Linguistically based scales for assessment of young students' writing

This study addresses the question of how different aspects of students' writing achievement can be recognised and evaluated. We developed a linguistically based framework for criteria-based assessment, anchored in a functional view of language and language learning. The framework was used to determine what traits characterise texts at different Proficiency Groups based on comparative judgement and what traits characterise texts assessed differently. Altogether, 100 texts (written by students ages 6-9) representing four text genres were assessed and ranked using both comparative judgement (holistic assessment) and criteria-based analysis. The results indicate that texts generally are assessed as stronger (i.e., placed in a higher Proficiency Group) when comparative judgement is used than what the assessment of a specific language resource indicates. The results also indicate that assessment differences might be a result of different quality expectations for different genres. This points towards the need for genre- and subject-specific assessment criteria to scaffold students in their emergent disciplinary writing development.