VIRAJ

EVALUATION OF A PREVENTION PROGRAM
FOR VIOLENCE IN DATING RELATIONSHIPS
INVOLVING HIGH SCHOOL STUDENTS

Fancine Lavoie, Martine Hébert
Université Laval,
Québec, Canada

Communication presented at the
American Educational Research Association Annual Meeting
Montréal, Québec
April 22, 1999  

For more information:
Francine Lavoie,  École de psychologie, Université Laval, Québec, Qué., Canada G1K 7P4
E-mail: francine.lavoie@psy.ulaval.ca  

Research funded by Le Conseil québécois de la recherche sociale, Gouvernement du Québec.

The elimination of violence is a national priority in both Canada and the United States. One of the too few evaluations of a program to prevent dating violence during adolescence is presented here.  It is a three-year evaluation study using a patched-up cohort design, involving an assortment of comparison groups of Grade 10 students in two schools.  817 students were involved in the different comparison groups.  The objective of the evaluation were to test, with a quasi-experimental design, the net effect of the program, clear of the effects of maturational trends, interfering events and pretesting.  The outcomes were measured after five weeks and four months; and for a subsample, after 13 months.  The results of multiple analysis of variance show that the 2 1/2-hour program influences attitudes toward dating violence, even 13 months after completion, with rival hypotheses taken into account.  Students obtained better results after the intervention.  An exploratory analysis was done with female victims in a dating relationship and male aggressors; and results indicate that the program was also efficient for those at-risk populations on the attitudes scale.  In general, the program was not conductive to fostering changes in behavioral intentions, in self-efficacy of control of dating violence and or in peers approval.   A second objective was to verify, through interviews, the influence of the program six months later. 

The elimination of violence is a national priority in both Canada and the United-States.  The school setting is often considered to be one of the best levels of intervention.  One of the too few evaluations of a program to prevent dating violence during adolescence is presented here.

A) Objectives

The objective of the evaluation was to test, with a quasi-experimental design, the net effect of the program, clear of the effects of maturational trends, interfering events and pretesting, and to measure medium- term influences.  A second objective was to verify, through interviews, the influence of the program six months later. 

B) Perspectives

Epidemiological data indicates that in Canada and in the United States, more than 20% of all young people will have experienced incidents of dating violence before finishing high school (Foshee, 1996;  Malik, Sorenson & Aneshensel, 1997;  O'Keefe, 1997;  Poitras & Lavoie, 1995).   It is therefore imperative that measures be taken to eliminate this violence. This paper will describe the evaluation of a program to prevent violence and foster equal-partner relationships among 14- to 16-year-old adolescents going to school.  The program is different from the majority of programs available in the field, because it centers exclusively on dating relationships and considers themes of control and emotional, physical, and sexual abuse.  The program, which was initially developed by a team of researchers in psychology, has been endorsed by the Province of Quebec's Ministry of Education and has also been adopted by other French-speaking communities in Canada. An English version is also available.  An initial evaluation without a control group conducted with 517 students confirmed the program's effectiveness to in modifying attitudes (Lavoie, Piché, Boivin, & Vézina, 1995).

C) Methods

Content and objectives of the program

The school program (named VIRAJ) consists of two classroom sessions (total program duration:  150 minutes).

The objectives of the first session are:

1)         to distinguish self-control or control over one's environment from the abusive control of other people;

2)         to identify and denounce different forms of control, including physical and social control and emotional blackmail;

3)         to understand the importance of the problem of violence in dating relationships. 

The objectives of the second session are:

1)         to establish certain rights of each partner in a dating relationship;           

2)         to learn how to apply these rights in situations with a risk of abuse;        

3)         to learn that each partner is responsible for respecting the other's rights;

4)         to understand that the responsibility for abuse must not be attributed to the victim, but rather to the perpetrator.

Other resources available are:

-           a training program for first-line practitioners to identify and support victims and aggressors looking for help

-           a training session for schools wishing to implement the program

-           a videotape and posters.

Evaluation goal

-           Using questionnaires, assess the impact that the VIRAJ program has on the short and medium term, meaning five weeks, four months and in certain cases thirteen months after the end of the program.

-           Using interviews with a selection of students including some who had experienced violence, verify what is learned.

Participants

The study was conducted in two Quebec City area high schools. The mean age was around 15 years old, and 817 students were involved in different comparison groups.  Forty-eight young students were interviewed.

Design

-           a three-year evaluation study with a quasi-experimental design

-           a patched-up cohort design, involving an assortment of comparison groups

-           control of the effects of:

                        . pretesting

                        . maturational trends

                        . interfering events

-           measurement with validated instruments at 5 five weeks, 4 four months and for a subsample, 13 months after the program

-           assessment of attitudes, behavioral intentions, self-efficacy in preventing their own involvement in dating violence, peer approval of their involvement in dating violence

-           interviews completed six months after the completion of the program with a selection of students

A description of the quasi-experimental design:

P = Program
R = Random assignment
Oi=1 to 7  = test

Measures

Attitudes

A 34-item attitudes scale was built, based on results of work undertaken previously, and on the work of other researchers.  A Likert scale was used, with four response alternatives (ranging from “completely agree” to “completely disagree”).  Higher scores indicated attitudes against dating violence.  The alpha with a group of 997 students from a rural area was 0.86 (Lavoie et al., 1997).

Behavioral intentions as a victim/as a witness

Two items represented behavioral intentions:   "If I were the victim of violence inflicted by my partner, I would keep it secret" and "If I were to observe violent acts within a couple, I would tell them that I disagree with the use of violence.”  The same Likert scale was used. 

Peer approval

A 9-item scale of peer approval of violence in dating relationships was constructed.  Students had to evaluate how their peers would approve of them if they bullied their intimate partner in 9 couple-conflict situations.  Alpha was 0.86 with a group of 246 students from a rural area and 0.74 with a group of 971 students from urban areas (Lavoie et al., 1997).  Higher scores indicated that students perceive that their peers would approve them. 

Self-efficacy of control of dating violence

The same nine couple-conflict situations were presented and students had to evaluate how easy or how difficult it would be for them not to react with violence.  Alpha was 0.89 with a group of 250 students from a rural area and 0.89 with a group of 977 students from urban areas (Lavoie et al., 1997).  Higher scores indicated that students perceived that they demonstrated good self-efficacy of control of dating violence. 

Dating violence

A questionnaire on dating violence (psychological, sexual and physical) was developed to identify male aggressors and female victims (description of the instrument:  Lavoie & Vézina, submitted).

D) Results

Please note that in the following sections no distinction is made between the results obtained from males and females, because in all analyses of our main measurement, the attitudes scale, results indicated that even if females on a whole scored higher than the males, there was no interaction effect between gender and comparison situations. In other words, the same trends were observed for females and males in comparable situations.

Test of the equivalence of groups

Before submitting the data to the analyses of variance, the equivalence of groups was tested. Even if a cohort design presumes that there is a similarity between students of three different years in a same school when there is no documented changes in the institution recruitment policy, statistical tests could be used to verify this equivalence for cohorts 2 and 3.  On the other hand, equivalence of groups had to be checked between Schools 1 and 2 before comparing the students (Cook & Campbell, 1979).  Results (Table 1) indicate that in pretest (O2), Cohort Group 2 showed results comparable to those of the third cohort in  pretest (O5) and students of the second school during the first questionnaire (O6).  Therefore, one may conclude equivalence among the different comparison groups.

Table 1

Test of equivalence

Pretesting effects

The effect of the pretest was verified with the second cohort.  Half of the students were given an evaluation questionnaire during the pretest and the other half completed a questionnaire on a different subject (teachers at the school). The results obtained from both groups of students were the same during the post-test, indicating that the pretest had no effect. Moreover, post-test results of students not having completed the pretest were higher than the results obtained in the pretest of the other group of students in the attitudes scale.

Table 2

Pretesting effects

Impact  five weeks after the program

The impact of the program was verified through a series of comparisons between the information gathered during pretests and during post-tests.  It may be concluded that on four occasions the program was deemed efficient, but only with respect to attitudes. The next sections will address attitude measurement only.

Table 3

Impact five weeks after the program