Remedial Teaching

Abstract

Remedial teaching, which MeSH classifies under educational psychology, is instruction designed to close the gap between what a struggling student has learned and what the curriculum expects — targeted, intensified teaching for children who have fallen behind their peers. Its modern form rests on two findings. Bloom's demonstration that one-to-one tutoring lifts the average student two standard deviations above a conventionally taught class set an aspirational target for group instruction, and Stanovich's account of the Matthew effect showed why early reading failure compounds, making prompt remediation urgent. Out of these grew a diagnostic-prescriptive tradition, the explicit and systematic instruction that meta-analyses repeatedly favour, and response to intervention — the tiered framework that now organizes remedial provision in many school systems.

Keywords: remedial teaching, response to intervention, explicit instruction

Every classroom is paced for the middle of its group, and a child who arrives already behind is served no better by that pace than a child who is ahead — the instruction is mismatched to the learner, and the gap tends to widen rather than close on its own. Remedial teaching is the attempt to interrupt that widening: to diagnose what a struggling student has not yet mastered and to teach it directly, at greater intensity and with more feedback than the regular curriculum supplies. The field has moved a long way from the vague notion of “extra help,” converging over four decades on a body of evidence about which instructional methods reliably raise the achievement of low-performing students, and on frameworks for delivering that instruction before failure hardens into a diagnosed disability.

What Remedial Teaching Is

Remedial teaching is instruction addressed specifically to the skills a student has failed to acquire on the normal schedule, delivered with the intensity, structure, and feedback that the regular classroom cannot provide to one child. In the MeSH vocabulary the descriptor Remedial Teaching is filed under both Psychology, Educational and Teaching, reflecting its dual character: it is a form of teaching, and it is governed by the psychology of how struggling learners acquire skill. It differs from ordinary review in being diagnostic. Effective remediation begins by identifying the specific component a student lacks — the phonemic distinction they cannot hear, the number fact they have not automatized — rather than re-teaching the whole lesson at a slower pace (Torgesen, 2004).

The distinction matters because the alternative — more of the same, delivered more slowly — rarely closes a gap and can widen it. A student who has not mastered the prerequisites of a lesson gains little from repeating the lesson; they need the prerequisites taught directly. This is why the vocabulary of the field is one of targeting and intensity: remedial teaching increases instructional time, shrinks the group, sharpens the feedback, and narrows the content to exactly what is missing, converting a diffuse difficulty into a series of teachable objectives (Fletcher & Vaughn, 2009). The aim is not to label the child but to change the instruction, and the measure of success is the closing of the gap, not the completion of a program.

The Two-Sigma Problem and the Case for Mastery

The modern field has an origin point in Benjamin Bloom's 1984 analysis of what he called the 2 sigma problem. Comparing conventional whole-class teaching with two alternatives — mastery learning, in which students must reach a criterion on each unit before moving on, and one-to-one tutoring — Bloom reported that the average tutored student scored about two standard deviations above the average conventionally taught student, and the average mastery-learning student about one standard deviation above (Bloom, 1984). A two-standard-deviation gain moves a student from the 50th to roughly the 98th percentile: the average tutored student outperformed 98% of the conventional class. The Worked Example below re-derives these percentile shifts.

Bloom posed this as a problem because one-to-one tutoring is unaffordable at scale, and he set the field the task of finding group methods that approach the tutoring effect. His diagnosis of why tutoring works — continuous diagnosis of error and immediate corrective feedback, with the learner never allowed to accumulate misunderstanding — became the design brief for remedial instruction generally. Mastery learning contributed the complementary principle that a struggling learner should not advance on a fixed calendar but on demonstrated competence, so that the gaps which compound into failure are closed as they appear rather than carried forward (Bloom, 1984).

Why Early Remediation Is Urgent: the Matthew Effect

If Bloom set the field's ambition, Keith Stanovich supplied its sense of urgency. His 1986 account of Matthew effects in reading — from the Gospel line that the rich get richer — described a compounding mechanism: a child who reads a little worse early on reads less, and reading less, falls further behind in vocabulary, fluency, and knowledge, so that a small initial difference widens into a large one over years of schooling (Stanovich, 1986). The gap is not static; it grows, and it grows fastest in the early grades. This reframing changed the economics of remediation. If difficulties compound, then intervening early — before a first-grade decoding weakness becomes a fourth-grade comprehension collapse — is far cheaper and more effective than remediating the entrenched failure later (Torgesen, 2004).

The Matthew effect is the empirical rationale for the whole apparatus of screening and early intervention. It implies that waiting to see whether a struggling child “catches up” is precisely the wrong response, because the natural course is divergence, not convergence. The third demonstration below makes this compounding visible, and shows how an earlier start to remediation closes far more of the eventual gap than a later one of equal intensity.

Principles of Effective Remedial Instruction

Decades of intervention research converge on a consistent picture of what remedial teaching should look like, and it is strikingly uniform across subjects. The instruction that raises the achievement of struggling learners is explicit and systematic: skills are broken into components and sequenced from simple to complex; each is modelled clearly by the teacher; students practise with guidance before working independently; and errors are corrected immediately rather than allowed to stand (Rosenshine, 2012). Barak Rosenshine's synthesis of the instruction of effective teachers — small steps, high rates of successful practice, systematic feedback, and cumulative review — describes the same principles that intervention studies recover experimentally.

This picture is not a matter of opinion. In mathematics, a meta-analysis of instructional components for students with learning disabilities found the largest effects for explicit instruction and for teaching students to verbalize their problem-solving steps (Gersten et al., 2009). In reading, a best-evidence synthesis of programs for struggling readers found that the strongest results came not from any single packaged program but from structured, teacher-led instruction with continuous assessment (Slavin et al., 2011). The convergent evidence for Direct Instruction — the fully scripted, tightly sequenced approach — is among the field's most robust: a meta-analysis spanning half a century found consistent, substantial positive effects across subjects and populations (Stockard et al., 2018). Two older traditions anticipated these findings: Ogden Lindsley's precision teaching, which made the daily charting of a learner's response rate the basis for instructional decisions, embodied the same insistence on continuous measurement and responsiveness to the individual learner's data (Lindsley, 1992).

Tiered Delivery: Response to Intervention

Knowing which instruction works does not by itself tell a school how to deliver it to the right children at the right time. The dominant answer is response to intervention (RTI), a tiered service-delivery framework that emerged in the 2000s as both a way to prevent reading failure and an alternative to the discrepancy model of diagnosing learning disabilities (Fuchs et al., 2006). RTI organizes provision into levels of increasing intensity. Tier 1 is high-quality general instruction for all students, with universal screening to flag those at risk; Tier 2 adds small-group supplemental instruction for students who do not respond to Tier 1; and Tier 3 provides intensive, often individualized intervention for the few who do not respond to Tier 2 (Fletcher & Vaughn, 2009). Progress is monitored frequently, and a student's response — their measured rate of improvement — determines movement between tiers. The measurement tool that makes this possible is curriculum-based measurement (CBM), Stanley Deno's method of brief, standardized, repeatable probes of performance on curriculum content: because the same short probe can be given weekly and plotted over time, CBM turns a student's response to instruction into a visible slope a teacher can act on, and it is what supplies the “response” in response to intervention (Deno, 1985).

The framework's appeal is that it delivers remediation early and proportionately, reserving the most intensive resources for those who most need them, and it uses non-response to instruction, rather than a test-score discrepancy, as the signal of a genuine disability. The evidence for its Tier 2 component is solid in the early grades: meta-analyses of Tier 2 reading interventions in kindergarten through third grade find reliable positive effects on reading outcomes (Wanzek et al., 2016). But RTI is a framework, not a guarantee, and its large-scale implementation has been uneven. A federal evaluation raised the uncomfortable possibility that, as implemented in some schools, RTI was associated with weaker outcomes for certain students — a finding Fuchs and Fuchs argued reflected poor implementation and over-complex decision rules rather than a failure of the underlying logic, and which they used to argue for simpler, better-specified frameworks (Fuchs & Fuchs, 2017). Table 1 sets out the tiers and what distinguishes them.

Evidence Across Domains

The strongest evidence for remedial teaching is in reading, where the accumulation of controlled studies now supports firm conclusions. In the early grades, structured interventions delivering explicit instruction in phonemic awareness, decoding, fluency, and comprehension produce reliable gains, and the evidence base is strong enough to support practice guides (Wanzek et al., 2016). Beyond the early grades the picture is more sobering: a meta-analysis of interventions for struggling readers in grades 4 through 12 found positive but more modest effects, reflecting how much harder it is to remediate reading once the Matthew-effect gap has widened and the demands shift from decoding to comprehension of complex text (Scammacca et al., 2015). The lesson is not that later remediation is futile but that it is costlier and less complete than early intervention, exactly as the compounding model predicts.

Intensity matters as much as method. Reviews of intensive early reading interventions find that increasing dosage — more minutes, smaller groups, more sessions — yields larger gains for the students with the most severe difficulties, who do not respond to standard-intensity Tier 2 provision (Wanzek et al., 2018). In mathematics the evidence, though thinner than in reading, points the same way: explicit instruction, worked examples, and the deliberate building of number sense produce the largest effects for students with mathematics difficulties (Gersten et al., 2009). Across both domains, the recurring finding is that evidence-based practices exist but are unevenly adopted, and that closing the gap between what research supports and what classrooms deliver — the concern of implementation science — is now as important as discovering new methods (Cook & Odom, 2013).

Interactive demonstrations

The three demonstrations below let the reader work with the field's core ideas. The first builds Bloom's 2-sigma comparison, sliding a class's mean achievement upward to show how a one- or two-standard-deviation gain translates into a percentile rank against a conventional class; the second is a response-to-intervention triage, sorting a cohort into tiers as a screening cutoff moves; the third shows the Matthew effect as two diverging reading trajectories, and lets the reader start remediation earlier or later to see how much of the eventual gap it closes.

The 2-sigma problem: tutoring versus a conventional class

Slide the achievement gain that a remedial method delivers. The readout is where the average student in the shifted group ranks against the conventional class. Bloom put mastery learning near +1σ and one-to-one tutoring near +2σ.

The average student in the shifted group ranks at the 98th percentile of the conventional class — outperforming about 98 of every 100 conventionally taught students.

Note. Percentile = the normal-curve area below the shift. At +1σ it is 84th; at +2σ, 98th. Deterministic; computed from a closed-form error-function approximation.

Response to intervention: sorting a cohort into tiers

Universal screening flags the lowest-performing students for supplemental instruction. Set what share of 100 students screening sends beyond Tier 1, and what share of those needs the most intensive Tier 3.

Note. A common working split is roughly 80/15/5. Tier 3 cannot exceed the total screened beyond Tier 1, so it is capped accordingly. Deterministic; no random sampling.

The Matthew effect, and when remediation begins

A strong early reader pulls steadily ahead of a struggling one — the gap widens with time. Move the grade at which remediation starts and watch how much of the eventual gap an earlier start closes.

Final gap without remediation: 45 · with remediation from grade 2: 21 · gap closed: 24.

Note. A stylized model: an earlier start compounds over more grades, so it closes more of the gap than an equally intensive later one. Deterministic; illustrative of the mechanism, not calibrated to a test.

Worked Example

Bloom's 2-sigma claim is a statement about overlapping normal distributions, and its force is clearest when the percentile shift is computed directly. Model achievement in a conventionally taught class as a normal distribution with mean 0 and standard deviation 1 — so a score is simply its z-score, the number of standard deviations from the class mean. A student at the class average sits at z = 0, the 50th percentile, since the area of the normal curve below 0 is 0.500.

Now consider the average student taught by one-to-one tutoring. Bloom's finding is that this student scores two standard deviations higher, at z = 2.0, relative to the conventional distribution. The area below z = 2.0 is 0.9772, so the average tutored student ranks at the 98th percentile of the conventional class — they outperform about 98 of every 100 conventionally taught students. Apply the same arithmetic to mastery learning, Bloom's one-sigma condition: at z = 1.0 the area below is 0.8413, the 84th percentile. So mastery learning moves the average student from the 50th to the 84th percentile, and tutoring from the 50th to the 98th.

The gap between these two — 84th versus 98th percentile — is the space the field works in: mastery learning is deliverable to a whole class and captures much of the gain, while the remaining distance to the tutoring effect is what more intensive, individualized remediation must supply. The values are the exact normal-curve areas, re-derived in code to confirm the demonstration. Figure 1 shows the two shifted distributions against the conventional class.

Discussion

The arc of the field runs from an aspiration to a delivery system. Bloom's 2-sigma finding named the achievable ceiling and diagnosed why tutoring reaches it — continuous diagnosis and immediate correction — while Stanovich's Matthew effect explained why the intervention must come early, before small gaps compound into large ones (Bloom, 1984); (Stanovich, 1986). The intervening decades established, with unusual consistency, that explicit and systematic instruction is the method that transfers most of the tutoring gain to feasible group settings, and that this holds across reading and mathematics alike (Rosenshine, 2012); (Stockard et al., 2018).

Two tensions remain. The first is between framework and fidelity: response to intervention is a coherent way to deliver early, proportionate remediation, but its benefits depend on implementation, and large-scale rollouts have too often diluted the well-specified interventions that the research validated (Fuchs & Fuchs, 2017). The second is between early and late: remediation is most complete when it comes early, yet the students hardest to reach are often the older ones in whom the gap has already widened, and for whom even intensive intervention yields more modest gains (Scammacca et al., 2015). The practical conclusion the field has reached is that the known methods, delivered early and with fidelity at the intensity a student's severity demands, are what close gaps — and that the frontier is now as much implementation as discovery (Cook & Odom, 2013).

Current Directions

The most active current work concerns intensity and individualization for the students who do not respond even to standard supplemental instruction. Research on intensive interventions is establishing how much dosage, and what kind of adaptation, the most severe difficulties require, moving beyond the one-size-fits-a-tier model toward data-based individualization for non-responders (Wanzek et al., 2018). A parallel line scrutinizes specific branded approaches against the general principles: a recent synthesis of Orton-Gillingham reading interventions — long popular for dyslexia — found the evidence for their specific efficacy surprisingly thin, a caution against assuming that a well-known program is necessarily an evidence-based one (Stevens et al., 2021).

A second front is the honest reckoning with implementation. The federal evaluation that questioned RTI's real-world effects sharpened a field-wide interest in why validated interventions lose their potency at scale, and in how to specify frameworks simply enough to be implemented faithfully; Gersten and colleagues used the same evaluation to press a set of unanswered questions about how the framework had been operationalized (Gersten et al., 2017); (Fuchs & Fuchs, 2017). This is the concern of implementation science applied to special and remedial education: the recognition that identifying an evidence-based practice is only the first step, and that the systematic study of how practices are adopted, adapted, and sustained in real schools is now essential to the field's progress (Cook & Odom, 2013).

Common Misconceptions

“Remedial teaching just means going over the lesson again more slowly.”

Effective remediation is diagnostic and targeted: it identifies the specific missing skill and teaches it directly, at greater intensity and with more feedback, rather than repeating the whole lesson at a slower pace (Torgesen, 2004).

“It is best to wait and see whether a struggling child catches up.”

The Matthew effect means the natural course of an early gap is to widen, not close, so waiting is the wrong response; early intervention is far more effective and less costly than remediating entrenched failure (Stanovich, 1986).

“A popular, long-established program must be evidence-based.”

Not necessarily. A recent synthesis found the specific evidence for widely used Orton-Gillingham reading interventions to be thin; program popularity is not the same as demonstrated efficacy (Stevens et al., 2021).

“Adopting a tiered framework like RTI guarantees better outcomes.”

RTI's benefits depend heavily on faithful implementation of well-specified interventions; a federal evaluation found that poorly implemented RTI did not reliably help, and can even be associated with weaker outcomes (Fuchs & Fuchs, 2017).

Glossary

Curriculum-based measurement (CBM).

Brief, standardized, repeatable probes of a student's performance on curriculum content, used to monitor progress frequently and decide whether an intervention is working.

Direct Instruction.

A fully specified, tightly sequenced and often scripted instructional approach emphasizing clear modelling, high rates of successful practice, and cumulative review; among the most robustly evidenced methods for struggling learners.

Evidence-based practice.

An instructional practice supported by rigorous research demonstrating that it causes improved outcomes; distinguished from practices that are merely popular or traditional.

Explicit instruction.

Teaching in which skills are made overt — clearly modelled, practised with guidance, and corrected with immediate feedback — rather than left for the student to discover; the most consistent finding of intervention research.

Fidelity of implementation.

The degree to which an intervention is delivered as designed; low fidelity is a leading reason validated interventions underperform when scaled up in real schools.

Mastery learning.

An approach in which students must reach a criterion of competence on each unit before advancing, so that gaps are closed as they appear; in Bloom's analysis it lifts the average student about one standard deviation.

Matthew effect.

Stanovich's term for the compounding of reading ability: early readers read more and improve, while early strugglers read less and fall further behind, so small initial gaps widen over time.

Phonemic awareness.

The ability to hear, identify, and manipulate the individual sounds in spoken words; a foundational early-reading skill whose weakness is a common target of remedial instruction.

Precision teaching.

Lindsley's data-based approach in which a learner's response rate is charted daily and instructional decisions are made from that continuous measurement.

Remedial teaching.

Targeted, intensified instruction addressed to the specific skills a student has failed to acquire, aiming to close the gap between attainment and curricular demand.

Response to intervention (RTI).

A tiered service-delivery framework that provides increasingly intensive instruction to students who do not respond to lower tiers, using measured response to instruction to guide decisions and identify disability.

Tier 2 intervention.

Supplemental, small-group instruction added to core teaching for students flagged by screening as at risk; the level at which most remedial provision within RTI occurs.

Two sigma problem.

Bloom's observation that one-to-one tutoring lifts the average student two standard deviations above a conventional class, together with the challenge of achieving comparable gains through affordable group methods.

Universal screening.

The assessment of every student, rather than only those referred, to identify who is at risk and needs supplemental instruction before failure sets in.

Key Researchers

Benjamin S. Bloom (1913–1999). American educational psychologist who formulated the 2 sigma problem and championed mastery learning, setting the benchmark against which remedial and group instruction are still measured. Wikipedia · Wikidata

Douglas Fuchs. American special-education researcher who, with Lynn Fuchs, developed responsiveness-to-intervention frameworks and Peer-Assisted Learning Strategies for struggling learners. Faculty page · Google Scholar

Lynn S. Fuchs (1950–2025). American researcher who pioneered curriculum-based measurement and response-to-intervention research in reading and mathematics, shaping how schools monitor and remediate difficulty. ORCID · Wikipedia · Wikidata

Russell Gersten. American researcher whose meta-analyses of reading and mathematics interventions, and leadership of federal practice-guide panels, helped establish the evidence base for explicit remedial instruction. ORCID · Instructional Research Group

Ogden R. Lindsley (1922–2004). American psychologist who founded Precision Teaching, a data-based approach that made the daily measurement of a learner's response rate the basis for instructional decisions. Wikipedia

Keith E. Stanovich (b. 1950). Canadian-American psychologist who named and analysed the Matthew effect in reading, supplying the empirical rationale for early intervention. Wikipedia · Wikidata · Personal site

Sharon Vaughn. American researcher on reading interventions and multi-tier systems of support, whose meta-analyses and intervention studies underpin much of current remedial reading practice. Faculty page

Jeanne Wanzek. American special-education researcher whose syntheses of intensive early reading interventions established how dosage and intensity affect outcomes for the hardest-to-reach students. Faculty page

Frequently Asked Questions

What is remedial teaching?
It is targeted, intensified instruction addressed specifically to the skills a student has failed to acquire on the normal schedule, aimed at closing the gap between their attainment and what the curriculum expects. It is diagnostic and focused, not simply a slower repetition of the regular lesson (Torgesen, 2004).

How is it different from ordinary review or extra help?
Ordinary review re-teaches the whole lesson; remedial teaching first diagnoses the specific missing component — a decoding skill, an unlearned number fact — and teaches that directly, with more time, smaller groups, and immediate feedback than the regular classroom can provide (Fletcher & Vaughn, 2009).

What is the 2 sigma problem?
Bloom's finding that the average student taught one-to-one scores about two standard deviations above the average conventionally taught student — roughly the 98th percentile — together with the challenge of achieving comparable gains through group methods that schools can actually afford (Bloom, 1984).

Why does remediation need to happen early?
Because reading and other skills follow a Matthew effect: early strugglers read less and fall further behind, so a small early gap widens over years. Intervening early, before failure compounds, is far more effective and less costly than remediating later (Stanovich, 1986).

What kind of instruction works best?
Explicit and systematic instruction — clear modelling, guided practice, immediate correction, and cumulative review — is the most consistent finding across decades of intervention research in both reading and mathematics (Rosenshine, 2012); (Stockard et al., 2018).

What is response to intervention (RTI)?
A tiered framework that delivers increasingly intensive instruction to students who do not respond to lower tiers: high-quality core teaching for all, small-group supplemental instruction for those at risk, and intensive individualized intervention for the few who need it, with progress monitored throughout (Fuchs et al., 2006).

Does RTI reliably improve outcomes?
Its Tier 2 component has solid evidence in the early grades, but its benefits depend on faithful implementation; a federal evaluation found that poorly implemented RTI did not reliably help, which prompted calls for simpler, better-specified frameworks (Wanzek et al., 2016); (Fuchs & Fuchs, 2017).

Is remediation still effective for older students?
Yes, but less completely. Interventions for struggling readers in grades 4 through 12 produce positive but more modest effects than early intervention, because the gap has widened and the task has shifted from decoding to comprehending complex text — which is precisely why early action is preferred (Scammacca et al., 2015).

References