EDUCATION FACTORS: POLICY IMPLICATIONS
Farrington and Welsh (2007) emphasized that one of the most important individual factors predicting offending is low academic attainment, and our findings confirm that academic achievement is a differential predictor of violence. Given that low grades are so consistently found among violent youth and that school predictors of violence have been found to be strong mediators of other risk factors (e.g., Herrenkohl et al., 2001), intervention in the schools in childhood and adolescence has enormous potential for violence prevention.
Our review confirms that low school bonds are associated with both violent and nonviolent offending (with too little evidence comparing offenders to draw a reliable conclusion on this point) and school problems, academic attainment, and parent education are likely to be differential predictors of violence as well.
Of course, there are a huge number of programs and interventions designed to improve academic achievement, and many of them have been successful (e.g., Lipsey & Wilson, 1993). Nonetheless, there are still children and adolescents failing at school, and the media regularly report stories about the poor reading ability and math scores among American children. This suggests a much broader problem with the institution of education in the United States. School failure, though, happens to be concentrated in certain school districts where, not coincidentally, violent crime and other risk factors for violence are also concentrated. Thus, we know where to implement the programs.
School-based programs designed to prevent violence usually employ classroom-based curricula about drug use or violence, teaching social skills, using role playing, and interactive exercises. Some of these programs have been effective in reducing violence and related behavior in the schools (e.g., Howard, Flora, & Griffin, 1999). Examples of programs with “demonstrated effectiveness” are the Good Behavior Game, the Seattle Social Development Project, the Child Development Program, and Fast Track (Loeber & Farrington, 2000). A few educational programs have been evaluated for their effects on delinquency in older children and adolescents. Gottfredson, Wilson, and Najaka (2002) conducted a systematic review of an enormous number of school-based programs designed to prevent an array of problems such as drug use and bullying. Some of the programs, such as PACT (Promoting Achievement through Cooperative Learning) and Fast Track, have had moderate effect sizes on antisocial behavior and aggression. Gottfredson (2001), in her book Schools and Delinquency, points out that, on the whole, school-based prevention programs do not work as well as we might expect them to, given the enormous advantage they have in delivering services. We might speculate, given our own findings, that the problem might be due to the fact that these programs are not addressing the most salient correlates of violence which are academics, perhaps school bonds, and academic attainment (for young people and their parents). Many programs evaluated for their ability to prevent delinquency in school settings (like those described above), have multiple and myriad goals, and their approaches, such as focusing on classroom management, are likely to address other problems more directly (such as “perceptions of school safety” etc.). Incongruously, “school-based crime prevention programs” rarely focus on school grades or retention, which evidence suggests are consistent correlates of violence. We would return to the refrain by Loeber and Farrington and Stouthamer-Loeber and others that developmental prevention is likely to work best when it focuses on established risk factors.
Unfortunately, the many programs that focus on academic achievement are usually implemented for reasons apart from behavioral problems and are not evaluated for their influence on violent behavior. A few programs have been evaluated with positive results on other indicators of antisociality. Bloomquist and Schnell (2005) provided an overview of the enormous volume of research on academic and skill building interventions and their ability to help children with aggression and conduct problems. They organize the literature into three types: self-management training, instructional strategies, and home-school academic interventions. They conclude that many of these have benefits (and they list “best practices” for each). Other reviewers have concluded that findings are mixed. Wasserman and Miller (1998) point out that most programs using the academic approach have been offered to children who may be too old to benefit, having already begun to manifest significant academic and behavioral problems. Brunner (1993) makes strong recommendations about using literacy programs to reduce the chances of recidivism among antisocial adolescents, and although we do not find that the literature points to one academic area more than another in terms of its effects, poor reading ability is associated with violence, so literacy is a good place to start.
It must be highlighted that early academic interventions (preschool) have had substantial effects. Farrington and Welsh (2007) evaluate studies of preschool intellectual enrichment programs and report significant impacts on much later offending. The long-term benefits are clear for intensive interventions and not clear for “less intensive” programs. The most known of the intensive preschool programs is the Perry Preschool. This program was created in response to problems with school performance among poor children and provided one or two years of preschool education to disadvantaged children with the goal of improving cognitive and social outcomes in school. The program had early effects on IQ test scores and longitudinal effects on reading, arithmetic and language achievement, academic attainment, and adult earnings (Tremblay & Japel, 2003). The program cost just over $12,000 per participant, but the savings to government was estimated at twice that (Karoly et al., 1998). The Perry Preschool model has been strongly advocated by many scholars for addressing the risks of low intelligence and school problems (e.g., Farrington, 1990).
Other preschool interventions have also been successful. Reynolds and Chang (1998) evaluated the effectiveness of the Chicago Child-Parent Center program and found that extended academic intervention, from preschool or kindergarten into the primary grades, is needed to reap full crime prevention benefits. Although some earlier reviews of Head Start impact studies were unfavorable (e.g., Haskins, 1989), recent evaluations suggest that Head Start programs have had beneficial impacts on a wide array of outcomes (e.g., Vogel, Brooks-Gunn,
Martin, & Klute, 2013; Webster-Stratton, Reid, & Hammond, 2001). Interestingly, high quality, nonmaternal child care may also have a long-term influence on academic achievement (e.g., Vandell, Belsky, Burchinal, Steinberg, Vandergrift et al., 2010), with implications for the provision of day care for high risk families. The effect has been attributed to enhancing infants’ responsiveness to their environment (Burchinal, Campbell, Bryant, Wasik, & Ramey, 1997).
We were surprised to see that bonds to school were also consistently associated with violent outcomes (as well as nonviolent antisocial outcomes). This may provide an opportunity to intervene with children who struggle academically. We might imagine that inclusive programs in sports, clubs, mentoring, tutoring and the arts could engender bonds to school, even when academic performance is lacking. There is evidence that efforts to enhance school bonds reduce antisocial behavior. Losel and Bender (2003) found that school bonding and academic achievement were protective against antisocial behavior, even among children with below-average intelligence. Authors from the Seattle Social Development project emphasize commitment to school in their model and address it through school interventions (e.g., Hawkins & Herrenkohl, 2003). Some of the practices emphasized by Bloomquist and Schnell (2005) include a positive environment and promoting child interest, which may foster school bonding.
We might expect that a focus on high-risk high school students would be essential, and alternative schooling has had some modest effects on academic achievement (Cox, Davidson, & Bynum, 1995), but beneficial effects on delinquency have not been demonstrated (e.g., Cox et al., 1995). This is likely due to peer effects; adolescents placed in alternative schools will have a greater opportunity to befriend delinquent or troubled fellows, and this may generate offending and offset any gains on the academic side. Of course, evaluating this type of program is fraught with potential biases, since if random assignment is not used to determine which troubled high schoolers go to alternative schools.
An important point to address in educational policy for violence prevention is the challenge of dealing with juvenile offenders. School systems may be reluctant to work with juvenile offenders (Howell et al., 2012) and prefer to let the juvenile justice system handle it. Juvenile offenders often suffer from learning disabilities or academic problems. Maintaining continuity in schooling seems like an obvious minimal requirement, with the real goal of enhancing academic performance. Most incarcerated youths do not graduate from high school or earn a GED; those who do achieve these markers of academic success are less likely to end up back in prison. A meta-analysis by Wilson et al. found that participation in education programs during incarceration had a significant effect on reducing recidivism while employment training programs did not (Wilson, Gallagher, & Mackenzie, 2000).
In the same vein, our review suggests that academic attainment is a differential predictor of violence and high school diplomas have been associated with reduced recidivism in general (e.g., Ambrose & Lester, 1988). Evaluations of adult correctional education programs also suggest they are promising. Findings from a metaanalysis of studies of post-secondary correctional education (PSCE) programs suggest that they reduce recidivism (Chappell, 2004). Correctional education programs have also resulted in increased employment, and those findings just cited by Wilson et al. (2000) suggest that adult basic education, GED and postsecondary education may be more effective in reducing recidivism than work programs, though the authors caution that the studies included in their meta-analysis did not have the strongest empirical methods.