Page 90 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH
P. 90

Educational Measurement and Evaluation


                   Notes                (c)  This method cannot be used in speed tests, because the parts of the test are not
                                           independent.
                                      Merits
                                       (a)  The scores are obtained after administering the test only once under this method,
                                           which results in saving time and energy.

                                       (b) This method is quite wide.
                                  3.  Test-Retest Method : In the test-retest method, a test is administered to a definite group of
                                      students or individuals. After some time, that test is once again administered to that group.
                                      This is reliability coefficient. This is the simplest method. The higher the reliability coefficient
                                      (r) between two tests, the more reliable is the test. The assumption behind the test-retest
                                      method is that if the test is reliable, then the scores obtained by the group will be repeated
                                      the next time also. There may be some variation, but less the variation, the higher the
                                      reliability.
                                      This method too has certain limitations. The first demerit is that if the test is repeated soon
                                      after the first test, then students will continue to remember the answers of the first test and
                                      they will make an effort to write them in the second time. Thus, the scores in the retest will
                                      not be independent, rather they will be influenced by the first test. Consequently, the
                                      contact between the two scores will influence the reliability of the test. Its second demerit is
                                      that if the test is administered after a long interval of time, such as six months or more, then
                                      the second time scores will be influenced by the factor of time, because six months is a
                                      sufficient period to effect much mental development in students. This development has its
                                      influence on students’ performance, and in this case, their scores will be more than those of
                                      the first time. Thus, score coefficient (r) is decreased, and the reliability of the test takes a
                                      back seat. Its third demerit arises due to practice and experience. When students have solved
                                      such types of questions once, they have gained experience and practice, and they are benefited
                                      by it when they are administered the test the second time, which helps to increase their
                                      scores. It decreases the reliability of the test. It is clear that the method is certainly simple,
                                      but it is not more accurate.
                                  4.  Split-half Method : In the above two methods, test-rest method and parallel method, the
                                      test has to be administered twice, but it is possible that a test may not be administered twice
                                      or two formats of the test may not be available. In such a situation, split-half method is
                                      considered’ the most suitable. In this method, the test is not administered twice, rather : he
                                      test items are divided into two parts on the basis of the same test, and then separate scores
                                      of students are found out in order to ascertain their correlation coefficient.
                                      There can be several methods of diving a test into two parts; however, the commonest of
                                      them is the even-odd method. In this method, all the test are arranged in the ascending
                                      order of difficulty, and then even numbered items, such as 2, 4, 6,  8, etc. are segregated
                                      from the odd numbered items, such as 1, 3, 5, 7, etc. Thus, the test is divided into two parts.
                                      Then, the coefficient between two parts is calculated. It is not necessary that only odd and
                                      even numbered items be segregated. There is yet another method to do it which is called
                                      first half-second half method. In this method, the first half items are kept in the first test and
                                      the second half items are kept in the second test. But this method is not suitable, because if
                                      the questions have been arranged in the ascending or descending order of difficulty, then
                                      one part will have all the simple questions and another part all the difficult ones, and thus,
                                      the two parts will not be homogeneous, rather these two parts will become two different
                                      tests. Also, the amount of coefficient will be decreased.
                                      This problem is also faced when questions are divided on even and odd numbers, if the
                                      questions have been arranged in ascending order of difficulty; for example :





         84                                 LOVELY PROFESSIONAL UNIVERSITY
   85   86   87   88   89   90   91   92   93   94   95