status of salaries and level of student preparations, so they are dissatisfied with current methods of evaluation. In 1960 Robert F. Oliver discussed the Eternal and (Infernal) Problem of Grades,1 and many teachers in 1978 still con sider grading to be a problem. A survey of Volume 27 (1977) of the Education Index showed fifty articles ex amining the problem of assigning grades to students. In working toward a solution to the grading problem, the author has looked at past successes and mistakes, com paring how others assign grades with her own procedures. Traditional efforts to solve the eternal and infernal prob lem have clustered around two approaches to grading, norm-referenced and criterion-referenced evaluation. Working definitions of those two approaches are in order. Norm-referenced measurement evaluates a student's progress among classmates. When teachers approach the problem of grading armed with the bell-shaped curve and the various statistical methodologies designed to predict means and medians, deviations and, ultimately, grades, they are using norm-referenced strategies. Problems arise for accurate statistical analysis when class sizes are small. The author also observed an interest ing phenomenon in regard to competition. Students tended to compare rates and amounts of work and to drift toward a lower common denominator. Old habits die hard, so there would usually be one or two rate break ers who would encounter grumbles and groans from class mates when they pulled the top edge of the curve higher than those on the fringes had hoped. A similar phenomenon has been observed with pass/ fail systems: students tended to do the minimum amount of work required to avoid failure. Students also appeared
Read full abstract