نسیمه نوحی جادسی

Using a mixed methods design, the present study attempted to identify the factors influencing Iranian untrained EFL raters in rating group oral discussion tasks. To fulfil this aim, 16 language learners of varying proficiency levels were selected and randomly assigned to groups of four and performed a group discussion task. Thirty two untrained raters were also selected based on their volunteer participations. They listened to the audio files of the group discussions and assigned a score of one to six to each language learners based on their own judgments. They also provided comments on each language learners’ performance pointing to why they assigned such scores. The researchers had an interview with the raters after the rating session as well. The quantitative phase investigated whether linguistic features of accuracy, fluency, complexity and amount of talk were attended to by the raters in terms of having any relationship to the scores the raters assigned. Speech rate as an index of fluency and amount of talk turned out to be significantly correlated with the scores. Of more importance was the qualitative phase with the aim of identifying other factors that may account for the scores. The comments provided by the raters on each score and the interviews were codified based on Content Analysis (CA) approach. It was found that the raters attend not only to the linguistic features in rating oral group discussions, but they are also sensitive to the interactional features like the roles the participants take in groups tasks and the overall interaction patterns of the groups. The findings of this study may shed light on group oral assessment in terms of training the raters rating group oral tests and developing rating scales specific for group oral assessment.