CrimeSolutions.gov

Additional Resources:

Practice Profile

School-Based Child Sexual Abuse Prevention Programs

Evidence Ratings for Outcomes:

Effective - More than one Meta-Analysis Attitudes & Beliefs - Knowledge
No Effects - One Meta-Analysis Victimization - Sexual abuse/exploitation

Practice Description

Practice Goals/Target Population
School-based child sexual abuse (CSA) prevention programs are designed to reduce the occurrence of sexual abuse in children and adolescents. The prevention programs seek to improve students’ knowledge and skills in order to help them avoid unsafe situations and report incidences of sexual abuse. The programs target school-aged children and teach them about CSA and how to protect themselves from it.

Unfortunately, there is no agreed-on definition of CSA. Some research defines sexual abuse as instances of sexual body contact with a child, while other definitions include any sexual behavior in a child’s presence (Zwi et al. 2007). CSA may include child prostitution, child pornography, child molestation, child exposure to sexually explicit materials, and attempted or threatened sexual abuse (Chen, Fortson, and Tseng 2012).

Experiencing sexual abuse can have serious short- and long-term repercussions for children. For example, individuals with a history of CSA are at higher risk for alcohol and substance use, suicidal attempts, risky sexual behavior, and mental health difficulties, such as depression or posttraumatic stress disorder (Chen, Fortson, and Tseng 2012). School-based CSA prevention programs are one strategy that can be used to teach children the knowledge and skills to avoid and report abuse.

Practice Components
Generally, school-based CSA prevention programs teach children information about appropriate or inappropriate touches and ways to identify abusive situations, how to refuse inappropriate touches or get out of abusive situations, and how to tell a trusted adult about the incident. Although there are some basic concepts common to all CSA prevention programs, the programs can differ on numerous components, such as the format or method of presentation, the age of the targeted population, and the length of the program.

CSA prevention programs can use a variety of methods to present the information to children. Programs can be divided into two categories: information-based training and behavioral skills training (Chen, Fortson, and Tseng 2012). With behavioral skills training, the modes of presentation can be more active—for example, role playing or active skills rehearsal of self-protection skills, and modeling of skills by the presenter. With information-based training, the modes of presentation can be less active and include using methods such as videos or films, written materials, discussions or lectures, puppet shows, and dolls or stuffed animals.

The targeted population of the school-based prevention programs can also vary. Some programs target children as young as 3 in preschool, while other programs target older youths, such as those in fifth or sixth grade. CSA prevention programs usually target children in kindergarten and elementary school; only a few studies have looked at programs targeting adolescents in middle or high school (Zwi et al. 2007).

School-based CSA prevention programs tend to be implemented by teachers, but they may also be delivered by mental health professional, psychologists, college students, police officers, researchers, or other types of volunteers.

Meta-Analysis Outcomes

top border
Effective - More than one Meta-Analysis Attitudes & Beliefs - Knowledge
Looking at the results from 26 studies evaluating school-based sexual abuse–prevention programs, Davis and Gidycz (2000) found a significant and large effect size of 0.81. This means that children who participated in prevention programs performed significantly higher on measures of prevention-related knowledge and skills compared with children in the control groups. Similarly, Zwi and colleagues (2007), when looking at the results from nine randomized controlled trials, found an effect size of 0.59, showing a significant increase in knowledge of children in the prevention programs compared with control group children.
No Effects - One Meta-Analysis Victimization - Sexual abuse/exploitation
Zwi and colleagues (2007) aggregated the results from two randomized controlled trials and found a nonsignificant effect size (odds ratio=4.8). This means there was no significant difference in the odds of children disclosing or reporting their sexual abuse following participation in school-based programs. However, this result should be interpreted with caution because of the small number of studies that were analyzed.
bottom border

Meta-Analysis Methodology

top border
Meta-Analysis Snapshot
 Literature Coverage DatesNumber of StudiesNumber of Study Participants
Meta-Analysis 11985 - 1996278115
Meta-Analysis 21986 - 2003154347

Meta-Analysis 1
Davis and Gidycz (2000) conducted a meta-analysis of evaluations that looked at the effectiveness of school-based child abuse–prevention programs. Studies were included in the analysis if they 1) evaluated a school-based prevention program focusing on sexual abuse, 2) examined children ages 3 to 13, 3) rated effectiveness of the program using a knowledge-based or behavioral outcome measure, 4) included a control group comparison, and 5) were written in English. The research team conducted a comprehensive search of databases, article references, and relevant journals.

Twenty-seven studies that collectively included more than 8,000 participants were located for inclusion in the review. Of the 27 studies, 21 were published journal articles, 3 were unpublished doctoral dissertations, 2 were unpublished internal program evaluations, and 1 was an unpublished master’s thesis. The age groups ranged from 3 to 12 years old. However, no other information was provided on study characteristics or the characteristics of study participants.

The outcome of interest was prevention-related knowledge and skills of children. The outcomes from the studies were measured using questionnaires, vignette interviews, or behavioral observations. Effect sizes and their variances were calculated as d values for posttest comparison of control and treatment groups. The calculations used one effect size for each individual construct presented in each study. However, one study (Nemerofsky, Carran, and Rosenberg 1994) was removed from the analysis because of its unusually large effect size and very large sample size, causing the study to be heavily weighted. Moderator variables were also examined by breaking the studies into subsets based on the proposed variable and conducting a meta-analysis on each subset separately.

Meta-Analysis 2
Zwi and colleagues (2007) conducted a meta-analysis to assess whether school-based prevention programs were effective in improving knowledge about sexual abuse and self-protective behaviors, and whether participation resulted in an increase in disclosure of sexual abuse. The review included randomized control trials and quasi-experimental studies of school-based interventions designed to prevent child sexual abuse compared with another intervention or no intervention. Studies included in the analysis had to include at least one standardized outcome measure for both the intervention and control groups, preintervention and postintervention.

The authors conducted a comprehensive search of databases (such as PsychINFO and ERIC). To identify unpublished studies, the research team contacted experts in the field by letter and searched the National Child Protection Clearinghouses for Australia, Canada, the United Kingdom, and the United States. In total, 15 trials were found and included in the analysis. The 15 studies included more than 4,000 participants. The studies varied in the characteristics of participants and locations. All of the studies were randomized controlled trials (RCTs). Some were clustered RCTs, meaning randomization took place at the school or classroom level, not at the student level. The ages ranged from kindergarteners (approximately 5 years old) to sixth graders (approximately 11 years old). One study included 10th graders (but no average age of the study’s sample was reported). Most of the studies included an almost equal number of boys and girls. Twelve of the included studies took place in urban and rural settings in the United States. Two studies took place in Canada, and one took place in China. The race/ethnicity of study participants was not provided.

The authors looked at a variety of outcomes. For this review, the outcomes of interest were knowledge of sexual abuse and abuse prevention concepts (measured by questionnaires) and disclosure of sexual abuse by a child or adolescent during or after participating in the CSA prevention program. The effect sizes were calculated as standardized mean differences, and a random effects model was assumed in the analysis.
bottom border

Cost

top border
There is no cost information available for this practice.
bottom border

Other Information

top border
The meta-analysis by Davis and Gidycz (2000) included additional tests—called moderator analyses—to see whether any factors strengthened the likelihood that school-based child sexual abuse (CSA) prevention programs improved outcomes. For age, the authors looked at three age groups: 3.0 to 5.0 years, 5.1 to 8.0 years, and 8.1 to 12.0 years. The moderator analysis showed that the weighted mean effect size was highest for 3- to 5-year-olds (d=2.143) followed by the second oldest age group (d=1.243) and the oldest age group (d=0.770). This suggests that CSA prevention programs may be most effective on younger children. The review authors also looked at the number of sessions of the prevention programs and found a linear trend. As the number of sessions increased, the effect size increased (ranging from d=0.598 for 1 session to d=1.536 for three or more sessions). They also looked at the mode of presentation (for example, use of written materials, role play, discussion/lecture, video/film, puppet shows, dolls or stuff animals, and behavioral skills training). Most of the modes of presentation did not significantly predict effect size. However, behavioral skills’ training was found to be a significant predictor. Studies with behavioral skills training had a mean effect size twice as high as those without such training (d=1.210 and d=0.663, respectively).
bottom border

Evidence-Base (Meta-Analyses Reviewed)

top border
These sources were used in the development of the practice profile:

Meta-Analysis 1
Davis, M. Katherine, and Christine A. Gidycz. 2000. “Child Sexual Abuse Prevention Programs: A Meta-Analysis.” Journal of Clinical Child Psychology 29(2):257–65.

Meta-Analysis 2
Zwi, Karen, Sue Woolfenden, Danielle Wheeler, Tracey O’Brien, Paul Tait, and Katrina Williams. 2007. “School-Based Education Programs for the Prevention of Child Sexual Abuse.” Campbell Systematic Reviews 5.
http://campbellcollaboration.org/lib/project/28/
bottom border

Additional References

top border
These sources were used in the development of the practice profile:

Chen, Yi–Chuen, Beverly L. Fortson, and Kai–Wen Tseng. 2012. “Pilot Evaluation of a Sexual Abuse Prevention Program for Taiwanese Children.” Journal of Child Sexual Abuse 21:621–45.

Nemerofsky, Alan G., Deborah T. Carran, and Leon A. Rosenberg. 1994. “Age Variation in Performance Among Preschool Children in a Sexual Abuse Prevention Program.” Journal of Child Sexual Abuse 3(1):85–102.

Rispens, Jan, Andre Aleman, and Paul P. Goudena. 1997. “Prevention of Child Sexual Abuse Victimization: A Meta-Analysis of School Programs.” Child Abuse & Neglect 21(10):975–87.
bottom border


Practice Snapshot

Age: 3 - 12

Gender: Both

Settings: School

Practice Type: Classroom Curricula, Victim Programs

Unit of Analysis: Persons