Skip to content

Mastery Learning

Principle · All principles
Evidence · 21 claims (8 for, 11 mixed, 2 against) · 21 studies (5 causal, 5 quant-synthesis, 4 review, 4 theoretical, 2 qualitative, 1 associational), q1–q4 · 5 of 21 report an effect size · 17 claims rest on one study

Conditional relationship

In a sequence of units where later units depend on earlier ones (cumulative procedure, principle or complex-skill goals), a learner whose response on a unit check does not yet meet a stated criterion may do better on later units and on the course examination if that response triggers corrective instruction and another, comparable check before the learner advances, rather than advancing on a fixed schedule. The relationship is conditional on four things the evidence names: the check must represent the intended capability, the correction must respond to what the response showed rather than repeat the same material, the extra time must be available, and the outcome and horizon must be stated, since course-examination gains (general-achievement), attitudes, completion (persistence) and time (learning-time) move separately. The pattern specifies a reusable gate-correct-recheck policy; its local diagnostic branches remain proposals.

Fitting the design to a situation

Ask for four facts when the brief omits them: whether the units are cumulative, whether the class moves together or each learner sets the pace, what test the outcome will be judged on (a course test or a standardized one), and how much extra class time correction can take.

If the brief says… Then change… Basis
A course of cumulative units at college, high-school or upper-elementary level End each unit with a check against a stated criterion; below it, give corrective instruction and a parallel recheck before moving on Mastery learning improves outcomes [+M]
The criterion has not been set Set a high unit standard rather than a low one; programmes with higher standards had larger effects Effects vary with standard, pacing and test type [~M]
Learners will work at their own pace Prefer group-based pacing with individual correction; if self-paced, add deadlines or pacing support, because completion falls Self-paced mastery programmes reduce course completion [-M]; Effects vary with standard, pacing and test type [~M]
The class mixes stronger and weaker students Use it; both groups gain, and the spread of scores narrows Effects larger for lower-aptitude students [+M]
Elementary students Expect a larger effect than in high school or college Effects larger in elementary school [~M]
A science course Expect a smaller effect than in mathematics, social studies or language arts Effects smaller in science [~M]
Success will be judged on a state or standardized test Expect a much smaller effect than on the course's own tests, and check that gating does not stop learners reaching later tested topics Effects vary with standard, pacing and test type [~M]; Tension between mastery and curricular coverage [-W]
The course must keep to a fixed scope or calendar Budget class time for the correction loop (about a fifth of class time in one study) and decide in advance what happens at the time limit Trained teachers gave 20% of class time to the corrective loop [~W]; Mastery learning did not reduce pace or scope [~W]
The outcome that matters is retention after the course Recheck earlier units at a delay; gains held at about eight weeks Gains persist at follow-up [+M]
Teachers are adopting it for the first time Prepare teachers for feeling less confident at first, even as results improve Mixed effects on teacher variables [~M]

Observation, state and explanation

What a mastery check records is a response to one task under stated conditions (items, aids, time, attempt number, earlier exposures), not a state called "mastered". Below-criterion and at-criterion are observations against a locally chosen cut score; readiness for the next unit is an inference from them, and so is "has a gap". A second attempt is not independent of the first: the learner has seen the items or their kind, has had the correction, and may have learned the test rather than the capability. "Novice" here means novice on the unit's task, not in the subject.

Response Plausible explanations Observation that could change the interpretation
Below criterion on the first unit check Missing prerequisite from an earlier unit; current-unit relation not yet learned; check misaligned with what was taught; access or language demand of the items Probe the prerequisite separately from the unit task; compare an item on the same relation in a different representation; ask the learner to explain one wrong step.
Passes on the recheck after correction Learned the relation; recognised repeated or near-identical items; correction taught to the items; lower cut on the recheck Use a parallel form with changed surface and a changed relevant condition; check performance again at a delay or in the next unit's first task.
Passes the gate, then struggles in the next unit Unit check sampled a narrow part of what the next unit needs; forgetting between units; the next unit's demand is new rather than cumulative Compare the next unit's failed step with the previous check's items; re-probe the earlier capability at the point of struggle.
Repeated below-criterion attempts Correction repeats the original instruction; a deeper prerequisite gap; disengagement or reduced effort after repeated failure; the criterion is set above what the unit can teach Change the form of correction and observe the next attempt; probe an earlier prerequisite; ask about purpose and barriers; check how other learners fare on the same item set.
Slow progress, falling attendance or withdrawal Time cost of reattempts exceeds what the learner can give; the gate is experienced as punitive; the learner's goal differs from the unit criterion Record time per unit and attempt count; ask what the learner wants from the course; compare with completion in non-gated sections where available.

These rows are proposals for interpreting responses, untested with learners. A probe can shift confidence between explanations; one contrast does not identify a cause, and the probe itself is an exposure that can change the response. Record what remains unresolved.

Evidence and qualifications

  • Mastery learning programmes and examination performance [+M]: a meta-analysis of 108 controlled evaluations at college, high-school and upper-elementary levels (synthesis-experimental, mixed population bands) reports positive effects of mastery programmes on examination performance (general-achievement) against conventionally paced instruction, apparently stronger for the weaker students in a class, varying with the mastery procedures used, study designs and course content. It also reports more positive attitudes toward content and instruction, possibly more student time on instructional tasks, and, for self-paced programmes in college, often reduced completion. Checked by the judge on the abstract only; the claim page records no pooled effect size, interval or common horizon. It does not predict an effect for one learner, a delayed-retention gain, or which mastery procedure to use.
  • Formative assessment and achievement [+M]: syntheses of K-12 (pre-k-12) formative assessment interventions with control conditions report small positive overall effects (d = .29 over 33 studies; a weighted mean of .20 over 13 studies), with larger estimates where students were supported in self-assessment. This bears on the check-and-act loop that mastery learning builds on, not on the gate itself; the feature estimates are comparisons across study configurations, not randomised comparisons of features, and no common horizon is established. Two of three entries pass the judge; the third could not be confirmed from its abstract.
  • Contingent versus fixed support [+M]: in one-to-one tutoring of long division with fourth and fifth graders (n = 8 per condition), fully contingent support beat fixed, moderate, partly contingent and no support on immediate and one-month follow-up tests; a within-subjects tutoring study reports similar immediate outcomes but better transfer for interactive, contingent tutoring; a synthesis of dynamic assessment ranks explicit strategy training above contingent scaffolding. This supports correction that responds to the response rather than repeating it, in one-to-one settings; classroom contingency is largely undemonstrated. Not settled: none of the three judged entries could be confirmed from abstracts.
  • Earned credit in mastery and non-mastery sections [~W]: a retrospective analysis of one college system's records (correlational, postsecondary) found higher earned credit rates in mastery sections in eight of nine course-level comparisons, but students taking one mastery class had a lower rate (59%) than those taking none (64%) or two (66%), which the authors attribute to course selection. No effect size; assignment not controlled; not yet checked against its source. It shows that a section-level comparison can be confounded by who enrols where, not that mastery sections cause higher credit rates.

Keep population band, comparator (conventionally paced, group or self-paced), the mastery procedure, the instrument and the horizon with each finding. Do not combine the synthesis and the records analysis into one strength, and do not read "positive effects" as a size.

The evidence by question

Each question below links a general claim that gathers the studies behind it into one table, one row per distinct study, with each study's design and effect as recorded and no averaged verdict (eval/synthesis/mastery-learning.json, scripts/synthesize_claim.py). The count is of distinct studies across the claim and the narrower claims that bear on it.

Further evidence, not yet read against this model

Claims this page cited before it was rewritten as a conditional model. They are evidence about the relationship, but none has yet been re-read against the model above; each says how far its own sources have been checked.

Objective and learner-valued goal

The designer's objective in mastery learning is usually a criterion: perform the unit's task to a stated standard before the next unit. What the learner values may be something else: finishing the course this term, a credential, using one skill at work, or not being held back while others move on. Ask what the learner wants and why, record agreement or divergence as an observation, and do not infer it from compliance with retakes.

For example, a part-time college learner may value completing a statistics course within the term, while the designer's objective is criterion performance on each cumulative unit. The synthesis above reports that self-paced mastery programmes in college often reduced completion. A design that serves both aims states the criterion, the time it will cost, and what happens when time runs out (a capped number of attempts, a parallel route, or an explicit decision to advance with a recorded gap), rather than letting the gate decide by default.

What would revise this model?

The expectation should weaken if comparable learners, with valid unit checks and responsive correction, do not outperform conventionally paced comparisons on aligned and on independent outcomes, or if gains appear only on rechecks that repeat the first check's items. If differences between mastery and non-mastery sections disappear once enrolment and course selection are controlled, treat the section-level evidence as confounded. If correction that merely repeats instruction does as well as responsive correction, revise the claim that responsiveness is the active ingredient. If the time and completion costs exceed the achievement gain that matters to learners, the model is incomplete without them. Do not protect it by relabelling every failure as a misaligned test or an unmotivated learner.

Unit-check success, course-examination performance, delayed retention, transfer to new tasks and completion are separate claims. The present evidence does not establish an optimal cut score, a number of attempts, a unit size, or a forecast for an individual learner.

Claims

Examples

  • A unit where learners reassess after targeted feedback until they meet the rubric threshold.
  • A digital practice system that unlocks later tasks only after demonstrated proficiency on prerequisite skills.
  • Small-group reteaching for students who miss the first checkpoint while others move to extension work.

Key Sources

  • Bloom, B. S. (1971). Mastery learning. In J. H. Block (Ed.), Mastery learning: Theory and practice. Holt, Rinehart and Winston.
  • Guskey, T. R. (2007). Closing achievement gaps: Revisiting Benjamin S. Bloom's "Learning for Mastery." Journal of Advanced Academics, 19(1), 8-31. doi:10.4219/jaa-2007-704
  • Kulik, C. L. C., Kulik, J. A., & Bangert-Drowns, R. L. (1990). Effectiveness of mastery learning programs: A meta-analysis. Review of Educational Research, 60(2), 265-299. https://doi.org/10.3102/00346543060002265
Page metadata
Field Value
type principle
id mastery-learning
title Mastery Learning
description When a learner's response on a unit check falls short of a stated criterion, corrective support and a further check before advancing may raise later unit and course performance, qualified by the validity of the check, the quality of the correction, the time it costs and the horizon assessed.
status review
canonical True
generated by claude/unspecified on 2026-10-01
trust unverified — no one has confirmed this page's content
sources 1 — kulik-1990

Frontmatter is the page's machine-readable record — see the Schema & Guide.