This paper discusses the results of research on learning LOGO programming language by two groups of fifth grade students who were taught using either a structured tutorial mode or an unstructured discovery method. Previous results showed mastery differences between the structured and unstructured teaching methods with the results favoring the structured approach. Using classroom observational techniques, the present study examines how students in the LOGO classes compare with students in traditional situations for such factors as time on task, discipline, and organization. Additionally, this research asks whether techniques other than mastery measures can be used to differentiate LOGO classes taught by structured and unstructured techniques, and whether there are observable classroom behaviors that differ in classes taught using the two different teaching methods. Findings indicate that while traditional methods of classroom observations may be appropriate to differentiate traditional classrooms from LOGO classrooms, they may be inappropriate to differentiate various LOGO learning environments; and that, while it may be possible for LOGO classrooms to look as if adequate learning is taking place, mastery may not happen. It is suggested that research in LOGO to specifically measure mastery is needed before claims of transfer can be taken seriously. (JB)