Adaptive Tutoring Modalities for Supporting Learners’ Reflective Writing Practices

Paper Title

Adaptive Tutoring Modalities for Supporting Learners’ Reflective Writing Practices

Publication Info

• Topic area: Adaptive tutoring systems for reflective writing in education
• Keywords: Reflective writing, conversational tutoring, adaptive feedback, Gibbs’ cycle, metacognitive skills, AI in education, longitudinal study, HCI, learning technologies, LLM-based tutoring

Background and Problem

• Problem / challenge: Many learners struggle with structured reflective writing, often producing unstructured, descriptive texts that lack critical analysis. Traditional methods of providing personalized guidance and feedback are resource-intensive and not scalable for large classes.
• Significance: Reflective writing is critical for developing metacognitive skills, self-regulated learning, and deeper engagement with learning experiences. Addressing these challenges can improve educational outcomes and foster lifelong learning habits.
• Motivation and related work: Prior research has explored pre-scripted conversational agents and feedback systems for reflective writing, but these often lack adaptivity and flexibility. Few studies have investigated the longitudinal impact of conversational tutoring in reflective writing, leaving a gap in understanding its effectiveness over time.

Solution

• Proposed approach: MindBuddy, an adaptive tutoring tool combining conversational guidance and adaptive feedback, designed to support reflective writing based on Gibbs’ reflective cycle.
• Novelty:
  1. Development of a hybrid conversational and feedback-based tutoring system grounded in the Gibbs’ cycle.
  2. Iterative design process informed by two pilot studies and user feedback.
  3. Longitudinal evaluation of interactive conversational tutoring versus feedback-only modalities.
  4. Introduction of four design principles for adaptive conversational tutoring tools.
• Procedure and key techniques:
  • MindBuddy provides conversational guidance to introduce Gibbs’ cycle and a text editor for free-form writing.
  • Adaptive feedback is delivered iteratively, evaluating texts against the Gibbs’ cycle components.
  • Two pilot studies refined the system, addressing user needs and technical limitations.
  • A six-week field experiment compared two versions of MindBuddy: conversational tutoring (TG1) and feedback-only (TG2).

Results

• Concrete findings:
  • TG2 (feedback-only) achieved significantly higher scores in formal reflective structure at post-test (p = .025, d = .83).
  • Both groups showed similar perceived skill increases (TG1 = 70%, TG2 = 71%).
  • No significant differences in reflective writing quality or knowledge test scores were observed.
• Advantage over baselines: TG2 showed marginally better outcomes in reflective structure and quality, while TG1 facilitated greater engagement and ease of starting reflective writing.
• Experiments / evaluation:
  • Participants: 34 undergraduate students over six weeks.
  • Metrics: Reflective writing structure (Gibbs’ cycle), quality, perceived skills, and knowledge test scores.
  • Methods: Pre- and post-surveys, text annotations, and server log analysis.
• Limitations and future work:
  • Small sample size (n=34) limits statistical power.
  • Ceiling effects observed due to participants’ strong baseline reflective competence.
  • Study focused on a single reflective framework (Gibbs’ cycle) and a specific student demographic.
  • Future work should explore diverse frameworks, larger and more diverse samples, and multilingual learners.

Summary

This study introduces MindBuddy, an adaptive tutoring tool combining conversational guidance and feedback to support reflective writing. A six-week field experiment with 34 students compared conversational tutoring (TG1) and feedback-only (TG2) modalities. TG2 showed better outcomes in reflective structure, while TG1 facilitated engagement and ease of starting reflections. Both groups demonstrated similar perceived skill improvements. The study highlights the potential of adaptive feedback for enhancing reflective writing and provides design principles for developing conversational tutoring systems. Future research should address scalability, diverse learner needs, and alternative reflective frameworks.