The instructional design and student leaming literature is reviewed for guidelines for instruction which would encourage deep rather than surface learning. A taxonomy is presented which values student conceptions of key phenomena and skills for the self-discovery of knowledge as more important than the accumulation of infonnatinn. Strategy elements suggested for the selection and sequencing of content, therefore, focus on revealing the interrelationship between key concepts. Evidence is presented of the persistent nature of existing conceptions and the difficulty of changing conceptual frameworks. Diagnostic questions are suggested as a means of exposing existing conceptions. It then seems necessary to provide a challenge to revealed or anticipated misconceptions so that students pass through a di~quilibrimn phase before re-forming their existing conceptions. As there is growing evidence of a mis-mal~ between the goals and practice of teachers, action research is suggested as a method of implementation.