Answers are grounded in this paper and cite their sources.
Assessment feedback in higher education provides fine-grained evidence about students’ specific errors but remains underutilized as a data source for teaching improvement. This study proposes a method based on a large language model (LLM) that transforms unstructured feedback into curriculum-aligned error analysis in three steps, deriving an error taxonomy from the feedback, mapping each piece to curriculum content by retrieval-augmented generation, and aggregating the annotations into error patterns with quantified cohort coverage. The method was applied to 6072 pieces of feedback across four cohorts (2022–2025) of an undergraduate engineering course. The feedback resolved into recurring, curriculum-aligned error patterns, dominated by missing content and justification, clustered in identifiable modules and weeks, and more than half recurring across cohorts. The resulting annotations achieved 89.1% agreement with human judgment, substantially higher than the 74.7% obtained when an expert-designed taxonomy was applied to the same data. In a focus group, the teaching team confirmed the patterns against their marking experience, proposed course revisions from them, and set the bound that they serve as a reference and not as a direct measure of teaching quality. The study reports a process account of this team’s use of curriculum-aligned error analytics, built on sensemaking and pedagogical response. The analysis turns feedback written for students into evidence on the teaching that preceded it, by which a teaching team can check the alignment between intended outcomes, teaching, and assessment in students’ actual work, reflect on its own practice, and improve its teaching across cohorts.
No takes yet. Share an insight, caveat, or question.
Hua et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: