ABSTRACT Purpose To investigate the influence of order of reading tasks application on the reading comprehension performance by students with and without reading difficulties. Methods 40 students (4th and 5th grade) were classified according to the presence (Research Group-RG) or absence (Control Group-CG) of reading difficulties. RG-r and CG-r - 20 students (10 for each group) who retold the read text and responded to open-ended questions; RG-q and CG-q - 20 students (10 for each group) who responded to open-ended questions and then retold the read text. The analysis quantified the main idea, details and inferences retold, causal links and retelling reference standard (3-0) was also established from the best to the worst performance. Open-ended questions received one point for each correct answer. Results Open-ended questions influenced only the retelling performance of good readers. A better performance of CG-q was noted for the number of second level links retold (U=50.50, p=0.155), total of links retold (U=23,00, p=0.038) and retelling reference standard (U=24.50, p=0.039). Reading-monitoring strategies are laborious and tend to be less used by students with reading difficulties. This is because these compete directly with low-level skills (decoding and microstructure processing), losing efficiency or being abandoned in the very course of reading. Conclusion There was improvement on the retelling performance of students without reading difficulties when this task was preceded by the open-ended questions, possibly because of the use of monitoring strategies that allowed a better understanding of the link between the retained ideas, improving links and retelling reference standard.