4.2 Peer- and self-assessment
Peer- and self-assessment have substantial advantages (see section 5.4 for further reading) and these are summarised by Race (2001):
- Students are already self-assessing and peer assessing quite naturally.
- Tutor assessment is not sufficiently valid, reliable or transparent.
- Peer- and self-assessment deepen students' learning experiences.
- They let students into the assessment culture.
- They help students towards becoming autonomous learners.
- They help students develop skills relating to life-long learning.
- They help students to gain much more feedback than would otherwise be possible.
- Some of the drudgery can be saved by peer assessment.
- Self-assessment gets students to reflect on their own work, and can open up productive student-tutor dialogues.
- Students' performance in traditional assessments is enhanced.
A number of appropriate peer- and self-assessment checklists are included in Jaques (2000). One of these lists concentrates on students' skills in organising their learning:
- Contribution to group discussion.
- Carries out instructions and degree of supervision needed.
- Works easily with others.
- Response to criticism.
- Polite with colleagues.
- Ability to organise own work.
- Ability to supervise others' work.
- Speed of recognising essentials.
- Adaptability to new situations.
- Efficiency at solving problems.
A mark of 1-5 is applied on a scale that goes from a very positive assessment of their colleagues' ability to a very negative assessment.
Instead of checklists completed by all students, we can opt for forms that are completed by the project group leaders. Elliott (2001) adopts this method in a group project for first-year business economics, with statements being made about the work of each group member.
The argument for self- and peer assessment is not one-sided. There are dangers of student bias that need to be allowed for. Typically, bias may arise from:
- reticence about providing negative information on fellow students;
- reluctance to cover a full spectrum of marks, with realistic maxima and minima;
- collusion that disadvantages some individuals;
- reluctance to fail a student.
Design options to counter such bias include:
- initial discussion, input and commitment by students to assessment programme;
- checking uniformity of student opinion (e.g. individual questionnaires);
- observing anonymity in assessment (e.g. questionnaires completed in class);
- a clear procedure for arbitration on disputed assessments;
- a procedure for investigation of student participation (e.g. vivas, log books and exams).