Ten sections of Freshman Composition were randomly assigned to either the experimental or control condition. The experimental method, in order to motivate students to revise and rewrite, delayed grades until revisions were completed. The control method used incidental revision with immediate grades. Five instructors each taught one experimental and one control section. Instructional procedures,
... [Show full abstract] textbooks, writing assignments, and methods of marking were identical for both groups. Two tests of writing proficiency were applied in a pretest-posttest design. The results showed that for the English Expression Tests the control group was significantly better than the experimental group and that interaction between teacher/class and method was highly significant. For the essay test, the interaction was also highly significant, but there was no significant difference between the methods. (LL)