Measuring the immeasurable
Over the past year a General Teaching Council working party has been discussing the kind of partnership between schools, education authorities and teacher education institutions that is necessary for successful pre-service education of teachers in Scotland. A draft report will shortly be circulated for consultation, and is bound to stimulate debate inter alia about the assessment of student teachers on placement. This article is intended to contribute to that debate by outlining some of the thinking behind Stirling University's rather different approach to assessing student teachers.
Our department's current practices originated in intensive partnership with the profession during the Collaborative Project on Teacher Education in the 1980s. The project established a pattern of working parties with a majority of classroom teachers, which has kept our practices under review ever since. Two key principles were agreed: complete separation of academic assessment from assessment of teaching performance and the use of a profile of competences, without any traditional marks or grades. These principles have not gone unchallenged. Recently the GTC asked us to justify why we did not grade teaching proficiency, and include the grades in deciding honours classifications. It was argued that a student should not be able to get a first class honours "teaching" degree without being a first class teacher.
Our reasons for keeping the two assessments separate are both theoretical and practical: whenever we combine two different assessments, we lose information, and when the areas assessed are as different as the study of an academic subject and proficiency in classroom practice, then we risk seriously misrepresenting what students can do. Our objection to grading was also part theoretical, part practical. In principle, grading would involve writing grade-related criteria for the assessment of practice, and creating practical settings in which student teachers were given equal opportunities to display their competence. Neither is feasible.
In fact, examination of competence systems in teaching suggests it is not impossible to agree about broad categories of practice that would be accepted as "good". We can all agree that teachers should know their subject, should be able to keep order in the classroom, should be able to motivate their pupils, and so on. Problems arise when we try to specify exactly how these goals should be achieved. There are clearly many ways in which teachers can keep classes under control and stimulate them to learn, and these depend on the class itself, the school context and the skills and preferences of individual teachers.
Yet, despite previous experience, the national profile takes only the broad categories of the Scottish Office's competences, and supposes that teachers and tutors can be asked to assign grades to students that have national validity and comparability.
A conclusive argument against grading is that placement schools are so different: different in the demands they make and the opportunities they provide for students to acquire and display competence. Students respond differently: a student may find an "easy" placement more difficult than one generally reckoned to be harder. We cannot simply ask teachers and tutors to make allowances by adjusting grades up or down. That would simply add another uncontrolled variable.
We have to recognise that teaching is inseparable from its context, and that comparisons across contexts cannot be made "fair". Other professions accept this - neither nursing nor social work grades practical experience placements. So how can students be assessed fairly across a range of different contexts?
The Stirling approach is to emphasise the target of becoming a "competent teacher". A "competent teacher" will regularly, reliably, teach "competent lessons", ones in which substantial and appropriate goals are set and achieved, within the broad accepted consensus about "good" practice mentioned earlier, and across the range of contexts the school can offer. On these contexts, supervising teachers are the local experts.
We ask them to assess performance in context-related terms, bearing in mind the criterion "would you be happy for this person to join your staff next week?" They are not asked to guess at how students compare with others elsewhere, or with ones they vaguely remember from other years. (Grading implicitly requires such comparisons.) Visiting tutors are asked for a similar judgment - "is this student working successfully in this particular school setting?" - and not to speculate about how well the students might perform in other contexts where they have not been seen working.
This approach has many advantages. In reporting, it is easy to give equal weighting to school and university assessments, as required by the Scottish Office. Systems where grades are added or averaged can be problematic, particularly when, for example, grades from the teacher education institution are used to adjust those from schools. The Stirling system is simple: students must be judged competent by both parties. While either has doubts, the students does not pass.
Pressure to grade teaching arises partly from a tradition that "if something isn't graded it can't be important". We would argue differently. Classroom performance is too important to allow it to be distorted by pseudo-quantification, and lumped in with other things. But pressure also comes from claims that employers "need" grading "to aid selection" - or rather, to save them reading large numbers of reports they find bland and uninformative. The flaws in grading mean that grades have negligible predictive value, but they will be used in the absence of anything better.
The challenge for us is to write reports employers will want to read, because they describe what students know and can do at a useful level of detail. They would summarise school reports, making clear the kind of evidence on which these were based, and the contexts in which the performance occurred. Employers that took the trouble to study these reports would be compensated by having a far better basis for selecting staff.
During the revision of the national profile the need to improve written reports was accepted. There is plenty of relevant expertise available. What gets in the way is the notion that what we need is an instrument, one that will be precise and can be applied equally in every classroom, primary or secondary, across the country. Once this goal is conceived, a sort of "measurement madness" takes over. It follows that extensive training will be necessary because a central requirement of such a system is that supervising teachers must learn to ignore their own preferences and local knowledge (the very things that students find it so valuable to learn about). Either that or the instrument must be calibrated, to "correct errors", and provide a true context-free measure of student's innate teaching potential. It is a peculiar model of partnership.
To promote quality in teacher education we need to exploit what schools can offer, and that means accepting local diversity. Assessment must reflect this, even at some loss of uniformity and administrative convenience. The alternative has been weighed in the balance, and long been found wanting.
Peter Cope is director of the Initial Teacher Education Programme and Eric Drever chief examiner of the programme, both at Stirling University.