Factors17(2):

Mal J Nutr Associated

175 -with
187,Physical
2011 Activity Levels among Adolescents Attending School 175

Determination of Factors Associated with Physical Activity

Levels among Adolescents Attending School in Kuantan,
Malaysia
Dan SP, Mohd Nasir MT & Zalilah MS

Department of Nutrition and Dietetics, Faculty of Medicine and Health Sciences, Universiti Putra
Malaysia, 43400 UPM, Selangor, Malaysia

ABSTRACT

Introduction: Findings from the National Health and Morbidity Survey III (MOH,
2008) indicate a 43.7% prevalence of physical inactivity among Malaysian adults.
This sedentary lifestyle can also be observed among children and adolescents.
Methodology: A cross-sectional study was conducted to determine factors
associated with physical activity levels of four hundred, 13 year-old adolescents
in Kuantan, Pahang. Data on socio-demographic, health-related, and psychosocial
factors were collected using a self-administered questionnaire while physical
activity level was assessed using the Physical Activity Questionnaire for Older
Children (PAQ-C). Results: About one-third of the respondents were in the low
physical activity level category, 61.5% were in the moderate category and only
3.0% of the adolescents were in the high physical activity level category. Males
were more physically active than females (χ2=23.667, p=0.0001) with female
adolescents (45.1%) twice as likely as male adolescents (22.1%) to be in the low
physical activity level category. The associations between physical activity
level with socio-demographic and health-related factors, perception of weight
status and body parts satisfaction were not significant. However, physical activity
was found to be positively correlated with physical activity self-efficacy (r=0.496,
p=0.0001), peer influence (r=0.468, p=0.0001), family influence (r=0.298, p=0.0001)
and beliefs in physical activity outcomes (r=0.207, p=0.0001). Negative
relationships were found between physical activity with depression (r=-0.116,
p=0.021) and body size discrepancy (r=-0.143, p<0.01). Respondents who had a
better perception of their current health status were more physically active
(χ2=21.062, p=0.0001). Multivariate analyses for the prediction of physical activity
showed that physical activity self-efficacy, sex and peer influence were the most
significant contributors in explaining physical activity among adolescents.
Conclusion: Physical activity interventions should include psychosocial
components as mediator variables in interventions designed to promote regular
physical activity in adolescence.

Keywords: Adolescents, physical activity, psychosocial factors, health-related

factors and socio- demographic factors

INTRODUCTION as it exerts many health benefits and has

protective effects on several chronic diseases
Engaging in physical activity during as well as premature mortality (Lahti-Koski
childhood and adolescence is an imperative et al., 2002). However, many studies have

* Correspondence author: Dr Mohd Nasir Mohd Taib; nasir@[Link] / [Link]@[Link]

176 Dan SP, Mohd Nasir MT & Zalilah MS

shown marked declines in physical activity national schools in Kuantan, Pahang. These
during adolescence (Jago et al., 2005; Pahkala schools were randomly selected since they
et al., 2006). The US National Youth Risk were all multi-ethnic and co-educational in
Behaviour Survey indicates that there was a their student composition. All Form 1
dramatic decline of youth participating in students in the two schools were invited to
daily physical educational classes from participate in the study with the final sample
42.0% in 1991 to 27.0% in 1997 (CDC, 1997a). consisting of 400 adolescents.
Similarly, a high percentage of low physical This study was approved by the Medical
activity level was also observed in Malaysia. Research Ethical Committee, Faculty of
Findings from the National Health and Medicine and Health Sciences, Universiti
Morbidity Survey III (MOH, 2008) indicate
Putra Malaysia and the Ministry of
the prevalence of physical inactivity among
Education. Permission was obtained from the
Malaysian adults to be 43.7%. This
sedentary lifestyle can also be observed school principals before the inception of the
among children and adolescents. Lim (2005) study and a written informed consent was
reported that of 75 adolescents studied, obtained from the respondents prior to their
about 44% were in the sedentary category. involvement.
Sedentary behaviour often originates in
childhood and adolescence, may persist into Measurements
adulthood, and may lead to many adult
Data collection was carried out using a self-
chronic diseases (Monge-Rojas et al., 2002).
Adolescence is a particularly unique period administered questionnaire. Anthropometric
in life since it is a time of intense physical, measurements of height and weight were
psychosocial, and cognitive developmental assessed by the researchers and all
changes (WHO, 2000). Each of these measurements were carried out in the school
changes influence adolescent participation setting.
in physical activity (McCabe & Ricciardelli, The questionnaire used for data
2004). collection was translated into Malay.
As numerous studies have shown that Forward translation from the original English
there are dramatic declines in physical version was performed by a university
activity during adolescence (Jago et al., 2005; professor and a university student who were
Pahkala et al., 2006), young people at this both fluent in the Malay language. Back
stage of life are an important target group to translation was also performed by the
focus on. To optimally promote physical university professor and the final version of
activity, there is a need to study and
the questionnaire was derived from a
understand factors or predictors that affect
physical activity levels of adolescents. reconciliation of the original, back and
Physical activity levels can be influenced by forward translations. A pre-test of the
a broad range of factors including biological, questionnaire among thirty-six Form 1
socio-demographic, health-related factors as students was carried out prior to the
well as psychosocial factors (Schmitz et al., scheduled data collection. This was to
2002; Aarnio et al., 2002). Thus, the objective determine whether the students understood
of this study was to determine the factors the items and the questions in the
that predict physical activity levels of questionnaire and to measure the reliability
adolescents. (internal consistency) of the scales in the
questionnaire.
METHODOLOGY
Assessment of physical activity
Study population
The Physical Activity Questionnaire for
This was a cross-sectional study and Older Children (PAQ-C), which was
subjects were Form 1 students from two modified and adapted from Kowalski,
 Factors Associated with Physical Activity Levels among Adolescents Attending School 177

Crocker & Donen (2004) and Ernst & Anthropometric measurements

Pangrazi (1999) was used to assess general
Weight and height of the adolescents were
levels of physical activity of the adolescents.
measured using a TANITA digital weighing
The physical activity items listed out in the
scale (Tanita, Japan) and body meter (Seca,
questionnaire were modified according to
Germany), respectively. Each measurement
the physical activities proposed in the
was obtained twice with the weight
school co-curriculum. The instrument
recorded to the nearest 0.1 kg and the height
consisted of 10 items. Item 1 was on leisure
to the nearest 0.1 cm. Body Mass Index (BMI)
time activity during the last 7 days. The
in kg/m2 was calculated and categorised
adolescents were required to respond to an according to BMI-for-age reference (WHO,
activity checklist and were scored on a 5- 1995).
point scale ranging from ‘no’ activity being
scored as 1 and ‘7 times or more’ being scored Socio-demographic background
as 5. The mean score of all activities on the
activity checklist was calculated to form a This section consisted of four questions that
composite score for item 1. were used to assess the socio-demographic
Items 2 to 8 related to activities of the background of respondents, which included
adolescents during physical education (PE) the date of birth, sex, ethnicity and
class, recess, lunch, immediately after school, household size. Besides, this section also
evenings, weekends, and leisure times. The included questions on respondents’ parents
answer options for each item began from the such as occupation, total years of schooling,
lowest activity response (score 1) and educational attainment and monthly
progressed to the highest activity response income.
(score 5). For item 9, the adolescents were
asked on the frequency of participating in Health-Related Factors
daily physical activity in the previous week. Eating Behaviours
A score of 1 was assigned to ‘none’ and a
score of 5 was assigned to ‘very often’. The The Eating Attitude Test (EAT-26), which
mean physical activity score for all days of was adapted and modified from Garner &
the week was calculated to form a composite Garfinkel (1979) consisted of 26 items and
score for item 9. Finally, item 10 was used to was used as a screening tool to measure
identify adolescents with unusual activities symptoms and concerned characteristics of
during the previous week. eating disorders.
Once the sum of the scores from items 1 An individual with a total score greater
to 9 was calculated, the final PAQ-C activity than or equal to 20 points was categorised
summary score could be obtained by taking as being prone to eating disorders.
the mean of these 9 items. All the adolescents Conversely, an individual with a total score
were classified into three categories (low, of less than 20 points was categorised as
moderate and high) based on their mean total having no symptoms of eating disorders.
physical activity score (Kowalski et al., 2004). The internal consistency for this measure
An adolescent with a mean total score was high (Cronbach’s alpha = 0.93).
ranging from 1 to 2.33 was categorised as
having low physical activity level, 2.34 to Psychosocial Factors
3.66 as moderate physical activity level, and This section was divided into 6 parts as
3.67 to 5 as high physical activity level. The follows:
internal consistency (Cronbach’s alpha) of
1. Perception of current health status
this assessment, calculated from an
This section recorded the respondent’s
independent sample of 36 Form 1 students,
perception of current health status. Five
was 0.79.
178 Dan SP, Mohd Nasir MT & Zalilah MS

choices were given and the respondent a total of 20 items which provided a
was required to choose only one answer score ranging from 0 to 60. A high score
that best represented the current health indicates more depressive symptoms.
status. The internal consistency (Cronbach’s
2. Physical Activity Self-Efficacy alpha) of this scale was 0.83.
This scale which was adapted from 6. Body Image
Saunders et al. (1997) consisted of 17 a. Perception of body size (shape)
items to assess perceived self-efficacy or The respondent was required to
confidence in one’s ability to be choose figures from the ‘Contour
physically active. The respondent was Drawing Rating Scale’ (Thompson
required to respond based on a 5-point & Gray, 1994) to represent his or her
scale from ‘not at all confident’ to perceived current body size, ideal
‘extremely confident’ for each item. A body size and healthiest body size.
high score indicates a high self-efficacy A discrepancy score was deter-
for physical activity. The internal mined based on the difference
consistency (Cronbach’s alpha) of this between perceived current body size
scale was 0.88. and perceived ideal body size.
3. Social Influences for Physical Activity b. Perception of body weight status
This scale for physical activity was This section assessed the correctness
adapted from Saunders et al. (1997). It of the respondents’ perception of
consisted of 8 items to assess peer their current weight status (Koay,
influence and family influence on 1998; Simko, Cowell & Hreha,
physical activity, respectively. The 1989). A respondent’s perceived
internal consistency of this scale was current body weight status was
0.87. Further, the internal consistencies compared with his or her actual
for peer influence (Cronbach’s current body weight status (Body
alpha=0.87) and family influence Mass Index). The respondent was
(Cronbach’s alpha=0.82) subscales were classified into one of the three
also calculated separately for the categories which were correct-
purpose of the study. estimator, over-estimator or under-
estimator.
4. Beliefs in Physical Activity Outcomes
A total of 16 items adapted from c. Body parts satisfaction
Saunders et al. (1997) were used to assess There were a total of 10 main body
beliefs in physical activity outcomes. parts in this section, which was
The respondent was required to answer adapted and modified from Koay
each of the items on a 5-point scale. A (1998). The respondent was re-
high score indicates a strong positive quired to indicate level of
belief in the outcomes of physical satisfaction with each of the ten
activity. The internal consistency body parts listed. Respondents were
(Cronbach’s alpha) of this scale was classified into either one of the two
0.67. categories which were satisfied (>
65) and dissatisfied (<65) based on
5. Depression
The Center for Epidemiologic Studies- the total score.
Depression Scale (CES-D) (Radloff, 1977)
Statistical Analysis
was used to measure self-reported
symptoms associated with depression Data were analysed using the SPSS for
experienced in the past week. There was Windows version 14.0. Descriptive statistics
 Factors Associated with Physical Activity Levels among Adolescents Attending School 179

Table 1. Characteristics of the subjects

Characteristics n (%)

Sex
Male 167 (41.8)
Female 233 (58.2)
Ethnicity
Malay 225 (56.2)
Chinese 168 (42.0)
Indian 7 (1.8)
Body Mass Index (kg/m2) (Mean+SD) 20.18+4.61
Underweight 29 (7.3)
Normal Weight 280 (70.0)
At risk of overweight 91 (22.7)
Parents’ Years of Education
Father (Mean+SD) 12.91+4.07
Mother (Mean+SD) 12.29+3.39
Total Household Income (Mean+SD) 3238.12+2740.55
< RM 500 6 (1.6)
RM 501-RM 1000 112 (28.0)
RM 1001-RM 2000 79 (19.7)
RM 2001-RM 3000 49 (12.2)
> RM 3000 154 (38.5)
Household Size (Mean+SD) 5.74 + 1.60
<5 84 (21.0)
5-7 260 (65.0)
>8 56 (14.0)
Physical Activity Level (Mean+SD) 2.56+0.57
Low 142 (35.5)
Moderate 246 (61.5)
High 12 (3.0)

such as frequencies, means, standard categorical variables. Pearson Product-

deviations, and percentages were used to Moment Correlation was used to determine
describe variables such as physical activity, the relationship between two continuous
sex, ethnicity, body mass index, household variables. Multiple Linear Regression was
income, household size and parents’ used to measure the amount of influence an
education level. Normality test was independent variable had on a dependent
conducted using graphical methods (e.g. variable. A statistical probability level of p<
histogram, Q-Q plot), descriptive statistic 0.05 was considered as significant.
and frequentist test (e.g. Shapiro–Wilk test,
Kolmogorov-Smirnov Test). RESULTS
Cronbach’s alpha was used to assess
internal consistency of each scale in the Table 1 shows that a total of 400 adolescents
questionnaire. The Chi-square test was used (167 males and 233 females) with a mean
to determine associations between age of 13.23+ 0.31 years participated in the
180 Dan SP, Mohd Nasir MT & Zalilah MS

present study. They comprised 225 (56.2%) respondents who were prone to eating
Malay, 168 (42.0%) Chinese, and 7(1.8%) disorders (42.1%) were in the low physical
Indian adolescents. The distribution of activity level as compared to respondents
respondents by physical activity level who had no eating disorder symptoms
indicated that a majority of the respondents (34.4%). However, no significant associa-
(61.5%) were in the moderate physical tion was found between physical activity
activity level category while only 3.0% of the level and EAT-26 score category (χ2=1.267,
respondents were in the high physical p>0.05).
activity level category. Moreover, about one-
third of the respondents (35.5%) were in the Physical activity and psychosocial factors
low physical activity level category. Table 2 also shows the distribution of
respondents according to perception of
Physical activity and socio-demographic current health status for physical activity
factors level. There was a significant association
Table 2 shows males to be more physically between physical activity level and
active than females (χ2= 2.296, p<0.001). perception of current health status
Most of the male (77.9%) and female (54.9%) (χ2=15.399, p<0.001) (Table 2). In other
respondents were in the moderate/high words, respondents who have a better
physical activity level. However, there were perception of their current health status were
twice as many female adolescents (45.1%) more physically active.
who were in the low physical activity level For physical activity self-efficacy, there
compared to male adolescents (22.1%). was a positive and moderate linear
In the analysis of physical activity levels relationship between physical activity self-
by ethnicity, Indians were excluded since efficacy and physical activity scores (r=0.496,
they comprised only a small proportion of p<0.001) (Table 2) based on Guilford’s Rule
the respondents. Results showed that there of Thumb (Hinkle, Wiersma &Jurs, 1981). In
was no association between physical other words, as physical activity self-efficacy
activity and ethnicity (χ2=1.550, p>0.05). increased, physical activity also increased.
Additionally, no association was found As shown in Table 2, there was a positive
between socio-demographic factors and and moderate linear relationship between
physical activity except for father’s total peer influence and physical activity score
years of schooling. There was a positive and (r=0.468, p<0.001). On the other hand, a
weak association between father’s total years positive and weak relationship was found
of schooling and physical activity score (r = between family influence and physical
0.105, p = 0.036). activity score (r=0.298, p<0.001). Most of the
respondents strongly agree that if they were
Physical activity and health-related physically active most days, it would help
behaviour them to be healthy (65.5%), control their
weight (44.5%) and give them energy
As shown in Table 2, the distribution of
respondents with low and moderate/high (53.7%). In contrast, about 50.0% of the
respondents strongly disagree that being
physical activity levels was almost similar
physically active would be boring, may hurt
whether they were underweight, normal
weight or at risk of overweight. No them physically and may embarrass them
in front of others. As a whole, most of the
association was found between physical
respondents agreed that being physically
activity and weight status (χ2=0.023, p>0.05).
For eating behaviours, 14.3% were active would give them positive outcomes.
Further, there was a positive and weak
categorised as prone to eating disorders. A
correlation between beliefs in physical
close inspection of Table 2 reveals that more
 Factors Associated with Physical Activity Levels among Adolescents Attending School 181

Table 2. Bivariate correlation between physical activity and socio-demographic factors, health
related factors & psychosocial factors

Factors Physical activity Level χ2 r value

Low Moderate/High
n(%) n(%)

Socio-demographic Factors
Sex 2.296***
Male (n=167) 37 (22.1) 130 (77.9)
Female (n=233) 105 (45.1) 128 (54.9)
Ethnicity 1.550
Malay (n=225) 86 (38.2) 139 (61.8)
Chinese (n=168) 54 (32.1) 114 (67.9)
Father’s total years of schooling 0.105*
Mother’s total years of schooling 0.08
Household income 0.08
Household size 0.03
Health Related Factors
Body Mass Index 0.023 -0.043
Underweight (n=29) 10 (34.5) 19 (65.5)
Normal Weight (n=280) 100 (35.7) 180 (64.3)
At risk of overweight (n=91) 32 (35.2) 59 (64.8)
Eating Behaviours 1.267
No symptoms of eating disorders (n=343) 118 (34.4) 225 (65.6)
Prone to Eating disorders (n=57) 24 (42.1) 33 (57.9)
Psychosocial Factors
Perception of health status 15.399***
Good (n=204) 54 (26.5) 150 (73.5)
Moderately good (n=175) 77 (44.0) 98 (56.0)
Poor (n=21) 11 (52.4) 10 (47.6)
Physical activity self-efficacy 0.496 ***
Peer influence 0.468***
Family influence 0.298***
Beliefs in physical activity outcomes 0.207***
Depression score -0.116**
Perception of body weight status 1.366
Correct-estimator (n=238) 79 (33.2) 159 (66.8)
Under-estimator (n=95) 37 (38.9) 58 (61.1)
Over-estimator (n=67) 26 (38.8) 41 (61.2)
Body Size Discrepancy Score -0.143**
Body Parts Satisfaction 2.666
Satisfied (n=191) 60 (31.4) 131 (68.6)
Dissatisfied (n=209) 82 (39.2) 127 (60.8)

* p<0.05, ** p<0.01, *** p<0.001

182 Dan SP, Mohd Nasir MT & Zalilah MS

Table 3. Multiple linear regression of physical activity

Variables Unstandardised Standardised t p-value

coefficients coefficients

B Beta

Constant 11.430 12.660 0.0001

Self-efficacy 0.151 0.338 7.245 0.0001
Sex (male) 2.366 0.229 5.571 0.0001
Social influence (peer) 0.339 0.248 5.255 0.0001

Notes: R= 0.599; R 2= 0.359; Adjusted R 2 = 0.354; F=74.012, p=0.0001

activity outcomes score and physical activity were more dissatisfied with their body size
score (r=0.207, p<0.001) (Table 2). were more likely to be physically inactive.
The Center for Epidemiologic Studies For perception of body weight status, the
Depression Scale (CES-D) which was distribution of the respondents by physical
adapted from Radloff (1977) was used to activity categories was almost similar for
assess depressive symptoms. Table 2 shows under-estimators, correct-estimators and
a negative and weak relationship between over-estimators. Hence, there was no
depression and physical activity score (r= - significant association between perception
0.116, p<0.01). This finding indicates that of weight status and physical activity level
respondents with more depressive category (χ2=1.366, p>0.05).
symptoms were more likely to be physically For body parts satisfaction, more
inactive. respondents who were satisfied with their
Respondents’ body image was assessed body parts (68.6%) fell in the moderate or
in three parts which included perception of high physical activity level categories.
body weight status, perception of body size, Nevertheless, no association was found
as well as body parts satisfaction. For between physical activity level and body
perception of body size, a majority of the parts satisfaction (χ2 = 2.666, p>0.05) (Table
respondents with low physical activity level 2).
chose a smaller figure (Figure 3) of the
Contour Drawing Rating Scale to represent Multivariate analysis
their current and ideal body sizes while most
The multiple linear regression was used to
respondents with moderate and high measure the amount of influence
physical activity level chose Figure 4 or 5 (of independent variables had on a dependent
the Contour Drawing Rating Scale) to
variable. Based on the stepwise method
represent their current and ideal body sizes. used, only three variables were found to be
Further, body size discrepancy score of the significant in explaining physical activity
respondents was determined based on the
(Table 3). These three variables were
difference between perceived ideal body size physical activity self-efficacy, sex and peer
and perceived current body size. Results influence.
showed that there was a significant negative
The R-square of 0.359 implied that about
association between discrepancy score and 35.9% of the variance in physical activity was
physical activity score (r= -0.143, p=0.004) explained by physical activity self-efficacy,
(Table 2). In other words, respondents who
sex and peer influence. Based on the
 Factors Associated with Physical Activity Levels among Adolescents Attending School 183

Guildford’s rule of thumb, the R value of et al., 2009). Family members may restrict
0.599 indicated that the relationship between female children from playing outside by
the three variables (physical activity self- shielding them in the home in order to
efficacy, sex and peer influence) and physical protect them, while boys may be allowed to
activity was moderate. play outside because they are traditionally
Furthermore, Table 3 shows that viewed as being physically tough (Gomez et
physical activity self-efficacy (t = 7.245, p = al., 2004).
0.0001), sex (t = 5.571, p =0.0001) and peer Several studies have shown that
influence (t = 5.255, p=0.0001) contributed physical activity patterns varied by race
significantly toward physical activity at the (Schmitz et al., 2002; Gordon-Larson,
0.05 level of significance. In other words, an McMurray & Popkin, 2000). However, no
increase in physical activity self-efficacy, association was found between physical
peer influence and being male tended to activity levels and ethnicity in this study. In
increase the physical activity levels among contrast, studies done on US adolescents
the adolescents. reveal ethnic differences in physical activity.
More non-Hispanic black and Hispanic
DISCUSSION adolescents were found to engage in
moderate to vigorous physical activity
Many studies have reported on children and compared to Asian adolescents. Likewise,
adolescents being physically inactive NHMS III (MOH, 2008) showed that the
(Monge-Rojas et al., 2002; Pahkala et al., highest prevalence of physical inactivity was
2006). Similarly, the present study showed observed among Chinese, followed by
that more than one-third of the adolescents Indians, Other Bumiputras (indigenous
were physically inactive. This is consistent people) and Malays.
with a study carried out by Pahkala et al. The BMI status was not correlated with
(2006) which showed that the amount of physical activity level in the present study.
leisure-time physical activity produced by However, Lahti-Koski et al. (2002) reported a
one-third of 13-year-old girls was extremely negative relationship between BMI and
low. Further, Jago et al. (2005) also reported physical activity level whereby overweight
that high levels of sedentary activity were subjects tended to be more physically
found among adolescent boys and girls inactive. In contrast, overweight Malaysian
during all time periods. adolescents were found to be more
Sex was one of the important predictors physically active than normal weight
of physical activity (Kjelsas & Augestad, adolescents (Pon, Mirnalini & Mohd Nasir,
2004). A number of studies reveal that 2004). The discrepancy among these studies
physical activity levels vary by sex whereby may be due to the selection of different
male adolescents are more physically active segments of the study population, for
as compared to female adolescents (Aarnio, example, adolescents versus children.
2003; Wenthe, Janz & Levy, 2009). Similarly, Socio-demographic factors were found
our study also showed that males were more to be related to physical activity in many
physically active than females whereby previous studies (Aarnio, 2003; Gordon-
female adolescents were twice as likely as Larson et al., 2000; Lasheras et al., 2001). In
male respondents to fall in the low physical contrast, the present study found none of
activity level category. Boys are more likely the socio-demographic factors to be
to be encouraged to participate in physical correlated with physical activity except
activity by family and friends than girls, father’s total years of schooling.
while girls are less likely than boys to be Nevertheless, other studies reported that
rewarded for their physical activity (Wenthe children and adolescents whose parents had
184 Dan SP, Mohd Nasir MT & Zalilah MS

higher educational levels were more In line with self-efficacy, social influence
physically active than those children and and support from family and peers were
adolescents whose parents had a lower associated with greater involvement in
educational level (Schmitz et al., 2002; physical activity (Springer, Kelder &
Gordon-Larson et al., 2000; Lasheras et al., Hoelcher, 2006; Wenthe et al., 2009). This
2001) study also showed a positive relationship
Dieting in adolescents is commonly between family and peer influence with
believed to be associated with negative physical activity. However, findings showed
health behaviours. A study of 1482 that friends exerted more influence on the
university students in Norway which was adolescent to be involved in physical activity
conducted by Kjelsas & Augestad (2004) as compared to family influence.
found that disordered eating behaviour was This study indicated that beliefs for
a significant predictor of physical activity physical activity outcomes were positively
(p<0.05). The present study showed that correlated with physical activity. On the
14.3% of the adolescents were prone to eating other hand, depressive symptoms have been
disorders but no association was found shown to predict higher sedentary leisure
between EAT-26 score category and physical habits (Schmitz et al., 2002) and consistent
activity level. with this, the present study showed a
A variety of psychosocial factors have negative relationship between depression
been studied as to their influence on physical and physical activity. This reflects that
activity in adolescents. The present study adolescents with depressive symptoms are
showed a significant relationship between more likely to be less physically active.
perception of current health status and Body image perception and
physical activity level among adolescents, dissatisfaction were found to be associated
a reflection that respondents with a better with physical activity. Nevertheless, no
perception of their current health status were association was found between perception
also more physically active. This was of weight status and body parts satisfaction
consistent with other studies which also with physical activity level while a
indicated that adolescents who perceived significant and negative relationship was
their current health and health improvement found between body size discrepancy and
as important were more physically active physical activity in the present study. This
(Aarnio et al., 2002; Aarnio, 2003). Likewise, indicated that respondents who were
Schmitz et al. (2002) also reported that a dissatisfied with their body size were more
higher value given to health predicted higher likely to be physically inactive. Likewise,
physical activity and lower sedentary leisure Anton, Perri & Riley (2000) reported that
habits in adolescents. larger discrepancies between ideal body size
Numerous studies have shown that self- and actual body size were found to be
efficacy is strongly correlated with physical associated with lower physical activity.
activity (Lewis et al., 2002; Wenthe et al., Monge-Rogas et al. (2002) also reported that
2009). Adolescents with higher self the influence of body image was the most
confidence are more likely to participate in important factor in explaining physical
physical activities (Monge-Rojas et al., 2002). activity among Costa Rican adolescents.
In line with other studies, the present study The multivariate linear regression
showed that physical activity self-efficacy analyses showed that physical activity self-
was positively and moderately correlated efficacy, sex and peer influence were
with physical activity among adolescents. significant predictors of physical activity
In other words, adolescents who have a among adolescents. Based on these findings,
higher self-confidence in physical activity an increase in physical activity self-efficacy
were more likely to be physically active. and peer influence will likely increase the
 Factors Associated with Physical Activity Levels among Adolescents Attending School 185

physical activity levels of adolescents. This ACKNOWLEDGEMENTS

shows that adolescents who had more
confidence in physical activity such as The authors would like to express their
having skills needed to be physically active appreciation to all parties involved in this
or were able to overcome the barriers of collaborative effort - Ministry of Education
physical activity (homework and time (MOE), State Education Department of
constraints) were more likely to be involved Pahang and Universiti Putra Malaysia. We
in physical activity. Moreover, adolescents also wish to extend our gratitude to the
with greater encouragement from friends school principals, teachers and all the
were more physically active as well. Sex also participants for their support throughout
played an important role in physical activity the duration of the study.
whereby male adolescents were more
physically active. This is consistent with REFERENCES
numerous studies which have also shown
similar results whereby self-efficacy is a Aarnio M, Winter T, Kujala U & Kaprio J
strong predictor of physical activity (Strauss (2002). Associations of health related
et al., 2001; Monge-Rojas et al., 2002). Some behaviour, social relationships, and
studies also showed that peer participation health status with persistent physical
in physical activity and support from peers activity and inactivity: a study of Finnish
was positively correlated with physical adolescent twins. Br J Sports Med 36: 360-
activity level (Springer et al. 2006; Wenthe et 364.
al., 2009). Aarnio M (2003). Leisure-time physical
activity in late adolescents: a cohort
CONCLUSION
study of stability, correlates and familial
aggregation in twin boys and girls. J
The primary findings in the present study
Sports Sci Med 2: 1-41.
show that self-confidence in performing
physical activity, social influences as well Anton SD, Perri MG & Riley JR (2000).
as sex are the main factors that predict Discrepancy between actual and ideal
physical activity levels among adolescents. body images impact on eating and
These results provide further support that exercise behaviors. Eating Behaviors 1:
self-efficacy and peer influence play 153-160.
important roles in influencing changes in Center for Disease Control and
physical activity among adolescents. As Prevention(CDC) (1997a). Guidelines for
such, these should be potential areas to school and community programs to
embark upon when developing physical promote lifelong physical activity
activity interventions to further establish an among young people. J Sch Health 67:
active lifestyle for adolescents. Besides, male 202-210.
and female adolescents exhibit different
physical activity levels with female Ernst MP & Pangrazi RP (1999). Effects of a
adolescents being more physically inactive. physical activity program on children’s
This should be taken into account when activity levels and attraction to physical
planning and implementing physical activity. Pediatr Exerc Sci 11:393-405.
activity programmes or interventions. Garner DM & Garfinkel PE (1979). The
eating attitudes test: an index of the
symptoms of anorexia nervosa. Psychol
Med 9: 273-279.
186 Dan SP, Mohd Nasir MT & Zalilah MS

Gomez JE, Johnson BA, Selva M & Sallis JF Lewis BA, Marcus BH, Pate RR & Dunn AL
(2004). Violent crime and outdoor (2002). Psychosocial mediators of phy-
physical activity among inner-city sical activity behaviour among adults
youth. Prev Med 39: 876-881. and children. Am J Prev Med 23: 26-35.
Gordon-Larson P, McMurray RG & Popkin Lim SF (2005). Pengambilan nutrien, aktiviti
BM (2000). Determinants of adolescents fizikal dan status kesihatan tulang di
physical activity and inactivity patterns. kalangan pelajar sekolah menengah.
J Pediatr 105: 1-8. Bachelor in Science thesis. (Pemakanan
Hinkle DE, Wiersma W & Jurs SG (1981). dan Kesihatan Komuniti). Universiti
Applied Statistics for the Behavioral Putra Malaysia, Serdang, Malaysia.
Sciences. Rand McNally College McCabe MP & Ricciardelli LA (2004). A
Publishing Company, Chicago. longitudinal study of pubertal timing
and extreme body change behaviours
Jago, R, Anderson, CB, Baranowski, T &
Watson, K (2005). Adolescent patterns among adolescent boys and girls.
of physical activity: differences by Adolescence 39: 145-166.
gender, day, and time of day. Am J Prev Ministry of Health Malaysia(MOH) (2008).
Med 28: 447-452. The Third National Health and
Morbidity Survey (NHMS III) 2006, Vol
Kjelsas E & Augestad LB (2004). Gender,
eating behaviour, and personality 2. Institute for Public Health, Ministry
characteristics in physically active of Health Malaysia.
students. Scand J Med Sci Sports 14: 258- Monge-Rojas R, Nunez HP, Garita C & Chen-
268. Mok M (2002). Psychosocial aspects of
Costa Rican adolescents’ eating and
Koay CY (1998). Imej tubuh dan amalan
pemakanan di kalangan remaja sekolah physical activity patterns. J Adolesc
di Ipoh. Bachelor in Science Health 31: 212-219.
thesis(Pemakanan dan Kesihatan Pahkala K, Heinonen OJ, Lagstrom H,
Komuniti). Universiti Putra Malaysia, Hakala P, Sillanmaki L & Simell O
Serdang, Malaysia. (2006). Leisure-time physical activity of
Kowalski KC, Crocker PRE & Donen RM 13-year-old adolescents. Scand J Med Sci
(2004). The Physical Activity Sports :1-7.
Questionnaire for Older Children (PAQ- Pon LW, Mirnalini K & Mohd Nasir MT
C) and Adolescents (PAQ-A) Manual. (2004). Body image perception, dietary
practices and physical activity of
Lahti-Koski M, Pietinen P, Heliovaara M &
Vartiainen E (2002). Association of body overweight and normal weight
Malaysian female adolescents. Mal J
mass index and obesity with physical
Nutr 10(2): 131-147.
activity, food choices, alcohol intake,
and smoking in the 1982-1997 FINRISK Radloff LS (1977). The CES-D scale:a self-
studies. Am J Clin Nutr 75: 809-817. report depression scale for research in
Lasheras L, Aznar S, Merino B & Lopez EG the general population. Applied Psych
Meas 1: 385-401.
(2001). Factors associated with physical
activity among Spanish youth through
the National Health Survey. Prev Med
32: 455-464
 Factors Associated with Physical Activity Levels among Adolescents Attending School 187

Saunders RP, Pate RR, Felton G, Dowda M, Thompson MA & Gray JJ (1994).
Weinrich MC, Ward DS, Parsons MA & Development and Validation of a New
Baranowski T (1997). Development of Body Image Assessment Scale. In:
questionnaires to measure psychosocial Handbook of Assessment Method for Eating
influences on children’s physical Behaviour and Weight Related Problem-
activity. Prev Med 26: 241-247. Measures Theory and Research: Assesment
Schmitz KH, Lytle LA, Phillips GA, Murray of Body [Link] JK. Sage
DM, Birnbaum AS & Kubik MY (2002). Publication International Education and
Professional Publisher, pp.119-148.
Psychosocial correlates of physical
activity and sedentary leisure habits in Wenthe PJ, Janz KF & Levy SM (2009).
young adolescents: the teens eating for Gender similarities and differences in
energy and nutrition at a school study. factors associated with adolescent
Prev Med 34: 266-278. moderate-vigorous physical activity.
Pediatr Exerc Sci 21(3): 291-304.
Simko MD, Cowell C & Hreha M (1989).
Practical Nutrition: A Quick Reference World Health Organization (1995). Physical
for the Health Care Practitioner. Aspen Status: The Use and Interpretation of
Publishers, Rockville, 282-286. Anthropometry. WHO Technical Series
Springer AE, Kelder SH & Hoelscher DM Report No. 854. WHO, Geneva.
(2006). Social support, physical activity World Health Organization (2000).
and sedentary 6th-grade girls: a cross- Nutrition for Health and Development:
sectional study. Int J Behav Nutr Physical A Global Agenda for Combating
Activity 3: 1-14. Malnutrition. WHO, Geneva.
Strauss RS, Rodzilsky D, Burack G & Colin
M (2001). Psychosocial correlates of
physical activity in healthy children.
Arch Pediatr Adolesc Med 155: 897-902.