A critical assumption made in Kapur's ("Instr Sci" 40:651-672, 2012) productive failure design is that students have the necessary prerequisite knowledge resources to generate and explore solutions to problems before learning the targeted concept. Through two quasi-experimental studies, we interrogated this assumption in the context of learning a multilevel biological concept of monohybrid inheritance. In the first study, students were either provided or not provided with prerequisite micro-level knowledge prior to the generation phase. Findings suggested that students do not necessarily have adequate prior knowledge resources, especially those at the micro-level, to generate representations and solution methods for a multilevel concept such as monohybrid inheritance. The second study examined how this prerequisite knowledge provision influenced how much students learned from the subsequent instruction. Although the prerequisite knowledge provision helped students generate and explore the biological phenomenon at the micro- and macro-levels, the provision seemingly did not confer further learning advantage to these students. Instead, they had learning gains similar to those without the provision, and further reported lower lesson engagement and greater mental effort during the subsequent instruction.