Agentic education · Learning science

Actionable companion · Self-regulation & metacognition

Learning with agency: From Predictors to Mechanisms: Interpretable Artificial Intelligence Evidence on Mathematics Achievement and Cognitive Learning Systems

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paperSearch paper agentCC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Age / stage not established. Verify the participant population and task before transferring the findings.

Original authors’ abstract, reproduced under the linked source license:

Abstract Understanding academic achievement requires moving beyond the identification of influential factors toward explaining how these factors are organized into functional learning and cognitive mechanisms. Although prior research has extensively documented the roles of socioeconomic status, student attitudes, and learning behaviors, less attention has been paid to how these elements interact within structured pathways that reflect underlying learning intelligence across educational systems. This study adopts a mechanism-oriented perspective to examine mathematics achievement using data from PISA 2018. Focusing on high-performing regions in East Asia and Western countries, it integrates interpretable artificial intelligence methods with structural modeling to investigate how contextual, psychological, and learning-process factors jointly shape achievement outcomes. The findings show that high achievement is not governed by a single set of dominant predictors, but by distinct organizational mechanisms of learning intelligence. In East Asian systems, achievement follows a chain-like convergent structure, in which socioeconomic background is systematically translated into academic outcomes through sequential psychological and self-regulatory processes. Psychological factors, particularly educational expectations and self-beliefs, function as key mediating mechanisms that organize learning engagement and strategy use. By contrast, high-performing systems in Europe and North America exhibit a parallel configuration, in which multiple cognitive and behavioral factors independently contribute to achievement through more decentralized pathways, reflecting a distributed structure of learning intelligence. Across regions, learning processes such as reading engagement and digital literacy show consistently positive associations with achievement. However, their roles vary depending on how they are embedded within broader system-level structures. These results suggest that self-regulation operates not merely as an associated factor, but as an organizing mechanism of learning intelligence that structures the translation of background resources into performance. By reconceptualizing prediction as a means of revealing the organization of learning intelligence, this study proposes a unified analytical framework that links interpretable artificial intelligence with theory-driven explanation. The findings contribute to a deeper understanding of how achievement systems function and highlight that high performance can emerge through multiple, structurally distinct pathways, with important implications for educational research, cognitive theory, and policy design. Keywords: mathematical achievement, learning intelligence, interpretable artificial intelligence, cognitive mechanisms, self-regulated learning, PISA, cross-regional comparison

What kind of evidence is this?

Supporting AI learning science · Design requires full-text appraisal

Stage: Age / stage not established. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Limitations and uncertainty · source passage

First, this study is based on cross-sectional PISA 2018 data. Therefore, the analysis cannot establish temporal ordering or causal direction with certainty. The findings should be interpreted as associations and organizational patterns consistent with the proposed analytical framework, rather than as evidence of strict causal mechanisms. Future research using longitudinal data would be valuable for examining how these pathway structures develop over time.

Design the system before the pilot

Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.

  • Use plan–attempt–check–reflect prompts and compare confidence with actual performance. Keep learner choice and reflection visible.
  • Limit simultaneous instructions and unnecessary dialogue. Compare task performance and a suitable mental-effort measure; easier-feeling work is not automatically better learning.
  • Measure engagement separately from achievement. Check whether novelty-driven participation persists and whether it translates to independent performance.

A 90-day district pilot

  1. Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
  2. Days 16–30 · Design. Treat the paper as background evidence until its population is verified in the full text. Do not assign an age-specific intervention or extend adult results to children. Name an accountable owner and set the baseline, success threshold and budget before implementation.
  3. Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
  4. Days 61–90 · Decide. Extract the participant ages or education stage, recruitment, task and comparator before selecting a local pilot. If these cannot be established, retain the paper for design discussion rather than age-band efficacy claims. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Danyang Meng; Alan T K Wan. From Predictors to Mechanisms: Interpretable Artificial Intelligence Evidence on Mathematics Achievement and Cognitive Learning Systems. Journal of Intelligence (2026-05-25). https://doi.org/10.3390/jintelligence14060091

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.