A study investigated the effects of two methods of teaching college level composition on writing development and, secondarily, the effects the two methods had on overall improvement by gender. The treatment method involved the same syllabus, texts, assignments, and workshop approach as the control method, with the exception of the independent variable--the treatment sections received eight class periods of 50 minutes each in guided peer group evaluation training and practice in revision. Participants included 65 college bound high school seniors. In both methods, the instructors discussed the assignment, analyzed suitable topics with the students in small groups, engaged them in prewriting activities, helped them develop strategies, and discussed an appropriate organization for the modes and aims under consideration before they had the students attempt a rough draft at home. The holistic and analytic scores of student pretests, assignments, and posttests agreed on four variables: (1) the instructors' performances were extremely even; (2) the treatment and control groups were reasonably close in ability; (3) the treatment (training and extensive practice in peer evaluation) had no noticeable power by itself over the control (no peer evaluation); and (4) the females outperformed the males. (HOD)