Phil Race

Quality of assessment

Phil Race
Ch4 from Never Mind the Teaching Feel the Learning. SEDA Paper 80, 1993. Reproduced with permission.

Abstract

What our learners really pay for (or what society pays us for) are their Degrees or their Diplomas or their Certificates. Yet we put much more emphasis on trying to teach our students than we normally put into designing assessment - particularly the 'formal assessment' that counts towards final qualifications. Assessment is where learners often get a raw deal. In this chapter, I will propose the following ten 'worries' about assessment, and explore ways of leading to a better quality of assessment.

Ten worries about assessment

1. Assessment is often done in a rush, to meet exam board deadlines . It's rarely done under the best of conditions!
2. Assessment is often done by bored people, tired of reading the same answers to the same questions (and seeing the same mistakes).
3. Assessment tends to be governed by 'what is easy to assess'. Therefore, traditional written exams (relatively straightforward to assess) are used. These measure students' skills at tackling traditional written exams.
4. Students rarely know the intimate details of the assessment criteria, and how we interpret them.
5. How should we develop students' unassessable qualities? Should we refrain from developing them because we can't measure them?
6. Almost all assessment processes in common use foster student competition rather than collaboration. No wonder our educated people are so bad at working in teams.
7. What competences are measured by assessment anyway? Are they ' can do ' competences? Or are they simply 'did do, once' ghosts?
8. If we were to introduce ' Quality in Assessment BS5750A what should the criteria look like? What evidence of competence should assessors demonstrate?
9. 'If you can't measure it, it doesn't exist. If you can measure it, it isn't it'. What should we be trying to measure?
10. Where stops the buck? Whose fault is it that assessment is so artificial? HoDs? Employers? Assessors? Validators? The Government? Yours? Mine?

Is Competence enough?

In 1991, the Association of Educational and Training Technology (AETT) International Conference had the theme 'Developing and Measuring Competence'. As one of the editors of the proceedings of that conference (Saunders and Race, 1992), I became well acquainted with the many views on competence which were aired by the various contributors. It was heartening to see that many of the approaches to the development of competence focused on defining the evidence which would be sufficient to demonstrate competence.

However, I was left with some concerns about whether competence would often end up as minimum competence rather than high competence. Also, I already hinted at some of my concerns towards the close of the previous chapter. The following two figures pose some questions about 'shades' of competence, and about a range of descriptors which may be needed to provide more information than simply 'can do'.

 

Figure 1. Shades of Competence

 

Figure 2. Some competence descriptors

Towards mass higher education

Higher education in the United Kingdom is being moved 'in fits and starts' towards greater participation rates, despite apparent U-turns in Government fiscal policies on expansion. In the long term, for an sorts of reasons - and at whatever cost - the pressures for higher participation rates in higher and further education will be irresistible. This means not only rapidly increasing numbers of school leavers, but also many more non-traditional entrants. Already, many lecturers and tutors are seeking help in handling large classes. It is clear that the workload of staff in higher education is increasing, as they become responsible for more and more learners.

In some ways, teaching can be 'done to' large numbers almost as easily as to small numbers - it's almost as easy to lecture to 300 as to lecture to 100. However, teaching greater numbers may have little connection with learning by greater numbers. And when it comes to assessment, there are no short cuts. It usually takes three times the amount of time to assess 300 learners as it would to assess 100 learners. The main challenge we face as we move towards a mass higher education system is merely maintaining the quality of assessment we presently have - not to mention improving it.

Assessment and learning: how do they mix?

In Chapter 1 I introduced the simple model of learning summarised again below:

• wanting to learn (motivation)
• learning by doing (practice, learning from mistakes, and so on)
• learning through feedback (to develop positive feelings about the learning)
• digesting (taking stock, making sense of the experience, and of the feedback).

How does assessment usually relate to these steps in the learning process? Let's take the most common form of assessment - the written examination.

• wanting to learn (motivation): Not many people like exams! The fact that there is an exam coming along at the end of the road is not the strongest motivator for most people. The inevitability of traditional forms of assessment is a key factor in preventing many people from participating in learning. Assessment processes are at cross purposes with learning processes.
• Learning by doing (practice, learning from mistakes, and so on): Learning by doing can indeed happen during assessments, including written exams. Mistakes are indeed made during assessment - plenty of them! But learning by doing while being assessed is hardly the best way of using experiential learning. Besides, it is usually presumed that the learning should have taken place before the assessment event. Again, assessment processes are at cross purposes with learning processes.
• Learning through feedback (to develop positive feelings about the learning): All assessment results in some sort of feedback. However, it's often the absolute minimum of feedback - for example a mere score - and that weeks or months after the event! There is little or no real feedback, and chances to learn from the feedback are minimal. Once more, assessment processes are at cross purposes with learning processes.
• digesting (taking stock, making sense of the experience, and of the feedback): Exams are better known for producing indigestion than for allowing people the chance to consolidate their learning. As I've said, the feedback is usually very limited in scope (and often delayed in time) and is not a useful means towards 'digesting'. Such feedback as there is, tends to be one-way feedback. There's little or no chance to discuss the details or negotiate what best to do next. Yet again, assessment processes are at cross purposes with learning processes.

Perhaps traditional forms of assessment have only one real contribution towards learning - people are frightened (shamed) into doing some learning so that they may minimise their chances of being shown to be 'lacking'. Much intensive learning is done just before exams - but most of it is of a superficial nature and soon forgotten again.

Ten worries - and some suggestions

In my Abstract at the start of this Chapter, I listed ten worries I have about assessment. I would like now to repeat them, giving a little more detail about my concerns, and offering some suggestions about how the problems may be minimised. Many of my suggestions point towards involving learners in their own assessment. This can be achieved by letting learners get their hands on the assessment criteria. It can be achieved even better by letting learners apply the assessment criteria, in self-assessment and peer-assessment. It is achieved best by helping learners to formulate the criteria, then apply them. An illustrated discussion of uses of self- and peer-assessment is given later in this chapter.

1 Assessment is often done in a rush, to meet exam board deadlines. It's rarely done under the best of conditions!

This is because assessment tends to be done to learners - not by them. Assessment tends to be done at the end of learning something, rather than as a means to help the learning processes. In public exams, examiners often face piles of some hundreds of scripts, which all need to be finished within only a week or two.

Suggestions:

• allow much more time for assessment, so that it can be done well.
• allow learners to use self- and peer-assessment, so they can learn by assessing.
• set questions or tasks where there will not be too many ways for learners to interpret them, making it easier to assess their work, and giving you more time to assess it objectively.

2 Assessment is often done by bored people, tired of reading the same answers to the same questions (and seeing the same mistakes).

Examiners get thoroughly fed-up as they wade through hundreds of scripts. They get discouraged when they see things they hoped their learners would have mastered, only to find that messages have not got across . Any tedious or repetitive task causes people to change their mood. If assessors' moods plunge, the objectivity of assessment is likely to be affected accordingly.

Suggestions:

• decrease the emphasis on traditional written exams altogether.
• allow learners to learn by reading their own mistakes, and those of their peers.
• 'double-blind' marking is expensive, but will make assessment more objective, not just because of the second-opinion, but also because of the effect of being a little more careful to mark objectively because of the likelihood of subjectivity being detected!

3 Assessment tends to be governed by 'what is easy to assess' . Therefore, traditional written exams (relatively straightforward to assess) are used. These measure learners' skills at tackling traditional written exams.

There is still not enough attention being paid to what should constitute the evidence upon which to base awards. Many important competences are simply not assessable by traditional methods. While it is perfectly possible to use traditional methods to measure recall of facts and information, it is not-at-all easy to use such methods to measure innovation, judgment, or personality. Fig.3 shows an overhead transparency I use to alert students to the various agendas that may be served by traditional exams.

Suggestions:

• look carefully at exactly what is being measured by each form of assessment.
• refrain from measuring the same things all the time - especially recall. People who can find and apply information are usually more valuable than people who simply happen to remember a lot of it.
• discuss with your learners exactly what you are going to try and measure, so they can prepare in a well-focused way for assessment.

4 Learners rarely know the intimate detail of the assessment criteria, and how we interpret them.

There really is no excuse for this. The reason may be sinister - that those who design the assessment criteria are not sufficiently confident about them to show them to the learners! Assessors often fear that learners may demand to know 'why did I get 65% for this, when my friend got 75%?' Surely, they have every right to ask this sort of question - and to learn from the feedback they should be given by way of a response to the question.

Suggestions:

• give learners the opportunity not only to see the intimate details of assessment criteria, but also to use the criteria.
• as a class exercise, get learners to apply a past assessment framework to specimen past answers, then to compare scores or grades, and discuss the reasons for differences.
• get learners to design an assessment framework for a given task and give them feedback about the objectivity and practicality of the assessment criteria they devise.

5 How should we develop learners' unassessable qualities? Should we refrain from developing them because we can' t measure them?

'Don't bother to learn anything, when you can't see how they can ask you about it at the end of the day': this is a perfectly rational view taken by learners, deciding what to learn and what not to learn.

Suggestions:

• bring the unassessable qualities firmly onto the agenda. Explain to learners why they 're important and work out with learners what kinds of evidence can be linked to these qualities, and how the demonstration of that evidence can be built in to assessment procedures.
• devise assessed tasks of the sort where learners necessarily develop desirable qualities en route to an assessed piece of work. For example, assess a 'learning log' of reflections on participation in a team project, as a means to cause learners to develop their collaborative skills and to reflect on them.

6 Almost all assessment processes in common use foster learner competition rather than collaboration. No wonder our educated people are so bad at working in teams.

Learners preparing for exams are often quite secretive about the work they do. No-one likes to be thought of as 'a swot'! However, it's more sinister than this: we actually compound the competition by using norm-referenced assessment far too much. In other words, only a certain proportion of learners are allowed to receive 'A' grades, or lst-class Honours degree classifications. Therefore, learners are in competition.

Suggestions:

• use criterion-referenced assessment only - abolish the use of norm-referencing.
• help learners to feel that they can help each other prepare to demonstrate their competence, without disadvantaging one another.
• build-in group tasks to the overall assessment plan. For example, allow members of a group to agree on how to split up the tutor-assessed overall mark for the product of the work of the group.

7 What competences are measured by assessment anyway? Are they 'can do' competences? Or are they simply 'did do, once' ghosts?

Exams tend to measure 'did once' competences. At their worst, they still measure 'knew once' competences rather than 'did once'!

Suggestions:

• increase the proportion of assessment schedules allocated to continuous assessment which measure 'is doing' competences rather than 'did once' ones.
• involve learners in self-assessment and peer-assessment of traditional-type exams (with due moderation where necessary). This at least allows learners to benefit from the learning payoff associated with reflecting on their own (and each other's) triumphs and disasters - better than just receiving a good or bad grade.

8 If we were to introduce 'Quality in Assessment BS5750A what should the criteria look like? What evidence of competence should assessors demonstrate?

At present, it is automatically assumed that anyone appointed to a post involving teaching or lecturing is blessed with all the skills needed to design assessment schemes and implement fair assessment. People are appointed to teaching (and assessing) posts not on the basis of how well they can do either task, but often on the record of their own academic performance.

Suggestions:

• assess the assessors. Have a system of 'licences' to assess, and police the system thoroughly. Moves in this direction are already underway in the world of training.
• increase the uses of self-assessment and peer-assessment, which depend far less on subjectivity of assessors, and allow far greater amounts of feedback to contribute towards successful learning experiences.
• provide training in assessing. Usually, a new lecturer is simply thrown in at the deep end. Participating with colleagues in a series of workshops on assessment-design can be much less intimidating, and allows discussion and feedback from experienced assessors.

9 'If you can' t measure it, it doesn't exist. If you can measure it, it isn't it' . What should we be trying to measure?

It has been said that one of the main faults of our education and training systems is that we tend to teach people things that are already understood, instead of equipping them to understand new things. Assessment reflects this.

Suggestions:

• use self- and peer-assessment as an inherent part of learning processes, with the emphasis on learning rather than assessment outcomes.
• increase the use of a portfolio-approach to assessment. A balanced portfolio represents much more than the relatively trivial ability to answer examination questions against the clock.

10 Where stops the buck? Whose fault is it that assessment is so artificial? Heads of Department? Employers? Assessors? Validators? The Government? Yours? Mine?

If you imply that there is something suspect about people's abilities to assess, it is badly received! Assessment is something that is usually done privately rather than publicly and people go to great lengths to ensure that they retain privacy. Is such privacy really needed mainly because of the 'put down the number you first thought of' syndrome?

Suggestions:

• be brave - be prepared to experiment with new ways of assessing your learners particularly when these involve learners themselves getting to know how assessment works.
• remember that there is much you can do without anyone else casting critical eyes over it. Traditional assessing is normally a rather private activity - take advantage of the privacy to try non-traditional ways of assessing.

Learning through assessing

In much of my discussion so far, I've been focusing on the dangers when assessment is 'done to' people, and hinting at the benefits which can result when learners themselves are allowed to be intimately involved in their own - and each other's - assessment.

Self-assessment and peer-assessment may lack some of the precision of the best of 'formal' assessment - where (some) assessors have a great deal of experience and (sometimes) assess fairly and conscientiously. However, what may be lacked in terms of precision is more than compensated for by the benefits of deeper learning, which go hand in hand with the act of learners themselves assessing.

Close encounters with assessment criteria

This is the crucial difference between formal assessment and self-assessment or peer-assessment. Learners find out a lot about any subject simply by applying assessment criteria to examples of work in that subject (whether the examples are self generated, made by other learners or issued by a teacher). Previously, assessment criteria have seemed to learners to be the property of examiners. There has been a tendency for teachers to regard assessment criteria as quite private.

For many years, I marked O-level and A-level scripts for two of England's public examinations boards. The marking schemes were invariably labelled 'Secret' and sometimes I had to sign a declaration to the effect that after the marking was completed I would destroy all information relating to the marking scheme.

In Higher Education, it is usual to have some sort of external moderation of assessment, usually involving sending draft examination papers to lecturers in other institutions for feedback or approval. Even where model answers and marking schemes have been required to be sent to external examiners or moderators, the vital information in such schemes has seldom been shared with learners, and until relatively recently hardly ever applied by learners themselves. Yet when learners have the chance to get their hands on assessment criteria, they seem to develop a thirst for the information they can derive from them - leading to much deeper learning.

Self-Assessment and peer-assessment are not just self-testing

These forms of assessment when well-developed involve several processes:

• involving learners in identifying standards or criteria to apply to their work
• agreeing with learners the relative importance of different criteria, by developing a suitable scale of 'weightings'
• allowing learners to make judgments about the extent to which they have met these standards and criteria
• providing the opportunity for learners to discuss with each other (and with their tutors) matters arising from their assessment of their own work or each other's work
• facilitating discussion of the fairness and objectivity of the assessment

In other words, self-assessment and peer-assessment have a tangible learning payoff, not least that associated with the quantity and quality of feedback which learners gain about their own performances and (in peer assessment) about each other's performances.

Assessment criteria: black and white or shades of grey?

In subjects like maths, science and engineering, things are often either right or wrong - and it is relatively easy to devise assessment criteria for tests and exercises. However, even in subjects such as law or social studies, there are identifiable hallmarks of a good answer or an unsatisfactory answer to a question. Such hallmarks can be turned into checklists of a flexible kind, which enable the characteristics of good and less-good answers to be compared and contrasted. Students can benefit by learning in the act of applying assessment criteria to their own and each others' work.

Benefits to learners of close encounters with assessment criteria

Learners can quickly find out about incorrect assumptions they have been making. They are able to find out more and more detail about the answer to the crucial question: 'What am I expected to become able to do?'

There are, of course, many more benefits, depending on how we involve learners in using assessment criteria - including helping learners themselves formulate the criteria (when this is possible or appropriate) - leading to the most obvious form of ownership of assessment.

Some examples of self- and peer-assessment mechanisms

Self assessment is not confined to the variety that is widely used in open and distance learning (though of course that is one powerful form of it). Self- and peer-assessment processes can include any of the following:

• providing learners with assessment criteria devised by tutors, a marking scheme and allowing them to mark their own work
• as above, but then allowing learners the chance to compare their mark with that of a 'professional' marker
• as above, but also giving learners feedback about the quality of their self-assessment
• enabling individual learners to generate assessment criteria and use them to assess their own work
• enabling a group of learners to generate assessment criteria, and so on
• allowing learners to use core criteria generated by a group, plus additional criteria specific to their own pieces of work, with an agreed weighting
• groups of learners can be issued with criteria to apply to each others' work
• groups of learners could produce criteria and apply them to each others' work

There are further combinations of these. There is also the additional matter of whether the grades or scores contribute in a formal way to the performance records of learners.

Facilitating learners' ownership of assessment criteria

Where it is possible to draw assessment criteria from learners themselves, especially in group situations, the sense of ownership which learners develop is very powerful, and leads to them using the criteria with considerable enthusiasm and commitment when self- or peer-assessing. It is of course important to explain to learners the benefits they can draw from taking part in assessing their own or each other's work before proceeding further.

I have found that the following approach gives useful results for groups of 10-20. I have spelled out the process in some detail below, but you will often be able to take short cuts and simplify the task in the light of your own experience. Nonetheless, I have found that the whole of the sequence below can be completed in less than an hour with groups of around 25 learners.

Suppose learners are about to do task 'x' (where 'x' could be to write an essay, give a presentation, prepare a project, write a report, design a handout, and so on). After a general introduction to clarify the nature of the task, the time available and the general format of the final product, proceed as follows.

1. Ask each learner to privately list (e.g.) six things you would expect of a good 'x'
2. Ask learners to go into groups of 3 or 4 and discuss the criteria they have devised.
3. Encourage the groups to 'refine' the criteria, by (for example) turning them into checklist questions, and avoiding 'subjective' words such as 'interesting' or 'scholarly'.
4. Ask the groups to make a list of the most important of the criteria and to prioritise them.
5. Make a flipchart of criteria, starting by asking for the most important criterion from each group in turn, then the next most important, and so on. Naturally, expect overlaps and duplication among the criteria from different groups, and cluster them together as well as possible on the flipchart, or combine overlapping criteria while building up the list.
6. Ask the whole group whether anything important is missing from the flipchart list. There will usually be some further criteria worth adding at this stage, or some further adjustments to be made to the list.
7. Tidy up the flipcharted items if necessary, or even edit the list onto a new flipchart, and then number the criteria, e.g. 1-8
8. Ask each learner to privately distribute (e.g. 30) 'marks' among the criteria.
9. Where the group size is 30 or less, ask each of the learners in turn to speak out the score they allocated to 'criterion 1', writing the scores beside the criterion on the flipchart. Then do the same with 'criterion 2' and so on.
10. Normally, it is very clear which of the criteria are the most important ones from the scores allotted to them by most of the learners. Also, there may be one or more criteria attracting very low scores and it is often worth deleting these altogether.
11. Either average the scores out, or allow each of the learners to apply their own weightings in the peer-assessments to follow.
12. Brief the learners then go off to prepare to perform the task to be assessed (individuals or groups)
13. Produce and copy a grid with their criteria and weightings listed on it, ready for peer-assessment. (See Fig. 4 for a typical grid structure).
14. All learners then enter scores onto their copies of the grid, relating to each of the successive 'performances' (e.g. presentations or multiple-assessments of photocopied written work). It is particularly useful to additionally allow learners to enter scores self-assessing their own performances, allowing them in due course to compare their self-assessment with that given by their peers.
15. The grids can be signed by learners, or handed in anonymously and average scores can be computed for each of the overall performances - and for each of the criteria too if it is wished.

Peer Assessment Grid

The raw data are presented in Table 1.

SETFY Presentations: 2 December: 1300-1700

Allowing for individual differences

The sort of peer-assessment described above is suitable for tasks such as presentations, where many people can assess the same piece of evidence, and where scores can then be compared and discussed by the group. For individual tasks such as essays, reports, projects, dissertations and so on, it is likely that each piece of work will reflect slightly different criteria (or even very different criteria) and then it is often best to allow for some 'agreed' criteria, and some 'idiosyncratic' criteria so that each learner can exercise more ownership of the assessment criteria. An example of a grid that can be adapted for such purposes is shown in Fig.6.

Self Assessment Grid

The most important outcome of involving learners in the formulation of self-assessment or peer-assessment criteria is that learners address the task with criteria in their minds and the quality of their work seems to be much higher than it may otherwise have been. In other words, there is no longer the 'hidden agenda' of the criteria in the mind of the assessor. Instead, when learners use self-assessment and peer-assessment, they are already in possession of the 'rules of the game' and will strive particularly hard to live up to these rules, particularly when they have helped to make them in the first place. Furthermore, being involved in devising and applying assessment criteria gives learners a useful insight into how other parts of their work (for example, formal examinations) are likely to be assessed, and helps them to work out more about the 'rules' (or 'wiles'?) of exam marking too.

Throw away the numbers or grades?

I've often suggested to learners after a peer assessment exercise that the numbers or grades they awarded (if not contributing to their overall assessment) were only a vehicle to help them do learning of a higher productivity. However, I've found they usually want to hang on to the numbers - good or bad. Perhaps this is evidence of the sort of ownership we're aiming for?

How well can students assess themselves, and each other?

In general, students are quite accurate in their assessing. I have found that when students are asked to 'guess' their own performance scores just after completing an exam, around 90% of students 'guess' within 5% of their actual scores. It is useful to identify the 10% who had an inaccurate perception of how they had done - they usually benefit from a discussion to probe the causes. Those 10% may be over-anxious and underestimate their achievements, or over-confident and over-estimate their achievements. When discrepancies in self-assessment occur, they are usually due to one of the following causes:

• there is some tendency for learners to over-rate themselves in areas to which they are new; this tends to happen with the weaker members of a group.
• there is a tendency for some learners to under-rate themselves in areas in which they are experienced; this tends to happen with the more-skilled members of the group.

Peer-assessment and self-assessment can be usefully combined. Peer-assessment can be conducted 'blind' so that 'arranged' scoring is avoided. If the peer-assessment mark or grade is equal to the corresponding self-assessment mark (e.g. within 5%) then the self-assessment marks go forward into the assessment system - possibly with a staff 'scan' to ensure that fair play is in operation. (It is far quicker to scan a piece of work to check whether the assessment is fair, than it is to mark the work from scratch). When self- and peer-scores differ, negotiation or staff intervention maybe necessary (but this happens surprisingly rarely in practice).

 

But can't it all go wrong?

There are several things that can go wrong with self-assessment and peer-assessment and to use either process successfully, it's worth knowing the potential dangers.

'You're paid to assess me - why should I assess myself?'

There will always be some learners who regard self-assessment or peer-assessment as an abdication from duty by tutors. My own reply is along the lines 'Certainly, I'll assess your work if you wish. But the real aim of involving you in assessing is not to save me work, but to allow you to gain much greater understanding not only of your subject material, but also of how assessment works'.

'But we haven't the expert knowledge to use to assess'

This, of course, is a real concern of many students faced for the first time with the task of assessing their own - or their colleagues - work. Tutors are regarded as 'expert witnesses' blessed with the experience with which to make authoritative and valid judgments. The key to solving problems of this nature is clarification of assessment criteria. When the criteria are phrased in language that learners can readily understand, their reservations about being able to make judgments based on the criteria are rapidly dispelled. In addition, with peer assessment, it is wise to spell out the advantages of a multiple judgment rather than the single one that might be given by a tutor. The average of several people's opinions is always likely to be more meaningful than a single opinion - even when the single opinion comes from a 'figure of authority' i.e. a tutor.

'What about passengers in team work?'

Often, when the products of team work are being self- or peer-assessed, the matter of 'contribution' comes up. One way is to award a score or grade to the overall product of the work of the team, whether it is a presentation, a report, an exhibition, or so on. Then, the team can be asked to 'split' the overall score between its members, in terms of the contribution each member made to the final product. Obviously, it's best if this can be done 'publicly' with all team members reaching agreement about the split of the award. Alternatively, however, it can be done by 'secret ballot', with the possibility of tutor intervention if necessary. Surprisingly perhaps, it is seldom necessary to invoke such extremes, especially if the issue of 'passengers' has been addressed by the whole group in advance and some ground rules on 'contribution' drawn up with the approval (better still, ownership) of the group as a whole.

'Is it the novelty that makes it work?'

Usually at present, when self-assessment or peer-assessment are introduced into a course, they have 'novelty value'. Because they are different, students tend to take them seriously and engage fully in the right spirit with their increased responsibilities. If such methods of assessment become commonplace, they would be regarded as part of the normal system, and human nature dictates that students would seek ways of 'playing the system' as in many other aspects of life. So, to some extent, it has to be admitted that the genuine enthusiasm and objectivity that students display for well-arranged self-assessment or peer-assessment has something to do with the novelty-value.

'What if the group just does not gel?'

Especially when assessing the product of group-work, there is always the possibility of a particular group having problems, possibly due to a single member. One way round this is to allow new groups to be formed for successive tasks, so that the effect of any disruptive or uncooperative individual is spread rather than concentrated. However, this can lead to what a colleague of mine referred to as 'sieving' of students, ending with the weakest all being left in a single group.

'But won't students be far too kind to themselves when they self-assess?'

Tutors are genuinely concerned, sometimes, that left to their own devices, students win an award themselves first-rate scores or grades. However, it is often the case that students' assessments are 'harder' than staff assessments. Admittedly, there are a few students who will self-assess their own efforts unrealistically 'high', but such instances are rare, and stand out clearly. If we are honest about it, students are in a better position to self-assess their own efforts than anyone else is. Even if they rate themselves too high, they almost certainly know they are doing just that. The real payoff is in the reflection that they engage in when making their judgments about their own work, whether rightly or wrongly. In other words, self-assessment is not a means to an end, but an end to a means - i.e. a way of helping students to reflect deeply. In the model of learning I use throughout this book, self-assessment is a very useful process which aids 'digesting'.

The discussion above shows that self-assessment and peer-assessment can certainly 'go wrong'. If either process is introduced by tutors who do not believe in the benefits that can be achieved, or to learners who are reluctant to engage fully with the processes, it is not hard to see how either process will be doomed to failure. It is important not to 'impose' innovations such as self- or peer-assessment against the will of tutors. Tutors are highly intelligent professionals and if forced to do something they don't trust will ensure that it demonstrably fails to work!

Self-Assessment, Peer-Assessment and how people learn

Throughout this chapter, I've advocated the benefits of helping learners become intimately involved in processes of assessing and I've pointed to the hazards of traditional assessment procedures. To conclude, let's look once again at the four main processes of learning I introduced in Chapter 1 and mentioned briefly earlier in this chapter and see how the processes of self- and peer-assessment relate to how people learn.

• wanting to learn (motivation)
  Motivation can be improved by early success. Self-assessment in particular can be used with the comfort of privacy and learners can gain confidence by finding that they are 'doing alright' long before they need to prove so publicly or formally. Also, motivation is greatly increased by 'knowing the rules of the game' of assessment. Self-assessment and peer-assessment take away many of the 'fear of the unknown' feelings about assessment.
• learning by doing (practice, learning from mistakes, and so on)
  There's no better way to find out about one's successes and failures than by finding them out for oneself in the comfort of privacy - or having a peer help one do this, rather than an 'authority figure' like a tutor or examiner. The very act of assessing is intrinsically 'learning by doing' - it involves the application of criteria, decision-making, judgment, and reflection. In other words, assessing is a 'deep' activity rather than a ' surface ' one and avoids the passivity which can pervade many less-active forms of learning.
• learning through feedback (to develop positive feelings about the learning)
  The 'showstopper' of formal assessment has to be the dreadfully limited feedback that is the norm - i.e. just receiving a number or a grade. Peer-assessment can allow for a great deal of feedback - far more than could ever be given by a tutor or assessor (especially when class sizes are increasing). In addition, the feedback gained in peer-assessment is usually far less threatening than that from 'professional' assessors. It is therefore received in a more-relaxed way, without the heightened emotions often involved in receiving feedback from a figure of authority. Indeed, peers will often argue and debate issues, further deepening the usefulness of the feedback exchanges they receive.
• digesting (taking stock, making sense of the experience, and of the feedback)
  Both self-assessment and peer-assessment can help learners make sense of their learning experiences - and of the feedback they gain. Furthermore, the time lag between the learning and the feedback can be much less than with traditional methods of assessment. Therefore, the feedback is much more actively received and the learning thereby enhanced.

Reflecting on assessment

It can be argued that people who need a 'tester' are inadequately prepared to be sent out into the world outside. Self-assessment and peer-assessment can both be important parts of the learning process. The learning experience resulting from such forms of assessment is more important than the result of the assessment. Self-assessment or peer-assessment do not necessarily have to lead to any 'formal' (recorded) assessment. The aim can be purely as a learning experience, with the 'marks 'simply part of the process through which that experience is facilitated.

Self-assessment and peer-assessment are skills, and become more reliable with practice. Receiving feedback on the quality of these forms of assessment is vital if learners are to derive the maximum benefit from engaging in them. Self-assessment and peer-assessment should be introduced early -for example during the first term rather than being left till the final year. Late in a course students may see little point in embracing new ways of learning.

'Ownership' is the most crucial aspect of successful learning and both self-assessment and peer-assessment are closely connected to the development of ownership of learning.

Not all students warm to the 'exposure' of self- or peer-assessment. They may begin their studies with expectations that they will be assessed by professionals. 'What's in it for me?' they naturally may ask. They need to be convinced that self- and peer-assessment have direct benefits for themselves, and do not represent an abdication from duties on the part of tutors. Some tutors, however, feel it is dangerous to 'lose control' of assessment. If such tutors try to employ self- or peer-assessment, but constantly safeguard their right to step in 'should things go wrong', the whole concept of such forms of assessment is undermined.

So what about teaching? Admittedly, our students learn from us. But they probably learn more on their own, and they probably learn even more from each other. Much of their learning occurs in the immediate run-up to assessment of one kind or another - so the role of assessment is an important factor in the circumstances which accelerate learning. Therefore, perhaps our biggest contribution to our students' learning is directly associated with the quality of the assessment our students encounter - and has less than we'd like to think to do with our teaching activities. I believe there is a strong case for using self- and peer-assessment not primarily to assess, but as processes to enhance learning.

Finally, if and when we must resort to traditional, formal assessment, I believe that there is a great deal of room for improvement. The 'ten worries' I expressed at the beginning of this chapter may help to set an agenda for improving the quality of assessment. I end this chapter with ten recommendations.

What can we do to improve assessment?

1. Extend and develop the competence-assessment approach, where people accumulate evidence of the things they do, and when they're ready, bring the evidence forward for assessment.
2. Use assessment as a learning opportunity. Give people feedback on their performance, and help to improve it. Make the primary purpose of assessment that of helping people find out how their learning is going, so that they can adjust their strategies accordingly.
3. Use assessing as a learning process - help learners to self-assess their own work and peer-assess their colleagues ' work (whether or not the marks or grades are to 'count'). Let learners gain familiarity with the nature of the assessment criteria which they need to live up to. Help them find out what sort of performance they need to be able to give.
4. Change the culture where any professional can sit down and scribble down an exam question in a few minutes, to a culture where all important exams are made up entirely of questions which have been piloted, tested, evaluated and proved to measure desirable things.
5. Where learners need to take traditional assessment forms such as written examinations, build-in to their courses detailed advice on how best to prepare for exams, and how to tackle the task of giving written answers against the clock.
6. Develop team approaches to question setting and marking. Introduce at the very least double-marking as a standard for exams - and when possible multiple-marking.
7. Move away from the type of exams which mainly measure recall. For example, move towards open book exams, where learners' ability to 'make sense of information and apply it' can be tested.
8. Develop external moderation of assessment much further, for example by paying external examiners to 'blind-mark' a selection of examination scripts and compare marks with the internal assessors.
9. Pay people well for assessing, instead of just paying them for teaching. Teachers often earn 'holiday money' by doing extra exam-marking. It's all the wrong way round! Assessment is so important that perhaps it should be that assessors can earn 'holiday money' by doing a little teaching! Then require them to continuously demonstrate their competence at the job of assessing.
10. Take the results of traditional assessment less seriously, especially when selecting candidates for jobs. There are countless tales of 'well qualified' people turning out to be entirely unsuitable for the real world of employment, yet almost every application form devotes a lot of space for educational records.

Reference

Saunders D and Race P (1992) (eds) Developing and Measuring Competence: Aspects of Educational and Training Technology XXV Kogan Page, London.