This study investigates meaning-making processes in language and literacy assessments. Using a social semiotic perspective, it examines how adult second language learners with emerging literacy self-articulate their understanding of multimodal elements and components utilized in low-stakes assessments, and the strategies they use to make meaning in such tests. The study examines how the intended meaning of assessment prompts aligns with the perceived meaning on the part of participants. To do so, we used an existing assessment tool developed for refugee-background adults in the United States as well as two experimental assessment texts created as part of the research. Data was collected via semi-structured interviews with 14 participants. Interviews were analyzed using a critical multimodal social semiotic approach, informed by systemic functional linguistics. Data was coded according to metafunction and theme, with results organized by test genre element and multimodal component. Results exposed tensions between participants' responses to textual and visual prompts and the expectations of test designers. Findings reveal that textual composition and assessment practices may be inadvertently biased against individuals with diverging literacy profiles. Results have implications for assessment of this population such that language, literacy, and test socialization should be considered during test design and evaluation framework development.