To evaluate composition programs based on peer feedback, a study compared the improvement in written composition skills of two sets of eighth grade students in 11 classes at eight Dutch schools. The control group was enrolled in teacher feedback programs, the experimental group was enrolled in peer feedback programs. Students wrote a discursive essay and responded to an attitude inventory questionnaire before, during and after the course. Writing aloud protocols were elicited from a small number of students. A random sampling of each class yielded 792 essays which were typed out and evaluated by two raters for the writing achievement variables of goal orientation, audience orientation, organization and style. The hypotheses predicted that, as compared with the control group, the experimental classes would exhibit (1) higher test scores on the writing achievement variables, (2) a lower level of writing apprehension, (3) more planning activities such as self-instruction and generation of ideas, (4) more frequent interruptions of the writing process by a choice process, and (5) more frequent evaluation and revision in the revision process. These hypotheses were not supported. The results showed no differences between the two groups, except for some differences in the writing process variables; self-instruction rose in the experimental group and fell in the control group; interruptions by a choice process rose in the control condition, and fell in the experimental. (A 19-page bibliography includes listings of numerous articles in Dutch.) (JG)