Assessments of literacy used in global monitoring are typically monolingual, even in contexts that have adopted bilingual education. Preliminary research shows that bilingual assessments – or assessments that aim to capture the full linguistic repertoire of bilingual students – may be fairer and more appropriate than monolingual assessments. However, the world is far from using bilingual assessments in higher stakes circumstances. There is not a clear path for transitioning from monolingual to bilingual assessment frameworks. More evidence is needed to consider bilingual assessments for evaluation purposes such as global monitoring. This paper uses data from a bilingual assessment to evaluate different representations of bilingual literacy in a linguistically diverse context like the Democratic Republic of the Congo. In addition, this study examines whether bilingual assessments of literacy provide equivalent information for students who come from different linguistic backgrounds, across multiple provinces. The results show that complex multidimensional conceptualizations of literacy may best represent bilinguals’ proficiency. Such complex representations can only be captured via bilingual assessments. However, this study suggests that bilingual assessments are not inherently equivalent across subgroups of students. The outcomes vary by provincial and language groups, raising doubts as to whether bilingual assessments could serve highstake large-scale uses. Policy recommendations are provided, considering the current challenges that bilingual education faces in contexts like the DRC. Most notably, countries should consider using bilingual assessments as a part of their efforts to strengthen bilingual education programs. This paper has been commissioned as part of the GEM Report Fellowship Programme in 2021.
Languages: