between 54 – 59, 13 student had score between 60 – 65 and 1 student had score
between 66 – 70 and 1 student had score 71
It also shows that the mean score of post-test in controlled class is 64.80 with 72 as the highest score and 54 as the lowest score. Then, the frequency
distribution table of post-test result also shows that there are 1 student had score between 50
– 55, 4 students had score between 55 – 60, 7 students had score between 61
– 65, 5 students had score between 66 – 70 and 3 students had score 71 - 75
In addition, the calculation also shows that the total of gained score of pre-test and post-test in controlled class is 75 and the mean of gained score of pre-test and
post-test in controlled class is 3.75 , with the highest student’s gained score is 11
and the lowest student’s gained score is -8.
B. Data Analysis
1. Normality Test of the Data
The result of normality test on both experimental and controlled class’ pre- test and post-test was gained from Lilliefors test using SPSS 22. This test is used
to determine whether the distribution of the data from the sample is normal. If the normality is more than the level of significance
α 0.05, scores will be normally distributed.
The following table was the results of normality test of the experimental and controlled class’ pre-test:
Table 4.3 Normality Pre-test Results between Experimental and Controlled Class
Tests of Normality
teknik dictogloss Kolmogorov-Smirnov
a
Shapiro-Wilk Statistic
Df Sig.
Statistic df
Sig. Nilaites
Experimental ,161
20 ,184
,960 20
,535 Controlled
,125 20
,200 ,964
20 ,631
. This is a lower bound of the true significance. a. Lilliefors Significance Correction
The result of the normality test above shows that the significance of the experimental class is 0.184 and 0.535. The controlled class is 0.200 and 0.631, it
means that the significance of both experimental and controlled class was higher than the degree of significance 5
α = 0.05. Therefore, it could be concluded that the data of both experimental and the controlled class’ pre-test is
normally distributed. The following table was the results of normality test of the experimental
and controlled class’ post-test:
Table 4.4 Normality Post-test Results between Experimental and Controlled Class
Tests of Normality
teknik dictogloss Kolmogorov-Smirnov
a
Shapiro-Wilk Statistic
Df Sig.
Statistic df
Sig. Nilaites
Experimental ,109
20 ,200
,973 20
,824 Controlled
,147 20
,200 ,949
20 ,354
. This is a lower bound of the true significance. a. Lilliefors Significance Correction
The result of the normality test above showed that the significance of the experimental class were 0.200 and 0.824. As for the control class were 0.200 and
0.354. It means that the significance of both experimental and controlled class was higher than the degree of significance 5
α = 0.05. Therefore, it could be concluded that the data of both experimental and the controlled class’ post-test is
normally distributed.
2. Homogeneity Test of the Data
After doing the normality test, the researcher continued to do homogeneity test. Homogeneity test is used to know the similarity of the two condition or
population. In other word, homogeneity test is used to know whether both groups, experimental class and controlled class, are homogeneous or not. To do
homogeneity test, the researcher used Levene Statistic Test from SPSS 22. The steps are:
The following table contained the result of homogeneity from pre-test score between experimental and controlled class and followed by the result of
homogeneity from post-test score between experimental and controlled class.
Table 4.5 Homogeneity of Pre-test Results between Experimental and Controlled Class
Test of Homogeneity of Variances
nilaites Levene Statistic
df1 df2
Sig. ,307
1 38
,583
Table 4.6 Homogeneity of Post-test Results between Experimental and Controlled
Class
Test of Homogeneity of Variances
nilaites Levene Statistic
df1 df2
Sig. ,607
1 38
,441
From the result of the Levene Statistic Test above, it could be seen that the significance of the data from experimental and controlled class pre-test score is
0.583 and the post-test is 0.441. It means the significance of the data is higher than the significance degree
α = 0.05. The result of homogeneity tests shows that pre-test and post-test between experimental and controlled class had homogeneity
distribution and can be tested using t-test.
3. Hypothesis Test t-test
In this part, the researcher calculated the data to test the hypothesis that whether there is significant
difference between the students’ writing of descriptive text by
using dictogloss technique in experimental class and the students’ writing of descriptive text without using dictogloss technique in control class. The
researcher calculated the data using t-test formula. Two classes were compared,
the experimental class was X variable and the control class was Y variable.