33
A4 53 70 B4 83 80
A5 50 63 B5 73 66
A6 60 66 B6 70 86
A7 43 60 B7 73 80
A8 50 60 B8 70 66
A9 40 53 B9 56 63
A10 53 56 B10 73 76
A11 73 66 B11 83 80
A12 80 93 B12 66 70
A13 80 86 B13 80 90
A14 40 53 B14 66 70
A15 60 66 B15 60 63
A16 53 66 B16 80 86
A17 56 63 B17 50 56
A18 63 80 B18 66 66
A19 83 80 B19 86 86
A20 76 76 B20 60 60
A21 73 80 B21 56 60
A22 60 50 B22 50 53
A23 40 60 B23 53 53
A24 66 56 B24 50 56
A25 56 66 B25 73 66
A26 76 80 B26 60 73
A27 60 63 B27 83 80
A28 50 60 B28 66 76
A29 46 63 B29 76 66
A30 50 53 B30 46 50
A31 66 63 B31 66 63
A32 60 70 B32 76 76
A33 60 73 B33 56 50
A34 56 70 B34 76 70
A35 76 80 B35 76 73
A36 46 63 B36 66 73
A37 76 86 B37 70 70
B38 66 70
B39 66 63
∑ 2239 2538 ∑ 2585 2661
Mean 60.51 68.59 Mean 66.28 68.23
Min 40 50 Min 46 50
Max 83 93 Max 86 90
Based on Table 4.1, it is presented that in the experimental class, the highest pre-test score is 83 while the lowest pre-test score is 40. Meanwhile, the students of control class got 86 as the highest pre-test score, while they got 46 as the lowest pre-test score. Moreover, in the experimental class, the average score is 60.51. On the other hand, in the control class, the average score is 66.28.
Table 4.1, also shown the post-test score of the experimental and control class. The highest post-test score in the experimental class is 93 while the lowest pre-test score is 50. Meanwhile, the highest post-test score of the control class was 90 while the lowest score is 50. In addition, the average score of the experimental class is 68.59. On the other hand, the average score of the control class is 68.23.
b. Gain Score
Gain score is a score that is gained by finding the difference between the post-test score and the pre-test score. The description of the gain score is presented in table 4.2 as follows:
Table 4. 2
Gained Score of Experimental and Control Class No Experimental Class Control Class
1 13 4
2 13 7
3 10 6
4 17 -3
5 13 -7
6 6 16
7 17 7
8 10 -4
9 13 7
10 3 3
11 -7 -3
12 13 4
13 6 10
14 13 4
15 6 3
16 13 6
17 7 6
18 17 0
19 -3 0
20 0 0
21 7 4
22 -10 3
23 20 0
24 -10 6
25 10 -7
26 4 13
27 3 -3
28 10 10
29 17 -10
30 3 4
31 -3 -3
32 10 0
33 13 -6
34 14 -6
35 4 -3
36 17 7
37 10 0
38 4
39 -3
∑ 299 92
Mean 8.08 2.30
On Table 4.2, it is shown that the highest gain score from the experimental class is 20 and the lowest score is -10. Furthermore, the highest gain score of the control class is 16 and the lowest score is -10. This calculation result indicates that some students, both in the experimental class and control class, gained better results in the post-test than in the pre- test. However, on the other hand, the score of some students is worst in their post-test than in their pre-test. In addition, the average gain score for the experimental class is 8.08, while the average score of the control class
is 2.30.
2. Data Analysis
The data in this research was analyzed using t-test to confirm the statistical hypothesis. However before it was tested using t-test, the data must be analyzed using the normality test and homogeneity test first. Normality test is a test to check whether the data is normally distributed or not, while the homogeneity test is conducted to check the data is in the same variant or not. After the data was confirmed that is normally distributed and in the same variant, the t-test can be conducted. The test in this research was conducted by using manual calculation.
a. Normality Test
Normality test is a requirement before conducting a parametric test. If the data was not normally distributed, parametric test cannot be conducted, in other word, the researcher need to conduct non-parametric test.
In this research, the data was analyzed using Kolmogorov-Smirnov and Shapiro-Wilk. The data is normally distributed when the score was higher than 0.05. The result of normality using IBM SPSS 24 in this result is presented in a table 4.3 as follows :
Table 4. 3 Tests of Normality
Class
Kolmogorov-Smirnova Shapiro-Wilk Statisti
c df Sig.
Statisti
c df Sig.
Read ing
Pretest Experiment
.113 37 .200* .957 37 .158
Com prehe nsion
Posttest Experiment
.159 37 .019 .950 37 .095
Pretest Control .133 39 .077 .948 39 .069 Posttest Control .097 39 .200* .972 39 .436
*. This is a lower bound of the true significance.
a. Lilliefors Significance Correction
Table 4.3 shows that Kolmogorov-Smirnov score of pretest in experimental class is 0.200 and in control class is 0.077. Both of significance score shows that the score is higher than 0.05. It can be concluded that the pretest score of experimental and control class are normally distributed. Furthermore, the table above shows the significance score of post-test as well. Kolmogorov-Smirnov score of experimental class is 0.019 and control class is 0.200.It proved that the data was normally distributed as the significance score of post-test is higher than 0.05.
b. Homogeneity Test.
After the normality test was conducted, it is necessary to prove that the data is in same variant or homogeneous, by conducting Homogeneity test.
Homogeneity test was employed to pre-test score and post-test score. The result of homogeneity test of post-test score is presented in table 4.4 as follows.
Table 4. 4
Homogeneity Test of The Pre – Test Levene
Statistic df1 df2 Sig.
Pre-test Based on Mean .514 1 73 .476
Based on Median .437 1 73 .511
Based on Median and with adjusted df
.437 1 71.895 .511
Based on trimmed mean
.507 1 73 .479
Table 4.4 shows that the degree of significance based on the mean in the pre-test is 0.476, that is higher than 0.05. Therefore it can be concluded that both groups in the pre-test are homogenous or in the same variant.
In other hand, the analysis of the homogeneity variances of both groups in the post-test was conducted as well. The result of the homogeneity test of the post test can be seen at table 4.5 as follows
Table 4. 5
Homogeneity Test of The Post - Test
From the result of homogeneity test in the Table 4.7, it can be seen that the degree of significance based on mean of post test score is 0.558 which is higher than 0.05. From the table above, it can be proved that that groups,
Levene
Statistic df1 df2 Sig.
Post-test Based on Mean .345 1 75 .558
Based on Median .128 1 75 .721
Based on Median and with adjusted df
.128 1 74.628 .721
Based on trimmed mean
.320 1 75 .573
experimental and control class, in the post-test are homogenous.
c. Hypothesis Test
The normality test and homogeneity test that was explained before proved that the data in this research are qualified to be analyzed by using parametric test. The writer used t test formula to find the empirical evidence and to prove the hypothesis of this research. T-test formula was employed to find the existence of significance effect using Socratic Questioning on students‘ reading comprehension in analytical exposition text.
The hypothesis test was conducted by using manual calculation. Here is the manual calculation of the t-test.
Table 4. 6
Comparison of Students’ Gain Score between Students in Experimental Class and Students in Control Class
No X Y x = X - x y = Y - x x2 y2
1 13 4 9.32 7.02 86.8624 49.2804
2 13 7 9.32 7.29 86.8624 53.1441
3 10 6 6.32 4.02 39.9424 16.1604
4 17 -3 13.32 11.29 177.4224 127.4641
5 13 -7 9.32 7.02 86.8624 49.2804
6 6 16 2.32 0.29 5.3824 0.0841
7 17 7 13.32 11.02 177.4224 121.4404
8 10 -4 6.32 4.29 39.9424 18.4041
9 13 7 9.32 7.02 86.8624 49.2804
10 3 3 -0.68 -2.71 0.4624 7.3441
11 -7 -3 -10.68 -12.98 114.0624 168.4804
12 13 4 9.32 7.29 86.8624 53.1441
13 6 10 2.32 0.02 5.3824 0.0004
14 13 4 9.32 7.29 86.8624 53.1441
15 6 3 2.32 0.02 5.3824 0.0004
16 13 6 9.32 7.29 86.8624 53.1441
17 7 6 3.32 1.02 11.0224 1.0404
18 17 0 13.32 11.29 177.4224 127.4641
19 -3 0 -6.68 -8.98 44.6224 80.6404
20 0 0 -3.68 -5.71 13.5424 32.6041
21 7 4 3.32 1.02 11.0224 1.0404
22 -10 3 -13.68 -15.71 187.1424 246.8041
23 20 0 16.32 14.02 266.3424 196.5604
24 -10 6 -13.68 -15.71 187.1424 246.8041
25 10 -7 6.32 4.02 39.9424 16.1604
26 4 13 0.32 -1.71 0.1024 2.9241
27 3 -3 -0.68 -2.98 0.4624 8.8804
28 10 10 6.32 4.29 39.9424 18.4041
29 17 -10 13.32 11.02 177.4224 121.4404
30 3 4 -0.68 -2.71 0.4624 7.3441
31 -3 -3 -6.68 -8.98 44.6224 80.6404
32 10 0 6.32 4.29 39.9424 18.4041
33 13 -6 9.32 7.02 86.8624 49.2804
34 14 -6 10.32 8.29 106.5024 68.7241
35 4 -3 0.32 -1.98 0.1024 3.9204
36 17 7 13.32 11.29 177.4224 127.4641
37 10 0 6.32 4.02 39.9424 16.1604
38 4 -5.71 32.6041
39 -3 -2.3 5.29
∑ 299 92 159.16 74.59 2825.429 2330.396
The statistical calculation to conduct t test was done in the followingsteps:
1) Determine the Mean of Variable X with formula:
Mx = ∑
=
= 8.08
2) Determine the Mean of Variable X with formula MY = ∑
=
= 2.30
3) Determining the Standard Deviation Score of Variable X with formula : SDX =√∑
= √
= 8.74
4) Determine the Standard Deviation Score of Variable Y with formula : SDy =√∑
= √
5) Determine Standard Error of Mean of Variable X with formula : SEMX =
√
=
√
6) Determine Standard Error of Mean of Variable Y with formula : SEMY =
√
=
√
7) Determine Standard Error of Difference with formula : SEMX - SEMY √ –
= √ –
= √
8) Determine t0 with formula : t0 =
= 2.99
9) Determine degree of freedom (df) with formula : df = N1 + N2 – 2
= 37 + 39 – 2
= 74
It can be seen, from the result of manual calculation above, that the value of t0 is 2.99 and the degree of freedom (df) is 74. The value of t-table in the degree of freedom of 74 and at the degree of significance 5% is 1.99. It can be conclude that the ttest is higher than ttable.
The theoretical hypothesis in this research are :
a) Null Hypothesis (H0) = There is no significant different result of using Socratic Questioning and without using it on students‘ reading comprehension in analytical exposition in the eleventh grade of SMA PGRI 56 Ciputat.
b) Alternative Hypothesis (H1) = There is significant different result between using Socratic Questioning and without using it on atudents‘
reading comprehension in analytical exposition in the eleventh grade of SMA PGRI 56 Ciputat.
The criteria to prove the hypothesis from the obtained data of experimental class and control class are:
a) If ttest ≤ ttable, in significant degree of 5%, the Null Hypothesis (H0) is accepted and the Hypothesis Alternative (Ha) is rejected. It can be inferred that there is no significant effect of Socratic Questioning on students‘ reading comprehension in analytical exposition text in the eleventh grade of SMA PGRI 56 Ciputat.
b) If ttest ≥ ttable, in significant degree of 5%, the Null Hypothesis (H0) is rejected and the Hypothesis Alternative (Ha) is accepted. It can be inferred that there is significant effect of Socratic Questioning on students‘ reading comprehension in analytical exposition text in the eleventh grade of SMA PGRI 56 Ciputat.
The value of ttable in degree of significance 5% (0.05) and tdegree of freedom 74 is 1.99. By comparing the result of ttable and ttest, in the degree of significance of 5%, the hypothesis can be interpreted in the following formula:
Ha = t0 ≥ ttable, = 2.99 ≥ 1.99
According to those results, a conclusion the research can be drawn that the H0 was rejected meanwhile the Ha wasaccepted. It means that there is significant effect of Socratic Questioning on students‘ reading comprehension in analytical exposition text in the eleventh grade of SMA PGRI 56 Ciputat.
10) . Effect Size
To identify the strength of the effect, the writer analyzed the data by using Cohen‘s formulation. The steps of calculation are :
d =
Determining pooled standard deviation with formula :
SDpooled = √ ( )
√
√ √
Determining the effect size with the formula :
d =
To interpret the result of effect size, the Cohen‘s d effect size criteria was used to figure out the level of the effect strengthens. The criteria can be seen as follows :
0 – 0.20 = weak
0.20 – 0.50 = modest effect 0.50 – 1.00 = moderate effect
>1.00 = strong effect
The result of the effect size is 0.05. Referring to the Cohen‘s effect size, 0.05 is on the interval 0 – 0.20. It can be inferred that the effect size of Socratic Questioning on Students‘ reading comprehension in analytical exposition text at eleventh grade of SMA PGRI 56 Ciputat has weak effect.