Description:
<jats:p> In this article 'covert' language assessment refers to the implicit assessment of non-native speaking and writing that takes place when the focus of the assessment is on an examinee's understanding of an academic subject such as geography, history, biology or physics, and not on the examinee's language proficiency as such. Drawing on a larger study of assessment practices at the University of the Witwatersrand in Johannesburg, South Africa, the authors demonstrate that there was in some cases a lack of clarity in the extent to which proficiency in written English was taken into account in the assessment of the academic writing of second language students at the university. They argue that if universities are to be accountable to students, the criteria used in the assessment of assignments and examinations should be made explicit. More specifically, second language students need to know to what extent their performance on academic writing tasks is determined by the quality of their ideas and to what extent it is determined by the quality of their written expression - notwithstanding the complex relationship between them. Recommendations are made for the assessment of students in ter tiary institutions with multilingual student populations. </jats:p>