The validity and reliability of essay scores were examined by comparing the mean scores assigned to good and poor quality essay responses of different lengths written by high school sophomores. In-service and pre-service social studies teachers graded essay responses to a test question requiring knowledge of the Constitutional provisions for police power and the relationship of police power to the 14th Amendment. A scoring key was used. Mean scores were analyzed using t-tests. Comparisons between responses of different qualities and different lengths indicated that the long response had a significantly higher mean score than short or moderate length responses; that good quality papers were graded significantly higher than poor quality responses; and that the mean for the poor quality long response was not significantly different from the mean responses for good quality short and moderate length responses. Results indicated that the length of the response--rather than knowledge of the subject--influenced scores, and that a scoring key alone was not sufficient to insure that quality of response was the main determinant of the grade assigned. (Author/MH)