THE EFFECTIVENESS... Nur Laila Molla Mater’s Program in Linguistics, Diponegoro University
©2010, UNDIP Institutional Repository
researchers were aware of the importance of collecting data on what actually went on in the classroom, and built into the study systematic classroom
observation. In this research, there are two groups, treatment experimental group and control group. First of all, pre-test was applied for both. During
research, the researcher implemented GTM Plus at experimental group and RSBI GTM implemented at control group. In the end of research post-tests
were applied for both groups and then compared the result.
B. Data Analysis
The researcher analyzes data quantitatively. Instrument of analysis as follows:
1. Validity of The Test
Validity of a test can be known from the result of mind process and experience. The first thing to get is logical validity and the second is empirical
validity. Those two things become a basic of test validity grouping Arikunto, 2009:65. The writer uses Pearson Product Moment formula to measure
validity see appendix 14.
2. Reliability of The Test
Reliability is related with trust. A test can be treated as high reliability if the test can give constant result. It means that test reliability is related with
consistency of test result. Or, if the result changes, the change itself can be said unworthy. A test is reliable to the extent that is measured consistently,
from one time to another in different occasions. The writer measured the reliability of test using split-half method and then continued by Spearman-
Brown formula to measure the whole test reliability. Committing this research,
THE EFFECTIVENESS... Nur Laila Molla Mater’s Program in Linguistics, Diponegoro University
©2010, UNDIP Institutional Repository
the writer split the item into two parts: odd and even numbers see appendix 15.
3. Difficulty Level
Every item of the test was tested. The writer measured the difficulty level of every item. According to formula, there are several classifications of
difficulty. Arikunto 2009:210 explains that according to general principle, difficulty index is often classified as follows:
-Test number with P from 0.00 to 0.30 is difficult, 0.30 to 0.70 is moderatefair, and 0.70 to 1.00 is easy see appendix 16.
4. Discriminating Power
Discriminating power measures how the test item is arranged. This means to identify the possible differences for students’ competence. Typical
classification of discriminating power is stated as follows: - D: 0.00
– 0.20: poor, 0.20 – 0.40 : satisfactory, 0.40
– 0.70: good, 0.70 – 1.00: excellent, D
: negative, it is not good, the test item should be eliminated Arikunto, 2009:213-218.
See appendix 17.
Preliminary Analysis
Before the hypotheses are tested, there are two items tested, normality and homogeneity. Subana and Sudrajat 2009:159 says that things to figure out
related with inferential statistic to test treatment to compare are the condition of data normality and variant homogeneity.
1. Normality Test