1. Demographic Characteristics
The data on using demographic characteristics in matching is mixed. Older adolescents are more likely to terminate the mentoring relationship early, as are lower income, minority, female, and married mentors in their mid-to-late 20's. However, the effect of martial status is attenuated by a strong relationship with youth, as are minority and female effects when the mentors share common interests with their youth (Grossman & Rhodes, 2002). Overall, perceived common interests appear to be of greater importance than ethnicity, race or gender (Ensher & Murphy, 1997). However, for JUMP youth, boys paired with male mentors were more likely to avoid drugs and gangs. In addition, greater improvement was reported when JUMP youth were paired with mentors of the same ethnicity or race (Novotney, et al., 2000). It may be that these demographic characteristics are more important for youth at-risk for delinquency.

Several programs have attempted to address these concerns. OJJDP’s Safefuture Initiative report (Office of Juvenile Justice and Delinquency Prevention, 2000) notes that cultural differences can interfere in the implementation of programs. Faith-based programs face difficulties recruiting young people with similar religious beliefs. In some cultures, mentoring is a foreign concept for families; parents feel uneasy about their children spending time with strangers. Assigning a mentor from the same ethnic group or cultural background can alleviate their concern, but it is often difficult to find volunteer mentors with the time to participate in the program. To address these issues, Safefutures uses group instead of one-to-one mentoring (Office of Juvenile Justice and Delinquency Prevention, 2000). Blechman (1992) has suggested that mentors for high-risk minority youth be bi-culturally competent.

The BBBS program in New York City developed a culturally sensitive mentoring program for immigrant youth called the New American Partnership. Many of the training sessions include cultural education to increase awareness of cultural diversity. Even though mentors are well-trained to enter cultural environments different from their own, it is still difficult to break the cultural barrier. For immigrant youth, the biggest challenge is adapting to mainstream American culture.  The New American Partnership program matches immigrant youth with first-generation Americans or someone whose cultural background is similar. The project is designed to empower immigrant youth by matching them with adults who have overcome the same difficulties and understand the youth’s situation. The mentors assist them to adapt to the new culture and at the same time, enjoy sharing their own culture (BBBS NYC, 2004).

When implementing a program, it is important to consider culture. Styles of kinship, lifestyle, and family dynamics are culturally diverse.  For example, in Asian communities, extended family members often have strong bonds and close relationships with members of the Asian community. In these communities, there are natural mentors to guide and support young people. A recent study indicates that children with natural mentors have better attitudes toward school and fewer behavioral problems (Zimmerman, Bingenheimer, & Notaro, 2002).

2. Role of Specific Behaviors for Improvement
The Guide for Implementing the Comprehensive Strategy for Violent, and Chronic Juvenile Offenders, published by OJJDP, suggested that the use of contingency reinforcement was more effective in mentoring programs than just the use of the mentoring relationship (Howell, 1995). This recommendation was based on a study of a mentoring program called the Buddy System showing greater improvements in specific behaviors when contingency reinforcement was used (Fo & O'Donnell, 1974). However, the BBBS study showed improvement without the use of contingency reinforcement (Tierney, et al., 1995).

Pitting the importance of the mentoring relationship against a focus on improvement in specific behaviors, using contingency reinforcement or other means, is a pseudo-issue that has led some researchers into an intellectual cul-de-sac. In the Buddy System, all mentors were trained in developing their relationships with their mentees. Having good rapport and a strong relationship with their youth was emphasized. In addition, when contingency reinforcement was used to target specific behaviors, most of the behaviors showed improvement (Fo & O'Donnell, 1974). Indeed, it is difficult to imagine a successful mentoring program without strong relationships among mentors and youth. Relationships are the heart of the mentoring process.

Therefore, can improvements occur in mentoring programs that focus only on developing strong relationships? Of course they can. A strong relationship can be a powerful influence. Given a strong relationship, can a focus on improving specific behaviors be useful in facilitating improvements in behavior? Of course it can. For example, the academic gains reported in school-mentoring programs occurred because of their focus on academic behavior.  In another study, gains were reported through the use of participation in a physical fitness program as a reward (Galbavy, 2004). It may be especially important to focus on specific behaviors with youth who are referred for specific problem behaviors or who are at-risk for engaging in illegal activities. In a study of a diversion program, skill training showed better results than mentoring (Blechman, Maurice, & Buecker, 2000). Focusing on specific behavior change is not only compatible with developing strong mentoring relationships, but can be an important component in the success of the mentoring relationship. The greater use of a focus on specific behaviors in mentoring may increase effects beyond the modest gains found in the meta-analysis of mentoring programs (DuBois, Holloway, et al., 2002).

Back to Top

3. Youth Involved in the Justice System
Perhaps the most important finding from the Buddy System was that mentoring for delinquency prevention could be harmful to some youth. Youth, who had not been arrested for a major offense, were more likely to be arrested if they participated in the program than if they were in the randomly assigned control group. In contrast, youth who had been arrested were less likely to be arrested again if they participated in the program. These results were interpreted as a peer-network effect, where youth who had not been arrested formed relationships during the program with other youth who had been arrested. The results were first reported using a small sample (Fo & O'Donnell, 1975), then confirmed with the complete sample (O'Donnell, Lydgate, & Fo, 1979).

Subsequently, the results of several other delinquency prevention programs, from the famous Cambridge-Somerville study (McCord, 1978) to Scared Straight (Buckner & Chesney-Lind, 1983; Finckenauer, 1979) were reviewed and their results were found to be consistent with the peer-network interpretation (O'Donnell, 1992; O'Donnell, Manos, & Chesney-Lind, 1987). Recently, a theoretical model was proposed, in which the important effects, both positive and negative, that parents, schools, and communities can have on delinquency, is mediated by their effect on peer networks (O'Donnell, 1998, 2000, 2003).

A recent reanalysis of the Cambridge-Somerville study supports the peer-network theory. The reanalysis indicated that negative effects occurred among those who were sent to summer camp more than once during the intervention. Friendship was built among boys during the summer camps, which led to negative effects into their adult lives (Dishion, Poulin, & McCord, 1999).

One way to avoid these potential harmful effects is to just use one-to-one mentoring without providing program youth opportunities to have contact with each other. All mentoring programs, especially those that use group mentoring, need to monitor the possible peer relationships forming within the program and to assess the peer networks of their youth. Programs that find potentially harmful peer networks forming or existing with peers outside of the program need to make this concern a priority and take steps to reduce the potential harm. Facilitating pro-social peer networks, whenever possible, should be a goal of all mentoring programs.

In addition, greater study of natural mentors is needed. In a provocative study, youth with natural mentors were found to be less affected by the negative behavior of their peers. It was concluded, “natural mentors may encourage young people not to befriend peers who engage in problem behaviors” (Zimmerman, et al., 2002, p. 238). Learning more about how this process works could have great impact.

The peer-network effect is of particular importance to mentoring programs for youth in the juvenile justice system or who have a parent who is incarcerated. In addition to the academic and social problems of other at-risk youth, children of prisoners “are six times more likely than other children to be incarcerated at some point in their lives” (Farley, 2004, p. 1). Clearly, successful prevention programs for these youth are vitally important.

Some mentoring programs focus on these youth. One is Amachi, which reports youth improvements in self-confidence and academic performance (Farley, 2004). Other programs help young people who are transitioning out of the juvenile justice system to support their reentry. Volunteers in Prevention, Probation and Prisons(VIP) is the pioneer of mentoring juvenile offenders, for youth convicted of misdemeanors (VIP, 2004). The program has been replicated nationwide. In one of the VIP sites, Contra Costa County in California, mentors are trained to serve two roles: to provide personal attention to youth, which probation officers usually cannot provide, and to takeover part of the probation officer’s tasks, such as making telephone calls, providing transportation, etc.

The mentoring program launched by the Seattle Office of the Juvenile Rehabilitation Administration in 1996 report preliminary findings showing the mentoring group with a 34% lower felony recidivism rate compared to the control group. In this program, mentoring begins four to six months before release and mentors build trusting relationships during confinement by visits, e-mail, and telephone calls. During the reentry process, mentors help youth to find employment or schooling and housing all at a cost of $3,000 per person (Washington State Institute for Public Policy, 2002).

The importance of developing successful prevention programs for youth in the juvenile justice system and with incarcerated parents cannot be overestimated. Additional research on the development of such programs should be a major priority.

Back to Top