The authors investigated whether the amount of instructional guidance affects science learning and self-efficacy. Sixty 9- and 10-year-old children were randomly assigned to one of the following three instructional conditions: (a) guided instruction consisting of examples and student-generated explanations, (b) direct instruction consisting of a lecture and examples, and (c) minimal instruction consisting of student directed discovery. Children who received guided instruction designed a greater percentage of experiments correctly and self-reported greater changes in science self-efficacy than children in the other conditions. No statistically significant differences were observed between direct and guided instruction on outcome measures of cued recall, application and evaluation. However, both conditions performed statistically higher on these outcome measures relative to the minimal instruction condition.