Discriminating power Validity of the Test

33 numbers: 4, 12, 14, 12, 18, 20, 22, 23, 25, 35, 57, 61, 72, 77, 84, 85, 93, 98, 101 should be discarded unused. The result of index of difficulty can be seen in appendix 9.

3.4.6 Discriminating power

The writer in the analysis selected 27 of the upper group and 27 of the lower group of the students’ results in order to find the discriminating power. Then the number of correct answer made for each item coming from both the lower and the upper group results were counted. The discriminating power of an item can be calculated by applying the following procedure recommended by Haris, 1969:106, as follows: Step1. Separate the highest and the lowest 27 percents of the papers. Step2. For each item, subtract the number of lows who answered the item correctly, from the highs who answer correctly. Step3. Divide the result of step 2 by the number of papers in each group, highs and lows, to obtain the item discrimination index. The following is the formula of calculating the discriminating power: DP = FH + FL N Where DP = Index of discrimination power FH = The number of students in the upper group who answered an item correctly ; 34 FL = The numbers of students in the lower group who answered an item correctly ; N = The total number or of students in one group. For example, the computation of the formula for item test number 1 is as follows: FH = 8 FL = 5 DP = 8 – 3 = 0.30. 10 The result of computation the discriminating power of the try out test can be seen in appendix 8. After the try out test was arranged, the test items that did not fulfill the requirements or had low index of discrimination were considered to be invalid.

3.4.7 Validity of the Test

Validity is the most important quality of any test. A test is said to be valid when it actually measures what is intended to measure; Gronlund, 1976:79. Related to the validity, Harris also states that validity of the test is usually distinguished into three kinds. They are contents validity, empirical validity or statistical validity and face validity. In order to find out the validity of the test, the writer took content validity to measure the test items. Content validity is one that depends on a careful analysis of the language being tested and the particular course objectives. A test is said to have high content validity if each item used to gather data has relevance to established criteria or objectives and covers representative materials. To find out whether the 35 test has content validity, the writer compared the test items with the materials dealing with question tag. Related to this case, in constructing the instrument the writer had to look first at the references and find the materials that the students had learned. Empirical or statistical validity is the way a test shows how the test scores obtained by the testes are related to some outside criterion such as marks they get at the end of a course. Face validity is the way a test looks to examiners, teachers, and examinees or if a test looks right to other testers, teachers or testes.

3.4.8 Reliability of the Test