To relate the design of open education programs to their outcomes, the authors surveyed 72 research studies on the effects of open education and examined the methods and findings of three previous literature surveys on this topic. The 72 studies were divided by an objective statistical test into those showing large effects of open education and those showing small effects. The researchers then compared seven features and four aspects of open education with six student outcome variables, controlling for nine characteristics of the research studies' design and methodology. Among the open education features were multigrade student groupings, open space, individualized instruction, and team teaching; open education aspects included the number and extensiveness of open education features in each study. Outcome variables comprised students' self-concept, creativity, attitude toward school, and reading, math, and language achievement. The results showed that the size of open education effects varied for different features and outcomes. For instance, for student self-concept, open education programs showing large effects averaged two more features than those showing small effects, while for student creativity, programs with larger effects had one more feature. (RW)