Abstract

This study compared the effects of an experimental technique of grading papers before the errors were marked with a control technique of grading papers after the errors were marked. The effects were measured in terms of writing improvement and student satisfaction in a developmental English course for technical students. Data on writing improvement were collected in the form of T-unit length and frequency of errors per 100 words, and data on student satisfaction were collected in the form of scores on a semantic differential. Although statistical analyses indicated only slight differences favoring the experimental group on between-group and most within-group comparisons, a statistically significant reduction in errors per 100 words did occur within the experimental group but not within the control group.