Metacognition¶
Learner Variable · All learner variables
Evidence · 132 claims (85 show a difference, 29 mixed, 18 no difference) · 96 studies (39 causal, 17 associational, 13 review, 11 theoretical, 7 quant-synthesis, 7 qualitative, 2 design),q1–q4· 28 of 96 report an effect size · 131 claims rest on one study
Description¶
Metacognition is knowing about and managing one's own cognition. It has two sides: knowledge (what a learner knows about how they learn, what they do and do not understand, and which strategies suit a task) and regulation (planning an approach, monitoring progress and comprehension, and adjusting when it is not working). It is usually measured by self-report inventories of strategy awareness, by parent and teacher rating scales, and by comparing a learner's confidence judgements with their actual performance (calibration).
Metacognition appears across reading, writing and mathematics. In reading it shows up as comprehension monitoring, noticing that a passage has stopped making sense; in writing as planning and revising; in mathematics as choosing a strategy, predicting, and checking a solution. Accounts treat it as one of the capacities behind problem solving, decision making and critical thinking. It develops through childhood and is thought to reach maturity by late adolescence, though adults continue to use and benefit from it, and older adults' metacognitive strategies are tied to how they study and remember.
It overlaps heavily with self-regulation, of which monitoring and control of thinking are usually treated as a core part, and it depends on executive function: the two are described as separate but interdependent. Peer work, discussion and caregivers who prompt children to reflect are described as occasions for it to be practised.
Implications¶
Context¶
- Matters wherever learners work without step-by-step structure: independent reading, extended writing, multi-step problems, self-paced or online study.
- Can be prompted and taught: planning, self-questioning and reflection prompts, and small-group problem solving give learners occasions to use it.
- Reading on screens may encourage overconfident judgements of understanding in adults.
Target Learners¶
- Learners from early childhood through adulthood; monitoring and control are still developing through adolescence.
- Learners who misjudge what they know, or who keep re-reading rather than changing strategy.
- Learners whose access to language was delayed (for example some deaf or hard-of-hearing learners), for whom metacognitive skills may be less developed.
Target Learning Objectives¶
- Reading comprehension and comprehension monitoring.
- Planning and revision in writing.
- Strategy selection and checking in mathematics problem solving.
- Accurate self-assessment and independent study.
Claims¶
- Access to AI advice roughly doubled reported confidence in answers in one experiment [+M] — instruction changes it
- Adult learners generally find pronunciation more difficult than children and are unlikely to achieve native-like pronunciation, though experience and self-monitoring offset this somewhat [~M] — learners who differ on it differ in outcomes
- AI-assisted task performance can rise without corresponding gains in underlying knowledge, metacognitive processing, or independent reasoning [-W] — instruction changes it
- Better awareness calibration co-occurs with lower incorrect-advice acceptance and lower attribution bias, interpreted associationally only [+M] — learners who differ on it differ in outcomes
- Both ChatGPT feedback and peer feedback prompted doctoral students to actively exercise hard evaluative judgement in statistical analysis, through different challenges [+M] — instruction changes it
- Cognitive offloading tendency is negatively associated with critical thinking, technical problem-solving, and metacognitive self-regulation [+M] — learners who differ on it differ in outcomes
- Collaborative text co-construction with the teacher as scribe gives five-year-olds autonomy of composition and judgement and sustains long focused engagement [+M] — instruction changes it
- Developmental change in offloading reflects metacognitive control, not knowledge [+M] — learners who differ on it differ in outcomes
- Experiential instruction produces larger immediate pre-post gains than instructional lecture in AI engagement and metacognitive knowledge of cognition [+M] — instruction changes it
- Feedback improves confidence calibration, and prior lure exposure inflates confidence in intrusions [~M] — instruction changes it
- Formal concept analysis identifies self-monitoring as the primary developmental gateway and multiple viable pathways to expertise [+W] — learners who differ on it differ in outcomes
- Frequent generative AI use for essay writing can foster metacognitive laziness, reducing intrinsic motivation and learning autonomy [+W] — instruction changes it
- Full reasoning traces worsened calibration bias and Brier scores relative to both other conditions [+M] — instruction changes it
- GenAI feedback limitations require students to exercise evaluative judgement [+W] — instruction changes it
- Harm profiles are subject-dependent: metacognitive harm crosses 90% in all three subjects in multi-turn, peaking at 97.44% for Qwen2.5-32B in Mathematics [~W] — instruction changes it
- High-proficiency EFL learners use more metacognitive strategies, effort regulation, and coping with problems than low-proficiency learners [+M] — learners who differ on it differ in outcomes
- Higher levels of student involvement in self-monitoring lead to stronger intervention effects (preliminary evidence) [+W] — an instructional effect differs with it
- Human–GenAI collaborative interactions show promise for supporting metacognition and sociometacognition, based on an exploratory expert case study [+W] — instruction changes it
- In a 10-student pilot, EduGuard produced higher post-test accuracy, larger normalized learning gain, lower calibration error, and lower overreliance than GPT-4o-mini Tutor [+M] — instruction changes it
- Instructional lecture produced no immediate significant change in metacognitive knowledge or regulation of cognition [-M] — instruction changes it
- Lack of knowledge of cognition, regulation of cognition, and intrinsic motivation hinder learners' metacognitive strategy use [+W] — learners who differ on it differ in outcomes
- Metacognitive activity alone does not equate to effective regulation; unaccompanied metacognitive processes may be counterproductive [~M] — learners who differ on it differ in outcomes
- Metacognitive awareness shapes whether learners use ChatGPT reflectively or copy shallowly, and metacognitive support sustains self-regulated learning [+M] — learners who differ on it differ in outcomes
- Metacognitive confidence about credit correlates with financial behaviors and advice giving, but knowledge performance does not correlate with behavior [+M] — learners who differ on it differ in outcomes
- Metacognitive self-regulation mediates the relationships between all three AI interaction mechanisms and both higher-order outcomes [+M] — learners who differ on it differ in outcomes
- Metacognitive strategies and prior coding experience significantly predict CT profile membership, with metacognition contributing more strongly [+M] — learners who differ on it differ in outcomes
- Metacognitive strategies and products of metacognitive processes were more diverse and frequent in reading strategy instruction groups than in the no-instruction group [+M] — instruction changes it
- Metacognitive strategies correlate strongly with problem-solving skills, and behavioural metrics correlate with task performance [+M] — learners who differ on it differ in outcomes
- Metacognitive Strategies Improve Learning [+S] — learners who differ on it differ in outcomes
- Metacognitive strategies, effort regulation, and coping with problems significantly predict Japanese EFL learners' proficiency [+M] — learners who differ on it differ in outcomes
- Metacognitive strategy use correlates positively with EFL reading comprehension achievement (r = .54) [+M] — learners who differ on it differ in outcomes
- Monitoring ideas was the metacognitive process with the greatest strategy variety, and its use was highest in the peer tutoring instruction group [+W] — instruction changes it
- Monitoring lowers accuracy for unlearnable rules: preposition usage was significantly less accurate in the more monitored style [~W] — learners who differ on it differ in outcomes
- Monitoring significantly improves a second language learner's accuracy on learnable rules (subject-verb agreement, regular past tense) in the most monitored style [~M] — learners who differ on it differ in outcomes
- No trace format supported calibrated self-evaluation: perceived-achieved correlations were low and non-significant in all conditions [-M] — instruction changes it
- Peer feedback promoted critical reflection, collaboration and engagement, but students doubted its reliability because peers were statistics beginners [~M] — instruction changes it
- PLC leaders' recognition of their own metacognition impacted the type of work they led in the PLC [+M] — learners who differ on it differ in outcomes
- Practice condition did not significantly affect learners' subjective confidence about credit knowledge or ability to manage credit [-M] — instruction changes it
- Qualitatively, Socratic prompts supported reflective monitoring while directive feedback was associated with overload and resistance [+W] — instruction changes it
- As cited second-hand in an AI-literacy framework paper: ChatGPT use improved short-term essay scores but triggered metacognitive laziness with no significant knowledge-transfer gains [+M] — instruction changes it
- Self-regulation and metacognition correlate strongly with mathematics achievement (r = .42 and .37) [+W] — learners who differ on it differ in outcomes
- Sketching during reading yielded the most accurate comprehension monitoring, with significantly less absolute error than photograph, no-image, and animation conditions [+M] — instruction changes it
- Social presence shows the strongest association with metacognition in an online case-based course, while teaching presence shows no significant relationship [~M] — instruction changes it
- Socratic feedback is associated with higher task orientation and comprehension monitoring than directive feedback; on feedback regulation it has no main effect, but its effect depends on agent type [~M] — instruction changes it
- Spatial thinking and generative activity reduced monitoring error while higher interest increased it [~M] — instruction changes it
- Students can only improve if they develop the capacity to monitor the quality of their own work during actual production [+W] — learners who differ on it differ in outcomes
- Students made minimal reference to higher-level cognitive or metacognitive processing benefits of group work, departing from prior PBL research [-W] — instruction changes it
- Students report heightened autonomy in English learning through self-directed regulation, personalized feedback, and active evaluation of ChatGPT suggestions [+W] — instruction changes it
- Students who use the ePSRL Management System post more scaffold reflections and self-reflections per month than non-users [+M] — instruction changes it
- Students' JoL ratings became more consistent over time and evolved from under-rating to over-rating to accurate judgement [+M] — instruction changes it
- The combined training-plus-diary condition improved specific SRL components: SRL knowledge, metacognitive strategies, cognitive strategies (elaboration, organisation, critical thinking), and resource management strategies (time-management, effort regulation) [+M] — instruction changes it
- The paper reports that metacognition—students thinking about how they understand their own learning—supports more efficient and effective learning and improved outcomes [+W] — learners who differ on it differ in outcomes
- The professional learning community is an environment in which teacher metacognition can be nurtured [+M] — instruction changes it
- With peer feedback groups added to web-based SRL training (Group TDP), students showed statistically significant gains in self-motivation, volition and reflection, beyond the planning and metacognition gains also seen in Group TD; goal-setting and elaboration gains stayed non-significant [+M] — instruction changes it
- Without instructional support, learners show low cognitive strategy use and almost absent metacognitive strategies in journal writing, and realized strategy use correlates with learning outcomes [+M] — learners who differ on it differ in outcomes
- Better metacognitive awareness of trial-to-trial VSTM performance is associated with lower error, higher precision, and fewer misbinding errors, but not with item capacity, and this association is invariant across age [~M] — learners who differ on it differ in outcomes
- Metacognitive postdiction skills are related to mathematical accuracy throughout the whole elementary school period (Grades 1–6) [+M] — learners who differ on it differ in outcomes
- The mixed model holds for metacognitive quality in both age groups: metacognitive skillfulness contributes uniquely to math performance on top of intelligence [+M] — learners who differ on it differ in outcomes
- The roles of intelligence and metacognitive quality as predictors of math performance turn around between ages 13 and 15 [~M] — learners who differ on it differ in outcomes
- Children showing higher metacognition and self-regulation felt the task would be difficult but that they would be able to solve it [+M] — learners who differ on it differ in outcomes
- Communicating feelings of difficulty is significantly associated with six-year-olds' deployment of metacognition and self-regulation, with the exception of planning [~M] — learners who differ on it differ in outcomes
- Numerical metacognition, but not emotion metacognition, predicts children's standardized mathematics achievement [~M] — learners who differ on it differ in outcomes
- Success in PBL requires students to set goals, monitor, reflect, and sustain motivation across the whole project, yet many students do not do these processes naturally [+W] — learners who differ on it differ in outcomes
- The context-learning effect did not significantly interact with reading comprehension ability, though the trend favored more able readers [-M] — an instructional effect differs with it
- A collaborative script embedding solitary reading between planning and review episodes shaped a group's regulative activity, but only low-level regulation [+M] — instruction changes it
- Script-supported groups mainly engaged in accepted intra-group metacognitive interaction, with co-constructed interaction infrequent and highly variable between groups [~M] — instruction changes it
- The 'Y-read?' collaboration script elicited mainly low-level metacognitive regulation during pre- and post-task discussions, with deep-level activities rare [+M] — instruction changes it
- Short-term prompting in weekly learning journals increases university students' self-reported metacognitive strategy use over time [+M] — instruction changes it
- Short-term prompting increases the quantity of cognitive strategy use in open-ended learning journals, but not the quantity of metacognitive strategy use [-M] — instruction changes it
- Drawing after learning, both from scratch and with spatial scaffolds, improves absolute monitoring accuracy (calibration) compared with a time-matched mental reactivation control condition [+W] — instruction changes it
- The inventing group showed more metacognitive regulation during future learning, contrary to hypothesis, without benefiting transfer [+M] — instruction changes it
- Metacognitive scaffolding during guided invention improved conceptual features, dataset ranking accuracy, and learning-strategy behaviors over unguided invention [+M] — instruction changes it
- Rereading raises judgements of learning after each reading while multiple text reading raises them only after the second reading (significant interaction, ηp2 = 0.07) [~M] — instruction changes it
- Students lack metacognitive awareness of the benefits of errorful learning: the three learning groups did not differ in perceived competence, judgments of learning, or rated method effectiveness even after tests [-M] — instruction changes it
- Constantly high complexity lowers reported intrinsic cognitive load and metacognitive strategy use compared with gradually increasing complexity [+M] — instruction changes it
- Intrinsic cognitive load is a strong predictor of metacognitive strategy use [+M] — instruction changes it
- All metacognitive strategy factors and critical thinking subfactors were significantly and positively correlated with each other and with academic writing [+M] — learners who differ on it differ in outcomes
- Both metacognitive awareness dimensions correlate positively with deep approach and organized studying and negatively with an unreflective approach [+M] — learners who differ on it differ in outcomes
- Learners request additional information mainly on items where their (adjusted) confidence is low, indicating confidence-based metacognitive regulation [+M] — learners who differ on it differ in outcomes
- Metacognition is significantly and positively associated with EFL students' critical thinking skills (Model 2) [+M] — learners who differ on it differ in outcomes
- Metacognition is significantly and positively related to EFL students' academic writing performance (Model 1) [+M] — learners who differ on it differ in outcomes
- Metacognitive judgement positively predicts perceived individual performance in CPS, which mediates perceived group performance [+M] — learners who differ on it differ in outcomes
- Pre-structured diagrams improved metacomprehension accuracy while self-assessment cues showed no differences [~M] — instruction changes it
- Simulation feedback predicts sadness and neutral facial expressions and metacognitive judgement, but not perceived task difficulty [+M] — instruction changes it
- Students in the Reflection condition became more confident in their grade estimates over the school year, unlike students in the Practice and Grade comparison conditions [+M] — instruction changes it
- The level of reflection support (practice-only, grade comparison, or reflection prompts) did not differentially affect the improvement of secondary students' calibration accuracy over time [-M] — instruction changes it
- Time pressure weakens monitoring-based self-regulation of study time: gamma correlations between judgments of learning and study time are stronger in the untimed than the timed condition [~M] — learners who differ on it differ in outcomes
- Trait metacognitive awareness positively predicts perceived group performance while perceived task difficulty negatively predicts it [+M] — learners who differ on it differ in outcomes
- Adding diagram self-assessment to diagram completion does not increase diagnostic cue utilization or metacomprehension accuracy beyond diagramming alone [-M] — instruction changes it
- Children of both age groups improved monitoring and control accuracy from T1 to T2 for both test formats, mainly for incorrect rather than correct answers [+M] — instruction changes it
- Diagramming conditions show no significant differences from control in regulation measures or reading comprehension test scores for causal relations [-M] — instruction changes it
- Elementary school children monitor and control their answers to open-ended text comprehension questions more accurately than answers to true-false questions [+M] — instruction changes it
- In this sample the best-performing university students calibrated slightly worse than the lowest-performing students, so no unskilled-and-unaware effect was supported [~M] — learners who differ on it differ in outcomes
- Monitoring accuracy predicts comprehension-based regulation accuracy, and cue availability also has an indirect effect on regulation accuracy via monitoring accuracy [+M] — learners who differ on it differ in outcomes
- Relative comprehension-based regulation accuracy is higher in the diagram-only condition than in the name-only condition (direct effect) [+M] — instruction changes it
- Subjective confidence ratings did not differ between performance quartiles for university or secondary school students, failing to replicate Miller and Geraci [-M] — learners who differ on it differ in outcomes
- Teachers' monitoring accuracy does not differ significantly between the name-only condition and either diagram condition [-M] — instruction changes it
- Teachers' monitoring accuracy is higher in the name+diagram condition than in the diagram-only condition, opposite to the blinding hypothesis [+M] — instruction changes it
- Self-reported metacognitive regulation of reading contributes to expository digital text comprehension in a home e-learning environment [+M] — learners who differ on it differ in outcomes
- Five of seven situation-model interventions significantly improved relative metacomprehension accuracy, concept mapping only marginally (p = .097), and self-explaining not significantly [~W] — instruction changes it
- Goal-driven manipulations weaken the negative effort–monitoring correlation, which is nonsignificant under incentives, time pressure or self-agency ratings taken singly [~M] — an instructional effect differs with it
- Classroom-level instructional support, organization, and emotional support are each positively associated with upper elementary students' metacognitive strategy use [+M] — instruction changes it
- Higher teacher-student conflict is associated with lower levels of specific metacognitive skills, while closeness is positively associated with metacognition in upper elementary school [+M] — instruction changes it
- High-performing readers use more monitoring strategies while under-performing readers rely more on cognitive strategies like memorization [+W] — learners who differ on it differ in outcomes
- Whole-task, metacognitive knowledge, and external standard interventions significantly improve monitoring accuracy in problem solving [+M] — instruction changes it
- Evidence on long-term persistence of self-regulation intervention effects is mixed: no fadeout for metacognitive strategy interventions versus substantial fadeout in higher-quality RCT syntheses [~M] — instruction changes it
- GenAI fluency corrupts metacognitive monitoring by producing cognitive ease that is misread as comprehension, weakening the regulation needed to keep offloading beneficial [+W] — instruction changes it
- Informing learners about the to-be-monitored cue before the problem-solving phase improves self-assessment accuracy compared with informing them only afterwards [+M] — instruction changes it
- High Overall Engagement and Metacognitive-but-not-Constructive profiles show the highest immediate and delayed knowledge, with metacognitive engagement marking the key difference from lower performers [+M] — learners who differ on it differ in outcomes
- Metacognitive strategy use correlates with comprehension non-monotonically across grades, which the model attributes to unconscious self-regulation in experienced learners [~W] — learners who differ on it differ in outcomes
- Effortful generative learning activities during problem-solving example study provide cues that improve students' item-level self-monitoring accuracy [+M] — instruction changes it
- Self-explaining does not improve retrospective monitoring accuracy, while problem complexity reduces it [-M] — instruction changes it
- Metacognitive awareness is weakly correlated with intelligence and independently predicts performance (review-reported) [+M] — learners who differ on it differ in outcomes
- Delayed diagram completion improves students' relative monitoring accuracy of text comprehension compared with no diagramming [+M] — instruction changes it
- For inference questions, lower reading skill, higher self-efficacy, and higher self-reported monitoring strategies were associated with stronger overestimation, with openness associated only for predictions [+M] — learners who differ on it differ in outcomes
- Generative activities improve students' monitoring accuracy but not their actual learning after restudy [+M] — instruction changes it
- Giving teachers access to students' diagrams improves teachers' regulation accuracy but not their monitoring accuracy [~M] — instruction changes it
- Mind wandering can function as a resource for regulation when learners have metacognitive awareness of attentional shifts [~M] — an instructional effect differs with it
- Offloading information to external stores is associated with lower memory performance, moderated by awareness of a later test [~M] — an instructional effect differs with it
- Regular and explicit teacher addressing of metacognitive strategies is linked to students applying more metacognitive strategies over time [+M] — instruction changes it
- Consistently high complexity raises meta-awareness in Round 1 but not in later rounds [~M] — instruction changes it
- Metacognition correlates with meta-awareness in the gradually increasing complexity group in Round 2, but not in the consistently high complexity group [~M] — learners who differ on it differ in outcomes
- Technology-enhanced adaptive teaching shows no significant effects on monitoring accuracy, subject-specific interest, or self-efficacy [-W] — instruction changes it
- Seventh-graders in remote inquiry learning accurately monitor their own learning, with no differences between experimentation conditions [-M] — instruction changes it
- Metacognitive prompting did not significantly change the overall frequency of metacognitive SRL activities in main categories and subcategories [-M] — instruction changes it
- Metacognitive prompting increases domain-general metacognitive SRL activities (goal specification and strategic planning) in young learners in a CBLE [+M] — instruction changes it
- Students who did not report the discrepancy in peer feedback showed longer first-pass reading times [+M] — learners who differ on it differ in outcomes
- Model answers and self-assessment instructions do not improve monitoring accuracy after delayed diagramming [-M] — instruction changes it
- Model answers combined with self-assessment instructions interactively improve monitoring-based regulation accuracy, with two model answers most effective [~M] — instruction changes it
- Model answers improve both monitoring-based and comprehension-based regulation accuracy [+M] — instruction changes it
- Higher order cognitive skills (inference, perspective taking, monitoring) are differentially related to writing quality, writing productivity, and correctness in writing in Grade 2 [+W] — learners who differ on it differ in outcomes
- Both intervention versions raise declarative metacognitive knowledge about retrieval practice equally, and this knowledge declines over three weeks regardless of version [+M] — instruction changes it
Measures¶
- Behavior Rating Inventory of Executive Function (BRIEF) (Gioia et al., 2000): parent and teacher questionnaires whose metacognition index covers ages 5 to 18.
- Metacognitive Awareness of Reading Strategies Inventory (MARSI) (Mokhtari & Reichard, 2002): self-report survey of how often students use global, problem-solving and support strategies while reading.
Related Learner Variables¶
- Self-Regulation: monitoring and controlling one's thinking is treated as a core component of regulating one's own learning and behaviour.
- Affect Regulation: reflecting on one's own thinking is one of the tools learners use to manage emotions such as frustration.
- Working Memory: holding information in mind supports comprehension monitoring, and metacognitive strategies may help compensate for memory limits in adults.
- Self-Efficacy: beliefs about one's own ability, including self-efficacy, are tied to which study strategies learners choose.
- Access: learners with hearing loss whose early access to language was limited often show less developed metacognitive skills.
- Attention: directing attention is part of noticing whether a text is being understood.
- Communication Skills: explaining and discussing reasoning, as in group mathematics talk, brings metacognitive thinking into the open.
- Adverse Experiences: early adversity is linked to unhelpful thinking habits such as rumination, and weak metacognition may worsen its effects on reading.
- English-Learner Status: bilingual learners may show earlier or stronger awareness of their own language use.
- Academic and Critical Literacy: awareness of one's reading goals and knowledge gaps supports critical and academic reading.
- Executive Function: inhibition and cognitive flexibility support accurate reflection on one's own thinking, and the two are described as interdependent.
- Critical Thinking: monitoring one's own reasoning is described as underlying higher-order and critical thinking.
Examples¶
Key Sources¶
- Digital Promise. Learner Variability Navigator: Metacognition (Adult Learner, Building Literacy 4-6, Building Literacy 7-12, Building Math 3-6, Building Math 7-10, Portrait of a Learner 4-8, Portrait of a Learner 9-12, Portrait of a Learner PK-3). lvp.digitalpromiseglobal.org. Licensed CC BY-NC 4.0.
Page metadata
| Field | Value |
|---|---|
type |
learner-variable |
id |
metacognition |
title |
Metacognition |
description |
A learner's awareness of their own thinking and their ability to monitor and steer it: planning an approach, checking understanding and changing strategy when something is not working. |
status |
review |
canonical |
True |
generated |
by claude/unspecified on 2026-10-16 |
trust |
unverified — no one has confirmed this page's content |
Frontmatter is the page's machine-readable record — see the Schema & Guide.