Main Article Content
Problem solving is a transversal skill that transcends individual disciplines and explores the applicability of knowledge. Mapping certain aspects of this skill which had not been investigated previously became a reality with computer‐based testing. The aim of log file analyses is to provide both a qualitative and quantitative description of exploration strategies used in mapping minimally complex, simulated problems. The participants in the study were sixth‐ to eighth‐graders (n=2226) and first‐year university students (n=1259). A latent profile analysis was conducted to ascertain the exploration strategies employed by the latent classes as they worked to understand the problems. Solutions were examined in two to eight classes. Based on the findings of this study, a distinction can be made between (A) six qualitatively different groups in the primary age group: (1) cannot process even the most fundamental systems; (2) grasps the simplest systems at a low level; (3) grasps simple problems with some success; (4) learns fast; (5) grasps simple systems well, but underperforms when working with more complex systems; and (6) uses advanced strategies; and (B) four qualitatively different groups at the university level: (1) low-performing strategy users; (2) proficient strategy users; (3) slow learners; and, finally, (4) rapid learners, the most valuable group from an educational point of view. No so-called intermediate strategy users were detected. The analyses bear out the assumption that school selection can be significantly explained by the developmental level of students’ problem solving abilities. The true benefit of the latent profile analysis‐based approach is that it has confirmed the hypothesis that development can be described not only with quantitative change, but also with qualitative change. An exclusively quantitative analysis is insufficient, as it would lead to false conclusions in this case.