34
0.890 see Appendix II, if the instrument test refers to the value of reliability coefficient
α r table, the research instrument can be regarded as reliable.
3.5 Data Analysis Technique
To find out the category of learning achievement for the pre- and the post- test results, the researcher used the ideal mean and the ideal standard deviation.
Nurgiyantoro 1988: 395 states that for the achievement test, the ideal mean is 60 from the highest score and the ideal standard deviation is 25 from the ideal
mean. There were 10 items of the oral test in the form of questions
. It was a test,
which has the value 10 for the correct answer or based on the rubric of the speaking performance. So, in this research the highest score for the test is 100.
The ideal mean is 60 x 100 = 60. The ideal standard deviation is 25 of 60 equal to 15
. Thus the category of students’ speaking mastery can be put according to:
Table 4 . The Category of Students’ Learning Achievement
Score class Category
90 Excellent
75 – 89
very good 60
– 74 Good
45 – 59
Poor 30
– 44 very poor
29 extremely poor
The data from the procedure of data collecting show the score of test before the treatment and after the treatment. The score of test, which is made after
the treatment, indicates whether there is an improvement in speaking skill or not.
35
The statistics used in the data analysis in the quantitative research are descriptive and inferential analysis.
a. The Descriptive Analysis The descriptive analysis of the variables under this study is based on their
computation of the mean, standard deviation, and the lowest and highest scores. The formula for the computation of the mean
is
�
= �¡
�
Me = mean �¡ = total scores
N = the number of students
Sugiyono, 2010: 49 The formula of standard deviation is:
= �¡ −� ²
� − 1 = Standard deviation
� = mean � − 1 = degree of freedom
Sugiyono, 2010: 49 b. Test of Normality
Test of normality is used to determine whether the distribution of scores is normal or not. It indicates the discrepancy between the obtained frequencies and
the expected frequencies. In this case, chi square statistic formula can be used.
х
2
=
�−� 2 �
х
2
= Chi square O = Obtained frequency
E = Expected frequency
36
Weinbergh and Schumaker, 1969: 212 c. Test of Homogeneity
The test of homogeneity aims at knowing or not the sample variance is homogeneous, that is whether the scores of one group have homogeneous
variance with the scores of the other groups or not. For this, the F-test is applied. The F-test formula is as follows.
F
=
� � �
MST = Mean square treatmentBetween groups MSe = Mean Square errorWithin groups
Santosa and Ashari, 2005:68 d. Inferential StatisticsHypothesis Test
In the inferential statistics, the researcher utilized the statistical t-test. The test was utilized to uncover the difference between the scores of the speaking skill
test obtained in the pre-test and those in the post-test. T-test formula:
t =
� −�
� ² + � ² � +� −2
1 �
+
1 �
� = mean of X1 � = mean of X2
n = the number of students in control group n = the number of students in treatment group
Suharto, 2002:70
3.6 Research Procedure