Article Text

Download PDFPDF

Review: school based child sexual abuse prevention programmes are effective for improving prevention skills and knowledge
  1. Peter Fonagy, PhD, DipPsych FBA
  1. University College London London, UK

Statistics from

OpenUrlCrossRefPubMedWeb of Science

QUESTION: Are school based programmes effective for increasing children's skills and knowledge in preventing sexual abuse?

Data sources

Studies were identified by searching PsycINFO, Medline, ERIC, HealthSTAR, and Dissertation Abstracts; scanning bibliographies of relevant articles; and handsearching recent volumes of relevant journals (eg, Child Abuse and Neglect).

Study selection

Studies were selected if they evaluated a school based sexual abuse prevention programme in children 3 to 13 years of age, rated effectiveness using a knowledge based or behavioural outcome measure, included a control group, and were presented in English.

Data extraction

Data were extracted on age, mode of presentation, (eg, written material, puppets, role plays), level of child participation, (ie, physical, verbal, or no participation), qualifications and sex of instructor, programme duration, number of sessions, time period between programme and post-test, type of outcome measure used (ie, behavioural or knowledge based), and methodological quality. Effect sizes were calculated for post-test comparisons of control and treatment groups.

Main results

27 studies (8115 participants) met the selection criteria (21 journal articles, 3 unpublished doctoral dissertations, 2 unpublished internal programme evaluations, and 1 unpublished masters thesis). Only 2 studies used behavioural outcome measures, and there were no studies on abuse prevalence. 73 separate effect sizes were calculated (range 0.1 to 2.9), and all but 1 showed a benefit for prevention programmes. The weighted mean effect size was 1.07, which means that {36% of children in the control group did worse than the average child in the prevention programme group}*. When a large study (n=1339) with the largest effect size (2.93) was excluded from the analysis, the overall mean weighted effect size decreased to 0.81. Multiple regression showed that methodological factors such as type of control group, research design, blinding of interviewers, random assignment, and number of items on the outcome measure were significant predictors of effect size and accounted for 43.5% of the total variance. Studies with poorer methodology generally had larger effect sizes. When only randomised controlled trials were included in the analysis, the effect size was 0.73. Programmes presenting >4 sessions that allowed children to become physically involved produced the highest effect sizes.


School based child sexual abuse prevention programmes are effective for increasing children's prevention skills and knowledge.


The review by Davis and Gidycz concerns mainly US studies on the prevention of child sexual abuse. Many states in the US now mandate educational programmes, including role playing, participant rehearsals, films, and lectures aimed at improving prevention knowledge and skills. This carefully conducted and well reported meta-analysis shows that, on the whole, educational programmes aimed at enhancing knowledge and skills that might help children to avoid victimisation achieve what they set out to do. Wisely excluding a large, poorly controlled study which unduly influenced the results, the analysis revealed that longer (≥4 sessions) programmes that involved physical participation (skills training) by using role play or drama and behavioural training were particularly effective on behavioural and questionnaire outcome measures in 5–8 year old children. Children >8 years of age are relatively less affected by such programmes because pre-post test differences are likely to be reduced by greater knowledge at pre-test. We should be cautious in interpreting the findings because nearly one half of the variance in effect size could be accounted for by methodological quality, although this included the aberrant study. As usual, studies with poorer methodology had larger effect sizes and thus the overall effect size may be somewhat inflated. Nevertheless, these findings are encouraging and suggest that prevention programmes should be implemented as part of universal prevention initiatives.

Of course, the literature on attitude change in healthcare contexts is clear that bringing about changes in knowledge or skills does not necessarily result in behavioural change. Thus, improvements in prevention related knowledge may not translate into reduction of abuse prevalence. Because two thirds of children in the US are likely to be exposed to a prevention programme, one might anticipate a dramatic decline in sexual abuse if these programmes were truly effective. Studies that attempt to create in vivo simulations with strangers show that children's risk taking behaviour is diminished by prevention programmes. Such in vivo studies involve strangers, whereas threat tends to come from adults who are known to the child. The extent to which school based prevention programmes generalise to these contexts remains unclear.


  • Source of funding: not stated.

  • For correspondence: Dr C A Gidycz, Department of Psychology, Ohio University, 231 Porter Hall, Athens, OH 45701, USA. Fax +1 740 593 0579.

  • *Number calculated from mean effect size.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.