RESUMO
Productive Failure (PF) facilitates students' conceptual knowledge by delaying instruction until after problem solving. While PF is well investigated in middle and high school students, little is known about its effectiveness in younger students. Studies in younger samples, which implemented delayed instruction designs similar to those used in PF studies, showed mixed results. However, these studies did not implement two core design components of PF: (1) contrasting and comparing student-generated solutions and the canonical solution during the instructional phase (contrasting activity), and (2) students' collaboration in small groups during the initial problem-solving phase. Both components can be expected to contribute to the effectiveness of PF. In a quasi-experimental study with 228 fifth graders, we implemented the first component (contrasting activity) with all students to establish whether under this condition, problem solving prior to instruction would be more effective for younger students' conceptual knowledge acquisition than direct instruction (i.e., problem solving after instruction). Further, we experimentally tested the effect of the second component (collaborative vs. individual problem solving) on students' conceptual knowledge and the number of solution ideas generated during initial problem solving. We found no empirical support for either of our hypotheses. To explore the extent to which students' collaboration actually achieved its potential and relates to students' conceptual knowledge and solution ideas in PF, we conducted analyses of collaborative processes. Our study adds to the mixed results regarding the superiority of problem solving prior to instruction for young students, thus opening the discussion about age-related prerequisites as boundary conditions for PF.