validity test of post-test, there were 26 valid questions from 30 questions which were gained. Based on the standard of reliability test, 0.4-0.59 means the
reliability of the test was medium, meanwhile 0.6-0.79 means the reliability of the test was strong, and 0.8-1.0 means the reliability of the test was very strong. In
this research, the score of reliability test of pre-test was 0.8 which means that the reliability of pre-test was very strong, meanwhile the score of reliability test of
post-test was 0.66 which means the reliability of post-test was strong.
E. Data Collection Technique
Technique of collecting data in this research was pretest and posttest that was given to both experimental class and control class. In this research, the
researcher gave pre- test to see students’ basic capability in reading before using
Question Generation Strategy in reading class. The next test was post-test which measured
the difference of the students’ score between the experimental class and control class after given the treatment.
F. Data Analysis Technique
The researcher used statistical calculation to find out the differences between students’ score in experimental class 8B which was taught using
Question Generation Strategy with students’ score in control class 8A which
was taught using silent reading method. In this research, there are four tests which were calculated; normality test and homogeneity test, hypothesis test, and effect
size test.
1. Normality Test
Normality test was done in both classes; the experiment class and control class. Normality Test was used to know whether the data from the sample group
which was examined comes from the population of normal distribution or not. The normality test was calculated using Kolmogorov-Smirnov technique with
requirement as follows: - If the significance scores of Asyim Sig 2 tailed 0.05, the data came
from the normal population
- If Asyim Sig 2 tailed 0.05, the data did not come from normal population.
2. Homogeneity Test
After normality test gave indication that data was distributed normally, so it needed to do homogeneity test. Homogeneity Test was used to know the
similarity of the two conditions or population. The researcher calculated the homogeneity test using Levene Technique with One Way ANOVA with
requirement as follows: - If the significance scores of Asyim Sig 2 tailed 0.05 meant that the
students’ English level in experimental class and control class were at the same
- If Asyim Sig 2 tailed 0.05 meant that the students’ English level in
experimental class and control class were not at the same level.
3. Hypothesis Test
For The Hypothesis test, the researcher used t-test to find out whether there is significant difference between experimental class which was taught using
Question Generation Strategy and control class which was not. The researcher did the t-test from post-test score of both classes and from the gained score of both
classes. The researcher used SPSS v.20 to find the result in a form of significance of t-test. After gaining the significance of t-test, the researcher compared the
significance of t-test with the significance degree which was 0.05. Testing hypothesis used criteria with significance degree 0.05. The conclusion was gained
as follows: If the significance of t-test 0.05, H
is rejected or H
1
is accepted If the significance of t-test 0.05, H
is accepted H
: There is no effect of Question Generation Strategy on the students’
achievement in reading comprehension of narrative text. H
1
: There is an effect of Question Gener ation Strategy on the students’
achievement in reading comprehension of narrative text.