Internet Learning Volume 6, Number 2, Fall 2017/Winter 2018 | Page 33
Internet Learning Journal
In Table 1, both the Kolmogorov–
Smirnov and Shapiro–Wilk tests indicate
that the sample data do not meet
the normality assumption (Sig. <0.05).
Table 1: Tests of normality
Tests of Normality
Kolmogorov–Smirnov a
Shapiro–Wilk
Statistic df Sig. Statistic df Sig.
0.150 691 0.000 0.892 691 0.000
0.125 121 0.000 0.906 121 0.000
A visual inspection of the histograms
shown in Figures 1 and 2 indicates
that the data do not appear to
follow the normal distribution. Furthermore,
an examination of the skewness
and kurtosis values can be used to
determine if a sample approximates a
normal distribution (Corder & Foreman,
2014). Table 2 shows the SPSS
output table for the kurtosis and skewness
of the test score data.
Table 2: Descriptive statistics
Descriptive Statistics
N Minimum Maximum Mean Std.
deviation
Skewness
Statistic Statistic Statistic Statistic Statistic Statistic Std.
Error
Statistic
Kurtosis
Std. Error
Score 812 12 100 78.50 17.803 −1.242 0.086 1.468 0.171
Valid N
(listwise)
812
Corder and Foreman (2014) indicate
that the z-scores for the kurtosis
and skewness must be manually computed.
The z-score for kurtosis is found
by subtracting zero from the kurtosis
statistic in SPSS and dividing the result
by the standard error.
The z-score for skewness is found
by subtracting zero from the skewness
statistic in SPSS and dividing the result
by the standard error.
In order for the test score data
to meet the normality assumption, the
z-score values must fall between −1.96
and +1.96 (with α = 0.05). Neither values
falls within this range, so we can
32