Principles of Assessment Design
Download transcript for videoLink opens in a new window
Assessment tasks and associated criteria must test student attainment of the intended learning outcomes effectively and at the appropriate level. Where learning outcomes state skills and attitudes as well as knowledge, this should be appropriately reflected in the chosen assessment methods. This is known as constructive alignment.
Constructive alignment and learning outcomes:
Assessment must be aligned to learning outcomes; we tell our learners what we expect and then test them to see if they match, and to what level, those expectations. It is, therefore, essential to define learning outcomes effectively, efficiently and at the appropriate level as these will direct the method(s) by which you assess learning and will form the basis of your assessment criteria. Although you will be assessing against university-wide standards, the specific assessment criteria for your module need to define characteristics and standards of performance in line with the learning outcomes that you are assessing.
Intended learning outcomes capture the answer to the essential questions:
-
What do you want your students to know or to be able to do?
-
What will the student do that demonstrates learning?
-
What is the context within which that learning will be demonstrated?
-
How well will that student be required to demonstrate that learning?
Therefore the very first thing to determine when constructing learning outcomes is the knowledge, understanding, competencies, behaviours and attributes that you wish your student to demonstrate.
Reflection point: Thinking about one of the modules you are teaching on, how do you ensure that students are familiar with the intended learning outcomes?
The main point of intended learning outcomes is to make clear to learners what is expected of them; the intention is to share a common [to teachers and learners] understanding of expectations. This is more than just listing outcomes in the module and programme documentation; we need to discuss the outcomes with learners, to ensure they understand what is expected, and remind them that the outcomes should be guiding their learning and will be used to measure their progress. To do this it is helpful to have a standard ‘language’ and approach and the most common one is based on the Bloom taxonomy, as described below. Having made your expectations clear to learners, you then need to decide how (and when) to assess their achievement of the outcomes.
A common language for outcomes:
When we assess learners we invariably do so by asking them to do something: write an essay or report; calculate answers; analyse information; present an argument; exhibit a behaviour; demonstrate a competence; etc. This means that the verbs in learning outcomes assume central significance.
Some verbs describe fairly straightforward achievement - for example, "to describe". Others can be more complex - for example, "to compare". A learner can only "compare" if s/he first "describes" both things that s/he is comparing. It follows, then, that comparing is more complex than describing. Hierarchies of cognitive learning outcomes based on their complexity and derived from ideas have evolved, and are now widely used. They draw upon a framework for categorising educational achievement, devised in 1956 by educational psychologist Benjamin S. Bloom with collaborators Max Englehart, Edward Furst, Walter Hill, and David Krathwohl, in order to promote higher order thinking. This device, commonly known as Bloom's Taxonomy, is a useful tool when authoring learning outcomes, particularly when linking outcomes to level descriptors. The taxonomy was revised by Anderson and Krathwohl (one of the original collaborators) in 2001, and the new model is illustrated below.

Image available from Information Technology, University of FloridaLink opens in a new window
Verb tables threaten to fetishise language at the expense of authentic pedagogic thinking. Generating formulaic outcomes appears to be more about bureaucracy than pedagogy. You are not educational bureaucrats, you are educational leaders; deciding what it is the next generation of thinkers and doers in your discipline will be learning. Learning outcomes are powerful tools for designing learning and should be considered carefully. The table below provides examples of measurable verbs which align to each level of Bloom's Taxonomy to help structure learning and assessment.
Bloom | Key question | Associated verbs |
---|---|---|
Creating | Can the student create a new point of view or product? |
Assemble, construct, create, design, develop, formulate, write |
Evaluating |
Can the student justify a stand or position? |
Appraise, argue, defend, judge, select, support, value, evaluate |
Analyzing |
Can the student use information in a new way? |
Appraise, compare, contrast, criticise, differentiate, distinguish, examine, experiment, question, test |
Applying |
Can the student use information in a new way? |
Choose, demonstrate, dramatise, employ, illustrate, locate, recognise, report, select, translate, paraphrase |
Understanding | Can the student explain ideas or concepts? |
Classify, describe, discuss, explain, identify, locate, recognise, report, select, translate, paraphrase |
Remembering |
Can the student recall or remember information? |
Define, duplicate, list, memorise, recall, repeat, state |
Reflection point: Consider one of the modules you are teaching on, is it clear how the assessment has been structured to test the intended learning outcomes?
As noted above our expectations of learners include skills and attitudes as well as cognitive achievement. In the same way we describe cognitive outcomes using the sorts of verbs listed above, we should be clear about the skills that we expect (and to what level) and the attitudes and approaches that learners should be developing. Assessing skills and attitudes is, generally, more difficult that testing cognitive outcomes. How fair do you think your driving test was? How do we judge empathy, bedside manner, critical approach, etc.? Clarity over expectations and linking these to our assessment criteria makes the assessment process more transparent, fair and equitable.
For a more detailed discussion of these ideas see Butcher et al 2019.
Designing assessment that supports the use of AI
Members of the ‘AI in Education’ subgroup of the WIHEA Diverse Assessment Learning Circle have developed helpful guidance on designing assessment that supports the use of AI - see 'Principles of good assessment design that support the use of artificial intelligenceLink opens in a new window' to access this resource.