Frequency Percent
Valid Percent Cumulative
Percent 72
80 2
1 5.3
2.6 5.3
2.6 94.7
97.4 88
1 2.6
2.6 100.0
Total 38
100.0 100.0
Beside the table of the frequency distribution, it also can be described by a diagram which is presented as follows:
Picture 4.4 Diagram of Posttest Result of Controlled Class
In addition, in analyzing the data, the experimental class and the controlled class were compared to one another. It was done in order to prove statistically
whether there is any significant difference between the two classes. There were two things needed to analyze before calculating statistical hypothesis; test of
normality and homogeneity. After getting the data which are the results of the students
’ reading comprehension of both classes, then the data were analyzed by using statistic calculation of the test formula.
3. Normality Test
The normality test was performed by using Kolmogorov-Smirnnov and Shapiro-Wilk. The test was done to determine if the distribution of the data from
the sample is normal or not. Thus, SPSS version 22 was used. If the normality is more than the level of significance
α 0.05, the test is normally distributed. The
1 2
3 4
5 6
7
40 48
52 56
60 64
68 72
80 88
result was gained as follows: Analyze → Descriptive Statistics → Explore. Insert PretestPosttest in Dependent List and Class in Factor List. Click Plots and
Checklist Normality plots with tests → Continue → OK.
The results of normality test of the data are presented as follows:
Table 4.11 Normality of Pretest Result between Experimental and Controlled Class
Tests of Normality
Class Kolmogorov-Smirnov
a
Shapiro-Wilk Statistic
df Sig.
Statistic Df
Sig. Pretest
Experimental .131
38 .099
.968 38
.346 Controlled
.123 38
.154 .953
38 .109
a. Lilliefors Significance Correction
From the Table 4.11, it can be seen on Kolmogorov-Smirnov that the significance of the pretest in experimental class is 0.099. It can be concluded that
the data are normally distributed because of 0.099 0.05. Meanwhile, the significance of the pretest in controlled class is 0.154. Therefore, the data are also
normally distributed because 0.154 0.05. In other words, the data of the pretest
in both classes are normally distributed.
Table 4.12 Normality of Posttest Result between Experimental and Controlled Class
Tests of Normality
Class Kolmogorov-Smirnov
a
Shapiro-Wilk Statistic
Df Sig.
Statistic df
Sig. Posttest
Experimental .130
38 .106
.960 38
.184 Controlled
.116 38
.200 .963
38 .231
. This is a lower bound of the true significance. a. Lilliefors Significance Correction
From the Table 4.12, it can be seen on Kolmogorov-Smirnov that the significance of the posttest in experimental class is 0.106. It can be concluded that
the data are normally distributed because of 0.106 0.05. Meanwhile, the significance of the posttest in controlled class is 0.200. Therefore, the data are also
normally distributed because of 0.200 0.05. In other words, the data of the pretest in both classes are also normally distributed.
4. Homogeneity Test
The homogeneity test was used to test whether the data from both classes have the same variant so that the hypotheses can be tested by t-test. SPSS version
22 was also used to do homogeneity test. Homogeneity test was calculated by using Levine. The result was gained as follows: Analyze
→ Compare means → One Way Anova
→ Put PretestPosttest in Dependent list and Class in Factor List → Click option and Checklist Homogeneity of variance test → Continue → OK.
The following tables contained the results of homogeneity test between both classes:
Table 4.13 Homogeneity of Pretest Result between Experimental and Controlled Class
Test of Homogeneity of Variances
Pretest Levene Statistic
df1 df2
Sig. 2.508
1 74
.118
Table 4.13 shows that the significance of pretest results between experimental and controlled class is 0.118. Therefore, it can be concluded that
both classes have the same variant because of 0.118 0.05.
Table 4.14 Homogeneity of Posttest Result between Experimental and Controlled Class
Test of Homogeneity of Variances
Posttest Levene Statistic
df1 df2
Sig. .829
1 74
.365
Table 4.14 shows that the significance of posttest results between experimental and controlled class is 0.365. Therefore, it can be concluded that
both classes have the same variant because of 0.365 0.05. In other words, the data of pretest and posttest in both classes are homogen.
5. Hypothesis Testing
The last calculation was testing the hypothesis. This was the crucial calculation to answer the problem formulation of the research that whether
Directed Reading-Thinking Activity DR-TA strategy gives influence on students’ reading comprehension of narrative text or not. Comparative technique
was used to test the hypothesis in which the result tests pretest and posttest of both classes experimental and controlled class were compared as follows:
Table 4.15 The Comparison Score between Students in Experimental
Class X and Students in Controlled Class Y No
X Y
x = X-Mx y = Y – My
x
2
y
2
1 20
12 3
3.79 9.00
14.36 2
28 12
11 3.79
121.00 14.36
3 28
-4 11
-12.21 121.00
149.08 4
20 12
3 3.79
9.00 14.36
5 28
-8 11
-16.21 121.00
262.76 6
24 4
7 -4.21
49.00 17.72
7 16
12 -1
3.79 1.00
14.36 8
2 12
-15 3.79
225.00 14.36
9 8
8 -9
-0.21 81.00
0.04 10
16 4
-1 -4.21
1.00 17.72
11 8
8 -9
-0.21 81.00
0.04 12
8 12
-9 3.79
81.00 14.36
13 16
8 -1
-0.21 1.00
0.04 14
12 12
-5 3.79
25.00 14.36
15 12
4 -5
-4.21 25.00
17.72 16
8 -4
-9 -12.21
81.00 149.08
17 20
8 3
-0.21 9.00
0.04