Reliability of the Test

32 c Face Validity Face validity covers the looks of the test. If it looks as if it measures what it is intended to measure then it has face validity Hughes, 1989. Since the test used in this study measures students’ mastery in translating relative clauses from Indonesian into English, it was constructed in form of sentences. The students were to rewrite while translate the test items from Indonesian into English. From the result of the pilot test, the writer found that the participants could understand clearly how to do the test and thus were able to do it. It means that the participants knew what to do when they got the test sheets. The looks of the test can be seen in the appendices See appendix F.

2. Reliability of the Test

Discussing reliability of a test, Brown 2004 suggests that it should be “consistent and dependable”. While validity concerns on accuracy, reliability is about consistency. A test is said to be reliable when it gives a more or less the same overcome when it is administered to the same subjects on different occasion. Ebel 1979 states that “there are at least five methods have been used for obtaining the independent measurements necessary for estimating test reliability.” The methods are reader reliability, test-retest, equivalent forms, split-halves, and Kuder-Richardson. The test that the writer conducted used the split-halves method. According to Ebel 1979, in split-halves method the test was split into two equivalent halves which then produce two independent scores used to PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI 33 estimate reliability. This method is simpler than the other methods because it requires only one administration of one test. A test is said to be reliable if it yields the same result when the test administered to the same students or matched students on two different occasions Brown, 2004. Thus, the split-halves method, although only administered once, gives the students two scores on the test. The two scores are derived from the two halves of the test. The first half scores are from odd numbered items, which then labelled X. The other half scores are from even numbered items, then labelled Y. The distribution of the test items can be seen in the blueprint of the test See Appendix E. In determining the reliability of the test used in this study, the writer used the Pearson product moment coefficient of correlation to estimate the correlation between the two-halves. The formula is shown below: where r = Pearson r ΣX = sum of scores in X-distribution ΣY = sum of scores in Y-distribution ΣXY = sum of the products of paired X- and Y-scores ΣX 2 = sum of the squared scores in X-distribution ΣY 2 = sum of the squared scores in Y-distribution N = number of paired X- and Y-scores subjects PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI 34 After estimating the correlation between the two-halves, the writer estimated the reliability of the split-half test using the Spearman-Brown split-half reliability formula. The following is the formula of the Spearman-Brown split-half reliability estimate: where: = the Spearman-Brown split-half reliability estimate = the correlation between the two halves To determine the reliability of the test, the writer conducted a pilot test. The preliminary test was conducted in November 19, 2009 in Structure V Class. The time allocation was 45 minutes from 9 o’clock until 9.45 in the morning. There were thirty students carried out the test, including the shoppers. Out of the thirty students there were eighteen students from the fifth semester. The other twelve were shoppers students other than the fifth semester. The writer used the test result of the eighteen students to estimate the reliability of the test. The result of the pilot test showed that reliability score for the test was 0.31, which means the test was reliable. The participants’ score in the pilot test and the computation of the test reliability can be seen in the appendices See Appendix H and Appendix I. PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI 35

3. The Scoring Procedure for the Test