Rater-task interaction effects on testing EFL learners’ written performances

Abstract

This study examines the interaction effect between rater groups and tasks on evaluating EFL learners’ written performances. Fifty raters took part in this study. The experienced rater group (n=25) and the novice rater group (n=25) judged sixty essays (30 narratives and 30 argumentative writing modes) written by third-year English students. Raters’ decision-making behaviours, in terms of scores assignment and written comments, were diagnosed based on different quantitative and qualitative tools. Scores were analysed based on FACETS to examine the effects of rater-task interaction on raters’ severity and internal consistency and the analytic scale’s functionality. Qualitative data were also analysed to diagnose which aspects of writing were deemed more important than others across rater groups and task types. The analysis revealed that both raters and tasks were substantially influential factors. The majority of expert raters displayed more severity in assessing narrative essays than argumentative essays. Different qualitative judgments are also detected across raters and tasks due to rating experience and task requirements. The findings of this study reflected implications not only for testing learners’ writing proficiency but also for test validation research in the task-based writing performance assessment field.