This article examines the problem of learning to learn from the perspective of research on the organization and quality of students' knowledge. This approach is based on the assumption that students' learning competencies can be studied through the analysis of the outcomes of schooling. The article synthesizes findings of a long-term research programme assessing the development of components of propositional and procedural knowledge, and relationships between them. A model was devised classifying indicators of school outcomes on four levels: (1) teachers' grading; (2) objective knowledge tests; (3) indicators of quality of knowledge (application, understanding); and (4) higher order thinking skills. A set of tests and questionnaires were administered to Year 7 and Year 11 students in two consecutive projects; then, relationships between them and differences between the two cohorts were analysed. As results indicate, a number of points in schooling should be changed in order to improve schools' impact on learning to learn, for example, inadequate feedback given by grading, and shifting attention from reproductive to productive learning. The instruments developed in this programme assess science application, science misconception, mathematical understanding, historical reasoning, and inductive, deductive, spatial and critical reasoning. They can also be utilized for monitoring learning to learn. (Contains 2 tables and 2 figures.)