49
r
kk
=
2
1 1
t
s pq
k k
The item is reliable if r
o
is higher than r
t
.
The coefficient of reliability of reading test is 0.921. It is higher than r-table for N = 40 at 5 level of significance 0.312. It
means that the reading test is reliable. The computation of reliability of reading test can be seen in Appendix 9 page 288.
The valid and reliable items are used to get the data of the experimental and control class. Then, the instruments are administered
to 27 of upper group high self-esteem group and 27 of lower group low self-esteem group from both classes, so there are twenty -
two students from the experimental class and twenty -two students from the control one 27 x 40 = 11 for upper group, 27 x 40 = 11
for lower group Anas, 2007: 398-400.
E. Technique of Analyzing the Data
The technique used in analyzing the data is descriptive analysis and inferential analysis. Descriptive analysis is used to know the mean, median,
mode, and standard deviation of the scores of the reading test. To know the normality and the homogeneity of the data, the writer uses normality and
homogeneity test. The normality and homogeneity tests are done before testing the hypothesis. Inferential analysis used is multifactor analysis of
variance 2 x 2. H
o
is rejected if F
o
is higher than F
t
. If H
o
is rejected, the analysis is continued to know which group is better using Tukey test. The
design of multifactor analysis of variance is as follows:
50
Teaching M ethod A
Self-esteem B Teams-Games-Tournament
A
1
Lecture A
2
High B
1
A
1
B
1
A
2
B
1
B
1
Low B
2
A
1
B
2
A
2
B
2
B
2
A
1
A
2
Figure 3. The Design of M ultifactor Analysis of Variance
Note: A
1
B
1
: the mean score of reading test of students having high self-esteem who are taught by using Teams-Games-Tournament.
A
2
B
1
: the mean score of reading test of students having high self-esteem who are taught by using lecture.
A
1
B
2
: the mean score of reading test of students having low self-esteem who are taught by using Teams-Games-Tournament.
A
2
B
2
: the mean score of reading test of students having low self-esteem who are taught by using lecture.
A
1
: the mean score of reading test of experimental class which is taught by using Teams-Games-Tournament.
A
2
: the mean score of reading test of control class which is taught by using lecture.
B
1
: the mean score of reading test of students having high self-esteem. B
2
: the mean score of reading test of students having low self-esteem.
51
The data are analyzed using the following ways: 1. The total sum of square
N X
X x
t t
t 2
2
2. The sum of squares between groups
N X
n X
n X
n X
n X
x
t b
2 4
2 4
3 2
3 2
2 2
1 2
1 2
3. The sum of squares within groups
2 2
2 b
t w
x x
x 4. The between-columns sum of squares
N X
n X
n X
x
t c
c c
c bc
2 2
2 2
1 2
1 2
5. The between-rows sum of squares
N X
n X
n X
x
t r
r r
r br
2 2
2 2
1 2
1 2
6. The sum-of-squares interaction
2 2
2 2
int br
bc b
x x
x x
7. The number of degrees of freedom associated with each source of variation:
a. df for between-columns sum of squares = C -1 b. df for between-rows sum of squares = R
– 1 c. df for interaction C
– 1 R – 1 d. df for between-groups sum of squares = G - 1
e. df for within-columns sum of squares =
1 n
f. df for total sum of squares = N – 1
Note: C = the number of columns
52
R = the number of rows G = the number of groups
n = the number of subjects in one group N = the number of subjects in all groups
8. Tukey test is used to know which teaching method is more effective or better to teach reading
a. Between column q = n
iance error
c X
c X
var
2 1
b. Between column
HI
q = n
iance error
r c
X r
c X
var
1 2
1 1
c . Between column
LI
q = n
iance error
r c
X r
c X
var
2 2
2 1
or q =
n iance
error r
c X
r c
X var
2 1
2 2
9. The statistic test is found by dividing the difference between the means by the square root of the ratio of the within group variation and the sample
size. ST : q =
n s
X X
w j
i 2
53
CHAPTER IV THE RES ULT OF THE S TUDY
