Page 307 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH
P. 307

Unit 26: Achievement Test: Concept, Types and Construction


                     Example                                                                         Notes
                     Suppose, the following data are obtained in respect of question number 27 of a test :
                          R  = 60,       W  = 30,     N = 100,  NR = 10,
                           H         H        H          H
                          R  = 20,        W  = 50,      N  = 100,   NR  = 30,  K = 4
                           L         L        L           L
                     So, the difficulty level of question number 27 will be :
                                          1 ⎡    30 ⎞   1    ⎛     50 ⎞  1   ⎤⎛
                                     DL =  ⎢  ⎜  60 –  ⎟    +  ⎜  20 –  ⎟    ⎥
                                          2 ⎝    4 – 1 100 – 10  ⎝  4 – 1 100 – 30 ⎦  ⎠
                                                    ⎠ ⎣
                                          1 ⎡       1  ⎛  10 ⎞  1 ⎤
                                        =  ⎢  (   )60–10  +  ⎜  ⎟  ⎥
                                          2 ⎣      90  ⎝  3 ⎠  70 ⎦
                                          150    1  ⎤  ⎡ 138 ⎤
                                            ⎡
                                        =   290  +  21 ⎥  ⎦   =   ⎢  ⎣ 263 ⎥  ⎦
                                            ⎢
                                            ⎣
                                          19
                                        =     = .30
                                          63
                     Difficulty level (in %) = 30%
                     Note :
                     Though no norm has been determined as regards difficulty level for the selection of
                     items in the construction of a test, yet most specialists consider the selection of items of
                     different level from 30% to 70% as suitable. Some of the specialists are of the opinion that
                     it should be limited from 40% to 60%. Katiz has written in this context that if only
                     proficiency of the students in a subject is to be found out, then the test should be constructed
                     using less difficult items; but if the students have to be divided into different classes as
                     high, medium and low then comparatively difficult items should be used to construct a
                     test. He has demonstrated it by the following table :


                              Form of Item           Difficulty level (in %)
                              Essay type                     50
                              Five choice type               60
                              Four choice type               62
                              Three choice type              66
                              Two choice type                75

                (b)  Discriminating Power of Items : Discriminating power of a question or item intimates
                     how far it can discriminate the students of higher and lower achievement. If a question
                     has not this quality, it is not selected for the test. If a question or item in a test is solved
                     by students of higher and lower levels in an equal or almost equal percentage, then its
                     discriminating power is zero and it is not considered worth for inclusion in the test.
                     However, if the percentage of the students of higher level is sufficiently more than the
                     students of lower level, the positive discriminating power of the question is high and
                     such a question is selected. The determination of high and low levels is done on the basis
                     of obtained scores of the test.
                     Finding out Discriminating Power of the Items : The following methods and formulae
                     are used to find out the discriminating power of the items :
                     First Method : When the group of students is of small size, the number of students
                     solving an item is known in percentage and the discriminating power is found out using
                     the following formula :




                                               LOVELY PROFESSIONAL UNIVERSITY                                    301
   302   303   304   305   306   307   308   309   310   311   312