This study reports a practice-based inquiry into student-led interdisciplinary projects in secondary computer science education, using a lesson study (LS)–informed practitioner inquiry stance to examine how teacher mentoring sustains student agency and motivation when experimental outcomes are unsuccessful.
The study draws on a structured LS process conducted across one academic year by four subject teachers. Three cycles combined in-class diagnostic observation, joint observation of extracurricular project sessions and end-of-year synthesis. Five data streams were analysed through a three-pass iterative interpretive process with cross-source triangulation.
The focal case showed different patterns of student reasoning from the technically successful comparison case when mentoring reframed experimental failure as an epistemological resource. Motivation was sustained through autonomy, disciplinary relatedness and personal meaning rather than outcome validation. Competition outcomes provided external corroboration. After non-qualification for a national final, focal-case students independently diagnosed a robotics competence gap and developed a new project, providing evidence of self-directed learning agency following sustained LS-informed mentoring.
Four mentoring principles are proposed: agency-focused observation, failure-reframing language, affirmation of personal meaning and structured LS reflection on mentoring decisions.
The study extends LS to long-term mentoring contexts, reframes experimental failure as a productive site of teacher learning and identifies self-directed learning agency as a documentable mentoring outcome under uncertainty.
