This intervention study examined the effectiveness of instructional support tailored toward two techniques (i.e., random associations and constraint identification) to strengthen children's creative problem-solving skills within upper-elementary science education. Five inquiry-based science lessons with ample opportunity for creative problem-solving (i.e., divergent and convergent thinking) were provided. Children were assigned to a condition with instructional support (n = 107) or without (n = 134). Domain-general and specific measures of divergent and convergent thinking were included, and reading comprehension as well as mathematical ability were taken into account. Repeated measures multivariate analyses of covariance revealed how all children improved in terms of domain-general convergent thinking, with a larger increase for children who performed better in mathematics. This shows a promising premise for future research focusing on the domain generality of convergent thinking and for the potential of transfer across domains. No additional improvement based on instructional support was found and children did not improve in terms of divergent thinking. The constraint identification and random associations technique might not be suitable for elementary school children, yet future research is necessary to validate such claims. Meanwhile, teachers could possibly support convergent thinking by simply providing exercises for divergent and convergent thinking.