Statistical Hypotheses RESEARCH METHODOLOGY
Table 4.1 Descriptive Statistics of Creative Thinking CT Ability Data Section 1 of
the Two Raters
CT_Section1 _Rater1
CT_Section1 _Rater2
N Valid
26 26
Missing Mean
50.32 56.73
Std. Error of Mean 3.587
4.530 Median
50.00 56.25
Mode 25
a
75 Std. Deviation
18.292 23.096
Variance 334.615
533.439 Skewness
.178 -.682
Std. Error of Skewness .456
.456 Kurtosis
-.899 .029
Std. Error of Kurtosis .887
.887 Range
63 88
Minimum 21
Maximum 83
88 Sum
1308 1475
Percentiles 25
36.46 40.63
50 50.00
56.25 75
66.67 75.00
a. Multiple modes exist. The smallest value is shown
Based on Table 4.1 above, the central tendency distribution of creative thinking ability data of 26 students of MA Khazanah Kebajikan Academic Year
20152016 assessed by the two raters is indicated by the mode, mean, and median. First, the most frequently scores mode of the first and second raters found
respectively are 25 in this case, actually there are two other modes rated by the first rater of which frequency is similar to the mode of 25, i.e., 54 and 42, though
only the smallest mode is presented in Table 4.1 above; also see Appendix XI and
Appendix XII for the further detail results, and 75. Next, it is found that the middle point median and average score mean of the first rater is lower than the
second rater median
1
median
2
= 50.00 56.25 and mean
1
mean
2
= 50.32 56.73. By taking account of the statistics above, it can be interpreted that based
on the assessment conducted by the first rater, most of the students’ creative thinking ability are categorized as not evident or emerging see the further criteria
in Chapter III which is indicated by the modes that are under or near the mean; on the other hand, based on the assessment conducted by the second rater, most of
the students’ creative thinking ability is categorized as expressing because it is indicated by the mode that is above the mean.
In addition, according to Table 4.1 represented above, the variability of data distribution between the first and second raters is indicated by the scores of
variance, standard deviation, range, skewness, kurtosis, and percentiles. First, by comparing the scores of variance and standard deviation found [i.e., variance =
334.615 and standard deviation = 18.292 the first rater; variance = 533.439 and standard deviation = 23.096 the second rater], the data of the first rater can be
considered more homogenous than the data deriving from the second rater because the variance and standard deviation scores from data of the first rater is
lower than the data from the second rater. Moreover, the range between the maximum and minimum scores of data deriving from the first rater is found to be
lower than the data from the second rater. Next, based on the scores of skewness of data from the first rater and the second rater 0.178 and -0.682 and kurtosis of
data from the first rater and the second rater -0.899 and 0.029 as well as the skewness ratio and kurtosis ratio which are obtained by dividing the skewness and
kurtosis scores with their standard error scores i.e. skewness ratio of the first rater = 0.39 and skewness ratio of the second rater = -1.50; kurtosis ratio of the first
rater = -1.01 and kurtosis ratio of the second rater= 0.03, both of the data of the first rater and the second rater can be considered to be normally distributed
because their skewness and kurtosis as well as their ratios scores are included in the reasonably accepted scores of normal data distribution, i.e., between -2 and 2.
Another indicator of variability which is based on the range of the middle 25, 50,
75 percent of the test scores are shown by percentiles scores, i.e., 36.46, 50.00, and 66.67 for data of the first rater, and 40.63, 56.25 and 75.00 for data of the
second rater. Next, to see the reliability of the data sets between the two raters, the inter-
rater reliability of the two raters is calculated. However, to determine the kind of the statistical test used, i.e., whether it is calculated through a parametric test or
non-parametric test, the linearity and normality data distribution are examined first.
1 Test of Linearity The linearity of the data between the two raters is examined through the
scatter plot presented in Figure 4.1 as follows:
Figure 4.1 Scatter Plot of CT Test Section 1 between Rater 1 and Rater 2
The scatter plot presented in Figure 4.1 above reveals that the data of rater 1 and rater 2 tend to have a fairly high relationship which is indicated by most of
the observed dots that are closely located to the linear assumption line that is drawn through the dots. Moreover, it may be interpreted that the data from the two
raters are considered to have a positive relationship because the dots in the plots show an indication from down left side to the up right side. Also, the data can be