The aim, in this qualitative semi-structured interview study, is to analyze what modalities a primary school teacher in a multicultural school context stresses when assessing her third grade students’ meaning making in their multimodal text productions. The study has a socio-cultural perspective and the theoretical points and analysis tools are based in the socio semiotic multimodal theories, and in the multiliteracies research. The results show that the teacher on the one hand mainly focuses on the semiotic resources, image, color and writing, and to some extent squares, symbols and speech bubbles in student petitions, but on the other hand her statements points out the complexity of the task formulation gives rise. It proves to be problematic for the teacher to allow students to freely interpret and independently design their task where she at the same time intends to make an overall assessment of how the content is presented. In practice, the teacher more often are focused on assessing abilities that show how thoroughly and carefully the students are documenting than developing the knowledge. It points to the difficulty of reconciling assessment requirements in the context of a multimodal text production.