In this study, reading comprehension tests of narrative and informational text were created using common theoretical models and by following specific design steps (DeVellis, 2003). They were administered to 401 French-speaking 8-year-old students. The data collected underwent item, factorial and multilevel analyses. It led to examine the success rates for each test, to better determine their components and to compare them, in order to understand their usefulness for teachers and researchers. Results reveal that the two tests have a similar factorial structure. However, open-ended questions targeting anaphoric, causal and lexical inferences are more difficult to achieve, especially in the test using an informational text. These new instruments can help identify teaching avenues and promote a better transition between the first and second cycles of primary school in reading.