This comparison of correction of written papers by peers with traditional correction by teachers in composition courses tested the hypothesis that a statistically significant difference would exist in favor of a peer-graded group. The theoretical basis of study was the influence on adolescents of the opinions of their peers and the motivating effect of writing for an audience of peers. In the experiment, 153 ninth-grade students were taught writing in three experimental and three control classes. Students in the experimental group were trained to evaluate each other's writings, both individually and in group discussions; they frequently used prepared guide sheets to plan the oral and written criticism. Control group writing was corrected by teachers after classes. All pupils took the STEP Essay and Writing Tests before and after instruction. The Essay Test scores were discarded as being statistically unreliable, and no significant difference was found between the groups in mean score gains in the Writing Test. The conclusion was that since no difference existed between the two methods of paper-evaluation, teachers should expand peer correction and should spend the released time on other aspects of composition teaching. A recommendation was made that the STEP Essay Test be updated. (Author/LH)