How to gather and use information on three aspects of our students' background that can help us in teaching them.
How to make quality assessments
Have you ever sat down to assess students work and realised you are not completely sure what, exactly, you are assessing? If so, you are not alone. A key reason is that “the goal” of our teaching is rarely a single thing.
In higher education, goals exist at several levels, from programme aims to course aims, to intended learning outcomes, and right down to the detailed and assessable sub-outcomes you might target in a single exercise. If those levels are not clearly connected, assessment becomes fuzzy, feedback becomes generic, and students struggle to understand what “good performance” even looks like.
In this spotlight we present a practical way of thinking about quality assessment by focusing on three questions: what learning goal are you looking for, how can you collect quality data about students’ learning, and what do you use the data for. By being able to answer these questions you can make existing assessment practices more targeted, transparent, and useful for student learning.
The three questions
The three questions are partially related to the notion of constructive alignment, where learning outcomes, teaching and learning activities, and assessment are deliberately aligned so that assessment that each element helps strengthen student learning.
What are the learning goals?
Quality assessment begins when you can describe, in plain language, “this is what I am looking for today,” and you can show how it relates to course and programme aims.
You cannot assess everything at once, and so it is central to be able to delimit what we want to focus on when talking about student learning.
At programme level you often define learning goals as broad aims. At course level these goals become a little more specific, and these goals can be further broken down into more discreet units of learning outcomes.
To have goals, that you are actually able to assess, you need even more specific sub-outcomes. In any given teaching situation, you can select one or two sub-outcomes as focus in order to plan the lesson and make it clear what you are assessing – to the students and to yourself.
How do you collect quality data about students’ learning?
Most of us already use a wide repertoire, including quizzes, tests, dialogue, observations, peer feedback, reports, presentations, concept maps, and many other creative methods. The issue is not lack of methods; it is whether we use them systematically enough to generate evidence we can trust, whether for grading or for formative feedback.
One particularly powerful idea here is performance or authentic assessment: assessing students while they are doing something that resembles what you ultimately expect them to be able to do in the future. For example, we educate coming high school teachers at our department and as part of one of our courses we have our students teaching high school students an actual lesson as part of the course and examination. But it could be demonstrating any practical competence in a realistic setting such a lab, in a clinical setting or during field work. When the task itself mirrors the types of situations that the students are expected to perform in, the data you collect tends to be more meaningful.
How do you use the data?
Assessment data can serve summative purposes, such as documenting achievement in an exam situation. It can also be used formatively to support learning. Being explicit about the intended use matters, because students respond to assessment differently to assessment depending on what it will be used for.
If students are to learn from the assessment the quality of the feedback they receive is important. Research shows that effective feedback helps students understand what a good performance is, can compare their current performance to that standard, and can act to close the gap. This means that whether you are responsible for a program, a course, or a single teaching session, you can strengthen the quality of assessment and thereby student learning by making the learning goals at different levels more explicit.
Using rubrics to strengthen assessment and learning
A rubric is a tool that will help you answers the three question above in order to make quality assessments. A rubric is in essence a scoring guide that makes quality visible by emphasising the criteria you care about and describing different levels of performance for each criterion. Rubrics can be used to do three things: communicate what the learning targets are, collect consistent data about student performance, and help students give feedback that supports progression rather than merely signalling success or failure.
What is a rubric
A rubric is basically a table or matrix, where you list a short number of chosen criteria (usually 3-5) down one axis and a chosen number of different categories of performance. This typical include 3-5 categories ranging from insufficient to excellent. For each of the resulting cells in the matrix, you find a detailed description of what a performance of each category of quality looks like for each criteria.
A practical way to build a rubric is to start from your course learning outcomes and treat them as your first draft criteria. Taking an hypothetical course “Advanced Ecological Modelling,” the learning outcomes could include explaining key theories, developing and implementing models, analysing ecological data and evaluating model predictions, and communicating modelling results. Those can become four criteria for succeeding in the course.
Now zoom into one criterion to see how the rubric creates clarity. Take “Analyse ecological data and evaluate model predictions.” A rubric might describe performances from “insufficient” to “Excellent” as below.
| Insufficient | Passable | Good | Excellent | |
|---|---|---|---|---|
| Criterion: Analyse ecological data and evaluate model predictions | Data analysis is incomplete or incorrect. Model results are reported but not meaningfully evaluated. No sensitivity analysis or validation. Interpretation is unsupported. | Basic data analysis using appropriate methods. Model predictions are described and partly interpreted. Limited attempt at validation or sensitivity analysis. Conclusions partly supported. | Systematic data analysis using appropriate techniques. Model predictions are evaluated against data or diagnostics. Sensitivity analysis or robustness checks are conducted and interpreted. Limitations are recognised. | Rigorous and well-justified data analysis. Model predictions are critically evaluated using multiple approaches (e.g., validation, sensitivity or uncertainty analysis). Interpretation demonstrates strong ecological understanding and clearly justified conclusions. |
Why use a rubric
A rubric clearly communicates the learning target by making the criterion and its performance standards explicit. Students can see what you mean by “evaluate model predictions” because you have described what evaluation looks like in practice.
Secondly, it improves the quality of the data you collect, because you now have a shared lens for noticing evidence in student work.
Third, it supports feedback, because you can point to the current level and describe the next step in terms of observable features, not general encouragement.
Rubrics are especially useful when you want to think in terms of progression. Moving from “insufficient” to “excellent” in this example reflects increasing correctness of analysis, increasing quality of model evaluation, deeper interpretation, and more explicit treatment of robustness and uncertainty. That progression framing is pedagogically valuable because students are not merely told they are “wrong”; they can see what development looks like.
Rubrics also have wider uses beyond teacher grading. They can support peer assessment and self-assessment, precisely because they clarify the criteria students should apply when they judge work. They can support planning of teaching activities by reminding you what to prioritise in a given teaching situation, especially when there is a risk of trying to cover too much. They can increase transparency by being shared beforehand, so students prepare with the criteria in mind and the rubric becomes a guide for learning, not only an instrument for judgement.
Tips for creating a rubric
- Avoid criteria that rely on vague terms such as “understanding.” If a descriptor says “limited understanding” versus “deep understanding,” students may not know what action would actually improve their work. In practice, you often need to replace that kind of language with something more behavioural and evidential, so it is clear what a better performance would involve.
- Include examples. For instance, showing a weak answer to a complex question and explaining precisely why it is insufficient can make the standards far more concrete than adjectives alone. This kind of exemplification reduces guesswork and helps students internalise quality.
- Make it clear whether the rubric is primarily for grading, for feedback, or for both. Rubrics can help achieve consistency and fairness in grading by offering a shared language for distinguishing performance levels rather than relying on general impressions, which can be particularly helpful in contexts like oral examinations or when assessing a thesis.
- Be aware that even well-written descriptors can remain interpretive, meaning different teachers may rate the same performance differently unless you calibrate your assessments.
- Focus on one or two criteria in a given activity. Using a full rubric in real time can also be cognitively demanding and no rubric captures everything; you always prioritise some dimensions of learning over others.
- To get started, pick a teaching activity and try to draft 2 or 3 criteria, write performance descriptors that avoid vague terms, and test the rubric by applying it to assess and give feedback on the students’ performance.