This paper describes the methods and outcomes of a 12-year longitudinal study into the effects of an early intervention program, while reflecting back on changes that have occurred in approaches to research, learning and instruction since the preliminary inception stages of the study in the mid 1960s. We began the study to challenge the prevailing consensus at the time that primary school children were either "preoperational" or "concrete operational" in their cognitive development and they could not learn abstract concepts. Our early research, based on Ausubelian theory, suggested otherwise. The paper describes the development and implementation of a Grade 1?2 audio tutorial science instructional sequence, and the subsequent tracing over 12 years, of the children's conceptual understandings in science compared to a matched control group. During the study the concept map was developed as a new tool to trace children?s conceptual development. We found that students in the instruction group far outperformed their non-instructed counterparts, and this difference increased as they progressed through middle and high school. The data clearly support the earlier introduction of science instruction on basic science concepts, such as the particulate nature of matter, energy and energy transformations. The data suggest that national curriculum standards for science grossly underestimate the learning capabilities of primary-grade children. The study has helped to lay a foundation for guided instruction using computers and concept mapping that may help both teachers and students become more proficient in understanding science.