lesaux and SiegelLesaux & Siegel, 2003; Reference Linan-Thompson, Vaughn, Prater and CirinoLinan-Thompson et al., 2006). Chinese, as a morphosyllabic script, differs substantially on orthography–phonology correspondences from alphabetic languages (i.e., Korean, English) or languages based on alphasyllabaries (i.e., Japanese). The link between phonological awareness, particularly phonemic awareness, and reading is less straightforward due to the lack of grapheme–phoneme correspondences in Chinese. However, given the use of Pinyin and phonetic radicals to denote character pronunciations, some aspects of phonological awareness, such syllable and tone awareness, may still play an important role in Chinese reading (e.g., Reference Ruan, Georgiou, Song, Li and ShuRuan et al., 2018). It has been found that learning a phonetic alphabet such as Pinyin enhances phonological awareness among Chinese children. Children from Mainland China and Taiwan perform better on phonological-awareness measures compared to their peers from Hong Kong, who read without the assistance of a transliteration system (Reference Huang and HanleyHuang & Hanley, 1995; Reference McBride-Chang, Bialystok, Chong and LiMcBride-Chang et al., 2004). Furthermore, research has shown that syllable and tone awareness are both significant predictors of word reading in Chinese (Reference McBride-Chang, Tong and ShuMcBride-Chang et al., 2008; Reference Tong, Lee, Lee and BurnhamTong et al., 2015). For example, syllable awareness was found to act as a stronger predictor of Chinese word reading than onset-rime awareness in the early grades (Reference Pan, Song and SuPan et al., 2016; Reference Shu, Peng and McBride‐ChangShu, Peng, & McBride-Chang, 2008). By contrast, research examining onset-rime awareness has generated mixed findings. While some studies found onset-rime awareness to predict Chinese word reading (e.g., Reference Ho and BryantHo & Bryant, 1997; Reference Wang, McBride-Chang and ChanWang et al., 2014), others did not (e.g., Reference McBride-Chang, Tong and ShuMcBride-Chang et al., 2008).

9.5.2 Role of Morphological Awareness

Morphemes are the smallest units of meaning. They are combined to build complex words in a language. A root carries the meaning of a word, whereas affixes are auxiliary components that attach to a base.Footnote ⁸ Affixes can be further divided into inflectional and derivational affixes. Inflectional affixes denote grammatical functions (e.g., book-books). Derivational affixes form a new word from the base unit, and this new word often belongs to a different word class (e.g., book-bookish). The morphological system in Chinese has many distinct features. Chinese does not have grammatical agreement, morphological paradigms, and morphophonemic alternation (Reference PackardPackard, 2000). About 70 percent of Chinese words are lexical compounds formed by combining two roots (e.g., cow: beef meat) (Reference PackardPackard, 2000). In Chinese, one morpheme is typically represented by one syllable in its spoken form and one character in print, for example 天/tian1/ “sky” (Reference DeFrancisDeFrancis, 1984). Occasionally, loan words consist of more than one syllable, for example, 咖啡/ka1fei1/ “coffee,” because characters are used to indicate sounds. In addition to homophonic morphemes, Chinese contains homographic morphemes, which have the same pronunciation and orthographic form. In other words, the same character can correspond to different morphemes. For example, 面 /mian4/ means flour in 面粉/mian4fen3/ (flour powder) but surface in 桌面/zhuo1mian4/ (desk top).

Morphological awareness can be defined as the awareness of the morphological makeup of words and the ability to reflect on and manipulate their constituent morphemic parts. Due to the prominence of lexical compounding and a high density of homophony in Chinese, researchers have examined the role of compound awareness and homophone/homograph awareness in children’s reading development. A morphological construction task is often used to measure compound awareness (Reference Chen, Hao, Geva, Zhu and ShuChen et al., 2009; Reference McBride-Chang, Shu, Zhou, Wat and WagnerMcBride-Chang et al., 2003). Children are first presented with the definition of a familiar word and are then asked to create a novel word that has the same morphological structure. For example, “Striped horse (zebra) is a horse with stripes on its body. What should we call a cow with stripes on its body?” (斑马是身上有斑纹的一种马, 那么身上有斑纹的牛我们叫什么?). The correct answer is “striped cow” (斑牛). Homophone/homograph awareness is often assessed with a morpheme-judgment task. For example, in a homophone-awareness task, children are aurally presented with a word containing the target morpheme (e.g., /hua4/ in 画家 /hua4jia1/ – “paint artist: painter”) and then are asked to choose the word that contains the same morpheme from several options (e.g., 话剧/hua4ju4/– “speech opera: drama”; and 画布 /hua4bu4/ – “paint fabric: canvas”) (Reference McBride-Chang, Shu, Zhou, Wat and WagnerMcBride-Chang et al., 2003; Reference McBride-Chang, Cho and LiuMcBride-Chang et al., 2005; Reference Shu, McBride-Chang, Wu and LiuShu et al., 2006). In a homograph-awareness task, children are aurally presented with a compound word containing a target morpheme (e.g., /hua1/ in 花园 /hua1yuan2/ – “flower yard: garden”), then asked to produce two words with the target morpheme having different meanings in the words (e.g., /hua1/ in 花朵 /hua1duo3/ – “flower blossom” and /hua1/ in 花钱 /hua1qian2/ – “spend money”) (Reference Xie, Zhang, Wu and NguyenXie et al., 2019).

Both lexical compounding awareness and homophone/homograph sensitivity have been found to predict reading outcomes among Chinese children (Reference Chen, Hao, Geva, Zhu and ShuChen et al., 2009; Reference Tong, Tong and McBrideTong, Tong, & McBride, 2017). As the majority of Chinese words are compound words, it is crucial to understand how morphemes can be legally combined to form morphologically complex words (Reference ChanChan, 2013). Studies have found that after controlling for other relevant variables, compound awareness has significant effects on Chinese character reading (e.g., Reference Liu and McBride-ChangLiu & McBride-Chang, 2010), vocabulary (e.g., Reference Liu and McBride-ChangLiu & McBride-Chang, 2010; Reference Liu, McBride-Chang, Wong, Shu and WongLiu, McBride-Chang et al., 2013), and reading comprehension (e.g., Reference Li, Zhang and XiaLi et al., 2017; Reference Xie, Zhang, Wu and NguyenXie et al., 2019). In addition, the high density of homophones and homographs requires children to distinguish among characters that have the same pronunciation (and sometimes also the same orthographic form) but represent different morphemes. Research shows that children who are more sensitive to homophones (Reference Chung and HuChung & Hu, 2007) and homographs (Reference Li, Dong, Zhu, Liu and WuLi et al., 2009) perform better on Chinese character reading and vocabulary tasks (Reference Cheng, Wu, Liu and LiCheng et al., 2017; ; Reference Ku and AndersonKu & Anderson, 2003; Reference Liu, McBride-Chang, Wong, Shu and WongLiu, McBride-Chang et al., 2013; Reference Shu, McBride-Chang, Wu and LiuShu et al., 2006; Reference Wu, Anderson and LiWu et al., 2009). Due to the one-to-one mapping between syllables and morphemes in Chinese, studies comparing morphological and phonological awareness simultaneously often report that the former plays a larger role in reading (e.g., Reference Li, Anderson, Nagy, Zhang, Li, Gaffney and PackardLi et al., 2002; Reference McBride-Chang, Cho and LiuMcBride-Chang et al., 2005; Reference Pan, Shu, Wang and YanPan et al., 2015). This pattern is confirmed by a meta-analysis by Reference Ruan, Georgiou, Song, Li and ShuRuan et al. (2018), who reported that morphological awareness was more strongly associated with word reading and reading comprehension than with phonological awareness in Chinese. Interestingly, the same study found phonological awareness to play a stronger role than morphological awareness in reading English.

9.5.3 Role of Orthographic Knowledge

Characters are composed of strokes, which are the smallest units of written form in Chinese. A stroke is usually a line (一) or a dot (丶). A stroke can be straight or curved (丿), vertical (丨), horizontal, or diagonal (), and it can contain a “hook” (亅) (Reference Taylor and TaylorTaylor & Taylor, 1995). The visual complexity of a character is usually identified by the number of individual strokes within the character. Characters of higher frequency usually contain fewer strokes, whereas those of lower frequency have more strokes (Reference Li, Zhang and EhriLi et al., 2016; Reference Shu, Chen, Anderson, Wu and XuanShu et al., 2003). On average, the total number of strokes in characters written in the simplified script is approximately 22.5 percent fewer than those written in the traditional script (Reference Gao, Kao and KaoGao & Kao, 2002). Interestingly, readers who have been exposed to traditional Chinese characters showed higher visual-perception skills than those who have been exposed to simplified characters (Reference Chang, Chen and PerfettiChang, Chen, & Perfetti, 2018). Radicals, formed by stroke combinations, are functional units of Chinese characters. There are approximately 800 phonetic and 200 semantic radicals in Chinese (Reference Hoosain, Chen and TzengHoosain, 1992). Most radicals are independent characters with their own meanings and pronunciations; a small number of them are in bound forms that can only appear as a part of a character (Reference Shu, Chen, Anderson, Wu and XuanShu et al., 2003). Among the four common types of characters, pictographs and ideographs are simple characters without an internal structure. Semantic-phonetic and semantic-semantic compound characters are complex characters made up of radicals. The majority of compound characters consist of a left–right structure (e.g, 彩 /cai3/); others have a top–bottom structure (e.g, 菜 /cai4/) or an inside–outside (surrounding) structure (困 /kun4/) (Reference Zhang, Li, Dong, Xu and SholarZhang et al., 2016).

Orthographic knowledge in Chinese refers to knowledge of the internal structure of Chinese characters (Reference Ho, Ng and NgHo, Ng, & Ng, 2003). This knowledge includes the position, function, and legal combination of phonetic and semantic radicals that form compound characters. With respect to position, semantic radicals usually appear on the left (75 percent) or top (15 percent) of a character (Reference Feldman and SiokFeldman & Siok, 1999), whereas the common positions for phonetic radicals are right or bottom. Most radicals always appear in the same position, but a small number of them appear in different positions across different characters. Functionally, semantic and phonetic radicals provide cues to meaning and pronunciation respectively. The Chinese writing system is more accurate in representing meaning than phonology. Semantic radicals are generally informative of character meaning. In contrast, a phonetic radical has the same syllable as the character in only about 40 percent of the characters (Reference ShuShu, 2003). The percentage is even lower when tone is considered. Finally, not all semantic and phonetic radicals can combine to form characters. Simply combining a semantic radical and a phonetic radical in their legal positions may or may not form a real character together.

Due to the unique characteristics of the Chinese writing system, a variety of measures have been designed to assess orthographic knowledge. For example, in an orthographic judgment task, children are presented with different types of “novel” characters and asked to judge whether each one could be a real character (Reference Li, Shu, McBride‐Chang, Liu and PengLi et al., 2012). These items may include line drawings with no conventional stroke patterns (), noncharacters with real radicals in illegal positions (), pseudo-characters with real radicals that appear in legal positions but do not combine together (), pseudo-characters with subcomponents that do not exist in Chinese (). Furthermore, specific tasks have been designed to assess children’s awareness of the position and function of phonetic (Reference Ho, Yau, Au, McBride-Chang and ChenHo, Yau, & Au, 2003; Reference Yeung, Ho, Chan and ChungYeung et al., 2016a) and semantic radicals (Reference Yeung, Ho and ChikYeung et al., 2011; Reference Zhang, Li, Dong, Xu and SholarZhang et al., 2016). For example, in a phonetic radical knowledge task, a pseudo-compound character () is paired with three real characters, one with the same semantic radical (破), one with the same phonetic radical (味), and one that does not share any radicals in common (埋). Children are asked to choose the character (味) that has the same pronunciation as the target character. In a semantic radical knowledge task, each item is paired with four pictures, and children are asked to choose the picture that best represents the meaning of the semantic radical (Reference Ho, Ng and NgHo, Ng, & Ng, 2003).

Research has shown that children start to form mental representations of radical forms and positions in Grade 1, and these representations are well established by Grade 3 (Reference Chan and NunesChan & Nunes, 1998; Reference Ho, Yau, Au, McBride-Chang and ChenHo, Yau, & Au, 2003; Reference Li, Shu, McBride‐Chang, Liu and PengLi et al., 2012). Children are faster at rejecting noncharacters that violate legal radical positions than pseudo-characters that follow the legality of radical positions in lexical-decision tasks (Reference Qian, Song, Zhao and BiQian et al., 2015; Reference Wang, Perfetti and LiuWang, Perfetti, & Liu, 2005). Children are able to derive meanings of unfamiliar compound characters using familiar semantic radicals (Reference Anderson, Li, Ku, Shu and WuAnderson et al., 2003; Reference Ho, Ng and NgHo, Ng, & Ng, 2003; Reference Shu and AndersonShu & Anderson, 1997) and pronunciations using phonetic radicals (Reference Anderson, Li, Ku, Shu and WuAnderson et al., 2003; Reference Ho and BryantHo & Bryant, 1997). The pronunciation of an unfamiliar character (e.g, 蜻 /qing1/) can also be predicted by making an analogy to a character (e.g., 清 /qing1/) with the same phonetic radical (e.g., 青 /qing1/) (Reference Ho, Ng and NgHo, Ng, & Ng, 2003). Importantly, research shows that orthographic knowledge of both semantic and phonetic radicals is related to word reading (Reference Leong, Tse, Loh and KiLeong et al., 2011; Reference Peng, Li, Yang and ChenPeng, Li, & Yang, 1997; Reference Yeung, Ho, Chan and ChungYeung et al., 2016b) and spelling (Reference Yeung, Ho, Chan and ChungYeung et al., 2016b). For example, two studies (Reference Ho, Ng and NgHo, Ng, & Ng, 2003; Reference Yeung, Ho and ChikYeung et al., 2011) observed that semantic radical knowledge predicted word reading in Grades 1 and 3, and phonetic radical knowledge predicted word reading in Grades 1, 3, and 5. Reference Yeung, Ho, Chan and ChungYeung et al. (2016a) found that both types of radical knowledge predicted word spelling in Grades 2 and 4. Finally, knowledge of semantic radicals has been found to contribute to reading comprehension (Reference Ho, Ng and NgHo, Ng, et al., 2003). Because the majority of the compound characters introduced in the elementary school curriculum are semantically transparent, semantic radicals provide reliable information about character meaning (Reference Shu, Chen, Anderson, Wu and XuanShu et al., 2003). For example, Reference Yeung, Ho, Chan and ChungYeung et al. (2016a) reported that semantic radical knowledge uniquely contributed to reading comprehension among Chinese children in Grade 4.

9.5.4 Word Reading and Spelling

According to the lexical-quality hypothesis (Reference PerfettiPerfetti, 2007), word identity consists of orthography, phonology, grammar, and meaning, and these features are bound together to secure coherence (binding). Studies have provided evidence in support of the lexical-quality hypothesis in Chinese (Reference Guan and WangGuan & Wang, 2017). Accordingly, phonological awareness, morphological awareness, and orthographic knowledge have all been found to contribute to word reading in Chinese children. With respect to phonological awareness, research has shown that both syllable and tone awareness play an important role in Chinese word reading due to the salience of these linguistic units (Reference McBride-Chang, Tong and ShuMcBride-Chang et al., 2008; Reference Tong, Lee, Lee and BurnhamTong et al., 2015). In terms of morphological awareness, both lexical compounding awareness (Reference Liu and McBride-ChangLiu & McBride-Chang, 2010) and homophone/homograph sensitivity (Reference Chung and HuChung & Hu, 2007) predict word reading because Chinese consists of a large number of compound words and homophones/homographs. In addition, the knowledge of the position, function, and legal combination of phonetic and semantic radicals of compound characters is a reliable contributor to Chinese word reading (Reference Ho, Yau, Au, McBride-Chang and ChenHo, Yau, & Au, 2003; Reference Leong, Tse, Loh and KiLeong et al., 2011; Reference Peng, Li, Yang and ChenPeng et al., 1997; Reference Yeung, Ho, Chan and ChungYeung et al., 2016a). Metalinguistic skills have also been found to contribute to spelling. Previous studies have shown that phonological awareness (Reference Tong, McBride-Chang, Shu and WongTong et al., 2009; Reference Yeung, Ho and ChikYeung et al., 2011), orthographic processing (Reference Ho, Ng and NgHo et al., 2003; Yeung et al., 2015), and morphological awareness (Reference Tong, McBride-Chang, Shu and WongTong et al., 2009; Reference Yeung, Ho, Chan and ChungYeung et al., 2016a) are all related to spelling acquisition in Chinese.

Because Chinese characters are visually complex, repetition of character writing, following the correct stroke order, is a popular approach to fostering children’s literacy development (Reference Lam and McBrideLam & McBride, 2018; Reference Pine, Ping’an and Ren SongPine, Ping’an, & Ren Song, 2003). Several recent studies have examined the role of visual-motor integration in reading and spelling Chinese characters. Visual-motor integration refers to the ability to coordinate visual information and hand movement (Reference Longcamp, Anton, Roth and VelayLongcamp et al., 2005) and is measured by character-copying tasks in Chinese children. This small body of research has provided evidence that visual-motor integration is related to both reading (Reference McBride-Chang, Zhou and ChoMcBride-Chang et al., 2011; Reference Meng, Wydell and BiMeng, Wydell, & Bi, 2019) and spelling (Reference Lam and McBrideLam & McBride, 2018) in Chinese. For example, Reference Wang, McBride-Chang and ChanWang, McBride-Chang, and Chan (2014) observed that after controlling for age and nonverbal intelligence, the ability to copy unfamiliar scripts (Korean, Vietnamese, and Hebrew) uniquely contributed to spelling in kindergarteners (Reference Wang, McBride-Chang and ChanWang et al., 2014). Reference McBride-Chang, Zhou and ChoMcBride-Chang et al. (2011) reported that this ability differentiated Chinese children in Grades 3 and 4 with and without dyslexia and uniquely contributed to reading in the combined sample. In addition, orthographic working memory, measured using a delayed copying task (where children were briefly presented with unfamiliar characters and then asked to write them down from memory), was found to predict both word reading (Reference Chung, Ho, Chan, Tsang and LeeChung et al., 2011) and spelling (Reference Mo, McBride and YipMo, McBride, & Yip, 2018; Reference Wang, McBride-Chang and ChanWang et al., 2014) among kindergartners in Hong Kong.

9.5.5 Reading Comprehension

The simple view of reading model (SVR) proposes that reading comprehension is the cross product of two main components, decoding and listening comprehension (Reference Gough and TunmerGough & Tunmer, 1986; Reference Hoover and GoughHoover & Gough, 1990). Skilled decoding is defined by Reference Hoover and GoughHoover and Gough (1990) as the ability to “rapidly derive a representation from printed input that allows access to the appropriate entry in the mental lexicon” (p. 130), whereas listening comprehension refers to the ability to use phonological, semantic, syntactic, and discourse information to understand language (Reference Hogan, Adlof and AlonzoHogan et al., 2014). Although Chinese is represented by a logographic writing system, there is evidence supporting the SVR model in Chinese (Reference Ho, Zheng and McBrideHo et al., 2017; Reference Joshi, Tao, Aaron and QuirozJoshi et al., 2012; Reference Yeung, Ho, Chan and ChungYeung et al., 2016b). Since characters cannot be read through grapheme–phoneme correspondences, the decoding component in the SVR model is measured by character- and word-reading accuracy and fluency. As for listening comprehension, studies have measured children’s vocabulary knowledge, morphological awareness, morphosyntactic skills, and discourse skills (Reference Chik, Ho and YeungChik et al., 2012; Reference Yeung, Ho, Chan, Chung and WongYeung et al., 2013). Generally speaking, studies examining the SVR model in Chinese have found that decoding and listening comprehension explain a significant amount of variance in reading comprehension, confirming the applicability of the SVR model (Reference Ho, Zheng and McBrideHo et al., 2017;Reference Yeung, Ho, Chan and ChungYeung et al., 2016b; Reference Zhang, McBride-Chang and TongZhang et al., 2012). For example, in Reference Yeung, Ho, Chan and ChungYeung et al. (2016b), the best-fitting model with decoding and listening comprehension explained 83 percent of the variance in reading comprehension among Grade 1 children in Hong Kong, with decoding and listening comprehension accounting for 37 and 46 percent of the variance respectively.

As mentioned in the previous sections, research has examined the contribution of metalinguistic skills (morphological awareness, phonological awareness, and orthographic knowledge) to reading comprehension in Chinese. Studies have demonstrated that morphological awareness, including compound awareness and homophone awareness, contribute to both concurrent and longitudinal reading comprehension, even after controlling for phonological awareness and speed naming (Reference Tong, McBride-Chang, Shu and WongTong et al., 2009; Reference Xie, Zhang, Wu and NguyenXie et al., 2019; Reference Zhang, McBride-Chang and TongZhang et al., 2012; Reference Zhang, Lin, Wei, Anderson, Chen, Wang and YangZhang et al., 2014). Phonological awareness is also shown to play an significant role in reading comprehension among Chinese children (Reference Lau and ChanLau & Chan, 2007; Reference Ruan, Georgiou, Song, Li and ShuRuan et al., 2018; Reference Zhang, McBride-Chang and TongZhang et al., 2012; Reference Zhang, Lin, Wei, Anderson, Chen, Wang and YangZhang et al., 2014). Orthographic knowledge, especially knowledge of the semantic radical, is related to reading comprehension in Chinese due to its contribution to character meaning (Reference Ho, Ng and NgHo, Ng, & Ng, 2003; Reference Yeung, Ho and ChikYeung et al., 2011). In addition to metalinguistic skills, cognitive skills such as working memory have been found to predict reading comprehension among Chinese children (Reference Jing and LuJing & Lu, 2009; Reference Lu and ZhangLu & Zhang, 2007). Reference Jing and LuJing and Lu (2009), for example, observed that higher working-memory capacity was associated with better Chinese reading comprehension.

10.2.1 History of Aboriginal Culture and Languages

The first European visitors to the Australian continent were Dutch, and the European name for the continent from about 1620 until well into the 1800s was New Holland. One of the earliest British maps of New Holland was made in 1790, two years after the British settlement at Port Jackson. The map shows details of the geographical features and European settlements of the coastal and an empty interior. The British administration was aware of the significant Aboriginal presence on the continent, but nonetheless deemed it to be “terra nullius” – “nobody’s land” – a legal mechanism that afforded lawful annexation and settlement, and that declared, effectively, that whatever it was that the new settlers were doing, they were not invading (Reference Goot and RowseGoot & Rowse, 2007).Footnote ¹

Evidence from archaeology and genetics (e.g., Reference Clarkson, Smith and MarwickClarkson, Smith Marwick et al., 2015; Reference Malaspinas, Westaway and MullerMalaspinas, Westaway, Muller et al., 2016) indicates that, beginning about 60,000 years before the visitors arrived from Europe, the first Australians crossed from the Sunda Peninsula, later to become the Indonesian archipelago, into Sahul, a single continent subsuming what are now mainland Australia, Tasmania, and Papua New Guinea. Genetic research suggests that these first Australians traveled around the coast in both directions with surprising rapidity, filled the interior, and then largely remained in situ; the gene-flow picture of Aboriginal Australia is one of remarkable stability (Reference Malaspinas, Westaway and MullerMalaspinas, Westaway, Muller et al., 2016).

The tentative coastal outlines of the earliest maps made by Europeans provide the initial impressions of outsiders peering into a continent from the edge. For the most part the travelers and settlers who followed them over the next century inherited the restricted gaze of the newcomer, literally unable to see beyond their coastal settlements, and metaphorically unable to see the locals’ unfamiliar social organizations, or their sophisticated systems of land tenure, religious belief, and cultural practice. As for talk and inscription, effectively, what was heard across the land in 1788, and for a long time afterwards, was lingua nullius, nobody’s voice. Two centuries after settlement we now understand that, throughout Australia, Aboriginal people had developed thriving cultures, and filled the continent with at least 250 different languages, each as sophisticated as all other human languages. Figure 10.1 shows linguists’ and historians’ best current mapping of the regions and languages of Australia in 1788.

Figure 10.1 Aboriginal languages map, Australian Institute of Aboriginal and Torres Strait Islander Studies, AIATSIS, Reference HortonHorton (1996)

This map attempts to represent the language, social or nation groups of Aboriginal Australia. It shows only the general locations of larger groupings of people, which may include clans, dialects, or individual languages in a group. It used published resources from 1988 to 1994 and is not intended to be exact, nor the boundaries fixed. It is not suitable for native title or other land claims.

David R. Horton (creator), © Aboriginal Studies Press, AIATSIS, 1996. No reproduction without permission. To purchase a print version visit: www.aiatsis.ashop.com.au/.

This map shows the extraordinary linguistic diversity and complexity evident on the continent before the European arrival. For tens of thousands of years, the first Australians needed to be competent in several languages in order to communicate and intermarry with neighboring communities, and to accommodate additional language expectations among their children (Reference ClyneClyne, 2005, Reference Clyne, Jupp and Clyne2011; Reference Clyne, Jupp and ClyneJupp & Clyne, 2011). Recent research suggests that the stability of populations, the development of rich narrative cultures, and culture-specific mechanisms of intergenerational transmission, have combined to make Australian Aboriginal people the custodians of the world’s oldest orally transmitted memories, recording, for instance, late-Holocene inundations of the continental shelf that occurred about 10,000 years ago (Reference Nunn and ReidNunn & Reid, 2016), a narrative feat of human memory probably unparalleled on other continents.

In his encyclopedic history of writing, Fischer defined it as “the sequencing of standardized symbols (characters, signs or sign components) in order to graphically reproduce human speech, thought and other things in part or whole” (Reference FischerFischer, 2001, p. 12). For over 50,000 years Aboriginal Australians have produced inscriptions that depicted land features, origin myths, and ideas about land, ancestors, social organization, customs, and moral responsibilities. While not representing speech phonemically, they used painting and sculpture on rocks, wood, and leaves, employing symbols that retained similar meanings over long periods of time within and across regions (e.g., Reference TaylorTaylor, 1996). Depictions of animals that were extinct almost 40,000 years ago have been discovered, as have paintings of the arrival of ships, presumably from Europe less than a few hundred years ago (Reference Kleinert and NealeKleinert & Neale, 2000).

These rich precolonial linguistic assets of Australia, however, were quickly lost. Estimates vary, but the Australian Aboriginal population prior to the arrival of the British settlers seems to have been between 500,000 and 750,000. Since then, as Reference BurnleyBurnley (1988) pointed out,

Most estimates are that the Aboriginal population on the continent returned to precolonial levels only after about 220 years. But by 2018, of the 250 languages of 1788, more than 100 had become extinct, about 130 were critically endangered (spoken only by the oldest generations and no longer learned by children), and only 13 were currently being learned as first languages by children (AIATSIS, 2019). Those thirteen languages are often classed, in Australia, as “healthy,” using “intergenerational language transmission” as a diagnostic of “health”; but crucially, all thirteen, at the time of writing, fail to meet most of the other key diagnostic factors employed in UNESCO’s Language Vitality and Endangerment report (UNESCO, 2003) for determining language vitality – in particular, “absolute number of speakers” and “proportion of speakers existing within the total (global) population.” They can, therefore, be accurately classified as “critically endangered.”

Researchers have documented grammars, dictionaries, and collections of texts in most of the languages still spoken (Australian Research Council Centre of Excellence on the Dynamics of Languages, e.g., Reference Simpson, McConvell, Thieberger, Arthur and MorphySimpson, McConvell, & Thieberger, 2019; Reference SimpsonSimpson, 2019; Reference Eades, Koch and NordlingerEades, 2014). A rich understanding has emerged of historical relationships within the “Australian language family” (Reference Bowern and KochBowern & Koch, 2004; Reference SimpsonSimpson, 2019). Theoretical linguists around the world have long mined the richness and complexity of Aboriginal languages in pursuing answers to theory-driven questions about the nature and limits of human language in general (e.g., Reference Evans and SasseEvans & Sasse, 2002; Reference Koch and NordlingerKoch & Nordlinger, 2014).

For much of the last two centuries Aboriginal people and their languages were largely ignored. White Australian nation building advanced on the assumption that Aboriginality would be absorbed and assimilated into the new Anglo-Australia. Instead, we now find a vibrant and growing Aboriginal population across Australia, ranging from metropolitan, urban living people to those living in rural and remote locations, speaking one or more of the following languages and/or language varieties:

• Traditional Aboriginal languages, of which about 130 are still spoken (Reference Purdie, Frigo and OzolinsPurdie et al., 2008; Reference SimpsonSimpson, 2019);

• Standard Australian English, a descriptor used for a variety of English spoken by most Australians (Reference Collins, Peters, Kortmann and SchneiderCollins & Peters, 2004);

• Aboriginal English, a nonstandard variety of English, usefully conceived of as a continuum ranging from light acrolectal varieties (closer to standard), with only subtle differences from Standard Australian English, to heavy basilectal varieties that are closer to creoles (Reference Kaldor, Malcolm and RomaineKaldor & Malcolm, 1991; Reference Eades and RomaineEades, 1991);

• One of two creole languages, Kriol and Broken, largely English-lexified, but with substrate influence from Aboriginal languages (Reference Harris and SandefurHarris & Sandefur, 1984).

10.2.2 Aboriginal Literacy and Schooling

Despite the obvious cultural importance of Aboriginal languages to the identity of modern Australia, there has generally been little provision or sustained political support for the learning of Aboriginal languages by non-Aboriginal Australians (Reference Oldfield, Lo Bianco, Rennie and HarperOldfield & Lo Bianco, 2019; Reference SimpsonSimpson, 2019). As we outline below, additional language learning in Australian schools, for most of the last century, has remained mostly focused on French, German, Italian, Japanese, Indonesian, and Chinese. Only since 2006 has there been an effort to teach Aboriginal languages in public schools. As a result, in 60 target schools in New South Wales, for instance, about 5,000 primary and secondary students, of whom two thirds are non-Aboriginal, were learning an Aboriginal language in 2008 (Reference Purdie, Frigo and OzolinsPurdie et al., 2008), with recent increases to around 7,000 in 2018.

In 2019, of the top fifteen non-English languages learned in New South Wales schools, three were Aboriginal languages (NSW Department of Education, 2019). At the national level, the Australian curriculum includes language-specific curricula for so-called “world” languages, and a Framework for Aboriginal Languages and Torres Strait Islander Languages (ACARA, 2016a, b; Reference DisbrayDisbray, 2019; Reference Simpson, Disbray and O’ShannessySimpson, Disbray, and O’Shannessy, 2019; and see First Languages Australia, 2020), which aimed to increase students’ learning of Aboriginal and Torres Strait Islander languages. By 2008 there were 16,000 Aboriginal students and 13,000 non-Aboriginal students in 260 Australian schools participating in Aboriginal language and literacy programs (Reference Purdie, Frigo and OzolinsPurdie et al., 2008). These programs ranged from bilingual programs to second-language-learning programs to revival and language-awareness programs. Here we focus on the history of bilingual education and the use of creoles and Aboriginal English in schools, before turning to literacy outcomes.

There were sporadic attempts through the 1800s, often by missionaries, to teach Aboriginal children through the use of their own languages, notably Kaurna, Diyeri, and Arrernte (Reference Gale, Hartman and HendersonGale, 1994). Up until the 1970s, however, most Aboriginal children were educated in English-only schools, and faced the task of acquiring literacy skills in a language other than their mother tongue. The 1970s saw the introduction of bilingual education in about twenty communities, mostly in the Northern Territory, and typically in public schools (Reference DevlinDevlin, 2009; Reference Gale, Hartman and HendersonGale, 1994). These were transitional programs, intended to introduce literacy and numeracy using the child’s first language and then to transfer those skills to English, via graduated increases, across each school year. All of these programs explicitly nominated English literacy, rather than traditional language maintenance, as their immediate goal (Reference Gale, Hartman and HendersonGale, 1994). They exemplify what Reference Baker and WrightBaker and Wright (2017, Reference Reid and SenguptaReid, 2010) described as the “subtractive” model of bilingualism, in which the first language gradually disappears.

Bilingual schooling of Aboriginal children in Australia has a troubled history. Funded by the Australian government but administered by the Northern Territory’s Education Department, the scheme’s support from governments has vacillated. As a scheme affecting only a relatively small number of students, its rationale and details have often been poorly articulated and understood, including by many of those who have made decisions about its direction and funding. Administrators of bilingual education have often failed to understand that learning an Aboriginal language, like learning any additional language, complements rather than hinders developing proficiency in English (Reference Simpson, Caffery and McConvellSimpson, Caffery, & McConvell, 2009). Critically, support for bilingual education never fully embraced the principle that children learning English as an additional language (EAL) need to be taught by EAL-trained teachers (Reference NichollsNicholls, 2005), preferably bilingual teachers who understand the processes entailed in additional language learning (Reference EllisEllis, 2016). Nationally standardized English literacy assessments for children in bilingual schools have failed to assess students’ first-language literacy. Further, evidence that these students were performing at levels lower than nonbilingual schoolchildren in Standard Australian English literacy testing has regularly been used to discredit the scheme by those seeking to reduce its funding (Reference Simpson, Caffery and McConvellSimpson et al., 2009).

Bilingual schools, mostly in remote locations, have struggled to find enough appropriately trained first-language-speaking teachers, have been frequently staffed by non-Aboriginal teachers whose professional training did not prepare them specifically for bilingual classroom practices, and have experienced levels of teacher turnover far higher than mainstream schools (Reference Simpson, Caffery and McConvellSimpson et al., 2009). Despite high levels of multilingualism and oral literacy, the parents of most children in bilingual schools have low levels of written literacy themselves, and, for many children in these remote communities, books and writing play no role in their lives outside of school. Since 2008 the Northern Territory Education Department has mitigated its involvement in a bilingual approach by reintroducing a “Talk English” focus and calling on bilingual schools to focus on English literacy for the first four hours of each school day. Some bilingual programs have complied while others have ignored the department’s directive and insisted on maintaining a “two-way education” balance. In these latter cases, the general aim has been to value both languages equally (Reference DevlinDevlin, 2009; Reference Sellwood and AngeloSellwood & Angelo, 2013). There are promising aspects of this story, with some individual bilingual schools achieving excellent results in the national English literacy tests, the only measure for literacy outcomes available. But bilingual and biliterate education for Aboriginal Australia has a history of policy somersaults and fluctuating levels of funding, support, and commitment.

Most Aboriginal children speak Standard Australian English, Aboriginal English, or a creole variety, whereas the bilingual programs described above cater to only the minority of Aboriginal children who speak an Aboriginal language as their first language. There is clear evidence that Aboriginal English and the two creole varieties, Kriol and Broken, while predominantly English-lexified, are distinct from Standard Australian English. The degree of distinctiveness is greater in the case of the creoles, which in their “heavier” forms are not mutually intelligible. At the acrolectal end of the Aboriginal English spectrum, the lexical, grammatical, and phonological differences from Standard Australian English are comparatively minor (Reference Kaldor, Malcolm and RomaineKaldor & Malcolm, 1991). But a range of more subtle pragmatic differences, ranging from forms of knowledge organization (Reference SharifianSharifian, 2005) to culturally significant ways of using silence and eye contact and to methods of eliciting information through questioning techniques (Reference Eades and RomaineEades, 1991) still create difficulties in communication between Standard Australian English and Aboriginal English speakers.

Apart from a single Standard Australian English–Kriol bilingual program that ran at one school in the 1980s, there has been little use of either creoles or Aboriginal English as mediums of education in Australia. Due to the apparent similarities of these languages with Standard Australian English, speakers of the creoles and Aboriginal English are typically not identified as learning Standard Australian English as an additional language (Reference Wigglesworth, Billington and LoakesWigglesworth, Billington, & Loakes, 2013). A history of attitudes that characterize Aboriginal English as “bad English,” held by both teachers and speakers of Aboriginal English, has further functioned to make Aboriginal English unacceptable as a language of instruction in many community schools. Because Standard Australian English is widely associated with vocational advantage (Reference Disbray and LoakesDisbray & Loakes, 2013), and because there is a continuum of creole-to-Aboriginal English varieties that makes it difficult to choose one language variety over another for schooling purposes (Reference SiegelSiegel, 1999), Standard Australian English has remained unchallenged as the language of both instruction and assessment in Australian schools.

10.2.3 Aboriginal Literacy Outcomes

Aboriginal children in remote communities are outperformed on standardized literacy tests across all English literacy skills by non-Aboriginal children in metropolitan centers (ACARA, 2017). According to the Closing the Gap report (Australian Department of Prime Minister and Cabinet, 2015), school attendance rates have been as low as 14 percent in some remote schools, and in 2014 only 34 percent of Aboriginal students in remote communities met the national minimum standard for Year 7 reading. Similar findings come from the 2017 NAPLAN results: 25 percent of students in remote parts of Australia are achieving at or above minimum standards for reading in Year 3, increasing only to 27 percent in Year 9 (ACARA, 2017). In the “very remote” geographical location category, only 15 percent are at minimum standard by Year 9; and over 80 percent of Aboriginal students across the Northern Territory achieved below the minimum standard for Year 3 reading, spelling, grammar and punctuation, results that are lower than preceding years (ACARA, 2017).

The Programme for International Student Assessment (PISA 2012) report showed a 2.5–3-year gap between non-Aboriginal and Aboriginal literacy rates in Australia (Reference Thomson, De Bortoli and BuckleyThomson, De Bortoli, & Buckley, 2012). Similarly, the Progress in International Reading Literacy Study (PIRLS 2016), drawing on data from fifty countries, points to significant literacy gaps between Indigenous and non-Indigenous Australians: “Australia’s Aboriginal and Torres Strait Islander students attained an average score of 483 points, which was 67 points lower than the average score for non-Indigenous students … In 2016 … students with an Aboriginal or Torres Strait Islander background had almost four times the odds of being a poor reader compared to non-Indigenous students” (Reference Thomson, Hillman, Schmid, Rodrigues and FullartonThomson et al., 2017). Findings such as these have been used by successive governments to criticize bilingual education, and to justify the shift of resources to developing English-only literacy skills. Australian linguists and literacy specialists have noted that the real issue is “the cost and complexity of lifting both performance standards and attendance rates across remote schools. Bilingual education is a decoy” (Reference DevlinDevlin, 2010, p. 3). The recent removal of bilingual programs in many areas has not yielded any increase in English literacy levels.

Those schoolchildren speaking Indigenous forms of English and creoles have also long been recognized to be disadvantaged (Reference Hudson and Taylor-HenleyHudson & Taylor-Henley, 1995), and their failure to achieve expected outcomes, especially in literacy, is well documented (OECD, 2011; Reference SharifianSharifian, 2001; Reference Wigglesworth, Billington and LoakesWigglesworth et al., 2013). Prominent among the drivers of disadvantage are structural differences between Aboriginal English and Standard Australian English, such as:

• gender undifferentiated third singular pronoun “ee”;

• hypercorrective initial “h” as in “huncle Henry”; and

• zero copula usages like “ee gone.”

Such grammatical differences are routinely interpreted by speakers of Standard Australian English as “errors,” and can result in teachers overlooking demonstrated content knowledge (Reference DixonDixon, 2013; Reference Wigglesworth, Simpson and LoakesWigglesworth, Simpson, & Loakes, 2011).

A meaningful response to the challenges and opportunities involved in engaging Aboriginal Australian school students will emerge when the current invisibility of Aboriginal English and creoles in Australian education systems is recognized. Reference MurtaghMurtagh (1982) assessed the single Kriol bilingual program and found that explicit teaching of, and in, both languages resulted in increased oral proficiency in both English and Kriol (Reference MurtaghMurtagh, 1982, cited in Reference SiegelSiegel, 1999, and see the collection of studies in Reference Rennie and HarperRennie & Harper, 2019). Similarly, various researchers have pointed out that the directions that will help the Australian schooling system address the disadvantage that Aboriginal students currently experience include: teaching that explicitly addresses the differences between Standard Australian English and the students’ own variety, be it a creole, Aboriginal English, or a traditional Aboriginal language; teaching that understands and reinforces the contextual appropriateness of each (e.g., Reference Shinkfield, Jennings, Rennie, Harper, Bianco and WileyShinkfield & Jennings, 2019); and teaching that supports students in learning Standard Australian English as an additional variety (e.g., Reference Sellwood and AngeloSellwood & Angelo, 2013; Reference SiegelSiegel, 1999; Reference Wigglesworth, Billington and LoakesWigglesworth et al., 2013).

To summarize, the subtractive focus of bilingual education, the overlooking of creoles and Aboriginal English as varieties to be used in school, and the unfounded suspicion that the teaching of traditional languages, creoles, and Aboriginal English is incompatible with English literacy acquisition, have together contributed to a large gap between the literacy outcomes of Aboriginal and non-Aboriginal Australians. An acknowledgment of the language and literacy resources and needs of Australia as a country calls for a much more nuanced reaction to the realities of European–Aboriginal history, rather than the apathy, tokenism, and capricious educational and social policies that have characterized much of Anglo-Australia’s reaction for generations.

10.3.1 Historical Perspectives

Analogous patterns in practice and policy arise in any consideration of community and migrant education in literacy. British colonials arriving in Australia promoted the norm of English monolingualism and monoliteracy as both a gift and an instrument of empire (Reference Clyne, Jupp and ClyneClyne, 2011; Reference EllisEllis, 2008). The expectation that English would be the national language was such that the need has never been felt to declare an official language for Australia: English is the language of politics, law, education, and of everyday social life more broadly, and so the de facto national language.

The history of Australian migrant settlement, however, tells a different story. Apart from the original hundreds of Australian Aboriginal languages described above, significant numbers of the earliest migrant settlers spoke Irish, Gaelic, and Welsh as first languages, and during the nineteenth century settlers came from across Europe and Asia, attracted by discoveries of gold and readily available land. There were many urban and rural communities that used spoken and written languages other than English (Reference Clyne, Jupp and ClyneClyne, 2011). Multilingual and bilingual educational activities were more prevalent in the nineteenth century than at any other time in Australia’s history. Many migrant communities used their languages for everyday and business purposes, and to teach their children in schools (Reference Lo Bianco and SlaughterLo Bianco & Slaughter, 2009).

During the seven decades following the First World War, a period in which the population of Australia grew from about 5 to 14 million, the use of languages other than English was considered undesirable, and their active usage decreased. The dominant political and social paradigm was “integration,” a project to which the learning of English was seen as central. Over this period waves of large-scale immigration to Australia from nations across Europe, the Middle East, and Asia brought a wide range of languages and cultures, but these migrants were actively encouraged to learn English, and discouraged from using their own languages, even in their homes (Reference Clyne, Jupp and ClyneClyne, 2011; Reference Lo Bianco and SlaughterLo Bianco & Slaughter, 2009).

The large influx of migrants to Australia that followed the Second World War led to efforts to link migration with English literacy capability, and thereby to Australia’s most consequential language policy instruments, the Adult Migration Education Program and the subsequent 1969 Child Migrant Education Act. Via these, a national policy of teaching English to all children was systematically enacted (Reference Hajek and SlaughterHajek & Slaughter, 2014). Additional language education practices “favoured choices and methods of instruction dictated by attachment to the western canon of literary prestige, principally for reading and cultivation rather than active use” (Reference Lo Bianco and SlaughterLo Bianco & Slaughter, 2009, p. 15). So French and German, as well as Latin and Classical Greek, as languages of literature and academic prestige, were popular in elite schools and universities. There was little effort to promote or retain the use of active community languages (e.g., Modern Greek, Italian, German, Chinese, Polish, Turkish, Lithuanian), despite the substantial communities of migrants for whom these were first languages.

Since the late 1970s overseas military and political conflicts have led to a surge in immigration from southern Asian nations (e.g., Vietnam, Cambodia, East Timor, and China), and stimulated the emergence of a more pluralistic cultural policy. Australians developed an image of their nation as a “multicultural” nation, but nonetheless monolingual (Reference Clyne, Jupp and ClyneClyne, 2011), celebrating this image in the popular media and in policy (Reference Hajek and SlaughterHajek & Slaughter, 2014). It is a self-representation that has persisted to the present and is frequently announced to the global community as an example of the peaceful nature of such multicultural identity building. It ignores, however, the multilingual or plurilingual possibilities that arise from such diversity.

By the late 1980s increasing support for learning the languages of neighbors and new trade partners, the balance of which had shifted to Asia from Britain following Britain’s accession to the Common European Market, was beginning to take hold in schools and universities (Reference Hajek and SlaughterHajek & Slaughter, 2014; Reference Lo Bianco and SlaughterLo Bianco & Slaughter, 2009; Reference Morgan, Kohler and HarbonMorgan, Kohler, & Harbon, 2011). The Federal Government began to implement policies promoting and supporting Asian language studies. Japanese surpassed French as the most studied additional language, and Chinese, Indonesian, Japanese, and Korean were promoted as priority languages (Reference HarringtonHarrington, 2012).

The study of “community” languages (Reference ClyneClyne, 2005) was pushed out of mainstream schools into “community” or “ethnic” schools operating outside of school hours. Community groups tried to maintain heritage languages and cultures by running these schools. Despite pressures to reduce the use of heritage languages, by 2018 there were about eighty languages offered by the Ethnic Schools Association for students from Foundation to Year 10 and over forty languages at Years 11 and 12. These were offered across all states and territories, with almost 120,000 students participating in over 1,400 venues (Australian Federation of Ethnic Schools, 2016; Community Languages Australia, 2019; Migration Heritage Centre, NSW, Powerhouse Museum, n.d.; Reference Scrimgeour, Morgan, Cruikshank and HajekScrimgeour et al., 2019). Such practices remained politically and professionally marginalized, operating principally through “Saturday schools,” and attracting little political or media recognition. In 2019 an effort was made to formalize these community provisions and to begin a process of educationally credentialing the community members doing the teaching work, via the creation of the Sydney Institute for Community Language Education (SICLE, n. d.). The Institute’s initiatives include introducing community languages into mainstream primary schools, with programs in nine community languages – Arabic, Assyrian, Chinese, Greek, Hindi, Korean, Macedonian, Tamil, and Turkish – and the development of full syllabuses for New South Wales schools in Hindi, Macedonian, Persian/Dari, Punjabi, and Tamil. Over 400 languages are spoken in homes and communities across Australia (Australian Bureau of Statistics (ABS), 2016a; Reference Scrimgeour, Morgan, Cruikshank and HajekScrimgeour et al., 2019; Reference Slaughter, Hajek, Hajek and SlaughterSlaughter & Hajek, 2014). No current data exist on levels of written literacy for these languages, however, and children commencing school who are fluent in spoken community languages vary substantially in how well they read and write in those languages.

There has been a gradual acknowledgment of multilingualism and its value for literacy generally (see Reference Wilks-SmithWilks-Smith, 2017), and multicultural policies to promote cultural diversity exist at all levels of government. The prevalent message of the media, and of populist political comment in relation to literacy capability, however, remains firmly focused on English. The view is that English literacy, assessed through standardized testing programs, remains key to Australia’s economic success, and should therefore be the educational imperative. In some jurisdictions it occupies, by policy mandate, the prime teaching time in elementary schools, in so-called “literacy blocks” each morning. This is at the expense of learning areas such as additional languages, or other subjects taught through additional languages.

Australia’s history indicates that it is one of the world’s major immigrant nations, with over 7.5million settlers since 1945 in a national population of about 26 million (Reference Phillips and Simon-DaviesPhillips & Simon-Davies, 2016). The estimated overseas-born population in 2016 was 28.6 percent, with a further 21 percent having at least one overseas-born parent (in total, half the population), a rate considered high compared to those of other OECD nations (ABS, 2016a; Reference Phillips and Simon-DaviesPhillips & Simon-Davies, 2016). It is therefore a prime site for historical analyses of immigrant processes, and for educational research and development relating to the teaching and learning of language and literacy.

Historically, most migrants have come to Australia from Europe, but there are increasing numbers from Asia. Of the 189,700 migrants arriving in 2016–2017, around 21 percent were from India, 15 percent from China, 6.3 percent from the Philippines, 3.5 percent from Pakistan, and 3 percent each from Vietnam and Nepal. China and India, both at about 8 percent, now rank highest as Australians’ countries of birth after the UK and New Zealand (ABS, 2016b). Many of these immigrants are first-generation Chinese and Indians. Table 10.2 summarizes Australian ancestry data, showing high rates of first-generation Chinese (74.3 percent) and Indian (79.8 percent) citizens, who bring their first languages with them and who also have large existing communities of speakers.

Humanitarian migration for refugees and others who cannot return to their country of origin currently accounts for an average of around 14,000 people a year (out of 60,000–70,000 annual applications). These come from a range of countries, with an additional 12,000 from the ISIS conflict for the year 2015–2016 (Australian Government, Department of Immigration and Border Protection, 2016a, 2016b). While some of the 14,000 or so in the standard program were currently, controversially, housed offshore, their needs remained the responsibility of the Australian government. These needs entail the maintenance of education and health, and communication in their own languages. While these refugee numbers are low in terms of the overall Australian population, their linguistic and cultural diversity contributes significantly to the complexity of the educational challenges facing Australian language and literacy educators.

10.3.2 Mainstream Literacy, Multilingualism, and Schooling

Recent research from a variety of sources – magnetic resonance imaging, psychological and cognitive testing, literacy testing at national and international levels, and across a range of curriculum areas – together indicates that bilingual or multilingual capacity is a significant aid to improved literacy outcomes (see also Nakamura & Holla, Chapter 8 and Schwartz, Chapter 19 in this volume). As well as establishing the benefits in personal, cognitive, academic, and social domains, many studies have identified language use and general literacy benefits from bilingual education (e.g., ACARA, 2014; Reference AlbanAlban, 2016; Reference Bak, Nissan, Allerhand and DearyBak et al., 2014; Reference Baker and WrightBaker & Wright, 2017; Reference BialystokBialystok, 2014; Reference BialystokBialystok, 2016; Reference Bialystok, Poarch, Luo and CraikBialystok et al., 2014; Reference Clyne, Hunt and IsaakidisClyne, Hunt, & Isaakidis, 2004; Reference DelistratiDelistrati, 2014; Reference PinterPinter, 2015; Reference Rubio-Fernández and GlucksbergRubio-Fernández & Glucksberg, 2012). These benefits include increased communicative capacity, including alternate expression in both (or all) languages; increased engagement in moving across languages, and agility in lexical choice; increased literacy skills in first and additional languages; equivalent or better performance in standardized literacy and numeracy tests; enhanced understanding of symbolic representation of print; and enhanced sense of identity and self-esteem (Reference Morgan, Scrimgeour, Farmer, Dodd and SaundersMorgan et al., 2016).

Despite these identified benefits, policymakers’ remedy of choice for any apparent shortcomings in literacy standards has generally been to increase time allocations to English literacy instruction. This view compounds the challenges presented by such a linguistically and culturally diverse nation, where learners need to engage with a globalized, multilingual and “super diverse” world (Reference VertovecVertovec, 2010).

The challenges for educators, including literacy researchers, are to increase awareness of the benefits of a bilingual or multilingual approach to literacy in mainstream schooling, to address the needs of an increasingly diverse population, and to give recognition to Aboriginal and migrant populations and their language and literacy needs that go beyond English. As we have seen, assessment results for Aboriginal students in English literacy testing remain low and are compounded by social history and lack of access to learning in their first languages in the critical first few years of schooling. This situation, however, is not reflected in migrant communities: Results from the national assessment program, NAPLAN, regularly show a consistent advantage for students from non-English-speaking backgrounds (termed ‘Language background other than English’ [LBOTE] in these tests). These students or their parents are born overseas. Table 10.3 summarizes the 2019 results.

The findings from the standardized national assessment are consistent and unequivocal: Students with an additional language, who are likely to speak that language at home, rather than “underperforming” in fact surpass (or, in one case match) non-LBOTE on all of these literacy measures. Left out of this program are assessments of the substantial cognitive, social, and cultural advantages, as we have seen, that LBOTE students enjoy.

10.4.1 International Test Results

Along with national processes for assessing literacy among school students, Australia has taken part in OECD’s international literacy assessments programs for both high-school students (PISA) and adults (PIAAC, the Programme for the International Assessment of Adult Competencies). The results of these are widely reported in the media and are often brought to bear in debates about teaching, policy, and research. The rankings of the OECD nations scoring significantly above the OECD mean for reading in PISA, round 2018, and for PIAAC, round 2012, along with some related national characteristics, are summarized in Table 10.4.

Comparing the Australian profile, we see Australia rating at about the midpoints on some measures (relationship between literacy score and socioeconomic status, and expenditure per student as percent of GDP spent on schooling), and above average on other measures (overall literacy, percentage of adults with tertiary education, and inequality-adjusted education index). In its PISA 2018 “Country Note: Australia,” the OECD highlighted these issues:

The rightmost columns of Table 10.4 refer to the literacy rankings of Australian adults on the tests administered as part of OECD’s PIAAC program. Clearly, Australian adults rank highly on the PIAAC literacy test. In 2015 Australia’s national government announced the implementation of a program of literacy testing for initial teacher-education students, as a reaction to the declining performance of Australian school students in the immediately preceding PISA tests. The government claimed that students’ unsatisfactory performance was partly explained by inadequate literacy levels among Australian schoolteachers. Reference Freebody and FreebodyFreebody and Freebody (2017) examined this claim using OECD’s PIAAC literacy test, selecting the literacy scores of participants, from Australia and the twenty-three other participating OECD countries, who nominated “teaching” as their profession, or “teacher education” as their highest qualification. They also compared Australian teachers’ scores with the scores of those of Australian participants from five comparable professional groups. They drew two main conclusions: “first, that Australian teachers, as determined by their highest qualification or their current professional status, are in the highest performing group of countries on OECD’s literacy tests, and second, that Australian teachers have literacy test scores statistically comparable to other professional practitioners in Australia” (2017, p. 5). It is informative to examine OECD’s own interpretation of Australian students’ performance on PISA and on the national assessment program, NAPLAN. The OECD made these six recommendations concerning the teaching and assessment of literacy in Australia (OECD Country Report Australia, OECD 2011, pp. 152ff):

• establish national strategies for strengthening the linkages to classroom practice within the overall evaluation and assessment framework;

• promote greater national consistency while giving room for local diversity;

• reinforce the assessment validity of the national testing program and establish safeguards against an overemphasis on it;

• strengthen teachers’ capacity to assess student performance against the Australian curriculum and use student assessment data;

• maintain the centrality of teacher-based assessment while ensuring the diversity of assessment formats;

• clearly establish what the country takes to be the fundamental purpose of external school evaluation – either it can “bring about general improvement across all schools or, more narrowly, it can focus on ‘failing schools’” (p. 156).

Remarkable here is the notice taken of this set of recommendations by Australian governments responsible for literacy education over the intervening years: These authorities have, by acts of both omission and commission, continued to move in precisely the opposite direction on all six counts.Perhaps the most striking aspect of the findings from PISA and PIAAC is how deeply Australian literacy educators – classroom practitioners, school leaders, teacher educators, researchers, and policymakers – are divided on the validity of these metrics and comparisons, and on their usefulness in improving the quality of literacy efforts (e.g., Reference Freebody, Lian, Kell, Black and LieFreebody, 2017, Reference Freebody, Cox, Feez and Beveridge2019; Reference Lingard, Thompson and SellarLingard, Thompson, & Sellars, 2016). The degree of attention paid by policymakers, politicians, and the media to literacy has resulted in the spectacle of recurring crises and “literacy wars” in Australia (Reference Castles, Rastle and NationCastles, Rastle, & Nation, 2018; Reference Lo Bianco and FreebodyLo Bianco & Freebody, 2001; Reference SnyderSnyder, 2008). But there have also been collaborations, summarized in Subsection 10.4.2, that have resulted in attempts to “balance” various approaches in light of the varying needs of Australian students (e.g., Reference Cox, Feez and BeveridgeCox, Feez, & Beveridge, 2019; Reference Wyatt-Smith and GunnWyatt-Smith & Gunn, 2007).

10.4.2 A History of Opportunities, Challenges, and Dynamism

The history of Australia’s institutional efforts in education has presented distinctive challenges and opportunities to literacy educators. The inconsistent management of Aboriginal and migrant education figures prominently. More recent, abrupt institutional changes in education over the period of the late 1980s and early 1990s were also influential: The institutes responsible for teacher education in Australia were amalgamated into universities, or, in some cases, combined to form aggregations that were renamed as universities. These amalgamations meant that academics in faculties or schools of educational studies in most existing universities were joined by significant numbers of colleagues from teacher-education institutions. In this relatively sudden way, different kinds of professional expertise were conjoined under the one set of institutional expectations and incentives. While teacher education and research had gone on in both university schools and colleges, the institutional emphases had been generally different; so the amalgamations called for new kinds of - and new priorities for – undergraduate teaching for some staff, and research activity and supervision for others.

These new configurations, mixing school-based “guild” experience and disciplinary-based approaches to teaching and research, brought to the surface differences around basic questions such as the nature of good teaching, how best to prepare teachers for the practical work of schools (Reference ConnellConnell, 2009), and the research–practice relationship. New forms of debate and new opportunities for collaboration arose. Literacy education was center stage in many of these debates partly because long-standing research programs in reading and writing had developed in some university schools of psychology, linguistics, and English, as well as in some schools of education.

These amalgamations also led to ownership skirmishes across disciplinary boundaries. At worst, discipline- or site-based groups have tended to become aware of conceptual, practical, or policy developments in their professional neighborhoods. At best, more productive cross-discipline and cross-method approaches to literacy education have occurred than in some other comparably sized research settings (e.g., in the psychology of reading, Reference Byrne, Freebody and GatesByrne, Freebody, & Gates, 1992; in reviews of literacy education, e.g., Australian Department of Employment, Education and Training, 1991, and in extended research collaborations such as Reference Morgan, Comber, Freebody and NixonMorgan et al., 2014).

The competitive dynamism that has developed within and across the various scholarly disciplines that address literacy education is evident in the volume and range of published Australian empirical and theoretical work. A well-developed line focuses on the cracking of graphic codes as they map onto speech and the educational implications of those processes (e.g., phonics teaching, alphabetic knowledge, phonemic awareness). This area has been developed largely by educational psychologists (e.g., Reference Buckingham, Beaman and WheldallBuckingham, Beaman, & Wheldall, 2014; Reference ByrneByrne, 1998, Reference Byrne, Bates and Maack2010; Reference Castles, Rastle and NationCastles et al., 2018; Reference Coltheart, Snowling and HulmeColtheart, 2005; Commonwealth of Australia, 2005; Reference Cox, Feez and BeveridgeCox et al., 2019; Reference Fielding-Barnsley and HayFielding-Barnsley & Hay, 2012). There has also been a body of research in theory within the Australian setting based on constructivist or so-called whole language approaches to the teaching of reading and writing, emphasizing Piagetian and other growth and development approaches and arguing against the pervasive use of explicit teaching and segmented, atomized skills development (e.g., Reference CambourneCambourne, 1988; and see the summary in Reference Wyatt-Smith and GunnWyatt-Smith & Gunn, 2007).

A distinctive contribution to literacy education has arisen from the application of systemic functional linguistics (SFL), developed by Reference HallidayHalliday (1985) and his students and colleagues. This contribution includes documentations and analyses of the formal curriculum-based reading and writing demands on students across the school years (most prominently Reference Christie and DerewiankaChristie & Derewianka, 2008), and applications of approaches based on SFL to classroom work (Reference GibbonsGibbons, 2014; Reference GrayGray, 2007; Reference HammondHammond, 2016; Reference MartinMartin, 1985). One application of SFL to classroom teaching and teacher professional development, called Reading to Learn, is summarized in Reference Rose, Miller, McCardle and ConnellyRose (2017). He outlined these key elements of the program: (i) engagement of students in curriculum texts that may be beyond their independent reading levels, (ii) interrogation of passages from texts that models comprehension processes, (iii) recognition of language usages and their related purposes, (iv) recruiting these language resources into students’ writing, and (v) constructing effective, purposeful, and appropriately organized texts.

Reference Rose, Miller, McCardle and ConnellyRose (2017) summarized several research application phases demonstrating the efficacy of this program over five grade-level bands from kindergarten to Grade 8. Accuracy levels on five features of students’ writing were assessed: overall generic structure, appropriate use of register, discourse, grammar, and graphic features. Composite assessment over these phases showed not only major gains at all grade levels after six months of the program’s use, but also a reduction of the gaps between previously designated literacy achievement bands within each grade level (see MacRae et al., 2000; Rose & Martin, 2013). Systemic functional linguistics has also been applied to questions about how multiple modalities and media can be put to work in literacy learning settings (e.g., Reference Macken-Horarik, Love, Sandiford and UnsworthMacken-Horarik et al., 2017; Reference Painter, Martin and UnsworthPainter, Martin, & Unsworth, 2013). Reference Unsworth and Macken-HorarikUnsworth and Macken-Horarik (2015), for example, documented three progression points among school students’ understandings of picture–language relations in multimodal texts across the elementary and secondary years:

• a “tactical orientation”: early-years students describe in a variety of nonrational ways how illustrators choose to represent the language (the tactic is simply finding something apparently relevant to say);

• “diegetic orientation”: images are interpreted in terms of the apparent reality of the story’s world, where illustrators’ choices are tied to the thoughts or feelings of the characters; and

• a “semiotic orientation”: students orient to the explicit crafting of the text, the deliberate construction of the multimodal to shape the responses of readers, and an awareness of the kinds of options the illustrator exercised to represent emotion and power relations.

These researchers concluded that teachers need to work with the learners’ “shifts from more idiosyncratic responses to images, on from those that attend primarily to the experiential world ‘inside’ the text, and toward those that respond to the shaping power of the text itself” (p. 75). Sociological approaches have also been evident in the Australian tradition of literacy research. These have included critical approaches to literacy materials used in schools (Reference Baker and LukeBaker & Luke, 1991; Reference GilbertGilbert, 1989; Reference Green and BeavisGreen & Beavis, 2012; Reference LukeLuke, 1988, Reference Luke2018) as well as approaches based on the rapidly changing social, technological, and vocational conditions in which schools operate, and in which increasingly diverse groups of students need to participate (Reference Cope, Kalantzis and HawkinsCope & Kalantzis, 2013; Reference Farrell, Baynham and PrinslooFarrell, 2009; Reference Kalantzis, Cope, Chan and Dalley-TrimKalantzis et al., 2016; Reference Lankshear, Knobel, Leung and StreetLankshear & Knobel, 2014; Reference Luke, Dooley and WoodsLuke, Dooley, & Woods, 2011).

As a large-scale example, a series of studies in this tradition was conducted by Reference Hill, Comber, Louden, Rivalland and ReidHill et al. (2002, and see Reference Hill, Comber, Louden, Rivalland and ReidHill, 2002). Through a compilation of detailed case studies, they documented the literacy learning experiences of 100 children from a range of cultural and socioeconomic backgrounds, including a remote Aboriginal community. They first followed the children from preschool into school, and followed up on the same children five years later. Among many other findings, some key conclusions were that:

• most children increased their literacy capabilities substantially;

• nonetheless the range of reading and writing achievement within age groups was extensive;

• significantly lower levels of literacy were found overwhelmingly in schools serving poor communities;

• findings overall indicated that, within those patterns, the teaching that has a substantial positive effect on children’s literacy learning, first, does not rely on a narrow set of teaching strategies or prepackaged curriculum content, and second, is dynamic and based on a detailed, “data-driven responsive” focus on individual students’ needs. (Reference Hill, Comber, Louden, Rivalland and ReidHill, 2002, p. 1)

In her 2002 summary, Hill concluded:

Collaborations across these fields have also included multidisciplinary focuses on the interactional details of classroom activity and learning (e.g., Reference Edwards-Groves and GrootenboerEdwards-Groves & Grootenboer, 2015; Reference Martin and MatonMartin & Maton, 2013), on new technologies, critical online analysis, and multimodal literacy activities in the middle years of schooling (e.g., Reference Morgan, Comber, Freebody and NixonMorgan et al., 2014), and on technology and diversity-oriented pedagogies (e.g., Reference Mills, Stornaiuolo, Smith and PandyaMills et al., 2018; New London Group, 1996; SICLE, n. d.). Rapidly growing is a research base aimed at capitalizing on students’ out-of-school activities in digital, online, and mobile contexts. This includes work on how classrooms can be constructed that capitalize on the new literacy affordances of these technologies (e.g., Reference Reimann, Bull, Kickmeier-Rust, Vatrapu and WassonReimann et al., & Wasson, 2016; Reference Wells, Auld, Doecke, Auld and WellsWells & Auld, 2014); how new media can reshape literate communications (e.g., Reference Green, Beavis, Hall, Cremin, Comber and MollGreen & Beavis, 2013); and how literacy learning can be reliably evaluated in online, “big data” settings (e.g., Reference MillsMills, 2019).

So-called C21 lines of work continue to develop along with “literacy education,” conventionally construed. The real conceptual, methodological, and professional advances on offer from an integration of print and digital literacy efforts, especially in Aboriginal, community, remote, and equity settings, are yet to fully emerge in Australian education. These topics present educators and researchers with challenges in learning from literacy education’s past and applying that knowledge to its future, in a unique combination of community, state, and national settings. Finally, how the various curriculum areas put literacy to work in gradually more specialized and divergent ways has assumed increasing prominence in Australian work (e.g., Reference Christie and MatonChristie & Maton, 2011; Reference Martin, Maton and DoranMartin, Maton, & Doran, 2019; Reference Muspratt, Freebody and KhineMuspratt & Freebody, 2013). This may be one outcome of increased collaboration among the various groups of scholars, policymakers, and classroom practitioners, in periods of national and state curricular reform.

The influential Alice Springs (Mparntwe) Declaration (Education Council, 2019), for instance, identifies languages, especially Asian and Aboriginal languages, as key learning, along with English. In the national languages curriculum, a strong case is made for the benefits of students’ literacy improvement in all languages through engagement in learning and using additional languages, as well as for the rights of Australian children to bilingual capabilities (ACARA, 2011, 2020). While programs managed at state and territory levels vary widely, there are positive signs: The increase of bilingual school programs is the biggest growth area in languages programs. This growth is most evident in programs that use a Content and Language Integrated Learning approach, in which specific, high-status curricular areas are taught in a second or additional language (Reference CrossCross, 2015; Reference Morgan, Scrimgeour, Farmer, Dodd and SaundersMorgan et al., 2016). In a recent initiative, the Australian government contracted the Australian Federation of Modern Language Teachers Associations (AFMLTA, lead investigators Morgan & Scrimgeour) to develop a national plan and strategy for languages education. These initiatives indicate emerging levels of understanding of the need for bilingual and multilingual programs to support literacy learning.