Blog

Teaching Reading With Social Justice

The following content was originally published on the Language on the Move research website. bua-lit was given permission to share the content on our blog. Editor’s note: Narrow views of reading as purely a mechanical skill have recently gained renewed traction. Literacy programs that strip away the social and cultural contexts of reading and writing run the danger of offering “poor education for the poor,” as the authors of the response report “Teaching Reading With Social Justice” show. Authored by Professors Judy Kalman (DIE, Cinvestav, México), Laura Eisner (Universidad Nacional de Río Negro, Argentina), Patricia Valdivia (Universidad Autónoma Metropolitana-Lerma, México), Citlalli López Rendón (DIE, Cinvestav, México), and María del Carmen Lorenzatti (Universidad Nacional de Chilecito y Universidad Nacional de Córdoba, Argentina), “Teaching Reading With Social Justice: What the evidence also shows” offers a critical response to the World Bank’s 2025 report “Effective Reading Instruction in Low- And Middle-Income Countries (ERI-LMIC): What the Evidence Shows.” The following constitutes the executive summary of “Teaching Reading With Social Justice: What the evidence also shows.” The full text is available for download in English and en Español. The 2025 report Effective Reading Instruction in Low- and Middle-Income Countries: What the Evidence Shows (ERI-LMIC) validated by the Global Education Evidence Advisory Panel (GEEAP) and commissioned by the Foreign, Commonwealth, and Development Office (FCDO), the World Bank, and UNICEF, proposes a model for teaching reading centered on the Science of Reading (SoR) and the Simple View of Reading (SVR). The document reignites an old debate about reading instruction methods, once again promoting the teaching of letters and sounds in isolation as a universal solution for early literacy. The ERI-LMIC is based on an alarming statistic: before the pandemic, 57% of 10-year-old children in low- and middle-income countries could neither read nor understand a simple story—a condition it terms “reading poverty” (ERI-LMIC p. 13). Based on this assessment, the document proposes interventions grounded in experimental studies and meta-analyses, without considering differences in context, pedagogical traditions, regional contributions, the coexistence of multiple languages, or the perspectives of the learners themselves. The problem we identify is not merely methodological: it is also political and epistemological. The report, produced and disseminated by multilateral organizations with significant international influence, has a strong impact on the educational policies of low- and middle-income countries. Its Simple View of Reading (SVR) is based on what it generically refers to as “the Science of Reading (SoR),” although it draws only on certain strands within this field of research. The proposals are based almost exclusively on evidence from quantitative experimental studies. As a result, they ignore decades of qualitative, anthropological, sociolinguistic, psychogenetic, historical, and sociocultural research that offers a much broader, more comprehensive, and more complex understanding of what it means to learn to read and write in different contexts. Our text responds to the ERI-LMIC with a critical review of its arguments, drawing on a broad body of research on literacy, written culture, and the learning to read and write, to demonstrate the relevance of this knowledge for educational policy, pedagogical guidelines, and teacher training. Our goal is to bring into focus urgent issues that the ERI-LMIC leaves out: the complexity of literate practices, their historical, social, and culturally situated nature, and the inequalities that shape the literacy experiences of large segments of the population. By foregrounding these dimensions, we aim to open a broader and more inclusive discussion that recognizes the diversity of languages, communicative resources, and reading and writing practices across social contexts. Our paper is organized into the following sections: Section 1. A Look at the ERI-LMIC Report. We examine the document’s general approach: its simplistic view of reading, its linear teaching model, its circular use of psychometric evidence, and some historical experiences of implementing the phonics method. Section 2. Rethinking the Arguments of the ERI-LMIC. We analyze the report’s central assumptions in detail: the contrast between natural learning and direct instruction; the relationship between decoding and comprehension; literacy skills and practices; multilingual environments; sensitivity to the learning process; and the individual role and social participation in the classroom. Section 3. Learning to Read and Write: Other Possible Perspectives. We present some alternative perspectives: New Literacy Studies and literacy as a social practice, the sociocultural perspective of learning, pedagogical experiences in Latin America, the role of teachers as mediators, the phases of early reading development, reading roles and dispositions, and situated teaching interventions. Section 4. Final Thoughts: Towards Open Dialogue. We highlight the implications of this analysis for social justice in education and call for greater representation of peripheral regions in decision-making regarding educational policies.

Rethinking the Arguments of the ERI-LMIC Proposal Based on the Science of Reading

Our analysis of the ERI-LMIC proposal examines its main arguments, premises, and sources. This review includes contributions and observations recently published by renowned scholars working in other research areas related to the Science of Reading that were not addressed in the report. The ERI-LMIC draws on several SoR principles and largely adopts the Simple View of Reading (SVR). It defines two main components of reading: decoding and comprehension. From this perspective, learning to decode is equivalent to learning to read, and comprehension follows automatically once decoding is in place. This stance disconnects decoding from meaning, promotes the training of isolated skills (phonological awareness, fluency, letter-sound recognition), and establishes a linear, fixed sequence that ignores the complexity of the reading process. Even within neuroscience and cognitive psychology—disciplines the ERI-LMIC draws upon—their reduced model is contested. Researchers such as Share (2011), Seidenberg et al.  (2020), Tierney and Pearson (2024), and Neuman (2023) argue that reading does not follow a linear progression, that context and meaning are central to decoding, and that there is no single “Science of Reading” with a capital “S,” but rather multiple sciences of reading with diverse approaches and findings. Restricting the evidence to studies that confirm a predetermined stance narrows our understanding of deeply complex processes and excludes substantial findings that could inform the early literacy development of many children, youth, and adults.
1. Natural Learning and Direct Instruction The ERI-LMIC rests on two premises: first, that spoken language acquisition is a “natural” process that occurs simply through exposure; and second, that learning to read requires direct and structured instruction. Research has challenged this contrast for decades. Language acquisition, far from being natural, is a deeply social process, one that involves young children’s interaction with other speakers in diverse contexts (Ochs and Schieffelin, 2017). Psycholinguistic research shows that children do not learn isolated phonemes; they actively participate in everyday communicative situations, reconstructing the structure, functions, meanings, and social relationships of language, building their own linguistic representation system and linking sounds to meanings. Research from sociolinguistics, anthropology, and cultural studies, among other fields, indicates that learning to read (contrary to what the report proposes) follows similar patterns as spoken language acquisition in everyday contexts — it relies on inferring general patterns of sound-letter correspondence and meaning rather than applying rules taught in advance. These studies also show that reading and writing are learned in complex social contexts that reflect the values, social positions, and identities of communities. This is a strong argument for situating early literacy within authentic literacy experiences that involve ongoing interaction with different types of texts and other readers.
The ERI-LMIC teaching model starts from the logic of the established reader — what expert readers do when faced with a text — and applies it directly to early literacy instruction, bypassing the conceptual processes learners actually go through. Ferreiro and Teberosky’s (1979) pioneering work demonstrated that children construct successive hypotheses about the writing system before mastering grapheme-phoneme correspondence. Learning to read is not a mechanical process; it is a complex conceptual one, and teaching it requires pedagogical designs that take the learner’s perspective as their starting point. The ERI-LMIC also reduces the teacher’s role to executing pre-established sequences with structured materials. Insisting on comprehensive reading instruction packages undermines teachers’ professional agency and their ability to adapt to the specific contexts of their classrooms.
The evidence supporting the ERI-LMIC comes almost exclusively from experimental and quasi-experimental studies that measure isolated skills (speed, accuracy, word recognition) through standardized assessments. The methodology is circular: skills are defined, mastery is measured, evidence is generated, that evidence justifies curriculum design or interventions, and then assessment begins again (without ever asking whether what is being measured captures how people actually read and write complete texts in their lives). The ERI-LMIC overlooks 45 years of situated qualitative research — with diverse theoretical frameworks and methodologies — that offers comprehensive descriptions of how children acquire the writing system and learn to read, and of how teachers and students interact in the classroom around written texts. Case studies, ethnography, psychogenesis, and sociocultural studies are not recognized as valid evidence for the design of educational policies, despite meeting the field’s standards for methodological rigor.
The ERI-LMIC casts the problem of reading in terms of deficit: poor children, speakers of non-dominant languages, and those from communities with little schooling experience are defined by what they lack relative to a normative student model. Standardized assessments attribute low performance to individual ability, without considering educational trajectories, prior exposure to literate culture, or structural inequalities. The “back to basics” proposal for so-called “poor readers” deepens inequality rather than addressing it: instead of offering culturally rich environments, it narrows learning opportunities even further. Zavala (2019) argues that equal opportunity discourse cannot function on the illusion that we are all equal, when society actively imposes structural disadvantages on certain groups.
The ERI-LMIC treats multilingualism as a “linguistic mismatch”: speaking a language other than the chosen language of instruction is cast as a barrier and a deficiency. While it advocates for teaching reading in students’ first language, the specific recommendations center on additional support and intensive training to bring students into alignment with the language of instruction (in many cases, the former colonial language). This technocratic and depoliticized view ignores that languages are not learned in isolation from cultural socialization, that multilingualism is a resource rather than a deficit, and that students’ multilingual repertoires are a source of knowledge that schools should build on.
The ERI-LMIC rests on two premises: first, that spoken language acquisition is a “natural” process that occurs simply through exposure; and second, that learning to read requires direct and structured instruction. Research has challenged this contrast for decades. Language acquisition, far from being natural, is a deeply social process, one that involves young children’s interaction with other speakers in diverse contexts (Ochs and Schieffelin, 2017). Psycholinguistic research shows that children do not learn isolated phonemes; they actively participate in everyday communicative situations, reconstructing the structure, functions, meanings, and social relationships of language, building their own linguistic representation system and linking sounds to meanings. Research from sociolinguistics, anthropology, and cultural studies, among other fields, indicates that learning to read (contrary to what the report proposes) follows similar patterns as spoken language acquisition in everyday contexts — it relies on inferring general patterns of sound-letter correspondence and meaning rather than applying rules taught in advance. These studies also show that reading and writing are learned in complex social contexts that reflect the values, social positions, and identities of communities. This is a strong argument for situating early literacy within authentic literacy experiences that involve ongoing interaction with different types of texts and other readers. The ERI-LMIC teaching model starts from the logic of the established reader — what expert readers do when faced with a text — and applies it directly to early literacy instruction, bypassing the conceptual processes learners actually go through. Ferreiro and Teberosky’s (1979) pioneering work demonstrated that children construct successive hypotheses about the writing system before mastering grapheme-phoneme correspondence. Learning to read is not a mechanical process; it is a complex conceptual one, and teaching it requires pedagogical designs that take the learner’s perspective as their starting point. The ERI-LMIC also reduces the teacher’s role to executing pre-established sequences with structured materials. Insisting on comprehensive reading instruction packages undermines teachers’ professional agency and their ability to adapt to the specific contexts of their classrooms. The evidence supporting the ERI-LMIC comes almost exclusively from experimental and quasi-experimental studies that measure isolated skills (speed, accuracy, word recognition) through standardized assessments. The methodology is circular: skills are defined, mastery is measured, evidence is generated, that evidence justifies curriculum design or interventions, and then assessment begins again (without ever asking whether what is being measured captures how people actually read and write complete texts in their lives). The ERI-LMIC overlooks 45 years of situated qualitative research — with diverse theoretical frameworks and methodologies — that offers comprehensive descriptions of how children acquire the writing system and learn to read, and of how teachers and students interact in the classroom around written texts. Case studies, ethnography, psychogenesis, and sociocultural studies are not recognized as valid evidence for the design of educational policies, despite meeting the field’s standards for methodological rigor. The ERI-LMIC casts the problem of reading in terms of deficit: poor children, speakers of non-dominant languages, and those from communities with little schooling experience are defined by what they lack relative to a normative student model. Standardized assessments attribute low performance to individual ability, without considering educational trajectories, prior exposure to literate culture, or structural inequalities. The “back to basics” proposal for so-called “poor readers” deepens inequality rather than addressing it: instead of offering culturally rich environments, it narrows learning opportunities even further. Zavala (2019) argues that equal opportunity discourse cannot function on the illusion that we are all equal, when society actively imposes structural disadvantages on certain groups. The ERI-LMIC treats multilingualism as a “linguistic mismatch”: speaking a language other than the chosen language of instruction is cast as a barrier and a deficiency. While it advocates for teaching reading in students’ first language, the specific recommendations center on additional support and intensive training to bring students into alignment with the language of instruction (in many cases, the former colonial language). This technocratic and depoliticized view ignores that languages are not learned in isolation from cultural socialization, that multilingualism is a resource rather than a deficit, and that students’ multilingual repertoires are a source of knowledge that schools should build on.

Final Thoughts: Towards Open Dialogue

Learning to read and write sits at the center of debates about development and equity. We welcome the fact that this topic is now on the agenda of international organizations, and that there is consensus on the importance of funding policies to strengthen reading instruction. This is a historic opportunity to debate proposals that genuinely put the interests of children first. Our position does not question the importance of teaching the relationship between letters and sounds; there is broad consensus on that. What needs reevaluation is how this instruction is conceived: as a sequence of isolated, decontextualized skills taught mechanically. To genuinely support mastery of the writing system (essential for understanding and producing written texts), we need to ask how we do this: in what context, for what purpose, with what materials, and through what learning activities.

Against the ERI-LMIC proposal, we recommend a broader and more inclusive vision, one that:

  • Recognizes reading and writing as situated social practices, not only as cognitive skills;
  • Starts from the learner’s perspective and the conceptual processes they construct;
  • Integrates grapheme-phoneme relationships into authentic literacy activities with communicative purpose and meaning;
  • Recognizes students’ linguistic and cultural repertoires as resources for learning;
  • Values the tradition of regional and local research and pedagogical experiences;
  • Offers opportunities for teacher training and professional development and trusts teachers to generate pedagogical knowledge, rather than simply implementing pre-packaged programs;
  • Insists that the communities and regions affected by these policies have a seat at decision-making tables.
Historically, literacy programs for marginalized populations have been pared down to their most rudimentary forms, resulting in “poor education for the poor” (García-Huidobro, 1994). To avoid replicating this logic—this time on a global scale—we need to listen to the diverse voices from varied theoretical and geopolitical backgrounds. This will foster a broad and inclusive science of reading that deepens — rather than simplifies — what it means to read and to become a reader. Download the full response here.