42
b. Item Discrimination
Heaton 1975: 173 stated that index of discrimination shows the extent to which the item discriminates between the testees, separating the more able testees
from the less able ones. Item discrimination can be formulated as follows. D = correct U – correct L
n D = discrimination index
U = upper half L = lower half
n = number of students in one group
Discrimination index ranges from +1 to -1. If an item has index discrimination -1, the item discriminates the students in entirely the wrong way. In
contradiction, if an item has index discrimination +1, the item discriminates perfectly. It shows perfect correlation with the testees result on the whole test.
Item discrimination 0 shows that the item does not discriminate in any way at all.
3. Validity, Reliability, and Authenticity
a. Validity
Validity consists of some aspects, namely content validity, criterion related validity, construct validity, and face validity. However, in this study the
researcher did not include all those aspects. The aspects the researcher included in the study were content validity, construct validity and face validity. To know
whether the test has already met those three types of validity, researcher interviewed three experts. They were an English teacher of RSBI seventh grade
43 class of SMP Pangudi Luhur I Yogyakarta, a lecturer from English Education
Study Program of Sanata Dharma University Yogyakarta, and a test constructor from Language Teaching Institution Yogyakarta. Their educational backgrounds,
professions, and experiences made the researcher believe that they were so much competent to evaluate the test in this study.
b. Reliability
Measuring reliability is measuring consistency. A reliable test is consistent and dependable Brown, 2004: 20. In this study the researcher used a method
proposed by Harris 1969. Harris 1969: 15 stated that to estimate the reliability of a test can be done by giving a single administration of one form of the test and
then, by dividing the items into two halves usually by separating odd and even numbered items and by obtaining two scores for each individual. By such “split-
half” procedure commonly known as Spearman-Brown formula, the researcher could obtain the correlation coefficient. The coefficient result became the
reliability of one group test.
c. Authenticity
To see whether the test already fulfilled the aspect of authenticity, the researcher asked three experts to judge the tests. The experts who judged the
authenticity of the test were the experts who also judge the validity of the test. Moreover, the researcher also relied on the theoretical review to prove that the test
has already met the authenticity criteria.
44
F. Research Procedure
The researcher organized the steps she took in this study into four main steps as follows.
1. Need-analysis In the first main step, the researcher asked for a letter of permission from
Sanata Dharma University to conduct a need-analysis to the RSBI seventh grade students of SMP Pangudi Luhur I Yogyakarta. Having the letter of permission
from the university, the researcher asked for a permition from the head master of SMP Pangudi Luhur I Yogyakarta to conduct the need-analysis in the school.
Then, having the permition from the headmaster, the researcher conducted an interview with the English teacher of SMP Pangudi Luhur I Yogyakarta. Further,
in this step the researcher also gathered sources and references needed to design the test.
2. Test Design The second main step was designing the test. After the test was designed,
the validity and authenticity were judged by three experts. Then, some revisions were made from the feedbacks given by the experts.
3. Field-Testing After the test was completely revised, the test was field-tested to 27 RSBI
seventh grade students of SMP Pangudi Luhur I Yogyakarta. 4. Scoring and Test Analysis
The fourth major step in this study was scoring and test analysis. The scores of the test helped the researcher to see whether the test was in accordance