Now they're talking : testing oral proficiency in a language laboratory

Efforts to increase the teaching of oral FL skills in the Finnish senior secondary schools have often been less successful, mainly because oral testing has not been part of the influential national school-leaving examination. Therefore, an attempt was made to develop and try out an oral test of FL English, which could be used to test all secondary school-leavers - more than 30 000 at a time - simultaneously. In order to develop a valid test the nature of oral proficiency was analyzed. Bachman's model of language ability was chosen to be the basic framework, and as criteria of proficiency the following features were used: pronunciation, fluency, coherence, amount of information provided, appropriateness of the language. For the instrument of assessment a SOPI type of test, called the LLOPI or Language Laboratory Oral Proficiency Test, was designed. The test has a contextual communicative frame and consists of six parts: warming up, reading aloud a letter, interpreting the Finnish part of a dialogue into English, conveying a Finnish newspaper story in English, reporting on the Finnish school system, and coping with everyday situations and expressing opinions. The whole test lasts 40 minutes, and the recorded sample of the student's speech about 20 minutes. The subjects, 60 school-leavers from two schools, were also tested with the ACTFL interview. The main research task was to find out whether the LLOPI was a reliable, valid, and efficient means of testing the students. It was also explored whether the LLOPI could be validated with the ACTFL interview. Student attitudes towards speaking and testing English were investigated, as well as the effect of spending time abroad on oral proficiency. The LLOPI proved to test the students reliably and validly. The correlation coefficient of the LLOPI with the ACTFL interview was .78, and 60% of the subjects received the same result in both tests. The LLOPI turned out to be more efficient than the interview, but the efficiency could be further increased by shortening the test. It was discovered that the ACIFL interview is not a perfect means to validate the LLOPT, because the two tests highlight partly different aspects of proficiency. The students' attitudes towards speaking and testing the foreign language were positive. It could not be shown that staying abroad would have had significant influence on the speaking skill. The investigation indicates that it would be both feasible and beneficial to start testing FL oral proficiency in the school-leaving examination. At least in the first foreign language the most practicable means would be a language laboratory test