It is possible that the student himself may be a valuable resource in the individualization of instruction in the elementary school. This study paired a fourth grade student with a sixth grade student who had a similar reading skill level. The pair then used a kit of auto-instructional materials designed to teach elementary school science. Students at a control school received conventional instruction in science. The study showed that fourth and sixth grade students can learn from auto-instructional materials without teacher assistance. Analysis of data from the control and experimental students shows that there was no difference between the two groups in relation to achievement levels or self-concept scores. Neither sex, reading level, nor race affected achievement. Only social position affected achievement significantly. The report discusses these and some secondary findings and outlines the research methodology of the study. A reference list is appended. (JY)