Mentoring Interventions to Affect Juvenile Delinquency and Associated Problems: A Systematic Review
Crime and Justice
Mentoring has drawn substantial interest from policymakers, intervention theorists, and those interested in identifying promising and useful evidence-based approaches to interventions for criminal justice and child welfare outcomes (Grossman & Tierney, 1998; Jekliek et al., 2002). Mentoring is one of the most commonly-used interventions to prevent, divert, and remediate youth engaged in, or thought to be at risk for, delinquent behavior, school failure, aggression, or other antisocial behavior (DuBois, Holloway, Valentine, & Cooper, 2002, DuBois, Portillo, Rhodes, Silverthorn, & Valentine, 2011). One account lists over 5000 organizations within the United States that use mentoring to promote youth wellbeing and reduce risk (MENTOR/National Mentoring Partnership, 2006). Definitions of mentoring vary, but there are common elements. For the purpose of this review, mentoring was defined by the following 4 characteristics: 1) interaction between two individuals over an extended period of time, 2) inequality of experience, knowledge, or power between the mentor and mentee (recipient), with the mentor possessing the greater share, 3) the mentee is in a position to imitate and benefit from the knowledge, skill, ability, or experience of the mentor, 4) the absence of the role inequality that typifies other helping relationships and is marked by professional training, certification, or predetermined status differences such as parent-child or teacher-student relationships. A total of 46 topic and methodologically eligible studies (out of 164 outcome reports) were identified for inclusion in the meta-analysis on delinquency and outcomes associated to delinquency: aggression, drug use, and academic achievement.
This systematic review had the following objectives:
a) To statistically characterize the evidence to date on the effects of mentoring interventions (selective and indicated) for delinquency (e.g. arrest, reported delinquency), and related problems of aggression drug use, school failure.
b) To attempt to clarify the variation in effects of mentoring related to program organization and delivery, study methodology, and participant characteristics.
c) To help define mentoring in a more systematic fashion than has occurred to date to, in turn, help clarify how intervention processes suggested as compromising how mentoring has effects and other important considerations for future research..
d) To inform policy about the value of mentoring and the key features for utility.
This is an update of a review completed 4 years ago. In the original review search we benefitted from the authors of three meta-analyses on mentoring or related topics (1) DuBois et al. (2002) on mentoring in general, 2) Lipsey and Wilson (1998) on delinquency interventions in general, and 3) Aos et al. (2004) on interventions for delinquency and associated social problems) who provided databases on reports and coding approaches. In addition, we searched various databases including PsychINFO, Criminal Justice Abstracts, Criminal Justice Periodicals Index, Social Sciences Citation Index (SSCI), Science Citation Index (SCI), Applied Social Sciences Indexes and Abstracts (ASSIA), MEDLINE, Science Direct, Sociological Abstracts, Dissertation Abstracts, Database of Abstracts of Reviews of Effectiveness, and ERIC (Education Resources Information Center) and the Social, Psychological, Educational and Criminological Trials Register (SPECTR- in original search), the National Research Register (NRR, research in progress), and SIGLE (System for Information on Grey Literature in Europe). Finally, the reference lists of primary studies and reviews in studies identified from the search of electronic resources were scanned for any not-yet identified studies that were relevant to the systematic review. For this update we searched the same databases (except SPECTR as it no longer existed), surveyed pertinent journals and the reference lists of primary studies and reviews.
1. Studies that focused on youth who were at risk for juvenile delinquency or who were currently involved in delinquent behavior. Risk is defined as the presence of individual or ecological characteristics that increase the probability of delinquency in later adolescence or adulthood.
2. We included interventions focusing on prevention for those at-risk (selective interventions) and treatment (indicated interventions) that included mentoring as the intervention or one component of the intervention and at least measured impact of the program. We excluded studies in which the intervention was explicitly psychotherapeutic, behavior modification, or cognitive behavioral training and indicated provision of helping services as part of a professional role.
3. We required studies to measure at least one quantitative effect on one of the four outcomes (delinquency, aggression, substance use, academic achievement) in a comparison of mentoring to a control condition. Experimental and high quality quasi-experimental designs were included.
4. The review was limited to studies conducted within the United States or another predominately English-speaking country and reported in English and to studies reported between 1970 and 2011. We did not have resources for translating reports not reported in English.
DATA COLLECTION AND ANALYSIS
All eligible studies were coded using a protocol derived from three related prior meta-analyses, with 20% double-coded. The intervention effect for each outcome was standardized using well established methods to calculate an effect size with 95% confidence intervals for each of the four outcomes (if included in that study): delinquency, aggression, drug use and academic achievement. Meta-analyses were then conducted for each independent study within a given outcome (delinquency, aggression, drug use, and academic achievement). Effect sizes for each study were scaled so that a positive effect indicated a desirable outcome (i.e., lower delinquency, drug use, and aggression or higher academic achievement).
A total of 164 studies were identified as meeting inclusion criteria as focused on delinquency and mentoring. Of these, 46 met the additional criteria for inclusion in the quantitative analyses. 27 were randomized controlled trials and 19 were quasi-experimental studies involving non-random assignment, but with matched comparison groups as was described above. Twenty-five studies reported delinquency outcomes, 25 reported academic achievement outcomes, 6 reported drug use outcomes, and 7 reported aggression outcomes.
Main effects sizes were positive and statistically significant for all four outcomes. Some studies showed effects that were not significant and a few reported negative effects. For each outcome there was substantial variation in effect size, too. Average effects were larger for delinquency than for other outcomes. When moderation was tested, there was considerable variation in effect sizes of studies that were similar in regard to the presence of a given moderating influence.
We compared effect sizes of those studies that were random assignment experimental designs with those that were quasi-experimental using meta-regression and found no evidence of differences in effect sizes. We conducted moderator analyses to determine whether effects found differed by 1) criteria for selecting participants, 2) presence of other components along with the mentoring intervention, 3) motivation of mentors for participation, or 4) assessment of quality or fidelity of implementation of the intervention. We also conducted moderator analyses to test for outcome differences by the presence or absence of four theorized key components of mentoring interventions. The relatively limited information about potential moderating characteristics extractable from many reports and the limited number of reports with extractable information led us to combine effects across all four outcomes to enable adequate power and in combination to our directional expectations for moderators to test significance using a one-tailed test (p < .05). For these analyses, we averaged effect sizes within a given study if more than one outcome of interest was reported. We also conducted analyses to check for bias in effects due to type of outcome, and found no suggestion of bias.
We found evidence for moderation when professional development was a motive for becoming a mentor. There was also moderation of the effect size when mentoring programs emphasized either of two theorized components: emotional support or advocacy. Effect sizes did not differ by whether or not the program emphasized the other two key components: modeling/identification or teaching, nor by whether other components were used, how risk was defined (environmental versus individual characteristics) or if fidelity/adherence of implementation features were assessed.
This analysis of 46 studies on four outcomes measuring delinquency or closely related outcomes of aggression, drug use, and academic functioning suggests mentoring for high-risk youth has a modest positive effect for delinquency and academic functioning, with trends suggesting similar benefits for aggression and drug use. Effect sizes varied more for delinquency and academic achievement than for aggression and drug use. We did not find a significant difference in effect size by study design (RA vs. QE) or by whether or not fidelity was assessed. We identified some characteristics that moderated effects that provide additional understanding for further studies and program design. Effects tended to be stronger when professional development was an explicit motive for participation of the mentors. Of four processes theorized as comprising the methods of effects in mentoring, we found evidence for significantly larger effects when emotional support and advocacy were emphasized. Although these findings support viewing mentoring as a useful approach for intervention to lessen delinquency risk or involvement, limited description of content of mentoring programs and substantial variation in what is included as part of mentoring efforts detracts from better understanding about what might account for the benefits. The valuable features and most promising approaches cannot be ascertained with any certainty. In fact, the body of studies is remarkably lacking in description of key features, program design organization, and theorized processes of impact that are typically provided in empirical reports of intervention effects. Our judgment is also that there does not seem to be much progression in quality of details in reports over the time period studied here. Given the popularity of this approach, the promise of benefits should be seen as a strong argument for a concerted effort through quality randomized trials to specify the theoretical and practical components for effective mentoring with high-risk youth. Concordantly, lacking such features, further trials may not add useful knowledge.