When Grades Replace Growth: Inside Bangladesh’s Struggle to Make Continuous Assessment Work

Continuous assessment was supposed to be higher education’s answer to the high-stakes final exam: a steady drumbeat of assignments, quizzes, and feedback designed to keep students learning all semester long rather than cramming at the end. But a new study from Khulna University in Bangladesh suggests that, in practice, the approach often delivers the paperwork of ongoing evaluation without the learning it promises. Researchers Bijon Kumar Roy and Md. Fajlay Rabbi of Khulna University’s Education Discipline examined why continuous assessment falls short at their own institution, and their findings, published in SN Social Sciences, paint a picture of a system that has been reduced to little more than a distributed grading machine.

The study took a qualitative descriptive approach, drawing on semi-structured interviews with 26 participants purposively selected from eight academic schools across the university: 10 teachers and 16 students. Rather than measuring outcomes with surveys or test scores, the researchers wanted to understand the lived experience of continuous assessment from both sides of the lectern. The interview transcripts were analysed using thematic analysis, the widely used qualitative method formalised by Virginia Braun and Victoria Clarke, in which researchers code data and iteratively refine codes into overarching themes. The team anchored its interpretation in Bronfenbrenner’s bioecological theory, which frames individual development as the product of nested systems ranging from the immediate classroom to institutional policy, an apt lens for a problem that spans pedagogy, workload, technology, and governance.

The central finding is deceptively simple: continuous assessment at the university is implemented as a grading-oriented process rather than a learning-oriented one. In theory, the distinction matters enormously. Assessment for learning, as scholars such as Dylan Wiliam and Eleanore Hargreaves have argued, uses ongoing tasks to surface misconceptions, adjust teaching, and build students’ capacity for self-regulation. Assessment of learning, by contrast, merely samples performance and assigns marks. When a semester’s continuous assessment consists of a fixed schedule of quizzes and class tests whose scores are simply aggregated into a final grade, the formative function evaporates. Students learn to treat each task as a mini-examination to be survived rather than an opportunity to be diagnosed and improved, and the pedagogical promise of continuous monitoring is quietly discarded.

Feedback, the mechanism that is supposed to convert assessment into learning, emerged as a key weak point. The study identified ineffective feedback as one of the core pedagogical problems, alongside limited learner-centred assessment and a persistent reliance on traditional assessment methods. This aligns with a broader international literature: research on formative feedback in universities has repeatedly shown that feedback delivered late, delivered as a bare score, or delivered without guidance on how to improve does little for motivation or achievement. Recent work published in Assessment and Evaluation in Higher Education in 2025, for instance, found that timely formative feedback measurably boosts university student motivation, underscoring that timing and quality, not the mere existence of feedback, are what count. At Khulna University, teachers and students alike described feedback practices that failed to meet that bar.

Why would educators, many of whom presumably understand the theory, implement continuous assessment in such a hollowed-out form? The study points to a cluster of institutional constraints. Workload pressures loom large: designing varied, learner-centred assessment tasks and providing substantive feedback on each takes time that heavily loaded faculty schedules do not allow. Limited technological resources compound the problem, since modern assessment design often depends on learning management systems, online quizzing tools, and digital submission platforms that reduce the administrative burden of frequent evaluation. Research on how technology shapes assessment design, including a 2016 study of university teachers published in the British Journal of Educational Technology, has shown that the available infrastructure directly constrains what assessment designs teachers can realistically sustain. Policy and structural constraints complete the picture: when institutional rules prescribe assessment formats and weightings, individual teachers have little room to experiment.

Students, for their part, reported a distinctive set of harms. The study found inconsistent engagement, assessment overload, and increased dependence on generative artificial intelligence. The overload finding echoes earlier international research: a 2006 study by Trotter on student perceptions of continuous summative assessment found that frequent graded tasks can heighten stress and encourage surface learning strategies, and a 2024 study in the International Journal of Management Education described how continuous assessment can become a stressful sum of grades rather than a scaffold for meaningful learning. When every small task carries marks, students optimise for scores rather than understanding, and the sheer volume of deadlines crowds out the reflection that formative assessment is meant to provoke.

The finding on generative AI is perhaps the most contemporary and the most sobering. As tools such as large language models have proliferated, students facing a relentless stream of graded written tasks have an obvious shortcut available, and the study documents increased reliance on these tools as part of the student experience of continuous assessment. This creates a feedback loop of its own: the more assessment is grading-oriented and the more it relies on traditional written formats, the more incentive students have to outsource the work to machines, and the less valid the resulting grades become. A 2026 systematic review in Frontiers in Education on generative AI’s impact on authorship, pedagogy, and integrity documents the scale of this disruption across higher education. The Khulna study suggests that assessment reform is now inseparable from the AI question: diversifying assessment methods is both a pedagogical improvement and an integrity measure.

Notably, the problems the researchers identified are not unique to Bangladesh. Studies from Ethiopian public universities, Danish science courses, Dutch higher education, and chemical engineering programmes in the European Higher Education Area have all documented gaps between the theory of continuous assessment and its practice, including challenges of class size, teacher workload, and student perceptions of frequent summative testing. What the Khulna study adds is a fine-grained, dual-perspective account, teachers and students together, within a single institution, organised by an ecological framework that makes clear how individual classroom practices are embedded in, and constrained by, larger institutional structures. That framing matters for anyone hoping to fix the problem: exhorting individual teachers to give better feedback will not work if workload policies, technology provision, and assessment regulations remain unchanged.

The study’s recommendations follow directly from its themes. The authors call for enhancing assessment literacy among academic staff, so that teachers understand the distinction between formative and summative functions and can design tasks that serve both; improving feedback practices, so that students receive timely, actionable information rather than bare scores; diversifying assessment methods, moving beyond the traditional quiz-and-class-test repertoire toward formats that are harder to outsource and better suited to different kinds of learning; and strengthening institutional support, including workload relief, technological infrastructure, and policy frameworks that enable rather than prescribe. Each recommendation targets a different level of Bronfenbrenner’s nested systems, from the microsystem of the classroom to the exosystem of university administration, reflecting the study’s theoretical commitment to seeing assessment as an ecological problem.

For university administrators in Bangladesh and beyond, the message is that continuous assessment is not a self-executing reform. Simply mandating frequent assessment distributes the defects of summative examination across the semester while leaving its benefits, diagnosis, adjustment, and self-regulated learning, unrealised. The Khulna findings offer a practical checklist for redesign: audit whether current tasks generate feedback students actually use, whether the assessment mix rewards understanding over mark accumulation, whether faculty have the time and tools to sustain good practice, and whether policies treat assessment as a learning technology rather than an accounting system. Grades without growth, the study’s title asks pointedly, and the answer from 26 teachers and students at one Bangladeshi university is that growth requires far more than a full gradebook.

Subject of Research: Implementation of continuous assessment in tertiary education in Bangladesh

Article Title: Grades without growth? Rethinking continuous assessment at the tertiary level in Bangladesh

Article References: Roy, B. K., & Rabbi, M. F. (2026). Grades without growth? Rethinking continuous assessment at the tertiary level in Bangladesh. SN Social Sciences, 6(9), Article 430. https://doi.org/10.1007/s43545-026-01736-7

Image Credits: AI Generated

DOI: 10.1007/s43545-026-01736-7

Keywords: continuous assessment, formative assessment, higher education, feedback, assessment literacy, Bangladesh, Khulna University, generative AI, student engagement, qualitative research, assessment overload, educational policy