Explicit instruction and modelling

• Identity and version: Name the component or programme, its edition or protocol where relevant, and the purpose being evaluated.
• Source role: Separate direct comparative evidence, descriptive implementation findings, guidance, theoretical support and firsthand experience.
• Population and setting: Retain the studied age, communication/access characteristics and setting. State when an intended application lies outside them.
• Intended and actual delivery: Record the planned components and dose separately from what was received, by whom, and with what support or interruptions.
• Comparison and other support: Name the actual comparison and concurrent changes. If there is no comparison group or it is unreported, say so rather than supply an assumed usual-care condition.
• Outcome and reporter: Record the outcome, time point and reporter. Keep the young person’s account separate from proxy judgement, and access or participation separate from task performance.
• Fidelity and adaptation: Record whether the core components were delivered and describe adaptations; distinguish an adapted application from the protocol tested.
• Burden and unwanted effects: Record time, effort, practical access, cost or workload where known, alongside refusal, fatigue, distress, restriction and other unwanted consequences. Unreported burden is not evidence of no burden.

A020 concerns teaching specified knowledge or strategies through accessible explanation, modelling, responsive practice and review. It is not a named curriculum, a prescription for all lessons or a programme-specific fidelity claim. The important distinction is what an adult teaches and what the learner is enabled to do with it.

The complete bounded earlier article is preserved from the 29 August 2026 conversation. That message reached the capture limit, but A020 ends before the following named entry. The 30 August evidence-profile message is below the cap; its A020 material and ensuing evidence-framework pilot discussion are retained exactly. This proves completeness of those captures, not that either is the current canonical master.

The older profile assigned Strong ratings, cited several research syntheses through internal handles and proposed a general evidence framework. Those are historical editorial judgements, not automatically current evidence grades. This review verifies the Education Endowment Foundation (EEF) guidance connection and reconstructs a narrower corpus. The unverified large mathematics effect, review-of-reviews and later writing/comprehension leads are not imported as confirmed findings. Their exact historical wording remains in the claim register, including the framework proposals outside the article itself.

Two sources establish guidance or identity. EEF supports a general teaching approach; National Institute for Direct Instruction (NIFDI) describes its named model. The remaining four are original comparative studies or evaluations: a science-learning experiment, a fractions package trial and two versions of a structured numeracy intervention. The last pair informs attribution, comparison group and implementation limits; it does not directly test the broad A020 family.¹²³⁴⁵⁶

This selection deliberately does not present a stack of positive programme results as independent replications of the same method. It addresses the inherited claims through a small number of inspectable comparisons. It is a targeted review, not a literature census, complete domain synthesis or formal certainty rating. A source's public accessibility does not make its population or design suitable for every intended application.

The science experiment deliberately made discovery unusually unguided. Its acquisition finding therefore concerns that particular contrast. The subsequent comparison of children classified by achieved mastery conditions on a post-treatment result: those groups no longer preserve the original random allocation. Similar later scores among selected learners do not establish statistical equivalence of assigned teaching methods or settle the place of supported inquiry.²

The fractions study tests a richer package than its label alone suggests. Its comparison group also contains teaching, examples and practice. The package alters content emphasis and instructional arrangements as well as explanation, so the study cannot identify which feature is sufficient. Likewise, a moderator or mediator analysis is not a randomized dismantling of the component it names.⁶

Catch Up is relevant precisely because both evaluations considered alternative supported teaching. Their results should not be simplified to either programme success or teaching failure. Differences between a specific product and active provision are different from differences against ordinary provision. A null package advantage leaves the contribution of shared teaching ingredients unresolved.³⁴

The science report begins with 112 children but removes eight already-expert performers from the Day 1 acquisition analysis; the resulting arms each contain 52. The fractions abstract's 259 refers to complete cases after initial allocation of 290. These denominators must not be interchanged. Neither sample was defined as representative of all special educational needs and disabilities (SEND) learners.²⁶

The first Catch Up report's allocation diagram states 336 pupils; a simple multiplication of six by 54 misses two dual-site schools. Its Methods text calls the setting England, while the geographical table includes Wales. This review preserves that inconsistency rather than resolving it by assumption. The report calls itself an effectiveness trial; a later EEF webpage describes the first trial as developer-led. The dated original design and delivery record governs the source profile.³

For 2019, 150 is the school-randomisation count and 1,481 the primary analysed pupil count. Recruitment, post-test, administrative linkage and model counts differ within the report. The principal effect interval is Bayesian credible, not a frequentist confidence interval. The negative free school meals (FSM) subgroup estimate also spans zero, with an upper-bound discrepancy between summary and appendix; no subgroup harm claim is inferred from it.⁴

The later Catch Up version allowed termly reassessment and stopping of support, whereas the earlier trial required a fixed period. The nominal weekly schedule is therefore not a shared received dose. In the later trial, control delivery also differed in grouping and possibly time. These are material interpretation limits, not reasons to silently substitute one version's results for another's.³⁴

An immediate answer, a practised procedure, a novel example and later use in an ordinary setting are distinct outcomes. This is the basis for the article's review questions; it is not an assertion that all inspected studies measured them. The science study's delayed task was a supported poster critique. The fractions study did not establish subsequent-grade durability. A broad standardised mathematics test asks another question again.²⁴⁶

The article's writing example is a constructed illustration of the adult action. It is not an observed outcome, a named writing programme or a claim that the inspected studies evaluated that exact sequence. The five core components likewise describe an editorial family scope, not a newly validated protocol.

Making a process visible is a reasonable instructional purpose. Reduced cognitive load, increased confidence or improved executive functioning must remain distinct hypotheses unless an appropriate source actually tests them. The fractions study's teacher-rated attention comparison was null, illustrating why an academic package gain does not establish broad attentional improvement.⁶

The first numeracy evaluation's accounts of unwanted withdrawal from valued lessons come from staff. The later pupil attitude questionnaire adds direct self-report on a narrow outcome, with uncertainty; it is not a detailed interview about instructional consent or everyday experience. The current corpus does not supply representative children and young people (CYP) views of generic explicit instruction. That gap is retained rather than filled with adult enthusiasm.³⁴

Atlas's requirements to maintain access, permit requests for help or breaks, avoid public humiliation and review opportunity costs are editorial safeguards. They are not claims that a particular trial demonstrated a quantified harm. Timed games or reward arrangements in a tested package should remain visible, without becoming defaults for all instruction.⁶

Fidelity needs a named object. A score for sampled essential teaching points cannot establish learner acceptability or completeness of every session. High school-level implementation can coexist with missed sessions, selection problems and uncertain exposure. Observation of the control matters too: a described curriculum is not direct observation of what a pupil received.⁴⁶

Full domain-specific reading, writing and mathematics synthesis reconstruction remains outside this bounded pass. The 2022 historical meta-analysis review and the 2024 fraction-component lead were located but did not yield substantive opened primary text, so neither supplies a current effect claim. No commercial programme is admitted by association with A020, and no generic method is presented as a substitute for relevant subject or clinical expertise.

Current canonical master/source-version selection, human specialist and recipient review, editorial adoption and publication remain pending. Separate AI editorial review is recorded in the batch support records. No professional or recipient consultation is claimed. Future work should resolve specific unanswered claims or intended applications, rather than seek a single grade for everything called explicit instruction.

A020-S01 — Special Educational Needs in Mainstream Schools — Education Endowment Foundation

Access: Full public guidance opened; teaching, scaffolding and teaching assistant (TA) sections inspected. Locator: Printed pp.24–27 and pp.36–37.

• Identity and version: Generic explicit instruction: demonstration, supported practice and later independent use.
• Source role: Official educational guidance drawing on reviews; no new trial.
• Population and setting: Mainstream pupils with SEND; heterogeneous needs, no single study population or n.
• Intended and actual delivery: Clear examples/non-examples, sequenced concepts, checking misconceptions; no common dose.
• Comparison and other support: No assigned comparison group; source reviews vary.
• Outcome and reporter: Supports a teaching recommendation, not a family effect size or universal superiority claim.
• Fidelity and adaptation: Assessment and adaptation expected; no generic fidelity score.
• Burden and unwanted effects: Staff knowledge and pupil access matter; no pooled harms or direct acceptability estimate.

Source limits: Guidance endorsement does not independently validate every component or named programme. This review does not reconstruct every underlying reading/writing/maths synthesis. ¹

A020-S02 — The Equivalence of Learning Paths in Early Science Instruction: Effects of Direct Instruction and Discovery Learning — Klahr and Nigam

Access: Full public author manuscript opened; Methods, Results and Discussion read. Locator: Manuscript pp.2–7; footnote 3; Day 1/Day 2 analyses.

• Identity and version: Generic control-of-variables instruction; explicitly distinguished from Engelmann programmes.
• Source role: Individually randomised experiment; subsequent mastery-group comparison not randomised.
• Population and setting: 112 US third/fourth graders, ages 8.4–10.6; four schools, not a SEND-selected sample.
• Intended and actual delivery: One teaching/exploration phase, immediate test, poster evaluation about a week later; actual minutes unreported.
• Comparison and other support: Modelling/examples/explanations versus deliberately unguided exploration; both manipulate apparatus.
• Outcome and reporter: Eight baseline experts excluded from acquisition analysis; 40/52 versus 12/52 reached criterion. Mastery groups later performed similarly on poster critiques.
• Fidelity and adaptation: Later interviewer blinded; no separate implementation-fidelity estimate reported.
• Burden and unwanted effects: No learner acceptability or adverse-outcome measure.

Source limits: Extreme unguided comparison group does not represent all inquiry learning. Short follow-up, selected post-treatment mastery groups; no formal equivalence or long-term SEND transfer inference. ²

A020-S03 — Catch Up Numeracy: Evaluation Report and Executive Summary — Rutt, Easton and Stacey / National Foundation for Educational Research (NFER)

Access: Full original evaluation PDF opened; design, impact and process sections read. Locator: PDF pp.4–5,8–14,17–27.

• Identity and version: 2012–13 fixed-duration Catch Up Numeracy version; structured assessment/individual numeracy support.
• Source role: Three-arm pupil-randomised trial and adult process interviews; contextual package evidence for A020.
• Population and setting: 336 pupils, 54 schools with two dual sites; Years 2–6. Geography table includes 11 Welsh schools although Methods says England.
• Intended and actual delivery: Two 15-minute sessions/week for 30 weeks planned; received frequency varied. Trained teaching assistants (TAs); resources and preparation.
• Comparison and other support: Matched-time TA numeracy support and usual provision; possible within-school contamination.
• Outcome and reporter: Analysis 108 Catch Up/102 matched/108 usual; gains over usual, no significant package advantage over matched support.
• Fidelity and adaptation: Early recruitment unobserved by external evaluator; delivery substitutions, missed sessions and cross-arm knowledge.
• Burden and unwanted effects: Staff reported workload/stress and some pupils resenting withdrawal from valued lessons or being singled out; proxy accounts.

Source limits: Not a clean test of modelling or explicit instruction. Catch Up versus usual post-hoc comparison unadjusted for multiplicity; full generalisation/harms not established. ³

A020-S04 — Catch Up Numeracy: Evaluation Report and Executive Summary — Hodgen, Adkins, Ainsworth and Evans

Access: Full original report opened; design, results, compliance and process sections inspected. Locator: Report pp.7–9,13–17,25–28,31–38,40–51.

• Identity and version: Later flexible version with termly reassessment and stopping support when eligibility judged resolved.
• Source role: School-cluster randomised trial and implementation study; contextual package evidence.
• Population and setting: 150 schools randomised; Years 3–5, ages 7–10; 1,481 in primary model. Selection did not consistently reach intended lowest-attaining group.
• Intended and actual delivery: Two 15-minute weekly sessions, one to three terms planned; logs suggest unequal received exposure, missing/noncomparable data.
• Comparison and other support: Active TA support informed by EEF guidance; more group delivery and possibly greater time.
• Outcome and reporter: Delayed maths outcome g=-0.04, 95% credible interval [-0.21,0.13]. Pupil attitude result weak/uncertain; adults also reported perceived gains.
• Fidelity and adaptation: High school-level implementation alongside selection/dose deviations; 17% pupil attrition, baseline imbalance and floor effects.
• Burden and unwanted effects: Assessment time, staff preparation and displaced provision mattered; qualitative burden mainly adult-reported.

Source limits: No usual-provision arm; no isolated explicit-instruction effect. Raw/post/model denominators differ; 1,481 is analysed n, not recruited n. FSM estimate negative but interval spans zero; subgroup is not proof of harm or a prescribing rule. ⁴

A020-S05 — The History of DI Research — National Institute for Direct Instruction

Access: Full substantive professional webpage opened. Locator: The History of DI Research; programme description.

• Identity and version: Capital-D Direct Instruction programmes developed by Engelmann and colleagues.
• Source role: Provider identity and design description; promotional effectiveness claims not adopted.
• Population and setting: Describes school programmes; no defined sample in this page.
• Intended and actual delivery: Curriculum design, sequenced content and detailed teaching guidance; programme/version-specific delivery needed.
• Comparison and other support: No assigned comparison on this descriptive page.
• Outcome and reporter: No independent outcome estimate taken from provider assertions.
• Fidelity and adaptation: Fidelity to a particular programme cannot be inferred from generic explanation/modelling.
• Burden and unwanted effects: Training, materials and recipient burden require programme-specific review; not quantified here.

Source limits: Programme identity source only, not a comparative review or admission of a new Atlas programme. ⁵

A020-S06 — Improving At-Risk Learners’ Understanding of Fractions — Fuchs and colleagues

Access: Full original advance-publication PDF opened; methods, results and limitations read. Locator: PDF pp.4–8,14–16; Fraction Intervention; Fidelity of Intervention.

• Identity and version: Fraction Challenge manual, 2010, used in this trial.
• Source role: Individually randomised package trial; 290 allocated, 259 complete cases.
• Population and setting: US fourth grade, 13 schools/53 classes; calculation <35th percentile, excluded below 9th percentile on both Wechsler Abbreviated Scale of Intelligence (WASI) subtests; 5% special education in each arm. Exact ages unreported.
• Intended and actual delivery: 36 thirty-minute lessons/12 weeks planned; 3:1 graduate-research tutors, training/supervision, conceptual teaching, practice, speed games and rewards; individual received dose unreported.
• Comparison and other support: School curriculum also included models/guided practice; different content and delivery, broadly similar instructional time.
• Outcome and reporter: Fraction outcomes favoured package; no subsequent-grade follow-up. Teacher attention ratings at six weeks showed no significant difference.
• Fidelity and adaptation: 97.69% essential points in sampled 20% of recordings; control lessons not observed.
• Burden and unwanted effects: No direct learner acceptability/harms measure; speed/reward ingredients and instructional time retained.

Source limits: Abstract n=259 compresses complete-case versus initial allocation. Multi-component package and selected population do not isolate modelling or establish general SEND effects. Mediator/moderator analyses do not prove component causation; study authors acknowledge model specification limits. ⁶

1. 1.

   Special Educational Needs in Mainstream Schools — Education Endowment Foundation. Guidance based on 2020 evidence review; opened file path dated 2025. https://d2tic4wvo1iusb.cloudfront.net/production/eef-guidance-reports/send/eef_special_educational_needs_in_mainstream_schools_guidance_report_2025-04-10-110432_klxp.pdf. Accessed 2026-09-12; Full public guidance opened; teaching, scaffolding and TA sections inspected. Locator: Printed pp.24–27 and pp.36–37.

   ↩1 ↩2
2. 2.

   The Equivalence of Learning Paths in Early Science Instruction: Effects of Direct Instruction and Discovery Learning — Klahr and Nigam. 2004; opened prepublication author manuscript uploaded 2020. https://www.researchgate.net/publication/8266887_The_Equivalence_of_Learning_Paths_in_Early_Science_Instruction_Effects_of_Direct_Instruction_and_Discovery_Learning. Accessed 2026-09-12; Full public author manuscript opened; Methods, Results and Discussion read. Locator: Manuscript pp.2–7; footnote 3; Day 1/Day 2 analyses.

   ↩1 ↩2 ↩3 ↩4 ↩5
3. 3.

   Catch Up Numeracy: Evaluation Report and Executive Summary — Rutt, Easton and Stacey / NFER. February 2014. https://files.eric.ed.gov/fulltext/ED581123.pdf. Accessed 2026-09-12; Full original evaluation PDF opened; design, impact and process sections read. Locator: PDF pp.4–5,8–14,17–27.

   ↩1 ↩2 ↩3 ↩4 ↩5 ↩6
4. 4.

   Catch Up Numeracy: Evaluation Report and Executive Summary — Hodgen, Adkins, Ainsworth and Evans. February 2019; delivery 2016–17. https://d2tic4wvo1iusb.cloudfront.net/staging/documents/projects/Catch_Up_Numeracy.pdf?v=1757291769. Accessed 2026-09-12; Full original report opened; design, results, compliance and process sections inspected. Locator: Report pp.7–9,13–17,25–28,31–38,40–51.

   ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8
5. 5.

   The History of DI Research — National Institute for Direct Instruction. Undated; accessed 12 September 2026. https://www.nifdi.org/research/history-of-di-research. Accessed 2026-09-12; Full substantive professional webpage opened. Locator: The History of DI Research; programme description.

   ↩1 ↩2
6. 6.

   Improving At-Risk Learners’ Understanding of Fractions — Fuchs and colleagues. 29 April 2013 advance online publication. https://files.eric.ed.gov/fulltext/ED552737.pdf. Accessed 2026-09-12; Full original advance-publication PDF opened; methods, results and limitations read. Locator: PDF pp.4–8,14–16; Fraction Intervention; Fidelity of Intervention.

   ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8