A study at Lansing, Michigan Community College investigated whether the current policy of grading the students according to the number of errors in their compositions was actually helping the students to make fewer errors in their writing. Error data were collected from 4 graded essays from each of 311 students in 2 writing courses (one on exposition, the other on argument). Results showed a 50% reduction in errors, with the students making the most impressive gains in the first course, indicating that the rigorous error-count system of the school is getting positive results. Conclusions of this study are that marking papers does decrease error--just as was always thought. (Two appendixes showing error-reduction data and the grading procedures are attached.) (PRA)