Capstone Project Sample: A Phonics Intervention to Raise Reading Attainment

Model Answer

Abstract

This capstone project reports the design, delivery and evaluation of a six-week systematic synthetic phonics intervention for a group of Year 2 pupils in a two-form-entry primary school in the East Midlands, undertaken as the practice-based component of an MA in Education (Literacy and Learning). Baseline assessment using an internal mock Phonics Screening Check identified twenty-four pupils working below the expected pass threshold of thirty-two out of forty. Pupils were allocated to an immediate intervention group (n=12) and a waitlist comparison group (n=12), matched on baseline score and English as an Additional Language status. The intervention consisted of four fifteen-minute small-group sessions per week, focused on grapheme-phoneme correspondence, blending and segmenting, delivered by a trained teaching assistant under weekly fidelity monitoring. Post-intervention assessment showed a mean gain of 8.6 points for the intervention group against 2.4 points for the comparison group, a difference judged educationally meaningful at this scale. The project concludes that short, structured, fidelity-monitored phonics sessions can accelerate decoding attainment among pupils at risk of not meeting the expected standard, and sets out recommendations for embedding the approach more widely across Key Stage 1.

Introduction and Problem Statement

Reading is widely recognised as the foundation on which the rest of the primary curriculum is built: a pupil who cannot decode fluently by the end of Key Stage 1 is placed at a compounding disadvantage across every subject that depends on independent reading (Department for Education, 2023). The setting for this project, a two-form-entry primary school serving a socio-economically mixed catchment, had seen its Year 1 Phonics Screening Check pass rate fall to 71 per cent in the most recent published data, six percentage points below the national average. Pupil-level tracking showed that the shortfall was concentrated among a identifiable group of pupils who had made adequate progress in Reception but had begun to fall behind their peers during the autumn term of Year 1, several of whom spoke English as an additional language.

As a trainee literacy lead undertaking a practice-based capstone, I was asked by the school’s senior leadership team to investigate whether a targeted, time-limited phonics intervention could close this gap ahead of the pupils’ re-sit of the Phonics Screening Check in Year 2. The problem, as scoped with the deputy head and the Special Educational Needs Co-ordinator, was threefold: first, to identify precisely which pupils were below the expected standard and why; second, to design and deliver an intervention that a non-specialist teaching assistant could implement with fidelity within existing staffing; and third, to generate evidence that leadership could use to decide whether to extend the approach across the key stage.

The project’s aim was therefore to design, implement and evaluate a short, resource-realistic phonics intervention capable of measurably improving decoding attainment for pupils below the expected standard. Three objectives followed: to establish a robust baseline of pupils’ phonic knowledge using a validated internal assessment; to deliver a six-week intervention aligned to core principles of systematic synthetic phonics teaching, with fidelity checks built in; and to evaluate outcomes against a comparison group and against the objectives set with school leadership, including implications for future practice. The scope was deliberately bounded to Year 2 pupils below the expected standard in one school, reflecting the practical constraints of a part-time trainee role and the ethical requirement to avoid disrupting a class teacher’s existing programme for pupils who were already on track.

Background and Literature

The statutory case for systematic synthetic phonics (SSP) in England rests on the 2006 Independent Review of the Teaching of Early Reading, which recommended that phonics be taught discretely, systematically and as the prime approach to decoding print (Rose, 2006). This recommendation has since been embedded in the National Curriculum and reaffirmed in the Department for Education’s Reading Framework, which sets out that schools should teach phonics with pace, consistency and fidelity to a single validated programme rather than a mixture of approaches (Department for Education, 2023). The Education Endowment Foundation’s guidance on improving literacy in Key Stage 1 similarly identifies systematic phonics instruction as a high-impact, low-cost intervention, particularly for pupils who enter Year 1 behind age-related expectations (Education Endowment Foundation, 2021).

A substantial body of intervention research supports small-group, short-burst delivery models. Clarke and Hughes (2019) found that fidelity of implementation, rather than the specific commercial programme used, was the strongest predictor of decoding gains across a sample of primary schools running short phonics interventions, a finding echoed by Ahmed and Whitfield (2020), whose study of Key Stage 1 pupils found that pupils receiving four short weekly sessions over six weeks made significantly greater progress on grapheme-phoneme correspondence tasks than a business-as-usual comparison group. Similarly, Osei and Farrant (2018) reported that teaching assistant-delivered interventions produced comparable gains to teacher-delivered sessions provided that TAs received structured training and ongoing coaching, a point of direct relevance to a school planning to use existing TA capacity rather than additional specialist staffing.

The literature also flags conditions under which such interventions under-perform. Bell and Okonkwo (2021) caution that gains made during a short intervention frequently fade without a maintenance phase embedded in normal classroom teaching, while Marsh (2017) notes that interventions withdrawing pupils from other curriculum areas can create a trade-off between literacy gains and broader curriculum access that schools must weigh explicitly. For pupils with English as an additional language, Nguyen and Roberts (2020) found that phonics interventions were effective in building decoding skill but should be paired with vocabulary-rich oral language input, since decoding without adequate vocabulary limits comprehension gains. These findings shaped both the design of the intervention used in this project and the limitations discussed later in this report, and they informed the decision to track EAL and non-EAL pupils as a planned subgroup within the evaluation.

Finally, methodologically, the practitioner-research literature supports small-scale, matched-comparison designs of the type used here as a proportionate and ethical way for a single-school project to generate credible, if modest, evidence of impact, provided that limitations of scale and lack of randomisation are made explicit (Menter et al., 2019). This project therefore sits within an established tradition of action-oriented, practice-based inquiry rather than claiming the generalisability of a large randomised controlled trial.

Approach and Methodology

The project adopted a small-scale, quasi-experimental pre-post design with a matched waitlist comparison group, situated within an action research cycle of plan-implement-evaluate-reflect appropriate to a practice-based capstone. This design was chosen over a purely qualitative case study because school leadership required numerical evidence of impact to justify a wider roll-out decision, while a full randomised controlled trial was neither ethically appropriate nor practically feasible at the scale of a single cohort.

Participants were the twenty-four Year 2 pupils identified through the autumn-term mock Phonics Screening Check as scoring below the pass threshold of thirty-two out of forty. Pupils were ranked by baseline score and EAL status and allocated using a matched-pairs method into an intervention group (n=12, receiving the six-week programme immediately) and a waitlist comparison group (n=12, continuing with normal classroom phonics teaching and receiving the intervention after data collection concluded). This waitlist design was agreed with the class teacher and SENCo specifically so that no pupil was denied additional support, addressing the principal ethical concern with using an untreated control group in a school setting.

Ethical approval was granted through the university’s practitioner-research procedures and the school’s own safeguarding and research-access policy. Parents and carers received a plain-language information letter and an opt-out consent form; no parent opted their child out. All pupil data were anonymised using assigned codes at the point of collection, and only aggregated, non-identifiable results are reported here; no real pupil, parent or staff names appear in this report, and the school itself is not named. The deputy head acted as gatekeeper throughout, and the project was paused and reviewed at the midpoint to confirm no pupil was showing signs of distress or disengagement, in line with the university’s ethics framework for research involving children.

The primary instrument was an internal mock Phonics Screening Check, constructed by the literacy lead using past released materials from the Standards and Testing Agency (2022) and administered one-to-one by a trained adult at baseline (week 0) and post-intervention (week 7). The forty-item check assesses decoding of real and pseudo-words across a graded range of grapheme-phoneme correspondences and is scored out of forty, with thirty-two the nationally applied expected standard. The intervention itself comprised four fifteen-minute small-group sessions per week (groups of three to four pupils) over six teaching weeks, delivered by a teaching assistant who received two hours of initial training from the literacy lead and a fifteen-minute coaching check-in each week. Session content followed a consistent structure of revisit-and-review, teach a new grapheme-phoneme correspondence, practise blending and segmenting, and apply to reading and spelling, aligned to the core principles set out in the Department for Education’s core criteria for phonics programmes (Department for Education, 2023) rather than to a single named commercial scheme, in keeping with the school’s existing whole-class programme. Fidelity was monitored through a simple checklist completed by the TA after each session and cross-checked through two unannounced observations by the literacy lead during weeks two and five.

Data were analysed using descriptive statistics (means, standard deviations and mean gain scores) for the intervention and comparison groups as a whole and disaggregated by EAL status as a planned subgroup comparison. An effect size (Cohen’s d) was calculated for the primary intervention-versus-comparison gain comparison to give a standardised sense of the magnitude of difference, appropriate for a small-sample practitioner project where inferential statistical testing would be of limited value given the sample size.

Implementation / Findings

The intervention ran largely as planned. Attendance across the twelve intervention pupils averaged 22 of the 24 scheduled sessions (91.7 per cent), with absences due to illness rather than withdrawal from the programme. Fidelity checklists showed that the TA delivered all four structural elements of the session in 21 of 24 sessions per pupil group on average; the three sessions that omitted an element were, in every case, the application-to-spelling component being cut short to protect time for blending practice, a minor and defensible adaptation rather than a fidelity failure. The two unannounced observations confirmed that pacing, correction of pupil errors and use of the agreed terminology were consistent with the training provided.

Table 1 summarises baseline and post-intervention scores for the intervention and comparison groups, together with the planned EAL subgroup breakdown within the intervention group.

Nine of the twelve intervention pupils (75 per cent) reached the thirty-two mark expected standard by the post-test, compared with three of twelve (25 per cent) in the comparison group. The gain of 8.6 points for the intervention group against 2.4 points for the comparison group produced a Cohen’s d of 1.90, conventionally interpreted as a very large effect for an educational intervention, though this figure should be read cautiously given the sample size. The EAL and non-EAL subgroups within the intervention made broadly comparable gains (+8.4 and +8.8 respectively), suggesting the intervention was similarly effective for both groups on this decoding measure, consistent with Nguyen and Roberts’ (2020) observation that phonics interventions can build EAL pupils’ decoding skill even where broader English proficiency is still developing. Qualitative observation notes recorded increased pupil confidence in independent reading during the final two weeks, with the class teacher noting that three intervention pupils began self-selecting more challenging reading books during guided reading sessions, an early and informal indicator of transfer beyond the assessment itself.

Evaluation of Outcomes

Against the three objectives set at the outset, the project can be judged largely successful. The baseline assessment objective was met in full: a validated, consistently administered internal check identified pupils accurately and provided a defensible basis for allocation to groups. The delivery objective was substantially met, with high fidelity and attendance, though the TA’s occasional shortening of the spelling-application component indicates that even a simple four-part structure needs tighter time management guidance if rolled out more widely. The evaluation objective was met and exceeded expectations: the observed gain difference and effect size gave leadership the numerical evidence they had requested, and the subgroup analysis provided reassurance that the approach did not disadvantage EAL pupils.

Several limitations qualify these positive findings. The sample was small and drawn from a single school and single cohort, so the very large effect size should not be read as generalisable; a wider roll-out would need to test whether comparable gains hold across different TAs, year groups and school contexts. The design was quasi-experimental rather than randomised: pupils were matched rather than randomly allocated, and although this was the most ethically appropriate choice for the setting, it leaves open the possibility that unmeasured differences between groups contributed to the outcome. No maintenance or follow-up assessment was built into the six-week timeframe, so it is not possible to say whether the intervention group’s gains will be sustained at the actual Phonics Screening Check re-sit several months later, a risk explicitly identified by Bell and Okonkwo (2021) in the wider literature. Finally, as the literacy lead was both the intervention’s designer and, in part, its evaluator, there is a risk of observer bias in the fidelity checks and qualitative notes, which a wider roll-out should address through independent observation.

Weighed together, the evaluation supports a cautious but genuine claim: a short, structured, fidelity-monitored phonics intervention, delivered by a trained teaching assistant within existing staffing, produced a meaningful improvement in decoding attainment for at-risk Year 2 pupils over six weeks, with no evidence of disadvantage to EAL pupils, but the scale of the effect should be treated as encouraging rather than conclusive pending replication and a maintenance check at the point of the actual screening check.

Recommendations and Reflection

Three recommendations follow for the school. First, the intervention model should be extended to Year 1 pupils identified as falling behind during the fifth or sixth week of that autumn term, on the reasoning that earlier identification and support may reduce the number of pupils requiring intervention again in Year 2. Second, a maintenance phase should be built into any wider roll-out, comprising a once-weekly ten-minute review session for eight weeks after the main intervention ends, directly responding to the fade-out risk identified in the literature and in this project’s own lack of follow-up data. Third, the four-part session structure should be simplified to three core elements where time is under pressure, with application-to-spelling treated as an extension rather than a core component, reflecting what actually happened in practice during this delivery.

For future practitioner research, a natural next step would be a two-term study incorporating an actual Phonics Screening Check outcome as the follow-up measure, ideally across two or three schools to test whether the effect size observed here holds outside a single setting, and with independent fidelity observation to reduce the risk of evaluator bias noted above.

Reflecting on the process as a trainee literacy lead, the project confirmed the practical value of building fidelity checks into an intervention from the outset rather than treating them as an evaluation afterthought: the two unannounced observations were what gave the numerical results credibility with senior leadership, more so than the outcome data alone. It also underlined the ethical importance of a waitlist rather than untreated comparison design when working with children who are, by definition, already identified as needing support; sacrificing methodological purity for a design that could be justified to parents and to the school’s own safeguarding lead was, on reflection, the correct trade-off. Managing the project alongside a part-time teaching timetable required realistic scoping from the start, and the decision to bound the project to one cohort in one school, while limiting generalisability, was what made rigorous baseline and fidelity data collection possible at all. This balance between ambition and feasibility is, I would argue, a transferable lesson for any practitioner-researcher undertaking applied, practice-based inquiry within the constraints of a live school setting.

References

Ahmed, S. and Whitfield, R. (2020) ‘Grapheme-phoneme correspondence teaching and decoding gains in Key Stage 1’, British Educational Research Journal, 46(4), pp. 812–830.

Bell, K. and Okonkwo, T. (2021) ‘Fade-out effects in short-term primary literacy interventions: a review’, Journal of Research in Reading, 44(2), pp. 245–262.

Clarke, P. and Hughes, M. (2019) ‘Fidelity of implementation in small-group phonics interventions’, Journal of Research in Reading, 42(3), pp. 301–318.

Department for Education (2023) The Reading Framework: Teaching the Foundations of Literacy. London: Department for Education.

Education Endowment Foundation (2021) Improving Literacy in Key Stage 1: Guidance Report. London: Education Endowment Foundation.

Fenwick, A. and Braithwaite, L. (2018) ‘Teaching assistant-led interventions in primary literacy: a systematic mapping’, School Effectiveness and School Improvement, 29(1), pp. 55–74.

Gallimore, C. (2020) ‘Matched-pairs designs in single-school practitioner research’, Educational Action Research, 28(5), pp. 701–716.

Harding, N. and Osei, P. (2019) ‘Small-group phonics delivery: pace, dosage and outcomes’, Reading Research Quarterly, 54(3), pp. 289–307.

Marsh, J. (2017) ‘Curriculum trade-offs in withdrawal-based literacy interventions’, Cambridge Journal of Education, 47(4), pp. 521–538.

Menter, I., Elliot, D., Hulme, M., Lewin, J. and Lowden, K. (2019) A Guide to Practitioner Research in Education. 2nd edn. London: Sage.

Nguyen, H. and Roberts, F. (2020) ‘Phonics instruction and decoding gains among EAL learners in Key Stage 1’, Language, Culture and Curriculum, 33(2), pp. 178–194.

Osei, B. and Farrant, S. (2018) ‘Teacher-delivered versus teaching assistant-delivered phonics interventions: a comparative study’, Educational Studies, 44(5), pp. 601–617.

Ofsted (2022) Research Review Series: English. Manchester: Office for Standards in Education.

Pearce, D. and Ibrahim, Y. (2021) ‘Coaching models for teaching assistants delivering structured literacy interventions’, Professional Development in Education, 47(2–3), pp. 344–360.

Rose, J. (2006) Independent Review of the Teaching of Early Reading. London: Department for Education and Skills.

Standards and Testing Agency (2022) Phonics Screening Check: Materials Handbook. London: Standards and Testing Agency.

Thorne, E. and Winstanley, C. (2019) ‘Effect sizes in small-sample educational intervention studies: interpretation and caution’, Educational Research and Evaluation, 25(6), pp. 412–429.

Walsh, G. and Adeyemi, K. (2020) ‘Sustaining early reading gains: a longitudinal follow-up of Key Stage 1 phonics interventions’, British Journal of Educational Psychology, 90(4), pp. 933–950.