that the significacance level or probability value p of the data was higher than the degree of significance
α = 0.05. Based on the criteria of homogeneity hypothesis testing the researcher determined, the criteria of
homogeneity hypothesis testing were mentioned that if the significance levelprobability value
≥ the degree of significance α = 0.05, the sample data came from population which had homogenous variance. Conversely,
the sample data came from population which did not have homogenous variance if the significance levelprobability value the degree of
significance α = 0.05. Therefore, it could be interpreted that the sample data came from population which had homogenous variance. The
complete statistical calculation was presented in the APPENDIX 8.
c. Test Analysis
After finishing the statistical analysis of prerequisite test including the normality and homogeneity test, the researcher continued to do
analytical testinghypothesis testing. Due to both the experimental and the control class’ pre and posttest which met the requirement of normality and
homogeneity test, the researcher was able to use independent sample t-test to test the hypothesis. However, the independent sample t test was not
only used to see the significant difference in both the experimental and the control class’ posttest after the treatment was given, but it was also done
to see if there was no extreme difference in both the experimental and the control class’ pretest before the treatment was given.
The independent sample t test was also performed in both the experimental and the control class’ pretest to show that both of the classes
had similar ability in narrative writing skill before the several writing instruction was given to the control class and the treatment was given to
the experimental one. The following is the result obtained from the
calculation of Independent-Sample T Test in both the experimental and the control class’ pretest :
Table 4.7 The Mean, the Standard Deviation, and Standard Error Mean of the Experimental and Control Class’ Pretest Score
Group Statistics
Class N
Mean Std. Deviation
Std. Error Mean Pretest
Score Experimental
26 40.481
13.4911 2.6458
Control 26
43.558 16.7507
3.2851
The table 4.7 above showed that the mean score of pretest in the experimental class was 40.48 whereas in the control class was 43.55. It
indicated that the mean score of the control clas s’ pretest score was higher
than the experimental one. It meant that there was no extreme difference in both the experimental and control class before the treatment was given.
Table 4.8 The Result of Independent Sample t-Test of the Experimental and Control Class’
Pretest
Independent Samples Test
Levenes Test for
Equality of Variances
t-test for Equality of Means
F Sig.
T Df
Sig. 2-
tailed Mean
Difference Std.
Error Difference
95 Confidence Interval of the
Difference Lower
Upper Pretest
Score Equal
variances assumed
1.847 .180 -.729
50 .469
-3.0769 4.2181 -11.5492
5.3953 Equal
variances not
assumed -.729
47.828 .469
-3.0769 4.2181
-11.5587 5.4049
In interpreting the result of Independent Sample T test presented in Table 4.8 above, it had to see the column Equal variances assumed first. It
was obtained that significance level or probability value of F 0.180 was higher than si
gnificance degree α = 0.05. This result meant that variance of data in both the experimental and the control class was homogenous.
Because of obtaining homogenous variance data, the researcher had to use t value t = -.729 and df = 50.
Based on the result of independent sample t-test for pretest above, the result was gained from the calculation that t
count
is -.729 whereas t
table 0.025
is 2.009 gained based on df = 50 and ⁄
α = 0.025
. Because of obtaining Sig. 2-tailed from the result of independent sample t-test, the
degree of significance α = 0.05 had to be divided by 2. It meant that t
count
-.729 was lower than t
table
2.009. Similarly, the value of Sig. 2- tailed was 0.469. It meant that the significance level 0.469 or
probability value p was higher than the degree of significance ⁄
α = 0.025. The result of Independent-Sample T Test on pretest p = 0.469.
To interpret the result of the independent sample t test of pretest explained above, the resercher formulated the criteria of hypothesis testing
which stated that there was no extreme difference in narrative writing skill between the experimental and the control class before the treatment was
given; if t
count
t
table
or Sig 2-tailed
⁄
α
.
Meanwhile, if Sig 2-tailed
⁄
α or t
count
t
table,
it could be assumed that there was difference in narrative writing skill between the experimental and the control class
before the treatment was given. Hence, the result of the independent sample t test of the
experimental and control class’ pretest could be concluded that there was no extreme difference in both the experimental and the
control class’ narrative writing skill before the treatment was given. In other words, both
the classes had equal ability in writing narrative text before the treatment was given.