Introduction

In recent years, mentoring has become a popular strategy for fostering health development and preventing negative outcomes for youth, particularly those considered at risk. It is estimated that three million youth are in mentoring relationships based in formal programs in the U.S. (Rhodes & Lowe, 2009). Additionally, government agencies have invested money in programs that provide mentoring to at-risk children. Enthusiasm for mentoring as a practice has, at times, run ahead of the research base regarding its effectiveness. Increasingly, however, researchers are focusing on documenting program outcomes and effective practices in mentoring relationships. We explain the concept of mentoring children as a prevention strategy, including a review of theory, research, and practice.

Much of the research in youth mentoring has focused on adolescents or has used samples that range from childhood to adolescence. However, there are issues unique to mentoring children; Cavell and Smith (2005) discuss both the potential barriers and benefits in mentoring children rather than adolescents. They suggest that a potential disadvantage of focusing on younger children is that the children’s verbal and cognitive capacities can limit the scope of mentoring (e.g., make communication more difficult or not allow for explorations of identity). Advantages are that most potential mentors will find children (rather than adolescents) to be more approachable, receptive, and malleable, and that prevention – before significant problem behaviors and patterns exist – will be more effective than intervention after the fact.

Definitions and Scope

Rhodes (2002) provided a useful definition of mentoring as “…a relationship between an older, more experienced adult and an unrelated, younger protégé – a relationship in which the adult provides ongoing guidance, instruction, and encouragement aimed at developing the competence and character of the protégé” (p. 3). Mentoring may occur informally or in structured programs. In this entry, we focus on formal mentoring programs, because much of the research on childhood mentoring is done in the context of formal programs. Instead of relying on mentoring relationships to form naturally, programs use a systematic approach to help youth and adults develop partnerships. Formal mentoring programs may use evidence-based and standardized policies, or best practices, to match adults and children and to ensure that mentoring relationships effectively meet the needs of the population being served (Britner & Kraimer-Rickaby, 2005; Thompson & Kelly-Vance, 2001).

Formal programs mostly fall into two categories: community-based and school-based. In community-based mentoring, the mentor and protégé usually determine the frequency of their meetings, which may be an hour a week or a little as once a month, as well as the location. In-person meetings can take place at school or other community settings, such as a movie theater or a recreational center.

Alternatively, in school-based programs, mentoring takes place during the school day; typically the mentor and protégé spend 1 h a week together engaging in academic and social activities (Herrera, 2004). Compared to community-based programs, school-based programs are more likely to serve children who are having academic or behavioral difficulties in school and/or who have repeated a grade (Herrera, Sipe, & McClanahan, with Arbreton, & Pepper, 2000). In this endeavor, school-based mentoring programs are often more structured and place greater emphasis on improving students’ academic performance than community-based programs (Portwood & Ayers, 2005). Traditionally, most mentoring programs have been community based, but the simplified logistics (i.e., taking place during school hours and using the schools as a meeting place) of mentoring within schools combined with schools’ needs to find ways to help individual students succeed has led to an increase in school-based mentoring.

Theories

Several models have been presented to describe the process of how mentoring might “work.” The most widely cited of these models is Rhodes’ “pathways of mentoring influence” model. The model of youth mentoring proposed by Rhodes (2002) suggests that there is a crucial step to mentoring in which mentor and youth protégé form an emotional bond. Theoretically, only then can the mentor influence the youth’s developmental outcomes by (a) enhancing social skills and emotional well-being, (b) improving cognitive skills through dialogue, and (c) serving as a role model and advocate for the youth.

Positive mentoring relationships can foster social and emotional development by challenging the children’s views of themselves in relationships, by providing a “corrective experience” that generalizes to other relationships, by helping children regulate their emotions, and by alleviating relationship stress. Mentoring relationships can enhance identity development through positive role modeling and also enhance cognitive development through dialogue and listening (Rhodes, 2002).

Whereas Rhodes’ model is the most popular, other researchers have proposed ideas to explicate how mentoring works. For example, Parra, DuBois, Neville, Pugh-Lilly, and Povinelli (2002) presented a structural model in which mentoring, followed by the development of a significant bond, widens the youths’ social support networks, which in turn bolsters self-esteem. By encouraging trust, autonomy, and initiative, mentors enhance resilience in development. Many mentoring researchers place a strong emphasis on self-esteem as a key variable for determining resilience in childhood (see entry on “Self-Esteem During Childhood”). Larose and Tarabulsy’s (2005) sociomotivational perspective portrays structure, involvement, and support for autonomy as key elements for student motivation and academic achievement in mentoring programs for academically at-risk children. Keller’s (2005) systemic conceptual model considers the interdependent network of relationships established between mentor, child, parent/guardian, and caseworkers against the ecological backdrop of agency policies and procedures. Ironically, this important “relationship” focus has rarely been applied to studies of mentoring as embedded within larger social systems (Britner & Kraimer-Rickaby, 2009).

Current Research

The prevalence of and support for mentoring programs has grown in recent years at a faster rate than the body of research on their practices and outcomes. When studies have been conducted in the past, many have suffered methodologically from an over-reliance on anecdotal, observational, and/or self-report data and a failure to conduct rigorous statistical analyses (Rhodes, 2002). However, there is an increasing body of literature on the efficacy and effectiveness of various mentoring programs to meet the needs of children.

In this section, we begin by presenting broad results from two meta-analyses on the impact of mentoring. We then examine some specific studies that address child outcomes and mentoring. Many studies of mentoring utilize samples that include both children and adolescents. Therefore, we conclude this section with a brief discussion of what is known about the differences in mentoring children versus adolescents.

Meta-analyses

DuBois, Holloway, Valentine, and Cooper (2002) published a widely cited and comprehensive meta-analysis of the impact of mentoring in studies published through 1998. Five outcomes were examined: emotional/psychological, problem/high-risk behavior, social competence, academic/educational, and career/employment. Mentoring was found to have a measurable, significant effect on children and adolescents; however, the effect size was relatively small (0.14 and 0.18 under the assumptions of fixed and random effects, respectively). In particular, it was found that career development outcomes, academic outcomes, and problem behaviors were influenced by mentoring (DuBois et al., 2002).

More recently, DuBois, Portillo, Rhodes, Silverthorn, and Valentine (2011) completed a meta-analysis of 73 independent evaluations of mentoring programs published from 1999 to 2010 (i.e., the era after the first meta-analysis). Findings were largely consistent with the earlier analysis, with additional support for gains in social competence or emotional well-being. The authors conclude that modest effects (i.e., approximately a 9 % point advantage for mentored children over non-mentored children) are typical, across domains, within the programs profiled in the analysis.

An important consideration in using mentoring as a prevention strategy is its usefulness with certain at-risk populations. The DuBois et al. (2002, 2011) meta-analyses both found that children from backgrounds of moderate levels of risk (defined broadly) have the most to gain and tend to benefit the most from mentoring, especially when best practices are employed and strong relationships are formed. This finding supports the use of mentoring as a prevention strategy.

Specific Study Outcomes

Whereas a variety of outcomes were examined in the meta-analyses cited above, most of the individual studies on mentoring in childhood focus on academic outcomes. These are discussed here in two categories: school performance and academic self-concept. There have also been some findings relating to social development, which are also discussed.

School performance. Mentoring is often promoted as a remedy for children who are academically at risk or already failing in school (Larose & Tarabulsy, 2005). Research exploring this assertion has yielded mixed results. Thompson and Kelly-Vance (2001), using a small sample of 12 boys involved in Big Brothers Big Sisters and 13 boys who were on the waiting list for the program, reported that mentored children made significantly higher gains on a standardized intelligence test than the control group, even after controlling for ability.

Herrera, Grossman, Kauh, and McMaken (2011), with a large sample of 1,139 youth, conducted a random assignment study of the Big Brothers Big Sisters School-Based Mentoring program; 565 youth were assigned to the treatment group to receive mentoring and 547 youth were placed on agency wait lists to serve as a control group. Youth were followed for 1.5 schools years; after the first year, the mentored youth performed better academically (as measured by teacher reports) than did the control group youth.

Not all studies find improvement in academic performance. Kolar and McBride (2011) studied a sample of 160 at-risk students in first through sixth grade who participated in a Big Brothers Big Sisters mentoring program. At the end of the program, there were a number of improvements on some psychological variables, but no improvements in grades were found. Similarly, Karcher (2008), in a study of 516 (mostly Latino) youth randomized to supportive services or supportive services plus mentoring, found no effect on grades but several positive social outcomes (which varied by sex and age).

Academic self-concept. Previous research suggests that mentored students may benefit academically from the presence of a mentor by a change in their academic self-perceptions. A meta-analysis suggests that although the effect is small, positive self-beliefs, in fact, predict later achievement, particularly for the academic domain (Valentine, DuBois, & Cooper, 2004). Guay, LaRose, and Boivin (2004) found that school-age children’s academic self-concept predicted educational attainment level 10 years later.

In the Herrera et al. (2011) Big Brothers Big Sisters School-Based Mentoring study, the randomly assigned treatment group reported greater improvement in perceptions of their academic abilities than youth in the control group. Given the promise of past research and the constant pressure on public schools to “leave no child behind,” school-based mentoring and its possible link to enhanced academic outcomes should continue to be explored.

Social and emotional development. Elledge, Cavell, Ogle, and Newgent (2010) studied the impact of a school-based lunchtime mentoring program for bullied children in grades 4 and 5. The 24 students in the treatment group met with a college student mentor twice a week for one semester. In the study, there were two control groups: a “same” control group made up of 12 students and the same school and a “different” control group of 12 children from another school. At the conclusion of the program, treatment children experienced greater reductions in peer reports of peer victimization than children in the “different” control group. However, there were no statistically significant differences between treatment children and the “same” control group; the small sample sizes might have made it difficult to detect changes, but until further research is done, it is unclear what the full impact one-on-one mentoring might have on children who are bullied.

In the Kolar and McBride (2011) study, there were improvements on a number of psychological variables. Pre-/post-analyses found that children liked school better, had better classroom behavior, got along better with peers, and had increased self-esteem after a year spent with a mentor. The authors conclude that a relatively short period of school-based mentoring can have important impacts on at-risk youth; while these results are positive, they should be interpreted with caution, as there was no control group included in the study.

Focusing on different risk populations, several recent studies have found effects of mentoring, often combined with other skills training, on social and emotional outcomes. Jent and Niec (2009) randomly assigned 86 children (ages 8–12 years) accessing services at a community mental health center to a group mentoring intervention or wait-list control group; the mentored children reported improved social problem-solving skills, and their parents reported that the children had decreased internalizing and externalizing behavior problems. Kuperminc, Thomason, DiMeo, and Broomfield-Massey (2011) found that urban girls in a youth development program who were also mentored were more likely than the program’s non-mentored girls to experience gains in social acceptance and body image. Taussig and Culhane (2010) employed a randomized control trial design to test the effectiveness of the Fostering Healthier Futures (FHF) intervention with 9- to 11-year-old children in foster care. The intervention group (n = 79) received an assessment, a manualized skills curriculum, and a 30-week mentoring program with graduate social work students. At a 6-month follow-up, the intervention group – relative to the control group (n = 77) – had fewer mental health problems and symptoms of dissociation, and they were also less likely to have received mental health treatment.

Research on Childhood Versus Adolescent Mentoring

The age and developmental characteristics of the mentee can impact mentoring relationship. There is a larger literature on adolescents (see entry “Mentoring During Adolescence,” this volume). Some researchers have found that the mentoring relationships of children are closer than those of older adolescents (Britner & Kraimer-Rickaby, 2009; Herrera et al., 2000). Additionally, mentoring relationships with children have a lower risk of early termination (Grossman & Rhodes, 2002) and generally have a longer duration (Rhodes & Lowe, 2009) than those of older adolescents. In their meta-analyses, however, DuBois et al. (2002, 2011) found that effect sizes were similar across different mentee age groups (childhood/early adolescence vs. middle/late adolescence).

Overview of Strategies

Research on mentoring has primarily been concerned with whether specific programs are associated with improved outcomes for children. There are many different mentoring programs; they range from programs with a single site to programs that are widely administered across many sites. Due to this variability in the nature of programs, very few have been able to conduct large-scale, high-quality evaluations to provide the needed evidence to conclude they “work.” These programs are presented in the next section. There has been a recognition among researchers that much of the impact of mentoring depends on the quality of implementation (DuBois et al., 2002, 2011). Moving beyond simply determining if a program “works” or not, many researchers have focused on determining program-related practices that contribute to the success of mentoring relationships. Because these studies often do not focus on any one particular program, they provide generalizable findings that can be implemented in a variety of mentoring practice settings.

What Works

A search of the research literature did not uncover an intervention that met the standard of three successful trials. Nevertheless, there is evidence for other interventions and strategies that – while not meeting the above criteria – yield considerable amounts of evidence and therefore are widely considered as empirically based. These interventions and strategies are described in the next section. Meanwhile, efforts continue to conduct rigorous research on mentoring during childhood.

What Is Promising

There are some general program characteristics that seem to enhance the likelihood of producing positive outcomes. The DuBois et al. (2002, 2011) meta-analyses found that programs demonstrate a larger effect for mentoring when certain best practices are present. Examples of these practices, which increased relationship intensity and longevity, include more than 2 h of intensive training for mentors, structured group activities for mentors and youth, and a system for ongoing monitoring of the mentor-protégé match (DuBois et al., 2002). Other research supports the impact of “best practices.” For example, practices such as mentor screening, support, and supervision were identified as predictors of more positive outcomes for mentored children (Grossman & Tierney, 1998; Herrera, 2004).

Most program-level mentoring strategies can be considered “promising practices.” Many have shown positive results, but only in a few studies or in studies that have methodological flaws and/or small sample sizes. Some have shown mixed results across studies and more research is needed to clarify the circumstances under which the strategies are most effective. To date, two programs have been studied frequently enough with consistently positive results to confidently say they stand out as “promising” during childhood: Big Brothers Big Sisters and Across Ages. Presented below are summaries of these two programs and an overview of the research that supports their effectiveness.

Big Brothers Big Sisters

Big Brothers Big Sisters is a highly structured program that facilitates one-on-one mentoring relationships for 5- to 18-year-olds who primarily come from single-parent families. Big Brother Big Sisters is purely a mentoring program; this makes it unique from other programs which provide mentoring as one part of an array of services. Big Brothers Big Sisters is one of the most widely recognized programs; the agency estimates that it serves 245,000 youth across 370 separate agencies (Big Brothers Big Sisters, 2009).

Big Brothers Big Sisters is the focus of the most influential study on the impact of one-on-one mentoring on the lives of youth conducted to date. The study followed 959 youth (487 received mentors; 472 were waiting list controls) between the ages 10 and 14 years over a period of 18 months, while the treatment youth were participating in the Big Brothers Big Sisters program (Tierney, Grossman, & Resch, 1995). Many of the youth had experienced several parental factors that placed them at risk, including parental divorce or separation or a family history of substance abuse or domestic violence (Grossman & Tierney, 1998).

Following the intervention, youth in the treatment group showed significant differences relative to the waiting list control group in the domains of antisocial behavior, academic achievement, and family and peer support. Compared to youth who did not have a mentor, youth who had mentors for at least 1 year were 46 % less likely to initiate drug abuse (70 % less for minority youth) and 27 % less likely to start using alcohol. These youth were also 32 % less likely to hit someone, and they skipped, on average, 42 % fewer days of school (Tierney et al., 1995).

In the evaluation of the school-based Big Brothers Big Sisters model (Herrera et al., 2011), mentees exhibited higher academic achievement (as measured by teacher report) and more positive perceptions of their academic abilities. Herrera et al. did not find any differences between mentees and wait-list youth in perceptions of their self-worth, relationships with their parents, or rates of problem behavior. They suggest that the school-based programs may impact primarily school-related outcomes, whereas community-based models may influence a broader range of outcomes.

Mentoring Plus: Across Ages

According to Sipe (1996), combining mentoring with other services may increase the efficacy of the mentoring relationship to promote positive outcomes for youth. The most widely cited example of a “mentoring plus” model is Across Ages. Across Ages is an intergenerational and comprehensive drug prevention program that provides youth with older mentors (aged 55 and older). In addition to mentoring, Across Ages provides youth with community service activities, positive youth development (e.g., life skills development), and parental involvement. It is the only program with a mentoring component that has been identified as a “model program” through the Substance Abuse and Mental Health Administration’s National Registry of Evidence-based Programs and Practices (Substance Abuse and Mental Health Service Administration, 2008).

LoSciuto, Rajala, Townsend, and Taylor (1996) investigated the effectiveness of the Across Ages program using a random assignment pretest-posttest design; 159 youth were divided into three groups: group C, the control group; group PS, which was comprised of youth who received positive youth development, participated in community services activities, and had parent workshops; and group MPC, which was made up of youth who received mentoring in addition to positive youth development, community services activities, and participated in parent workshops. Youth were followed for one academic year. Youth who received mentoring in combination with other services had fewer absences from school, used drugs less frequently, and had a more positive attitude toward school, the future, and elders.

In another study of the Across Ages program (Aseltine, Dupre, & Lamlein, 2000), the most dramatic changes in self-confidence also were attributable to the combination of mentoring, community service, and positive youth development curriculum. Perhaps because of the experimental research findings and the program’s flexible delivery (in school- or community-based settings), Across Ages has maintained its popularity, while some other models have fallen out of favor.

What Does Not Work

Poor mentoring relationships and matches that disrupt within 3 months or matches that meet inconsistently can have negative emotional, psychological, and cognitive consequences (Grossman & Rhodes, 2002; Sipe, 1996). For example, Grossman and Rhodes (2002) found that Big Brothers Big Sisters youth whose matches terminated within the first 3 months suffered significant declines in their global self-worth and their perceived scholastic competence; conversely, relationships that lasted more than 12 months reported the greatest number of positive outcomes (Grossman & Rhodes, 2002). In a study of the Big Brothers Big Sisters School-Based program (using the same samples as the Herrera et al. study discussed above), Grossman, Chan, Schwartz, and Rhodes (2011) found that although treatment children on average had academic improvements, children whose match terminated prematurely did not experience any such improvement. Additionally, children who were rematched after a failed mentoring relationship actually showed negative impacts. In order to detect any benefit, Rhodes (2002) has suggested that mentors meet with their protégés for a minimum of 4 h per month for at least 1 year.

Summary

Whereas some questions remain about the overall effectiveness of mentoring in childhood, we do know that when implemented effectively mentoring can be an effective prevention strategy. Both community- and school-based programs have demonstrated impacts on children’s academic performance, academic self-concepts, and social/emotional development, though there are often limits to the effectiveness of any one program (e.g., attitudes toward school improve but grades do not). Mentoring usually works best for children who have a moderate level of risk, making it an appealing prevention strategy. Additionally, the idea has been put forth that mentoring that begins in childhood, as opposed to in adolescence, might be more effective as it serves a protective function and can help prevent children from going on to develop negative behaviors.

Perhaps most important to keep in mind when reviewing the effectiveness of mentoring is that most of the benefits of mentoring accrue only after mentor and mentee have spent a significant amount of time together; short-term or failed matches can have no effect or even negative effects. When mentoring takes place within formal programs, as is the case in many instances, it is important that programs follow best practices as far as recruiting, training, and assigning mentors.

See Also

Mentoring During Adolescence

Resiliency During Childhood

Social and Emotional Learning During Early Childhood