Mother tongue on paper, English in practice: The language policy and pedagogy disconnect in Grade 3 classrooms in Kiambu County, Kenya

1.0 INTRODUCTION

1.1 Background of Study

The acquisition of basic literacy, numeracy, and socio-emotional skills in the early years of schooling is fundamental to a child's educational trajectory, future employment prospects, and lifelong success (World Bank, 2020). In Kenya, while primary enrollment rates are high, with 9 in 10 children attending school, learning outcomes present a profound system-level challenge: national monitoring frameworks indicate that only 45% of Grade 3 learners meet expectations in reading English text, and 60% in solving numeracy problems (KNEC, 2020, 2022a). Over a quarter of Grade 3 pupils fail to meet basic competency expectations in English and Kiswahili writing, with these foundational weaknesses persisting into upper primary grades and contributing to low Grade 6 proficiency levels (KNEC, 2022b, 2024). These outcomes highlight a critical gap between educational access and meaningful learning achievement across the Kenyan basic education sector.

The challenges facing foundational literacy in Kenya are multifaceted and structurally entrenched. Despite national language policies promoting mother tongue or local catchment language instruction in early grades (Grades 1–3) and a Competency-Based Curriculum (CBC) advocating for learner-centered, high-impact teaching strategies, evidence suggests substantial disconnects between policy intent and daily classroom practice (Wawire, 2020; Piper et al., 2018). Neither the prescribed medium of instruction (language policy) nor the endorsed method of instruction (pedagogy) is translating effectively into expected literacy gains (Abiyo, 2025).

1.2 Statement of Problem

Kenyan early grade classrooms are characterised by a dual disconnect: first, a substantial gap exists between official language policy intent and the actual language used for instruction (KNEC, 2020); second, a parallel gap persists between the pedagogical methods teachers are trained in and those they sustain in daily practice (KNEC, 2022a; Cheruiyot, 2024). The language of instruction is a fundamental pedagogical choice, and the efficacy of literacy pedagogy is contingent upon the linguistic medium through which it is delivered (Vygotsky, 1978; Street, 1984). While previous studies have examined general CBC implementation barriers (Mauti et al., 2025; Muchira et al., 2023), there is limited empirical research directly investigating how actual classroom language practices align with national policies and how teachers navigate the intersection of language choice, high-impact pedagogy, and structural classroom constraints in early grade settings (Rose et al., 2020; David & Nsengimana, 2022). Therefore, this study aimed to investigate the extent to which language practices in Kenyan early grade classrooms align with national language-in-education policy, and how teachers navigate the intersection of language choice, high-impact literacy pedagogy, and structural classroom constraints in their daily instruction.

1.3 Objectives

This study aimed to describe the prevalence and nature of the disconnect between national/school language policy, actual classroom language use, and the application of high-impact teaching practices in Early Grade classrooms in Kiambu County, Kenya.

1.4 Research Questions

The study addressed the following specific research questions:

What is the extent of teachers’ and school leaders’ awareness and understanding of the national language-in-education policy for Grade 3?

What primary languages and language strategies are actually used during Grade 3 literacy instruction?

What is the level of adoption and daily application of endorsed high-impact pedagogical practices (phonics, guided reading, comprehension strategies, differentiated instruction) in Grade 3 classrooms?

What systemic, structural, and assessment-related contradictions reinforce the policy-practice-pedagogy disconnect?

2.0 LITERATURE REVIEW

2.1 Empirical Review of Literature

International and regional literature underscores persistent policy-practice gaps in early grade language instruction (Rose et al., 2020; Hadley et al., 2022; Justice et al., 2021). In Sub-Saharan Africa, studies reveal that English-medium mandates often conflict with learners' linguistic realities, severely limiting conceptual acquisition (David & Nsengimana, 2022; Evans & Mendez, 2020). In Kenya, empirical work highlights that pre-service teacher education programmes provide inadequate Pedagogical Content Knowledge (PCK) for teaching emergent readers in multilingual settings (Wawire, 2020; Wawire et al., 2020). Magaju et al. (2024) found that only 20% of lower primary teachers in Meru County received in-service training on mother tongue education. Furthermore, Abiyo (2025) and Abiyo & Oostendorp (2025) documented that unsupportive classroom environments and resource deficits force teachers to resist or abandon official language policies. While the CBC rollout aimed to shift instruction toward active learning (Nyakiti et al., 2025), implementation is hampered by severe resource deficits, large class sizes, and inadequate continuous professional development (Mauti et al., 2025; Cheruiyot, 2024; Muthoka & Mbirithi, 2025; Ireri et al., 2024).

2.2 Theoretical Framework

The study integrates four complementary theoretical frameworks to explain how national language policies translate into classroom practice. The Policy Enactment Theory reframes policy implementation not as a mechanical, top-down transmission, but as active interpretation and translation by teachers as policy actors across situated, material, professional, and external contexts (Ball et al., 2012). The Ecological Systems Theory situates teaching practice within interacting nested systems, from micro-level classroom dynamics to macro-level cultural attitudes toward English and national assessment policies (Bronfenbrenner, 1979). Teacher Cognition and Pedagogical Content Knowledge (PCK) highlights teachers' internal mental lives, beliefs, and Pedagogical Content Knowledge as essential mediators between official curriculum mandates and daily instructional choices (Shulman, 1986; Borg, 2006; Wawire et al., 2020). Finally, the Sociocultural Theory of Learning positions language as the primary tool for cognitive mediation and literacy as a culturally embedded social practice, demonstrating that instructing children in an unfamiliar language undermines cognitive engagement (Vygotsky, 1978; Street, 1984).

2.3 Conceptual Framework

The conceptual framework in Figure 1 assumes that policy does not translate into practice automatically. Teachers interpret it within classroom conditions that may enable or constrain their choices. The independent variable, policy intent, comprises the national language-in-education policy, which prescribes the mother tongue or catchment-area language as the medium of instruction in Grades 1 to 3, and the Competency-Based Curriculum's expectation of learner-centred literacy pedagogy. Drawing on policy enactment theory (Ball et al., 2012) and teacher cognition research (Borg, 2006), the framework positions teacher mediation as the intervening variable. It comprises teachers' awareness of the policy, their language beliefs, their pedagogical content knowledge for local-language literacy, and the transfer of training into practice. Teachers who favour English or feel unprepared to teach in the local language are likely to deviate from policy.

These factors shape enacted Grade 3 classroom practice, where the disconnect becomes visible in the language of instruction, code-switching and translanguaging, and the literacy methods used. The outcomes are policy-practice alignment and learners' reading fluency and comprehension. Structural constraints act as moderating variables and limit both what teachers can do and what they choose to do. These include large classes, limited time, scarce mother-tongue materials, English-dominant assessments, limited coaching, and the linguistic diversity of Kiambu's schools.

Figure 1: Conceptual Framework

3.0 METHODOLOGY

3.1 Research Design

The study adopted a concurrent mixed-methods design, integrating quantitative survey data and qualitative field data simultaneously (Creswell & Creswell, 2018). This design was chosen because language choice, pedagogical adoption, and system-level constraints are interdependent phenomena that require real-time empirical integration to capture both statistical prevalence and contextual nuance.

3.2 Target Population and Sample Size

The study was conducted in Kiambu County, a cosmopolitan region featuring 12 sub-counties (7 urban/peri-urban, 5 rural) with diverse linguistic and socio-economic profiles (Thuo et al., 2024). The target population comprised public primary schools offering Grade 3, Grade 3 literacy teachers, heads of institutions, Curriculum Support Officers (CSOs), and education policy makers. Quantitative sample size was determined using Cochran's formula for categorical data and applying a conservative Design Effect (DEFF = 1.5) for cluster sampling, which yielded a sample size of 327 schools. Incorporating a 15% non-response buffer resulted in a target sample of 376 public primary schools (37 urban, 339 rural) selected via stratified multistage cluster sampling across Regular, Special Unit, and Integrated school categories. A total of 370 Grade 3 teachers successfully completed the quantitative survey. For qualitative data, purposive sampling was used to select 20 key informants (9 CSOs, 10 Heads of Institution, 1 County Director of Education) and 24 live Grade 3 direct classroom observations.

3.3 Research Instruments

Six standardized instruments were deployed: (1) Structured Grade 3 Teacher Questionnaire; (2) Grade 3 Learner Literacy Assessment Test (comprehension and oral reading fluency); (3) Semi-structured KII Guide for Heads of Institution; (4) KII Guide for CSOs; (5) KII Guide for Education Policy Makers; and (6) Structured Grade 3 Classroom Observation Checklist evaluating 37 observable indicators across reading practices, language medium, differentiation, and learning resources.

3.4 Data Collection Procedures and Quality Control

Quantitative survey data were collected using mobile hand-held tablets running KoBoCollect with built-in logical validation constraints. Qualitative interviews were audio-recorded, transcribed verbatim, and verified against field notes. Field enumerators underwent two days of intensive participatory training including role-play simulations and pilot testing in non-sample schools.

3.5 Data Processing and Analysis

Quantitative data were exported to Stata for descriptive statistics (frequencies, percentages, means, standard deviations) and cross-tabulations. Qualitative transcripts were imported into NVivo for inductive and deductive thematic analysis guided by Policy Enactment and System Coherence frameworks. Methodological triangulation was systematically executed by cross-validating self-reported survey responses against direct lesson observation logs and key informant testimonies.

4.0 RESEARCH FINDINGS AND DISCUSSION

4.1 Knowledge of National Policy and Informal School Rules

Kenya's national language-in-education policy prescribes mother tongue or the language of the school's immediate catchment area as the primary medium of instruction for lower primary (Grades 1–3), with English introduced as a subject and becoming the primary medium from Grade 4 (KNEC, 2020). As shown in Table 1, teacher survey findings reveal an extreme policy knowledge discrepancy. A sizeable 67.3% of teachers incorrectly identified English as the official policy-mandated language of instruction for Grade 3, while only 14.3% correctly identified mother tongue.

Table 1: Teachers' Reported Knowledge of the Policy-Mandated Language of Instruction

Understanding among other education actors was similarly uneven in the qualitative interviews. Several Curriculum Support Officers (CSOs) showed accurate knowledge of the policy, noting that English and Kiswahili account for most Grade 3 instruction, with mother tongue reserved for interpreting difficult words. Heads of Institution, however, showed only partial understanding, with several believing that the urban or cosmopolitan character of their schools exempted them from the mother tongue provision. As one explained: "Mother tongue is important according to government policy. But because it is cosmopolitan, we might not be very good at the indigenous one... so we do both English and Kiswahili" (Head of Institution 1).

The interviews further confirmed that schools actively formulate local language regimes that explicitly deviate from national provisions. For example, one school enforced a strict 'Monday Kiswahili, Tuesday-to-Friday English' rule, enforced by student 'language monitors' and punishment systems for 'language offenders', who were required to write out sentences as a correction. Another school used Kiswahili daily across Grades 1 to 3 and English from Grade 4 upward, again bypassing the mother tongue provision. These school rules appeared to enjoy parental support: "Parents are very happy when they hear children talking Swahili and English. Even at home, most parents communicate with learners using English or Kiswahili" (Head of Institution 2). CSOs acknowledged actively discouraging mother tongue instruction, citing assessment realities: "Exam instructions are in English and Kiswahili, not in local language... so we cannot tell teachers to use vernacular in class" (CSO 3).

4.2 Actual Classroom Language Practices

In daily practice, mother tongue instruction is virtually absent across Kiambu County classrooms. As presented in Table 2, in the teacher survey, 55.7% of teachers reported English as their primary medium of literacy instruction, 22.2% used a mixed approach, 21.9% used Kiswahili, and only 1 out of 370 teachers (0.3%) reported using mother tongue as the primary medium.

Table 2: Teachers' Self-Reported Primary Language of Literacy Instruction

Reported frequency of language use followed the same pattern. English was used 'often' or 'always' by 90.0% of teachers (mean 4.39/5) and Kiswahili by 64.1% (mean 3.77), compared with only 4.6% for mother tongue (mean 1.91). Direct lesson observations across 24 Grade 3 classrooms revealed an even stronger dominance of English: English was recorded as the primary medium of instruction in 87.5% of observed lessons, with Kiswahili accounting for the remaining 12.5%. Zero lessons used the mother tongue as the primary medium. Teachers used English in 91.7% of lessons, learners responded in English in 95.8% of lessons, while local mother tongue was used to support comprehension in only 12.5% of lessons and in group discussion in just 8.3% of lessons.

All three data sources therefore agree that mother tongue is almost absent. They differ, however, on how dominant English is: the gap between self-reported (55.7%) and observed (87.5%) English dominance suggests that teachers who code-switch frequently may perceive their practice as more mixed than an outside observer records it to be. Effective teacher professional development is most successful when it is sustained, practice-focused, and embedded in teachers' day-to-day instructional contexts (Darling-Hammond et al., 2017). This suggests that improving implementation of language-in-education policy will require continuous classroom-based coaching and support rather than one-off workshops alone.

4.3 Strategic Code-Switching and Translanguaging

Beneath official policy mandates and rigid school rules, code-switching and translanguaging emerge as the true functional reality of early grade instruction. In the teacher survey, 58.1% of teachers reported code-switching 'often' or 'always' (mean score 3.66/5), and 51.4% reported translanguaging (mean 3.52/5). Direct observation confirmed strategic code-switching in 75.0% of lessons, and teachers used Kiswahili to explain concepts in 79.2% of lessons. Notably, 91.4% of surveyed teachers agreed that code-switching improves learner comprehension, and 90.8% viewed translanguaging as a legitimate, helpful instructional tool. As one CSO explained: "Code-switching happens almost everywhere... It is not about pressure from anyone; it is the teacher responding for clarity" (CSO 4). Another described it as a deliberate teaching move rather than an error: "It is not a mistake; it is how the concept lands" (CSO 5). Code-switching therefore functions as an informal bridge between policy intent and classroom reality, allowing teachers to support comprehension within English- and Kiswahili-dominant lessons.

4.4 High-Impact Pedagogical Adoption

Regarding pedagogical practice, the findings demonstrate high overall adoption of general learner-centred methods endorsed under the CBC framework (Nyakiti et al., 2025). In the teacher survey, formative assessment checks were reported by 96.2% of teachers, reading comprehension strategies by 93.2%, guided reading by 91.1%, explicit word work by 90.3%, and collaborative group work by 89.2%. Direct lesson observations confirmed active learner participation and teacher questioning in 100% of visited classrooms and collaborative group/pair work in 95.8% of lessons. Table 3 presents the self-reported frequency of each practice.

Table 3: Teachers' Self-Reported Frequency of High-Impact Pedagogical Practices

However, differentiated instruction represents a sharp technical bottleneck, standing as a clear outlier at only 69.7% adoption, 26.5 percentage points below formative assessment. Qualitative findings explain that differentiated instruction is severely constrained by large class sizes, particularly in urban primary schools where Grade 3 classes average 57.5 learners, compared with 41.7 in rural schools (and reach up to 90 learners per class). As one CSO observed: "In a class of 90 learners, the teacher cannot give adequate work, mark it, and maintain control. Group work cannot function in a class of 90" (CSO 1). Teachers themselves attributed their inconsistent use of learner-centred methods mainly to structural rather than personal constraints, citing limited lesson time (44.3%), large class sizes (39.5%), and inadequate teaching materials (32.7%). The pedagogical side of the disconnect is therefore narrower than the language side. General learner-centred practice is widespread, and the gap is concentrated in more technically demanding, resource-dependent practices such as differentiated instruction and explicit phonics.

4.5 Systemic Contradictions Entrenching the Disconnect

The policy-practice gap is reinforced by two severe structural contradictions. The first is a severe instructional material deficit. A substantial 86.0% of surveyed teachers reported that local-language reading materials simply do not exist in their schools and reading materials in multiple languages were observed in only 37.5% of classrooms. As one CSO put it: "We do not have any material in mother tongue. I have never seen charts sold or bought that are in the local catchment language" (CSO 9).

The second is assessment incoherence. National monitoring examinations (KNEC KPSEA/MLP) are set exclusively in English and Kiswahili, creating a powerful systemic disincentive against mother tongue instruction (KNEC, 2022b). Teachers confirmed this pressure, with 37.3% citing the language of assessment and 21.4% citing textbook availability as reasons for their choice of instructional language. In addition, school-level grading of learners persists despite the Ministry's no-grading directive, contradicting the CBC's competency-progression model.

These structural pressures are reinforced by teachers' beliefs. While 98.1% of teachers agreed that early introduction of English improves long-term academic performance, only 33.8% believed learners read more effectively when first taught in their mother tongue, and only 36.2% felt confident teaching literacy in the local language. Yet 56.2% of teachers identified Kiswahili as the language in which learners comprehend best, while only 21.9% used it as their primary medium of literacy instruction. Instruction therefore largely proceeds in English, which teachers themselves judge to be learners' weaker comprehension language. Together, these findings show that the language policy and pedagogy disconnect is sustained less by teachers' lack of pedagogical knowledge than by missing materials, misaligned assessment, and beliefs about English as the language of academic advancement.

5.0 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS

5.1 Summary of Findings

The study demonstrates that the policy-practice gap in Kenya's early grade literacy classrooms is structural rather than a knowledge deficit in general pedagogy. While teachers demonstrate high awareness and execution of general active learning methods (formative checks 96.2%, group work 89.2%), mother tongue instruction is nearly non-existent due to a 67.3% policy misconception rate, complete absence of local-language materials (86.0%), and examination pressures. Strategic code-switching (75.0% observed) functions as an essential, teacher-led pedagogical bridge for learner comprehension.

5.2 Conclusions

The national language policy requiring mother tongue instruction for Grade 3 has suffered near-total operational collapse in Kiambu County, having been effectively replaced by English-medium instruction and informal school language regimes. Furthermore, code-switching and translanguaging do not represent instructional failures, but rather function as pragmatic, evidence-based pedagogical adaptations that enable essential cognitive mediation in multilingual early grade classrooms. Ultimately, pervasive system incoherence, specifically the complete absence of local-language textbooks, English-only national assessments, and large urban class sizes serves as the primary driver of the policy-practice disconnect across foundational literacy settings.

5.3 Recommendations

From the study findings, the following recommendations to different stakeholders are made:

The Ministry of Education should formally revise language guidelines to recognize structured translanguaging as a legitimate pedagogical tool in multilingual early grade classrooms.

The Kenya Institute of Curriculum Development and the Ministry of Education should prioritize the development and timely delivery of supplementary readers and digital learning aids, ring-fencing capitation funds before term commencement.

In-service training of teachers should must shift from generic CBE workshops to practical, language-specific structured literacy routines focusing on phonics and large-class differentiation.

5.4 Areas for Further Research

Further research may be conducted on ongitudinal tracking of early grade literacy trajectories across the Grade 3-to-4 transition under translanguaging vs. monolingual instruction.

ETHICAL CONSIDERATIONS AND DECLARATIONS

This study received formal ethical approval from African International University's (AIU) Ethics and Scientific Research Committee (Approval Ref: AIU/ESRC/2026/04) and research licensing from NACOSTI (Licence No. NACOSTI/P/26/4186669). Written informed consent was secured from heads of institutions and teachers, and written institutional consent and child assent were obtained prior to learner assessments. All data were fully anonymized using numeric participant codes. The authors confirm that this manuscript is original, not under consideration elsewhere, and free of plagiarism. The authors declare zero conflicts of interest. Funding support from ESSA, REAL Centre, Gates Foundation, and GPE KIX is gratefully acknowledged (Grant Ref: FLAFRICA001).