HEALTH EDUCATION RESEARCH

Theory & Practice

Vol.19 no.2 2004

Pages 185197

Evaluation of a theoretically based AIDS/STD peer

education program on postponing sexual intercourse
and on condom use among adolescents attending high
school1
1,4

F. Caron , G. Godin , J. Otis and L. D. Lambert

Abstract

1Public

Health Direction, Planning and Evaluation,

Regional Health Board Monteregie, Longueuil, Quebec
J4K 2M3, 2Research Group on Behaviors in the Field of
Health, Faculty of Nursing, Laval University, Quebec
G1K 7P4 and 3Department of Sexology, University of
Quebec at Montreal, Quebec H3C 3P8, Canada
4Correspondence to: F. Caron;
E-mail: f.caron@rrsss16.gouv.qc.ca

use condoms on a regular basis than those in

the control group. Program effects occurred
among both sexes, but a few differences in
response were observed among males and
females. Results suggest this type of theorybased program is effective in modifying psychosocial variables related to postponing sexual
intercourse and related to condom use among
adolescents. Personal involvement in designing
intervention appears to be effective in modifying the behavior of peer educators.

Introduction
Adolescents are at risk for STD and HIV infection.
In Canada, genital chlamydial infections are most
prevalent among teenage girls aged 1519 years
(Laboratory Center for Disease Control, 1999).
When untreated, chlamydia can cause serious
health problems. It is estimated that up to 40% of
women with untreated chlamydia will develop
pelvic inammatory disease, and, of these, 20%
will become infertile, 18% will experience debilitating pelvic pain and 9% will have a lifethreatening tubal pregnancy (Centers for Disease
Control and Prevention, 2001a). Furthermore,
recent research suggests that women with chlamydia have a 3- to 5-fold increased risk of acquiring
HIV, if exposed (Centers for Disease Control and
Prevention, 2001a).
In the US, the AIDS incidence from 1996
through 2000 has declined in most populations, but
increases were observed notably among women
and persons infected through heterosexual contact.
Among females aged 1319 years who were
infected with HIV, but did not have AIDS,

Health Education Research Vol.19 no.2, Oxford University Press 2004; All rights reserved

DOI: 10.1093/her/cyg017

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

The effectiveness of an intervention based upon

the theories of Ajzen and Triandis was evaluated among 698 junior and 306 senior high
school students. The intervention to juniors
was offered by senior students who were
trained during a course integrated into the
school curriculum. Respondents in the control
and experimental groups completed a questionnaire at baseline and 9 months after the program. Compared to junior respondents in the
control group, those in the experimental group
positively modied their attitude, perceived
behavioral control, personal normative beliefs,
perceived role beliefs, anticipated regret and
intention with respect to postponing sexual
intercourse and with respect to condom use, as
well as perceived self-efcacy to negotiate both
behaviors. Compared to senior respondents in
the control group, those in the experimental
group showed a signicant positive modication of all the above variables except perceived
behavioral control (indirect measure), anticipated regret and intention with respect to postponing sexual intercourse. At post-test, seniors
in the experimental group were more likely to

F. Caron et al.

186

example, adolescents who believe their friends are

not in favor of sexual intercourse for teenagers are
more likely not to have engaged in intercourse
(DiIorio et al., 2001) and those who perceive their
peers as having a favorable attitude toward condom
use are more likely to use a condom themselves
(Turner, 2001). However, even though peer-led
programs have received positive anecdotal reviews
from program coordinators, peer educators and
young audiences, only a few programs have been
evaluated for their impact.
Recent studies do suggest that peer education
can bring about short-term positive changes in
attitude (Caron et al., 1998), self-efcacy (Smith
et al., 2000) and behavioral intention regarding
condom use (Caron et al., 1998), attitude towards
abstinence (Caron et al., 1998), self-efcacy to
refuse sex (Aarons et al., 2000), more conservative
sexual norms (Mellanby et al., 2000), delay in the
initiation of sexual behaviors (Aarons et al., 2000),
the frequency of intercourse (Jemmott et al., 1998)
and the use of condoms among sexually active
teens (Smith et al., 2000). Such effective programs
tend to be theory-based, highly structured and led
by well-trained peers (Smith et al., 2000).
A few studies have compared the effectiveness
of peer-led sex education programs with adult-led
interventions. One such study found that peer
leaders were more effective in establishing conservative norms and attitudes related to sexual
behavior than were adults, but they were less
effective than adults in imparting factual
knowledge (Mellanby et al., 2000). In contrast,
students who attended another peer-led program
had signicantly higher knowledge scores on HIV
prevention than did those who participated in the
same intervention led by community health nurses
(Dunn et al., 1998). In a review of 13 studies that
compared peer-led and adult-led health education
programs, peer leaders were found to be at least as,
or more, effective than adults in most studies
reporting positive outcomes (Mellanby et al.,
2000).
There does appear to be a shortage of welldesigned studies that have evaluated the mediumand long-term effects of AIDS/STD peer education

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

heterosexual contact was the most frequently

reported risk factor (Centers for Disease Control
and Prevention, 2001b).
From 1991 to 1997, condom use at last intercourse signicantly increased (from 46 to 57%),
birth control pill use decreased (from 21 to 17%)
and use of withdrawal decreased signicantly
(from 18 to 13%) among high school students in
the US (Everrett et al., 2000). Despite recent
decreases in sexual risk taking (Centers for Disease
Control and Prevention, 1998), adolescents still
face a potentially high risk of STD and HIV
infection because of their behavior. Half (49.9%)
of the high school students in the Youth Risk
Behavior Survey reported sexual intercourse and
58% reported the use of a condom at last
intercourse. Furthermore, 50% of high school
students had drunk alcohol during the 30 days
preceding this survey and 26.7% had used
marijuana (Kann et al., 2000). Studies have
shown that marijuana use is associated with
unprotected sexual intercourse among adolescents
(Kingree et al., 2000), and substance abuse is
related to earlier onset of sexual activity, more
sexual partners, less consistent use of condoms and
more STDs among adolescent detainees (Tapert
et al., 2001).
Over the past decade, schools have implemented
numerous STD/HIV prevention programs, but
results have been mixed (Song and Pruitt, 2000;
Kirby, 2001). Peer education programs have also
grown in popularity internationally and have
targeted a wide range of youth in a variety of
settings. One reason why peer education has
become so popular may be an underlying belief
that youths are effective educators because they are
credible and inuential role models. In fact, there is
theoretical evidence for this belief since, according
to Social Cognitive Theory, adolescents are more
likely to enact modeled behavior if they perceive
the models as warm, supporting, and similar to
themselves with respect to such characteristics as
gender, ethnicity and age (Bandura, 1986).
Research ndings also suggest that peer norms
inuence the process of sexual initiation (Carvajal
et al., 1999) and condom use (Kirby, 2001). For

Impact of peer education on adolescents' safe sex behaviors

Table I. Salient model beliefs of respondents targeted by the program
Condom use

Postponing sexual intercourse offers a way of

Being responsible
Respecting oneself
Showing respect to one's partner
Finding out if one really loves one's partner
Getting to know oneself better
Discovering each other sexually at own rhythm
Being more ready for intercourse later
Protecting oneself from pregnancy
Protecting oneself from STDs
Protecting oneself from AIDS
Obstacles to postponing sexual intercourse:
One's partner insists on having intercourse
One's partner is against one's decision to postpone
sexual intercourse
One's friends insist one should have sex
One's female friends are against one's decision to
postpone sexual intercourse
One's male friends are against one's decision to
postpone sexual intercourse

Using a condom offers a way of

Being responsible
Respecting oneself
Showing respect to one's partner
Having a sense of security
Protecting oneself from pregnancy
Protecting oneself from STDs
Protecting oneself from AIDS
Obstacles to condom use:
One's friends insist one should have intercourse
without a condom
One is afraid that sex would be less natural
One is afraid that sex would be less romantic
One is afraid the boy might lose his erection
One is afraid one will have less pleasure
One is afraid the condom would break
One is already taking the birth control pill
One would feel shy when buying condoms

among both the peer educators themselves and

their target audience. This paper presents the
evaluation results of a school-based peer education
program. The goal was to determine the extent to
which this peer education program was effective in
changing medium-term behavior and its underlying social cognitive determinants among both
senior and junior high school students with respect
to postponing sexual intercourse and with respect
to condom use.

Overview of the program

The Protection Express Program is based on the
theories of Planned Behavior (Ajzen, 1991) and
Interpersonal Behavior (Triandis, 1980), and a
teaching model based on Social Cognitive Theory.
According to the theories used, individual's behavior is determined by intention and a number of key
social cognition constructs. In the Protection
Express Program, the content is developed in
order to affect attitude, perceived social norm,
perceived behavioral control, personal normative

belief, role beliefs and the beliefs underlying these

main constructs. The information used to dene the
content of the Protection Express Program is based
on a previous study that identied the salient modal
beliefs as well as the main determinants of the
targeted behaviors among the population under
study (Caron et al., 1998). These beliefs are listed
in Table I.

The training of peer educators

Peer educators were senior high school students
who participated in a 25-hour training program
integrated into regular Moral and Religious
Education classes. The training program was
offered by the teacher with support from the
school nurse or a supervised undergraduate student
in Sexology.
The peer-training program was interactive and
highly structured. Students worked in teams of ve
individuals to develop an educational presentation
on one of ve topics: postponing sexual intercourse, communication and assertiveness in rela187

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Postponing sexual intercourse

F. Caron et al.

Method
Design and site of study
The impact evaluation design was quasi-experimental and included an experimental group composed of respondents who attended six high
schools located in two regions in Quebec,
Canada. One of these schools was located in a
rural setting while the remaining ve were all
situated close to a major city. Two schools were
large in size with over 2000 students enrolled, two
others were of medium size with approximately
1300 students and the nal two were small high
schools with less than 600 students. The comparison groups were also composed of respondents
attending high schools that were matched in size
and location. The experimental group received the
Protection Express Program, and the comparison
group received their usual sex education program
and were not involved in the development or
delivery of the program being evaluated. The
design of the study allowed for the evaluation of
the program at two levels: among the senior
respondents (i.e. those trained to offer the intervention to junior students) and among the junior
respondents (i.e. those receiving the peer education
component of the program). The impact of the
intervention was evaluated with a self-administered questionnaire completed by respondents
188

during three observation periods: (1) before the

Protection Express Program, (2) 2 weeks following
the end of the program and (3) 9 months later.
Also, two behaviors were targeted by this evaluation: (1) postponing sexual intercourse and (2)
condom use.

Population and sample

A total of 1166 students at the junior level and 549
at the senior level composed the potential pool of
participants. A total of 945 (junior level) and 477
(senior level) students agreed to participate. This
represented approximately 87 and 81%, respectively, of the junior and senior targeted populations.
Nine months after the intervention, the nal sample
of senior respondents consisted of 147 and 159
respondents, respectively, in the experimental and
control groups. The nal sample of junior respondents consisted of 369 juniors in the experimental
group and 329 juniors in the control group. For
both groups, 29.4% were lost at follow-up or
excluded from the analysis since all three questionnaires (pre-test and both post-tests) had to be
included in the analysis. There was also a fairly
high student absentee rate at either the rst or the
second post-tests. Students retained in the study
were more likely to be female (P < 0.001), younger
(P < 0.001), sexually inactive (P < 0.001) and to
have higher scores on many of the dependent
variables evaluated.

Procedure for data collection

Data was collected by the research team between
November 1999 and February 2000 (T0), 2 weeks
following the end of the program (T1), and 9
months later (T2: February 2001May 2001). All
questionnaires were completed in the classroom
within approximately 60 min. Respondents were
given a code to ensure condentiality and allow for
the matching up of questionnaires. Written parental
consent was obtained prior to the study for students
under the age of 14 years. All respondents signed a
consent form on a voluntary basis. Approval was
obtained from the Ethics Review Board of Laval
University.

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

tionships, equality in relationships, conditions to a

healthy relationship, and condom use. Each team
developed their own sketches, which were often
based on their personal experience. However, all of
the sketches on a particular topic dealt with the
same content (e.g. advantages of postponing sexual
intercourse). The content was therefore consistent
from one school to another, but it was presented in
various ways by different teams. Content delity
was monitored by using formative evaluation tools.
Thus, all peer presentations were evaluated twice
by both teachers and peers, and only those judged
to be educational according to the pre-established
criteria utilized in the formative evaluations were
subsequently presented to junior students.

Impact of peer education on adolescents' safe sex behaviors

Instrumentation for impact evaluation

189

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

The questionnaire used to evaluate the impact of

the Protection Express Program was an adaptation
of validated instruments based on the theories of
Planned Behavior and Interpersonal Behavior
(Godin et al., 1997). Questions were formulated
to take into account the sex of the respondents, and
all questions were pilot tested to ensure they were
clear and that phrasing of items was appropriate for
the population under study. Sexual intercourse was
dened as including the penetration of the penis
into the vagina or the anus.
The primary outcomes were self-reported sexual
behaviors in the previous 3 months, including
postponing sexual intercourse (no, yes, no partner
in last 3 months), sexual intercourse (no, yes, no
partner in last 3 months), frequency of condom use
with a regular boyfriend or girlfriend or with a
partner who is not a boyfriend or a girlfriend (a
`one night stand'). Frequency of condom use was
rated on a ve-point scale: never, 25% of the time,
50% of the time, 75% of the time and always.
The direct measures of attitude toward postponing sexual intercourse and toward condom use
were each assessed by means of ve items on a
ve-point differential semantic scale. Respondents
were asked to answer to the following questions.
`If in the next 3 months, I was with a boy and I
postponed having sexual intercourse with him (or
we used a condom during intercourse) it would
beunpleasant/pleasant,
frustrating/satisfying,
disagreeable/agreeable, useless/useful, unsafe/
safe'. Cronbach's coefcient a was 0.76 for the
subscale on postponing sexual intercourse and 0.79
for the condom use subscale.
The remaining items were assessed on ve-point
Likert scales (agreedisagree). The indirect measure of attitude was based on respondents' beliefs
regarding the advantages and disadvantages of
postponing sexual intercourse (10 items: a = 0.81)
and of using condoms (eight items: a = 0.74).
Three items were used to assess the direct
measure of behavioral control. Respondents indicated to what degree they could easily (or not
easily) postpone sexual intercourse (a = 0.59) or

use a condom (a = 0.78). The direct measure of

behavioral control was based on respondents'
beliefs regarding their ability to overcome (or
not) the obstacles associated with postponing
sexual intercourse (ve items: a = 0.89) and with
condom use (eight items: a = 0.88). For example,
`If in the next 3 months I was with a boy I would be
able to postpone sexual intercourse with him even
ifthe boy insisted we have intercourse, my
friends insisted I should have intercourse, etc.'.
Similarly, `If in the next 3 months, I was with a
boy, I would be able to use a condom even ifI
was already taking the birth control pill, I was
worried the condom would break, etc.'.
Five self-efcacy items measured respondents'
perceived ability to negotiate postponing sexual
intercourse (a = 0.78) and to negotiate condom use
(a = 0.84). For example, `If a boy insists I have sex
and I have already decided to postpone sexual
intercourse (if a boy insisted we have intercourse
without a condom), I would be able tosay no and
maintain my decision, tell him how I feel, refuse to
discuss it further, etc.'.
Role beliefs refer to one's perception of how a
person of one's own age, gender or status should
behave (Triandis, 1980). Three items measured to
what degree respondents felt it was normal (or not)
for a person of their age and gender to postpone
sexual intercourse (a = 0.76) and use a condom
(a = 0.81).
Personal normative beliefs are feelings of personal responsibility regarding the performance or
not of a given behavior (Triandis, 1980). Three
items measured to what extent respondents felt that
postponing sexual intercourse (a = 0.90) and use of
a condom (a = 0.84) were in keeping with their
own values.
Anticipated regret refers to the various worries
and regrets that one takes into account before
making a decision (Richard et al., 1998). Three
items were used to measure this variable. `If in the
next 3 months I had sex even though I had already
decided to postpone sexual intercourse (I had
sexual intercourse without a condom)I would
worry afterwards, I would regret it the next day, I
would feel preoccupied in the days that followed.'

F. Caron et al.
Cronbach's a values were 0.84 (postponing intercourse) and 0.92 (condom use). Finally, three items
were used to assess intention to postpone sexual
intercourse (a = 0.88) and condom use (a = 0.88).

Analysis

190

Results
Demographic characteristics of the
respondents
Descriptive data of senior and junior respondents
in the experimental and control groups are presented in Tables II and III, respectively. At both
high school levels, the experimental and control
groups were similar. The only exception was a
signicant difference in the proportion of respondent ethnicity; at both high school levels, compared
to control groups, there were in the experimental
groups a higher proportion of respondents born
outside Quebec/Canada.

Impact on behavioral and psychosocial

variables
Senior respondents: impact on postponing
sexual intercourse
Nine months after the end of the program, experimental and control groups did not differ in
behavior with respect to postponing sexual intercourse, nor did they differ regarding intention,
perceived behavioral control (indirect measure)
and anticipated regret. The experimental group,
however, scored higher than the control group on
the remaining psychosocial variables (Table IV,
top portion), i.e. attitude (P < 0.001, direct and
indirect measure), the direct measure of perceived
behavioral control (P < 0.001), perceived normative beliefs (P < 0.01), role beliefs (P < 0.001) and
perceived self-efcacy (P < 0.001). The impact of
the program on all variables was similar for both
sexes.

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

In order to take into account the clustering effects,

a hierarchical model procedure (with the SAS
PROC MIXED procedure or GLIMMIX macro
provided by SAS for dichotomous variables) was
applied. For the junior respondents, no clustering
effects were detected and the analyses based on
groups or individuals yielded similar signicant
results. As for the senior respondents, only three
variables were affected by clustering effect, but
they were not signicant. For the remaining
variables, the analyses based on students yielded
similar signicant results. Consequently, the results based on individuals are reported.
For the impact evaluation, frequencies were
computed to determine the demographic characteristics of the respondents. The c2-test was
used to compare the demographic characteristics
of the experimental groups with those of the
comparison groups. Because the variable concerning the postponement of sexual intercourse
in the last 3 months was dichotomous, the
GENMOD procedure (SAS version 8.02) was
used to measure the impact of the program on
this variable. The remaining hypotheses were
tested using covariance analysis (ANCOVA)
whenever the F-test showed parallelism of
regression slopes (Cook and Campbell, 1979).
When parallelism of regression slopes could not
be demonstrated, hypothesis testing was accomplished using the JohnsonNeyman method of
analysis (Huitema, 1980). In this case, the
P value refers to results obtained when the
effects of the program became signicant.
These methods of regression analysis were
used because the focus was to determine the
effect of the program on the response variables
adjusted for the presence of the control variables in the model (score on pre-test and sex of
respondents). To determine the short-term impact of the program, pre-test data was compared

to data from the rst post-test (T0T1). Finally,

pre-test data was compared to data collected 9
months after the end of the program (T0T2).
Data analysis was completed with SAS 6.32.
Because the short-term impacts (T0T1) were
similar in pattern to those observed 9 months
later, only the results contrasting T0T2 are
presented.

Impact of peer education on adolescents' safe sex behaviors

Table II. Demographic characteristics and sexual behaviors of senior respondents on pre-test
Variables

Seniors

aThere

Control
[% (N = 159)]

16.0
44.2
55.8
92.5
7.5

15.9
44.0
56.0
97.5
2.5

29.7
15.9
54.5

29.1
15.2
55.7

16.5
34.5
49.0

18.9
28.9
52.2

57.8
42.2

55.6
44.4

38.1
61.9

50.0
50.0

was a signicant difference between experimental and control groups P < 0.05.
among those who are sexually active.

bPercentage

Junior respondents: impact on postponing

sexual intercourse
Nine months after the program, experimental and
control groups did not differ in behavior with
respect to postponing sexual intercourse in the last
3 months. The experimental group, however,
scored higher than the control group on all of the
psychosocial variables (Table VI, top portion):
intention (P < 0.001), attitude (P < 0.001, direct
measure), perceived self-efcacy (P < 0.001), role
beliefs (P < 0.001) and anticipated regret (P <
0.001) with respect to postponing sexual intercourse. As shown in Table VII (top portion), the
impact of the program on the following variables

was greater among respondents with the lowest

pre-test scores: indirect measure of attitude (P <
0.03), indirect measure of perceived behavioral
control (P < 0.02) and personal normative beliefs
(P < 0.02). The program also had a greater impact
among boys than among girls on the direct measure
of perceived behavioral control (boys P < 0.02;
girls P < 0.02).

Senior respondents: impact on condom use

Nine months after the intervention, respondents in
the experimental group were more likely to use a
condom consistently with their regular or occasional sexual partner (P < 0.01) than respondents in
191

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Age (years)
Male
Female
Respondent born in Quebec or Canada
Respondent born outside of Quebec or Canadaa
In the last 3 months, did you postpone sexual intercourse with a boy (girl)?
No
Yes
I have not dated a boy (girl) in the last 3 months
In the last 3 months, did you have sexual intercourse with a boy (girl)
(penetration of the penis in the vagina or the anus)?
No
Yes
I have never had sexual intercourse
In the last 3 month, when you had sexual intercourse with the boy (girl)
you are dating (your boyfriend, girlfriend), how often did you use a condom?b
Not always
Always (100% of the time)
In the last 3 months, when you had sexual intercourse with a boy (girl)
who is not your boyfriend (girlfriend) (i.e. a one night stand), how often
did you use a condom?b
Not always
Always (100% of the time)

Experimental
[% (N = 147)]

F. Caron et al.
Table III. Demographic characteristics and sexual behaviors of junior respondents on pre-test
Variables

Juniors

aThere

Control
[% (N = 329)]

14.1
48.5
51.5
93.8
6.2

14.2
42.3
57.7
97.0
3.0

25.2
15.1
59.7

23.3
15.5
60.7

21.8
15.5
62.7

22.1
12.6
65.3

54.5
45.5

46.2
53.8

45.2
54.8

50.0
50.0

was a signicant difference between experimental and control groups P < 0.05.
among those who are sexually active.

bPercentage

Table IV. Senior respondents: LSmean scores on psychosocial variables (ANCOVA)

Variablesa

Experimental group (N = 147)

Pre-test

Postponing sexual intercourse

Attitude direct measure
Attitude indirect measure
Perceived behavioral control direct measure
Personal normative beliefs
Role beliefs
Perceived self-efcacy
Condom use
Perceived behavioral control indirect measure
Perceived self-efcacy
aAll

3.24
3.65
3.45
3.09
3.93
3.89
4.43
4.16

Post-testb
3.38
3.80
3.91
3.11
4.16
4.19
4.62
4.46

scales vary from 1 (strongly disagree) to 5 (strongly agree); sex = 1 (male).

at 9-month follow-up.

bPost-test

192

Control group (N = 159)

Pre-test

Post-testb

3.08
3.47
3.40
2.85
3.83
3.76

2.99
3.31
3.48
2.64
3.71
3.71

0.001
0.001
0.001
0.01
0.001
0.001

4.29
3.94

4.12
3.84

0.001
0.001

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Age (years)
Male
Female
Respondent born in Quebec or Canada
Respondent born outside of Quebec or Canadaa
In the last 3 months, did you postpone sexual intercourse
with a boy (girl)?
No
Yes
I have not dated a boy (girl) in the last 3 months
In the last 3 months, did you have sexual intercourse with a boy
(girl) (penetration of the penis in the vagina or the anus)?
No
Yes
I have never had sexual intercourse
In the last 3 month, when you had sexual intercourse with the boy
(girl) you are dating (your boyfriend, girlfriend), how often
did you use a condom?b
Not always
Always (100% of the time)
In the last 3 months, when you had sexual intercourse with a boy
(girl) who is not your boyfriend (girlfriend) (i.e. a one night stand),
how often did you use a condom?b
Not always
Always (100% of the time)

Experimental
[% (N = 369)]

Impact of peer education on adolescents' safe sex behaviors

Table V. Senior respondents: LSmean scores on psychosocial variables (JohnsonNeyman technique)
Variablesa

Experimental group (N = 147)

Pre-test

Post-testb

Pre-test

Post-testb

4.60
4.75
4.66
4.00
4.00
4.33
4.00

4.44
4.68
4.70
3.91
4.59
4.68
4.45

4.60
4.75
4.66
4.00
4.00
4.33
4.00

4.21
4.49
4.42
3.63
4.26
4.46
4.09

0.05
0.05
0.01
0.03
0.01
0.05
0.02

aAll

scales vary from 1 (strongly disagree) to 5 (strongly agree); sex = 1 (male).

at 9-month follow-up.
impact of the program was greater among respondents with the lowest pre-test scores. The means reported are the results
obtained when the impact of the program became signicant.
dThe impact of the program was greater among senior females than among senior males.
bPost-test
cThe

Table VI. Junior respondents: LSmean scores on psychosocial variables (ANCOVA)

Variablesa

Postponing sexual intercourse

Intention
Attitude direct measure
Perceived self-efcacy
Role beliefs
Anticipated regret
Condom use
Intention
Roles beliefs
aAll

Experimental group (N = 369)

Control group (N = 329)

b

P
b

Pre-test

Post-test

3.48
3.53
4.12
4.11
3.44

3.09
3.17
3.84
3.99
3.11

2.97
3.07
3.84
3.76
2.99

0.001
0.001
0.001
0.001
0.001

4.70
4.75

4.63
4.69

4.51
4.58

0.001
0.001

Pre-test

Post-test

3.29
3.24
3.84
4.00
4.00
4.49
4.57

scales vary from 1 (strongly disagree) to 5 (strongly agree); sex = 1 (male).

at 9-month follow-up.

bPost-test

the control group. The experimental group scored

higher on all psychosocial variables related to
condom use. As shown in Table IV (lower portion),
the results indicated a positive change in perceived
behavioral control (P < 0.001; indirect measure)
and perceived self-efcacy (P < 0.001) among the
experimental group. The impact on the remaining
variables was greater among respondents with the
lowest pre-test scores (Table V): attitude (P < 0.05,
direct and indirect measure), perceived normative
beliefs (P < 0.01) and anticipated regret (P < 0.03).

A greater impact on females was also observed

with respect to three variables: perceived behavioral control (direct measure: girls P < 0.001; boys
P < 0.01), role beliefs (girls P < 0.001; boys P <
0.05) and intention (Figure 1: girls P < 0.001; boys
P < 0.02).

Junior respondents: impact on condom use

Nine months after the intervention, experimental
and control groups did not differ in behavior with
respect to condom use. However, compared to the
193

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Condom use
Attitude direct measurec
Attitude indirect measurec
Personal normative beliefsc
Anticipated regret c
Perceived behavioral control direct measurec,d
Role beliefsc,d
Intentionc,d

Control group (N = 159)

F. Caron et al.
Table VII. Junior respondents: LSmean scores on psychosocial variables (JohnsonNeyman technique)
Variablesa

indirect measurec
direct measurec,d

direct measurec
indirect measurec

Control group (N = 329)

Pre-test

Post-testb

Pre-test

Post-testb

4.50
4.61
4.66
5.00

4.11
4.16
3.88
4.36

4.50
4.61
4.66
5.00

3.93
3.97
3.58
3.96

0.03
0.02
0.02
0.02

4.80
4.75
4.76
4.76
5.00
4.65
4.33

4.48
4.54
4.66
4.60
4.76
4.38
4.15

4.80
4.75
4.76
4.76
5.00
4.65
4.33

4.28
4.40
4.49
4.47
4.61
4.20
3.99

0.01
0.02
0.01
0.02
0.02
0.02
0.05

aAll

scales vary from 1 (strongly disagree) to 5 (strongly agree); sex = 1 (male).

at 9-month follow-up.
cThe impact of the program was greater among respondents with the lowest pre-test scores. The means reported are the results
obtained when the effects of the program became signicant.
dThe impact of the program was greater among junior males than among junior females.
bPost-test

remaining results indicated a greater impact among

respondents with the lowest pre-test scores
(Table VII, lower portion): the direct measure of
attitude (P < 0.01), the indirect measure of attitude
(P < 0.02), the direct measure of perceived
behavioral control (P < 0.01), the indirect measure
of perceived behavioral control (P < 0.02), perceived normative beliefs (P < 0.02), perceived selfefcacy (P < 0.02) and anticipated regret (P <
0.05). The impact of the program on males and
females was similar.
Fig. 1. Scale varies from 1 (strongly disagree) to 5 (strongly
agree). The short vertical bar indicates that the impact of the
program was signicant among male respondents who had
lower values.

control group, the experimental group scored

higher on all psychosocial variables related to
condom use. As shown in Table VI (lower portion),
the results indicated a positive change in respondents in the experimental group regarding intention
(P < 0.001) and role beliefs (P < 0.001). The
194

Discussion
The evaluation results indicate the Protection
Express Program had a positive effect on most
evaluated psychosocial variables related to postponing sexual intercourse, and related to condom
use among both senior peer educators and junior
respondents. The former were also more likely to
use condoms on a regular basis than senior
respondents in the control group. All of these
effects were still observable 9 months following

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Postponing sexual intercourse

Attitude indirect measurec
Perceived behavioral control
Personal normative beliefsc
Perceived behavioral control
Condom use
Attitude direct measurec
Attitude indirect measurec
Perceived behavioral control
Perceived behavioral control
Personal normative beliefsc
Perceived self-efcacyc
Anticipated regretc

Experimental group (N = 369)

Impact of peer education on adolescents' safe sex behaviors

These positive effects of the Protection Express
Program among both male and female junior
respondents may be related to the approach utilized
and the relevance of the content based on the
theories of Planned Behavior and Interpersonal
Behavior. Presentations prepared by peers were
transmitted in a language that was age and culture
appropriate, and teaching methods employed such
as sketches and role-plays were interesting and
pertinent to both sexes because they were created
by a mixed team of peer educators. The presentations were also highly structured since all activities
prepared by the peer educators were laid out in a
detailed lesson plan. Previous research suggests
that strong classroom organization can enhance the
efcacy of peer education, particularly if it is
enacted in the presence of high positive regard for
peer educators (Ozer et al., 1997).
In most evaluated programs, peer educators
were opinion leaders who were selected after an
interview or referral process (Smith et al., 2000). In
contrast, the peer educators in the Protection
Express Program were not screened. Our ndings
suggest that slightly older, regular students can be
effective peer educators if they are well trained. In
this regard, peer educators in the Protection
Express Program received substantial training
(i.e. 25 hours), and much support and supervision
from the adults involved.
Even though a positive response to the
Protection Express Program occurred among both
male and female respondents, a greater impact on
perceived behavioral control with respect to postponing sexual intercourse was observed among
junior boys, and a greater impact on perceived
behavioral control, role beliefs and intention
related to condom use was observed among senior
girls. Moreover, the greatest response on many
psychosocial factors occurred among respondents
with the lowest pre-test scores. This suggests that
subjects who responded the most to the Protection
Express Program were those who were perhaps
most vulnerable to sexually transmissible infections as evidenced by weak scores on most
psychosocial factors related to safer sexual conduct. Overall, these ndings do not support the
195

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

the intervention, and both males and females

showed a signicant positive change.
As these results suggest, involvement by peer
educators in the development of creative and
educational activities for others brought about a
positive change in their own behavior with respect
to condom use. Although past research has found
peer education to be effective, few evaluations
have measured the impact of this type of program
on the behavior of the peer educators themselves.
In one such study, exposure to a 36-hour theorybased training program resulted in a positive longterm impact on perceived self-efcacy and consistency of condom use (Smith et al., 2000). Our
results are in keeping with this study, and lend
support to the mounting evidence regarding the
effectiveness of programs that are theory-based
and employ teaching methods designed to involve
participants and have them personalize the information (Kirby, 2001).
However, senior peer educators and junior
respondents in the experimental groups did not
modify their behavior with respect to postponing
sexual intercourse following exposure to the
Protection Express Program. These results are
similar to those of other program evaluations
where sexual intercourse was not delayed but a
positive impact was found on condom use (Coyle
et al., 1999). In fact, evidence for the effectiveness
of abstinence programs is meager (Kirby, 2001). In
another literature review, only six of the 19 nonpeer-led programs examined were found to have an
impact on postponing sexual intercourse (Santelli
et al., 1999).
The impact of the Protection Express Program
among junior respondents was encouraging, given
the entire peer intervention lasted only 150180
min. Previous research suggests that brief peer
interventions can modify psychosocial variables
related to condom use (Dunn et al., 1998). The lack
of impact on behavior regarding condom use was
not surprising, given the short duration of the
intervention and the fact that being younger, they
were, as a group, less sexually active than their
older peers.

F. Caron et al.

Conclusion
To conclude, the Protection Express Program
produced positive results over a period of 9 months
among both male and female peer educators, and
their target audience of junior high school students.
196

Also, compared to the control groups, the peer

educators were more likely to use condoms on a
regular basis. The results of this study suggest that
this well-designed theory-based, highly structured
and supervised AIDS/STD peer education program
was effective among adolescents in a high school
setting.

Acknowledgements
F. C., G. G. and J. O. are members of a provincial
research team funded by the Quebec Council for
Social Research (CQRS). This project was funded
by a grant from the Quebec Council for Social
Research RS-3251.

References
Aarons, S.J., Jenkins, R., Raine, T.R., El-Khorazaty, M.N.,
Clark, M.C. and Wingrove, B.K. (2000) Postponing sexual
intercourse among urban junior high school studentsa
randomized controlled evaluation. Journal of Adolescent
Health, 27, 236247.
Ajzen, I. (1991) The theory of planned behavior. Organization
Behavior and Human Decision Processes, 50, 179211.
Anderson, C., Cheadle, A., Curry, S., Diehr, P., Shultz, L. and
Wagner, E. (1995) Selection bias related to parental consent
in school-based survey research. Evaluation Review, 19,
663674.
Bandura, A. (1986) Social Foundations of Thought and Action:
A Social Cognitive Theory. Prentice-Hall, Englewood Cliffs,
NJ.
Caron, F., Otis, J. and Pilote, F. (1998) Evaluation of an AIDS
peer education program on multiethnic adolescents attending
an urban high school in Quebec Canada. Journal of HIV/
AIDS Education and Prevention for Adolescents and
Children, 2, 3153.
Carvajal, S.C., Parcel, G.S., Basen-Engquist, K., Banspach,
S.W., Coyle, K.K., Kirby, D. and Chan, W. (1999)
Psychosocial predictors of delay of rst sexual intercourse
by adolescents. Health Psychology, 18, 443452.
Centers for Disease Control and Prevention (1998) Trends in
sexual risk behaviors among high school studentsUnited
States, 19911997. Morbidity and Mortality Weekly Report,
47, 749752.
Centers for Disease Control and Prevention (2001a) Chlamydia
in the United States. National Center for HIV and TB
Prevention, Division of Sexually Transmitted Diseases,
CDC, Atlanta, GA.
Centers for Disease Control and Prevention (2001b) US HIV
and AIDS cases reported through June 2001. HIV/AIDS
Surveillance Report, 13, 141.
Cook, T.D. and Campbell, D.D. (1979) Quasi-experimentation:
Design and Analysis Issues for Filed Settings. Rand
McNally, Chicago, IL.

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

need for separate, gender-specic interventions as

suggested by Aarons et al. (Aarons et al., 2000),
who found gender differences in various study
outcomes among adolescents exposed to the peerled `Postponing Sexual Involvement Program'.
To ensure the effectiveness of the Protection
Express Program, teachers and nurses must be well
trained to stimulate cooperative learning among
peer educators and supervise the quality of their
presentations to junior students. Thus, adequate
training and access to a competent program
consultant are absolutely required. It is also
important to create a balanced team of peer
educators, when working with regular students
who are not necessarily opinion leaders.
Homogenous groups should be avoided, i.e. putting all the students with strong social skills
together on the same team. Reinforcement by
teachers of the messages left by peers is needed to
increase the impact on behavior among junior
students.
Finally, the quasi-experimental design utilized
does not allow generalization of these results to
other groups of adolescents; too many factors may
not have been fully controlled. Also, the number of
groups involved in the study was small and may
not be representative of the whole student population. Careful interpretation of these results is also
required because of differences found between
students lost to the program and students who
remained in the nal sample. In addition, the
required parental consent procedure may have
introduced bias into the sample of junior respondents as students who obtained parental consent may
have a different sociodemographic prole and
differ with respect to their risk behavior than those
who did not receive parental consent (Anderson
et al., 1995).

Impact of peer education on adolescents' safe sex behaviors

Laboratory Center for Disease Control (1999) Genital
chlamydia in Canada. STD Epi Update, May, 15.
Mellanby, A.R., Rees, J.B. and Tripp, J.H. (2000) Peer-led and
adult-led school health education: a critical review of
available comparative research. Health Education
Research, 15, 533545.
Ozer, E.J., Weinstein, R.S., Maslach, C. and Siegel D. (1997)
Adolescent AIDS prevention in context: the impact of peer
educator qualities and classroom environments on
intervention efcacy. American Journal of Community
Psychology, 25, 289323.
Santelli, J.S., DiClemente, R.J., Miller, K.S. and Kirby, D.
(1999) Sexually transmitted diseases, unintended pregnancy
and adolescent health promotion. Adolescent Medicine: State
of the Art Reviews, 10, 87108.
Smith, M.U., Dane, F.C. Archer, M.E., Devereaux, R.S. and
Katner, H.P. (2000) Students Together Against Negative
Decisions (STAND): evaluation of a school-based sexual
risk reduction intervention in the rural south. AIDS
Education Prevention, 12, 4970.
Song, E.Y. and Pruitt, B.E. (2000) A review of sexuality
education programs for adolescents, 19601997. Journal of
Health Education, 31, 225230.
Richard, R., De Vries, N.K. and Van Der Pligt, J. (1998)
Anticipated regret and precautionary sexual behavior.
Journal of Applied Social Psychology, 28, 14111428.
Tapert, S.F., Aarons, G.A., Sedlar, G.R. and Brown, S.A.
(2001) Adolescent substance use and sexual risk-taking
behavior. Journal of Adolescent Health, 28, 181189.
Triandis, H.C. (1980) Values, attitudes and interpersonal
behavior. In Page, M.M. (ed.), Nebraska Symposium on
Motivation. Beliefs, Attitudes and Values. University of
Nebraska, Lincoln, NE, vol. 1, pp. 195259.
Turner, T.J. (2001) Using psychosocial factors and the theory
of reasoned action to predict sexual behavior in adolescent
females. Dissertation Abstracts International A, 61, 95011
95163.
Received on June 27, 2002; accepted on January 21, 2003

197

Downloaded from http://her.oxfordjournals.org/ by guest on November 29, 2015

Coyle, K., Basen-Engquist, K., Kirby, D., Parcel, G., Banspach,

S., Harrist, R., Baumler, E. and Weil, M. (1999) Short-term
impact of safer choices: a multicomponent, school-based
HIV, other STD and pregnancy prevention program. Journal
of School Health, 69, 181188.
DiIorio, C., Dudley, W.N., Kelly, M., Soet, J.E., Mbwara, J.
and Sharpe Potter, J. (2001) Social cognitive correlates of
sexual experience and condom use among 13 through 15year-old adolescents. Journal of Adolescent Health, 29, 208
216.
Dunn, L., Ross, B., Caines, T. and Howorth, P. (1998) A
school-based HIV/AIDS education program: outcomes of
peer-led versus community health nurse-led interventions.
Canadian Journal of Human Sexuality, 7, 339345.
Everrett, S.A., Warren, C.W., Santelli, J.S., Kann, L., Collins,
J.L. and Kolbe, L.J. (2000) Use of birth control pills,
condoms and withdrawal among US high school students.
Journal of Adolescent Health, 27, 112118.
Godin, G., Fortin, C., Michaud, F., Bradet, R. and Kok, G.
(1997) Use of condoms: intention and behaviour of
adolescents living in juvenile rehabilitation centres. Health
Education Research, 12, 289300.
Huitema, B.E. (1980) The Analysis of Covariance and
Alternatives. Wiley, New York.
Jemmott, J.B., III, Sweet Jemmott, L. and Fong, G.T. (1998)
Abstinence and safer sex HIV risk-reduction interventions
for African American adolescents. Journal of the American
Medical Association, 279, 15291536.
Kann, L., Kinchen, S.A., Williams, B.I., Ross, J.G., Lowry, R.,
Grunbaum, J.A. and Kolbe, L.J. (2000) Youth risk behavior
surveillanceUnited States, 1999. Morbidity and Mortality
Weekly Report, 49, 132.
Kingree, J.B., Braithwaite, R. and Woodring, T. (2000)
Unprotected sex as a function of alcohol and marijuana use
among adolescent detainees. Journal of Adolescent Health,
27, 179185.
Kirby, D. (2001) Emerging Answers: Research Findings on
Programs to Reduce Teen Pregnancy. National Campaign to
Prevent Teen Pregnancy, Washington, DC.