Knowledge Building is an educational approach, pioneered within the field of computer-supported collaborative learning, in which classrooms operate as knowledge-creating communities. Rather than receiving information passively, students identify questions that matter to them, propose ideas, refine those ideas through discussion, and take collective responsibility for advancing the community’s understanding. In practice, this work is often supported by Knowledge Forum, an online platform where learners post notes, build on each other’s contributions, and connect fragments of ideas into progressively deeper explanations. Teachers design scaffolds — tools, prompts, and strategies — and gradually hand over epistemic agency to the students themselves.

Implementing Knowledge Building is not without friction. Prior studies have documented resistance among novice students, particularly those accustomed to individual, teacher-directed learning, and early confusion is a common response when learners confront unfamiliar pedagogy and complex subject matter. This is where emotions enter the picture. The study draws on control-value theory, which holds that students experience epistemic emotions — curiosity, enjoyment, surprise, confusion, frustration, anxiety, and boredom — when they confront needs or challenges tied to knowledge itself, such as detecting a misconception, weighing conflicting evidence, or coordinating competing viewpoints. Positive epistemic emotions have generally been linked to engagement, self-regulated learning, and achievement, while negative emotions tend to undermine them.

To investigate whether reflective assessment could shape these emotions and, in turn, the quality of collaborative discourse, the researchers conducted a quasi-experimental study with 113 undergraduates enrolled in an educational research course over sixteen weeks in a blended format. Students had no prior experience with Knowledge Building pedagogy or its technology. Two intact classes were compared: a control group of 61 students and an experimental group of 52. Both groups were taught by the same instructor, followed the same syllabus, used identical materials and the same Knowledge Forum environment, and attended equivalent sessions — a design that strengthens the ecological validity of the comparison.

The course unfolded through structured modules. Both groups learned the principles of Knowledge Building and the platform, identified significant questions related to action research, and cycled through phases of question decomposition, online inquiry, and pooling of findings in mini-conference-style presentations. Only the experimental group, however, completed a module of concurrent reflective assessment. In these sessions, students received a nominal report with visual analytics drawn from a prototype analytical dashboard, showing their individual and community progress against two key Knowledge Building principles: moving from egocentric contributions toward collective cognitive responsibility, and recognizing that all ideas are improvable. They then completed portfolio templates combining Likert-scale items and open self-assessment questions, asking themselves, in effect, how they were doing and how they could do better.

Emotional data came from a validated short form of the Epistemically-Related Emotion Scales, administered at weeks 6, 11, and 16, capturing self-reported surprise, enjoyment, curiosity, confusion, frustration, anxiety, and boredom. Discourse quality was assessed from 1,174 notes written on Knowledge Forum, each coded on four dimensions — questioning, theorizing, scientific sophistication, and use of authoritative sources — with two independent analysts reaching 93.79 percent inter-rater agreement. A composite quality-of-contributions score was computed for each student by averaging the four dimension means.

The correlational and structural equation modeling results were striking. Positive epistemic emotions correlated positively with contribution quality (r = 0.674), negative emotions correlated negatively (r = −0.637), and the two emotion clusters were themselves strongly inversely related (r = −0.748). In the structural model, positive emotions showed a significant positive effect on discourse quality (β = 0.49) while negative emotions exerted a significant negative effect (β = −0.26). Notably, positive emotions were the stronger predictor, suggesting that cultivating curiosity and enjoyment may matter more than merely suppressing frustration when the goal is richer collective knowledge work.

The comparative analyses revealed the catalytic role of reflection. Both groups improved over time, but the experimental group improved earlier and more consistently, with large effect sizes across all three outcome variables (for example, 0.439 for positive emotions and 0.207 for contribution quality), and significant differences in every pairwise time comparison. The control group showed significant gains only from the third measurement point onward. Between-group comparisons confirmed that the groups started equivalent, diverged significantly at weeks 11 and 16, and then converged toward the end of the course — a pattern the authors interpret as evidence that reflective assessment accelerates students’ adaptation to Knowledge Building, rather than producing permanently superior outcomes. Once both groups internalized the collaborative and self-regulatory routines, the gap naturally closed.

For educators, the practical implications are concrete. The authors recommend that instructors implementing Knowledge Building with novices schedule reflective sessions concurrently with inquiry, using analytics-based reports to help students see how their discourse is evolving and to identify concrete improvements — seeking better sources, asking for help, or engaging experts. The findings also suggest reflection can help neutralize negative emotions such as confusion by restoring students’ sense of control. The study’s authors caution that the sample was predominantly female, that prior academic achievement was not controlled, and that self-reported emotion measures may carry response biases, and they call for replication across contexts and educational stages, including secondary education where adolescents’ still-developing self-regulation may make reflective scaffolds even more crucial. Future work, they note, could embed emotion-expression tools directly into Knowledge Forum to capture emotional dynamics in real time.

Subject of Research: The effect of reflective assessment on epistemic emotions and discourse quality in Knowledge Building classrooms

Article Title: Reflective assessment in knowledge building classrooms: fostering positive epistemic emotions and quality discourse on educational research

Article References: Gutiérrez-Braojos, C., Zhu, G., Rodríguez-Domínguez, C., & Rodríguez-Chirino, P. (2026). Reflective assessment in knowledge building classrooms: fostering positive epistemic emotions and quality discourse on educational research. Journal of New Approaches in Educational Research, 15(1), Article 16. https://doi.org/10.1007/s44322-026-00061-4

Image Credits: AI Generated

DOI: 10.1007/s44322-026-00061-4

Keywords: Knowledge Building, reflective assessment, epistemic emotions, collaborative learning, Knowledge Forum, control-value theory, computer-supported collaborative learning, quality of discourse, higher education, educational research, Reflective, assessment