INTRODUCTION There are several important reasons why society should be concerned with youth involved in gang activity. Gang members commit a disproportionate number of offenses and commit serious and violent offenses at a rate several times higher than non-gang-members (Howell, 1998). For instance, in Denver, Colo., gang members commit approximately three times as many serious and violent offenses as nongang youths do (Esbensen and Huizinga, 1993). In Rochester, N.Y., gang members commit more offenses at an even higher rate (seven times as many as nongang adolescents). Moreover, gangs members are far more likely than their nongang peers to commit certain types of crime-such as assault, carrying concealed weapons in school, auto and other theft, intimidating or assaulting victims and witnesses, participating in driveby shootings and homicides, and using, selling, and stealing drugs-even when the two groups have grown up under similar circumstances (Huff, 1998). Finally, studies indicate that the influence of gang membership on the level of youth violence is greater than the influence of other highly delinquent peers (Thornberry, 1998; Huizinga, 1997). Thus the nature of gangs and their involvement in serious crime and violence produces many risk factors for youth. Likewise, the greater the number of risk factors to which youths are exposed, the greater their risk of joining a gang. Longitudinal studies of adolescents in multiple sites (Seattle, Wash., and Rochester) have identified the causal risk factors for gang membership within each domain. The Rochester site found "youth who grow up in more disorganized neighborhoods; who come from impoverished, distressed families; who do poorly in school and have low attachment to school and teachers; who associate with delinquent peers; and engage in various forms of problem behaviors are at increased risk for becoming gang members" (Thornberry, 1998, 157). The researchers in Seattle found similar risk factors for gang involvement (Hill et al., 1996). While it appeared that the major gang "epidemic" of the 1960's had subsided in most major cities, increased gang activity (and attention to it) returned in the 1990's (Decker, 2002). Gang-related crime continues to be a pressing national issue in the United States. The 2000 National Youth Gang Survey, which surveys all cities with populations of 250,000 or more, reported that an estimated 3,330 jurisdictions in the United States experienced gang activity in 2000, compared with an estimated 3,911 jurisdictions in 1999 and 4,464 jurisdictions in 1998. The same study estimated that the United States had more than 24,500 gangs and 772,500 active gang members in 2000 compared with an estimated 26,000 gangs and 840,500 active gang members in 1999 and 28,700 gangs and 780,000 gang members in 1998 (Egley and Arjunan, 2002). There is no adequate definition of the term "gang." This difficulty is also compounded by the fact that many jurisdictions deny the existence of gangs or, conversely, characterize less serious forms of adolescent law-violating groups as gangs (Miller, 1992). Nevertheless, several qualified researchers have attempted to define "gang" in very different ways. The following definitions demonstrate the lack of consensus on the issue: A gang is an organized social system that is both quasiprivate and quasisecretive and whose size and goals have necessitated that social interaction be governed by a leadership structure that has defined roles; where the authority associated with these roles has been legitimized to the extent that social codes are operational to regulate the behavior of both leadership and rank and file; that plans and provides not only for the social and economic services of its members, but also for its own maintenance as an organization; that pursues such goals irrespective of whether the action is legal or not; and lacks a bureaucracy. [Jankowski, 1991] A gang has the following characteristics: a denotable group comprised primarily of males who are committed to delinquent (including criminal) behavior or values and call forth a consistent negative response from the community such that the community comes to see them as qualitatively different from other groups. [Klein, 1995] A gang is a group of individuals who have symbols of membership, permanence, and criminal involvement. A gang member is a person who acknowledges membership in the gang and is regarded as a gang member by other members. [Decker and Curry, 1999] A gang is a well-defined group of youths between 10 and 22 years old. [Huff, 1998] Despite the fact that there is little consensus on what a gang is, there are certain characteristics that typically define gangs. Research tends to use the following characteristics when identifying gangs: 1) formal organizational structure (not a syndicate), 2) identifiable leadership, 3) identified territory, 4) recurrent interaction, and 5) engaging in serious or violent behavior (Howell, 1994). Demographic Characteristics The average age of gang members is 17 to 18 years old (Curry and Decker, 1998) but tends to be older in cities where gangs have existed longer. The typical range is 12 to 24. The median age at which youths start hanging around gangs is roughly 13 years old. The median age for joining a gang is 14 (Huff, 1998). Although female gang membership is increasing (Klein, 1995), male gang members outnumber females by a wide margin (Miller, 1992). According to a national law enforcement survey, the ethnicity of gang members is 48 percent African American, 43 percent Hispanic, 5 percent white and 5 percent Asian (Curry, 1996). Certain offenses are related to different racial/ethnic gangs. For instance, African-American gangs are relatively more involved in drug offenses, Hispanic gangs engage in turf-related violence, and Asian and white gangs display a tendency toward property crimes (Spergel, 1990). It is important to point out, however, the disproportionate representation of minority groups in gangs is not a result of a predisposition toward gang membership, but rather that minorities tend to be overrepresented in areas overwhelmed with gang activity (Bursik and Grasmick, 1993). There are other ways to classify gangs other than by ethnicity. One way is viewing gangs along a continuum by degree of organization, "from youth groups who hang out together in shopping malls and other places; to criminal groups, small clusters of friends who band together to commit crimes such as fencing operations; to street gangs composed of groups of adolescents and young adults who form a semistructured operation and engage in delinquent and criminal behavior; to adult criminal organizations that engage in criminal activity primarily for economic reasons" (Howell, 1998). THEORETICAL CONTEXT One of the most important questions with regard to gang prevention programs is why youths choose to join gangs. Decker and Van Winkle (1996) argue that the act of choosing to join a gang can be described as a number of "pulls and pushes" on an individual. The "pull" refers to the attraction of gang membership to certain youths. On the one hand, gang membership can enhance prestige or status among friends and members of the opposite sex (Baccaglini, 1993), provide excitement (Pennell et al., 1994), and give the illusion of being lucrative through the selling of drugs (Decker and Van Winkle, 1996). On the other hand, social, economic, and cultural forces push many young people into gangs. For example, Decker and Van Winkle (1996) argue that protection from other gangs and general well-being are key factors in the decision to join or not join a gang. Wilson (1987) further argues that the underclass status of many minority youths can also "push" them into gang membership. Finally, gangs can provide a solution for social adjustment problems such as a need for "belonging." As a result of the findings showing gangs as a vital problem in society, researchers are led to consider important policy implications for prevention and intervention strategies. First, the identification of multiple risk factors reveals the direction for prevention and intervention programs. These programs need to address the risk factors for gang membership by providing protective factors in order to strengthen a youth's resilience toward gang involvement. Second, because youths who join gangs tend to possess a great need for "belonging" at about age 13, join 6 months later, and have criminal records by the time they are 14, research indicates that programs have a window of opportunity for when effective prevention is best used. This "underscores the need for effective gang-resistance education programs and other primary and secondary prevention and intervention initiatives directed at preteens, especially those prone to delinquent and violent behavior" (Huff, 1998). The findings also indicate that effective intervention should address the brief window between the early "belonging" stage and the age of first arrest. A second window of opportunity exists between the time gang members are arrested for their first property crime and when they might graduate to more serious violent offenses. This period lasts "about 1.5 to 2 years, affords a chance to divert young offenders from the gang subculture before they further endanger their own lives and victimize other citizens" (Huff, 1998). Unfortunately the research also suggests that most government and private programs for gang prevention have been left unevaluated and the few evaluated programs have either failed to decrease gang violence or have actually increased it (Sherman et al., 1997). Moreover, gang prevention programs have ignored the most likely causes of the recent growth of gangs (e.g., the community structure of urban neighbors). Nonetheless, successful methods for preventing gang violence are available. The next section reviews some available community-based programs for preventing gang violence. EVIDENCE OF IMPACT Spergel and Curry (1990) documented that although suppression was the dominant response to gangs, it was the least effective (Decker, 2002). Howell's (1998) review of the literature reveals that "nothing has been demonstrated through rigorous evaluation to be effective in preventing or reducing serious and violent gang delinquency, [although] a number of promising strategies are available" (Howell, 1998). Moreover, Spergel's (1995) independent review of the literature reaches the same conclusion: "traditional social intervention programs, whether agency-based, outreach or street work, or crisis intervention, have shown little effect or may even have worsened the youth gang problem" (Spergel, 1995). There is still a need for "high-quality evaluation research" on gang programs, both comprehensive, broad-based intervention and single-agency interventions (Decker, 2002). The most recent review of gang intervention programs (Reed and Decker, 2002), reports that Boston's Operation Ceasefire (Braga and Kennedy, 2002), which engaged a broad array of local, State, and Federal officials, as well as community and neighborhood leaders, proved to be an effective and efficient response to youth violence and gangs. The results of the evaluation of the Gang Resistance Education and Training (GREAT) Program proved inconclusive, but did show several small but important differences between students who participated in the GREAT program and those who did not. The results of the evaluation indicate modest short-term benefits from the GREAT initiative. Compared with students who did not participate in the program, GREAT participants reported significantly more prosocial behaviors and attitudes than those students who did not participate. Consequently the authors report that while this program is no "silver bullet," it does "represent one program that holds promise [of] making some difference, and such programs are hard to find" (Esbensen and Osgood, 1999). All of these favorable outcomes emerged 3 or 4 years after program completion (Esbensen et al., 2002). The Chicago Gang Violence Reduction program targeted two of the most violent gangs in Chicago. Approximately 200 hardcore gang youths, 17 to 25 years of age, in the two gang constellations have been targeted. The program consists of two coordinated strategies: 1) targeted control of violent or potentially hardcore, violent youth gang offenders, in the form of an increased probation department and police supervision and suppression and 2) provision of a wide range of social services and opportunities for targeted youths. Spergel and Grossman (1997) found that, compared with six other high gang-violence areas, the target area had the lowest level of increase in violent gang incidents (i.e., a combination of gang homicides, aggravated batteries, and aggravated assaults) over the 4-year project period, versus the 4-year pre-project period. Homicide, however, did increase as compared with several of the other areas. The target area also had the lowest rate of increase in violent gang incidents for the numbers of serious gang offenders, and the second lowest rate in increase in Hispanic/Latino serious violent gang offenders versus the other six areas over the same 4-year program period. This study also found that the project was most effective in assisting older youths to significantly reduce their criminal activities (especially violence) more quickly than would have been the case if no project services were provided. The project did not appear to be effective with younger youths. The Boys and Girls Clubs of America's (BGCA) Gang Prevention Through Targeted Outreach program, a major OJJDP gang initiative, is designed to prevent youths from entering gangs, intervene with gang members in the early stages of gang involvement, and divert youths from gang activities into more constructive programs. Although no formal impact evaluation has been conducted, a process evaluation reported that once enrolled in BGCA, 90 percent of the youths attended once a week or more, with 26 percent attending daily. The evaluation also found that a large percentage of these youths received recognition for in-club and outside civic activities and as many as 48 percent of the youths showed improvement in the academic arena. Specifically, more than 33 percent of the youths showed improved grades, and as many as 33 percent had better attendance (Feyerherm et al., 1992). Moreover, the Boys and Girls Clubs of America's Targeted Outreach site in Fort Worth, Texas, has become a popular model for gang prevention initiatives. The Boys and Girls Clubs of Greater Fort Worth currently administers the Comin' Up Program and acts as a bridge between gang-involved youths and community services, opportunities, and government agencies. During a 4-month period ending in February 1996, more than 16,500 members and visitors participated in the Comin' Up Program, representing an average of 4,140 young people receiving services each month (Boys and Girls Clubs of Greater Fort Worth, 1996).