Task-based language assessment represents a major component of task-based language teaching syllabi. Current perspectives emphasise the importance of tasks in the assessment process, suggesting that adherence to influential models of language production during task design yields predictable test outcomes. The current study contends that the significance of the rater has been overlooked, resulting in adverse consequences when employing current task-based frameworks. The article reviews literature on the rating process, with a focus on interactions between rater characteristics, rating scales, and rater effects. Drawing on the findings of the review, the author proposes a revised model of task-based language assessment that curbs the impact of task design on test scores. The revised model underscores the need to regard raters as active agents in the assessment process to enhance validity and fairness and makes recommendations about minimizing rater effects.