Page 231 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH
P. 231

Unit 19 : Grading - Need, Problems, Components and Methods



            This method involves simple procedure, and the meaning of each grade is understood. Criterion  Notes
            being the focal point; the student discover their strengths and weaknesses and each student can
            strive for attaining the highest grade. One limitation of this method is that scores are taken as
            such without considering the errors of measurement that creep in due to various subjective
            factors. Another limitation is that the cut-off scores for various categories are decided arbitrarily.
            Relative Grading
            This type of grading is called grading on the curve, which refers to normal distribution curve. In
            this case it is pre-determined what percentage of students are expected to receive the number of
            As, Bs, Cs, Ds and so on with reference to a particular group. That is why relative grading is also
            called ‘norm-referenced grading’. Steps involved in grading are as follows.
            (a)  Pupils are expected to acquire new competencies, which become the basis for testing of
                sampled behaviour outcomes.

            (b)  Tests developed provide evidences on students’ learning or achievement, which lead to
                award of marks.
            (c)  These marks when plotted on a graph leading to formation of distribution curve (obtained),
                which may be positively or negatively skewed. This obtained curve is converted into normal
                curve on the premise that there is a difference between the true level of achievement and
                the perceived level of achievement as yielded by tests. True level of achievement is expected
                to be normally distributed in a large population irrespective of curricular area.
            (d)  When perceived curve is transformed into a normal curve, the obtained scores can be
                classified into desired number of categories of grades. This is done by using mean of 5 and
                standard deviation of 2, and grade value will range from 1 to 9 categories. Two-tail categories
                at either end of distribution are combined to make 9 (Stanine scale) in place of 11 categories.
                In this way we will have grade-wise distribution of scores (Table 2).
  ∝
  ∝ σ σ σ σ σ σ σ σ σ σ σ σ σ σ σ
                            Table 2 : Grade-wise distribution of cases (9-point scale)


               Q.No.    Letter grade   Interval            No. of cases (%)  Grade  value

                 1           A         1.75  σ  to               4                9
                 2           B         1.25    to 1.75           7                8
                 3           C         0.75    to 1.25           12               7
                 4           D         0.25    to 0.75           17               6
                 5           E         – 0.25    to 0.25         20               5

                 6           F         – 0.75    to – 0.25       17               4
                 7           G         – 1.25    to – 0.75       12               3
                 8           H         – 1.75    to – 1.25       7                2
                 9           I         –   to – 1.75             4                1


            If 7-point grading is to be used, measurement scale may be divided into 7 equal parts. When
            5-point grading is to be used, measurement scale may be divided into 5 equal parts and grade
            values will range from 1 to 5 with a mean of 3 and S.D. of approximately 1 unit. Grade-wise
            distribution of cases will be as shown in Table 3.






                                               LOVELY PROFESSIONAL UNIVERSITY                                    225
   226   227   228   229   230   231   232   233   234   235   236