Improving Question Papers and Exams for Better Learning

Every semester, thousands of students sit for exams that are supposed to measure what they have learned. But how often do we stop to ask: are those exams actually doing their job? A poorly designed question paper doesn’t just frustrate students – it produces misleading results, rewards rote memorization, and undermines the entire purpose of education. Research published in Higher Education found that relatively few of the perceived academic benefits of high-stakes examinations rest on strong evidence, while the pedagogical drawbacks are well-documented. Improving how we design exams and question papers is, therefore, not a minor administrative task – it is central to fair, effective teaching.

What’s wrong with current exam practices? 🔗

Before we talk solutions, it helps to be clear about the problems. Exam systems in many universities struggle with three interconnected issues: frequency, language, and reliability.

Too much riding on too few exams 🔗

In many institutions, a student’s entire performance in a course is judged through a single end-semester examination. This is problematic for a straightforward reason: one test on one day cannot reliably capture everything a student has learned over months. Spreading content across multiple tests rather than a single high-stakes event is a well-established way to improve the accuracy of assessment. When assessment is concentrated in one final exam, students also tend to cram rather than learn – optimizing for the test rather than for understanding.

Ambiguous and inaccessible language 🔗

Question papers frequently suffer from poorly worded questions – instructions that are vague, unnecessarily complex, or open to multiple interpretations. Studies on examination paper errors show that a flawed question doesn’t just cost a student marks on that item – it disrupts their confidence and performance across the rest of the paper. Question language should be precise, accessible, and free of ambiguity. This is especially important in multilingual academic environments where the language of instruction may not be the student’s first language.

Reliability gaps in question paper construction 🔗

Research in academic assessment has consistently shown that in higher education, few faculty members receive formal training in how to construct examination questions, yet assessment is the dominant mode of evaluation. Without this training, question papers are often inconsistent in difficulty, skewed toward a narrow range of cognitive skills, and fail to represent the syllabus evenly. Educational assessment literature notes that question papers used by many universities are criticized for being inferior in quality and failing to perform their core educational function of meaningful evaluation.

Designing better question papers 🔗

Better exams begin before a single question is written. The foundation of a good question paper is a clear set of instructional objectives and a structured plan – called a blueprint – for how those objectives will be assessed.

Start with instructional objectives 🔗

An exam should measure what was actually taught and what students were expected to learn. This means question design must begin with a review of the course’s learning outcomes. The University of Tennessee’s guidelines on exam design emphasize that a valid and reliable test helps ensure students are assessed fairly and on intended outcomes. Before writing any question, educators need to ask: what knowledge, skill, or ability am I trying to measure here?

A useful framework for this is Bloom’s Taxonomy, which classifies learning outcomes from basic recall at the lower end to evaluation and creation at the higher end. Research has found that educators tend to construct test questions at the knowledge or recall level 80-90% of the time, rather than targeting higher-order thinking skills. This means most exams end up testing memory rather than understanding – a significant gap between what we claim to assess and what we actually measure.

The role of the question paper blueprint 🔗

A blueprint (also called a table of specifications) is a planning matrix that maps out how many questions will come from each topic, what types of questions will be used, and what cognitive level each question targets. An assessment blueprint is a tool that helps educators be intentional and reflective when creating exams, ensuring that the paper accurately reflects the content taught and the skills students are expected to demonstrate.

A well-constructed blueprint addresses several quality dimensions at once: relevance of questions to the syllabus, appropriate distribution of difficulty, variety of question formats, and proportional coverage of all major topics. Research on question paper quality identifies these as the top factors that determine whether a paper is fit for purpose. A common best practice is to maintain a difficulty distribution of roughly 30% easy, 50% moderate, and 20% difficult questions – enough challenge to differentiate performance without making the paper punishing for average students.

Question type matters 🔗

Not all questions serve the same purpose. The choice of question type should be deliberate and tied directly to what you are trying to assess. There are broadly two categories:

Selected-response questions – such as multiple-choice (MCQs), true/false, and matching – are efficient to administer and score, and are well-suited to testing factual knowledge and comprehension across a wide range of content. Constructed-response questions – short answers, problem-solving tasks, and essay questions – require students to produce rather than select answers, making them better suited for testing analysis, application, and higher-order thinking.

A review published in Higher Education found that short-answer questions and context-rich MCQs requiring knowledge application tend to enhance student learning more effectively than recall-based multiple-choice items. A well-designed question paper uses a deliberate mix of both types rather than defaulting entirely to one format.

Balancing objective and subjective testing 🔗

One of the ongoing debates in higher education assessment is how to balance objective tests (MCQs, true/false) with subjective ones (essays, short answers). Both have genuine strengths and real limitations.

Objective questions: strengths and cautions 🔗

MCQs are popular because they cover wide content quickly and can be scored consistently. But poorly written MCQs introduce their own problems. Research on item-writing guidelines shows that deviations from established construction principles – such as negatively worded stems, poorly designed distractors, or an overuse of “none of the above” – can make questions harder than intended and lead to an underestimation of students’ actual knowledge. MCQs work best when they are carefully written to require genuine application of knowledge rather than surface-level recall.

Assessment validity experts also caution that objective tests must be designed to avoid embedded bias – questions that inadvertently disadvantage certain student groups not because of knowledge differences, but because of how the question is framed.

Essay and short-answer questions: depth with discipline 🔗

Essay questions allow students to demonstrate reasoning, synthesis, and communication – skills that MCQs cannot easily capture. Short-answer questions occupy a useful middle ground: they require students to formulate a response, test comprehension and application, and are far more practical to grade than full essays at scale. The key to making essay and short-answer questions work is pairing them with clear, detailed marking rubrics. Without explicit criteria, different evaluators may assess the same answer very differently – a direct threat to fairness.

Ensuring fair and reliable evaluation 🔗

Even the best-designed question paper can produce unreliable results if the evaluation process itself is inconsistent. Reliability in assessment means that the same student’s knowledge should receive roughly the same score regardless of who marks their paper or when.

Training evaluators 🔗

For subjective assessments, having clear rubrics describing each performance level and training scorers before grading sessions are essential steps toward consistent evaluation. Inter-rater reliability – the degree to which different evaluators agree in their scoring – is a critical but often overlooked dimension of exam quality. Having multiple raters evaluate samples of student work, comparing their scores, and resolving disagreements through discussion before full-scale marking begins significantly reduces scoring variability.

Centralized and structured assessment processes 🔗

Centralized assessment – where answer scripts from multiple centers or sections are evaluated together under standardized conditions – helps reduce the influence of individual evaluator bias. Alongside this, regularly reviewing and updating tests as learning needs change ensures that assessments stay aligned with course objectives over time. Item analysis after each exam – examining how students performed on each individual question – is an invaluable tool for identifying questions that were either too easy, too difficult, or that failed to discriminate between students who understood the material and those who did not.

Multiple assessments over a single high-stakes exam 🔗

One of the most practical reliability improvements is also structural: spreading evaluation across multiple assessments rather than one terminal exam. Reliability can be improved by increasing testing time, separating content into multiple tests, and using a battery of assessments to measure the same competencies. Continuous assessment through class tests, assignments, and unit-end evaluations gives a more accurate cumulative picture of student learning than a single three-hour paper ever can.

Practical steps for educators 🔗

Improving exam quality does not require overhauling an entire institution overnight. Even small changes in practice can make a meaningful difference. Before setting any question paper, define your learning objectives clearly and prepare a blueprint. Vary question types so that different cognitive abilities are assessed. Write questions in plain, unambiguous language and have a colleague review the paper for clarity before it is finalized. After each exam, conduct a basic item analysis to flag questions that did not perform as expected. Train those responsible for evaluation on the marking rubric before grading begins. These are not extraordinary steps – they are simply good professional practice that too many institutions skip.

Exam blueprints are a strategy for equitable and effective assessment – but their real value is realized only when they are embedded within a broader culture of intentional, evidence-informed examination design. The goal is not just to produce a paper that is hard enough to rank students, but one that is fair enough to tell them – and their teachers – something true about what has been learned.

What do you think? If you design or moderate question papers, how systematically do you align your questions with stated learning objectives – and is the blueprint approach something your institution currently uses? Do you think spreading evaluation across multiple assessments rather than a single final exam would better reflect what students actually know?

References

1. https://link.springer.com/article/10.1007/s10734-023-01148-z
2. https://distancelearning.institute/curriculum-development/understanding-measuring-test-reliability/
3. https://www.tandfonline.com/doi/full/10.1080/03054985.2024.2308548
4. https://pmc.ncbi.nlm.nih.gov/articles/PMC3663625/
5. https://files.eric.ed.gov/fulltext/ED613841.pdf
6. https://teaching.utk.edu/exam-test-design/
7. https://www.anthology.com/blog/using-blueprints-to-align-course-objectives-with-assessments
8. https://slejournal.springeropen.com/articles/10.1186/s40561-021-00148-9
9. https://www.taotesting.com/blog/4-ways-to-improve-exam-content-validity/
10. https://pmc.ncbi.nlm.nih.gov/articles/PMC10666833/
11. https://www.facultyfocus.com/articles/educational-assessment/exam-blueprints-a-student-centric-approach-to-assessment/