Students judge different things differently: dimension-specific self-assessment calibration in introductory digital circuit design

Background and context Computing students must learn not only to produce working artefacts but also to judge their quality across multiple dimensions. This issue is especially important in introductory courses, where students are still learning what counts as good work and where functional correctness may mask weaknesses in understanding, design quality, or communication.Objective This study examined whether self-assessment calibration in computing is better understood as dimension-specific rather than as a single general skill. We asked whether calibration varied by student performance level and by rubric dimension, and whether voluntary self-assessment participation was associated with subsequent examination performance.Method We conducted a mixed-methods study in a first-year digital circuit design course. Self-assessment data from 323 students were compared with teaching-assistant ratings using the same five-dimension rubric, and interview data from 36 students were analysed to interpret observed calibration patterns.Findings Students showed a clear performance-related calibration pattern, with lower-performing students tending to overestimate more than higher-performing peers. More importantly, calibration was asymmetrical within the same assignment: students overestimated conceptually oriented dimensions but underestimated the open-ended, creative dimension (alternative solutions), whereas implementation quality was comparatively well calibrated. Implications: The findings suggest that self-assessment in computing should be designed as dimension-sensitive calibration work rather than as a generic reflective exercise. For computing education, this means helping students distinguish correctness from quality and making criteria visible through exemplars, prompts, and comparative feedback processes.