4

Preventing Juvenile Crime

Efforts to prevent juvenile delinquency have a long history, but generally they have not been subjected to scientific evaluation of their effects. The following sections discuss programs aimed at families, programs aimed at children and adolescents in schools, and community-based programs. This discussion is not intended as a catalog of all programs available around the country. A comprehensive review of the entire body of prevention program research was beyond the resource capacity of the panel. The literature analyzed for this report was culled largely from a number of published reviews (Barnett, 1995; Brewer et al., 1995; Catalano et al., 1998; Gottfredson, 1997; Gottfredson et al., 1998; Hawkins et al., 1995b; Hope, 1995; Karoly et al., 1998; Lipsey and Wilson, 1998; Sherman, 1997a, 1997b; Tolan and Guerra, 1998; Tremblay and Craig, 1995; Wasserman and Miller, 1998; Yoshikawa, 1995). The panel confined its discussions to prevention programs that have been subjected to experimental or quasi-experimental evaluation research.1 In selecting evaluations to include, the panel chose to look at programs that measured effects on risk factors for delinquency as well as on delinquent behavior itself. Few programs for families and young children have follow-ups that are long enough to test for delinquent behavior. Programs delivered to delinquent youth or that invoke juvenile justice sanctions are covered in Chapter 5.

1  

For an excellent discussion of the stages of evaluation research and the criteria for quasi-experimental and experimental designs, see National Research Council (1998, Chapter 3).



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 107
JUVENILE CRIME 4 Preventing Juvenile Crime Efforts to prevent juvenile delinquency have a long history, but generally they have not been subjected to scientific evaluation of their effects. The following sections discuss programs aimed at families, programs aimed at children and adolescents in schools, and community-based programs. This discussion is not intended as a catalog of all programs available around the country. A comprehensive review of the entire body of prevention program research was beyond the resource capacity of the panel. The literature analyzed for this report was culled largely from a number of published reviews (Barnett, 1995; Brewer et al., 1995; Catalano et al., 1998; Gottfredson, 1997; Gottfredson et al., 1998; Hawkins et al., 1995b; Hope, 1995; Karoly et al., 1998; Lipsey and Wilson, 1998; Sherman, 1997a, 1997b; Tolan and Guerra, 1998; Tremblay and Craig, 1995; Wasserman and Miller, 1998; Yoshikawa, 1995). The panel confined its discussions to prevention programs that have been subjected to experimental or quasi-experimental evaluation research.1 In selecting evaluations to include, the panel chose to look at programs that measured effects on risk factors for delinquency as well as on delinquent behavior itself. Few programs for families and young children have follow-ups that are long enough to test for delinquent behavior. Programs delivered to delinquent youth or that invoke juvenile justice sanctions are covered in Chapter 5. 1   For an excellent discussion of the stages of evaluation research and the criteria for quasi-experimental and experimental designs, see National Research Council (1998, Chapter 3).

OCR for page 107
JUVENILE CRIME PREVENTION OF JUVENILE DELINQUENCY THROUGH FAMILY INTERVENTIONS How parents interact with their children and, in particular, their disciplinary styles, can increase or decrease the likelihood of later antisocial and delinquent behavior by their offspring. As Patterson and his colleagues (Patterson, 1976, 1995; Patterson et al., 1984) indicate through their research, parents who nag or use idle threats are likely to generate coercive systems in which children gain control through misbehaving. So important does the family appear to be in terms of the development of youth crime that programs have been designed to help parents cope with their offspring. Kazdin (1997:1351) summarized this line of research by noting that parent management training “has led to marked improvements in child behavior on parent and teacher reports of deviant behavior, direct observation of behavior at home and school, and institutional records (e.g., school truancy, police contacts, arrest rates, institutionalization).” The following section discusses evaluations of programs that were wholly or in part focused on assisting and training parents. These programs are summarized in Table 4-1. Interventions with Parents and Young Children A strong case for interventions with expectant parents can be made because of the nature of human growth. Brain development during the fetal period has lifelong consequences (Carnegie Task Force on Meeting the Needs of Our Youngest Children, 1994) and can be altered by chemical agents (such as alcohol, nicotine, and drugs), by mothers' behavior and health, and by environmental effects on the mother (Coe, 1999; Wakschlag et al., 1997). Parents with a history of social adjustment problems are most likely to maintain risky behaviors during pregnancy. From this perspective, a number of preventive interventions have targeted pregnant adolescents. These experiments can often be considered interventions with disruptive adolescents in an effort to prevent the intergenerational continuity of antisocial behavior. Unfortunately, participants in these intervention studies generally have not been followed long enough to document the program's impact on the development of disruptive behavior for either the mother or the child. The Elmira Home Visitation study (Olds et al., 1997b, 1998) is an exception. Participants in this targeted prevention experiment were pregnant women with no prior live births and were either unmarried, adolescent, or poor. Other pregnant women were included in the study to prevent stigmatization. Three experimental groups were created by ran-

OCR for page 107
JUVENILE CRIME dom allocation. Women in the first group were visited weekly by a nurse for the first month after enrollment in the study, twice a month until birth, weekly for the first six weeks after birth, twice a month until the baby reached 21 months, and monthly until the child reached the end of the second year. Women in the second group received home visits only during pregnancy, while women in the third group had a screening interview after birth and free transport to the health clinic between the child's birth and the end of the second year. Mothers and children have been followed up to the child's 15th birthday. Fewer mothers in the first group were identified as perpetrators of child abuse and neglect. In addition, significant differences between the first group and controls have been observed when the comparison is limited to those women who were unmarried and had low incomes at initial enrollment. The unmarried, low-income mothers in the first group had fewer subsequent births, longer intervals between the birth of the first and second child, fewer substance abuse impairments, fewer self-reported and officially recorded arrests, and were less often on Aid to Families with Dependent Children (AFDC) than similar mothers in the control groups (Olds et al., 1997a). When the children reached 15 years of age, their delinquency involvement was assessed. Results showed that the adolescent children of the unmarried, low-income mothers who were visited by nurses during pregnancy and the first two years after birth, compared with the comparison groups, reported fewer instances of running away, fewer arrests, fewer convictions and violations of probation, fewer lifetime sex partners, fewer cigarettes smoked per day, and fewer days having consumed alcohol in the last six months (Olds et al., 1998). Olds et al. (1998) reported that the program cost was between $2,800 and $3,200 per family per year. They also estimated that the reduction of the number of pregnancies for low-income mothers and the related reduction in welfare costs were such that the costs of the intervention program were recovered four years after the birth of these women 's first child. The Rand Corporation estimated that when the children reached 15 years of age, savings had reached four times the cost of the program (Karoly et al., 1998). The Elmira nurse home-visiting model has been replicated in Memphis. The goal of the Memphis replication was to see the effects of the model when delivered through the existing health department with minimal input from the researchers (Olds, 1998). Low-income women experiencing their first pregnancy were recruited at the Memphis Regional Medical Center and were randomly assigned to one of four conditions: (1) free transportation to prenatal care appointments; (2) free transportation to prenatal care appointments plus developmental screening and referral services for the child at 6, 12, and 24 months; (3) free transportation,

OCR for page 107
JUVENILE CRIME TABLE 4-1 Family Interventions for Prevention of Juvenile Delinquency Intervention Evaluation Type Number of subjects at follow-up Program components Adolescent Transitions Program (Dishion and Andrews, 1995; Dishion et al., 1996) Random assignment to one of three treatments; quasi-experimental control group N (parent only) = 21 N(teen only) = 29 N(parent & teen) = 29 N(self-directed, control) = 26 N(control) = 36 Parent group education; teen group; family consultation; self-study materials Elmira Nurse Home Visitation Program (Olds et al., 1997a, b, 1998) Experimental (Randomized controlled trial) N(exp) = 97 N(control) = 148 Nurse home visiting Memphis Nurse Home Visiting (Kitzman et al., 1997) Experimental (Randomized controlled trial) N(exp) = 223 N(control) = 515 Nurse home visiting High/Scope Perry Preschool (Schweinhart et al., 1993) Experimental (Randomized controlled trial) N(exp) = 58 N(control) = 65 Preschool and home visiting Houston ParentChild Development Center (Johnson and Walker, 1987) Experimental (Randomized controlled trial) N(exp) = 51 N(control) = 88 Home visiting, center-based parent training, and day care

OCR for page 107
JUVENILE CRIME Age at treatment Length of treatment Length of follow-up Delinquency and antisocial behavior outcomes Other outcomes 10 to 14 years 12 weeks 1 year Teen-focus group had more teacher identified behavior problems at follow-up than those in other groups Parent-focus and teen-focus groups had improved family functioning; teen-focus group had significantly more tobacco use at follow-up than parent-focus or controls Prenatal to age 2 2 years 15 years Fewer arrests, convictions, and probation violations among treatment group; fewer incidents of running away from home among treatment group Less cigarette and tobacco use and fewer sexual partners among treatment group; lower rate of child abuse and neglect by treatment group parents Prenatal to age 2 2 years 0 years Fewer injuries and hospitalizations of program children; fewer subsequent pregnancies and less time on welfare for program mothers Age 3-4 1 to 2 years 24 years Significantly fewer arrests by age 27 among treatment group Treatment group had significantly higher high school graduation rates, significantly higher incomes, were significantly more likely to be home owners, and were less likely to receive social services Ages 1 to 3 years 2 years 5 to 8 years Program children had fewer acting out, aggressive behavior problems

OCR for page 107
JUVENILE CRIME Intervention Evaluation Type Number of subjects at follow-up Program components Mailman Center for Child Development (Field et al., 1982) Experimental (Randomized controlled trial) N(exp) = 64 N(control) = 30 Home visiting and day care Mailman Center for Child Development (Stone et al., 1988) Experimental (Randomized controlled trial) N(exp) = 31 N(control) = 30 Home visiting and day care Montreal Longitudinal Experimental Study (Tremblay et al., 1995) Experimental (Randomized controlled trial) N(exp) = 46 N(attention-control) = 84 N(no-contact control) = 42 Parent training and school-based social skills training Newcastleupon-Tyne (Kolvin et al., 1981) Experimental Group therapy, parent counseling and teacher consultation, and behavior modification (older children) or nurturing work (younger children) PARTNERS program with Head Start families (Webster-Stratton, 1998) Experimental (Random assignment of Head Start centers to treatment or control conditions) N(exp) = 189 N(control) = 107 Parent and teacher training

OCR for page 107
JUVENILE CRIME Age at treatment Length of treatment Length of follow-up Delinquency and antisocial behavior outcomes Other outcomes Birth to age 1 1 year 1 year Better growth and development among treatment group; treatment mothers had greater rate of return to work or school and fewer repeat pregnancies Birth to age 1 1 year 5 to 8 years No differences on academic, behavioral, or social-emotional skills Age 7 2 years 10 years Significantly less delinquent behavior among intervention group 7- to 8-year-olds, 11- to 12year-olds Parent counseling- up to 10 visits; group therapy-10 sessions; behavior modification-2 school terms; nurturing work-5 school terms 3 years after baseline assessments Less antisocial behavior among play group therapy treatment group; no difference between treatment and control groups for other treatment conditions Age 4 8 to 9 weeks 12 to 18 months Intervention children significantly decreased their misbehavior compared to no change in control group Intervention children significantly increased their positive affect compared to no change in control group.

OCR for page 107
JUVENILE CRIME Intervention Evaluation Type Number of subjects at follow-up Program components Seattle Social Development Project (Hawkins et al., 1999) Non-controlled Randomized trial N = 598 Teacher training, parent education, social competence training for children Syracuse University Family Development Research Program (Lally et al., 1988) Quasi-experimental (Matched Pair control group) N(program) = 65 N(control) = 54 Home visiting and day care developmental screening, and intensive nurse home visitation during pregnancy with one postpartum visit in the hospital and one postpartum visit at home; and (4) free transportation, developmental screening, and nurse home visitation during pregnancy continuing to child's second birthday. When the children were age 2, the group that received the full home visitation program (the experimental group) was compared with the group that had free transportation and developmental screening and referral services (the control). Mothers in the experimental group had fewer subsequent pregnancies and less time on welfare than mothers in the control group. Children in the experimental group had fewer injuries and ingestions and fewer hospitalizations for injuries and ingestions than the control children (Kitzman et al., 1997). Continuing follow-ups are planned. The Syracuse University Family Development Research Program (Honig, 1977; Honig and Lally, 1982; Lally and Honig, 1977) began providing services to low-income mothers during their last trimester of pregnancy and continuing for the first five years of the childrens' lives. The program consisted of weekly home visits by paraprofessionals to assist the family with issues of childrearing, family relations, employment, and community functioning. The program also provided day care services for

OCR for page 107
JUVENILE CRIME Age at treatment Length of treatment Length of follow-up Delinquency and antisocial behavior outcomes Other outcomes 1st to 6th grade 6 years 6 year (age 18) Significantly less violent delinquent behavior among experimental group Less heavy drinking, sexual intercourse, multiple sex partners, and pregnancy among experimental group; higher academic achievement among experimental group Prenatal to age 5 5 years 10 years Treatment group less likely to have juvenile justice system involvement the children for 50 weeks a year. Children from 6 to 15 months of age were provided with half-day care for 5 days a week. Full-day care was provided 5 days a week for children from 15 to 60 months of age. When the program children were 36 months old, a matched-pair control group was established. The control group children were matched in pairs with program children with respect to age, ethnicity, birth ordinality, sex, family income, family marital status, maternal age, and maternal educational status at the time of the child's birth. Lally et al. (1988) conducted a follow-up study of both control and program groups 10 years after program completion. Data on delinquency were collected from probation and court records. Children in the program group were less likely to have been involved in the juvenile justice system than were the control group children. Only 6 percent of program children, compared with 22 percent of the control children, had been processed as probation cases for delinquent behavior. Furthermore, the program children had committed less severe offenses than the control children. Although beginning parent training prenatally may be preferable to beginning postnatally, one would expect interventions with parents of infants to have a significant impact on their parenting skills, and thus on the socialization of their children.

OCR for page 107
JUVENILE CRIME The Mailman Center program (Field et al., 1982) study randomly allocated low-income, adolescent mothers to a program that offered home visitation over the first year of life, to a second program that added work as teacher's aides in their infants' nursery program, or to a control group that received no specific treatment. At age 2, children from the second intervention group had higher developmental scores, and their mothers had a higher rate of return to work or school and fewer pregnancies compared with the first intervention group and to the no-intervention control group (Field et al., 1982). However, at a later follow-up, when the children were between 5 and 8 years of age, no significant differences were observed between children in both intervention groups and those in the control group on academic, behavioral, and socioemotional assessments (Stone et al., 1988). Although the investigators assessed only half the families, no significant differences were found between the original sample and those followed up. The authors concluded that the low socioeconomic status of the mothers may have overridden the early positive effects of the interventions. Results of the Elmira and Syracuse studies that also targeted low-income adolescent mothers suggest that the lack of long-term effects of the Mailman Center program could be due to failure to include a prenatal component and to the short duration of the intervention. Programs for parents of infants seem to save money in the long run. Most of the reported savings in the Elmira program was due to increased employment and reduced welfare dependence among the mothers in the program. Karoly and colleagues (1998) noted that even more savings may be realized when information is available about employment of the children in the program. Furthermore, their study did not attempt to assign monetary value to other benefits of the program, such as increased IQ or less child abuse. Savings were not evident for the low-risk families who received services in the Elmira program. The authors point out that, at least from the perspective of government savings, this finding “underscores the need for matching the program to the population that needs its services” (Karoly et al., 1998:90-91). More recently, Webster-Stratton (1998) administered a parent training program that targeted risk factors for disruptive behavior in Head Start centers. Nine Head Start centers (64 classes) were randomly assigned to experimental (345 children) and control conditions (167 children). The 8 to 9 week program focused on teaching effective parenting skills, positive discipline strategies, and ways to strengthen children's social skills and prosocial behaviors to parents of the 4-year-olds attending the Head Start centers. Groups of parents (8-16) met weekly for two hours with a trained family service worker and a professional to view videotapes of modeled parenting skills and discuss parent-child interaction. Posttest

OCR for page 107
JUVENILE CRIME and one year follow-up assessments of parental competencies (mother reports and home observations) showed significant differences between the experimental and control condition parents. Teachers reported more parental involvement in the children's education and fewer behavior problems among the children whose parents had received the training. Experiments have tested the impact of quality day care centers on the development of high-risk children. The impact of day care without any other form of intervention, however, is not known because experiments generally include other forms of intervention, such as parent training and medical services. To the extent that cognitive development, emotional regulation, and peer interaction underlie the development of behavior problems, one would expect that quality day care programs would be an essential component of preventive efforts with at-risk infants and toddlers. The High/Scope Perry Preschool program (Schweinhart et al., 1986; Weikart and others, 1970), which targeted poor minority families, was successful in preventing crime through the age of 27, when those in the preschool program and in the comparison group were last traced (Schweinhart et al., 1993). The intervention included a high-quality preschool program for 3- and 4-year-olds and home visits by preschool teachers, during which the mothers were taught how to help their children with their preschool activities. Families were randomly assigned to the preschool or to a control group. At age 27, program participants were significantly less likely to have been arrested than were controls. Program participants also showed other positive outcomes. They were significantly more likely to have completed high school, earned significantly more money per year, and were significantly more likely to be home owners than members of the control group. The Houston Parent-Child Development Center Project (Johnson and Walker, 1987) randomly assigned low-income Mexican-American families with healthy 1-year-olds to a treatment or a control group. The treatment group received home visits by a paraprofessional for the first year of their involvement in the program. During the second year of program involvement, mother and child attended a center-based program four mornings per week. Mothers received classes in child management, child cognitive development, family communication skills, and other family life topics while children spent time in a nursery school. Teacher assessments of externalizing problems 5 to 8 years after the end of the program, when children were ages 8 to 11, showed a substantial positive impact. Children in the program were less likely than those in the control group to exhibit acting-out, aggressive problem behaviors. On the whole, there is good evidence for a positive impact of quality day care in preventing behavior problems for high-risk children, thus contributing to the long-term prevention of delinquency. The prevention

OCR for page 107
JUVENILE CRIME Age at treatment Length of treatment Length of follow-up Delinquency and antisocial behavior outcomes Other outcomes 5- to 15-yearolds 32 months 16 months Fewer police charges against experimental juveniles during treatment and at follow-up 6th, 7th, and 8th grades 3 years 0 Less onset and prevalence of alcohol use in treatment school districts   Not reported 1 year Patterns of significant changes apparent by neighborhood: Increased awareness of and participation in crime prevention meetings, reduced proportion of residents known by name, increased victimization levels, increased youth disorder, increased fear of crime and perceptions of local crime rates, decreased optimism about changes in the neighborhood   al. (2000) report on the status of Moving to Opportunity families in Boston. One to three and a half years after random assignment, 48 percent of the experimental group and 62 percent of the Section 8 comparison group had moved. Families in both the experimental group and comparison group were more likely to be residing in neighborhoods with low poverty rates and high education levels than were families in the control group.

OCR for page 107
JUVENILE CRIME Experimental group families who moved were more likely to be living in suburban areas than were comparison families who moved, but a higher percentage of experimental families did not move and remained in high poverty areas. Experimental and comparison group families who moved experienced improvements in multiple measures of well-being compared to the control group, including increased safety, improved health, and fewer behavior problems among boys. Katz et al. (2000:34) concluded that the short-term effects of providing vouchers to families who want to move out of high-poverty neighborhoods improves their well-being. In addition, “the MTO-Boston results strongly imply that neighborhoods have large impacts on the health and behaviors of children and adults from low-income families.” Policing Approaches and Community Mobilization Strategies Community-oriented policing has been embraced by national and local government throughout the United States. The enthusiasm for this approach, however, has been such that the term now encompasses a variety of meanings. In the mid-1980s, community policing seemed to mean getting police officers out of patrol cars and back in touch with ordinary citizens on a day-to-day to basis, rather than just when crises arose. The term problem-solving policing has also become popular, pointing to a process in which police, other public agencies, and local citizens in high-crime areas try to resolve recurrent problems in particular hot spots. Other recent trends in policing include zero-tolerance policies based on the broken windows theory of Wilson and Kelling (1982). The idea is that minor crimes that disturb the quality of life lead to more serious crime; therefore, cracking down on minor crimes can lead to decreases in all crime. All of these policing strategies involve working with the affected communities. Scientifically credible evaluations have not demonstrated the usefulness of these approaches. Efforts that may or may not involve working with police rely on organizing grassroots efforts to address community problems, including crime. Community organization, which was particularly popular during the 1980s, seemed more successful in affluent than in impoverished neighborhoods. Hope (1995) reports that in high-crime neighborhoods, people are wary of their neighbors and are therefore less likely to join community anti-crime efforts than are those in neighborhoods with less crime. The few credible evaluations of such activities as block watches have failed to show reductions in crime rates (Hope, 1995). Using a quasi-experimental design, Rosenbaum et al. (1986) tested a block watch program in middle-and lower-class neighborhoods in Chicago. Five experimental neighborhoods were selected. These neighborhoods had well-established volun-

OCR for page 107
JUVENILE CRIME teer community organizations, were interested in a block watch program, and had the resources to carry out the program. One set of comparison neighborhoods was selected on the basis of having well-established volunteer community organizations. Another set was selected randomly from among neighborhoods with demographic characteristics similar to the experimental neighborhoods. After a year of the intervention, no changes in victimization or in residents' crime prevention activities were found in the experimental areas. Compared with control neighborhoods, those in the experimental areas showed significant increases in fear of crime and decreases in attachment to their neighborhood. Another popular strategy is the imposition of curfews for juveniles. Curfew laws tend to be enforced primarily in high-crime areas in cities. Evaluation of curfew laws has found no consistent effect on crime reduction (Sherman, 1997a). In a statistical analysis of the effect of curfew enforcement in California, Macallair and Males (1998) found no correlation between strict curfew enforcement and juvenile arrests for crimes, with a few exceptions in areas where curfew enforcement was associated with an increase in juvenile arrests for offenses other than curfew violations. Community Economic Development and Situational Prevention Community economic development has grown enormously as a field, based primarily on the production and management of affordable low-income housing. As community groups have become substantial landlords and developers, they have been confronted with issues of public safety and have developed a number of ways of trying to make their housing developments and neighborhoods safer. These include hiring their own private security, working directly with local police, organizing their own tenants and other community residents, and redevelopment of blighted buildings and public areas (Keyes, 1992; Sullivan, 1998). Although not evaluated specifically for their effect on juvenile crime, some of these measures have been found to reduce crime, particularly property crimes— offenses in which juveniles are likely to be involved. Unfortunately, these evaluations seldom use experimental or even quasi-experimental designs. Closing of streets in high-crime areas has been another tactic aimed particularly at reducing prostitution and drive-by shootings. The one study with particular relevance to juvenile offenses took place in Los Angeles. In 1990, the Los Angeles Police Department installed 14 traffic barriers in a neighborhood with a high level of drug activity, shootings, and homicides, much of the crime gang-related. The traffic barriers were designed to make drive-up drug sales and drive-by shootings more difficult. The barriers were in place for a two-year period. Lasley (1996) compared violent crimes in the area for a year prior to the installation of

OCR for page 107
JUVENILE CRIME barriers and for the four years after their removal. In addition, crime rates in four adjacent areas were compared with the crime rate in the experimental area. Homicides declined 65 percent while the barriers were in place, then rose 800 percent in the follow-up period. Other violent crimes also fell during the period the barriers were in place, then rose when they were removed. The adjacent areas had steady crime rates during these same periods. Although the findings suggest that the barriers were helpful in reducing crime, the design of the study makes it impossible to rule out alternative explanations for the differences. Community Interventions Comprehensive community initiatives for children and families are a relatively new form of providing social services that emphasize not just individual programs, but also the coordination of different programs in particular local areas and the involvement of local residents in the process. The goals of coordinated community interventions are those of healthy development of children and adolescents, including prevention of delinquency and substance abuses (Catalano et al., 1998; Connell et al., 1995). As noted above, methods for evaluating comprehensive community interventions are still being devised. Although there have been no evaluations of the effects on juvenile crime of comprehensive community interventions, some more narrow community-based programs aimed at reducing drug, alcohol, and tobacco use among youngsters have found effects on substance use. Pentz et al. (1989a, 1989b, 1989c) reported on the Midwestern Prevention Project, which included mass media, school-based educational curricula, parent education and organization, community organization, and health policy. About one-third of the 6th and 7th grade students from all of the middle schools and junior high schools in Kansas City Standard Metropolitan Sampling Area (as designated in the U.S. census in 1980) were included in the baseline sample. All 6th and 7th grade classes in 16 schools and a random sample of classrooms in the remaining 34 schools were included. By the time the intervention was to be put in place, 6 of the 50 schools had closed and 2 could not schedule the baseline data collection. Of the remaining 42 schools, 8 were randomly assigned to either program or control conditions, 20 schools had the flexibility to institute the intervention programming and were assigned to the program condition, and 14 did not have such flexibility and were assigned to the control condition. Analyses were done using the school as the unit of analysis. In addition, 30 percent of the individual students were tracked, and analyses for them were also completed. Pentz and colleagues found significantly lower cigarette and drug use in intervention schools than in comparison schools

OCR for page 107
JUVENILE CRIME and among intervention students compared with control students. Since all students in the area were potentially exposed to the mass media, the researchers concluded that the other components of the program were more effective than the mass media alone (Johnson et al., 1990). A replication of the program using an experimental design in Indianapolis also found reduced cigarette and alcohol use among students who had already initiated use at baseline in the experimental schools (Chou et al., 1998). Perry et al. (1994) found significantly reduced tobacco use among students in communities in which a seven-year Minnesota Heart Health Program was implemented. The program combined a school-based component with community-wide activities, mass media, and adult education. The study employed a quasi-experimental design, in which a comparison community was matched to the experimental community on population size, socioeconomic makeup, and distance to a large metropolitan area. A similar program, Project Northland, for reduction of alcohol use among adolescents was also tested by Perry et al. (1996). The school districts in six primarily rural counties were blocked by size and randomly assigned to an intervention or control condition. The intervention included parent involvement and education, behavioral curricula for 6th, 7th, and 8th graders, peer leadership training, and community task forces. After three years, students in experimental areas were significantly less likely to use alcohol, tobacco, and marijuana than students in control areas. The researchers noted that the experimental and control communities had some differences at baseline, with more white students in the control communities and greater alcohol use in the experimental communities. Mentoring Evaluations of mentoring programs have consistently found that non-contingent, supportive mentoring relationships do not have the desired effects on academic achievement, school attendance, dropout, child behavior, or employment (Dicken et al., 1977; Goodman, 1972; McPartland and Nettles, 1991; Poorjak and Bockelman, 1973; Rowland, 1992; Slicker and Palmer, 1993). When mentors used behavior management techniques, Fo and O'Donnell (1974) found improvement in school attendance. Truancy was reduced when the mentor relationship included reinforcement contingent on appropriate behavior, but not when mentoring did not include contingent reinforcement. This evaluation suffered from a small sample size and short program length. Given the promotion of mentoring in recent years, the techniques used by mentors need more attention and evaluation, especially because there are some indications that some mentoring programs may increase delinquency (O'Donnell et al., 1979).

OCR for page 107
JUVENILE CRIME In the Cambridge-Somerville Youth Study, negative effects of treatment were evident. One boy from each matched pairs of boys, all of whom were under the age of 12 and living in poor neighborhoods, was randomly assigned to a treatment group. Those in the treatment group were assigned a counselor who visited them in their homes, at school, and on the streets approximately twice a month for more than five yeas. Boys in the treatment group also received a variety of services, including tutoring, medical and psychiatric care, and recreational activities. McCord (1978, 1981, 1992) found that those who had been in treatment were more likely to have a number of adverse outcomes by middle age, including more convictions for serious crimes. A more recent evaluation of mentoring reported positive effects on delaying the initiation of drug and alcohol use, on reducing the use of hitting, on improving academic performance, and improving relationships with family. Tierney et al. (1995) randomly assigned youngsters ages 10 to 16 accepted into the Big Brothers/Big Sisters program to be assigned a big brother or big sister or to an 18-month waiting list. Both program and control groups were assessed at time of random assignment and 18 months later through surveys administered to the youth, through surveys administered to the parent or guardian of the youth, and through data collected by the case managers. At 18 months, youngsters in the program group were 46 percent less likely than controls to have initiated drug use and 27 percent less likely to have initiated alcohol use. Program youth were 32 percent less likely than controls to have hit someone in the previous year. Program youth also had better school attendance and grades and better relationships with parent and peers than controls. Tierney et al. (1995) stressed the importance of the Big Brothers/Big Sisters infrastructure to the pairing of mentors and young people and the development and maintenance of their relationship. Volunteer mentors are carefully screened before being accepted. Training on program requirements, youth development, communication and limit-setting skills, and relationship-building skills are provided for the mentors. Youth and mentors are matched by gender, race (when possible), and interests. Case managers establish regular contact with mentors, parents, and youth. This infrastructure facilitates regular and sustained contact between mentor and young person. A earlier study of Big Brothers/Big Sisters found that 96 percent of mentor-youth matches had at least one meeting in the previous four weeks during the first year of their relationship and averaged 3.1 meetings per 4-week period (Furano et al., 1993). By comparison, studies of less structured mentoring programs found much lower rates of meeting—57 percent meeting on a somewhat regular basis among six campus-based mentoring programs (Tierney and Branch, 1992) and 40 percent of

OCR for page 107
JUVENILE CRIME scheduled meetings occurring between mentors and youth in a juvenile justice setting (Mecartney et al., 1994). Mentors are not the only ones carefully selected for Big Brothers/Big Sisters. In order to participate, a youngster must also be found eligible for the program. The screening process includes a written application, interviews with both parent and youngster, and a home assessment. In addition, both youngster and parent must agree to follow agency rules. Participating youngsters have a high level of motivation to enter the program. Although this evaluation accounts for that level of motivation by its random assignment of qualified youngsters to either program or waiting list conditions, it is possible that a mentoring program such as Big Brothers/ Big Sisters may be less successful for young people who are not as motivated. After-School and Nonschool-Hours Programs Increasing attention in recent years has been given to providing children and adolescents with supervised activities during nonschool hours. Although relatively few evaluations of such programs have focused on their effects on delinquency, it seems plausible that providing prosocial activities during the nonschool hours could deter delinquent behavior. Violent offenses by juveniles peak in the after-school hours (Snyder and Sickmund, 1999). Furthermore, parental monitoring and supervision, the lack of which is associated with delinquent behavior, may be missing during after-school hours. By 1997, 78 percent of mothers with children between the ages of 6 and 13 were in the labor market (Bureau of Labor Statistics, 1998), making many of them unavailable for child supervision after school. Several studies have found that children's and adolescents' involvement in unsupervised self-care in the company of peers was associated with behavior problems, particularly for juveniles in high-risk environments (Pettit et al., 1999; Vandell and Posner, 1999). One quasi-experimental study in Ottawa (Jones and Offord, 1989) showed that an after-school and weekend program in which special efforts were made to attract nonparticipants in one housing unit resulted in reduced security arrests, whereas the matched comparison housing unit with some recreational programs that made no attempts to attract non-participants had increasing security problems during the 32 months of program activity. Adult arrests were comparable in both areas. And 16 months after the program ended, the positive changes in the experimental area had diminished significantly. The program actively recruited all children in the housing complex to participate in structured courses for improving skills in sports, music, dance, scouting, and other nonsports activities. The program also encouraged participation in ongoing leagues

OCR for page 107
JUVENILE CRIME and other competitive activities in community. Despite cost-effectiveness in terms of reduced vandalism and other minor crimes, the program was not adopted into an ongoing funding stream, a problem that has typified even successful programs. As discussed in the section on peer group interventions above, grouping high-risk young adolescents, even with adult supervision, has the potential of exacerbating antisocial behavior (Dishion et al., 1999). Vandell and Posner (1999) studied the effects of participating in a formal after-school program among low-income 3rd to 5th graders. Although participating in the program was associated with positive behavioral outcomes for 3rd graders, for 5th graders program participation was associated with increased school and behavioral problems. Upon further analysis, it appeared that the 5th grade participants had been more troubled prior to the program than were the 3rd graders. The increasing salience of peers as children age, coupled with the association with a more troubled group of peers, may have been a factor in the negative outcomes for 5th graders. Feldman and colleagues (Feldman, 1992; Feldman et al., 1983) examined the group makeup of programs on their outcomes for antisocial boys. They placed antisocial boys who had been referred by juvenile courts, special schools, mental health facilities, and residential treatment centers into programs at a suburban community center. Boys who regularly participated in programs at the community center (nonreferred) were also included in the study. Boys were randomly assigned to a group composed of all referred boys, of all nonreferred boys, or in mixed groups composed of one or two referred boys in a group of nonreferred boys. Groups met once per week during the school year and participated in a wide range of recreational and leisure activities. Referred boys in the mixed groups significantly reduced their antisocial behaviors and increased their prosocial behaviors over the course of the year, but referred boys in the nonmixed group did not show decreases in antisocial behavior. Furthermore, the nonreferred boys in the mixed group were not negatively affected by their association with the referred boys; there was no significant difference between them and the nonreferred boys in the nonmixed group. This study showed the potential for improving the behaviors among at-risk boys by grouping a small number of them with a group of predominantly prosocial boys. Job Training Job training programs have long been a part of efforts to improve the conditions of the poor. Since lack of opportunities to earn money in legitimate jobs is a risk factor for participation in crime, improving the chances for legitimate employment may reduce crime rates among young

OCR for page 107
JUVENILE CRIME people. A recent evaluation of Job Corps (Schochet et al., 2000) found positive short-term effects of participation among 16- and 17-year-olds, including reduced rates of arrest, incarceration, and conviction. Job Corps is a comprehensive, residential program3 that provides academic and vocational training, health care and education, counseling, and job placement assistance. The program is available to economically disadvantaged youth, ages 16 to 24, who are high school dropouts or in need of additional education or training; are not on probation or parole; are free of serious medical or behavioral problems; and who come from disruptive environments. Participants may be enrolled in Job Corps for up to two years. A random sample of all first-time Job Corps eligible applicants between November 1994 and December 1995 were randomly assigned to the research treatment group or a control group. The control group members were not allowed to enroll in Job Corps for three years, but could participate in any other available programs. Both groups were interviewed 12 months and 30 months after random assignment. At the 30 month follow-up, the treatment group had increased their education and earnings more than the control group, and had significantly reduced their involvement in the criminal justice system. Results for 16- and 17-year-olds were particularly encouraging: their arrest rates dropped by 14 percent and incarceration rates by 26 percent (Schochet et al., 2000). A four year follow-up is planned to investigate longer term impacts. CONCLUSIONS Our review of attempts at intervention has turned up very few successful programs that have credible evaluations. Work with pregnant women, with preschool children and their mothers, with teachers and their parents in high-risk neighborhoods have been shown, using scientifically appropriate evaluations, to have clear benefits in terms of reducing delinquency. But several widely used intervention strategies have been found to increase delinquency. Many such programs rest on drawing young misbehaving adolescents together. The panel concluded that innovative approaches to delivering interventions, which avoid the danger of grouping misbehaving adolescents, should be encouraged. The most effective crime prevention programs, the panel concluded, addressed a range of difficulties. Thus, rather than targeting crime alone, successful preschool approaches helped young mothers teach their chil- 3   About 90 percent of the participants live in supervised dormitory housing; the remainder participate on a nonresidential basis.

OCR for page 107
JUVENILE CRIME dren, improved the health of pregnant women and their infants, and encouraged education in the classrooms. Based on its review of the literature on preventing delinquency, the panel drew the following conclusions: Programs that facilitate healthy births, infancy, and childhood appear to be effective crime prevention interventions. Programs that include multiple components for parents, youngsters, and environment (school or community) and that target multiple behaviors appear more beneficial than narrowly focused programs. Public concern about juvenile crime should be used to encourage adequate, scientifically credible evaluations of the programs instituted to address that concern. RECOMMENDATIONS A variety of interventions with infants, preschoolers, and elementary school children have been found to successfully reduce risk factors for delinquency. It is frequently assumed that reducing known risk factors associated with delinquency will result in reduced delinquency. Few of the studies, however, have long enough follow-up periods to assess whether criminal delinquency is actually reduced. Although reducing various risk factors may be a positive outcome in and of itself, it is important to know which preventive interventions have long-term effects on delinquency and crime. Mounting follow-up studies of participants from programs that were shown to be effective in reducing risk factors for delinquency is also a relatively inexpensive research strategy. Recommendation: Federal agencies concerned with the development of youth, in particular the Office of Juvenile Justice and Delinquency Prevention, the National Institute of Justice, the National Institute for Mental Health, and the National Institute on Child Health and Human Development, should encourage and fund studies of long-term outcomes for well-designed interventions that have shown short-term promise for reducing risk factors for delinquency. Although evaluation research has resulted in some information about what types of programs may be effective in preventing delinquency, much remains to be known. Is it best for a program to begin prenatally? Is there an ideal length of program delivery? Are some programs more effective for certain types of children or families or at certain ages? Which programs are counter productive? Some relatively well-evaluated programs, such as D.A.R.E. and shock incarceration programs (see Chapter 5) have

OCR for page 107
JUVENILE CRIME been shown to have little impact on the targeted behavior and even counter productive impacts among some populations. Until aspects of programs are systematically varied and well evaluated, these questions will remain. Recommendation: All publicly supported intervention programs should be evaluated for both safety and efficacy using scientifically credible methods for doing so. Adequate funding for such evaluations should be included in the public support of intervention programs. Funding for programs whose effectiveness is shown to be limited should be discontinued. Placing one or two antisocial juveniles in a group of primarily prosocial young people can decrease their antisocial behavior and increase their prosocial behavior without negatively influencing the prosocial youngsters. Some well-designed evaluations of treatments for at-risk juveniles found, however, that placing such youngsters in groups, even under careful adult supervision, had the undesired outcome of increasing their antisocial behavior. Recommendation: Federal and state funds should be used to develop treatments for misbehaving youngsters that do not aggregate aggressive or antisocial youth.