See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8187966

Predictors of Health Behaviors in College Students

Article  in  Journal of Advanced Nursing · January 2005

DOI: 10.1111/j.1365-2648.2004.03229.x · Source: PubMed

CITATIONS READS

232 6,383

5 authors, including:

Diane M Von Ah Sheryl Ebert

Indiana University-Purdue University Indianapolis Memphis VA Medical Center
66 PUBLICATIONS   1,283 CITATIONS    3 PUBLICATIONS   362 CITATIONS   

SEE PROFILE SEE PROFILE

Anchalee Ngamvitroj Chutitorn Na-Jin Park

Thai Red Cross College of Nursing University of Pittsburgh
3 PUBLICATIONS   298 CITATIONS    29 PUBLICATIONS   472 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Child and maternal factors that influence child blood pressure in preschool children View project

Long-term quality of life in breast cancer survivors and partners View project

All content following this page was uploaded by Diane M Von Ah on 18 January 2018.

The user has requested enhancement of the downloaded file.

I S S U E S A N D IN N O V A T I O N S I N N U R S I N G P R A C T I C E

Predictors of health behaviours in college students

Diane Von Ah PhD RN
Assistant Professor, School of Nursing, University of Louisville, Louisville, Kentucky, USA

Sheryl Ebert MA
PhD Candidate, Department of Psychology, University of Alabama at Birmingham, Birmingham, Alabama, USA

Anchalee Ngamvitroj MSN RN

PhD Candidate, School of Nursing, University of Alabama at Birmingham, Birmingham, Alabama, USA

Najin Park MSN RN

PhD Candidate, School of Nursing, University of Alabama at Birmingham, Birmingham, Alabama, USA

Duck-Hee Kang PhD RN

Associate Professor, School of Nursing, University of Alabama at Birmingham, Birmingham, Alabama, USA

Submitted for publication 20 November 2003

Accepted for publication 7 June 2004

Correspondence: VON AH D., EBERT S., NGAMVITROJ A., PARK N. & KANG D.-H. (2004)
Diane Von Ah, Journal of Advanced Nursing 48(5), 463–474
School of Nursing, Predictors of health behaviours in college students
University of Louisville,
Aim. This paper reports a study examining the direct effects of perceived stress,
Louisville,
perceived availability of and satisfaction with social support, and self-efficacy, and
Kentucky 40292,
USA. examines the intermediary roles of perceived threat (perceived susceptibility ·
E-mail: d0vona01@louisville.edu perceived severity), benefits, and barriers on alcohol behaviour, smoking behaviour,
physical activity and nutrition behaviour, general safety behaviour and sun-pro-
tective behaviour in college students.
Background. Health behaviours formed during young adulthood may have a sus-
taining impact on health across later life. Entering college can be an exciting, yet
stressful event for many adolescents and young adults as they face trying to adapt to
changes in academic workloads, support networks, and their new environment.
Coupled with these changes and new-found responsibilities, they have greater
freedom and control over their lifestyles than ever before. However, researchers
have shown globally that many college students engage in various risky health
behaviours.
Method. A cross-sectional sample of 161 college students enrolled in an introduc-
tory psychology course completed self-report questionnaires regarding stress; social
support; self-efficacy; and components of the Health Belief Model including per-
ceived threat, perceived benefits, perceived barriers; and common health behaviours.
Step-wise multiple regression analysis was conducted and significant predictors were
retained as modifiers in the path analysis.
Findings. Self-efficacy significantly predicted alcohol and smoking behaviour,
physical activity and nutrition protective behaviour, general safety protective
behaviour and sun-protective behaviour. Under high-perceived threat, self-efficacy
was mediated by perceived barriers for binge drinking and moderated by perceived
barriers for physical activity and nutrition behaviours. In addition, under

2004 Blackwell Publishing Ltd 463

D. Von Ah et al.

high-perceived threat, self-efficacy was moderated by perceived threat for alcohol

use at 30 days and 6 months. Under low threat, self-efficacy was mediated by
perceived barriers for smoking behaviour and general safety protective behaviours.
Conclusions. Future health promotion programmes with college students must use
interventions that maximize self-efficacy and ultimately reduce barriers to adopting
a healthy lifestyle.

Keywords: stress, social support, self-efficacy, health behaviours, health belief

model, nurse

has also been associated with other negative health behav-

Introduction
Entering college can be an exciting, yet stressful event for
many adolescents and young adults. Traditional college
students enter college immediately after high school (age 18
or 19 years) and are faced with trying to adapt to changes in
academic workloads, support networks, and their new
environment. Coupled with these changes and new-found
responsibilities, college students have greater freedom and
control over their lifestyles than ever before. Thus, this
transitional period is an opportune time to establish healthy
lifestyle behaviours (Dinger & Waigandt 1997). However,
researchers have shown globally that many college students
engage in various risky health behaviours, including alcohol
use, tobacco use, physical inactivity and unhealthy dietary
practices, ignore preventive safety habits such as wearing
helmets, seat belts and/or condoms, and engage in excessive
sun exposure, which may have long-term implications for
their health (Centers for Disease Control 1997, Steptoe &
Wardle 2001, Steptoe et al. 2002). Therefore, identifying
factors that influence health protective behaviours in college
students warrant further attention. Nurses, who often play a
vital role in developing health promotion and preven-
tion programmes, could then use this information to develop
and/or enhance programmes targeted at college students.
Excessive alcohol use among college students is a wide-
spread problem on many college campuses (Wechsler et al.
1998). Researchers have reported that approximately 63% of
Western European university students and 70% of Eastern
European university students categorized themselves as
occasional and/or regular drinkers (Steptoe & Wardle
2001). Recent United States of America (USA) studies
indicate the prevalence of binge drinking (drinking five or
more alcoholic drinks in one sitting) to be between 37Æ5%
(Presley et al. 1995) and 44% (Wechsler et al. 1998) among
college students, and many students report increased drinking
while attending college (Meilman et al. 1989). Binge drinking
iours including smoking (Jones et al. 2001), risky sexual
behaviour (Ichiyama & Kruse 1998) having multiple sexual
partners (Wechsler et al. 1995), injuries, and inadequate
seatbelt use (Everett et al. 1999). Smoking is also prevalent
among college students worldwide. Approximately, one
quarter of all college students in the USA smoke (Martinelli
1999), and 75% of those continue to smoke into later
adulthood (Flay 1993), placing them at greater risk of lung
and cardiovascular diseases. Similar trends were noted in
Western and Eastern European university students, with
22Æ9% and 19Æ8% reporting being regular smokers, respect-
ively (Steptoe & Wardle 2001). In addition, college students
universally often fail to meet current physical activity and/or
dietary intake recommendations (Centers for Disease Control
1997, Dinger & Waigandt 1997, Dinger 1999, Steptoe et al.
2002). In fact, the highest rate of decline in physical activity
occurs in the early adulthood period between 18 and 24 years
of age (US Department of Health and Human Services 2000).
This occurs despite the fact that physical inactivity and poor
dietary intake patterns are leading causes of death in the USA,
accounting for at least 300,000 deaths per year (McGinnis &
Foege 1993). This failure of college students to engage in
healthier lifestyle practices is not unique to the USA. Steptoe
et al. (2002) surveyed university students from 13 European
countries in 1990 regarding smoking, physical exercise, fruit
and fat intake, beliefs in the importance of behaviours for
health, and awareness of the influence of behaviours on heart
disease and repeated the study in 2000. Remarkably, smoking
prevalence increased and fruit consumption significantly
decreased during this period, while physical exercise and fat
intake remained stable. In addition, these authors reported
that changes in prevalence of behaviour were not correlated
with changes in risk awareness (Steptoe et al. 2002).
Similarly, in other studies knowledge about AIDS was not
translated into safe sex practices (Shapiro et al. 1999), nor
did knowledge about sun exposure curtail tanning, which is

464
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Issues and innovations in nursing practice
viewed as more attractive and even healthy (Robinson et al.
1997). Therefore, this study was designed to delineate
factors, which significantly contribute to health behaviours
in college students.
Background
Stress and health behaviours
Stress is known to influence health through its direct physio-
logical effect and its indirect effect via altered health behav-
iours. Stress occurs when a person appraises a situation
demand and/or challenge as exceeding available coping
resources (Lazarus & Folkman 1984). Researchers have
shown that stress among college students can have detrimental
effects on both academic performance and health (Campbell
& Svenson 1992).
College students, because of the transitional nature of
college life, are particularly prone to stress, and the majority
experience stress because of varying academic commitments,
financial pressures, and lack of time management skills
(Misra 2000). Misra found that students in their first year of
college were particularly vulnerable to stress because of the
inherent conflict and frustration of managing new responsi-
bilities and unfamiliar situations. In addition, they often lack
the strong social support networks and coping skills needed
to handle college stress effectively (Allen & Heibert 1991).
Inability to cope with stress has long been associated with
risky health behaviours such as smoking in adolescents and
young adults (Byrne et al. 1995). In fact, many young
smokers justify smoking as a means of dealing with their
stress (Mates & Allison 1992). Similarly, Steptoe et al. (1996)
reported that physical activity decreased significantly for
students undergoing examination stress, indicating its negat-
ive impact on health behaviours.
Social support, on the other hand, has been shown to have a
positive impact on preventive behaviours and health outcomes.
Social support is a multi-dimensional concept that describes
relationships individuals maintain with others. Emerging peer
groups may significantly influence college students’ health
behaviours (Flay et al. 1994). Hubbard et al. (1984) found that
perceived level of social support had a direct positive associ-
ation with participation in positive health practices including
adequate nutrition, exercise, relaxation, safety, substance use,
and health promotion. Conversely, college students with lower
levels of social support smoked significantly more and con-
sumed more alcohol than those with higher levels of social
support under examination stress (Steptoe et al. 1996). These
findings support the positive influence of social support on
healthy lifestyle choices.

2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Predictors of health behaviours in college students
Another important concept in health behaviour is self-
efficacy. Perceived self-efficacy has been increasingly associ-
ated with health behaviour and its change (Stretcher &
Rosenstock 1997). Perceived self-efficacy is an individual’s
belief that they are capable of achieving a goal. Bandura’s
Theory of Self-Efficacy (Bandura 1977) suggests that beha-
viour is better predicted by people’s beliefs in their capabil-
ities to do whatever is needed to succeed than by the
behaviour’s importance.
Martinelli (1999) reported that self-efficacy among college
students was the strongest predictor of health-promoting
behaviour. Self-efficacy for sun protection has been linked to
both intention to use sunscreen (Mahler et al. 1997) and sun
exposure (Reynolds et al. 1996). Increased self-efficacy has
been shown to be an important predictor for quitting
smoking among 18–29 year olds (Kvis et al. 1995). Con-
versely, evidence suggests that low self-efficacy contributes to
maladaptive health behaviour. Skutle (1999) surveyed 203
adult and young adult alcoholic men and found a significant
association between lower self-efficacy scores and experience
of greater psychological benefit from drinking. Likewise, it
has been reported that failure of college students to eat
healthy diets could be, in part, because of decreased self-
efficacy in making healthy food choices (Cusatis & Shannon
1996). Therefore, these findings collectively suggest the
importance of re-examining the effects of psychosocial
factors (stress, social support, and self-efficacy) on health
behaviour of college students.
Health belief model
The health belief model (HBM) is an explanatory model,
which is often used to determine the likelihood of performing
preventive health practices. Four main components of the
model are perceptions of susceptibility, severity, benefits, and
barriers. It is proposed that health-seeking behaviours or
behavioural changes are mediated by an individual’s level of
perceived susceptibility and perceived severity (or perceived
threat, which is the interaction effect of these two compo-
nents), and the belief that the protective-behaviour is bene-
ficial (i.e. perceived benefits outweigh perceived barriers).
In the study reported here, it was hypothesized that perceived
benefits and barriers would be stronger predictors of health-
promoting behaviour when perceived threat was high. In
contrast, when perceived threat was low, these would not be as
salient unless the benefits were seen to be exceptionally high.
Although researchers have evaluated perceived stress, social
support, and self-efficacy on multiple health behaviours, few
have included all these closely-linked variables simultaneously
(Kelly et al. 1991, Steptoe et al. 1996).
465
D. Von Ah et al.

The study Table 1 Demographic composition of sample (n ¼ 161)

Mean (SD ) n (%)

Aim
Age 19Æ7 (4Æ09)
The aims of the study were: Height (inches) 66Æ6 (5Æ79)
• to examine the direct effects of perceived stress, perceived Weight (pounds) 151Æ6 (37Æ11)
Race
availability of and satisfaction with social support, and
White people 70 (44)
self-efficacy on alcohol and smoking behaviour, physical Non-White people 90 (56)
activity/nutritional behaviour, general safety behaviour Missing 1 (<1)
(i.e. condom use, seat-belt use, etc.) and sun-protective Gender
behaviour; Male 43 (27)
• to explore the intermediary roles of perceived threat (per- Female 118 (73)
Marital status
ceived susceptibility · perceived severity), benefits, and
Single 146 (91)
barriers on these health behaviours in college students. Married 11 (7)
Divorced 4 (2)
Religion
Design Catholic 16 (10)
Protestant 31 (19)
A cross-sectional, descriptive design was used.
Buddhist 2 (1)
Muslim 1 (<1)
Other 111 (69)
Participants
Highest level education (mother)
A convenience sample of 161 of 400 (40% response rate) Less than High School 22 (14)
undergraduate students enrolled in an introductory psychol- High School Diploma 39 (24)
Some College 40 (25)
ogy course at the University of Alabama at Birmingham
Bachelor’s Degree 39 (24)
(UAB) volunteered to participate in this study. Data were Master’s Degree 17 (11)
collected on one occasion in mid-semester in October 2001. Doctoral Degree 3 (2)
The participants were 118 females and 43 males with a Highest level education (father)
mean age of 19Æ7 ± 4Æ09 years (typical age of incoming US Less than High School 20 (12)
High School Diploma 46 (29)
college students); 44% White people and 56% non-White
Some College 33 (21)
people. Table 1 presents their demographic information. In Bachelor’s Degree 42 (26)
addition, data were collected on immediate family (i.e. Master’s Degree 16 (10)
parents, siblings, and/or child/children) medical history; Doctoral Degree 4 (2)
cardiovascular disease was the most prevalent, with partic-
A high number of respondents indicated their religion to be ‘other’.
ipants reporting 44Æ1% with hypertension, 24Æ2% with high Given the geographical location of the sample, it is likely that many
cholesterol, and 11Æ8% with heart disease. of these were Baptist.

Perceived availability of and satisfaction with social

Instruments
Perceived stress was measured using the Perceived Stress Scale
(PSS) (Cohen et al. 1983), a 10-item inventory with a rating
scale from 0 (never) to 4 (very often) that measures feelings
and thoughts during the last month. Items were designed to
tap into how unpredictable, uncontrollable, and overloaded
respondents found their lives. The total score (range 0–40)
was used as an index of perceived stress, with a higher score
indicating a higher level. Empirical evidence supports the
validity of this global measure with community samples and
adequate internal and test–retest reliability in college students
(Cohen et al. 1983). The coefficient alpha in our study was
0Æ79.
support was measured by the 6-item Social Support Ques-
tionnaire (SSQ6) (Sarason 1987), which was adapted from
the original 27-item SSQ to reduce subject burden but
maintain equally high reliability and validity (Sarason et al.
1983). Each item has two parts: assessment of the number of
available others (number of perceived availability score) and
assessment of the degree of satisfaction (satisfaction score)
with perceived support in response to particular situations.
Scores represent the mean per item. Internal reliabilities for
the SSQ6 ranged from 0Æ90 to 0Æ93 for both dimensions in a
report by Sarason et al. (1983). In our data, the coefficient
alphas were 0Æ97, 0Æ92 and 0Æ88 for the total scale, availab-
ility of and satisfaction with social support, respectively.

466
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Issues and innovations in nursing practice
Perceived self-efficacy was measured using a 41-item self-
report instrument developed specifically for this study.
Based on Bandura’s Theory of Self-Efficacy (Bandura
1977), items were constructed to measure the respondent’s
confidence in performing health-promoting behaviours
related to alcohol consumption (five items), tobacco use
(four items), physical activity and nutritional intake (13
items), general safety protective behaviours (i.e. helmet,
seatbelt, and condom use, 13 items) and sun-protective
behaviour (six items). Higher scores were indicative of
higher levels of self-efficacy. Subscale scores of self-efficacy
(alcohol, smoking, physical activity/nutrition, general safety,
and sun self-efficacy) were used separately to predict the
respective health behaviours. The coefficient alphas of the
five subscales were 0Æ87, 0Æ90, 0Æ92, 0Æ85 and 0Æ79, for
alcohol, smoking, physical activity/nutrition, general safety,
and sun self-efficacy, respectively.
A 46-item Health Behaviour Questionnaire was also
developed specifically for this study. The Alcohol use
survey, an 11-item self-report instrument, measures the
quantity and frequency of alcohol consumption (beer) in
the past 30 days and 6 months (Windle 1996). Respondents
were first asked if they had ever consumed any alcoholic
beverage, and only those who had done so completed the
instrument. Frequency of beer intake was rated as
never ¼ 1 to every day ¼ 7 for the last 30 days and last
6 months. Second, respondents indicated how much they
had drunk (none ¼ 1 to more than eight cans/bottles ¼ 10)
for both the last 30 days and 6 months. The number of
episodes of binge drinking was assessed by asking, ‘How
many times did you drink five or more cans/bottles?’ The
higher score was indicative of higher alcohol use. Coeffi-
cient alpha for these data was 0Æ96.
Items on smoking behaviour (10-items), physical activity
and nutritional protective behaviour (13-items), and general
safety protective behaviours (i.e. helmet, seatbelt and
condom use) (eight items) were drawn from the Beha-
vioural Risk Factor Surveillance System (Centers for
Disease Control and Prevention 2001) and the Youth Risk
Behaviour Surveillance System (Centers for Disease Control
and Prevention 2001), accessible through the CDC,
National Center for Chronic Disease Prevention and Health
Promotion. No reliability and validity data were available
from these sources. A 5-point Likert-type scale was used
and the majority of items ranged from never ¼ 1 to all the
time/every day ¼ 5, with the higher score indicating higher
level of smoking and previously mentioned protective
health behaviours. Coefficient alphas in our study were
0Æ95, 0Æ71, and 0Æ55 for smoking behaviour, physical
activity and nutrition protective behaviour, and general

2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Predictors of health behaviours in college students
safety protective behaviour for each of the five health
behaviours, respectively.
Sun-protective behaviour was derived by four items of the
seven item Sun Protective Behaviour Scale (SPBS) developed
by Weinstock et al. 2000. Participants were asked how often
they engaged in sun-protective behaviours when out in the
sun. Items ranged from never ¼ 1 to all the time ¼ 5, with
higher scores indicating higher sun-protective behaviour.
Previously this instrument demonstrated acceptable reliability
(a ¼ 0Æ825) in a sample that included adolescents and young
adults (Weinstock et al. 2000). Coefficient alpha in our study
was 0Æ84.
In summary, alcohol and smoking behaviour scores reflec-
ted actual use of alcohol and smoking, while the other
(physical activity and nutrition, general safety, and sun)
represented positive and protective health behaviours.
The components of HBM – perceived susceptibility and
severity (perceived threat), perceived benefits, and barriers –
were assessed with a 102-item questionnaire constructed for
use in this study and based on the literature (Centers for
Disease Control and Prevention 2001). Each item was rated
on a 5-point Likert-type scale, with one indicating strongly
disagree and five indicating strongly agree. Thus, a higher
total score indicated a higher level of each of the four
components.
Perceived susceptibility was measured by 28 items (eight
items alcohol use, three items tobacco use, three items
physical activity/nutrition intake; nine items general safety,
and five items sun exposure) with a coefficient alpha
reliability of 0Æ92 for the total scale and 0Æ85, 0Æ89, 0Æ95,
0Æ85, and 0Æ94 for each subscale, respectively.
Perceived severity was measured by 27 items (six items
alcohol use, five items tobacco use, three items physical
activity/nutritional intake, 10 items general safety, and three
items sun exposure) with a coefficient alpha reliability of 0Æ92
for the total scale and 0Æ90, 0Æ83, 0Æ78, 0Æ79, and 0Æ88 for
each sub-scale, respectively.
Total scores from perceived susceptibility and severity were
then multiplied to obtain the level of perceived threat. The
higher the total score after multiplying, the higher the level of
perceived threat. The total threat score was then divided into
high and low levels using a median split, to form low and
high threat groups.
Perceived benefits were measured by a 21 item Likert scale.
The subscales were: five items alcohol use; six items tobacco
use, one item physical activity/nutritional intake, three items
general safety, and six items sun exposure. Coefficient alpha
reliability for the total scale was 0Æ91 and subscales were
0Æ86, 0Æ84, 0Æ83 and 0Æ92, excluding physical activity/nutri-
tion, which was measured with a single item.
467
D. Von Ah et al.

Perceived barriers were measured by a 29 item question- (P > 0Æ05), at which point the model was considered to be a
naire, which included alcohol use (four items), tobacco use good fit with the data (Hatcher 1994).
(six items), physical activity/nutritional intake (eight items),
general safety (nine items) and sun exposure (two items), with
Results
a coefficient alpha reliability of 0Æ84 for the total scale and
0Æ78. 0Æ87, 0Æ66, 0Æ76 and 0Æ82 for each of the sub scales,
Perceived stress, availability of and satisfaction with social
respectively.
support, and self-efficacy

Descriptive statistics were calculated to summarize levels of

Ethical considerations
The Institutional Review Board for UAB granted approval for
the study and informed consent was obtained prior to
administering the questionnaires. Participants received two
extra credit research points that could potentially improve
their course grade for completing the questionnaires, which
took on average 40 minutes. Confidentiality was assured and
data handling and storage was confined to the authors.
perceived stress, availability of and satisfaction with social
support, and self-efficacy for each of the five health behav-
iours (Table 2). Perceived stress levels appeared moderate
(19Æ56 ± 15Æ11), while satisfaction of social support was
very high (5Æ50 ± 0Æ66). Overall, self-efficacy for each of the
five health behaviours was also quite high, except for sun
self-efficacy, which was moderate.

Health behaviours
Data analysis
Of the 161 participants, 65Æ8% (n ¼ 133) reported having
A two-step hierarchical multiple regression analysis was ever consumed alcohol (beer, wine and/or liquor), while
used to examine the direct effects of perceived stress, 54Æ7% (n ¼ 88) reported having ever smoked (cigarette and/
perceived availability of and satisfaction with social sup- or cigars). However, about 45% (n ¼ 73) reported exercising
port, and self-efficacy on the identified health behaviours. for periods from at least 20 minutes or more frequently to
First, demographic variables (age, race and gender) were every day within the last 3 months. In the general safety
entered into the model. In step two, since none of the behaviour, which entailed helmet use, seat belt use, and
demographic variables were significant in any of the models, sexual activity, 51% (n ¼ 41/80) of those who routinely rode
only the predictor variables (perceived stress, perceived a bicycle and/or motorcycle reported never wearing a helmet,
availability of and satisfaction with social support, and self- whereas 75% (n ¼ 121) reported wearing seat belts every
efficacy) were entered into the model to determine their time they travelled in a car. Seventy-three per cent (n ¼ 117)
contribution in predicting each of the health behaviours. reported that they were sexually active, of whom most
Study variables used in this analysis were found to have reported becoming sexually active between 16 and 19 years
normal distributions and only weak to moderate correla- of age (n ¼ 70) and 84Æ4% (n ¼ 76) reported routine
tions and therefore meet the assumptions for the analyses condom use. Seventy-eight percent (n ¼ 127) reported rarely
used. or never using sunscreen with a sun protection factor (SPF) of
Structural equation modelling was used to test whether or 15 or higher, which is considered a minimum protection
not the effects of stress, availability of and satisfaction with against harmful ultraviolet rays (Weinstock et al. 2000).
social support, and self-efficacy on health behaviours were
moderated or mediated by perceived threat (perceived
susceptibility multiplied by perceived severity), perceived Table 2 Levels of perceived stress, availability of and satisfaction
benefits, and perceived barriers. Significant predictors from with social support for each of the five health behaviours
the regression analysis were carried forward into the path Variable Range n Mean SD
analysis. Each set of predictors was used in separate
Perceived stress 0–40 160 19Æ56 15Æ11
structural models to determine the pattern of prediction for
Availability of social support 0–9 160 4Æ21 7Æ80
low threat and high threat groups. After the first run of the Satisfaction with social support 0–6 159 5Æ50 0Æ66
structural equation model, all paths that were not significant Alcohol self-efficacy 0–50 161 45Æ88 8Æ43
were dropped from the model and the analysis was repeated. Smoking self-efficacy 0–40 161 36Æ27 8Æ39
Next, new paths were added as suggested from examining Physical activity/nutrition self-efficacy 0–130 161 83Æ36 24Æ99
General safety self-efficacy 0–130 161 107Æ03 23Æ54
the gamma and beta matrices. This process was continued
Sun self-efficacy 0–60 161 37Æ81 14Æ07
until the P-value for the chi-square became non-significant

468
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Issues and innovations in nursing practice Predictors of health behaviours in college students

Table 3 Impact of self-efficacy on each of the health behaviours examined

Health behaviour Self-efficacy predictor Coefficient SE t P value 95% CI

Alcohol use
30 days Alcohol
0Æ7285 0Æ1129
6Æ45 <0Æ0001
0Æ9498,
0Æ5072
6 months Alcohol
3Æ3140 0Æ6721
4Æ93 <0Æ0001
3Æ4457,
3Æ1823
Smoking use Smoking 0Æ1708 0Æ0306 5Æ59 <0Æ0001 0Æ1108, 0Æ2308
Physical activity and nutrition Physical activity and nutrition 0Æ1192 0Æ1757 6Æ78 <0Æ0001 0Æ0847, 0Æ1537
protective behaviour
General safety protective behaviour General safety 0Æ07254 0Æ0197 3Æ67 0Æ0004 0Æ0337, 0Æ1113
Sun-protective behaviour Sun 0Æ0807 0Æ0176 4Æ58 <0Æ0001 0Æ0461, 0Æ1152

components of the health belief model: perceived threat

Effects of stress, availability and satisfaction with social
support, and self-efficacy on health behaviours
As indicated in Table 3, self-efficacy emerged as the only
significant predictor for each of the five health behaviours.
Higher levels of self-efficacy resulted in decreased levels
of alcohol consumption at 30 days [F(4,99) ¼ 11Æ88,
P £ 0Æ0001] and 6 months [F(4,99) ¼ 7Æ08, P £ 0Æ0001], as
well as increased health-promoting behaviour for physical
activity and nutrition behaviour [F(4,156) ¼ 12Æ11,
P £ 0Æ0001], general safety behaviours [F(4,156) ¼ 4Æ00,
P £ 0Æ004], and sun-protective behaviours [F(4,156) ¼
7Æ94, P £ 0Æ0001]. However, higher levels of self-efficacy
resulted in increased smoking [F(4,84) ¼ 10Æ06, P £ 0Æ0001].
Contrary to previous findings, perceived stress and social
support did not significantly predict alcohol and smoking use
and physical activity, nutrition, general safety, and sun-
protective behaviours.
Role of perceived threat, benefits, and barriers
Structural equation modelling was used to explore the
predictors of specific health behaviours using the main
(perceived severity and perceived susceptibility) perceived
benefits and barriers. As the sample size was relatively small,
the analyses can only be considered preliminary, and should
be interpreted with caution and continued in studies with
larger samples. Figures 1–5 show the significant relationships
found in this analysis.
Alcohol behaviour was separated into three sub-categories:
within the last 30 days, 6 months, and binge drinking.
Interestingly, when perceived threat was low, the pathways
for alcohol behaviour were very similar for all three categ-
ories, and when perceived threat was high, the findings for
30 days and 6 months were virtually the same. Thus, Figure 1
shows the pathway significant for alcohol behaviour at
30 days under low and high threat and binge drinking under
high threat, which we believe to be representative of overall
alcohol behaviour. Under low perceived threat, only alcohol
self-efficacy had a significant and direct effect on low alcohol
consumption at both time points and on binge drinking.
However, under high perceived threat, alcohol self-efficacy
had a significant direct effect on alcohol behaviour, and was
also moderated by perceived threat for alcohol behaviour at
30 days and 6 months: higher self-efficacy resulted in lower

Low threat

Self-efficacy Alcohol use – 30 days

–0·66*

High threat
–0·56*

Figure 1 Alcohol use. Low threat: self-

efficacy directly impacted alcohol use when
–0·50* 0·33*
threat was low. High threat: perceived
Self-efficacy Perceived threat Alcohol use – 30 days
threat moderates the relationship
between self-efficacy and alcohol
use-30 days (R2 ¼ 0Æ36) when threat is
high. Binge drinking – high threat: Binge drinking – high threat
perceived barriers mediates the relationship
between self-efficacy and binge drinking Self-efficacy Perceived barriers Binge drinking –
6 months
(R2 ¼ 0Æ64) when threat is high. –0·68* 0·42*

2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474 469
D. Von Ah et al.

Low threat

Self-efficacy Perceived barriers Smoking behaviour

–0·68* 0·42* Figure 2 Smoking behaviour. Low threat:
perceived barriers mediate the relationship
High threat between self-efficacy and smoking
behaviour (R2 ¼ 0Æ64) when threat is low.
0·86*
Self-efficacy Smoking behaviour High threat: self-efficacy directly impacts
smoking behaviour when threat is high.

Low threat
0·43*

Perceived barriers Physical activity and

Self-efficacy Figure 3 Physical activity and nutrition
nutrition behaviour
behaviour. Low threat: self-efficacy directly
–0·59* –0·30 ns
impacts physical activity/nutrition protect-
High threat ive behaviour when threat is low. High
0·36* threat: perceived barriers moderates the
relationship between self-efficacy and
physical activity/nutritional protective
Physical activity and behaviour (R2 ¼ 0Æ29), when threat is
Self-efficacy Perceived barriers
–0·45* –0·29* nutrition behaviour
high.

Low threat
0·42*

Perceived General safety

Self-efficacy
–0·28* barriers –0·29* behaviours Figure 4 General Safety protective beha-
viour. Low threat: perceived barriers mod-
High threat erates the relationship between self-efficacy
and general safety protective behaviours
Self-efficacy
Perceived General safety (R2 ¼ 0Æ16), when threat is low. High
benefits –0·14 ns behaviours threat: self-efficacy directly impacts per-
ceived barriers, which in turn, impacts
–0·27* –0·37*
perceived benefits, however this fails to
Perceived significantly impact general safety
barriers protective behaviour.

Low threat As shown in Figure 2, the effect of self-efficacy on smoking

Sun-protective behaviour was mediated by perceived barriers under low
Self-efficacy
0·42* behaviour threat but was direct (without mediation) under high threat.
Under low perceived threat, higher levels of perceived self-
High threat
efficacy resulted in lower perceived barriers, and ultimately
0·64* Sun-protective less smoking. However, under high perceived threat, higher
Self-efficacy
behaviour
perceived self-efficacy resulted in an increase in smoking.
Figure 5 Sun behaviour. Self-efficacy directly impacts sun-protective Physical activity and nutritional protective behaviour was
behaviour. significantly and positively influenced by self-efficacy under
both low and high perceived threat conditions (Figure 3). As
predicted, under high perceived threat, perceived barriers
perceived threat, which in turn reduced the amount of served as a moderator, with the direct path from self-efficacy
alcohol consumed. Under high perceived threat, perceived to physical activity and nutrition behaviour remaining
barriers mediated the effect of self-efficacy on binge drinking. significant. The higher the self-efficacy and the lower the

470
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Issues and innovations in nursing practice
perceived barriers to physical activity and nutrition protective
behaviours, the more likely respondents were to engage in
physical activity and nutrition protective behaviour, such as
exercising three times a week for 20 minutes and/or eating a
balanced diet.
Self-efficacy had a significant direct effect on general safety
behaviour, which was moderated by perceived barriers under
low perceived threat (Figure 4). In addition, self-efficacy had
a significant, positive impact on perceived benefits. However,
contrary to predictions, this relationship was not maintained
when perceived threat was high. Under high perceived threat,
self-efficacy had a significant negative impact on perceived
barriers, which in turn had negative impact on perceived
benefits; however, there were no significant effects on general
safety behaviours.
Sun-protective behaviour was directly impacted by sun
self-efficacy when perceived threat was low (Figure 5). This
means that the higher the levels of perceived self-efficacy, the
more likely the individual was to engage in sun-protective
behaviour. Under high perceived threat, both sun self-efficacy
and satisfaction with social support had a significant impact
on sun-protective behaviour, as well as self-efficacy having a
positive impact on perceived benefits. However, perceived
threat, benefits and barriers did not have any significant role
in sun-protective behaviour.
Discussion
The direct impact of perceived stress, availability of and
satisfaction with social support, and self-efficacy was exam-
ined on selected health behaviours. In addition, the medi-
ating/moderating role of perceived threat, perceived benefits
and perceived barriers on alcohol and smoking behaviour,
and physical activity/nutrition behaviour, general safety
behaviour and sun-protective behaviour in college students
were explored. The most noted finding was that self-efficacy
was a significant predictor of all health behaviours examined,
and that it had a positive influence on each of these, except
for smoking. This means that the higher the perceived self-
efficacy, the less likely students were to drink, and the more
likely they were to engage in the health-promoting behav-
iours of physical activity, proper nutrition, general safety and
protection from the sun. The reason why higher self-efficacy
resulted in increased smoking remains unclear. Perhaps these
students were reporting confidence in their ability to quit
smoking in the future, not at the present time. Haddad and
Malak (2002) reported similar findings in a study of 650
university students in Jordan. In their study, two-thirds of
college students who smoked reported that they intended to
quit smoking sometime in the future. These findings are also

2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Predictors of health behaviours in college students
consistent with assertions by Rienzo (1992), who reported
that the incidence of smoking in young adults is complicated
by their belief that they will quit smoking in the future and
therefore be spared its long-term effects. Nevertheless, our
findings support previous research on the importance of
perceived self-efficacy as a cognitive factor affecting health
behaviour (O’Leary 1985). Also, since self-efficacy is
amenable to change, especially in college students whose
lifestyle habits are not firmly established, nurses must sponsor
health promotion programmes which incorporate methods
for increasing self-efficacy.
On the other hand, perceived stress did not have a
significant impact on the health behaviours evaluated. One
possible explanation for this may be the inadequacy and/or
insensitivity of the PSS in measuring stress among college
students. Some researchers have asserted that hassles or
everyday stressors are better predictors of health (Kanner
et al. 1981, Wagner et al. 1988) than general appraisal
instruments such as the PSS. Although internal reliability of
this instrument was adequate in this study, items may have
not been suitable to delineate stress specific to college
students. In addition, reported stress levels were lower than
anticipated. The timing of data collection, between mid-
term examinationss vs. immediately prior to an examination
period, may have also contributed to the lower stress
scores.
Availability of and satisfaction with social support did not
have a significant impact on health behaviours. The lack of
significance of social support in predicting health behaviours
may be attributed to the way in which social support was
conceptualized and measured. Only the direct positive impact
of social support was examined as a potential predictor of
health behaviour, and not the interactive, buffering effect of
social support on stress. The differential effects of social
support may become more evident when the level of social
support is low under high stress situations, which then may
become detrimental to health behaviour (Cohen et al. 2000).
In this study, however, students reported exceptionally high
satisfaction with social support under relatively low to
moderate stress. If the study had been conducted shortly
after their transition into college and prior to establishing
satisfying support relationships, or during a high academic
stress period, the results might have been different. Previous
research by Steptoe et al. 1996 found that social support
moderated the effects of examination stress on both alcohol
and smoking behaviour. Thus, further research during
different transitional points of college life may be needed to
determine the explanatory role of perceived stress, social
support, and their relationships with health behaviours in
college students.
471
D. Von Ah et al.
What is already known about this topic
• Many college students engage in various risky health
behaviours including alcohol use, tobacco use, physical
inactivity and unhealthy dietary practices, ignore pre-
ventive safety habits and engage in excessive sun
exposure, which may have long-term implications for
their health.
• Perceived stress, social support and self-efficacy may
impact separately on health behaviour.
• Studies of health behaviour in college students are
lacking and those that have been conducted have failed
to examine these closed-linked variables simulta-
neously.
What this paper adds
• Self-efficacy, but not perceived stress and availability
and satisfaction with social support, was a significant
predictor of alcohol and smoking use and physical
activity/nutrition behaviour, general safety and sun-
protective behaviours in college students.
• Perceived barriers had a significant and negative impact
on the health behaviours examined in college students.
• Self-efficacy plays an important role in reducing per-
ceived barriers to performing protective health behav-
iours.
Health belief model and its role in health behaviour
As predicted, under high perceived threat, perceived threat
served as moderator between self-efficacy and alcohol con-
sumption at 30 days and 6 months, with the direct path
remaining significant. In addition, under high threat self-
efficacy was mediated by perceived barriers to binge drinking
and moderated by perceived barriers to physical activity and
nutrition protective behaviours. Under low threat, self-effic-
acy was mediated by perceived barriers to smoking behaviour
and perceived barriers served as a moderator between self-
efficacy and general safety protective behaviours.
Interestingly, under both low and high perceived threat
conditions, perceived barriers was the most consistent
component of the HBM to show a significant negative
relationship with self-efficacy in almost all health behaviours.
These findings are congruent with Stretcher and Rosenstock’s
(1997) assertion that the ‘explanatory power’ of the HBM
would be further enhanced with the inclusion of self-efficacy
(Stretcher & Rosenstock 1997, p. 47). Our findings support
the importance of perceived barriers, as well as self-efficacy,
in predicting health behaviours in college students.
472
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Perceived barriers mediated the effects of self-efficacy on
binge drinking behaviour and smoking, whereas perceived
barriers moderated the effects of self-efficacy on physical
activity and nutrition behaviours and general safety behav-
iours. This indicates the important role that perceived
barriers play in mediating or moderating self-efficacy in the
various health behaviours. Typically, the higher the levels of
perceived barriers, the more likely participants were to
engage in more negative behaviours, such as more drinking
and smoking, less physical activity and adequate nutrition
behaviour and less general safety behaviours. In contrast,
perceived benefits had no mediation or moderating role in the
model. Thus, it appears that perceived barriers and self-
efficacy are the two most significant factors in predicting
various health behaviours in college students. Similar results
were reported by Grubbs and Carter (2002), who found that
reduction in perceived barriers to exercise was the most
influential factor for total exercise (minutes/week) in college
students (Grubbs & Carter 2002). Taken together, these
findings are especially interesting given the large number of
students who reported immediate family members with
cardiovascular diseases (i.e. hypertension, high cholesterol
and heart disease). Future studies incorporating risk evalua-
tions might shed more light on predicting health behaviour
and health behaviour change.
Study limitations
Although these findings give valuable insight into health
behaviours in college students, several limitations should be
noted. The use of self-report questionnaires, a cross-sectional
design, low response rate and small sample size for conduct-
ing path analysis, as well as limited generalizability, may have
affected the findings. In addition, further psychometric testing
of the questionnaires developed for this study is warranted.
Conclusions
Health behaviour established during adolescence and young
adulthood may have a significant impact on health behav-
iours and the occurrence of diseases later in life. College
students, who are notoriously known for engaging in risky
health behaviours, are prime candidates for health promotion
and prevention programmes. Our results affirm the signifi-
cant role of self-efficacy in reducing perceived barriers to
performing protective health behaviours. Thus, nurses, who
are on the frontline in many college student health centres,
should incorporate methods for enhancing self-efficacy and
reducing barriers to adopting a healthy lifestyle in future
health promotion programmes.
Issues and innovations in nursing practice
Acknowledgements
This study was supported by Sigma Theta Tau International –
Nu Chapter.
References
Allen S. & Heibert B. (1991) Stress and coping in adolescents.
Canadian Journal of Counseling 25, 19–32.
Bandura A. (1977) Self-efficacy: toward a unifying theory of
behavioral change. Psychological Review 84, 191–215.
Byrne D.G., Byrne A.E. & Reinhart M.I. (1995) Personality, stress
and the decision to commence cigarette smoking in adolescence.
Journal of Psychosomatic Research 39(1), 53–62.
Campbell R.L. & Svenson L.W. (1992) Perceived level of stress
among university undergraduate students in Edmonton Canada.
Perceptual and Motor Skills 75, 552–554.
Centers for Disease Control and Prevention (2001) Youth Risk
Behavior Surveillance System Questionnaire. US Department of
Health and Human Services, Centers for Disease Control and
Prevention, Atlanta, GA.
Centers for Disease Control (1997) Youth Risk Behavior Surveil-
lance: National College Health Risk Behavior Survey – United
States. Mortality and Morbidity Weekly Report 46(SS-6), 1–54.
Centers for Disease Control and Prevention (2001) Behavioral Risk
Factor Surveillance System Survey. US Department of Health and
Human Services, Center for Disease Control and Prevention,
Atlanta, GA.
Cohen S.K.T., Kamarck T. & Mermelstein R. (1983) A global
measure of perceived stress. Journal of Health and Social Beha-
vior 24, 386–396.
Cohen S., Gottlieb B.H. & Underwood L.G. (2000) Social relation-
ships and health. In Social Support Measurement and Intervention: A
Guide for Health and Social Scientists (Cohen S., Underwood L.G. &
Bottlieb B.H., eds), Oxford University Press, New York, pp. 3–25.
Cusatis D.C. & Shannon B.M. (1996) Influences on adolescent eating
behavior. Journal of Health and Social Behavior 18, 27–34.
Dinger M.K. (1999) Physical activity and dietary intake among
college students. American Journal of Health Studies 15(3),
139–149.
Dinger M.K. & Waigandt A. (1997) Dietary intake and physical
activity behaviors of male and female college students. American
Journal of Health Promotion 11(5), 360–362.
Everett S.A., Lowry R., Cohen L.R. & Dellinger A.M. (1999) Unsafe
motor vehicle practices among substance-using college students.
Accident, Analysis and Prevention 31, 667–673.
Flay B.R. (1993) Youth tobacco use: risks, patterns, and control. In
Nicotine Addiction: Principles and Management (Orleans C.T. &
Slade J., eds), Oxford University Press, New York.
Flay B., Hu F.B., Siddiqui O., Day L.E., Hedeker D., Petraitis J. &
Sussman S. (1994) Differential influence of parental smoking and
friends’ smoking on adolescent initiation and escalation of smoking.
Journal of Health and Social Behavior 35(3), 248–265.
Grubbs L. & Carter J. (2002) The relationship of perceived benefits
and barriers to reported exercise behaviors in college under-
graduates. Family Community Health 25(2), 76–84.

2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Predictors of health behaviours in college students
Haddad L.G. & Malak M.Z. (2002) ‘Smoking habits and attitudes
towards smoking among university students in Jordan’. Interna-
tional Journal of Nursing Studies 39(8), 793–802.
Hatcher L. (1994) A Step-by-Step Approach to Using the SAS System
for Factor Analysis and Structural Equation Modeling. SAS
Institute, Cary, NC.
Hubbard P., Muhlenkamp A.F. & Brown N. (1984) The relationship
between social support and self-care practices. Nursing Research
33, 266–270.
Ichiyama M.A. & Kruse M.I. (1998) The social contexts of binge
drinking among private university freshmen. Journal of Alcohol
and Drug Education 44(1), 18–33.
Jones S.E., Oeltmann J., Wilson T.W., Brener N.D. & Hill C.V.
(2001) Binge drinking among undergraduate college students in the
United States: implications for other substance use. Journal of
American College Health 50(1), 33–38.
Kanner A.D., Coyne J.C., Schaefer C. & Lazarus R. (1981) Com-
parison of two modes of stress management: daily hassles
and uplifts vs. major life events. Journal of Behavioral Medicine 4,
1–39.
Kelly R., Zyzanski S.J. & Alemagno S.A. (1991) Prediction of mo-
tivation behavior change following health promotion: role of
health beliefs, social support and self-efficacy. Social Science
Medicine 32(3), 311–320.
Kvis F.J., Clark M.A., Crittenden K.S., Warnecke R.B. & Freels S.
(1995) Age and smoking cessation behaviors. Preventative
Medicine 24, 297–307.
Lazarus R.S. & Folkman S. (1984) Stress, Appraisal, and Coping.
Springer Publishing Co., New York.
Mahler H.L.M., Fitzpatrick B., Parker P. & Lapin A. (1997) The
relative effects of a health-based versus and appearance-based
intervention designed to increase sunscreen use. American Journal
of Health Promotion 11, 426–429.
Martinelli A.M. (1999) An explanatory model of variables influen-
cing health promotion behaviors in smoking and nonsmoking
college students. Public Health Nursing 16(4), 263–269.
Mates D. & Allison K.R. (1992) Sources of stress and coping
responses of high school students. Adolescence 27, 461–474.
McGinnis J.M. & Foege W.H. (1993) Actual causes of death in
the United States. Journal of American Medial Association
270(18), 2207–2212.
Meilman P.W., Yanosfsky N.N., Gaylor M.S. & Turcon J.H. (1989)
Visits to college health service for alcohol-related injuries. Journal
of American College Health 37, 205–210.
Misra R. (2000) Academic stress of college students: comparison
of student and faculty perceptions. College Student Journal 21, 1–10.
O’Leary A. (1985) Self-efficacy and health. Behaviour Research and
Therapy 23(4), 434–451.
Presley C.A., Meilman P.W. & Lyerla R. (1995) Alcohol on Ameri-
can Campuses: Use, Consequences, and Perceptions of the Campus
Environment. Southern Illinois University, Carbondale.
Reynolds K.D., Blawn J.M., Jester P.M., Weiss H., Soong S. &
DiClemente R.D. (1996) Predictors of sun exposure in adolescents
in a southeastern US population. Journal of Adolescent Health 19,
409–415.
Rienzo P.G. (1992) Nursing Care of the Person Who Smokes.
Springer Publishing Co., New York.
473
D. Von Ah et al.
Robinson J.K., Rigel D.S. & Amonette R.A. (1997) Trends in sun
exposure, attitudes, and behaviors 1986 to 1996. Journal of
American Academy of Dermatology 17, 176–186.
Sarason I.G. (1987) A brief measure of social support: practical and
theoretical implications. Journal of Social and Personal Relation-
ships 4, 497–510.
Sarason I.G., Levine H.M., Basham R.B. & Sarason B.R. (1983)
Assessing social support: the social support questionnaire. Journal
of Personality and Social Psychology 44, 127–139.
Shapiro J., Radecki S., Charchian A.S. & Josephson V. (1999) Sexual
behavior and AIDS-related knowledge among community college
students in Orange County, California. Journal of Community
Health 24(1), 29–43.
Skutle A. (1999) The relationship among self-efficacy expectancies,
severity of alcohol abuse, and psychological benefits from drink-
ing. Addictive Behaviors 24(1), 87–98.
Steptoe A. & Wardle J. (2001) Health behaviour, risk awarenes and
emotional well-being in students from Eastern Europe and Western
Europe. Social Science & Medicine 53, 1621–1630.
Steptoe A., Wardle J., Pollard T.M., Canaan L. & Davies G.J. (1996)
Stress, social support and health-related behavior: a study of
smoking, alcohol consumption and physical exercise. Journal of
Psychosomatic Research 41(2), 171–180.
Steptoe A., Phil D., Wardle J., Cui W., Bellisle F., Zotti A., Baranyai R.
& Sanderman R. (2002) Trends in smoking, diet, physical exercise,
and attitudes toward health in European university students from 13
countries, 1990–2000. Preventative Medicine 35, 97–104.
474
2004 Blackwell Publishing Ltd, Journal of Advanced Nursing, 48(5), 463–474
Stretcher V. & Rosenstock I.M. (1997) The health belief model. In
Health Behavior and Health Education: Theory, Research and
Practice (Glanz K., Lewis F.M. & Rimer B.K., eds), Jossey-Bass,
San Francisco, CA.
US Department of Health and Human Services (2000) Healthy
People 2010: Understanding and Improving Health, 2nd edition.
US Government Printing Office, Washington, DC.
Wagner B.M., Compas B.E. & Howell D.C. (1988) Daily and major
life events: a test of an integrative model of psychosocial stress.
American Journal of Community Psychology 16, 189–205.
Wechsler H., Dowdall G.W., Davenport A. & CAstilo S. (1995)
Correlates of college student binge drinking. American Journal of
Public Health 85(7), 921–926.
Wechsler H., Dowdall G.W. & Maenner G. (1998) Changes in binge
drinking and related problems among American college students
between 1993 and 1997: results of the Harvard School of Public
Health College Alcohol Study. American Journal of College Health
47, 57–68.
Weinstock M.A., Rossi J.S., Redding C.A., Maddock J.E. & Cottrill
S.D. (2000) Sun protective behaviors and stages of change for the
primary prevention of skin cancers among beachgoers in South-
eastern New England. Annual of Behavior Medicine 22(4), 286–
293.
Windle M. (1996) Alcohol involvement typology for adolescents:
convergent validity and longitudinal stability. Journal of Studies on
Alcohol 57(6), 627–637.
View publication stats