The Development of Reflective Practice on a Work-Based Software Engineering Program: A Longitudinal Study

Technical Summary: The Development of Reflective Practice on a Work-Based Software Engineering Program

Problem Statement
While reflective practice is recognized as a critical component of experiential and work-based learning (WBL), there is a scarcity of long-term empirical studies evaluating its development within Software Engineering (SE) education. Existing literature suggests that reflection is essential for professional growth and continuous learning in SE, yet it is often difficult to teach, intimidating for students, and frequently absent in practice (e.g., within agile retrospectives). Furthermore, traditional SE degree programs often lack the authentic contexts required for deep reflection, and few studies track the evolution of reflective abilities over the duration of a degree. This study addresses the gap by examining how reflective practice develops longitudinally among students in a four-year, work-based Software Engineering apprenticeship program.

Methodology
The study employs a qualitative, longitudinal thematic analysis of reflective reports submitted by a single cohort of undergraduate students over four years. The program is an apprenticeship model where students spend approximately 80% of their time in the workplace and 20% on campus.

• Data Collection: The dataset consists of reflective reports from Year 1 to Year 4. In Years 1 and 2, students submit 750-word reports; in Years 3 and 4, they submit 1,500–2,000-word reports. The sample sizes vary slightly due to attrition and transfers (Year 1: n=36; Year 2: n=38; Year 3: n=38; Year 4: n=37).
• Analytical Framework: The analysis utilizes two established models of reflection as a priori coding schemes:
  1. Boud et al.'s Model of Reflective Process: Six categories (Attending to feelings, Association, Integration, Validation, Appropriation, Outcome of reflection).
  2. Bain et al.'s 5R Framework: A progressive five-stage scale (Reporting, Responding, Relating, Reasoning, Reconstructing).
• Coding Procedure: One author coded the entire dataset against both models. To refine the analysis for the 5R model, a "minimal" evidence category was introduced to capture partial demonstrations of a stage. A second and third author analyzed subsets of the data to ensure trustworthiness, with disagreements resolved through discussion. Ethical approval was obtained, and student anonymity was maintained via a coding system (S#/Y#).

Key Results
The longitudinal analysis reveals a clear trajectory of increasing sophistication in students' reflective abilities, with specific trends observed for each model:

• Boud et al.'s Model:

  • Increasing Prevalence: The average number of reflective elements per student increased from a mean of 2 in Year 1 to 5 in Years 3 and 4.
  • Association & Integration: These elements showed marked increases, reaching 100% prevalence in Years 3 and 4. Students increasingly linked university theory to workplace reality and synthesized new insights with existing knowledge.
  • Appropriation & Outcome: These "higher" forms of reflection, involving the internalization of skills and changes in future behavior, became significantly more prevalent in later years.
  • Validation: Contrary to other elements, the prevalence of Validation (assessing the authenticity of ideas) decreased over time. The authors suggest this may indicate students becoming more accepting of their apprenticeship context and less critical of it.
  • Attending to Feelings: While prevalent in later years (approx. 60%), it did not reach the same levels as Association or Integration. Early reports were dominated by negative emotions (fear, self-doubt), which shifted toward positive reflections on confidence and perspective in later years.

• Bain et al.'s 5R Framework:

  • Reporting & Responding: These foundational stages (describing events and emotional/cognitive reactions) were present in every report across all four years.
  • Relating: Strong evidence of connecting experience to prior knowledge existed from the start but became most robust in Years 3 and 4, where students explicitly linked university concepts (e.g., Scrum, design patterns) to workplace application.
  • Reasoning: Evidence of reasoning about experiences increased notably between Years 1–2 and 2–3. Early reasoning focused on solution effectiveness, while later reasoning addressed the complementary nature of university and workplace learning, including meta-reflection on the learning process itself.
  • Reconstructing: This highest level of reflection (drawing conclusions and planning future actions) was absent or minimal in Year 1. By Year 2, students began outlining lessons learned; by Year 3, over half demonstrated this stage; and by Year 4, all students exhibited future planning based on their accumulated experience.

Key Contributions

1. Longitudinal Evidence: The study provides rare empirical data tracking the development of reflective practice in SE education over a full four-year degree cycle, moving beyond cross-sectional snapshots.
2. Dual-Model Analysis: By applying both Boud et al.'s and Bain et al.'s models, the study validates findings across different theoretical frameworks, showing consistent patterns of growth (e.g., the parallel rise of "Outcome of Reflection" in Boud's model and "Reconstructing" in the 5R model).
3. Validation of WBL: The results highlight the complementary nature of workplace experience and university study, demonstrating that the work-based approach effectively fosters the transition from tacit to explicit knowledge and the development of "higher-order" reflective skills.

Significance and Claims
The authors claim that the study confirms the expectation that work-based learning modules successfully cultivate reflective, lifelong learners. The findings suggest that students in this program progressively move from descriptive reporting to sophisticated integration, appropriation, and reconstruction of their learning.

The paper modestly notes that it is an observational analysis of existing reports, not an experimental study, and therefore makes no claims about causation. It acknowledges limitations, such as the lack of standardized assignment prompts across years and the difficulty in distinguishing between certain reflective elements. However, the authors conclude that the observed development of sophisticated reflection is a positive outcome of the program. They suggest that these findings offer insight into how reflective practice evolves in SE and propose that traditional, non-work-based degree programs might consider incorporating more structured reflection to achieve similar developmental benefits. The study does not propose new experimental interventions but rather offers evidence to support the pedagogical value of the existing work-based model.