Self-Assessment¶
Element · All elements
Evidence · 65 claims (14 for, 38 mixed, 13 against) · 53 studies (26 causal, 11 associational, 6 review, 5 theoretical, 3 quant-synthesis, 1 qualitative, 1 design),q1–q4· 19 of 53 report an effect size · 61 claims rest on one study
Description¶
Self-assessment is the element in which learners judge their own work, understanding, or progress against explicit criteria. It is useful when the goal is not only better immediate performance, but stronger self-monitoring and more accurate judgments about what to improve next.
Design Implications¶
Context¶
Requirements¶
- Visible criteria: Learners need examples, rubrics, or success indicators to compare against.
- Reflection prompt: The task should ask learners to identify strengths, gaps, or next steps.
- Low-stakes use: Self-assessment works best when honesty is safer than impression management.
Constraints¶
- Weak calibration: Novices may misjudge quality without models or coaching.
- Performative reflection: Generic prompts can produce superficial responses.
Target Learners¶
- Learners developing self-regulation, revision habits, and metacognitive accuracy.
Target Learning Goals¶
- Improve monitoring, revision, and ownership of progress.
Affordances¶
Claims¶
- Prior Knowledge Needed For Accurate Self Assessment [+W]
- Self Assessment Improves Self Regulated Learning [+M]
- Students reported increased personal responsibility and described the reflective self-assessment process as confidence-building during their transition to university [+W] — attached 2026-10-11 from Pedlow et al. (2026), which proposed "GenAI-focused reflective self-assessment instrument for commencing university students".
- In one 2×2 experiment with 120 undergraduates, an Assessment-Performance Calibration Activity (APCA) gave the most accurate self-assessment and least overconfidence, and adding a feedback-literacy script did not improve calibration further [+W] — attached 2026-10-11 from Dai (2026), which proposed "APCA Assessment-Performance Calibration Activity"; tests this page's relationship.
- Guided reflection embedded in self-assessment supports commencing students to engage responsibly and confidently with GenAI while maintaining academic integrity [+W]
- A theoretical article, citing two simulation studies, reports that novice clinical learners overestimated their performance and that their confidence exceeded their accuracy [-W]
- Workflow condition is associated with students' self-assessment confidence [~W]
- Students' JoL ratings became more consistent over time and evolved from under-rating to over-rating to accurate judgement [~W]
- Students using the eliciting agent showed JoL ratings correlated with system learning metrics, strongest for mastery [+W]
- Adding diagram self-assessment to diagram completion does not increase diagnostic cue utilization or metacomprehension accuracy beyond diagramming alone [~W] — attached 2026-10-19 from Pijeira-Díaz et al. (2023), which proposed "Diagram self-assessment scaffold (prompting students to estimate completed and correct diagram boxes)"; tests this page's relationship.
- Prompted self-assessment decreases utilization of the low-diagnosticity commission-errors cue [-W] — attached 2026-10-19 from Pijeira-Díaz et al. (2023), which proposed "Diagram self-assessment scaffold (prompting students to estimate completed and correct diagram boxes)".
- Eliciting confidence ratings impairs rule-based category learning compared to both no-rating and judgment-of-learning conditions on ambiguous transfer items [-W]
- Bias scores did not improve over the school year, but school year level mattered: year 1 students became less overconfident while year 3 students improved their underconfidence [~W]
- Calibration (judgment accuracy) measures such as the Goodman-Kruskall gamma coefficient and signal detection theory's d' can be empirically compared, and calibration may vary across knowledge domains [~W]
- Correctness Confidence did not significantly predict recall accuracy, contrary to the authors' hypothesis [-W]
- Generating examples produces overconfident judgments of learning, especially for principles and procedures [-W]
- The Monitoring Exercise protected against overconfidence in the second exam score prediction but did not affect exam scores [~W]
- Paired flashcard learners show more accurate immediate metacognitive calibration than individual learners, who overestimate by roughly 20% [+W]
- Learners are substantially overconfident when evaluating the quality of their self-generated examples [-W]
- Response-time effects are limited to slower responding in the JOL condition during the first two training blocks [~W]
- In Experiment 2, neither self-assessment feedback plus correct answers nor plus contrasting own with correct answers improved test-phase self-assessment accuracy [-W]
- Providing a rubric during self-assessment improves the absolute accuracy of high school students' self-assessments of written scientific abstracts [+W]
- Calibration accuracy improvements on French did not transfer to German and Math courses without the intervention, arguing against a maturation explanation [+W]
- JOL resolution (relative accuracy) did not differ between younger and older adults [~W]
- Reactivity differs by rating scope: item-focused confidence ratings impaired rule learning while task-focused judgments of learning did not [~W]
- Self-assessments of diagram boxes are more accurate than comprehension judgments and strongly correlated with them in the diagramming paradigm [~W]
- Students tend to rely on non-diagnostic cues when judging their own performance, resulting in inaccurate judgments [~W]
- After self-scoring, the monitoring-accuracy gap between low- and high-performing students remains on multiplication (small) but disappears on division [~W]
- Scaffolded drawing showed no significant advantage over drawing from scratch for absolute monitoring accuracy [~W]
- Learners accurately judge that twice-presented items are more likely to be recalled than once-presented items [+W]
- The Bias Index significantly predicts all four WIAT-II achievement measures, with over-confidence associated with lower achievement [~W]
- Student teachers' self-assessed preparedness did not predict first-year effectiveness (r = 0.03), while mentor teachers' ratings of student teachers did (r = 0.24) [-W]
- Explicit metacognition has severely limited access to underlying cognitive processes, and metacognitive reports can be inaccurate [-W]
- Students' difficulty rankings of equation types partly mismatch their measured performance, though metacognitive accuracy improves with age [~W]
- JoL–mastery accuracy trajectories over the five weeks rose in the elicit condition, declined in the none condition, and were mixed in the tell condition [~W]
- Higher-achieving students tend to be more accurate but more underconfident in calibration than lower-achieving students [~W]
- Group settings and guidelines each promoted calibration accuracy and achievement in a high school biology factorial experiment [+W]
- Hard-easy effect: students are more accurate but underconfident on easy items and less accurate but overconfident on difficult items [~W]
- Incentives significantly improved calibration accuracy only among lower-achieving students in a college-course quasi-experiment [~W]
- Negative example standards do not reliably improve judgment accuracy and significantly decrease relative accuracy in evaluating self-generated examples [-W]
- Self-assessed creative performance and self-assessed idea creativity are moderately correlated but empirically distinct judgments [~W]
- The reduction in perceived difficulty does not mediate the benefits of rubrics on self-assessment accuracy [~W]
- Postdictions of performance are typically more accurate than predictions [+W]
- Self-assessment accuracy in PBL depends on achievement level: low achievers graded generously, high achievers accurately [~W]
- Young and older adults make accurate retrospective confidence judgments but not prospective ease-of-selection judgments about their arithmetic strategy selections [-W]
- Under sequential presentation with Likert ratings for every word, no memory benefit emerged for items judged likely to be forgotten; negative ratings predicted the poorest recall [-W]
- Tell and no-agent conditions showed no significant JoL–confidence correlations, with mixed results for progress and mastery [~W]
- Text topic moderates judgment bias: the CLT text produced higher overconfidence for application performance than the CVT text [~W]
- Children evaluated their own performance as worse when they were slower in Grades 3 and 4 [~W]
- Verbalized self-reported confidence is overconfident in the mid-confidence range, while consistency-based confidence shows good average calibration but localized failure regions [~W]
- Effort-rating item wording changes what learners reflect on: requirement wording correlates negatively and engagement wording positively with judgments of learning [~W]
- The relationship between invested mental effort and self-assessed understanding is weak and inconsistent, with most correlations non-significant and mixed directions [-W]
- The tribute reports students tend to overestimate their sport performance, and four-level model instruction may improve calibration accuracy [+W]
- Experiment 2 replicated the forward effect of cue-only JOLs over restudy with rock categories, and item-by-item feedback did not change the effect [~M]
- Prompted cue type and timing of cue prompt do not interact in their effect on self-assessment accuracy (Experiment 1) [~W]
- Mental effort correlates negatively with self-assessments regardless of whether learners receive up-front cue information [~W]
- Providing a standard answer to compare against improves the accuracy of students' retrospective monitoring judgments [+W]
- Boundary conditions: explicit relation-focused instructions and no access to products during judgment are needed for mapping and drawing to improve monitoring accuracy [~W]
- Effects of gender and self-perceived metacomprehension ability on bias arose mainly from judgment magnitude, whereas WMC and topic knowledge also affected performance [~W]
- Significant learner characteristics explained 17% and 14% of variance in prediction bias and 11% and 12% in postdiction bias, so characteristics were slightly more relevant for predictions [~W]
- Situation-model-approach interventions show a medium-sized positive effect on metacomprehension accuracy and a small direct effect on comprehension [~W]
- In the CRT experiment, confidence appraisals predicted solving accuracy more strongly than goal-driven effort appraisals [~W]
- Oral explaining yields more accurate meta-comprehension judgments than written explaining and retrieval practice [~W]
- Students feel quite confident about their monitoring and regulation judgments even before self-scoring [~W]
The evidence by question¶
- Should learners assess their own work, and can their self-ratings be trusted? Self Assessment Improves Self Regulated Learning [+M]
- Can learners' own monitoring be trusted, and can it be trained? Self Assessment Accuracy Is Low Without Training [~S]
Related Elements¶
Patterns That Use This Element¶
Examples¶
- Learners score a draft against a rubric before submission.
- Students rate confidence and identify one misconception after a quiz.
- Future self-assessment interventions should redirect students' attention from invalid cues toward valid, predictive cues rather than providing self-assessment accuracy feedback
Key Sources¶
- Andrade, H. (2019). A critical review of research on student self-assessment. Frontiers in Education, 4, 87. https://doi.org/10.3389/feduc.2019.00087
- Panadero, E., Brown, G. T. L., & Strijbos, J.-W. (2016). The future of student self-assessment: A review of known unknowns and potential directions. Educational Psychology Review, 28(4), 803-830. https://doi.org/10.1007/s10648-015-9350-2
Page metadata
| Field | Value |
|---|---|
type |
element |
id |
self-assessment |
title |
Self-Assessment |
description |
Self-assessment is the element in which learners judge their own work, understanding, or progress against explicit criteria. |
status |
review |
canonical |
True |
generated |
by codex/unspecified on 2026-04-07 |
trust |
unverified — no one has confirmed this page's content |
sources |
2 — andrade-2019, panadero-2016 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.