MPPU 1034

APPLICATION OF STATISTIC IN EDUCATIONAL RESEARCH

ASSIGNMENT 1

KAMARUL ARIFFIN BIN KHAMIS

MPP181122
SECTION 04

PREPARED FOR:
DR MEGAT AMAN ZAHIRI BIN MEGAT ZAKARIA
 PART A
1) Title of Research
Study on the factors affect pupil’s performance in Mathematics.

2) Research Design

In this research, survey was selected as a research design, which is one of quantitative
research method. Survey research is the collection of data attained by asking individuals
questions either in person, on paper, by phone or online. So that, for this study I had
distributed 37 questionnaires to pupils in year 4 to year 6 to know what are the factors affect
their performance in mathematics. I also want to find out is there any different of pupil’s
performance based on gender and classes.

3) Research Questions

RQ1 How many males and females involved in the study? How many of them are 10, 11 or
12 years old?

RQ2 Is there exists a significant different of pupil’s performance of final examination

between the gender?

RQ3 Is there exists a significant different of pupil’s performance of final examination

between classes?

4) Research Hypotheses

H0b : There is no different of pupil’s performance of final examination between male and
female.
H1b : There is a different of pupil’s performance of final examination between male and
female.
H0c : There is no different of pupil’s performance of final examination between classes.
H1c : There is a different of pupil’s performance of final examination between classes.
5) Research Instruments
Research instrument used for this study is questionnaire. There are 37 questionnaires
were distributed among 37 pupils form Sekolah Kebangsaan Kangkar Pulai 2, from year 4
to year 6. The information were collected from pupils are gender, age, class, marks of
final examination, and class. The questionnaire also use the Likert Scale to measure how
pupils feel the factors affect their performance. They were required to choose the rating
scale from 1 to 3 for each factor.

6) Research Data

Figure 1: Data Structure

The research data I collected for this study through questionnaire are gender, age,
class that categorized as nominal. I also collected ratio data such as marks of final
examination and ordinal data such as the factors affect the pupil’s performance in
mathematics.

7) Data Management (as necessary)

Figure 2: Data in SSPS

In SSPS data as shown above. The gender, age, class, marks of final examination and factors
of affecting performance of pupils were computed in SSPS.

8) Investigations of Assumptions on Data Normality (when applicable)

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
 
Statistic df Sig. Statistic df Sig.
 Marks of
final 0.197 37 0.001 0.903 37 0.003
examination
a. Lilliefors Significance Correction
Figure 3: Tests of Normality
Based on test of normality, marks of final examination is known as not normal data. So
that, non parametric test will be used to accept or to reject hypothesis null. Non parametric
tests include the Wilcoxon signed rank test, the Mann-Whitney U Test and the Kruskal-
Wallis test. The hypothesis used are :

Ho : Data is not normally distributed.

When testing for normality :

Probabilities ¿ 0.05 mean the data are not normal
Since the respondent not more than 2000 Shapiro-Wilk sig was measured for normality
test .So in this case, we have to reject Ho since the Sig of Shapiro-Wilk shows that it is less
than 0.05. Thus the data above is not normal.

9) Results

RQ1

How many males and females are involved in the study? How many of them are 10, 11 or 12
years old?

Valid Cumulative
  Frequency Percent Percent Percent
Vali male 18 48.6 48.6 48.6
d female 19 51.4 51.4 100.0
Total 37 100.0 100.0  

Figure 4: Table of Gender

Based on figure 4, there are 18 males and 19 females participated in this study. So that, total
of pupils involved are 37.
 BAR GRAPH
20
19
19
18
18
17
16
Total

15 Frequency
14
13
12
11
10
Gender

Figure 5: Bar Graph of Gender

Valid
Frequenc Percen Percen Cumulativ
  y t t e Percent
Vali 10 13 35.1 35.1 35.1
d yrs
11 11 29.7 29.7 64.9
yrs
12 13 35.1 35.1 100.0
yrs
Tota 37 100.0 100.0  
l

Figure 6 : Table of Age

Based on figure 5, there are 13 pupils age 10 years old, 11 pupils age 11 years old, and 13
pupils age 12 years old were participated in this study. So that, total of pupils involved are
37.
 Bar Graph
13.5
13 13
13

12.5

12
Total

11.5
11
11

10.5

10
10 yrs 11 yrs 12 yrs
Age

Figure 7: Table of Age

RQ2
H0b : There is no different of pupil’s performance of final examination between male and
female.
H1b : There is a different of pupil’s performance of final examination between male and
female.

Marks of final
 
examination
Mann-Whitney U 103
Wilcoxon W 334
Z -1.995
Asymp. Sig. (2-tailed)
0.046
Exact Sig. [2*(1-tailed
.047b
Sig.)]

Figure 8: Mann- Whiteny U test

When normality test for pupil’s performance of final examination run, found that the data
was not normal. Thus, Non-Parametric measurement was used to test the hypothesis null. In
this test, Mann- Whitney U was chosen since it is suitable for 1 dependent and 1 independent
variable. Table above shows that the respondents were consist of 21 male and 16 female
which almost equally uniform. As a conclusion, reject H0b because the asymptotic Sig for
this test is 0.046, which is less than 0.05. Thus, there is no different of pupil’s performance
between male and female.

RQ3
H0c : There is no different of pupil’s performance of final examination between classes.
H1c : There is a different of pupil’s performance of final examination between classes.

Marks of
final
examination
Chi-Square 27.857
df 5
Asymp. .000
Sig.

Figure 9: Kruskal Wallis Test

When normality test for pupil’s performance of final examination run, found that the
data was not normal. Thus, Non-Parametric measurement was used to test the Hypothesis
Null. In this test, Kruskal Wallis Test was chosen since it is suitable for 1 dependent and two
or more groups of independent variable. Table above shows that the respondents were consist
of 21 males and 16 females, which almost equally uniform. As a conclusion, reject H0c
because the asymptotic Sig for this test is 0.00, which is less than 0.05. Thus, there is
different of pupil’s performance between classes.
Because the value of significance for Kruskal Wallis Test is 0.000, which is less than
0.05, I need to run Man Whitney for post hoc comparisons in pair. Post hoc tests are run to
confirm where the differences occurred between groups, they should only be run when we
have a shown an overall statistically significant difference in group means. 

Sum
Mean of
class N Rank Ranks
Marks of Usah 2 12.50 25.00
final a
examinatio Rajin 11 6.00 66.00
n Total 13    
 Test Statisticsa
  Marks of final examination
Mann-Whitney U 0.000
Wilcoxon W 66.000
Z -2.177
Asymp. Sig. (2- .029
tailed)
Exact Sig. [2*(1- .026b
tailed Sig.)]
a. Grouping Variable: class
b. Not corrected for ties.

Figure 10: Post Hoc Test between Usaha and Rajin

Table above shows that the respondents were consist of 21 males and 16 females,
which almost equally uniform. As a conclusion, the asymptotic Sig for this test is 0.029,
which is less than 0.05. Thus, there is different of pupil’s performance between class Usaha
and Rajin.

Sum
Mean of
class N Rank Ranks
Marks of Usaha 2 12.25 24.50
final Jujur 11 6.05 66.50
examinatio Total 13    
n

Test Statisticsa
Marks of final
  examination
Mann-Whitney U .500
Wilcoxon W 66.500
Z -2.081
Asymp. Sig. (2-tailed) .037
Exact Sig. [2*(1-tailed Sig.)] .026b
a. Grouping Variable: class
b. Not corrected for ties.
Figure 11: Post Hoc Test between Usaha and Jujur
Table above shows that the respondents were consist of 21 males and 16 females,
which almost equally uniform. As a conclusion, the asymptotic Sig for this test is 0.037,
which is less than 0.05. Thus, there is different of pupil’s performance between class Usaha
and Jujur.
 Sum
Mean of
class N Rank Ranks
Marks of final Usaha 2 5.50 11.00
examination Tekun 4 2.50 10.00
Total 6    

Test Statisticsa
Marks of final
  examination
Mann-Whitney U 0.000
Wilcoxon W 10.000
Z -1.852
Asymp. Sig. (2-tailed) .064
Exact Sig. [2*(1-tailed Sig.)] .133b
a. Grouping Variable: class
b. Not corrected for ties.

Figure 12: Post Hoc Test between Usaha and Tekun

Table above shows that the respondents were consist of 21 males and 16 females,
which almost equally uniform. As a conclusion, the asymptotic Sig for this test is 0.064,
which is more than 0.05. Thus, there is no different of pupil’s performance between class
Usaha and Tekun.
Sum
Mean of
class N Rank Ranks
Marks of final Usaha 2 5.50 11.00
examination Gigih 4 2.50 10.00
Total 6    

Test Statisticsa
Marks of final
  examination
Mann-Whitney U 0.000
Wilcoxon W 10.000
Z -1.852
Asymp. Sig. (2-tailed) .064
Exact Sig. [2*(1-tailed Sig.)] .133b
a. Grouping Variable: class
b. Not corrected for ties.

Figure 13: Post Hoc Test between Usaha and Gigih

 Table above shows that the respondents were consist of 21 males and 16 females,
which almost equally uniform. As a conclusion, the asymptotic Sig for this test is 0.064,
which is more than 0.05. Thus, there is no different of pupil’s performance between class
Usaha and Gigih.

Sum
Mean of
class N Rank Ranks
Marks of final Usaha 2 6.50 13.00
examination Ikhlas 5 3.00 15.00
Total 7    

Test Statisticsa
Marks of
final
  examination
Mann-Whitney U 0.000
Wilcoxon W 15.000
Z -1.936
Asymp. Sig. (2-tailed) .053
Exact Sig. [2*(1-tailed Sig.)] .095b
a. Grouping Variable: class
b. Not corrected for ties.

Figure 14: Post Hoc Test between Usaha and Ikhlas

Table above shows that the respondents were consist of 21 males and 16 females,
which almost equally uniform. As a conclusion, the asymptotic Sig for this test is 0.053,
which is more than 0.05. Thus, there is no different of pupil’s performance between class
Usaha and Ikhlas.
 PART B
1) Title of Research
Study on the factors affect pupil’s performance in Mathematics.

2) Research Design

In this research, survey was selected as a research design, which is one of quantitative
research method. Survey research is the collection of data attained by asking individuals
questions either in person, on paper, by phone or online. So that, for this study I had
distributed 37 questionnaires to pupils in year 4 to year 6 to find out the relationship between
pupil’s performance in test 1 and final examination. I also want to find out is there any
different of pupil’s performance between test 1, mid-term examination and final examination
and the relationship between test 1 and final examination.

3) Research Questions

RQ1 Is there exists a significant different of pupil’s performance between test one and
final examination?

RQ2 Is there exists a significant relationship of pupil’s performance between test one and
final examination.

4) Research Hypotheses

H0a : There is no different of pupil’s performance between test one and final examination.

H1a : There is different of pupil’s performance between test one and final examination.
H0b : There is no relationship of pupil’s performance between test one and final
examination.
H1b : There is relationship of pupil’s performance between test one and final examination.

5) Research Instruments
 Research instrument used for this study is questionnaire. There are 37 questionnaires
were distributed among 37 pupils form Sekolah Kebangsaan Kangkar Pulai 2, from year 4
to year 6. The information were collected from pupils are gender, age, class, marks of
final examination, marks of mid-term examination, marks of test 1 and class. The
questionnaire also use the Likert Scale to measure how pupils feel the factors affect their
performance. They were required to choose the rating scale from 1 to 3 for each factor.

6) Research Data

Figure 1: Data Structure

The research data I collected for this study through questionnaire are ratio data such
as marks of final examination, marks of test 1 and marks of mid-term examination.
Ordinal data is used find the information about the factors affect the pupil’s performance
in mathematics.

7) Data Management (as necessary)

Figure 2: Data in SSPS

In SSPS data as shown above. The marks of final examination, marks of test 1, marks
of mid-term examination were computed in SSPS.
8) Investigations of Assumptions on Data Normality (when applicable)

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Marks of test 1 .186 37 .002 .903 37 .004
Marks of final .003
.197 37 .001 .903 37
examination
a. Lilliefors Significance Correction
Figure 3: Tests of Normality
Based on test of normality, marks of final examination is known as not normal data.
So that, non parametric test will be used to accept or to reject hypothesis null. Non
parametric tests include the Wilcoxon signed rank test, the Mann-Whitney U Test and the
Kruskal-Wallis test. The hypothesis used are :

Ho : Data is not normally distributed.

When testing for normality :

Probabilities ¿ 0.05 mean the data are not normal
Since the respondent not more than 2000 Shapiro-Wilk sig was measured for normality
test .So in this case, we have to reject Ho since the Sig of Shapiro-Wilk shows that it is less
than 0.05. Thus the data above is not normal.

9) Results

RQ 1
H0a : There is no different of pupil’s performance between test one and final examination.
H1a : There is different of pupil’s performance between test one and final examination.
Ranks
  Mean Rank
Marks of test 1 1.00
Marks of final 2.00
examination
 Test Statisticsa
N 37
Chi-Square 37.000
df 1
Asymp. Sig. .000
a. Friedman Test

Figure 4:Friedman test

When normality test for marks of test one and marks of final examination, I found that
the data were not normal. Thus, Non-Parametric measurement was used to test the
Hypothesis Null. The Friedman Test result is 0.000… (p¿ 0.05 ¿ which reject the hypothesis
null (H0a). It means that there is different of marks between test 1 and final examinations.
Thus, the respondents have either get higher marks or lower marks between the two
assessments.

RQ 2

H0b : There is no relationship of pupil’s performance between test one and final
examination.
H1b : There is relationship of pupil’s performance between test one and final examination.

Test : Spearman correlation

Marks of final
  Marks of test 1 examination
Spearman's rho Marks of test one Correlation 1.000 .942**
Coefficient
Sig. (2-tailed)   .000
N 37 37
Marks of final Correlation .942** 1.000
examination Coefficient
Sig. (2-tailed) .000  
N 37 37
**. Correlation is significant at the 0.01 level (2-tailed).

Figure 5: Correlation

The Spearmen correlation was used because the data are not normal. The test shows
that the significant value is 0.000, which is less than 0.05, and the correlation is 0.942. Hence
we have to reject H0b. Thus, it shows that there is significant relationship between test one
and final examination that are linear, positive and very strong.