Limitations to Prevention Strategies

Evaluations of these prevention strategies have brought to light a variety of problems, including recognizing developmental issues for children and adolescents, defining and measuring outcomes, and relating selection criteria and targeted outcomes to risk factor research, and other practical issues, all of which are discussed below. Careful consideration of these problems will help researchers and programmers design programs that better identify and address targeted risk factors and outcomes.

Recognizing Developmental Issues

Infants and toddlers. Programs for infants and toddlers focus on parent-child interactions through education and provision of emotional and practical support to the primary caregiver. They often target poor families and are delivered in the hospital or the family's home. However, these intervention programs have not been standardized and thus may not be highly reliable or easily disseminated. They do not include behavioral management techniques for parents that can reduce antisocial behavior among children and adolescents. Finally, long-term followup on child behavior has been limited, focusing instead on care of the child and child adjustment, especially in terms of health.

Preschool children. Most programs for preschool children have not included parent management training. Preschool interventions have not targeted peer relations, which are indicators of outcome. Although some of these programs have included long-term followup and have shown positive effects on antisocial behavior, researchers have not been able to attribute these positive outcomes to a specific cause. The complex social and economic changes of the past 20 years make analysis difficult (Yoshikawa, 1995). Current studies should therefore include the standardized interventions described in manuals and regular evaluations so that successful programs can be replicated.

Elementary school children. Programs for elementary school children are usually school based. Unfortunately, families are not often involved in these programs, and children thus tend to drop out. Even if a program offers parent groups, not all parents will take advantage of them. These programs also rarely assess or address ADHD. An additional problem in the design of school-based programs is the effect of "spillover," which occurs when both control and intervention classrooms are located within the same school. Control classrooms may adopt elements of the intervention (spillover), thus affecting the results of the study.

Adolescents. Programs for adolescents often focus on educational approaches but lack family components, especially in universal programs where it is difficult to involve parents of nonreferred adolescents. If the family is not involved, the youth may have difficulty internalizing and generalizing intervention lessons across settings. This may be one reason why conflict resolution programs show such uneven results and sometimes paradoxical effects. Evaluators of conflict resolution programs should supplement their usual psychoeducational approach with information from independent records, such as arrest data (Bry, 1982; Bry and George, 1980) or vandalism records,²² or raters who are blind to the intervention; otherwise, the success of the curriculums will not be measurable (Webster, 1993).

Defining and Measuring Outcomes

Defined targets. Programs should have clearly defined targets. These should be outcomes that are developmentally linked to antisocial behavior or to specific risk or protective factors. Targeting hypothesized risk factors such as moral reasoning may not result in a corresponding change in child behavior.

Effectiveness. When some risk factors, such as family process or early disruptive behavior, are targeted, effect sizes are small, even when the targeted processes are influenced and the changes are statistically significant. The effectiveness of the intervention may be called into question because the targeted process is influenced by a variety of factors that are highly interwoven and cannot be changed by a single-focus intervention. One exception is medication studies of children with ADHD, which show large effect sizes for inattention and disruptiveness or substantial gains 1 year after the treatment (Kazdin et al., 1987a; Kazdin, Siegel, and Bass, 1992).

Specificity. Specificity is another issue. Although antisocial children are likely to have certain deficits, not all of them will show the same pattern of problems. It is important to choose an intervention that directly addresses an identified deficit in a subgroup of children. Otherwise, the intervention or treatment evaluation may give the false impression that the program was only moderately effective, when in fact it was very effective with a subgroup of children. More research needs to be done to determine which interventions are most effective for specific risk factors.

Neglected outcomes. Studies also should address neglected outcomes. Most interventions focus on overt aggressive behavior, with insufficient attention paid to covert antisocial acts. Although physical aggression is easily observable, most later delinquency is covert in nature.

Long-term followup. Most studies have not included long-term followup, even at 1 year. Without it, long-term maintenance of treatment gains cannot be demonstrated.

Gender. Because the development of antisocial behavior may be different for boys and girls, gender must be considered in program design (Wasserman, 1996). Most interventions are designed with boys in mind and have been applied only to boys. When applied to girls, they may be less effective (Farrell and Meyer, 1997; Kellam et al., 1998). As a result, designers of prevention programs should consider new research on female antisocial behavior (Crick, Bigbee, and Howes, 1996; Crick and Grotpeter, 1995; Zoccolillo, 1993).

Relating Selection Criteria and Targeted Outcomes to Risk Factor Research

Categories of risk factors for delinquency and violence have been conceptualized in a variety of ways. Hawkins and Catalano (1993), for example, define categories of risk at the levels of community, family, school, and so on. However, these categories do not readily correspond to the selection criteria or outcomes studied in prevention research. That is, risk factors for serious, violent juvenile offending are not necessarily the outcomes targeted by preventive interventions. The problem lies in distinguishing between those factors used to select children at high risk and those factors that are themselves targets of intervention. For example, community risk factors such as poverty are commonly used as selection criteria for studies, but they are often inaccurate indicators of individual risk and are seldom targeted for intervention. The opposite tends to be true with family risk factors. Families are assumed to have family management problems if they are selected on the basis of other risk factors, such as the child's aggressive behavior or poverty.

School and individual risk factors that are used as selection criteria are often the focus of intervention. However, targeted outcomes based on these criteria may not be linked to offending. For example, social competency, peer relations, self-control, ethnic identity, or student-faculty communication have not been proven to contribute to antisocial behavior, nor have they been proven to be protective factors that will lessen the risk for antisocial behavior. Protective factors are not simply the opposite of risk factors.

Some risk factors have not been widely examined in prevention research. Family history of problem behavior is an excellent selection criterion,²³ and younger siblings of delinquents are an obvious, but underaddressed, target group. As discussed above, ADHD is an important selection criterion and intervention target, but programs targeting delinquency have yet to address the disorder systematically.

Practical Issues

Limitations of single-focus preventions. Antisocial behavior is rarely the result of a single risk factor.²⁴ Youth live in layered and complex environments that contain multiple risk factors at different levels. As a result, successful approaches to prevention must incorporate components directed at more than one type of risk factor.²⁵ Designers of prevention programs must consider:

• Available institutional resources. For example, is the school able to run the program?
  
  
• The family environment. For example, will the family become involved? Does the family have other problems that need to be addressed?
  
  
• The child's chronological age and developmental level. For example, what interventions are appropriate given the child's developmental level? Is the child facing a school transition, such as entry into a new school?

Antisocial behaviors and the impact of risk factors vary with a child's age and development. Transitions between school levels are important intervention points (Coie and Jacobs, 1993), partly because they reconstitute the child's peer group. At such transitions, other risk factors, such as neighborhood influences on the school, come into play.

Recruiting and retaining participants. Two practical problems hamper the design and implementation of research studies and programs that target serious and violent juvenile offending. The first problem is the feasibility of recruiting and retaining participants. Although school-based programs are popular because children provide a "captive audience," it is difficult to involve families in these programs. Outpatient child psychiatry clinics also have a high rate of missed appointments among families with antisocial children. Retaining control-group families is even more difficult.

In a pilot randomized clinical trial in New York City, Miller and Klein (1996) used a variety of methods to maintain parent participation. The program paid families for assessments and for the costs incurred in attending the clinic, provided food, performed initial assessments in families' homes, and sent holiday cards to families. Each family provided the telephone numbers and addresses of two family members or friends so that families involved in the program could be contacted, and family members in both the control and intervention groups received referrals for services. When appropriate, bilingual interveners conducted home visits in Spanish. The study had a low attrition rate of 10 percent and, using a consumer satisfaction questionnaire (Webster-Stratton, 1989), researchers found that satisfaction among parents was high.

Identifying necessary components. The second practical problem is identifying necessary components. As noted earlier, multifaceted interventions that target the development of chronic aggression are the most effective. Because home and peer settings establish and maintain antisocial behavior, interventions that include parent management training and a peer component are more successful. Programs also should enhance academic skills and provide for the treatment of conditions such as ADHD. If interventions take place early in the child's development and later during developmental transitions, such as those between school levels, the child will be better able to generalize one set of skills across all settings.

23 Cadoret, 1991; Patterson, 1984; Rowe and Gulley, 1992; Wasserman et al., 1996.

24 Elliott, Huizinga, and Ageton, 1985; Patterson, Reid, and Dishion, 1992; Simcha-Fagan and Schwartz, 1986.

25 Coie and Jacobs, 1993; Dodge, 1993; Tremblay et al., 1995.

Prevention of Serious and Violent Juvenile Offending	Juvenile Justice Bulletin • April 2000