Does Your Course Build Self-Regulated Learners? Metacognition and SRL as an Evaluation Lens

In brief: Self-regulated learning (SRL) — the cyclical process of setting goals, monitoring progress, and reflecting on outcomes — is one of the most robust predictors of academic achievement, and specific strategies such as time management, metacognition and effort regulation carry the strongest evidence (Zimmerman, 2002; Panadero, 2017; Broadbent & Poon, 2015). Most course evaluations ask whether students were satisfied; almost none ask whether the course taught them to learn. Adding a small set of SRL-oriented items reframes evaluation around whether a course built durable learning capacity — but SRL is measured by self-report and is a course-design outcome, not an instructor charisma score.

The question this answers

A course can be enjoyable, clearly delivered, and highly rated, and still leave students no better at managing their own learning than they were in week one. Conversely, a demanding course that forces students to plan, self-test, and reflect may feel harder — and score lower on satisfaction — while building exactly the capacity higher education is supposed to develop. This is the blind spot this article addresses: whether course evaluation should ask not "were you satisfied?" but "did this course make you a more capable, self-directed learner?"

What the research says

SRL is a well-specified, cyclical process. Barry Zimmerman's overview (2002, Theory Into Practice, 41(2), 64–70) describes self-regulation as three cyclical phases: forethought (goal setting and strategic planning), performance (self-control and self-observation), and self-reflection (self-evaluation and adaptive inference). The central claim is that self-regulation is not a fixed aptitude but a set of learnable processes — students can be taught to set goals, monitor comprehension, and adjust strategy. This matters for evaluation because it makes SRL a course outcome a teacher can influence, not merely a student trait.

The construct is mature and consolidated. Ernesto Panadero's review (2017, Frontiers in Psychology, 8:422) compares six major models of SRL — Zimmerman; Boekaerts; Winne and Hadwin; Pintrich; Efklides; and Hadwin, Järvelä and Miller — and shows substantial convergence: SRL integrates cognitive, metacognitive, motivational, and emotional components. The consolidation is important for a sceptical reader: SRL is not one lab's pet theory but a broad, cross-validated framework.

Specific SRL strategies predict achievement — but not equally. Broadbent and Poon's systematic review (2015, The Internet and Higher Education, 27, 1–13) synthesised studies of SRL strategies and academic achievement in online higher education. They found time management, metacognition, effort regulation, and critical thinking positively correlated with academic outcomes, while rehearsal, elaboration, and organisation had weaker empirical support; peer learning showed a notably strong association. They also caution that effects appear weaker online than in face-to-face settings — a warning against assuming SRL items transfer cleanly across modalities.

The synthesis across these sources: SRL is learnable, well-defined, and predictive of achievement, and the strategies that matter most (planning, self-monitoring, effort regulation) are precisely the ones a well-designed course can scaffold — and therefore the ones an evaluation could sensibly ask about.

Why it matters for course evaluation in practice

It reframes "quality" around durable capacity, not momentary satisfaction. A course evaluation dominated by satisfaction items rewards fluency and comfort — the very things the feeling-of-learning gap and desirable difficulties literatures warn can be inversely related to actual learning. SRL items ask a different question: did the course require and support planning, self-testing, and reflection? A course that scaffolds SRL may feel effortful precisely because it is doing its job.

It gives teachers a diagnostic, improvable signal. "Was the lecturer clear?" tells a teacher little about what to change. "Did this course help you monitor whether you actually understood the material?" points directly at a design lever — retrieval practice, low-stakes quizzing, reflective prompts. SRL items convert evaluation from a verdict into a design brief, which is the whole point of formative evaluation.

It complements, rather than duplicates, existing lenses. SRL sits alongside constructs the knowledge base already covers: cognitive load (whether the course managed working-memory demands — see cognitive load theory), deep-versus-surface approaches (see R-SPQ-2F), and self-reported learning gains (see SALG). SRL's distinctive contribution is the process framing — planning, monitoring, adapting — that these other lenses touch only partially.

Limitations and honest caveats

Self-report measures process imperfectly. SRL is typically measured by questionnaire (the MSLQ tradition and its descendants), and students are being asked to report on their own metacognition — a notoriously difficult thing to observe from the inside. Weak self-regulators may be precisely the students least able to accurately report their self-regulation, producing a calibration problem. Trace data (log files, self-testing behaviour) partially addresses this but is out of scope for most evaluations; see our note on triangulating with learning-analytics data.

Attribution is genuinely hard. If a student reports strong self-regulation at the end of a course, how much did the course cause it versus the student arriving already self-regulated? Broadbent and Poon's correlational designs cannot separate these, and much SRL-outcome evidence is cross-sectional. A retrospective or pre-post design (with its own response-shift problems) is needed to make even a weak causal claim.

Modality and discipline dependence. Broadbent and Poon found strategy-achievement links were weaker online, and SRL demands differ sharply between a lab-based science course and a reading-intensive humanities seminar. A single SRL item set may not be measurement-invariant across these contexts, so cross-course comparison of SRL scores should be treated with the same caution as any cross-group comparison.

It is not an instructor-performance metric. As with belonging, SRL outcomes reflect course design, assessment structure, and student characteristics as much as any individual's teaching. Use SRL items to improve courses, not to rank staff.

How Koji incorporates this

Koji for Education can operationalise SRL as an evaluation lens without bolting a 40-item MSLQ onto every survey.

• Targeted SRL items mapped to the phases. A short set of scale items can probe forethought ("This course helped me plan how to approach my learning"), performance-phase monitoring ("I regularly checked whether I actually understood the material"), and self-reflection ("The course prompted me to reflect on what was and was not working"). Keeping the set short respects questionnaire-length and satisficing concerns documented elsewhere in this knowledge base.
• Conversational probing of strategy. Because self-report of metacognition is fragile, Koji's AI-moderated interview can ask behavioural follow-ups — "What did you actually do when you got stuck?" — which elicit concrete strategy use rather than an abstract self-rating. This is designed to mitigate the calibration problem: what a student did is more verifiable than how self-regulated they feel.
• Thematic analysis of learning-strategy language. Automatic theming can quantify how often students describe planning, self-testing, or reflection, turning open text into a trackable SRL signal across cohorts and terms.
• Formative timing. Because SRL is a process a course can scaffold during the term, Koji supports mid-cycle collection so a teacher can adjust — adding retrieval practice or reflective checkpoints — before the course ends.
• Triangulation, not overclaiming. Koji is designed to surface SRL as perceived and reported strategy use; it does not claim to measure metacognition objectively, and pairs well with LMS trace data where an institution wants a behavioural check. Koji's core research platform at koji.so applies the same probing interview engine to customer and product research, where "what did you actually do?" is the same discipline in a different setting.

Framed honestly: SRL items are designed to reorient evaluation toward durable learning capacity and to surface strategy use — not to certify that a course produced self-regulated learners.

A worked example: the same lens, two very different courses

The value of an SRL lens is clearest when you contrast disciplines. In a problem-based science module, the self-regulation that matters is largely cognitive and behavioural: did students plan a route through a multi-step problem, monitor whether their working was sound, and regulate effort across a long lab report? Sensible items ask about planning under complexity and about noticing when an approach was not working. In a reading-intensive humanities seminar, the salient self-regulation is metacognitive and motivational: did students judge which readings deserved depth, monitor whether they had genuinely understood an argument rather than merely recognised it, and sustain engagement across a term without weekly tests forcing the pace?

The point is not to build two separate instruments but to recognise that a single generic SRL item ("This course helped me regulate my learning") will mean different things in each context and will not be measurement-invariant across them. A short, phase-anchored core (plan / monitor / reflect) plus one discipline-sensitive open-text probe travels better than a long borrowed battery. Reported strategy use should therefore be read within a course over time — is this cohort reporting more planning and self-testing than last year? — rather than ranked across dissimilar courses, where the construct itself shifts. That within-course, longitudinal reading is also where an SRL signal is most actionable: it points a teacher toward specific scaffolds (retrieval practice, planning templates, reflective checkpoints) that can be added and then re-measured.

Related Resources

• What Cognitive Load Theory Says Your Evaluation Should Ask
• Deep or Surface Learning? The R-SPQ-2F as an Evaluation Lens
• Desirable Difficulties: Why Better Teaching Can Lower Satisfaction
• The Feeling-of-Learning Gap
• SALG: Can Students Reliably Report Their Own Learning Gains?
• Triangulating Evaluations with LMS Learning-Analytics Data

References

• Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory Into Practice, 41(2), 64–70. https://doi.org/10.1207/s15430421tip4102_2
• Panadero, E. (2017). A review of self-regulated learning: Six models and four directions for research. Frontiers in Psychology, 8, 422. https://doi.org/10.3389/fpsyg.2017.00422
• Broadbent, J., & Poon, W. L. (2015). Self-regulated learning strategies & academic achievement in online higher education learning environments: A systematic review. The Internet and Higher Education, 27, 1–13. https://doi.org/10.1016/j.iheduc.2015.04.007
• Pintrich, P. R. (2004). A conceptual framework for assessing motivation and self-regulated learning in college students. Educational Psychology Review, 16(4), 385–407. https://doi.org/10.1007/s10648-004-0006-x