While the homogeneity of posttest of both groups is computed as follow:
F = 2.0072 The result was consulted with the value of F table with dk numerator v
1
= 27-1 = 26, dk numerator v
2
= 27-1 = 26 , and α = 0.05, squared to v
1
= 27and v
2
= 27 is 2.194. Because the F 1.74 2.194, then the data were considered to be
homogeny. The computation of homogeneity can be seen in Appendices 15 and 16.
4.3 Statistical Analysis
All of the test score have to be counted statistically. The first way to know the significant difference of the experimental and control group can be seen
through the differences of the two means. The following formula was used to get the means:
N Xe
Me
Depdikbud, 1994:29
Where: Me
= the mean score of the experimental group
Xe
= the sum of all scores of the experimental group
N
= the number of the subject sample
N Xc
Mc
Depdikbud, 1994:29
Mc
= the scores of the control group
Xc
= the sum of all scores of the control group
N
= the number of the subject sample The mean scores of the pre-test of the experimental and control group
was calculated as follows:
N Xe
Me
85 .
68 27
1857 Me
Therefore, the mean score of the experimental was 68.85
N Xc
Mc
85 .
62 27
1697 Mc
Therefore, the mean score of the control group was 62.85 The mean scores of the post-test of the experimental and control group
was calculated as follows:
N Xe
Me
67 .
84 27
2286 Me
Therefore, the mean score of the experimental was 84.67
N Xc
Mc
. 74
27 1998
Mc
Therefore, the mean score of the control group was 74.0 If I compared the two means, it was clear that the mean of the
experimental group was 84.67 and the control group was 74.0, so the difference between the two means was 8.39. To make the analysis more reliable, I analyzed
by using
t
-test formula. Using t-test formula could see the difference between the two means.
I analyzed the significant difference between teaching analytical exposition text with using Indonesia Now Video as a media and teaching
analytical exposition text without pictures video by using
t
-test. The
t
-test formula is as follow:
y x
y x
y x
N N
N N
y x
M M
t 1
1 2
2 2
Where:
t
=
t
-test
x
M
= the mean of the experimental group
y
M
= the mean of the control group
2
x
= Total number of individual scores deviation of experimental
2
y
= Total number of individual scores deviation of control group
x
N
= the number of subject of the experimental group
y
N = the number of subject of the control group
adopted from Arikunto, 1998: 306 The
t
-test of the pre-test of the experimental and control group was calculated as follows:
y x
y x
y x
N N
N N
y x
M M
t 1
1 2
2 2
Where:
85 .
68 27
1859 Nx
x Mx
4 .
65828 127996
193824 27
1859 193824
2 2
2 2
2 2
2
x x
x Nx
x x
x 85
. 62
27 1967
Ny Y
My
4 .
41746 10660
148406 27
1697 148406
2 2
2 2
2 2
2
y y
y Nx
Y Y
y
The analysis of the
t
-test was as follow:
y x
y x
y x
N N
N N
y x
M M
t 1
1 2
2 2
27 1
27 1
2 27
27 41
. 41746
4 .
65828 85
. 62
58 .
68 t
27 2
52 107575
00 .
6 t
24 .
153 00
. 6
t
4 .
12 00
. 6
t
48469 .
t
On =5 with fa = 27 + 27 – 2 = 52, the result
01 .
2
52 95
.
t
-2.01 0.485
2.01
Because
t
-estimation in the Ho area, so we can conclude that there is not significant difference between experimental and control group.
The
t
-test of the post-test of the experimental and control group was calculated as follows:
y x
y x
y x
N N
N N
y x
M M
t 1
1 2
2 2
Where:
667 .
84 27
2286 Nx
x Mx
276 193548
193824 27
2286 193824
2 2
2 2
2 2
2
x x
x Nx
x x
x 74
27 1998
Ny Y
My
554 147852
148406 27
1998 148406
2 2
2 2
2 2
2
y y
y Nx
Y Y
y Acceptance
area Ho
The analysis of the
t
-test was as follow:
y x
y x
y x
N N
N N
y x
M M
t 1
1 2
2 2
27 1
27 1
2 27
27 00
. 554
00 .
276 .
74 67
. 84
t
27 2
52 .
830 67
. 10
t
18234 .
1 67
. 10
t
09 .
1 67
. 10
t 81
. 9
t
On =5 with fa = 27 + 27 – 2 = 52, the result
01 .
2
52 95
.
t
-2.01 2.01
9.81
Because
t
-estimation is not in the Ho area, so we can conclude that there is significant difference between experimental and control group.
After getting
t
-value, I consulted the value of
t
-table to check whether the difference was significant or not. Before the experiment had been conducted, the
Acceptance area Ho
level significant to be used in the experiment had been divided first. I used the 5 alpha level significance as usually used in psychological and educational research.
The number of subjects in this experiment for experimental and control group were 27 with the degree of freedom fa 52, that was
2
2 1
n n
27+27- 2=52. Because it was a social research so the researcher used 5 alpha level
significance, the value of
t
-table was 2.01 while the value of
t
-test was 9.81
table tes
t t
. The statistical difference is significance; it means that the teaching analytical exposition text using Indonesia Now Video as a media had
influence in improving students‟ achievement. Based on the computation there was significant difference between teaching analytical exposition text using
Indonesia Now Video as a media and teaching analytical exposition text without Indonesia Now Video as a media. It can be seen by the result of the test where the
students that were taught using Indonesia Now Video as a media got higher grades than the students that were taught without Indonesia Now Video as a media.
4.3 Grade of Achievement