This study deals with the testing and grading of Swedish school leaving students’ oral proficiency in English, and with certain aspects of these students’ linguistic competence. The analyses and results are based on material drawn from an assessment project carried out at Gothenburg University in 1993. The 29 students taking part in the project were interviewed three times by three different interviewers in tests comprising three tasks, similar in structure but different in content. The interviewers were of three categories: school teachers of English, university teachers of English and native speakers of English. The student production was graded on a five-point scale according to a set of rating criteria. The interviewers assigned generally positive but often differing grades to the student performance. The grades were influenced by the students’ ability to communicate and speak with flow, and by gaps in vocabulary and by occurrences of grammatical errors. The students’ use of discourse phenomena and compensatory strategies was also of importance to the grades assigned. Many students were considered to have acceptable intonation and rhythm, but nevertheless an evident Swedish accent. The linguistic features studied comprised the verbal group, vocabulary, discourse markers and pronunciation. Differences could be observed between the members of the interviewer categories regarding the grades they assigned to student production. The school teachers seem to have paid special attention to grammatical accuracy, and the native speakers appear to have had a notion of communicative competence where accuracy plays a less important role. Differences in the grades assigned could also be explained by the order in which the interviews were made, by some students’ hesitant delivery, by the positive or negative effect of various fillers in the students’ speech, and by the interviewing methods used by the interviewers in the tests.

<p>This study deals with the testing and grading of Swedish school leaving students’ oral proficiency in English, and with certain aspects of these students’ linguistic competence. The analyses and results are based on material drawn from an assessment project carried out at Gothenburg University in 1993.</p><p>The 29 students taking part in the project were interviewed three times by three different interviewers in tests comprising three tasks, similar in structure but different in content. The interviewers were of three categories: school teachers of English, university teachers of English and native speakers of English. The student production was graded on a five-point scale according to a set of rating criteria.</p><p>The interviewers assigned generally positive but often differing grades to the student performance. The grades were influenced by the students’ ability to communicate and speak with flow, and by gaps in vocabulary and by occurrences of grammatical errors. The students’ use of discourse phenomena and compensatory strategies was also of importance to the grades assigned. Many students were considered to have acceptable intonation and rhythm, but nevertheless an evident Swedish accent. The linguistic features studied comprised the verbal group, vocabulary, discourse markers and pronunciation. </p><p>Differences could be observed between the members of the interviewer categories regarding the grades they assigned to student production. The school teachers seem to have paid special attention to grammatical accuracy, and the native speakers appear to have had a notion of communicative competence where accuracy plays a less important role.</p><p>Differences in the grades assigned could also be explained by the order in which the interviews were made, by some students’ hesitant delivery, by the positive or negative effect of various fillers in the students’ speech, and by the interviewing methods used by the interviewers in the tests.</p>