Familybased Prevention of Offending A MetaAnalysis David P
Family-based Prevention of Offending: A Meta-Analysis David P. Farrington & Brandon C. Welsh 2003 Jenna Ayers Radford University
Overview w Review of the effectiveness of family-based prevention programs in reducing offending and antisocial behavior by children and adolescents. w Family-based programs typically target family risk factors. w Goal: aims to prepare and maintain systematic reviews and to make them available electronically.
Previous Research w Narrative reviews of effect of family-based interventions to prevent delinquency and later criminal offending w Serketich & Dumas (1996) w Cochrane review completed by Woolfenden, Williams, & Peat (2002)
Systematic Review Why? w Use rigorous methods for: locating, appraising, and synthesizing evidence from evaluation studies w Explicit objectives w Explicit criteria for inclusion and exclusion of studies w Wide-ranging methods for searching for studies that are designed to reduce bias
Method w Selection of Evaluations (Inclusion) w 1. The family and family factors w 2. Outcome measure of delinquency or antisocial child behavior w 3. High quality methodologically w 4. Original sample size
Searching Strategies w Original aim: to update the review of family-based crime prevention in the 1997 Maryland Report w Began searches in 1997 w 1. Recent reviews w 2. Articles in major journals w 3. Youth Update w 4. Contacts
Measuring Effect Size w Aimed to measure 4 different effects in each study: w 1. Short-term effect on delinquency w 2. Short-term effects on child antisocial behavior w 3. Long-term effects on offending w 4. Long-term effects on antisocial behavior
Main Measure of Effect Size The standardized mean difference d, which summarizes the difference between the experimental and control groups in standard deviation units: w D= (Mc-Me)/s w Positive value of d indicates a desirable effect of the intervention
Family-based Prevention Programs w Home Visiting (4) w Day Care/Preschool programs (5) w Parent Training programs (10) w School-based programs (7) w Home/Community programs with Older Children (8) w Multi-systemic Therapy Programs (6)
Meta-Analysis Mean Effect Sizes Category A B C D E F Total Low CI. 111. 147. 274 -. 014. 056. 281. 278 High CI. 360. 371. 517. 149. 306. 548. 269 Mean. 235. 259. 395. 068. 181. 414. 223 Sig 2/4 4/5 5/10* 3/7 3/8 2/6* 19/40*
Mean Effect Sizes Delinquency Short-term Long-term Total Low CI. 132. 281. 250 High CI. 315. 464. 391 Mean. 224. 372. 321 Sig 5/12 6/10 10/19* . 157. 016. 146 . 252. 201. 246 . 204. 109. 196 11/26 2/7 11/27 Antisocial behavior Short-term Long-term Total
Results of Meta Analysis w Suggest that prevalence of offending could be reduced by about 10 -15% by implementing such programs. w More than half of all evaluations found a significant decrease in delinquency w Effects on delinquency persisted in long-term evaluation studies
Results w Most effective types of programs used behavioral parent training w Least effective were those based in schools w All other types of family-based programs were effective
Conclusions w 40 of the highest quality family-based crime prevention programs were reviewed w Programs grouped into 6 categories w These family-based programs had desirable effects in reducing delinquency and antisocial child behavior
Future Research w More large-scale evaluations are needed using randomized experiments w Ideally, programs focusing more clearly and more narrowly on family risk factors should be implemented and evaluated w More efforts should be made to determine links in the causal chain between family processes and offending w More long term follow ups should be carried out to establish the persistence of effects
Future Research w Important to investigate why effect sizes are greater in smaller scale studies than in larger scale ones. w Future experiments needed that attempt to disentangle the different elements of successful programs w Know more about the economic efficiency of family-based crime prevention programs
Bottom Line w Existing evidence suggests that family-based prevention programs are effective in reducing offending. w More of these types of programs should be implemented and evaluated.
- Slides: 18