ArticlePDF Available

Abstract and Figures

In determining the level of the Malaysian economy, one of the factors contributing to the economic development of the country is the availability of labour. Studies on the availability of female labour force by identifying female profiles in labor force participation were made using data from Labour Force Survey (LFS), Malaysia for reference year 1990, 2000, 2010 and the latest 2018. Referring to the latest year 2018, the recursive partitioning (RP) technique showed that four subgroup profiles of working women have been created equal to 12%, 43%, 57% and 78%, respectively (percentage of working women at the four subgroups). Majority of the working women are identified in the group of aged 20-59 and the highest certificate obtain was tertiary education (Certificate, Diploma and Bachelor’s Degree) with 78%. Comparison between year 2018 with others reference year, there was a difference in the profile of working women whose working age had increased from 54 to 59 years. There is also a difference in the highest qualification obtained from obtaining only UPSR or SRP qualification in year 1990 to Diploma and Bachelor’s degree certificate in year 2018. Therefore, efforts need to be enhanced by providing initiatives for this group so that those who are in this group or will join this group later will be motivated to join labour market.
Content may be subject to copyright.
226
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
RESEARCH ARTICLE
The Profile of Female Labor Force
Participation in Malaysia Based on
Recursive Partitioning Analyses
Abdul Hadi Alias
a
, Zamira Hasanah Zamzuri
b,
*, Nur Riza Mohd Suradi
a
a Department of Statistics Malaysia; b Department of Mathematical Sciences, Universiti
Kebangsaan Malaysia, Bandar Baru Bangi 43600, Selangor, Malaysia
Abstract
In determining the level of the Malaysian economy, one of the factors contributing to
the economic development of the country is the availability of labour. Studies on the availability of
female labour force by identifying female profiles in labor force participation were made using data
from Labour Force Survey (LFS), Malaysia for reference year 1990, 2000, 2010 and the latest
2018. Referring to the latest year 2018, the recursive partitioning (RP) technique showed that four
subgroup profiles of working women have been created equal to 12%, 43%, 57% and 78%,
respectively (percentage of working women at the four subgroups). Majority of the working women
are identified in the group of aged 20-59 and the highest certificate obtain was tertiary education
(Certificate, Diploma and Bachelor’s Degree) with 78%. Comparison between year 2018 with
others reference year, there was a difference in the profile of working women whose working age
had increased from 54 to 59 years. There is also a difference in the highest qualification obtained
from obtaining only UPSR or SRP qualification in year 1990 to Diploma and Bachelor’s degree
certificate in year 2018. Therefore, efforts need to be enhanced by providing initiatives for this
group so that those who are in this group or will join this group later will be motivated to join labour
market.
Keywords: Labour force participation, Working women, Recursive partitioning
Introduction
In determining the economic status of a country, one of the factors that will contribute to the economic
development of the country is the availability of labour. The structure of labour in a country will determine
whether the country is developed. According to the Department of Statistics, Malaysia (DOSM), the
definition of labour force refers to all persons in the working age who are either employed or unemployed.
The working age defined by DOSM refers to the household members age between 15 to 64 years who
are either in the labour force or outside the labour force (Labour Force Survey Report, Malaysia, 2019a,
DOSM). The United Nations Principles and Recommendations for Population and Housing Censuses
specify that the minimum age limit adopted for the census questions on economic activity should never
be higher than 15 years. The working age limit for a country is determined based on the age structure of
the economically active population of the country (Survey on economically active population,
employment, unemployment and underemployment: An ILO manual on concepts and methods, 1990,
International Labour Organisation). Child and youth recruitment is subject to the Children and Young
Persons (Employment) Act 1966. Act 350 defines children as "under 15" and "young people as underage
18". This labour information will be used as an input in analysing the labour market by the government
in formulating policies and planning, implementing and monitoring programs related to human capital
*For correspondence:
zamira@ukm.edu.my
Received: 9 Nov 2020
Accepted: 19 June 2021
© Copyright Alias. This
article is distributed under
the terms of the Creative
Commons Attribution
License, which permits
unrestricted use and
redistribution provided that
the original author and
source are credited.
227
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
development towards making Malaysia as developed nation in the future.
In 2019, the Malaysian population was 32.58 million people whose population structure is divided into
three age groups namely 14 years and less that is 7.71 million persons (23.3%), age group 15-64 years
that is 22.28 million persons (70.0%) and 65 years and above that is 2.10 million persons (6.7%). The
population distribution by gender indicates that the male population is 16.85 million persons while the
female population is 15.75 million persons (Current Population Estimates, Malaysia, 2019b, DOSM).
Among the female population, the highest population was 25-29 age group (1.54 million persons; 9.8%),
followed by 20-24 age group (1.52 million persons; 9.7%) and 15-19 age group (1.38 million persons;
8.8%). In 2018, the female labour force participation rate (LFPR) is 55.2 per cent, with the prime age
group of 25-34 (73.6%), 35-44 (67.9%) and 45-54 (58.8%) is higher than the national level (Labour Force
Survey Report, Malaysia, 2019a, DOSM). The older age of a woman, the less participation in the labour
force (Rahmah and Noorasiah, 2014).
Equality and balance are important principles to consider in planning the socio-economic development
of a country. Moving forward, more attention is required in order to ensure that all Malaysians have the
opportunity to enjoy a fair economic growth by improving quality and well-being. A study conducted by
Fatima and Sultana, (2009), on the relationship between female labour force participation rate with
economic development in Pakistan found that high economic development rates of 13.0 per cent (1993)
to 43.0 per cent (2002) had driven Pakistan women to participate in the workforce. The economy of
Malaysia in 2018 registered an annual growth in the gross domestic product (GDP) of 4.7 per cent, with
GDP at a constant price of 1361.5 billion as compared to GDP in 1990, at 9.0 per cent annual growth of
106.0 billion. In 2018, Services and Manufacturing sectors served as a major contributor to the Malaysian
economy with 56.7 per cent and 22.4 per cent respectively. The Malaysian female labor force
participation rate also increased from 47.8 percent to 55.2 percent over the same period (DOSM 2016,
2019d, 2019e). Nor Aznin and Norehan, (2010), in the study of women labour force participation in
Malaysia also found that the economic growth in the manufacturing and services sector would increase
women participation in the labour force. Therefore, to ensure that all Malaysians regardless of age,
gender, ethnicity, socioeconomic status and geographical position to enjoy the benefits of growth and
development in the country, the government has continued to focus on the importance of providing the
access to infrastructure, education, training and job opportunities for all Malaysian community.
The availability of highly skilled workforce is needed to support the transition of all economic sectors
towards knowledge-intensive activities to generate labour productivity and attract more investor to invest
in Malaysia. An efficient and inclusive labour market will ensure a balance between supply and labour
demand, and enable Malaysians to participate and benefit from the country's economic growth. This is
because the productivity of workers can be enhanced by the suitable education and training. In 2018,
the labour force in Malaysia is 15.3 million persons and the number has doubled compared to the labour
force in the last three decades (1987), which is only 6.5 million persons. The labour force participation
rate (LFPR) in 2018 was 68.3 per cent, which male account for 80.0 per cent while female was 55.2 per
cent (Labour Force Survey, Time Series Statistics by state, 2019d, DOSM Portal). Although in Malaysia
it is seen that female student enrolment in public universities in 2018 increased by 1.28 per cent
compared to 2000, but the female labour force participation increased only 0.74 per cent during that
period (Social Statistics Bulletin, multi-year). The fact that the increment in the female labour force
participation does not tally with the female student enrolment motivates a further investigation as
conducted in this paper.
Childcare factor is the other element that influencing women involvement in the labour market. Azid et
al., (2010), conducted a survey to 3911 respondents in Punjab, Pakistan and found that Pakistan
married women with large numbers of young children could decreased the possibility of women to join
the force labour (Norehan et al., 2012; Roopnarine and Ramrattan, (2012). In Malaysia, fertility rate
declined from 4.0 (1982) to 1.8 (2018), meanwhile female labour force participation rate improved from
44.5 per cent (1982) to 55.2 per cent (2018) (DOSM 2016b, 2019d). This indicates that there is an
inverse relationship between fertility and female labour force participation. Remarkably, if the woman
had more sons than the daughters, it would lower her chances of going out for business. Whereas if the
228
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
number of daughters more than son, the chances of women to get out for work is higher because they
felt the presence of daughters can help them to manage the home (Azid et al., 2010). This finding is also
supported by another study that mention those family have children aged over 18 age old, that children
can take care of younger siblings while their mother return from work (Norehan et al., 2012). The same
study by Norehan et al., (2012), also found that working women had problems doing their work at home
after getting back from working. They must carry a lot of workload at home after returning from work and
do not have enough time to do all the work at one time including taking care their family. A study
conducted by Haryana Rozana, (2015) with an interview against the 18 Malay women respondents who
are highly educated but not working or has resign, found that among the women who resign from work
because they feel the service facilities of social support is unconvincing in terms of safety which does
not meet certain criteria in terms of child safety and cleanliness of the area as well as the commitment
of childcare workers also raised concerns for women. At the same time, there have been numerous
reports of children being abused at childcare centre and some of that ended with death. This situation is
disturbing for parents as they must work outside to provide their family a good life as the cost of living is
relatively high especially in major cities. According to a report released by Kosmo (2018), during the
period January to May 2018, about 199 cases of abuse and neglect involving children in childcare centre
were reported nationwide. But this finding contradicts with the study conducted by Siah and Lee, (2015),
who used data from labour force participation, fertility and infant mortality rates from 1970-2010 found
that the presence of childcare centre led to a possible positive relationship between fertility and female
labour force participation.
Norehan et al., (2012), state that the high wage rates are among the factors for a married woman to
decide to join the workforce. As women feel that their husbands' income is small and insufficient to cover
their family expenses, women feel they need to work to help their spouses to reduce the burden of family
life. The low labour force participation rate in Malaysia is also due to the lack of support from husbands
and their families (Tan and Geetha, 2013). A study conducted by Tan and Geetha, (2013), found that
almost all female graduates are willing to join the workforce upon graduation, but they will not enter the
workforce if their family or husband forbids them to go out to work.
Sustainable human capital development is an important factor in generating and sustaining the growth
of the Malaysian economy to reach the level as a developed country. To ensure a sustainable source of
labour, the government has provided early access to education. The theory of Becker, (1993), state that
when student enrolment in the education system increases, participation rates in the labour market will
be high. According to the Ministry of Education (MOE), in 2018 the number of boys in primary schools
was 1.38 million compared to 1.31 million girls. While in high school, the number of boy students was
0.97 million and the girl students was 0.98 million. From these statistics, the number of girl students has
begun to surpass the number of boy students. This situation can also be seen in tertiary institutions
where the number of female students at tertiary institutions in 2018 was 746,012 (54.9%) over the
number of male students at 611,632 (45.1%). For higher education institutions such as Institute of
Teacher Education, female students account for 71.2 percent of the students. Similarly, in public
universities where the number of female students recorded 61.3 per cent compared to the male students
at 38.7 per cent only (Social Statistics Bulletin, Malaysia, 2019a, DOSM). Nor Aznin and Norehan, (2010)
found that the number of jobs opportunities is higher for peoples with higher education levels. Education
level also positively influence married women to be in the workforce (Rahmah and Noorasiah, 2014). But
a study by Suhaida and Mohd Faizal, (2014), found that the number of educated and skilled women
workforce was at worrying levels. Statistics show that women's participation in the labor force in Malaysia
is no more than 50 per cent over the last 30 years and was among the lowest among ASEAN countries.
The increase in the number of female graduates is not in line with the number of people who were in the
labour force.
Many previous studies, which wanted to see the relationship between participation in labour force
(response variables) and other estimator variables, have chosen to use logistics techniques in their
analyses (Norehan et al., 2012; Faridi and Rashid, 2014; Nor Amna A'liah and Rusmawati, 2014;
Suhaida and Mohd Faizal, 2014; Desta, 2017). This is because the technique is suitable for a binary
data that have only two levels, for example working or not working (Agresti, 2002). In this paper, we
229
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
apply other techniques derived from data science tools that can fullfill the same purpose as logistic
regression to the binary data. As no previous studies have applied recursive partitioning techniques
toward labour data, this study aimed to use recursive partitioning techniques as an alternative technique
that can be considered in analysing labour participation. According to Roman et al., (2009), recursive
partitioning would have an advantage if the true relationship between the variables and the outcome of
interest is nonlinear. Merkle & Shaffer (2011) stated that among lucrative features of the recursive
partitioning technique is that it does not require any assumption on the data distribution. Cho (2014) and
Gasper et al. (2000) agree that the recursve parttioning technique is easier to understand and interpret.
This proposed technique also has been widely used in other fields such as psychology (Kuroki, 2011;
Scott et al., 2011; Pearson, 2012) and medical (Van Hulst et al., 2015; Mahan et al., 2018).
Materials and methods
Materials
The data used in this study are micro data obtained from the Labour Force Survey (LFS) conducted by
the Department of Statistics, Malaysia (DOSM). A survey conducted covers women aged 15-64 years
(1990 and 2000) and 15 years and above (2010 and 2018) on their details in labour force involvement.
The number of samples for the selected year is as shown in Table.
Table 1. The sample size according to year
Year
Total
Selected sample
size (female only)
Male
Total
2018
170 993
339 258
126 311
50.4%
100.0%
75.1%
2010
199 532
397 449
141 982
50.2%
100.0%
71.7%
2000
109 690
217 869
67 124
50.3%
100.0%
62.0%
1990
67 830
136 219
40 106
49.8%
100.0%
58.6%
LFS comprises active and inactive populations in the economy of 15-64 years of age. The economically
active population consists of those who work and are unemployed, while those who are inactive are
classified as outside labour force. However, the scope of this study will focus on the involvement of
women in the labour force. The demographic characteristics of working and non-working women will be
explored in detail by looking at current patterns compared to those of the past. The demographic and
economic variables that will be used in this study are in Table 2.
Methods
The analytical approach for this study is to use the recursive partitioning (RP) techniques. Recursive
partitioning is used to identify sub-groups of women profiles in terms of their participation in the labour
force using RPART routines with R software. This RP technique has been the basis of two non-
parametric regression methods that is Classification and Regression Trees (CART) and Multivariate
Adaptive Regression Splines (MARS). The use of these RP techniques is becoming more and more
useful due to the use of large data sets with increasing variables. The decision tree turns as a prediction
model that come from a set of predictor variables, represented by trees to give conclusions about the
230
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
value of target variables, represented by nodes used in data mining and machine learning. This RP will
form the basis of a decision which will classify the population by dividing it into sub-populations based
on several independent variables. This process is known as recursive because each sub-population may
be split several times until the process of separation stop with a certain criterion being achieved (Zhang
and Singer, 2010).
Before this RP analysis is prepared, we need to form the basis of the decision. The decision tree is a
possible map of the relevant results. It can be used by someone to consider possible actions against
each other based on their cost, probability and benefits. It can also be used to map algorithms that predict
mathematical best choices. The tree has three layers of nodes. The first layer is called the unique root
node which is the main element in the tree. Then the internal node is in the second layer, followed by the
terminal node. Here, the root node can also be considered as an internal node. Both the root node and
the internal node are divided into two nodes in the next layer called the daughter node (left/right).
Occasionally, a daughter node can also be offspring nodes if the node can still be subdivided. Terminal
nodes are called leaf nodes, but these terminal nodes do not have offspring nodes. Each internal node
represents a filter on each selected attribute, while each branch represents the filter result and each
terminal node holds the class label. The result of these possibilities gives a form like a decision tree.
Let
!"
be the left node and
!#
is the right node,
$%&'
is the sample size in left node for jth class and
%('
is
the sample size inthe right node for the jth class, hence the formula for impurity in the daughter nodes
are given as:
)*!"+, -%&&
%&. /012%&&
%&.3 -%&(
%&. /012%&(
%&.3
)*!#+, -%(&
%(. /012%(&
%(.3 -%&(
%&. /012%((
%(. 3
Then the goodness measure of a split, s is given by:
45*67!+, )*!+- 89!":)*!"+-89!#:)*!#+
where
!
is the parent of
!"
and
!#
, and
;
*
!"
+ and
8
9
!#
:
$
are the probabilities that a subject falls into nodes
!"
and
!#
.
The decision tree is recommended when the data mining task contains a classification result (the class
to which the data belongs). Classification is the task of disseminating known structures to apply to new
data. For example, in this study we want to classify the female labor force as working or not working.
Data classification is a two-step process. The first step is learning, which consists of analyzing training
data by classification algorithms and then after learning, models or classifiers are represented in the form
of classification rules. The second step is classification, which is the test data used to estimate the
accuracy of classification rules.
Results and discussion
Overall finding from recursive partitioning
The results of this study were obtained from Labour Force Survey conducted by the Department of
Statistics, Malaysia for the reference year 1990, 2000, 2010 and 2018. In determining which female
profile will give the most decision on labour participation, the RP technique was used to classify female
who are most likely to work and not work.
The first analysis by using the RP techniques was to use all six selected variables namely age group,
ethnic group, marital status, relationship to the head of household, highest certificates obtain and strata
group. All these variables were demographic variables found in the labour force survey for all selected
reference year. From Figure 1, the percentage of female respondents aged 15-64 years who did not
work in 1990 and 2000 was 55 and 56 per cent, respectively. While the percentage of female
respondents aged 15 years and above in 2010 and 2018 was 60 per cent respectively. This result support
231
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
by the increasing of female life expectancy from 66.0 years in year 1966 to 77.1 years in year 2018
(DOSM, 2017, 2018). For the year 1990 and 2000 data, working details were recorded only for those
who are aged 15-64 years, while for year 2010 and 2018 working details for those aged 15 and above.
Based on Figure 1, it can be seen that working women have produced several subgroups; five subgroups
for year 1990 (Figure 1a), four subgroups for year 2000 (Figure 1b), seven subgroups for 2010 (Figure
1c) and four subgroups for year 2018 (Figure 1d). Using year 1990 as a reference year, five subgroup
profiles were made with the percentage of working women equal to group 1 (27%), group 2 (32%), group
3 (36%), group 4 (69%) and group 5 (72%) respectively. The highest percentage was group 5 which is
72% were working women those aged 20-54 years with their marital status were Other than married and
the highest certificate of UPSR (6 years schooling) and SRP (9 years schooling). As compared to the
latest year 2018, four subgroup profiles of working women have been created equal to group 1 (12%),
group 2 (43%), group 3 (57%) and group 4 (78%), respectively. The highest percentage of 78% were
working women with aged 20-59 and the highest certificate obtain was tertiary education (Certificate,
Diploma and Bachelor’s Degree). Looking at this group in detail (referring year 2018), it is found that
70% of those who are work come from skilled worker category. Most of these groups also work in sectors:
1) Education, 2) Human health and social work activities, and 3) Wholesale and retail trade. For those
who are highly educated but not working, they have a background education such as Social Science,
Business and Laws.
In terms of age, there is a variation in terms of age for entering and to exit from the workforce. In year
1990, 2000 and 2010 (refer to Figures 1a, 1b and 1c) women began to enter the labour force as early
as aged 20 years and left the labour force by the age of 54 years. However, referring to the recent year
of 2018, it is found that women still start working as young as aged 20 years, but the exit period extends
up to aged 59 years. This is also related to the government extension of retirement age from aged 55 to
60 years starting on January 1, 2012 (PSD Portal). Although age are found to have a positive impact on
the decision of married women to work but an increase of one year in the age of the married women will
reduce the probability of working by 4.3% (Rahmah and Noorasiah, 2014).
As for women's certificate factor, it has been changes in employment that in year 1990 women started
working as early as SPM (11 years schooling), but from year 2000, 2010 and 2018 women will start
working after having their Diploma or at least obtaining a post-SPM skills certificate. This indicates that
many women have begun to pursue their study into higher level of education. This finding is in-line with
other research findings, if level of educational was higher, the probability of women go to work was also
higher (Nor Aznin and Norehan, 2010; Rahmah and Noorasiah, 2014).
For the marital status factor, in 1990, 2000 and 2010 there was a pattern in which married women were
more likely to be unemployed. For the three periods mentioned, the variables of marriage were classified
into unmarried and married groups. This supports the findings from several previous studies that showed
that married women are more likely to be unemployed (Azid et al., 2010; Norehan et al., 2012;
Roopnarine and Ramrattan, 2012). However, for year 2018 the variables of marital status cannot be
classified as there are no differences between employment status and the category in marital status.
This indirectly indicates that more marriage women are working in the recent years.
Referring to the findings of year 1990 and 2018, there is a difference in the profile of working women
whose working age increased from 54 to 59 years. There is also a difference in the highest certificate
obtain from having only UPSR or SRP certificate to Diploma and Bachelor’s degree certificate. This may
indicate that the latest trend of working women shows that educated women will stay longer in the labour
market. This showed that the government initiative in providing education to all citizens at the higher
level has been successful. This finding supports the theory from Becker, (1993), which stated that as
student enrolment increases, labour force participation rates will also increase.
232
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
Table 2. List of Selected Variables.
No.
Variable
Abbreviation
Description
Code
i.
Age
U
15 years and above
ii.
Ethnic group
KE
Malay,
Others Bumiputera,
Chinese,
Indian,
Others
1
2
3
4
5
iii.
Marital status
TP
Never married,
Married,
Widowed,
Divorced,
Permanently separated
1
2
3
4
5
iv.
Relationship to the
head of household
PKIS
Head of household,
Spouse to the household head,
Unmarried children of the household head,
Married children of household head,
Son-in-law of household head,
Grandson of household head,
Father/Mother of household head,
Grandfather/Grandmother of household
head,
Siblings of household head,
Members of household head,
Maid,
Other members of household head
1
2
3
4
5
6
7
8
9
10
11
12
v.
Highest certificate
obtained
KS
No certificate (including those who have no
formal education)
UPSR,
PT3/SRP,
SPM,
STPM,
Certificate,
Diploma,
Degree
1
2
3
4
5
6
7
8
vi.
Strata group
ST
Urban,
Rural
1
2
vii.
Age group
KU
15-19 years,
20-24 years,
25-29 years,
30-34 years,
35-39 years,
40-44 years,
45-49 years,
50-54 years,
55-59 years,
60-64 years,
65-69 years,
70-74 years,
75-79 years,
80-84 years,
85 years and above
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
.
233
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
(a)
Female aged 15-64 years, 1990
(c)
Female aged 15 years and above, 2010
Bekerja
0.22
2%
KU = 2,3,4,5,6,7,8
KS = 6,7,8
PKIS = 1,8,9,11,12
KS = 4,5
KU = 3,4,5,6,7,8
TP = 1,3,4,5
Not Working
0.60
100%
Working
0.47
67%
Working
0.20
12%
Working
0.31
8%
Working
0.49
20%
Not Working
0.64
20%
Not Working
0.51
26%
Not Working
0.87
33%
Not Working
0.53
55%
Not Working
0.59
6%
Not Working
0.52
17%
Not Working
0.57
46%
Working
0.22
2%
KS = 1,4,5,7,8
KU = 2,3,4,5,6,7,8
Not Working
0.55
100%
Working
0.37
24%
Working
0.31
20%
Not Working
0.61
76%
Not Working
0.68
4%
Not Working
0.56
56%
Not Working
0.73
21%
Working
0.28
12%
Not Working
0.64
44%
KU = 2,3,4,5,6,7,8
TP = 1,3,4,5
234
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
(b)
Female aged 15-64 years, 2000
(d)
Female aged 15 years and above, 2018
Figure 1. Classification trees obtained from recursive partitioning technique for all selected variables.
KU = 2,3,4,5,6,7,8
TP = 1,3,4
KS = 7,8
Not Working
0.56
100%
Working
0.49
75%
Working
0.30
20%
Working
0.18
4%
Not Working
0.80
25%
Not Working
0.55
55%
Not Working
0.58
51%
235
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
Table 3. Summary of recursive partitioning (RP) by year
Year
Variable
KE
TP
PKIS
KS
ST
2018
2010
√√
2000
1990
Notes:
KE
TP
PKIS
KS
ST
KU
√√
Ethnic groups
Marital status
Relationship to the head of household
Highest certificate obtain
Strata
Age group
Categorized once
Categorized twice
Table 3 explains how many times the selected variable was classified into small group, for example in
year 2018 variable age group was partitioning twice by (1) group aged 20-59 years and (2) group aged
25-49 years, variable highest certificate obtain was partitioning once by (group Certificated, Diploma and
Bachelor’s Degree), and others variables can’t further partitioned into small groups. It can be seen from
the results of this RP analysis that the age factor is the most important factor for each reference year,
because age will breakdown into small age groups. Therefore, in the next sections, we will discuss further
in detail on the age factor and one selected profile group obtained from the recursive partitioning
classification.
Recursive partitioning by age
From the results obtained through recursive partitioning presented at earlier section, age group variables
can be classified in all reference years. Age group variables were classified twice almost all reference
years except 2000, which classified age group variables only one time. Therefore, focus on this age
factor will be discussed in this section.
Based on Figure 2, it can be seen that working women have produced several subgroups; five subgroups
for year 1990 (Figure 2a), four subgroups for year 2000 (Figure 2b), six subgroups for 2010 (Figure 2c)
and five subgroups for year 2018 (Figure 2d). Using the latest year 2018 as a reference year, five
subgroup profiles were made with the percentage of working women equal to group 1 (9%), group 2
(14%), group 3 (41%), group 4 (43%) and group 5 (63%) respectively. The highest percentage was group
5 which is 63 per cent were working women those aged 24-53 years. As compared to the year 1990, five
subgroup profiles of working women have been created equal to group 1 (18%), group 2 (30%), group 3
(36%), group 4 (49%) and group 5 (56%) respectively. The highest percentage of 56 per cent were
working women with aged 19-27 years. Also, from the result, women aged 27-52 years were 53 per cent
from the working age group (15-64 years).
236
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
(a)
Women aged 15-64 years, 1990
(c)
Women aged 15 years and above, 2010
U >= 20
U < 55
Not Working
0.60
100%
Not Working
0.55
86%
Working
0.47
67%
Not Working
0.90
14%
Not Working
0.84
19%
Not Working
0.61
7%
Working
0.45
60%
U >= 23
Not Working
0.51
27%
Working
0.40
33%
U < 39
Working
0.48
16%
Not Working
0.56
11%
U < 48
237
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
(b)
Women aged 15-64 years, 2000
(d)
Women aged 15 years and above, 2018
Figure 2. Classification trees obtained from recursive partitioning technique by age.
U >= 19
U < 51
Not Working
0.56
100%
Not Working
0.52
87%
Working
0.48
73%
Not Working
0.88
13%
Not Working
0.71
14%
Not Working
0.60
3%
Working
0.48
70%
U >= 20
U < 60
U >= 20
Not Working
0.56
100%
Working
0.49
82%
Working
0.42
71%
Not Working
0.86
18%
Not Working
0.91
12%
Not Working
0.59
9%
Working
0.40
62%
U < 54
Not Working
0.57
8%
Working
0.37
53%
U >= 24
238
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
Computing the misclassification rate
In order to check the reliability of the findings, we have computed the misclassification rate from
simulations conducted on the recursive partitioning (RP) technique employed to the data set. Table 4
depicts the summary on the average error rate, standard deviation and computational time for all four
data sets.
Table 4. Summary of the simulations
Year
Real
Sample
Time (mm:ss)
Misclassification
Rate
Standard Deviation
2018
126 311
00:27.0
30.4318
0.6554
2010
141 982
00:24.1
28.7932
0.6301
2000
67 124
00:26.0
32.4748
0.5054
1990
40 106
00:25.6
32.0684
0.5919
The results show that the misclassification rate for all years is around 28 to 32 per cent with a standard
deviation in the range of 0.5 to 0.6. With a sample number of 10 000 and number of iterations is 50, we
found that the prediction results of the recursive partitioning technique have an average misclassification
rate, around 30 per cent.
Testing for one selected group
From the previous studies (Azid et al., 2010; Norehan et al., 2012; Roopnarine and Ramrattan, 2012),
child factors play a role in women's work status. An in-depth analysis will be performed to see the
relationship between childcare factor and profile groups derived from the recursive partitioning in section
3.1. However, due to the constraints of the existing individual information that there was no information
on the number of children and need to be replicated from another variable “Relationship to the head of
household”, so only one profile group will be tested. The selected profile group is the group with the
highest percentage, 78% in the year 2018. This group consists of married women aged 20-59 years and
having "Certificate", "Diploma" and " Bachelor’s Degree" in the household.
Since both variables are categorical, we conduct a chi-square test to investigate the association between
number of children stay at the same household and working status. Based on Table, since the p-value
is smaller than 0.05, we conclude that there is an association between the two variables for the group of
married women aged 20-59 years and having "Certificate", "Diploma" and " Bachelor’s Degree" in the
household. This finding supported the previous study that mention those family have children aged over
18 age old, that children can take care of younger siblings while their mother return from work (Norehan
et al., 2012).
Table 5. Chi-Square test for association between number of children and working status
Value
df
p-value
Pearson Chi-Square
159.841
6
0.000
Figure 3 describes the distribution of number of children in the selected profile group. It can be observed
that in the households with two children and below, the percentage of married women in the household
who are not working exceeded the number of working women. In contrast, married women with three
children and more are going out for work may be due to two factors. First, based on a large number of
household members (five people and above); this mother needs to help their family economy by going
out for work. The second factor is that families with a large number of children are more likely to have
children age over 18 years old, and those children are able to take care of other younger siblings when
239
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
their mothers are out for work, as noted by Norehan et al. (2012). According to Noor Rahamah (2012),
many job opportunities, as well as the improving level of women's education, have resulted in women's
involvement in wide-open employment. Women are no longer considered housewives only but also
contributors to the socio-economic development of the country. However, women’s involvement in
employment does not eliminate their traditional roles as wives and mothers at home.
Figure 3. Percentage of employed status by number of children.
Conclusions
In the government efforts to make Malaysia a developed nation, the country economy needs to be
improved with the support of available resources. Malaysia working age (15-64 year) population in 2019
was 22.8 million people (male: 11.9 million, female: 10.9 million). Therefore, by utilizing this working age
population, it can greatly assist Malaysia in becoming a developed country. By attracting female labour
force participation especially from full-time housewives, attention should be given to specific age groups.
The latest trend of the year 2018 showed that, even though women began to enter the labour market as
early as the age of 20 years and leave the labour market by the age of 59 years, the most working age
group was 24-53 years. In year 2018, it was found that the variables of marital status were not classified
because there was no difference between employment status with the category in marital status. This
showed that nowadays many women who are married are working. This indirectly provides inputs on the
effectiveness of government incentives for returning women such as access to childcare centres, grant
and training, and individual income tax exemptions. These incentives need to be pursued and enhanced
in the future so that many women will be interested to join labour market. From this latest trend, it can
also be seen that the percentage of women with higher academic education has engaged in the labour
market. Looking at this group in detail, it is found that 70% of this group come from the skilled worker
category. Most of these groups also work in the sectors: 1) Education, 2) Human health and social work
activities, and 3) Wholesale and retail trade. This is verified by an increase in the contribution of the
services sector in Gross Domestic Product (GDP) from 46.8 per cent in year 1990 to 56.7 per cent in
year 2018 (DOSM, 2016, 2019e). This finding also support the study by Nor Aznin and Norehan, (2010),
Malaysia economic growth in the manufacturing and services sector would increase women participation
in the labour force. Therefore, the government should have a policy that is always open to new job
opportunities in the employment of skilled workers to cater for the arrival of women graduates so that the
knowledge they gain while pursuing higher education can be utilized as much as possible which will be
in line with the government efforts to make Malaysia more successful country. This is because we want
Malaysia to be a developed country with a lot of skilled labour as opposed to a lot of low skilled labour.
The compulsory retirement age also needs to be revisited so that extending the retirement period to 60
years or above will not affect youth employment opportunities. It is suggested that the age of this
compulsory retirement varies according to the level of occupation skills that are skilled, semi-skilled and
low skilled. Therefore, efforts need to be enhanced by targeting initiatives of this age group so that those
who are still unemployed will be motivated to start working.
240
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
Acknowledgments
The rst author thankfully acknowledges financial support from the Public Service Department (PSD) for
their trust in giving the sponsorship of “Hadiah Latihan Persekutuan (HLP)”. The authors thank to
Department of Statistics, Malaysia (DOSM) for the data use in this paper. The authors take responsibility
for the integrity of the data and the accuracy of the data analysis. The authors also thank to an
anonymous reviewer and the editor for their helpful comments.
References
Agresti, A. 2002. Categorical Data Analysis. Edisi kedua. New Jersey: John Wiley & Sons, Inc.
Azid, T., Khan, R. E. A., Alamasi, A. M. S. 2010. Labor force participation of married women in Punjab (Pakistan).
International Journal of Social Economics, Vol. 37 Iss 8 pp. 592612
Becker, G. S. 1993. Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education. Edisi
ketiga. United States: The National Bureau of Economic Research.
Cho, A. R. 2014. Comparisons of recursive partitioning analysis and conventional methods for selection of uncuffed
endotracheal tubes for pediatric patients. Pediatric Anesthesia 25, 698-704
DOSM. 2016. Malaysia Economic Statistics-Time Series 2016-National Account. https://www.dosm.gov.my/ [19
December 2017]
DOSM. 2017. Malaysia Economic Statistics-Time Series 2016-Population. https://www.dosm.gov.my/ [19 December
2017]
DOSM. 2018. Statistics-Malaysia @ a Glance-Malaysia. https://www.dosm.gov.my/ [14 November 2018]
DOSM. 2019a. Social Statistics Bulletin, Malaysia, 2019.
DOSM. 2019b. Current Population Estimates, Malaysia, 2018.
DOSM. 2019c. Labour Force Survey Report, Malaysia, 2018.
DOSM. 2019d. Labour Force Survey (LFS), Time Series Statistics by State, 19822018. https://www.dosm.gov.my/
[21 Jun 2019]
DOSM. 2019e. National Account, Gross Domestics Product, 2015-2018.
Desta, C. G. 2017. Do young children prohibit mothers from working? A study in the Amhara Region, Ethiopia.
International Journal of Population Studies 2017, Volume 3, Issue 2.
Faridi, M. Z., Rashid, A. 2014. The Correlates of Educated Women's Labor Force Participation in Pakistan: A
Micro-Study. The Lahore Journal of Economics 19:2 (Winter 2014): pp.155184.
Fatima, A., Sultana, H. 2009. Tracing out the U-shape relationship between female labor force participation rate and
economic development for Pakistan. International Journal of Social Economics Vol. 36 Nos 1/2, 2009pp. 182-198
Gasper, L. E., Scott,C., Murray, K. & Curran K. 2000. Validation of the Rtog Recursive Partitioning Analysis (RPA)
Classification for Brain Metastases. International Journal of Radiation Oncology Biol. Phys. 47 (4), 1001-1006.
Haryani Rozana, A. R. 2015. Dilema wanita berpendidikan tinggi: Faktor-faktor wanita berhenti kerja. SARJANA Vol.
30, No.1, Jun 2015, pp. 99-117.
ILO. 1990. Survey on economically active population, employment, unemployment and underemployment: An ILO
manual on concepts and methods, International Labour Organisation. First published 1990. Geneva, International
Labour Office.
Kuroki, Y. 2012. Recursive Partitioning Analysis of Lifetime Suicidal Behaviors in Asian Americans. Asian American
Journal of Psychology 2012, Vol. 3, No. 1, 1728
Mahan, V. L., Gupta, M., Aronoff, S., Bruni, D., Stevens, R. M., Moulick A. 2018. VVR Score Predicts Intensive Care
Unit Length of Stay in Patients Undergoing Re-entry Sternotomy. World Journal of Cardiovascular Surgery, 8, 7-
21. doi: 10.4236/wjcs.2018.81002.
Malaysia. 2011. Children And Young Persons (Employment) Act 1966 (Revised 2011). (Act 350).
Merkle, E.C. & Shaffer, V. A. 2011. Binary Recursive Partitioning: Background, methods and application to
psychology. British Journal of Mathematical and Statistical Psychology 64, 161-181
Noor Rahamah, A. B. 2012. Wanita bekerja dan pengurusan keluarga. Malaysia Journal of Society and Space 8 Issue
7 (155 162).
Norehan, A., Rahmah, I., Zulridah, M. N., Fariza, A. 2012. Kebarangkalian Bekerja Wanita Berkahwin di Malaysia.
Jurnal Ekonomi Malaysia 46(1) (2012) 107-117.
Nor Amna A'liah, M. N., Rusmawati, S. 2014. Malaysia's Labour Force Participation in Rural and Urban Areas. Asian
Economic and Financial Review, 2014, 4(10): 1461-1472.
Nor Aznin, A. B., Norehan, A. 2010. Labour Force Participation of Women in Malaysia. Jurnal Pembangunan Sosial
Jilid 13 (Jun) 2010: 115130
Online KOSMO. 2018. https://www.kosmo.com.my/ [28 Ogos 2018]
Pearson, M. R. 2012. Pathways to early coital debut for adolescent girls: A recursive partitioning analysis. Journal of
Sex Research, 49(1), 1326, 2012
PSD Portal. http://www.jpapencen.gov.my [1 Januari 2020]
Rahmah, I., Noorasiah, S. 2014. Married Women Labor Supply Decision in Malaysia. Asian Social Science Vol. 10,
No. 3; 2014
Roman, R., Charles, C., Benedicte, L., Chafika, M., Roberto I., Rene, N., Lajos, P. 2009. Direct comparison of logistic
and recursive partitioning to predict chemotherapy response of breast cancer based on clinical pathological
variables. Breast Cancer Res Treat. 117:325331.
Roopnarine, K. A., Ramrattan, D. 2012. Female labour force participation: the case of Trinidad and Tobago. World
Journal of Entrepreneurship, Management and Sustainable Development, Vol. 8 Iss 2/3 pp. 183193
241
Alias et al. | Malaysian Journal of Fundamental and Applied Sciences, Vol. 17 (2021) 226-241
Scott, S. B., Jackson, B. R., Bergeman, C. S. 2011. What contributes to perceived stress in later life-A recursive
partitioning approach. Psychology and Aging 2011. Vol, 26. No, 4. 830-843
Siah, A. K. L., Lee, G. H. Y. 2015. Female Labour Force Participation, Infant Mortality and Fertility in Malaysia. Journal
of the Asia Pacific Economy, 20(4), 613-629
Suhaida, M. A., Mohd Faizal P. R. 2014. Fenomena Kekurangan Tenaga Kerja Wanita Berpendidikan dan
Berkemahiran di Malaysia. Prosiding PERKEM ke 9 (2014) 269-277
Tan, P. L., Geetha, S. 2013. Perception of undergraduates towards female labour force participation. Procedia - Social
and Behavioral Sciences 105 (2013) 383390.
Van Hulst, A., Roy-Gagnon, M., Gauvin, L., Kestens, Y., Henderson, M., Barnett, T. A. 2015. Identifying risk profiles
for childhood obesity using recursive partitioning based on individual, familial, and neighborhood environment
factors. International Journal of Behavioral Nutrition and Physical Activity (2015) 12:17.
Zhang, H., Singer, B.H. 2010. Recursive Partitioning and Applications. Edisi kedua. New York: Springer.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Theoretical work relating economic effect of children suggests that labor market participation decreases for mothers with large number of young children and increases when children are adults. The majority of empirical studies find results consistent with this expectation, but there are some studies which fail to confirm this theoretical prediction for the developing countries. This paper used data from a household survey of rural and urban married women to test the theoretical prediction that labor market participation decreases for mothers with large number of young children and increases when children are adults. Results show that when all households are considered, children seem to have positive effects on the probability of the mother's work participation. However, when household lifecycle and rural-urban location differences are considered, coefficients are negative (but not statistically insignificant) for urban households with large number of young children and positive (and statistically significant) for those households with more adult children; whereas for rural households, these coefficient signs are reversed. Results from the quantitative data combined with qualitative narratives suggest that large numbers of young children do not prohibit rural mothers from working.
Article
Full-text available
This study attempts to determine the factors that affect educated women’s decision to participate in the labor force. Based on a field survey conducted in the district of Multan, we find that a number of factors have a positive and significant impact on women’s decision to work. These include women who fall in the age groups 35–44 and 45–54, the coefficients of all levels of education, the presence of an educated husband, marital status, family structure, and family expenditure. The presence of an educated father, being an educated married woman, location, distance from the district headquarters, the husband’s employment status and income, and ownership of assets significantly reduces women’s labor force participation. The results of the earnings equation show that variables such as women who live in an urban area and their level of education and experience are associated with a substantial increase in earnings with each additional year. The number of children has a negative and significant impact on women’s earnings. The hours-of-work model shows that age and the number of completed years of education have a positive effect on working hours, while the number of dependents and the number of hours spent on household activities have a negative effect on working hours.
Article
Full-text available
Few studies consider how risk factors within multiple levels of influence operate synergistically to determine childhood obesity. We used recursive partitioning analysis to identify unique combinations of individual, familial, and neighborhood factors that best predict obesity in children, and tested whether these predict 2-year changes in body mass index (BMI). Data were collected in 2005-2008 and in 2008-2011 for 512 Quebec youth (8-10 years at baseline) with a history of parental obesity (QUALITY study). CDC age- and sex-specific BMI percentiles were computed and children were considered obese if their BMI was ≥95(th) percentile. Individual (physical activity and sugar-sweetened beverage intake), familial (household socioeconomic status and measures of parental obesity including both BMI and waist circumference), and neighborhood (disadvantage, prestige, and presence of parks, convenience stores, and fast food restaurants) factors were examined. Recursive partitioning, a method that generates a classification tree predicting obesity based on combined exposure to a series of variables, was used. Associations between resulting varying risk group membership and BMI percentile at baseline and 2-year follow up were examined using linear regression. Recursive partitioning yielded 7 subgroups with a prevalence of obesity equal to 8%, 11%, 26%, 28%, 41%, 60%, and 63%, respectively. The 2 highest risk subgroups comprised i) children not meeting physical activity guidelines, with at least one BMI-defined obese parent and 2 abdominally obese parents, living in disadvantaged neighborhoods without parks and, ii) children with these characteristics, except with access to ≥1 park and with access to ≥1 convenience store. Group membership was strongly associated with BMI at baseline, but did not systematically predict change in BMI. Findings support the notion that obesity is predicted by multiple factors in different settings and provide some indications of potentially obesogenic environments. Alternate group definitions as well as longer duration of follow up should be investigated to predict change in obesity.
Article
Full-text available
In modern living, the participation of female in the labor market is becoming essential for development economy. The higher educational attainment among females makes it easier for them to find jobs and to be involved in the labor market. Nevertheless, the participation of women in the labor market is less prevalent than for men, especially for married women, where family responsibilities and household chores become obstacles for them. This paper attempts to identify the determinants of married women's participation in the labor market based on 3,520 data collected in 2011 through a field survey. The results from this study show that educational attainment, women's age and number of children are major determinants of the supply of married women labor. In contrast, husbands' wage and own wage are insignificantly determined the supply of married women labor in this study.
Article
This article analyses the women labor force participation in Malaysia. The rapid absorption of women into the labor market has been influenced by several factors. The rapid economic growth was due largely to important growth in the manufacturing and services sectors, where substantial and proportionally larger increase of female workers has been registered. Among all sectors of the economy, the manufacturing sector has recorded the highest growth rate during the last decade. The rising in the female labor force participation may also be attributable to improving economic incentives in employment and policies favoring the employment of women. In addition, the combined effects of the increase in years of schooling, access to family planning services, improved maternal and availability of child care, leading to arise in the average age at marriage, have allowed women to take advantage of the increased employment opportunities.
Article
The authors are Economists in the Research and Policy Department of the Central Bank of Trinidad and Tobago. The views expressed are those of the authors and not necessarily those of the Central Bank. The authors would like to thank the following persons for their support and technical comments: Dr Alvin Hilaire, Dr Reshma Mahabir, Ms Angela Henry, Mrs Tanisha Mitchell‐Ryan and Ms Rekha Sookraj, all Economists at the Central Bank of Trinidad and Tobago. The authors would also like to acknowledge and thank Mr Sterling Chadee from the Central Statistical Office and Dr Godfrey St Bernard from the University of the West Indies, St Augustine Campus. In addition, the authors are grateful to all those who were present and provided suggestions during the Research Department's Weekly Discussion Series held on February 16, 2011.