In the process of designing teaching materials for learning problem solving in technology education, domain-specific design specifications are considered important elements to raise learning outcomes with these materials. Two domain-specific design specifications were drawn up using a four-step procedure and were pplied to improve existing teaching–learning packages. The study focused on a construction problem (open-ended) and an explanation problem (constrained). Construction material (Fischertechnik) was used to solve the problems. In two experiments, these newly designed teaching materials were compared with the existing teaching materials. In all, 600 pupils participated in these experiments. In the experiment with the construction problem, no learning gains were made at all: the small gain in quality of the product made by the pupils cost too much time. In the experiment with the explanation problem, the quality of the pupils’ product was significantly better in less time. It is argued that strongly structured teaching materials for constrained problems are more suitable for learners with little experience with construction material.