COUNTRY ANALYSIS PROFILE: MALAYSIA

Introduction

Malaysia has been undertaking bold steps in improving the level of education of its
population over the past quarter century or so. Compared to other countries in East Asia, data
from the World Development Indicators, 1998 showed that Malaysia’ public education
expenditure as a percentage of GNP was highest compared with other countries in East Asia in
1980 and even in 1995 despite its decrease by 0.7%.

Yet despite the high priority given to education, and the efforts made by Malaysia to
ensure that the poor are relatively well educated, there are still some concerns that differences
are growing between urban and rural areas, in terms of access and quality. The Government of
Malaysia has also given special priority the improvement of science and math programs in their
schools while additional emphasis are continuing to improve the quality of and access to
schools in underserved areas.

Using the public release database of TIMMS (IEA’s Repeat of the Third International
Mathematics and Science Study at the Eight Grade), this study aimed to find out what other
variables (particularly school-related and policy-manipulable variables) could help improve
achievement in science and math. More specifically, the study aimed to examine such
questions as:

• Are schools providing the necessary inputs for students to attain higher
achievement in math and science?
• What teaching-learning practices could help improve achievement in math and
science?
• What school policies are contributing to higher math and science achievement?

The study also looked at students’ profiles and determined what among the individual-
and family-related variables could affect students’ achievement in math and science.
Questions, such as the following will be examined:

• How do family-related variables such as parents’ educational level, access to

computer at home, number of books at home and language spoken at home,
among others, affect students’ achievement in science and math?
• How does student’s self-concept affect their achievement in science and math?

The Malaysia Educational System

The Malaysia education system provides free education to students between the ages of
7 to 17 for a total of 11 years of universal education. Admission age to the first year of primary
schooling is usually seven and the graduating age for a first bachelor degree is about 22 years
old. In school year 2000-2001, primary education has a total enrolment of 3,017,902, which
registered a gross enrolment ratio of 98%. At the secondary level, school-age population was
3,136,000 of which 70% were enrolled. The country has a literacy rate of 93%, one among the
highest in the world.

By age group, the following is the grouping in the educational system:

2

• Pre-school education – from age 5 (for 2 to 3 years)

• Primary education from age 7 (for 5 to 7 years)
• Lower secondary education from age 13 (for 3 years)
• Upper secondary education from age 16 (for 2 years)
• Post-secondary education or Sixth Form from age 18 (for 1 to 1.5 years)
• Higher education
o Undergraduate studies from age 20 (for 3 to 5 years)
o Post graduate studies (for 1 to 5 years)

The primary and secondary levels follow a national curriculum that prepares students for
the common public examinations at the end of primary, lower secondary and upper secondary
levels. Bahasa Malaysia (Malay language) is the national language and medium of instruction
but English is a compulsory subject in the schools. To meet the needs of multi-ethnic groups in
the population, Mandarin or Tamil can be used as medium of instruction.

BACKGROUND

In 1999, the International Association for Educational Achievement (IEA) conducted the
Third International Mathematics and Science Study (TIMSS) measuring the achievement trends
in science and mathematics of eight-grade (junior secondary) students. Using the public
release database of TIMSS on school achievement among 13-year olds in Asia, the Asian
Development Bank commissioned SEAMEO INNOTECH to undertake a statistical analysis of
10 countries in the Asia region to map the learning process taking place in schools within each
country and to identify predictors of achievement and patterns in the distribution of that learning
among students. The study aimed to present country profiles of the following ten (10) countries
that participated in TIMSS (Philippines, Malaysia, Indonesia, Thailand, Singapore, Hong Kong,
Taiwan, Korea, Iran, and Japan).

Thirty-eight countries participated in TIMMS of which 11 were from the Asia-Pacific

region. The target population was the grade or year level where the majority of 13 year-olds are
enrolled in at the time of testing, and in the case of Malaysia was the lower secondary level after
completing five to seven years of primary education.

In the development of test items in science and math, the same curriculum framework in
TIMSS 1995 was used. The curriculum framework contains three dimensions or aspects. The
content aspect represents the subject matter content of school mathematics and science, the
performance expectations aspect describes, in a non-hierarchical way, the many kinds of
performance or behaviours that might be expected of students in school mathematics and
science, while the perspective aspect focuses on the development of students’ attitudes,
interest, and motivation in the subjects.

In addition to the science and mathematics achievement tests, the students were also
asked to answer questions pertaining to their attitude towards mathematics and science, their
academic self-concept, classroom activities, home background, and out-of-school activities.
The students were also asked to provide general home and demographic information.

The mathematics and science teachers of sampled students were also asked to
complete a teacher questionnaire comprised of two sections. The first section covered general
background information on the preparation, training, and experience of teachers and about how
teachers spend their time in school, and probed their views on mathematics and science. The
 3

teachers were also asked questions related to instructional practices in the class or classes
selected for the TIMSS 1999 testing. Information about how the curriculum was implemented
was also asked of the teachers such as the number of periods the class spent on a range of
mathematics and science topics, and about the instructional strategies used in the class,
including the use of calculators and computers. Teachers also responded to questions about
teaching emphasis on the topics in the curriculum frameworks.

The heads of schools also responded to questions about school staffing and resources,
mathematics and science course offerings, and support of teachers.

METHODOLOGY

From the TIMSS database provided, country files were extracted and set up as separate
data files. Initial runs on the data included frequencies, cross tabulations and breakdown to
determine which among the many variables would have significant effects on science and math
scores. The standardized math and science raw scores were used as dependent variables in
the analysis.

Regression analysis using stepwise method was used to identify which among the many
variables are good predictors of math and science achievement. In determining the best model
that fits the regression line, the condition index was used as a criterion. A model with variables
having condition indices of not more than 15 was taken as the best model that fits the
regression line.

The results of statistical analysis undertaken are presented below as country survey
results.

SURVEY RESULTS (TIMSS)

A. Profile of Sample Schools

• 152 schools were included as national sample

• 90 or 60% are located outside of town (rural areas) while 60 or 40% came from
urban areas (near center of town)
• There are 20% or 30 schools with less than 50 classroom teachers (classified as
small schools) and another 20% or 30 schools with 100 or more classroom
teachers (big schools)
• 90 or 60% of the total sample schools have 50 to 99 teachers (medium-sized
schools)
• Mean scores in math and science increase as school size increases, the
difference in scores among the three groups are significant.
• 120 or 78.9% of the sample schools have majority of their teaching staff who are
fairly new in the service (less than five years in the service)
• Mean scores in math of students in schools with teachers who have stayed for
more than 5 years are significantly higher compared to students in schools with
teachers who are fairly new in the service.
• 76 or 50.7% of the sample schools have from 4 to 6% of students who are
absent on a typical school day.
• Math and science mean scores of students significantly decrease as student
absentees in schools increase
4

School’s Capacity to Provide Instruction

• Except for AVR, it seems that schools’ capacity to provide instruction is not
affected by shortages in materials, classrooms, laboratory equipments and
qualified teachers.
• Students in schools with access to Internet have significantly higher mean scores
in math and science
• 26 or 17% of the total sample schools are a lot affected by shortage in AVR and
the science mean scores of these groups are significantly lower compared to the
other groups.

School Policies

• 13 schools indicated they admit pupils on the basis of interviewing the parents.
The mean scores in math and science of students enrolled in these schools are
significantly lower compared to the mean scores of students enrolled in 137
schools which do not based their admittance on this policy.
• 39 or 26% of the sample school based their admittance by giving preferences to
students whose older family members are already enrolled in same school. Mean
scores of students enrolled in these schools are significantly higher.

Classroom teaching practices

• Students who often do experiments or practical investigation in class performed

better in science test than students who rarely do same activity in class.
• Students who always or often work on worksheet or textbooks alone have
significantly lower scores in math compared to students who never work alone.
• Students whose teacher always gives homework performed better compared to
students whose teacher never gives them assignments.
• Students who always and often begin homework in class did better in Math than
students who never begin their homework in class.
• Students whose teacher always gives tests or quiz in class have significantly
higher scores in math.
• Students who rarely use Internet/emails in math and science projects have
significantly higher scores compared to student who often use Internet.
• Students who never or rarely copy on board performed better in math and
science than students who always/often do the same activity.
• Students who always work on math and science projects have significantly
higher scores in both subjects than students who never do the same activity.
• Students who rarely or never look at the textbook while teacher talks about the
new topic performed better compared to student who always look at the textbook
while the teacher is explaining the new topic.
• Students whose teacher always/often use computer to demonstrate ideas in
class performed better in science than students whose teacher never use
computers in class
• Students who always use things from everyday life in solving science problems
have significantly higher scores in science.
• Students whose teacher often use overhead projector in class and let her
students do the same performed better in science tests.
 5

B. Profile of Sample Students

• Total sampled students equal 5577 of which 3077 (55.17%) are girls and 2500
(44.83%) are boys
• By age group, there were 1259 (22.6%) students aged 14 and below and 4318
(77.4%) are above 14 years of age.
• On the average, girls performed better than boys in math test but the boys
outperformed the girls in science test. The differences in scores are statistically
significant.

Home background and family related variables

• How student perceives himself is an indicator of both math and science scores.
Those who expect to finish college are those students who got significantly
higher scores in both subjects.
• Student’s attitude towards math/science has an affirmative result on his
performance of the said subject.
• Students who do well in science to please their parents and themselves
performed really well in the science test.
• Presence of educational resources at home has a positive effect on the
performance of students in math and science.
• Educational background of the parents is an indicator of student’s math and
science score. The differences in mean scores across groups of students
categorized by level of educational attainment of their parents are significantly
different. Students whose parents have college degrees have significantly higher
mean scores in both science and math. Mean scores in tests of both subjects
increase as educational level of parents increase.

father mother
Mean
Highest Educational Mean Mean
Mean scores scores
Attainment of scores of scores of
of students of
Parents students students in
in science students
in math science
in math
Some primary sch. 48.63 48.20 47.94 48.72
Finish primary 48.48 48.54 48.19 47.63
Some high school 50.87 49.96 49.32 50.18
Finish high school 51.62 51.61 50.52 50.72
Some voc. schooling 54.96 55.99 52.68 52.62
Some coll. schooling 52.66 54.46 51.92 50.81
College graduate 57.06 57.49 56.85 56.30

• About 60% of the total sample students always use the language of the test at
home while 10% never use the language of the test at home. The latter group did
better in both tests compared to the first group.
• Students with little or no time to watch TV or video and those who are exposed to
TV for more than five hours everyday have considerably lower scores in math
and science.
6

• Students who have time to play computer games during school days performed
better in science and math tests compared to students who have little or no time
for the said activity.
• The more time these students spend taking extra lesson in Math or Science, the
better is their performance in both tests.

MATH SCIENCE
Number Number
Time spent in extra % of Mean % of Mean
of of
lesson or cramming Students scores Students scores
Students Students
No time 2998 54.57 49.64 3590 65.53 50.20
Less than 1 hour 475 8.65 48.23 457 8.34 48.39
1 – 2 hours 1676 30.51 51.18 1138 20.77 51.22
3 – 5 hours 300 5.46 54.19 269 4.91 50.01
more than 5 hours 45 0.81 55.34 24 0.45 54.24
T= 5494 T=100% T=5478 T=100%

• 4824 or 88.6% are considered full-time students because they have no time to
work in paid jobs. These students got significantly higher scores in both tests.

Results of Regression Analysis

In the stepwise regression analysis, there were 29 models for math. However, using the
condition index as the criterion to determine which among the models generated best fit the
regression line, model 7 was taken as the condition index of the last variable considered in the
model is 9.3127. Model 8 has a condition index of 21.164 on its 8th variable. The seven
predictor-variables for math scores with its coefficients can be grouped into the following and is
presented below:

Predictor Variables for Math

Coefficients
Std. Correlations
Variables* Beta S.E. + Sig. 0
Beta Partial Part
Order
Constant 52.06 0.71 72.37 .00
a. Students’ home background
(x1) Presence of computer at 5.08 0.40 .26 12.73 .00 .37 .29 .25
home
(x2) Speak language of test at -4.61 0.39 -.24 -11.70 .00 -.33 -.27 -.23
home

b. School policy
(x3) Percent of students absent -0.99 0.11 -.18 -9.23 .00 -.26 -.22 -.18
from class
(x4) Admittance policy: -5.54 0.66 -.17 -8.46 .00 -.22 -.20 -.17
interviewing parents

c. Student’s self-concept
(x5) Highest educational level 3.13 0.44 .143 7.10 .00 .21 .17 .14
sought
(x6) I usually do well in math 3.58 0.49 .15 7.33 .00 .18 .17 .14
 7

Coefficients
Std. Correlations
Variables* Beta S.E. + Sig. 0
Beta Partial Part
Order
Constant 52.06 0.71 72.37 .00
d. Teaching Practices
(x7) Student work from worksheet -3.87 0.58 -.13 -6.65 .00 -.18 -.16 -.13
or textbook alone
* Codes of these variables are presented as Attachments

From the model, the following is the regression line for mathematics:

Y = 52. 06 + 5.08X1 – 4.61X2 -0.99X3 – 5.54X4 + 3.13X5 +3.58X6 – 3.87X7 + e

For this model, R-square = .337, which means that about 34% of the variation in math
score is explained by the regression equation.

The regression equation means that:

• Average score in math in Malaysia is 52.06 when all the other variables in the
equation are zero.
• Average math score will increase by 5. 08, 3.13 and 3.58 points for every unit
increase in the following variables – presence of computer at home, student’s
desire to have college education, and student’s belief that they usually do well in
math, holding all other variables in the equation constant.
• An increase of 1 unit in the following variables in the equation would mean
corresponding decreases in the average score in math, holding all the other
variables in the equation constant.
o 4.61 points less for the variable “speak the language of test at home”
o about 1 point less for the variable “percent of students absent in a typical
class day”
o 5.54 points less for the variable “admittance based on interview of
parents”, and
o 3.87 points less for the variable “students working from worksheets or
textbooks on their own”.

For science, the model which best fits the regression line is model 12 with condition
index equivalent to 14.209. The predictor variables can be grouped into the following with their
corresponding coefficients:
8

Predictor Variables for Science

Coefficients
Std. Correlations
Variables* Beta S.E. + Sig. 0
Beta Partial Part
Order
Constant 47.57 0.95 49.86 .00
a. Students’ home background
(x1) Presence of computer at 3.72 0.45 .19 8.36 .00 .33 .20 .17
home
(x2) Speak language of test at -2.391 0.44 -.12 -5.47 .00 -.23 -.13 -.11
home
(x3) Highest educational level of 0.27 .53 .14 6.15 .00 .15 .14 .12
parents

b. School policy
(x4) Percent of students absent -0.78 0.11 -.15 -6.87 .00 -.22 -.17 -.14
from class
(x5) Admittance policy: -3.90 0.73 -.11 -5.38 .00 -.17 -.13 -.11
interviewing parents
(x6) Admittance policy: family 2.00 0.45 .09 4.45 .00 .14 .11 .09
members at school

c. Student’s self-concept
(x7) I usually do well in science 3.00 0.51 .12 5.95 .00 .15 .14 .12

d. Teaching Practices
(x8) Do experiments in science 2.26 0.42 -.12 5.39 .00 .21 .13 .11
(x9) Use computers to -9.99 2.32 -.09 -4.30 .00 -.09 -.10 -.09
demonstrate ideas
(x10) Explain new science topic 3.04 .69 .09 4.39 .00 .14 .11 .09
and rule

e. Others
(x11) Type of community 2.26 .42 .12 5.33 .00 .22 .13 .11
(x12) Gender -1.67 .41 -.09 -4.08 .00 -.04 -.10 -.08
* Codes of these variables are presented as Attachments

SCIENCE REGRESSION EQUATION

Y = 47. 57 +3.72X1 –0.78X2 +2.26X3 –2.39X4 +3.27X5 +3.00X6 +2.26X7 -.904X8

+2.00X9 -9.99X10 +3.04X11 -1.67X12 + e

R-square for this model is 0 .281 which means that about 28% of the variation in science
scores of student is explained by the regression equation. The equation further means that:

• Average score for science is 47.57 when all the other variables in the equation are
zero
• In the equation, there are 7 variables with positive effects on the average score in
science. This means that a one unit increase in these 7 variables will bring about
corresponding increases in the average score in science holding all other variables in
the equation constant. These are the variables and their corresponding increases:
o 3.72 points for an increase of one unit in the variable “PRESENCE OF
COMPUTER AT HOME”
 9

o 2.26 points for an increase of one unit in the variable “students who DO
EXPERIMENTS in class
o 3.27 points for an increase of one unit for the variable “students with
parents who finished college degree”
o 3 points for every one unit increase in the variable “Students who agree
that they USUALLY DO WELL IN SCIENCE”
o 2 points for an increase of one unit for the variable “location of schools”
o 2 points for an increase of one unit in the variable “students enrolled in
schools which GIVE PREFERENCE TO STUDENTS WITH OLDER
BROTHERS/SISTERS IN SAME SCHOOL”
o 3 points for an increase of one unit in the variable “teacher always
explains the rules and definitions when they begin a new topic in class”
• Variables that have decreasing effect on the average score in science, holding all
other variables in the equation constant are the following:
o 0.78 points decrease for a one unit increase in “percentage of
STUDENTS ABSENCE ON A TYPICAL CLASS DAY”
o 2.4 points decrease for the variable “SPEAK THE LANGUAGE OF THE
TEST AT HOME”
o 0.90 points decrease for a one unit increase of “students enrolled in
schools which based their ADMITTANCE solely in INTERVIEWING the
PARENTS”
o 10 points decrease in the average score in science for a one unit increase
in the variable “teacher always/often uses computer to demonstrate ideas
in science class”
o female students will have a decrease of 1.67 points in average score in
science

Implications

The regression equation for Math explains about 34% of the variation in Math score. This
means that about 66% of the variation in Math score can be attributed to factor-variables that
are not in the regression equation.

Below are some insights that can guide actions on how to improve student achievement in
science in Malaysia. While presence of a computer at home is the strongest predictor of science
achievement, the Ministry of Education cannot formulate a policy which requires every student
to have a computer at home. A feasible policy is for the Ministry of Education to provide a
computer laboratory in each school for the use of the students. Where this is not possible
because of resource constraints schools and school heads should be encouraged to collaborate
with parents in putting up a computer laboratory in the schools where their children are enrolled.
Where this is not possible, they should be encouraged to tap the support of the private sector,
particular business and industry, to adopt schools without computer laboratories. After all, it is
the schools that educate and train the future members of the labor force, who, in turn, are
absorbed by business and industry.

Speaking the language of the test at home is not advantageous to the students. Indeed, a one-
unit increase in this variable translates to 5.54 points reduction in student Math score, taking all
the other factors in the regression constant.

Efforts at improving the self-concept of students by motivating them to go to college and

believing in themselves that they will do well in math should be in every math teacher’s and
10

school head’s mind. Schools should provide co-curricular activities and learning experiences
aimed at increasing students’ educational aspirations and performing well in math. Both of these
are likely to boost student achievement motivation. Students believing in themselves to do well
in math is a self-fulfilling prophecy; they will achieve better in math achievement tests. If
students’ belief in their capacity to do math is strong and this is reinforced by positive
expectations by the teachers of their student performance, this will result in even greater student
academic achievement.

Self-instruction or self-learning on the part of the students facilitated by the use of worksheets or
textbooks alone have a depressing effect on student achievement. A one-unit increase in this
variable translates itself into almost 4 points of decrease in the students average math score.
School heads, therefore, should drive home the point to their teachers that unsupervised
seatwork in which students work by themselves using worksheets and textbooks is ineffective; it
is not a good teaching practice.

Likewise, school heads should not only consider interviewing parents as basis for their schools’
admissions policies. Results of parent interviews, student entrance examinations, past
academic records and the like will produce better results in terms of student achievement in
math.

In science, the variation in science scores explained by the regression equation with 12
independent variables is lower than in math (0.281 vs. .331).

A stricter school policy on student absences from class should be observed. An increase of 1 %
in “percentage of students who are absent from class” translates into 0.78 decrease in student
science scores. As in math, school heads should not only consider interviewing parents as basis
for student admission to their respective schools. However, it is good school policy to admit
students with siblings already studying in the school.

As in math, students who believe in themselves to do well in science, a measure of self-

concept, scored much higher than those who are low in their self-concept. Again, the self-
fulfilling prophecy principle works, prodding students to perform well in science.

Good science teaching practices include students performing experiments and teachers always
explaining rules and definitions when beginning a new topic in class. The first one exhorts every
teacher to shy away from the “textbook approach” to teaching the subject, with very limited
learning experiences provided for students to engage in scientific experimentation that enable
them to discover concepts and principles and construct their own learning rather than the
teacher constructing the learning for them. By itself, student experimentation is learning by
doing, with students engaged in active discovery and rediscovery, a very constructivist
approach to learning. It is, therefore, important for schools desiring to improve student science
achievement to expose their teachers in constructivist teaching and learning.

Demonstrating ideas in a science class with the use of a computer is not an effective teaching
practice. The more a teacher uses this teaching technique the lower student achievement
scores become. Why is this so? The computer is not the right medium for use in demonstrating
science ideas, the reason being that learning is passive, not active. The learner is reduced to
being a passive observer; he/she does not manipulate any tool or device that will make learning
more active and discovery-oriented.
 11

Lastly, if the school is located in an urban area the chances of such school performing better
science achievement tests is greater than when a school is located in a rural area. While school
heads and policymakers can’t do anything to change the location of the school from rural to
urban, they can do much in terms of science teaching practices that really influence student
achievement as pointed out above. They can do much in providing learning experiences and
activities that will improve students’ self-concept and in making their schools a conducive place
of learning.

Attachment 1
12

Codes for variables in Math Multiple Regression Analysis

A. Student’s home background

Presence of computer at home -X1 (coded 1 for yes and 0 for no)
Speak the language of test at home – X2 (coded 1 for yes and 0 for no)

B. School policies
Percent of student absent in a typical class day – X3 (actual values, maximum is 9%)
Admittance policy: interviewing the parents – X4 (1 for yes, 0 for no)

C. Student’s Self-Concept
Student’s expectation of self – X5 (1= college graduate, 0 = will not finish college)
Usually do well in math – X6 (1 for agree; 0 for disagree)

D. Teaching practices
Students work from worksheets or textbooks on their own- X7 (1 for always/often; 0 for
never or once in a while)
 13

Attachment 2

Codes for variables in Science Multiple Regression Analysis

A. Student’s home background

Presence of computer at home-X1 (coded 1 for yes and 0 for no)
Speak the language of test at home - X2 (1 for always, 0 for sometimes/never)
Parents educational level – X3 (1 for college graduate, 0 for lower than college graduate)

B. School Policy
Percent of students absent in a typical class day – X4 (actual values, maximum is 9%)
Admittance policy: interviewing parents – X5 (1 for solely basing on interview of parents,
0 for otherwise)
Admittance policy: give preference to students with siblings in the same school – X6 (1
for yes, 0 for no)

C. Student’s self-concept
Usually do well in science – X 7 (1 for agree, 0 for disagree)

D. Teaching practices
Students do experiments – X8 (1 for yes, 0 for no)
Teacher uses computer to demonstrate ideas in science – X9 (1 for always/often, 0 for
never)
Teacher explains rules and definitions when beginning a new topic – X10 (1 for always,
0 for never)

E. Others
Type of school community – X11 (1 for urban, 0 for rural)
Gender of students – X12 (1 for female, 0 for male)

Student’s Self-Concept
Student’s expectation of self - X12 (1= college graduate, 0 = will not finish college)