The Analysis of Test Validity The Analysis of Reliability Coefficient The Analysis of Difficulty Index

Notes: FV : Facility Value The difficulty index R : The number of the current answers N : The number of the students taking the test To give the interpretation of the difficulty index, Heaton 1988:179 suggests that there are some criteria to determine the degree of the difficulty index. The criteria of the difficulty index as follows, Table 3.3 Table of Heaton’s Difficulty Index Difficulty Index 0,30 0,30 – 0,70 0.70 Interpretation Difficult Fair Easy

3.5.2 The Result of the Try Out Test

The try out test was administered to one of the Eleventh Grade classes which did not belong to the research respondents experimental and control groups on January 13 th , 2015. The class was XI IIS 1 which was also chosen by lottery. It consisted of 32 students. The score of the try out test in this research covering the analysis of the test validity, the analysis of reliability coefficient and the analysis of difficulty index.

a. The Analysis of Test Validity

The content validity of the narrative text reading comprehension test were established by constructing the test items based on the curriculum used in SMAN 1 Jombang which was curriculum of 2014 for the Eleventh Grade Students. Moreover, the test items were consulted to both of the research consultants as well as the English teacher of SMAN 1 Jombang.

b. The Analysis of Reliability Coefficient

In finding the reliability coefficient of the narrative reading comprehension test, the researcher analyzed the scores statistically by using Split-Half Odd Even Technique through SPSS application. The researcher signed the odd number with X and the even number with Y. the distribution of odd and even numbers are enclosed on Appendix K. The result of the reliability coefficient is presented in the following 4.4 table: Table 3.4 The Output of Split-Half Odd Even technique SPSS Reliability Statistics Cronbachs Alpha Part 1 Value .365 N of Items 10 a Part 2 Value .160 N of Items 10 b Total N of Items 20 Correlation Between Forms .696 Spearman-Brown Coefficient Equal Length .801 Unequal Length .801 Guttman Split-Half Coefficient .807 a. The items are: N1, N3, N5, N7, N9, N11, N13, N15, N17, N19. b. The items are: N2, N4, N6, N8, N10, N12, N14, N16, N18, N20. From the table presented above, the reliability coefficient of the half test was 0.696 than the reliability of the whole test was 0.801. Cohen et al 2007:506 states that if the reliability coefficient is in the range of 0.80 until 0.90, it is highly reliable. It indicated that the researcher did not have to make any changes of the test items.

c. The Analysis of Difficulty Index

The try out test was in the form of multiple choice test. The students were asked to choose the correct answer by circling one of the provided options. There were 20 items in the try out test. It contained two different narrative texts as well. Each right item got 5 points, so the total score was accumulated as 100. To know whether each test items were too difficult, too easy or fair, the difficulty index of the test items were analyzed by using Facility Value formula from Heaton 1988:178. The result of the analysis informed that the difficulty index of the test items was in the range of 0.59 until 0.68. Heaton 1988:178 says that a test item which has the range from 0.30 until 0.70 is categorized as fair. It means that the test items in the try out test were already good, since all the items belonged to fair items. The detail results of difficulty index were written on Appendix L. Further, the time allocation of the test was categorized appropriate. The reason is all the students were able to do the test given within the provided time, so the test items of the posttest were the same as those in which administered on the try out test.

3.6 Data Analysis Method