66
The testing is done to find out whether the data are homogeneous. This test is important because homogeneity of the data shows that the population is well-
formed. To test homogeneity of data, test is used. If
2
o
  is lower than
2
t
  at the level of significance  α=0.05, the data are homogeneous.
Table 12. Homogeneity Test
Group n
i
– 1 1  n
i
– 1 s
i 2
Lns
i 2
n
i
– 1 Lns
i 2
1 10
0.1 40.273
3.696 36.957
2 10
0.1 19.473
2.969 29.690
3 10
0.1 57.655
4.054 40.545
4 10
0.1 54.655
4.001 40.010
40 0.4
147.202
2
=
 
 
 
 
 
 
 
 
 
 
 
 
 
 
k N
1 1
n 1
1 k
3 1
1 s
ln 1
n B
i 2
i i
=
 
 
 
 
 
 
 
 
40 1
4 .
1 4
3 1
1 202
. 147
461 .
150
= 3.128
2 0.  05
= 7.815 Because
2
o
 3.128 is lower than 815
. 7
2
t
 , the data are homogeneous.
C. Data Analysis
Hypothesis test can be done after the results of normality and homogeneity test are fulfilled. The test is done by using multifactor analysis of variance 2 x 2.
Ho  is  rejected  if  F
o
F
t
meaning  that  there  is  a  significant  difference  and  an
67
interaction.  If  Ho  is  rejected,  the  analysis  is  continued  to know which group is better by using Tukey test. The multifactor analysis of variance 2 x 2 and Tukey
test are described as below. 1.  Summary of a 2 x 2 M ultifactor Analysis of Variance
Table 13. M ultifactor Analysis of Variance Source of variance
SS df
M S F
o
F
t .05
Between columns model
Between rows Intelligence
Columns by rows Interaction
295.364
5324.000
1833.091 1
1 1
295.364
5324.000
1833.091 6.867
123.775
42.617 4.08
Between groups Within groups
7452.455 1720.545
3 40
2484.15
43.014
Total 9173.00
43
a.  Because  F
o
between  columns  6.867  is  higher  than  F
t  .05
4.08,  the difference  between  columns  is  significant.  It  can  be  concluded  that  the
model  used  differs  significantly  from  one  another  in  their  effect  on  their performance  of  the  subjects  in  the  experiment.  The  students’  mean  of  c
1
62.09 is higher than t he students’ mean of c
2
56.90, so the students who are  taught  by  using  reciprocal  are better in reading skill than students who
are taught by using direct instructional. b.  Because  F
o
between  rows  123.77  is  higher  than  F
t  .05
4.08,  it  means difference  between  rows  is  significant.  It  can  be  concluded  that  the
difference  between  the  achievement  of  those  subjects  having  high
68 intelligence  and  those  having  low  intelligence  is  significant.  The  students’
mean  of  r
1
70.50  is  higher  than  the  students’  mean  of  r
2
48.50,  so the students  who  have  high  level  of  intelligence  have  better  reading  skill  than
the students who have low intelligence c.  Because  F
o
interaction  42.617  is  higher  than  F
t  .05
4.08,  it  can  be concluded  that  there  is  interaction  effect  between  the  two  variables,  the
teaching models and the degree of intelligence. It means that the effect of the teaching model used on the achievement depends on the subjects’ degree or
level of intelligence. 2. Tukey Test.
Tukey test is done to find which model is more effective and which group is better.
q
is found by dividing the difference between the means by the square root of the ration of the within group variation and the sample size.
Table 14. Tukey Test
PAIR TUKEY
CRITICAL A
1
– A
2
3.92 2.95
A
1
B
1
– A
2
B
1
9.18 3.11
A
1
B
2
– A
2
B
2
4.06 3.11
B
1
– B
2
15.8 2.95
A
1
B
1
- A
1
B
2
17.1 3.11
A
2
B
1
- A
2
B
2
4.56 3.11
69
a.  Comparing  two  means  between  columns  to  find  which  model  is  more effective
1 Comparing two means between columns q
o
is 3.92. The value of q
t
for α = 0.05  and  n  =  22  is  2.95.  Because  q
o
3.92  is  higher  than  q
t
2.95 reciprocal  teaching  differs  significantly  from  direct  instructional  for
teaching  reading.  The  mean  score  of  students  who  are  taught  by  using reciprocal  teaching  62.09  is  higher than   who are taught by using direct
instructional  56.90,  so  reciprocal  teaching  is  more  effective  than  direct instructional for teaching  reading.
2 Comparing  two  means  between  cells  A
1
B
1
and  A
2
B
1
.  q
o
is  9.18.  The value of q
t
for α = 0.05 and n = 11 is 3.11. Because q
o
9.18 is higher than q
t
3.11  reciprocal  teaching  differs  significantly  from direct instructional for  teaching  reading  for students having high intelligence. The mean score
of  students  having  high  intelligence  who  are  taught  by  using  reciprocal teaching  79.54  is  higher  than  those  who  are  taught  by  using  direct
instructional  61.45,  so  reciprocal  teaching  is  more  effective  than  direct instructional for teaching  reading  for students having high intelligence
3 Comparing  two  means  between  cells group A
1
B
2
and group A
2
B
2
. q
o
is 3.42. The value of q
t
for α = 0.05 and n = 11 is 3.11. Because q
o
4.06 is higher  than  q
t
3.11  instructional  model  differs  significantly  from reciprocal  teaching  model  to  teach  reading  for  students  who  have  low
intelligence.  The  mean  score  of  students  having  low  intelligence who are taught  by  using  direct  instructional  52.36  is  higher  than those who are
70
taught by using reciprocal teaching 44.63, so direct instructional is more effective  than  reciprocal  teaching  for teaching reading for students having
low intelligence. b.  Comparing two means between rows to find which group is better
1 Comparing two means between rows group B
1
and group B
2
. q
o
is 15.8. The value of q
t
for α = 0.05 and n = 22 is 2.95. Because q
o
15.8 is higher than  q
t
3.95  the  students  who  have  high  intelligence  are  significantly different  in  reading  from  the  students  who  have  low  intelligence.  The
mean score of students having high intelligence 70.5 is higher than those who  have  low  intelligence  48.5,  so the students who have high level of
intelligence  have  better  reading  skill  than  the  students  who  have  low intelligence.
2 Comparing  two  means  between  cells group A
1
B
1
and group A
1
B
2
. q
o
is 17.7. The value of q
t
for α = 0.05 and n = 11 is 3.11. Because q
o
17.7 is higher  than  q
t
3.11  the  students  who  have  high  intelligence  who  are taught  by  using  reciprocal  teaching  are  significantly  different  in  reading
from  the  students  who  have  low  intelligence  who  are  taught  by  using reciprocal  teaching.  The  mean  score  of  students  having  high  intelligence
70.5  is  higher  than  those  who  have  low  intelligence  48.5,  so  the students  who  have  high level of intelligence have better reading skill than
the  students  who  have  low  intelligence  those  taught  by  using  reciprocal teaching.
71
3 Comparing  two  means  between  cells group A
2
B
1
and group A
2
B
2
. q
o
is 4.56. The value of q
t
for α = 0.05 and n = 11 is 3.11. Because q
o
4.56 is higher  than  q
t
3.11  the  students  who  have  high  intelligence  who  are taught  by  using  direct  instructional  model  are  significantly  different    in
reading  from  the  students  who  have  low  intelligence  who  are  taught  by using  direct  instructional  model.  The  mean  score of students having high
intelligence  61.4  is  higher  than  those  who  have  low  intelligence 52.4, so  the  students  who  have  high  level  of  intelligence  have  better  reading
skill  than  the  students  who  have  low  intelligence  those  taught  by  using direct instructional model.
D. Discussion