HEALTH EDUCATION RESEARCH
2 1996
Theory & Practice
Nutritional knowledge, beliefs and behaviours in
teenage school students
D.Gracey, N.Stanley, V.Burke, B.Corti1 and LJ.Beilin
Abstract
Three hundred and ninety-one adolescent West-
ern Australians, mean age 15.8 years, completed
questionnaires to determine nutritional know-
ledge and behaviours including stage of change;
health beliefs and values; barriers to change;
self-efficacy; locus of control; dietary patterns;
alcohol and smoking habits; television-watch-
ing; weight, height and body image. Highest
ranking beliefs and values regarding healthy
diets were improving health, feeling energetic,
feeling good about oneself, controlling weight,
lowering cholesterol, testing willpower and
improving appearance. Important barriers to
healthy eating were lack of suitable foods at
home and school, inability to influence food
choices at home, and ignorance about nutrients.
Nutritional knowledge, particularly concerning
fat, was deficient. Healthy eating related nega-
tively to television watching and alcohol, and
positively to self-efficacy, nutrition knowledge,
considering weight control and well-being as
important, and having influence over foods at
home. Of the 28% of boys and girls who drank
alcohol, 20% reported intake above 'safe' limits.
Twenty-four percent of boys and 22% of girls
smoked. Fifty-four percent of girls and 21% of
boys considered themselves overweight includ-
ing 20% of the leanest girls and 8% of the
leanest boys. Nutrition education for adolescents
University Department of Medicine, Royal Perth Hospital,
West Australian Heart Research Institute, Perth 6001 and
'Department of Public Health, University of Western
Australia, Nedlands, 6907, Australia
© Oxford University Press
Vol.11 no.
Pages 187-204
should incorporate self-efficacy, relevant health
values and barriers-to-change, education about
nutrients, and improved access to healthy foods.
Adolescent smoking and drinking should also
be targeted.
Introduction
Reductions in morbidity and mortality associated
with lifestyle diseases may be achievable if satis-
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
factory nutritional practices are adopted in early
life and maintained in the long-term (Gliksman
et al., 1987, 1993). During adolescence, young
people are assuming responsibility for their own
eating habits, health-related attitudes and behavi-
ours (Coates et at, 1982) and their growing inde-
pendence is often associated with unconventional
eating patterns (Truswell and Damton-Hill, 1981).
Although dietary behaviour during adolescence
may be transitory in some individuals, health-
related behaviours show tracking through adoles-
cence (Kelder et at, 1994). In a survey of teenagers
in Scotland, Sweeting and Anderson (1994) found
that dietary habits appeared to be established by
the age of 15. If habits acquired in adolescence
persist into adult life, behaviours established in
young people may have important long-term con-
sequences for health.
Knowledge about healthy food choices can be
a predisposing factor for the adoption of a healthy
diet (Thomas, 1994) but it is insufficient to motivate
healthy eating (Carmody et al., 1987) and psycho-
social factors must also be considered. Many
behavioural models have been applied to eating
behaviour. The Health Belief Model (Rodenstock,
1966) maintains that health habits are a function
187
 D.Gracey et aL
of perceived vulnerability to a disorder and the
belief that a particular health measure will be
sufficient to overcome this vulnerability. This
enables understanding of why people practice
health behaviours and prediction of some of the
circumstances under which their health behaviour
will change. However, undertaking many health
behaviours also requires a sense of personal control,
i.e. a belief that it is actually possible to perform
the health behaviour. Bandura (1977, 1986)
reasoned that an important determinant of the
practice of health behaviours is a sense of self-
efficacy which is the belief that an individual can
control their practice of a particular behaviour.
Self-efficacy affects a range of health behaviours
including weight control (Strecher et aL, 1986)
and quitting smoking (Prochaska and DiClemente,
1984). In reviewing the role of self-efficacy in
achieving health behaviour change, Strecher et aL
(1986) found strong relationships between self-
efficacy and both change and maintenance of
behaviour. Experimental manipulations designed
to increase feelings of self-efficacy can be success-
ful and enhancement of self-efficacy is related to
subsequent behaviour change (Strecher et aL.
1986).
Locus of control is similar to self-efficacy but
refers to a more general expectation relating to an
individual's interaction with the environment rather
than beliefs about the practice of a specific health
behaviour. Individuals with an 'internal' locus of
control are more likely to engage in preventive
health behaviours because they see these steps as
helping to protect them against poor health (for
review, see Strickland, 1978). However, there is
not a strong relationship between locus of control
and preventive health behaviour (Taylor, 1986; for
review, see Wallston and Wallston, 1984).
Ajzen (1985) and Ajzen and Madden (1986)
revised the theory of Fishbein and Ajzen (1975)
to formulate the Theory of Planned Behaviour.
They argue that, in addition to knowing an indi-
vidual's attitudes, subjective norms and behavi-
oural interactions with respect to a given behaviour,
it is necessary to know their perceived behavioural
188
control over that action. The Theory of Planned
Behaviour maintains that a health behaviour will
be adopted if an individual believes that the advant-
ages of success outweigh the disadvantages of
failure, if they believe they are expected to perform
the behaviour by people with whom they wish to
comply, and if they have enough control over
internal and external factors influencing the attain-
ment of the behavioural goal (Taylor, 1986). Thus,
feelings of perceived control and self-efficacy
appear to be important in demonstrating attitude-
behaviour consistency, even when there is a clear
behavioural intention to act on an attitude.
The uptake and maintenance of health-related
behaviour will also be determined by an indi-
vidual's readiness to change, as proposed by
Prochaska and DiClemente (1983). The stages of
change model, which relates to the transtheoretical
model of behaviour change (Prochaska and
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
DiClemente, 1992), classifies behaviours as pre-
contemplation (not interested in change), contem-
plation (thinking about changing), preparation
(planning to change), action (actively changing
behaviour) or maintenance (sustaining change). In
a large survey (Glanz et aL, 1994) there was a
stepwise association between dietary behaviour and
stage of change based on a seven-item algorithm.
Nutrition intervention programs are likely to
have greater success when both influences on food
choice and the theoretical framework applicable to
changes in health-related behaviours are taken into
account. Perception of risk, belief in a connection
between diet and disease, perceived benefits of a
healthy diet, and knowledge of which foods should
be chosen predispose to healthy dietary choices.
Recognition of barriers to change, the extent of
social support and the development of skills, e.g.
in selecting low-fat foods, are among enabling
factors affecting food choices. Assessment of the
costs and benefits of dietary change and dealing
with feelings of deprivation from restriction of
favorite foods also influence diet-related behaviour.
Sufficient knowledge about diet is needed for an
individual to assess the quality of their own diet
and their perception of dietary norms, particularly

in regard to people whose opinions they consider
significant, influences their classification with
regard to stage of change. Motivation to adopt
healthy eating patterns and self-efficacy are import-
ant determinants of behavioural change (Glanz
et ai, 1993).
If behavioural modification is to be achieved,
factors influencing that behaviour, such as habits,
attitudes, knowledge and barriers to change, must
be identified (Sallis et al., 1988). Understanding
the health beliefs of adolescents is particularly
important as, in this age group, there are consider-
able differences from those of adults, even when
adolescents are compared with their parents
(Nada-Raja et aL, 1994). In the present study, we
measured knowledge about nutrition and beliefs,
behaviours, and possible barriers affecting healthy
food choices in adolescents in Perth, Western
Australia. These findings were examined in relation
to eating habits and other lifestyle factors.
Subjects and methods
Subjects
The sample included 480 year 11 students (15-16
year olds) from two Perth metropolitan public high
schools and one co-educational private school.
High schools were randomly chosen from a list
of higher and lower socioeconomic status (SES)
schools classified according to the Australian
Bureau of Statistics' relative urban advantage
(McLennan, 1990) where higher SES relates to
such indicators as non-manual occupations and
higher income. A single co-educational college
was also selected to compare students from public
schools with a higher SES private school.
With the cooperation of the year 11 coordinators
at the three schools, questionnaires were distributed
to all year 11 students in mainstream classes,
comprising 110 students at the low SES public
school, 120 students at the private school and 250
from the high SES public school. Special education
classes and students learning English as a second
language were excluded because of possible liter-
acy problems in completing the questionnaire.
Nutritional knowledge and behaviour in adolescents
Questionnaire
The questionnaire was completed at each school
under teacher's supervision. It included three items
on self-reported age, weight and height, seven
items on stages of behaviour change, 18 items on
each of health belief and values, 18 items on self-
efficacy, five items on locus of control, 16 items
on barriers to change, eight items on nutritional
knowledge, and seven items on eating patterns. In
addition there was one item on each of alcohol,
smoking habits and television watching. Usual
intake of fatty foods, water and soft drinks was
determined using a 16-item questionnaire; the
variety of foods eaten was addressed using an
additional 22 items. Perceptions of body image
and whether usual eating habits were healthy were
assessed using one item for each and additional
items sought information about the use of special
diets.
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
Behavioural variables and construction of the
scales
Readiness to change behaviour was based on the
stages of behaviour change theory (Prochaska and
DiClemente, 1983) using questions adapted from
an algorithm by Greene et al. (1994). Students
were asked if they considered their diet to be
healthy and, if so, for how long they had eaten a
healthy diet (action and maintenance). Those who
considered their diet to be unhealthy were asked
whether they intended to adopt healthier eating
habits and, if so, how immediate was their intention
to change (pre-contemplation and contemplation).
A six-point Likert scale was used in the evalu-
ation of health beliefs, health values, barriers to
change and self-efficacy. For 18 health beliefs, the
scale ranged from true (1) to untrue (6). Health
values (18 items) and barriers to change (16 items)
ranged from important (1) to unimportant (6).
Questions relating to health beliefs were of the
form: 'If I were to adopt a healthier diet for the
next two weeks it would....' followed by statements
such as 'improve my appearance' or 'help lower
my cholesterol'.
189
 D.Gracey et al.
In the self-efficacy section students ranked items
according to their confidence that they could do
them (sure they could = 1, sure they could not =
6). Only 18 of the 61 items from those used by
Sallis et al. (1988) were chosen to keep the
questionnaire to a manageable length. This subset
had previously been validated by the Department
of Medicine, Royal Perth Hospital. These items
were of the form: 'How sure are you that you
could ....' followed by statements such as 'stick
to healthy foods when eating with friends' or
'substitute reduced fat milk for whole milk'.
Control over food intake was assessed with a
five-point Likert scale which ranged from almost
always (1) to never (5) (Williams et al, 1993).
Items were: (1) Do you have any say in what
foods are bought at home? (2) Do you have any
say in how food is prepared or cooked at home?
(3) Do you choose what goes onto your plate? (4)
Do you decide how much you eat? (5) Do you
decide what you eat for lunch at school? (6) When
you are out with friends do you decide for yourself
what to eat? (7) Do you eat breakfast? Locus of
control was measured On a four-point Likert scale
which ranged from very much or definitely (1) to
none at all or not at all (4) (Falconer et al, 1993).
Items asked whether students considered (1) that
diet affected future health; (2) that health depends
on how they take care of themselves; (3) that it is
possible to prevent sickness; (4) that they would
try foods recommended by health professionals;
and (5) that they were more concerned about diet
than they used to be. Perception of body image
was determined with one item asking if the students
considered themselves to be underweight, normal
weight or overweight.
Nutritional knowledge
Eight items about nutritional knowledge based on
the Australian dietary guidelines (NHMRC, 1993)
were included (Vandongen et al, 1995). These
items aimed to assess whether students knew
enough about the nutrients, particularly fat and
fibre, in common foods to be able to make healthy
food choices. The items are shown in Table VI
with the four possible responses for each, only one
190
of which was correct A total nutrition knowledge
score was obtained by adding the responses, scoring
one for each correct answer and zero otherwise.
Lifestyle factors
In the school time available, it was impossible to
use diet records or detailed food frequency ques-
tions to gauge nutrient intake. A 30-item food
variety score adapted for Western Australian chil-
dren (Milligan, 1994) was used as a measure of
eating diversity. In assessing food variety, indi-
viduals were asked to tick which of 30 food items
they had eaten in the past week. Responses do not
measure frequency of consumption but the number
of different foods eaten during the week. The food
variety score addresses the first dietary guideline
for Australians to 'enjoy a wide variety of nutritious
foods' (NHMRC, 1993). A high food variety score
indicates compliance with this NHMRC guideline
and has been shown in other populations to correl-
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
ate with healthy eating habits (Ries and Dachler,
1986).
Questions about fat intake, validated against
food frequency data (Kinlay et al, 1991), provided
a score indicating usual fat consumption. The items
included in the fat score related to the weekly
consumption of ice cream, cheese, whole milk and
low fat milk, biscuits, cake, chocolate, and the
type of spread (butter or margarine) used. Questions
also asked whether foods were usually fried and
whether individuals usually removed all, some or
none of the fat from meat and chicken before
eating. Additional questions assessed consumption
of fish, soft drinks and water, the addition of salt
to prepared foods and the use of special diets, as
well as the use of alcohol and cigarettes and the
amount of television watching on weekdays and
weekends.
Statistical methods
Data were analysed using SPSS/PC. For continuous
variables (age, weight, height) one-way analysis
of variances was used to compare schools and
gender. For interval variables Mann-Whitney U-
tests or Kruskal-Wallis tests were used to compare
groups, x 2 t e s t s were used for categorical data.
 Nutritional knowledge and behaviour in adolescents

Internal consistency was assessed using Cronbach's
OL As internal reliability for the self-efficacy scale
was acceptable (Cronbach's a = 0.8763), a total
score was calculated by summing the responses to
all self-efficacy questions. Responses were recoded
as 'most important' = 6 and 'least important' =
1, so that higher values indicated greater levels of
self-efficacy. Similarly a total score was calculated
for locus of control (Cronbach's a = 0.7683).
Spearman correlation coefficients were used to
examine relationships between psychological vari-
ables, nutrition knowledge and eating patterns, and
these relationships were explored further using
linear regression. A level of P < 0.05 was consid-
ered significant
Results
Subjects
Table I shows means and confidence intervals
for the age, weight, height and body mass index
(BMI) for the respondents. Girls at the high SES
school were significantly shorter and lighter than
those at the other school; however, there was no
significant difference for BMI. The mean BMI for
boys and girls at all schools corresponded to
reported normal values for this age group
(Hammer, 1991).
Stages of change
A significantly greater (P < 0.05) percentage of
boys at the low SES school (70%) considered that
they made healthy food choices compared to the
high SES school (57%) and the private school
(55.1 %). Significantly more girls (79.7%) than boys
(59.5%) tried to choose healthy foods (P < 0.05).
Among students in the maintenance phase (eat-
ing a healthy diet for more than 6 months) there

Downloaded from her.oxfordjournals.org by guest on May 26, 2011

Of the 480 eligible year 11 students, a total of 391
completed the questionnaire. Questionnaires were
completed by 85 (77.3%) of the 110 students, at
the low SES school, 104 (86.7%) of the 120 at the
private school and 202 (80.8%) of the 250 students
at the high SES school. Gender distribution was
almost equal (191 boys and 200 girls) with boys
making up 47% of participants at the low SES
school, 48% at the private school and 50% at the
high SES school.
was a significantly greater proportion of boys from
the high SES school (53.6%) compared with 38.8%
from the private school and 47.5% from the low
SES school (P < 0.05). For girls there was also a
significantly greater proportion in the maintenance
phase from the high SES school (P < 0.05), the
respective percentages being 52.9, 37.7 and 46.7%.
Significantly more girls from the private school
(39.6%) were in the action phase (changed to a
healthy diet in the past 3 months) compared with

Table I. Characteristics of the students according to school (95% Cl)

Variable Low SES school Private school High SES school Total

Age (yean)
boys 15.7 (153-15.9) 15.8 (15.6-16.0) 15.8(15.6-15.9) 15.8(15.7-15.9)
girls 15.6(153-15.8) 15.8 (15.6-16.0) 15.9(15.8-16.0) 15.8(15.7-15.9)
Weight (kg)
boys 64.2 (61.0-67.0) 69.8 (62.5-77.1) 64.5 (62.6-663) 65.6 (63.5-67.7)
girts 56.9(52.7-61.0) 56.0 (54.0-58.0) 52.5 (50.7-54.3)" 54.6 (53.2-56.0)"
Height (cm)
boys 175.8 (172.6-179.0) 176.3(173.5-179.2) 174.1 (172.2-176.1) 175.0(173.6-176.4)
girls 165.2(162.7-167.8) 169.2(167.0-171.3) 164.0(1623-165.4)" 165.8(164.6-166.9)
BMI (kg/m2)
boys 20.7(19.8-213) 22.6(20.1-25.1) 21.2(20.7-21.8) 213 (20.8-22.2)
girls 20.8(19.6-22.1) 19.7(19.0-20.4) 193 (18.9-20.2) 19.9 (19.4-20.4)*

'P < 0.05 in one-way ANOVA for differences 1between schools. Low SES. n = 85: private, n = 104; high SES, n = 202.

191
 D.Gracey et al.

the low SES (26.7%) and high SES (26.4%) schools
(/> < 0.05). Significantly more girls (30.3%) than
boys (15.7%) were classified in the action phase
(P < 0.05).
Amongst those who said they did not try to eat
a healthy diet, 63% were pre-contemplators who
did not intend ever to eat a healthy diet. A
significantly higher proportion of boys who were
precontemplators came from the high SES school
(54.7%) compared with 18.9% from the low SES
school and 26.4% from the private school
(P < 0.05). Similarly, for girls precontemplators
comprised 56.5% from the high SES school, 17.4%
from the low SES school and 26.1% from the
private school (P < 0.05). Boys made up a
significantly greater proportion (P < 0.05) of the
pre-contemplators (69.7%) than girls (30.3%).
Health beliefs and values
Overall, health beliefs considered to be true by the
largest proportion of students related to improving
health (81%), feeling energetic (63%), feeling
good (62%), lowering cholesterol (61%), testing
willpower (57%), weight control (53%) and
improving appearance (44%). Significantly more
girls than boys believed that the benefits of a
healthy diet related to improving health, feeling
good about oneself, weight control, improving
appearance, feeling energetic, testing their
willpower and saving money.
Table II shows the percentage of boys and girls
ranking health values in the highest two categories.
The health values considered important by the
greatest proportion of students overall were
improving health (82%), increasing energy (81%),
feeling good about oneself (80%), weight control
(61%), improving appearance (58%), saving money
(58%) and lowering cholesterol (51%). The propor-

Downloaded from her.oxfordjournals.org by guest on May 26, 2011

Table II shows the items used for health beliefs tion of girls ranking feeling good about oneself,
and values and the percentage of boys and girls weight control, improving health and appearance
ranking these items in the highest two categories. as important was significantly higher than for boys.

Table II. Percentage of responses ranked in the highest two categories for health beliefs and values

Items for beliefs and values relating to the effects of Responses in the highest two categories (%)
a healthy diet
Beliefs Values

Boys Girls Boys Girts

Improving health 74.6 88.2" 75.7 88.9"

Feeling energetic 57.1 68.5° 78.5 83.4
Feeling good about myself 52.6 73-5* 76.3 86.4"
Lowering cholesterol 58.4 64.1 53.6 49.5
Losing weight or maintaining desired weight 44.2 61.8" 49.0 72.2*
Testing willpower 52.6 61.3" 47.4 50.1
Improving appearance 36.0 52.5" 52.6 61.8a
Cutting down food choices 33.9 37.0 19.4 21.4
Saving money 26.8 35.5" 57.1 58.3
Being a good example for others 23.3 28.1 25.5 27.1
Craving certain foods 19.8 25.0 15.2 11.2
Changing mood and personality 20.6 21.6 27.1 34.8
Involving a lot of planning and inconvenience 20.9 14.7 16.8 19.6
Leading to failure 5.4 8.2 20.7 22.5
Offending people 6.3 6.6 20.5 20.8
Not being supported by my family 4.7 6.5 31.9 33.3
Limiting social life 4.2 7.1 28.4 21.7
My friends laughing at me 5.8 4.6 13.1 14.6

'P < 0.05 for differences related to gender.

192
 Nutritional knowledge and behaviour in adolescents

The response for health beliefs and health values
did not differ significantly between schools.
Barriers to change
Table HI shows the items used and the percentage
of boys and girls ranking barriers to change in the
highest two categories of importance. The barrier
to change considered important by the largest
proportion of students, combining responses from
boys and girls, was healthy food not being available
at home (45.1%). Other barriers related to healthy
foods being unavailable in school canteens
(41.4%), ignorance of calorie content (40.3%),
sugar and fat content (42.2%) and fibre content
(41.1%) of food, lack of control over buying of
food (40.2%), and problems in sticking to a healthy
diet (39.6%), particularly in girls. Girls were signi-
ficantly more likely than boys to report a number
of barriers including family support, ignorance
of fibre content, insufficient planning time and
differences in healthy food choices in relation to
school. Boys who considered their food choices as
healthy had significantly greater self-efficacy
scores than other boys. Similarly girls who reported
trying to eat a healthy diet had a higher self-
efficacy score than those reporting unhealthy diets.
Locus of control scores showed similar patterns
to those for self-efficacy, with significantly higher
values in girls (Table IV). For those reporting their
diet as healthy, locus of control score was higher
than in other students. There were no significant
differences in locus of control scores according
to school.
Control over food intake
Table V shows the proportion of boys and girls
who considered they 'almost always' or 'quite
often* had control over their food intake. Signific-
antly more girls than boys had some say over how

Downloaded from her.oxfordjournals.org by guest on May 26, 2011

problems in sticking to a healthy diet
Self-efficacy and locus of control
The self-efficacy score (Table IV) was higher in
girls than in boys, but there were no significant
their food is prepared, what is served on their plate
and what they eat for school lunch. There were no
significant differences in the overall proportion of
boys and girls eating breakfast; however, there were
significant differences between girls according to

TaWe IIL Percentage of responses in the highest two categories for items relating to perceived barriers to eating a healthy diet

Items used to determine barriers to change Percentage of responses ranked in the

highest two categories

Boys Girls

Foods that fit into a healthier diet are not available at home 45.4 44.9
I find it difficult to make healthy food choices at the school canteen 40J 42.0
I do not know which foods are best to reduce sugar and fat 38.3 45.2
I have no control over the foods available at home 37.4 42.9
1 do not know how many calories are in different foods 35.6 42.3
I do not know which foods are high in fibre 35.3 46.2*
I have trouble knowing how much I should eat 32.6 37.6
I have trouble sticking to a healthy diet 32.4 44.7*
I find it difficult to choose healthy foods for school lunches 31.6 35.7
I have trouble choosing healthy foods when I am out with family or friends 29.8 36.1
My family does not support my efforts to eat a healthier diet 29.4 35.9"
I don't see any benefit! from my efforts to eat a healthier diet 26.6 32.1
It is difficult to find time to plan healthy meals 26.3 36.91
It is difficult for me or my family to shop for the foods I need 24.5 30.5
I use food as a treat orrewardfor myself 20.2 26.0
I find that a healthy diet is too expensive 17.6 24.9

'P < 0.05 for differences related to gender.

193
 D.Gracey et aL

Table IV. Self-efficacy scores and locus of control scores (mean and 95% Cl) related to gender and to healthy or unhealthy food
choices

Variable Healthy food choices Unhealthy food choices Total

Self-efficacy
boys 76.4b 66.5 72.41
(73.2-79.6) (62.5-70.4) (69.9-74.9)
girts 80.2" 70.8 78.3
(77.7-82.8) (65.7-76.4) (76.0-80.6)
Locus of control
boys 16.6b 14.9 15.9°
(16.2-16.9) (14.2-15.5) (153-16.3)
17.2b 15 J 16.8
girls
(16.9-17.5) (14.2-16 3) (163-17.1)

V < 0.05 for gender difference; bP < 0.05 for difference between patterns of food choice (Mann-Whitney test).

Table V. Percentage of boys and girls ranking responses to questions related to eating patterns as 'almost always' or quite
often'

Downloaded from her.oxfordjournals.org by guest on May 26, 2011

Variable Percentage ranking response as 'usually' X2 value (P)

Boys Girls

Influence on foods bought at home 56.8 61.0 0.724 (03949)

Influence on food preparation at home 38.0 513 7.19 (0.0073)
Choosing food served onto plate 43.8 62.0 13.10(0.0003)
Choosing amount of food eaten 78.6 83.0 1.20 (0.2733)
Choosing school lunch 66.7 793 8.23(0.0041)
Choosing food when out with friends 88.0 90.5 0.629 (0.4276)
Eating breakfast 71.9 68.0 0.699(0.4031)

Table VL Percentage of boys and girls correctly answering nutrition knowledge questions

Nutrition knowledge item Percent of correct answers

Boys Girls

Which has carbohydrate? Meat/Butter/Bread/Cheese 75.0 82.0

Which has no fibre? Brown bread/Beans/White bread/Meat 62.0 613
Which is high in fat? Potatoes/Boiled sweets/Margarine/Cottage cheese 34.9 44.0
Which is high in fibre? Orange juice/Eggs/Rice/Meal 78.1 8X0
Which has most iron? Milk, Cheese and eggs/Nuts, dried beans and cauliflower/ 59.4 76.0
Liver, beef and breakfast cereal/Rice, Vegemite and biscuits
Which is low in fat? Hamburger with 'the lot '/Grilled fish and chips/Ham and cheese sandwich/ 42.7 42.0
None of these '
Which increases fibre in a spaghetti meal? Cooking less/Using canned tomatoes/ 77.1 87.0
Wholemeal spaghetti/Adding cheese
Which meal is lowest in fat? Bread and polony/Roast lamb and baked vegetables/Fried egg/ 503 513
Jacket potato and baked beans

194

school. Only 43.4% of girls at the low SES school
regularly ate breakfast, compared to 72.3% at the
private school and 73% at the high SES school
(X2 19.16, P = 0.0140). The respective proportions
for boys were 72.5, 74 and 72.7%. There were no
significant differences by school for boys.
Nutrition knowledge
Eight items pertaining to nutrition knowledge were
included in the questionnaire (Table VI). The mean
score of correct responses was significantly greater
(Mann-Whitney 2.54, P = 0.011) for girls (5.3,
Q 5.0-5.5), than for boys (4.8, CI 4.5-5.0). Table
VI shows the items and the percentage of boys
and girls who answered each item correctly. At
least 60% of students answered items correctly
except for the three questions relating to fat content
of foods.
A number of misconceptions were evident For
example, 22% of students believed that white bread
rather than meat contained no fibre, 22% considered
that boiled sweets rather than margarine were high
in fat, 42% thought a ham and cheese sandwich
was low in fat, and 25% thought bread and
polony was lower in fat than a baked potato and
baked beans.
Food variety score
Food variety score measures the number of differ-
ent foods eaten during the preceding week and is
not a measure of frequency of consumption. The
total food variety score did not differ significantly
according to gender but was significantly lower
(Kruskal-Wallis 6.74, P = 0.0344) at the low SES
school (mean 15.9, CI 14.9-17.0) than at the
private school (mean 17.0 CI 16.1-17.9) and the
high SES school (mean 17.1, CI 16.6-17.7).
The food group scores differed significantly
between schools for cereal, fish, fruit and nuts with
the low SES school having the lowest score. For
cereals the mean score for the low SES school
was 3.0 (CI 2.8-3.2) compared with a mean of 3.4
(CI3.2-3.6) for the private school and 3.3 (CI 3.2-
3.4) for the high SES school (Kruskal-Wallis 8.83,
P = 0.0121). The mean score for fish intake was
2.2 (CI 1.9-2.6) for the low SES school, 2.4 (CI
2.1-2.7) for the private school and 2.8 (CI 2.6-
Nutritional knowledge and behaviour in adolescents
3.0) for the high SES school (Kruskal-Wallis
14.97, P = 0.0000). For fruit the mean score was
3.0 ( O 2.6-3.2) for the low SES school, 3.4 ( a
3.1-3.7) for the private school and 3.3 (CI 3.1-
3.5) for the high SES school (Kruskal-Wallis
6.05, P = 0.0421). The low SES school scored
significantly lower in consumption of nuts (Krus-
kal-Wallis 6.49, P = 0.0389) with a mean score
of 0.4 (CI 0.3-0.6) compared with the private
school (0.7, Q 0.6-0.9) and high SES school (0.6,
CI 0.5-0.7). Separate analysis by gender for cereal
and fruit showed no significant differences for
boys but for girls there were significant differences
for both food groups due mainly to very low
consumption amongst the girls at the low SES
school (Kruskal-Wallis for cereal 5.98, P = 0.0288,
for fruit 8.90, P = 0.0117). The mean score for
cereals for girls was 3.0 (CI 2.8-3.2) at the low
SES school, 3.4 (CI 3.2-3.6) at the private school
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
and 3.3 (CI 3.2-3.4) at the high SES school. For
fruit the respective means for girls were 2.7 (CI
2.2-3.2), 3.4 (CI 3.0-3.8) and 3.4 (CI 3.2-3.7).
Boys who considered they had eaten a healthy
diet for at least 6 months had a mean food variety
score of 18.0 (CI 16.9-19.1) compared with a mean
of 16.1 ( a 15.0-17.2) in other boys
(Mann-Whitney 2.12, P = 0.0129). The respective
mean values for girls were 17.4 (CI 16.8-17.1)
and 15.6 (CI 14.7-16.5) (Mann-Whitney 3.22,
P = 0.0001).
Eating habits
Measurement of usual fat consumption using the
fat score (Kinlay et al., 1991) showed significantly
higher levels (Mann-Whitney 4.70, P = 0.0000)
for boys (mean 6.0, CI 5.7-6.4) than for girls
(mean 5.4, CI 5.2-5.7), but did not differ between
schools. There was a significantly lower fat score
(Mann-Whitney 1.99, P = 0.039) for those who
tried to eat a healthy diet (mean 5.3, CI 5.0-5.6)
compared with those who did not (mean 5.8, CI
5.4-6.2).
Because of reported dietary differences between
drinkers and non-drinkers (Herbeth et al, 1988)
as well as smokers and non-smokers (Veenstra
et al., 1993), the fat score was examined in relation
195
D.Gracey et al.
to smoking and drinking habits. A significantly
higher fat intake was found in both smokers
(Mann-Whitney, 2.05, P = 0.0407) and drinkers
(Mann-Whitney 2.52, P = 0.0116). The mean fat
score in smokers was 6.0 (CI 5.6-6.4), compared
with non-smokers 5.3 (CI 4.6-6.1), and was 6.1
(CI 5.4-7.0) in drinkers compared with 5.2 (CI
4.5-5.8) in non-drinkers.
Overall, 45.5% of boys and 53.3% of girls rarely
added salt to prepared foods. For girls at the private
school 71.2% rarely used salt compared with 55.6%
at the low SES school and 43.6% at the high SES
school (x 2 5.98, P = 0.026). A similar, but not
significant, trend was seen amongst boys with
54.0% at the private school, 46.2% at the low SES
school and 41.0% at the high SES school rarely
using salt Reported consumption of fish was
examined because of its association with lower
risk of cardiovascular disease (Vandongen et al,
1993). Only 20.0% of boys and 22.1% of girls ate
fish at least fortnightly, and there were no signific-
ant differences related to school.
Body image and special diets
Almost three times as many girls (54%) as boys
(20%) considered themselves overweight. Of those
who were in the upper quartile of BMI distribution,
68.3% of girls and 31.7% of boys thought they
were too fat while 7.3% of boys in the upper BMI
quartile thought they were too thin. Of students in
the lowest quartile of BMI distribution, 21.1% of
girls and 7.9% of boys thought they were too fat
There was no significant difference in perception
of body image by school.
Special diets were used by 7.7% of boys and
20.4% of girls (*2 15.40, P = 0.0040). Losing
weight was the reason given for the diet by 27.8%
of boys and 62.2% of girls, while 16.7% of boys
and 18.9% of girls reported themselves to be
vegetarians, 22.2% of boys and 2.7% of girls had
food allergies, and 33.3% of boys and 16.2% of
girls were on a diet for other reasons, most of
which were related to sports training. There were
no significant differences in type of special diet
eaten between schools.
196
Soft drink and water
Girls, compared With boys, consumed signific-
antly fewer glasses of soft drinks per week
(Mann-Whitney 4.90, P = 0.0000), with the lowest
intake among girls at the private school (Kruskal^
Walhs 15.41, P = 0.0005), the respective means
being 2.8 (CI 2.2-3.4), 1.8 (CI 1.5-2.1) and 2.7
(2.2-3.2) for the low SES, private and high SES
schools. For boys, the respective means were 3.7
( Q 3.8-4.7), 4.0 ( a 2.9-5.1) and 4.7 (CI 3.6-5.8)
glasses per week. Girls at the private school also
consumed significantly more glasses of water per
week (Kruskal-Wallis 6.97, P = 0.0307), with a
mean of 13.4 (CI 10.2-16.5) in the low SES school,
20.0 (CI 17.0-22.8) in the private school and 15.1
(CI 12.2-17.9) in the high SES school. Overall,
81.6% of girls and 83.8% of boys drank soft drinks
during the week of the survey.
Cigarettes and alcohol
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
Figure 1 shows cigarette use and alcohol consump-
tion by sex and school for students who drank or
smoked during the week of the survey. Overall,
28.6% of boys and 28.4% of girls drank alcohol,
and 24.2% of boys and 22.4% of girls were
smokers. There were no significant differences in
the proportions of drinkers or smokers according to
gender. The mean number of drinks and cigarettes,
derived only from those who smoked or drank,
showed significant differences between schools,
with the private school having the greatest percent-
age of smokers (%2 19.57, P = 0.000) and drinkers
(X2 36.94, P = 0.0000). For smokers, boys on
average smoked 27 cigarettes a week (CI 18.7-
35.7) and girls 28 a week (CI 20.6-35.3). For
drinkers, the mean alcohol consumption for boys
was 9.1 glasses a week (CI 6.2-12.0) and for girls
4.2 glasses a week (CI 3.2-5.2) (Mann-Whitney
2.92, P = 0.0035). Those at the high SES school
had the highest mean consumption of alcohol
(Figure 1) (Kruskal-Wallis 9.21 P = 0.0120).
Responses were classified according to the
recommended safe drinking limits of not more
than two drinks per day for females and not more
than four drinks per day for males (NHMRC,

LowSES Prints rtghSES
LowSES Prwtt HghSES
Fig. 1. Frequency of smokers (a) and drinkers (b) and mean cigarette (c) and alcohol (d) consumption for smokers and drinkers
by school for girls (SI) and boys ( • ) . *P < 0.005 for difference between schools.
1992). The alcohol intake in 10 boys (18% of male
drinkers) and one girl exceeded the safe drinking
guidelines. No information was available to distin-
guish a pattern of regular drinking from binge
drinking at weekends.
Television viewing
Both boys (Kruskal-Wallis 13.07, P = 0.0014)
and girls (Kruskal-Wallis 25.5, P = 0.0000) at the
private school watched significantly fewer hours
of television during the week and girls at the private
school also watched significantly less television at
weekends (Kruskal-Wallis 10.81, P = 0.0045).
The mean number of hours of television per day
during the school week was 3.4 (CI 2.6-4.2) for
girls and 3.7 (CI 2.9-4.5) for boys at the low SES
school, 1.5 ( O 1.1-2.0) for girls and 2.2 (CI 1.7-
2.6) for boys at the private school, and 2.7 (CI
2.3-3.1) for girls and 3.8 (CI 2.9, 4.7) for boys at
Nutritional knowledge and behaviour in adolescents
LowSES Private HghSES
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
LowSES Private rtghSES
the high SES school. For weekend television view-
ing the mean number of hours per day was 4.5
(CI 3.7-5.4) for girls and 5.0 (CI 3.9-6.0) for boys
at the low SES school, 3.1 (CI 2.6-3.7) for girls
and 4.9 (CI 4.0-5.9) for boys at the private school,
and 4.4 (CI 3.8-4.9) for girls and 6.3 (CI 5.0-7.6)
for boys at the high SES school. Girls watched a
mean of 2.6 h/day (CI 2.3-2.9) during the week
and 4.1 h/day (CI 3.7-4.4) at weekends. The
respective means for boys were 3.3 h/day (CI
2.8-3.9) and 5.6 h/day (CI 4.9-6.4). Differences
between boys and girls were significant for both
weekday and weekend viewing (Mann-Whitney
2.62, P = 0.0081; 3.65, P = 0.0001).
Television watching during the week correlated
negatively with nutrition knowledge scores (r =
-0.1170, P = 0.028). There were no significant
correlations between television watching and BMI,
nor between TV watching and fat score or food
variety score.
197
 D.Gracey et al.

TbMe VIL Regression models with measures of eating habits as dependent variables

Dependent variable Independent variables Regression Standard error P

coefficient (b) of*

Fat score* age -0.4678 0.1965 0.0179

(R2 = 0.2205) private school -0.5758 03429 0.0941
high SES school -0.5037 03025 0.0969
male gender 0.9168 0.2485 0.0003
self-efficacy -0.0430 0.0074 0.0000
influence over food bought at home -0.2473 0.1193 0.0390
drinker 0.6217 0.2810 0.0277
weekend TV 0.0614 0.0314 0.0513
constant 15.032
Food variety score age -0.5106 0.4575 0.1618
(R2 = 0.1163) private school 1.1172 0.6345 0.0791
high SES school 1.4162 0.5558 0.0112
male gender 0.8324 0.4575 0.0697
self-efficacy 0.0343 0.0139 0.0143
nurition knowledge 0.2595 0.1275 0.0426
weight control value 03765 0.1508 0.0130
influence over foods bought at home 0.5544 0.2187 0.0116
constant 19.978

Downloaded from her.oxfordjournals.org by guest on May 26, 2011

•Fat score according to Kinlay et aL (1991).

Regression models ficantly higher food variety score in students from

Kinlay's fat score or the food variety score were
used as dependent variables in linear regression
models and variables showing significant univariate
relationships with these variables were entered in
the model. All variables were continuous except
for gender (girls = 0, boys = 1), drinking and
smoking (0 = abstainer, 1 = drinker or smoker),
and two dummy variables which were used to
identify the private school and the high SES school
in relation to the low SES school. All models
controlled for gender, age and school.
Table VII shows that the fat intake score related
negatively to both self-efficacy and to a high
level of control over foods bought at home, and
positively to television watching, drinking alcohol
and being male. Nutrition knowledge score was
not significant in this regression model. With food
variety score as the dependent variable there were
significant positive relationships with self-efficacy,
nutrition knowledge, considering weight control as
an important health value, having control over
foods bought at home and male gender. In addition,
dummy variables indicating school showed a signi-
the high SES school compared with the low SES
school.
Discussion
This study of almost 400 15-year-old students
has provided results which have implications for
designing health promotion programs for adoles-
cents. Weight control, improving appearance,
lowering cholesterol, increasing energy, feeling
good and improving health were all considered to
be important health values. Perceived barriers to
change related to availability of healthy food at
home and in the school canteen and lack of control
over food at home. Ignorance about nutrient content
of foods was recognized as a barrier to change,
especially by girls, and the reality of this concern
was apparent in responses to nutrition knowledge
questions.
Perceived self-efficacy, i.e. the belief that an
individual can carry out a specific behaviour
(Bandura, 1977; Kingery, 1990), was strongly
related to healthy eating patterns. This finding is

198

consistent with reports showing self-efficacy to be
a key factor in the eating behaviour of middle
aged women (Shannon et aL, 1990). We found
that boys had significantly lower self-efficacy
scores than girls, in line with the report of Sallis
et aL (1988). Conversely, Kingery (1990) found
no differences in self-efficacy associated with gen-
der in his study of college students. Self-efficacy
did not differ according to SES in our study, again
consistent with the findings of Sallis et aL (1988).
Using the same locus of control questions as
Falconer et aL (1993), we obtained similar findings.
However, self-efficacy was a better predictor of
eating habits than locus of control as previously
reported (Kaplan, 1984).
The health-belief model suggests an association
between dietary behaviour and an individual's
perception of risk (Thomas, 1994) and has been
used to measure beliefs and values influencing
healthy behaviour patterns (Contento and Murphy,
1990). Similarly, the 'theory of reasoned action'
suggests that attitudes can be predicted from beliefs
and evaluative beliefs (Kristal et aL, 1990). Girls
ranked many health beliefs and values items as
more important than did the boys, particularly in
relation to weight control. This is consistent with
other Australian studies reporting the importance
of weight control to girls (Falconer et aL, 1993).
In regression models a high ranking for the health
values of weight control and feeling good about
oneself were significant predictors of a healthy
eating pattern suggesting possible targets for modi-
fying dietary behaviour in adolescents. However,
given concerns about anorexia in young women,
care should be taken in such a strategy.
Barriers to change must also be considered in
planning health education programs. The lack of
healthy food at home and in the school canteen
and ignorance of the nutrient content of various
foods were highly ranked as reasons preventing
students from eating a healthy diet. Girls ranked
barriers to change more highly than did boys and
identified some different barriers they considered
important. Concerns of females over the time
involved in planning a healthy meal has been
reported previously (Urban et al., 1992) and may
Nutritional knowledge and behaviour in adolescents
reflect their more realistic views of meal prepara-
tion. Greater understanding of the benefits of a
healthy diet and skills in preparing quick, easy,
nutritional meals may overcome the perceived
barrier of planning time. In regression models,
control over foods bought at home was a significant
independent predictor of healthy eating patterns
and more girls than boys were able to influence
what they ate, particularly relating to food prepara-
tion at home, choosing food served and choosing
school lunch. Encouraging boys to take part in
these processes may improve their food choices
and improved availability of healthy foods both at
home and at school is needed.
Students' self-perception of ignorance about
nutrient content of foods was supported by
responses to nutrition knowledge questions, par-
ticularly concerning fat content Serious miscon-
ceptions indicated difficulties in translating nutrient
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
advice into food choices. For example, white bread
was considered to be lower in fibre than meat and
boiled sweets were believed to be higher in fat
than magarine. Although nutritional knowledge
and eating patterns are not strongly correlated
(Carmody, 1987), knowledge is a predisposing
factor for eating behaviour (Thomas, 1994). In the
present study, knowledge score was a significant
predictor of the variety of "foods consumed.
Health-related behaviour is influenced by an
individual's readiness to change. Although most
students in the present study would be classified
as being in the maintenance stage (Prochaska and
DiClemente, 1983), a minority, particularly boys
and students at the high SES school were 'pre-
contemplators', i.e. they did not intend to adopt a
healthy diet Although this response may merely
reflect 'healthy' adolescent rebelliousness, further
information is needed about this group before
appropriate nutrition education material can be
designed to focus on their needs.
Interpretation of behavioural factors relating to
a healthy diet depends upon assessing eating habits.
In our study, a food variety and fat score were
used instead of diet records or food frequency
methods to keep the questionnaire of manageable
length. The food variety score has been validated
199
 D.Gracey et aL
in primary school children in Perth by comparing
results with diet records in the same subjects
(Milligan, 1994). In other populations, Ries and
Dachler (1986) have shown it to be of value in
dietary assessment The food variety score provides
some information about foods actually eaten and
targets the first of the national nutritional guidelines
for Australians, i.e. to enjoy a wide variety of
nutritious foods. Kinlay et al (1991) have validated
their fat score against food frequency methods and
recommend this technique as useful in population-
based studies.
The fat score in the present study was higher in
boys than girls and in students who considered
their food choices unhealthy. Drinkers and smokers
also had higher scores for fat intake which is
consistent with the reported association between a
high fat, low fibre diet, smoking (Fulton et aL,
1988) and drinking (Herbeth et aL, 1988). Less
healthy food choices in males are consistent with
the finding that women value health more highly
than men and more readily adopt preventive health
strategies (Kristiansen, 1990). Women also acquire
knowledge about healthy eating in response to a
nutrition education program more satisfactorily
than men (Finnegan et al., 1990). Gender differ-
ences in beliefs and behaviours need to be consid-
ered in designing health promotion programs.
The lower SES of the school was associated
with unhealthy behaviours, which is similar to
previous reports (Adler et aL, 1993). In the present
study, food variety score was lower at the low SES
school, especially for cereals and fruit, where
almost half of these girls reported that they did
not eat breakfast, a behaviour associated with
higher cholesterol (Resnicow, 1991) as well as
with nutritional deficiencies (Morgan et al, 1986)
and impaired academic performance (Simeon and
Grantham-McGregor, 1989). Less variety in foods
eaten is associated with lower SES and may reflect
disposable income. However, increased cardiovas-
cular risk in lower socio-economic groups (Alder,
1993) may also be related to other behaviour
patterns.
Smoking and drinking were more common
among private school students, but mean consump-
200
tion was lower than at the high SES school. The
NHMRC dietary guidelines (1993) recommend
water as the preferred drink for children. In our
study, girls at the private school conformed to this
guideline drinking the most water and little soft
drink. The high consumption of soft drinks by
other students suggests this is an area which could
be targeted for intervention. Greater intake of nuts,
less salt use, less consumption of soft drinks and
greater consumption of water in private school
students suggest that their dietary patterns are more
consistent with national dietary guidelines.
There is an increasing trend toward adolescent
dieting (Gortmaker et al, 1987), raising concerns
about developmental delay and increased incidence
of eating disorders. Although the student popula-
tion studied here conformed to previously reported
normal values for BMI (Hammer et al, 1991),
over half of the girls and 20% of boys considered
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
themselves overweight and 13% of girls were on
a weight-reducing diet These results are similar
to those among Californian school girls reported
by Huenemann et al. (1968), in which 46% wanted
to lose weight It is of particular concern that about
21% of girls and 8% of boys in the lowest BMI
quartile thought they were too fat
Twenty four percent of boys and 22% of girls
were self-reported smokers compared with 25% of
boys and 28% of girls among 16 year olds in a
nation-wide survey of Australian secondary school
students (Hill et aL, 1993). Hill et aL (1993) also
found boys to be heavier smokers than girls (mean
of 38 cigarettes per week versus 28 cigarettes per
week). In our survey the mean number of cigarettes
per week for smokers was similar for boys and
girls, 27 for boys and 28 for girls. The proportion
of students drinking alcohol was lower than
reported for 16 year olds by Hill et aL (1993) (44
versus 28%) but the mean alcohol consumption
was similar in both surveys. A significant positive
association was found between alcohol consump-
tion and smoking for both boys and girls, as
previously reported (NHMRC, 1993). It is dis-
turbing to find that almost one-fifth of drinkers
reported alcohol consumption above the safe
drinking guidelines of not more than two drinks

per day for females and not more than four drinks
per day for males (NHMRC, 1992). The pattern
of binge drinking common in young adults
(Wechsler and Isaac, 1992) places them at further
risk in the short term in relation to traffic accidents,
violence and other social problems. In the longer
term, cardiovascular risk is increased when associ-
ated with higher blood pressure (Wannamethee
and Shaper, 1991) and lower HDL cholesterol
(Gruchow et al, 1982). If health-related behaviour
is to be improved in this age group, smoking and
at-risk drinking must be targeted.
Television watching has also been linked with
an unhealthy diet, high cholesterol levels (Hei
et al, 1990) and obesity (Dietz and Gortmaker,
1985). This may be influenced by unhealthy nutri-
tion messages in commercials (Lank et al., 1992),
eating snack foods and decreased physical activity
(Robinson et al, 1993). In our study, television
watching was a significant independent predictor
of fat score in regression. We found no association
between television watching and BMI, similar to
the findings of Robinson et al. (1993). Less healthy
food choices associated with more hours of televi-
sion watching in the present study and fewer
leisure-time hours available for exercise suggest
that cardiovascular risk is likely to increase with
the duration of watching television.
Our findings indicate that the priorities for
nutritional health promotion amongst adolescents
should involve increasing perceived self-efficacy
to choose healthy foods (Kingery, 1990) given that
behaviour change can be achieved by strategies
to improve self-efficacy (Strecher et al, 1986).
Campaigns should focus on the health beliefs and
values considered most important, mainly relating
to improving health, increasing energy and feeling
good about oneself as well as to appearance and
weight control. However, priorities should be modi-
fied according to gender, particularly with regard
to weight control which girls ranked more highly.
Family involvement is also needed to ensure
that healthy foods are available at home and to
encourage participation in food buying, preparation
and serving, particularly for boys. In a controlled
trial in 10- to 12-year-old children, we have found
Nutritional knowledge and behaviour in adolescents
improved dietary practices with involvement of the
family in nutrition education programs compared to
entirely school-based programs (Vandongen et al,
1995). However, these programs, designed for
primary school children, are not immediately
applicable to older students. One possible approach
is the use of mailed material directed at educating
parents which Crockett et al (1989) have shown
to be effective in improving knowledge about
nutrition, modifying diet-related behaviours, invol-
ving children in nutritional activities at home and
influencing foods available at home. Healthy foods
should also be available at school canteens and
several schools are attempting to address this need
through parent-teacher groups. There is a need for
more widely accessible information applicable to
allowing healthy food choices in school canteens.
For example, in Western Australia, the recently
formed West Australian Canteen Association is
providing resources and support to facilitate
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
changes in school canteens.
Students need education about the nutritional
content of food, particularly fat, in order to make
appropriate food choices. Nutrition education in
schools is usually directed at younger children and,
given the academic pressures on older high-school
students, it is unlikely that teaching time would be
dedicated to this need. The students' lack of
knowledge about foods rather than nutrients sug-
gest that foods should be a focus for nutrition
education programs implemented in primary
schools. Health promotion materials to reinforce
these messages and to give older students an easily
available source of reference to check their food
choices may also be helpful. Smoking and at-risk
drinking are targeted in Western Australia, as in
many other areas, by government-funded programs,
and should be included in any strategy to improve
health-related behaviours in adolescents. Such a
global approach becomes particularly important in
view of the associations between smoking, drinking
and dietary behaviour reported previously (Herbeth
et al, 1988; Hebert and Kabat, 1990) and found
in the present study. Media campaigns have been
widely used to target smoking and drinking behavi-
our in teenagers and consideration should be given
201
D.Gracey et al.
to developing media campaigns publicising nutri-
tional messages appropriate to this age group.
The present study has shown trends in nutrition-
related behaviour as well as smoking, drinking and
television watching among senior high school
students in the Perth metropolitan area. In order
to devise appropriate health promotion programs,
educators need to understand these patterns of
behaviour as well as understanding the attitudes,
beliefs and values underlying the health-related
choices of adolescents. Our findings suggest the
need for a re-evaluation of health promotion pro-
grams for high school students, including assess-
ment of the place of media campaigns, while taking
account of differences related to gender and socio-
economic status at this critical age of behavioural
and physical development.
Acknowledgements
We thank the Year 11 co-ordinators who facilitated
this study and the students who took part The
study was supported by a Program Grant from the
Australian National Health and Medical Research
Council. We are grateful to Dr Stephen Houghton
for helpful advice in preparing the manuscript
References
AdlerJ^J.E., Thomas-Boyce.W., ChenseyJvt.A., Folkman,S. and
Leonard Syme,S. (1993) Socio-economic inequalities in
health; no easy solution. Journal of the American Medical
Association, 269, 3140-3145.
AjzenJ. (1985) From intentions to actions: a theory of planned
behaviour. In: KuhlJ. and BeckmannJ. (eds), Action-Control:
From Cognition to Behaviour. Springer, New York, pp. 11-39.
AjzenJ. and Madden.TJ. (1986) Prediction of goal-directed
behaviour attitudes, intentions, and perceived behavioural
control. Journal of Experimental Social Psychology, 22,
453-474.
Bandura^A. (1977) Self-efficacy: toward a unifying theory of
behavioural change. Psychological Review, 84, 191-215.
Bandura^V (1986) Social Foundations of Thought and Action.
Prentice-Hall, Englewood Cliffs, NJ.
Carmody.T.P., MatarazzoJ.D. and IstvanJ^A. (1987) Promoting
adherence to heart-healthy diets: a review of the literature.
Journal of Compliance in Health Care, 2, 105-117.
Coate»,TJ., PetersenA-C. and Perry.C. (eds) (1982) Promoting
Adolescent Health. A Dialog on Research and Practice.
Academic Press, New York.
202
ContentoJ.R. and Murphy.B.M. (1990) Psychosocial factors
differentiating people who reported making desirable changes
in their diets from those who did not. Journal of Nutrition
Education, 22, 6-14.
CrocketLSJ., Mullisjt., Perry.C.L. and LuepkerJtV. (1989)
Parent education in youth-directed nutrition interventions.
Preventive Medicine. 18, 475-491.
Dietz,W. and Gortmaker,S. (1985) Do we fatten our children
at the television set? Obesity and viewing in children and
adolescents. Pediatrics, 73, 307-312.
FalconerJl., Baghurst,KJ. and Rump.E.R (1993) Nutrient
intakes in relation to health-related aspects of personality.
Journal of Nutrition Education. 25, 307-319.
FinneganJ.R., Viswanath.K., RconeyJJ., McGoverruP.,
Baxter, J., Elmer, P., Graves, K., Hertog, } . , Mullis, R., Pirie,
P., Trenkner, L. and Potter, J. (1990) Predictors of knowledge
about healthy eating in a rural midwestem US city. Health
Education Research, 5, 421-431.
Ftshbein.M. and AjzenJ. (1975) Belief, Attitude, Intention
and Behaviour: An Introduction to Theory and Research.
Addison-Wesley, Reading, MA.
Fulton,M., Thompson.M, Elton,R.A., Brown,S., Wood.D.A.
and 01iver.M.F. (1988) Cigarette smoking, social class
and nutrient intake. Relevance to coronary heart disease.
European Journal of Clinical Nutrition, 42, 797-803.
GlanzJC. KrUtal,A.R., Sorenson.C Palombo.R., HeimendingerJ.
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
and ProberuC. (1993) Development and validation of measures
of psychosocial factors influencing fat- and fibre-related
dietary behaviour. Preventive Medicine, 22. 373-387.
GlanzX, PattersoruR.E., Kristal,A.R., DiClemente.C.C,
HeimendingerJ. Linnan.L. and McLerran.D.F. (1994) Stages
of change in adopting healthy diets: fat, fibre, and correlates
of nutrient intake. Health Education Quarterly. 21, 499-519.
GliksmanJvl.D., Dwyer.T. and Boulton.TJ. (1987) Primary
prevention of coronary heart disease: should intervention
commence in childhood? Medical Journal of Australia. 146,
360-361
GliksmanJvl.D.. LazarusJ*. and Wilson^. (1993) Differences
in serum lipids in Australian children: is diet responsible?
International Journal of Epidemiology, 22, 247—254
Gortmaker^.. Dietz,W.H., Sobolj\.N. and Wehler.C. (1987)
Increasing pediatric obesity in the US. American Journal of
Diseases of Children, 14, 535-540.
Greene.G.E., Rossi,S.R., Reed,G.R., Willey.C. and Prochaska,
JO. (1994) Stages of change for reducing dietary fat to
30% of energy or less. Journal of the American Dietetic
Association. 94. 105-110.
Gruchow.H.W., Hoffman.R-G., Anderson^AJ. and Barboriak,
JJ. (1982) Effects of drinking patterns on the relationship
between alcohol and coronary occlusion. Atherosclerosis, 43,
393-404.
HammerJ^D.. Kraemer.H.C, Wilson.D.M.. Ritter.P.L. »nd
Dombusch.S.M. (1991) Standardized percentile curves of
body-mass index for children and adolescents. American
Journal of Diseases of Children, 145, 259-263.
HebertJ.R. and Kabat,G.C. (1990) Differences in dietary intake
associated with smoking status. European Journal of Clinical
Nutrition, 44, 185-193.
Hei.T, GolAK., QaqundahJ5. and Wong.N. (1990) A com-
parison of television viewing habits and family history of

myocardial infarction as cholesterol screening indicators in
children. Circulation (Supplement III). 82, 111-227.
HerbethJ3., Didelot-BarthelemyX-. Lemoine,A. and Le
DevehatX. (1988) Dietary behaviour of French men
according to alcohol drinking pattern. Journal of Studies on
Alcohol 49, 268-272.
HiUDJ., White.VM., Williams,R.M. and Gardner.GJ. (1993)
Tobacco and alcohol use among Australian secondary school
students in 1990. Medical Journal of Australia, 158,228-234.
Huenemann,R.L., HamptoruM.C, ShapiroJLR. and
Bekuclce.A.R. (1968) Adolescent food practices associated
with obesity. Federation Proceedings, 25, 4-10.
Kaplan,R.M., Atkins,CJ. and Reinsch,S. (1984) Specific
efficacy expectations mediate exercise compliance in patients
with COPD. Health Psychology, 3, 223-242.
Kelder.S.H., Perry.C.L, Klepp.KX and LytleXX. (1994)
Longitudinal tracking of adolescent smoking, physical
activity and food choice behaviours. American Journal of
Public Health, 84, 1121-1126.
KingeryJ>M. (1990) Self-efficacy ind the self-monitoring of
selected exercise and eating behaviours of college students.
Health Education, 21, 26-29.
Kinlay.S., Heller,IU\ and HallidayJJV. (1991) A simple score
and questionnaire to measure group changes in dietary fat
intake. Preventive Medicine, 20, 378-388.
Knutsen,S. and Knutsen,R. The Tromso survey: the family
intervention study—die effect of intervention on some
coronary risk factors and dietary habits, a six year follow-
up. Preventive Medicine, 20, 197-212.
Kristal,A.K., Bowen,DJ., Curry,SJ., Shattuck^.L. and
Henry.HJ. (1990) Nutrition knowledge, attitudes and
perceived norms as correlates of selecting low-fat diets.
Health Education Research, 5, 467-477.
Krisu'ansen,CM. (1990) The role of values in the relation
between gender and health behaviour. Social Behaviour, 5
127-133.
LankJ>J.H.. Vickery.C.E., CotugruuN. and Shade.D.D. (1992)
Food commercials during' television soap operas: What is
the nutrition message? Journal of Community Health, 17,
377-384.
McLennan,W. (1990) Socio-economic Indexes for Areas.
Australia Bureau of Statistics. Catalogue No. 1356.0.
Milligan.R.A.JC (1994) Food variety score as a method of
dietary assessment in eleven year old children. Master of
Public Health Thesis, Curtin University of Technology, Perth,
Western Australia.
Morgan.KJ., Zabik>1.E. and Stampley.G.L. (1986) The role
of breakfast in diet adequacy of US adult population. Journal
of the American College of Nutrition, 5, 551-563.
Nada-Raja.S., McGee,R. and Williams^. (1994) Health beliefs
among New Zealand aolescents. Journal of Paediatrics and
Child Health, 30, 523-529.
National Health and Medical Research Council (1992) Is There
a Safe Level of Daily Consumption of Alcohol for Men and
Women? Australian Government Publishing Service,
Canberra.
National Health and Medical Research Council: Panel to
Review the Dietary Guidelines for Australians (1993) Dietary
Guidelines for Children. National Health and Medical
Research Council, Canberra.
Nutritional knowledge and behaviour in adolescents
ProchaskaJ.O. and DiClementcC.C. (1983) Stages and process
of self-change in smoking: Toward an integrative model of
change. Journal of Consulting and Clinical Psychology. 5,
390-395.
ProchaskaJ.O. and DiClementcCC. (1984) Self change
processes, self-efficacy and decisional balance across five
stages of smoking cessation. In EngstromJ'., Anderson J".N.
and MortensonX-E. (eds). Advances in Cancer Control:
Epidemiology and Research. Lits, New York.
ProchaskaJ.O. and DiClemente.C.C. (1992) Stages of change
in the modification of problem behaviours. In HersonX.,
Eisler,R.M. and Millerj^M. (eds). Progress in Behaviour
Modification. Brooks/Cole Publishing, Pacific Grove, CA;
vol. 28.
ResnicowJC (1991) The relationship between breakfast habits
and plasma cholesterol levels in school children. Journal of
School Health, 61,81-85.
Ries.CP. and Daehler,M.S. (1986) Evaluation of the nutrient
guide as a dietary assessment tool. Journal of the American
Dietetic Association, 86 228-233
Robinson.T.N., HammerX-D., KillenJX)., KraemerJH.C,
WilsonD.M., Hayward,C. and Taylor.C.B. (1993) Does
television viewing increase obesity and reduce physical
activity? Cross-sectional and longitudinal analyses among
adolescent girls. Pediatrics, 91 273-280.
RosenstockJ-M. (1966) Why people use health services.
Downloaded from her.oxfordjournals.org by guest on May 26, 2011
Milbank Memorial Fund Quarterly, 44, 94.
SallisJ.F., PiiukiJt.B^ GrossmanJt.M., Patterson.T.L. and
NaderJ?.R. (1988) The development of self-efficacy scales for
health-related diet and exercise behaviours. Health Education
Research, 3 283-292.
Shannon^., Bagbyjl., WangJvl.Q. and TrenkerX. (1990) Self-
efficacy: a contributor to the explanation of eating behaviour.
Health Education Research, 5 395-407.
SimeonJ5.T. and Grantham-McGregor.S. (1989) Effects of
missing breakfast on the cognitive functions of school
children of differing nutritional status. American Journal of
Clinical Nutrition, 49 646-653.
Strecher.VJ., De VelIis,B.M., BeckerJvI.H. and RosenstockXM.
(1986) The role of self-efficacy in achieving health behaviour
change. Health Education Quarteriy, 12, 73-92.
Strickland.B.R. (1978) Internal-external expectancies and
health-related behaviours. Journal of Consulting and Clinical
Psychology. 46, 1192-1211.
Sweeting,H. and AndersoruA. (1994) Sorio-demographic
correlates of dietary habits in mid to late adolescence.
European Journal of Clinical Nutrition, 48, 736-748.
Taylor,S.E. (1986) Health Psychology. McGraw-Hill, New
York.
ThomasJ. (1994) New approaches to achieving dietary change.
Current Opinion in LJpidology, 5, 36-41.
TruswelU.S. and Damton-Hill.I. (1981) Food habits of
adolescents. Nutrition Reviews, 39, 73-88.
Urbanjv)., White.E, Anderson.G., Curry,S. and Kristal,A.
(1992) Correlates of maintenance of a low-fat diet among
women in the Women's Health Trial. Preventive Medicine.
21, 279-291.
Vandongen,R., JennerJi.A., Thompson.C, Taggart^.C,
SpicketUi.R, Burke,V., BeilinXJ., MilIigan,Rj\. and
DunbarJXL. (1995) A controlled evaluation of a fitness and
203
 nutrition intervention program on cardiovascular health in
10- to 12-year-old children. Preventive Medicine. 24, 9-22.
Vandongen.R., Mori.TA., Burke.V., BeilinXJ-, MorrisJ. and
RitchieJ. (1993) Effects on blood pressure of co3 fats in
subjects at increased risk of cardiovascular disease.
Hypertension, 22, 371-379.
VeenstraJ., SchenkeUA.A., van Erp-BaaruA.MJ., Brants,
H.AJvt., HulsbofJCFA.M., Kistemaker.C, Schaafsma,G.
and Ockhuizen,Th. (1993) Alcohol consumption in relation
to food intake and smoking habits in the Dutch National
Food Consumption Survey. European Journal of Clinical
Nutrition, 47, 482-489.
WallstoivB.S. and WaJlston.lCA. (1984) Social psychological
modeb of health behaviour An examination and integration.
204
In Baum,A., Taylor^.E. and SingerJ.E (eds). Handbook of
Psychology and Health. Erlbaum, Hillsdale. NJ, pp. 23-54.
Wannamethee.G. and Shaper.A.G. (1991) Alcohol intake and
variations in blood pressure by day of examination. Journal
of Human Hypertension, 5 59-67.
Wechsler.H. and Isaacs^!. (1992) 'Binge' drinkers at
Massachussets colleges. Journal of the American Medical
Association, 267, 2929-2931
WiIliams,H.M., Woodward,D.R., BallJU., CummingJJ.,
Homsby,H. and BoonJ-A. (1993) Food perceptions and
food consumption among Tasmanian high school students.
Australian Journal of Nutrition and Dietetics. 50, 156-162.
Received on May 16. 1995; accepted on September 10, 1995
Downloaded from her.oxfordjournals.org by guest on May 26, 2011