Your Course Evaluation Measures Satisfaction, Not Emotion. Pekrun Says That's a Problem

In brief

Reinhard Pekrun's control-value theory of achievement emotions holds that students continuously experience distinct emotions — enjoyment, hope, pride, anxiety, boredom, hopelessness — that arise from how much control they feel over their learning and how much they value it, and that these emotions causally shape effort, self-regulation and achievement. Standard course evaluations measure global satisfaction, which flattens this emotional texture into a single valence. Adding validated achievement-emotion items — grounded in Pekrun's theory and measured by the Achievement Emotions Questionnaire (AEQ) — gives a quality office an earlier, more diagnostic signal than "overall, how satisfied were you," because a bored-but-satisfied class and an anxious-but-satisfied class need very different responses.

What the research says

Pekrun (2006), in the Educational Psychology Review, set out the control-value theory: achievement emotions are triggered by two appraisals — subjective control over an activity or its outcome (can I do this? can I influence the result?) and the subjective value of that activity or outcome (does it matter to me?). Different combinations produce different emotions: high control plus high value yields enjoyment; low control plus high value yields anxiety or, if outcomes seem unavoidable, hopelessness; high control plus low value yields boredom. Critically, the theory is reciprocal and causal — emotions are not just after-effects of a course but inputs to learning, influencing attention, motivation, use of learning strategies and ultimately performance.

To measure these emotions, Pekrun, Goetz, Frenzel, Barchfeld & Perry (2011) developed and validated the Achievement Emotions Questionnaire (AEQ) in Contemporary Educational Psychology. The instrument contains 24 scales covering enjoyment, hope, pride, relief, anger, anxiety, shame, hopelessness and boredom, across three settings — being in class, studying, and taking tests. Validated on a sample of university students (N = 389), the scales showed good reliability, a confirmatory factor structure consistent with the theory, and the predicted relationships with students' control-value appraisals and with learning outcomes. The AEQ has since been adapted across subjects, languages and education levels, and control-value theory has become one of the most widely used frameworks in the psychology of learning; Pekrun and colleagues have more recently extended it toward a general theory of human emotions covering epistemic, social and existential emotions as well.

The evaluation-relevant implication is that valence is not enough. Two emotions with the same negative valence — anxiety and boredom — have opposite appraisal origins (too little control versus too little value) and call for opposite remedies (build competence and lower threat versus increase relevance and challenge). A satisfaction score, or even a positive/negative sentiment split of open text, cannot distinguish them. Discrete-emotion measurement can.

Why it matters for course evaluation in practice

Almost every institutional course-evaluation instrument is built around satisfaction and perceived quality: "the teaching was good," "I would recommend this course," "overall satisfaction." These items capture a summary judgement but discard the emotional information that predicts whether students actually engaged and learned. A course can score a respectable 3.8 for satisfaction while quietly boring most of the class — a state that control-value theory links to disengagement and shallow processing — or while making a substantial minority anxious to the point of avoidance. Neither pattern is visible in the satisfaction number, yet each has a clear pedagogical fix that a QA process could act on.

Measuring achievement emotions also gives evaluators a diagnostic map rather than a verdict. Because each emotion is theoretically tied to control and value appraisals, an emotion profile points toward a cause: pervasive boredom suggests a value/challenge problem (content feels irrelevant or unchallenging); pervasive anxiety suggests a control problem (workload, assessment design or pacing exceed students' sense of capability); hopelessness is an urgent signal that students have stopped believing effort will pay off. This is far more actionable than "assessment scored low," and it connects the evaluation directly to the mechanisms that teaching enhancement actually targets.

Finally, emotions are a leading indicator. Because control-value theory treats emotions as causes of subsequent effort and achievement, capturing them mid-course — not just at the end — offers an early-warning system: rising boredom or anxiety can be addressed while the cohort is still enrolled, rather than diagnosed retrospectively after grades and satisfaction have already been depressed.

Limitations and honest caveats

Several caveats deserve to be raised before an institution rebuilds its survey around emotions. Self-reported emotion is still self-report, subject to recall bias, mood at the moment of completion, and the peak-end distortions that affect all retrospective evaluation; asking students to remember how anxious they felt weeks ago is not the same as sampling the emotion in situ. Length and burden. The full AEQ is long; bolting all 24 scales onto an end-of-term survey is impractical and would worsen response rates, so institutions must use validated short forms or a targeted subset, accepting some loss of coverage. Construct boundaries. Emotions correlate with, but are not identical to, motivation, workload perceptions and satisfaction; careful item design is needed to avoid simply relabelling existing constructs. Cultural and linguistic equivalence cannot be assumed — emotion terms and display norms differ across languages and cultures, so a translated emotion scale needs the same measurement-invariance checks as any cross-cultural instrument. Correlation is not licence for blame. Finding that a course elicits anxiety does not establish that the teaching caused it; prior preparation, concurrent courses and personal circumstances all contribute, and emotion data should inform inquiry rather than settle attribution. And measuring emotion does not by itself improve it — the value comes only if the institution acts on the profile.

How Koji incorporates this

Koji is designed to capture the emotional texture that a satisfaction survey discards, and to do so without the length penalty of administering a full emotions battery.

• Discrete-emotion probing, conversationally. Rather than 24 fixed scales, Koji's AI-moderated conversational interview can ask how a course felt and then follow up to disambiguate valence into specific emotions — distinguishing "this was boring" (a value/challenge signal) from "this made me anxious" (a control signal) through targeted open_ended probes, the exact distinction control-value theory says matters.
• Structured emotion items where you want them. For teams that prefer validated scale items, Koji supports scale and single_choice questions, so a short, theory-grounded emotion subset (for example enjoyment, boredom and anxiety) can sit alongside the conversational probes.
• Automatic thematic analysis mapped to appraisals. Koji's thematic analysis of open text can surface why an emotion is present — linking expressed boredom or anxiety back to the control and value appraisals (workload, relevance, pacing, assessment) that theory predicts — turning an emotion into a diagnosable cause.
• Mid-cycle and formative collection. Because emotions are leading indicators, Koji's support for mid-cycle collection lets a programme sample achievement emotions while the cohort is still enrolled, enabling response before the term ends rather than a post-mortem.
• Bias-aware, no-overclaim reporting. Koji frames emotion findings as diagnostic signals to investigate, not as proof that the teaching caused the emotion, keeping attribution appropriately cautious.

These mechanisms are designed to mitigate the emotional blind spot in conventional evaluation, not to reproduce a full validated administration of the AEQ, which remains the gold standard for formal research. Koji's core research platform at koji.so applies the same AI-moderated interview engine to product and customer research, where surfacing the emotion behind a satisfaction score is an equally common need.

Frequently asked questions

Isn't a satisfaction score already capturing how students feel? Only its overall valence. Satisfaction cannot distinguish boredom from anxiety, which share a negative sign but arise from opposite appraisals — too little value versus too little control — and require opposite remedies. Discrete-emotion measurement makes that distinction.

What are the core achievement emotions I should consider measuring? Pekrun's framework covers enjoyment, hope, pride, relief, anger, anxiety, shame, hopelessness and boredom. For a lean course survey, enjoyment, boredom and anxiety already span the key control-value combinations and are the most diagnostically useful.

Why do control and value matter for a QA office? Because they point to the cause. Boredom signals a value or challenge problem; anxiety and hopelessness signal a control problem (workload, pacing, assessment). That converts an emotion reading into a specific teaching-enhancement target.

Can I just run sentiment analysis on open-text comments instead? Sentiment gives you positive/negative, which is valence again. It does not separate the distinct negative emotions that matter. Aspect- and emotion-aware analysis, or direct probing, is needed to recover the discrete emotions.

Does capturing emotion improve teaching by itself? No. Like any evaluation data, it only helps if acted upon. Its advantage is that the emotion profile points more directly at a cause than a satisfaction mean does, making the required action clearer.

Related resources

• Students Rate the Classes Where They Learn the Most Lower: The Feeling-of-Learning Gap
• The Warmth Halo: How Instructor Immediacy Shapes Course Evaluations
• Is Your Course Pushing Students Toward Deep or Surface Learning? The R-SPQ-2F
• What Can Open-Text Student Comments Tell You That Likert Scores Cannot?
• It's Not Just Quality — It's Met Expectations: Expectancy-Disconfirmation
• Teacher Clarity Predicts Learning Better Than Charisma

References

• Pekrun, R. (2006). The control-value theory of achievement emotions: Assumptions, corollaries, and implications for educational research and practice. Educational Psychology Review, 18(4), 315-341. https://doi.org/10.1007/s10648-006-9029-9
• Pekrun, R., Goetz, T., Frenzel, A. C., Barchfeld, P., & Perry, R. P. (2011). Measuring emotions in students' learning and performance: The Achievement Emotions Questionnaire (AEQ). Contemporary Educational Psychology, 36(1), 36-48. https://doi.org/10.1016/j.cedpsych.2010.10.002
• Pekrun, R. (2024). Control-Value Theory: From Achievement Emotion to a General Theory of Human Emotions. Educational Psychology Review, 36, 83. https://doi.org/10.1007/s10648-024-09909-7