Language teachers are beginning to adjust to the idea that, from 1998, GCSE examinations will restrict the use of English in the questions and increase the use of the target language by pupils in their answers. The new criteria contrast starkly with those which originally paved the way for the introduction of the examination 10 years ago, in particular with respect to the recommendation at that time that comprehension, as a discrete skill, should be tested through English.

Some claim that insistence on assessment in English has created a backwash effect on classroom practice, especially in Year 11 as the examinations draw near. It certainly resulted in the publication of teaching materials overloaded with English. The fact that 10 per cent of questions and answers in the examination overall may be in English goes some way to appeasing those most vociferously opposed to the new criteria but many still fear that there will be less variety in the types of tests used and that those which are approved will be less “authentic”.

Extensive exposure to the target language through course books and other media, and judicious use of the language by the teacher are generally accepted as “good practice”. Discussions now centre on how to implement such a principle across the full ability range, how to promote target language use among teachers and learners and the definition of situations in which the use of English is both advisable and justified. While the theoretical debate has continued, very little empirical research has been carried out in schools and there is still considerable anxiety in the profession about the whole issue of target language testing.

The extent and nature of teachers’ anxiety was one of the aspects examined recently in a research project at the University of Warwick, funded by the School Curriculum and Assessment Authority. One hundred and twenty teachers in a range of schools in the Midlands completed a questionnaire designed to elicit attitudes about target language teaching and testing. While they were generally favourably disposed to the use of the language in the classroom, responses to testing in the target language were rather more negative. By far the most frequently cited causes for concern were that target language tests would be unfair to less able pupils and would put extra pressure on teachers. Some also feared that it would be impossible to discriminate between levels of performance. Those who considered the introduction of target language testing as a positive development argued that it would further promote communication skills and that it would raise standards.

The main focus of the study, however, was the evaluation of the effectiveness of a range of test types in listening and reading in French and German in 12 secondary schools. In total, 27 tests were created in French and 31 in German and more than 600 pupils were involved in the testing itself. It was decided to test not only Year 11 pupils but also pupils in Years 10 and 9, using the same material. This made enormous demands on the younger pupils but, in the event, some managed the experience extremely well. These year groups, of course, have been taught throughout their contact with French and German within the national curriculum and consequently are more likely to have been exposed to target language instructions by teachers and in textbooks.

The fact that some pupils in Years 10 and 9 were subjected to tests really designed for use at the end of the GCSE course was not the only unusual feature of the research. Pupils had no specific preparation for the tests, which were entirely in the target language, and were meeting some of the test types for the first time. From 1998 candidates will be allowed to use dictionaries in reading and responding, and writing tasks but these pupils were not allowed to use dictionaries or glossaries or rubric or instructions.

Perhaps most significant was the fact that the question booklets made no distinction, in terms of foundation tier or higher tier, between the tests. Some pupils with low grade predictions made very acceptable attempts at questions which were designed to challenge the most able.

All these factors made the pupils’ tasks more challenging than will be the case in the real examinations in two years’ time. Despite these difficult circumstances, however, most pupils tested performed broadly in line with teacher expectations as recorded by predicted grades. Where there were difficulties, they tended to arise from an understandable lack of knowledge of vocabulary in the stimulus material rather than in the instructions. Observation and “think-aloud” activities with individual pupils in this research indicate that scant attention is paid by pupils to the language used to set the context for the test. What is much more important is the transparency of the tasks to be done as demonstrated by the format of the test and the use of examples.

Target language testing may not be to everyone’s taste but this research shows that even without preparation pupils can perform well, some even better than expected, provided the tests are clear and logical in their format.

The main purpose of the research was to advise SCAA, GCSE examining groups and language teachers about the feasibility of a range of test types across the ability range. The project team set out to provide objective evidence which, it is hoped, will be of real value to those who are charged with producing the examinations under the new criteria. We also deemed it essential to support directly those whose responsibility it is to prepare pupils for examination at the end of key stage 4 and to ensure that their performance in the new style GCSE examinations is of the highest possible standard. The report therefore also contains a detailed checklist for devising target language tests and advice on how to prepare pupils for this new form of assessment.

Copies of the report may be obtained from: Customer Services, SCAA, Newcombe House, 45 Notting Hill Gate, London W11 3JB Dr Bob Powell is director of the Language Centre at the University of Warwick.