While selecting an appropriate assessment method is important, the design of the specific task is even more important. The quality of the questions we ask or the specific activities we design to assess students' learning is a key factor determining their ability to demonstrate and assess their learning towards the intended learning outcomes.
Race (2005) provides tips and pointers on the detailed design of fifteen types of assessment tasks.
One key point that Race makes is that as the assessment designer you do not try to do the detailed design of assessment tasks on your own. It is important to have a critical friend who will review tasks and specific questions to critique their clarity and appropriateness to the student group and the specific topic context. It also may be appropriate to involve students in the design of assessment tasks.
The Griffiths Institute Good Practice Guide emphasises the need to become aware of the common item flaws in question design such as:
- Multiple choice questions: e.g. grammatical cues, logical cues, use of absolute terms (‘never', always') or imprecise terms (‘many', ‘seldom'), a long correct answer may be a cue. (Case and Swanson, 2002)
- Short answer questions: e.g. answer expected does not match the question in level of detail required, ambiguous wording.
It is good practice to review the wording of your assignments or exams with colleagues.
What may seem very clear to you may well have considerable ambiguities. Have a colleague/s read your exam questions with answer guides and ask for feedback. Colleagues from a different area may pick up the ambiguities that students will see in a question, or a mismatch between the question and what is indicated in the answer guide.
Case and Swanson (2002) provides a comprehensive guide to detailed item design.
How much assessment is enough?
How much assessment is required will be shaped by the learning needs of the students and the learning objectives of the topic, as well as the need to grade students. The following should be taken into account:
- For assessment to enhance learning:
- students need to know where to focus their learning activities.
- students need a chance to risk failure before undergoing summative evaluation.
- assessment should not be a means of 'catching students out.
- the end result should be no surprise to students or assessor.
- A single mode or piece of assessment may advantage one group of students over another when it comes to demonstrating what they have learned,
- for example, will a memory-oriented, time-constrained exam paper favour some students more than, say, a researched essay?
- how might the context or timing of the assessment influence outcomes?
Phil Race on his web page muses about "if I were in charge" and notes several characteristics of his ideal assessment world. One of these is that there would be no more silly relationships between point values and word counts. Such relationships are arbitrary with no educational justification. They also promote unproductive behaviours like low level ‘word spinning' and padding.
In conclusion, to be useful assessment should address sufficient aspects of students' learning processes and outcomes appropriate to the topic objectives and the stage of the topic/course. The amount of assessment is driven by:
- students' need for a range of information for feedback about their learning
- assessors' need for a range of information for assessing learning outcomes.
So asking 'How much assessment is enough?' is like asking 'How long is a piece of string?'