Skip to content

Self-monitoring

Principle · All principles
Evidence · 29 claims (10 for, 16 mixed, 3 against) · 32 studies (12 causal, 8 quant-synthesis, 6 review, 3 associational, 3 theoretical), q1–q4 · 7 of 32 report an effect size · 20 claims rest on one study

Conditional relationship

This page owns checking one's own work while the task is still under way: a learner pauses at set points, compares what they are doing (their comprehension of a passage, the steps of a solution, a draft against a checklist, whether they are on task) with a stated criterion, records the result, and does something with it before going on. For a learner who has a criterion they can read, a record they can make quickly, and a concrete next move for each result (reread, fix the step, ask for help, carry on), adding taught and prompted self-monitoring to the task is expected to produce better performance on that task (immediate-performance, and for reading conceptual-understanding) and, where behaviour is the target, more time on task, than the same task without it. Whether the habit continues once prompts are removed, and whether it carries to other tasks, is a separate claim the evidence here does not settle.

The relationship is conditional on four things. Monitoring accuracy: learners' judgements of their own learning are often wrong, especially just after study and for learners with little task knowledge (novice), so a record built on a felt sense of "I get it" can steer them wrongly; monitoring checked against a concrete test (answer a question, compare with a worked answer, count the errors) is expected to work where a confidence rating is not. A next move: noticing a problem helps only if the learner knows what to do about it and has time to do it. The criterion: a specific standard, and for a novice one about the process (the steps of a strategy) rather than only the result, is expected to make the check more useful. Cost: frequent stops interrupt the task, and a record kept with no use (a diary on its own) has shown no gain.

No claim in the wiki isolates self-monitoring, with everything else held the same, against no self-monitoring on a learning outcome. The nearest evidence is: comprehension-monitoring strategy training against conventional teaching (one non-randomised classroom study), a meta-analysis of metacognitive strategy instruction of which monitoring is one part, self-recording of on-task behaviour in single-case studies (behaviour, not learning), and metacognitive prompts in computer-based environments.

How the converted siblings sit around it: Self-Regulated Learning holds the whole plan–monitor–act cycle and regulation training; self-monitoring is its middle phase, studied here on its own as a design move, and this page does not repeat that page's model of training and prompts. Goal Setting & Monitoring holds what the goal should be (specific, challenging, process or outcome, with an if-then plan) and how progress over days or weeks is made visible; this page takes the goal as given and is about checks within a task. Reflection holds examining a completed performance to change the next one; self-monitoring happens before the task is finished.

Default design, while the relationship is untested

The earlier page's guidance (criteria, checkpoints, a strategy menu, feedback to confirm the learner's judgement), kept as a concrete default a designer can act on and revise. Each step says what its evidence is.

  1. Name the criterion before the task, in the learner's terms. Give a short checklist, a rubric line, a worked answer or a model to compare against (for reading: "Can I say what this section said, in one sentence, without looking?"). For a novice, phrase at least some items as process ("Did I check the units?") rather than only the result ("Is my score above 80%?"). Basis: Process goals lead to better skill acquisition for novices than outcome goals. [~M] (high-school writing revision; a goal study, carried here to monitoring criteria); Specific, difficult goals lead to higher performance than easy or vague "do your best" goals. [~M] (a review of goal-setting research, mostly outside education). The checklist itself is an untested proposal.
  2. Set the checkpoints in advance, not only at the end. Reading: a stop after each section or each page (about every 5–10 minutes of reading). Problem solving: a stop after each solved item, before the next. Writing: a stop after each paragraph or after the draft, before submitting. On-task behaviour: a signal at intervals (the single-case studies summarised on the claim page used ratings mostly every 15 to 60 seconds with an app; for a whole class, a tone every few minutes is a proposal). The intervals for learning tasks are untested proposals; the behaviour interval is from Self Monitoring Improves On Task Behavior [+M].
  3. Make each check a test, not a feeling. At each stop the learner does something that can be wrong: answers a question about the section, reworks the last step another way, compares with a worked answer, counts the checklist items met. Ask for a confidence rating only after that attempt, and never use the rating alone as evidence of learning. Basis: Judgments of learning are often inaccurate [-M] (laboratory and video-lecture experiments with undergraduates: fluency raised predicted learning without raising recall).
  4. Attach a next move to every result. Give a short menu printed beside the checklist: if I could not answer, reread the section and try once more; if the step is wrong, find the first wrong line and fix it; if it is still wrong after one try, mark it and ask; if all is well, go on. The menu is an untested proposal; the comprehension-monitoring training in Metacognitive strategy instruction, which includes comprehension monitoring, improves academic performance by a medium amount, and in one non-randomised study comprehension-monitoring training raised low achievers' reading scores [+M] taught detection and repair together.
  5. Teach it before asking for it. In the first session, model the whole routine aloud on one task (stop, check, record, choose a move), then have learners do it on two or three items while you check their records against the work. Basis: Self Assessment Accuracy Is Low Without Training [~M] (80 Dutch pre-university students: those who studied modelling examples of self-assessment rated their own performance more accurately, ηp² = .10). The single-case behaviour studies trained learners for 20–45 minutes before self-recording began (Self Monitoring Improves On Task Behavior [+M]).
  6. Check the learner's record against an outside score, and show them the match. On some checkpoints, compare what the learner recorded with what a teacher, an answer key or software finds, and tell them where they matched and where they did not. Calibration feedback of this kind is described on the self-assessment claim pages as what improves accuracy, but in their Discussions, not their entries; untested proposal.
  7. Keep the record short and use it. One line or one tick per checkpoint; a record that nobody reads and that leads to no change is not enough. Basis: A daily learning diary alone (Group D) did not produce statistically significant pre-post gains on any measured outcome in an online mathematics preparation course [-M] (one randomised trial; no difference detected, equivalence not tested).
  8. Plan the hand-over across sessions. A default arc for a target several weeks away (untested proposal throughout):
  9. Sessions 1–2 (support high): the teacher models; learners use a printed checklist with three to five items and a fixed checkpoint after every section or item; the teacher co-checks every record.
  10. Sessions 3–5 (support shared): learners run the routine themselves; the teacher or software checks the match on about one checkpoint in three; the checklist stays visible.
  11. Session 6 onward (support low): checkpoints become learner-chosen ("stop where it got hard"), the printed checklist is removed for learners who no longer need it, and an outside check of the match happens once a week.
  12. Moving on: a learner moves to the next stage once their own judgement matches the outside score on about four of five checkpoints in two consecutive sessions, and returns a stage if the match falls below that. Learners whose first checks already match can start at the shared stage. Each segment of in-task checking should take under a minute at a reading or problem stop, so the checks take no more than about a tenth of task time.

When a condition is not met: if no outside check of the learner's judgement will ever happen, use items whose answers the learner can check themselves (an answer key, an inverse operation, a worked answer) rather than ratings; if the learner has no move to make after a check (no time, no help, no second attempt), drop the check, because it measures without helping.

Fitting the design to a situation

The three facts that most change the decision:

  • What the learner knows about this task. A novice cannot judge their own work well and is often overconfident; an advanced learner may monitor well enough without a checklist. If unstated, ask: when this learner says they understand, how often are they right? Try two items and compare.
  • What is monitored, and against what. Comprehension, the steps of a procedure, a draft against a rubric and on-task behaviour are different checks with different evidence. If unstated, ask: what exactly should the learner notice, and what would they compare it with?
  • Who can check the learner's judgement, and how often. Without an outside comparison, a learner's record cannot be corrected. If unstated, ask: will a teacher, a peer, an answer key or software ever compare the learner's record with the work, and when?
If the brief says… Then change… Basis
Learners are novice on the task, or low achievers Use a printed process checklist, checks that are tests (answer a question, rework a step) not ratings, every record co-checked in the first two sessions, and more sessions at the shared stage before hand-over Prior Knowledge Needed For Accurate Self Assessment [~M] (synthesis of 22 meta-analyses; undergraduates' exam predictions); Process goals lead to better skill acquisition for novices than outcome goals. [~M] (writing revision, high school)
Learners are advanced, or their first checks already match an outside score Start at the low-support stage: learner-chosen checkpoints, no printed checklist, a weekly outside check Untested; the age and expertise moderators on the self-assessment claims (Self Assessment Accuracy Is Low Without Training [~M]: closer agreement in advanced courses) suggest it
Learners are child or adolescent with, or at risk for, attention difficulties or disabilities, and the problem is staying on task Self-recording of on-task behaviour at short cued intervals (an app, a vibrating timer or a tone), after a training session; record the behaviour, not the learning, and check separately whether work completed changes Self Monitoring Improves On Task Behavior [+M] (single-case studies, grades 4–12, on-task behaviour only)
Goal is reading comprehension, learners are low-achieving secondary students Teach a comprehension-monitoring routine (stop, summarise, mark what is unclear, repair) over several weeks, with a stop at every section Metacognitive strategy instruction, which includes comprehension monitoring, improves academic performance by a medium amount, and in one non-randomised study comprehension-monitoring training raised low achievers' reading scores [+M] (one non-randomised study, 127 Nigerian senior-secondary students, raw gains 18.92 against 0.93; plus a meta-analysis of metacognitive strategy instruction, g = 0.50 at posttest and 0.63 at follow-up, of which monitoring is one part)
Setting is online-self-paced, no teacher present Build the checks into the software as task-specific prompts at natural pauses, each followed by an item with automatic feedback that shows the learner whether their judgement was right; do not rely on a free-text diary Metacognitive prompts improve learning [+M] (meta-analysis in computer-based environments, g = 0.40 on learning outcomes, larger with feedback, task-specific and adaptive prompts); diary alone [-M]
classroom with one teacher and many learners Run whole-class checkpoints on a common signal; every learner records on a one-line card; the teacher compares the cards of four or five learners per checkpoint with their work, rotating so each learner is checked at least weekly; whole-class feedback on the commonest mismatch Untested; the behaviour evidence is from individual students in single-case designs
Facilitator time is scarce: volunteers, busy shifts, a facilitator who drops in Keep the routine to two fixed pauses per session (one early, one midway) with a three-line card: what I am confident about and the check that shows it, what I am unsure about, my next move. Partners check each other's cards against the posted criteria (the task's own standard, or pictures of it); the group keeps one shared progress board; the facilitator looks at one pair's cards per pause instead of every record, and cards are kept, not graded Untested proposal; the partner check stands in for the outside score in step 6, which self-assessment accuracy is low without training [~M] says is needed
The objective itself names a self-check routine (pause, say what I know, what I am unsure of, my next move) Use that routine as the learner's script, word for word, at the checkpoints the brief sets; attach one quick check to the "what I know" line (redo the step, answer one item, compare with the standard) so a confident wrong answer is caught, and let the "unsure" line trigger the next move from the menu in step 4. Do not replace the requested routine with a different one Untested proposal; the check on the "know" line is what judgments of learning are often inaccurate [-M] calls for
Mixed mode: some learners remote, some in the room, or learners on phones only Use one shared digital record for everyone (a form or poll with one tick per checkpoint and the check item beside it), so remote and in-room records are checked the same way; on a phone, keep the checklist to three items and the check to one tap or one short answer Untested proposal
Time is one session only Teach and run the routine once on two or three items with every record checked; report the match between judgement and score; do not claim a habit or later effect Untested proposal
Learners have low confidence, or low literacy Use pictures or a three-point face scale instead of written criteria, let learners show the check by doing (point to the error, redo the step), and point out where their judgement was right as well as where it was wrong; underconfident learners may stop using a strategy that is working Untested proposal; underconfidence is described in the self-assessment accuracy [~M] claim's Discussion, not its entries
Adults in a work setting Write the checklist from the job's own standard (a procedure, a safety step, a quality criterion); keep records private to the learner and a supervisor, so the check is not read as surveillance; check against outcomes the work already produces (errors caught, rework) Untested proposal
A product or performance for a real audience Monitor drafts privately against the audience's criteria (what the reader or client needs) before anything is shown; the audience sees the product, not the checklist or the learner's ratings Untested proposal
Stakes are high, or the self-record will feed a grade or an adaptive system Never route a learner on a self-rating alone; use the check items' scores, and keep the self-ratings as data about calibration Judgments of learning are often inaccurate [-M] (laboratory and video-lecture experiments; carried here to adaptive and graded use)
Task is open-ended (an essay, a design) with no single right answer Give an annotated exemplar for each criterion and ask the learner to point to the evidence in their own draft for each, rather than score it Untested; the self-assessment claim's Discussion says calibration is weaker on open-ended work, but its entries do not test it

Five of the thirteen rows rest on a claim, and the advanced-learner row leans on a moderator reported in one. Only two of the five (comprehension monitoring for low-achieving secondary readers; self-recording of on-task behaviour) test monitoring itself, and one of those measures behaviour rather than learning; the others carry a neighbouring finding (goal setting, metacognitive prompts in general, judgement accuracy) and say so.

Observation, state and explanation

What can be observed is the learner's record at each checkpoint (a tick, a rating, a note, a chosen next move), what they did next, and the outside score on the same work. "Monitors accurately", "is overconfident" and "has internalised the routine" are inferred states. A filled-in checklist shows only that the checklist was filled in.

Response Plausible explanations Observation that could change the interpretation
Ticks "understood" at every checkpoint, then fails the check items Judging on fluency or familiarity; too little task knowledge to see the gap; the criterion was not understood; ticking to finish quickly Ask for a one-sentence summary or a worked step before the tick; give the learner an exemplar to judge against the same criterion; compare ticks on a familiar and an unfamiliar section.
Accurate record, same mistake on the next item No usable next move; no time to act; the move chosen does not fix this error Ask what they would do about the problem; offer the move menu and record which is chosen; give time for a second attempt.
Rates their work low when it is right Underconfidence; a criterion read as stricter than intended; anxiety about being judged Show the match between their rating and the score over several checkpoints; ask what they think the criterion requires.
On-task ratings rise, work completed does not Recording changed attention but attention was not what limited learning; ratings drift upward without real change Compare the learner's on-task record with an observer's on a sample of intervals; record work completed and accuracy separately.
Stops monitoring once prompts are removed Prompted performance, not an acquired habit; the routine costs more than it seems to give; the goal is not one the learner values Remove prompts on one task only and observe; ask what the checks have done for them; return to the shared stage with a shorter routine.
Spends long at each checkpoint, finishes less Too many checklist items; the stop interrupts a task that needs continuity; perfectionism Cut the checklist to three items, move stops to natural breaks, and compare the time per checkpoint and the work completed.

These rows are proposals, untested with learners as diagnostics. A check item is itself practice and changes what it measures, so record each one, and keep the explanations that remain open rather than choosing the one whose name fits.

Evidence and qualifications

No claim here compares the same task with and without self-monitoring alone on a learning outcome. The claims below test neighbouring parts of the relationship; they differ in learners, outcomes and designs, so do not rank them by effect label.

None of these measures whether a self-monitoring habit persists after prompts are withdrawn, or transfers to a new task, against a comparison.

The evidence by question

Further evidence, not yet read against this model

Objective and learner-valued goal

The designer's objective is usually better performance on the task now and, later, a learner who checks their own work without being told. Elicit separately what the learner wants from the checks: to finish faster, to avoid being caught out, to stop feeling lost, to be trusted to work alone. Agreement and divergence are both observations. A learner who values finishing will tick boxes quickly; one who fears being judged may rate everything low or leave the record blank; one who wants independence may welcome a routine that lets them drop the teacher's checks sooner. A blank or uniform record is not, on its own, a monitoring deficit.

For example, an adult apprentice electrician may want to pass a practical assessment and not be the one who caused a fault on site, while the designer's objective for the week is accurate self-checking of wiring against a step checklist. The checklist serves the designer's objective; asking the apprentice which faults they most fear making, and putting those steps first on it, serves theirs. Record both, so that a full checklist is not read as the safety the learner came for.

What would revise this model?

The expectation should weaken if, with comparable learners and the same task, adding trained, test-based self-monitoring with a next-move menu shows no gain over the task alone on an aligned outcome, or if gains appear only on self-reported regulation and never on scored work or behaviour traces. If learners whose judgements match outside scores do no better than those whose judgements do not, the mechanism (accurate detection leads to useful action) is wrong for that task. If gains vanish when prompts are withdrawn, the model describes prompted performance, not an acquired habit, and should say so. If on-task self-recording raises on-task behaviour without changing work completed or accuracy, behaviour monitoring should be treated as classroom management, not as a learning design.

Accuracy of judgement, on-task behaviour, immediate task performance, delayed retention, self-efficacy and continued use without prompts are separate claims. The present evidence establishes no checkpoint interval for learning tasks, no number of training sessions, no calibration threshold for hand-over and no durable transfer.

Claims

Examples

Key Sources

  • Bol, L., Campbell, K. D. Y., Perez, T., & Yen, C. J. (2016). The effects of self-regulated learning training on community college students' metacognition and achievement in developmental math courses. Community College Journal of Research and Practice, 40(6), 480-495. https://doi.org/10.1080/10668926.2015.1068718
  • Cho, K., Cho, M. H., & Hacker, D. J. (2010). Self-monitoring support for learning to write. Interactive Learning Environments, 18(2), 101-113. https://doi.org/10.1080/10494820802292386
  • Farrell, A., & McDougall, D. (2008). Self-monitoring of pace to improve math fluency of high school students with disabilities. Behavior Analysis in Practice, 1(2), 26-35. https://doi.org/10.1007/bf03391725
  • Ghanizadeh, A. (2017). The interplay between reflective thinking, critical thinking, self-monitoring, and academic achievement in higher education. Higher Education, 74(1), 101-114. https://doi.org/10.1007/s10734-016-0031-y
Page metadata
Field Value
type principle
id self-monitoring
title Self-monitoring
description For a learner with a readable criterion and a next move for each result, taught, test-based checks of their own work at set points during a task are expected to improve performance on it, but no claim here isolates self-monitoring against its absence on a learning outcome.
status review
canonical True
generated by claude/unspecified on 2026-10-05
trust unverified — no one has confirmed this page's content
sources 4 — bol-2016, cho-2010, farrell-2008, ghanizadeh-2017

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