Youth Empowerment Programs for Improving Adolescents’ Self-Efficacy and Self-Esteem

Abstract

Objectives: Assess the state of evidence regarding impacts of youth empowerment programs (YEPs) on adolescents’ (ages 10–19) self-efficacy and self-esteem, as well as other social, emotional, and behavioral outcomes. Method: Systematic searches of databases combined with an international outreach to identify experimental or quasi-experimental trials of community interventions that regularly involved youths in decision making. Results: Of the 8,789 citations identified, 3 studies met the inclusion criteria. None reported significant effects on the review’s primary outcomes. Data from only two studies on self-efficacy could be meta-analyzed (combined N = 167). Results found no significant intervention effect on self-efficacy (z = 1.21; 95% confidence interval [CI]: [−0.12, 0.49]). Significant effects were found for some secondary outcomes, but these were inconsistent across studies. Conclusions: The review reveals insufficient evidence of YEPs’ impacts. Further research is needed using well-implemented models with clear theories of change, larger samples, and rigorous impact study designs complemented by mixed-methods process evaluation.

Keywords

youth development empowerment adolescents impact social–emotional

For many individuals, adolescence represents a challenging developmental period. In particular, those adolescents experiencing multiple psychosocial risk factors, such as low self-esteem, hopelessness, exposure to trauma, and associations with negative peer groups, are significantly more likely to participate in antisocial behavior including drug use, violence, delinquency, and early sexual activity (Flannery, Singer, & Wester, 2001; Jessor, Van Den Bos, Vanderryn, Costa, & Turbin, 1995; Loeber, 1990). Although emotional and behavioral problem trends among adolescents for several countries, such as the United States, the United Kingdom, and the Netherlands, have begun to plateau or modestly decline in recent years, they remain at concerning levels (The Nuffield Foundation, 2009).

On the other hand, adolescence offers a dynamic developmental opportunity to foster young people’s strengths and thereby increase the likelihood of positive outcomes. Interventions that strengthen protective factors in adolescents’ attitudes, skills, and relationships may be able to have far-reaching impacts on young people’s ability to overcome adversity and make successful transitions to adulthood. Psychosocial protective factors can counteract the effects of risk. They predict higher avoidance of problem behaviors as well as higher educational success and healthy development (Aspy et al., 2004; Dumont & Provost, 1999; Jessor et al., 1995; Kia-Keating, Dowdy, Morgan, & Noam, 2011; Oman et al., 2004; Rutter, 1987).

Increasingly, funders and practitioners around the world promote youth empowerment—strength-based, participatory programming that engages young people in decision-making processes—as a strategy for increasing a wide range of potentially malleable protective factors among adolescents in order to help them reach greater well-being. Drawing on positive youth development frameworks (Roth & Brooks-Gunn 2003), empowerment theory (Zimmerman, 2000), and social learning and cognitive theories (Akers, Krohn, Lanza-Kaduce, & Radosevich, 1979; Bandura, 1993), youth empowerment programs (YEPs) aim to use highly participatory, youth-driven processes to help young people strengthen positive attitudes, skills, and behaviors that improve functioning across a range of life domains (Jennings, Parra-Medina, Messias, & McLoughlin, 2006; Wong, Zimmerman, & Parker, 2010). The extent to which data show whether, when, and how YEPs achieve such impacts on psychosocial protective factors, however, has been largely unknown (Gray & Hayes, 2008; Zeldin, McDaniel, Topitzes, & Callvert, 2000).

Self-efficacy and self-esteem in particular are frequently emphasized as targets of psychosocial change in youth development and empowerment literature, reinforcing their role as primary outcomes in this review (Anderson & Sandmann, 2009; Jennings et al., 2006; Oliver, Collin, Burns, & Nicholas, 2006; Roth & Brooks-Gunn, 2003; Sinclair, 2000). A number of other intended improvements in protective factors, such as positive connections, social skills, and prosocial behaviors, are also included as secondary outcomes in this review. Because YEPs aim to engage young people in interpersonal activities and participatory processes that exercise multiple aspects of social–emotional functioning, rather than focusing narrowly on addressing a single problem or disorder, multiple developmental outcomes need to be considered in reviewing the impacts of YEPs, as is the case with most complex interventions (Craig et al., 2008).

Defining YEPs

This review defines YEPs as interventions that regularly involve young people as participants in the decision-making processes that determine program design, planning, and/or implementation. With the support of caring adults, YEPs engage young people in program leadership as a characteristic of their involvement in safe, positive, and structured activities. Common examples of YEPs are found in youth councils, teen centers, and community-based participatory research programs. Structurally, this participation within programs often takes the form of advisory councils, committees, or staff positions.

Why It Is Important to Do This Review

The literature on youth empowerment overwhelmingly assumes beneficial effects. YEPs bring youths together, among other aims, to enhance participants’ self esteem, self-efficacy, and social skills. Yet, without clear empirical evidence to support the theory of change, it is possible that YEPs may be ineffective or even harmful. Some research has shown, for example, that programs aggregating youth with high levels of deviant behavior—even if for the purpose of positive interactions—can unintentionally increase antisocial behavior (Dishion, McCord, & Poulin, 1999).

Investigating the impacts of YEPs is all the more important to inform better understanding of the value of these programs in the context of limited economics resources. YEPs generally place a heavy emphasis on human resources, which can drive up intervention costs (Gray & Hayes, 2008). On the other hand, if YEPs are effective in improving social–emotional functioning for better productivity and reducing long-term behavioral problems, cost savings to tax payers could be substantial (Cunha, Heckman, Lochner, & Masterov, 2006; Scott, Knapp, Henderson, & Maughan, 2001). This review works toward a better understanding of the measured contributions of YEPs to youth outcomes.

Method

Eligibility

To be eligible for inclusion in this review, studies must have used either an experimental or quasi-experimental design with a prospectively assigned control group. Quasi-experiments needed to have taken steps to establish a reasonably credible counterfactual using either participant matching or statistical methods (e.g., propensity scores) to ensure between-group similarity at baseline. Control groups could be comprised of no-service, wait-listed or alternative service recipients, provided alternative services did not facilitate youth involvement in program decision-making or active leadership roles (i.e., empowerment components). Basic recreational or educational activities, such as instructional sessions or presentations, games, and informal athletic activities, for example, were considered appropriate comparison services to youth empowerment programming.

Types of Participants

Includable studies must evaluate the effects of YEPs on adolescents. Ages of adolescence, using the World Health Organization, United Nations, and World Bank's definition, include 10 to 19 (United Nations Population Fund [UNFPA], 2008; World Bank, 2003; World Health Organization, 2009). Following recommendations from Campbell Collaboration Social Welfare Group peer reviewers, studies were included only if at least 75% of the sample met the age criterion above.

Types of Interventions

YEPs are defined as intervention programs for youths that regularly involve youths in determining program design, activities, and/or implementation. Empowerment components could include participation in councils, committees, directory boards, workgroups, staff positions, or other such structures providing regular opportunities for engagement in program decision making. YEPs could also incorporate empowerment components through youths and adults serving together in formal leadership capacities or by reserving such roles for youths with adults restricted to supportive roles. All programs must have offered regular access to a supportive adult or older youth leader, though one-to-one mentoring is not necessary. Youth empowerment implies group interaction (Cargo, Grams, Ottoson, Ward, & Green, 2003; Jennings et al., 2006); consequently, exclusively one-to-one youth development interventions, such as most mentoring schemes, were not reviewed.

Delivery could have taken place in community-based or school-based settings but must have occurred outside of formal education and been convened regularly (i.e., not one-off events). Interventions primarily within formal education, juvenile justice, residential programs, therapeutic interventions, conferences, or workshops were not included.

Types of Outcomes

Studies must have measured at least one of the review’s primary or secondary outcomes. The review’s primary outcomes were:

Self-efficacy, including both general self-efficacy and task-specific self-efficacy (e.g., drug avoidance self-efficacy or sexual behavior self-efficacy). General and task-specific self-efficacy measures represent distinct constructs and were therefore considered separately for meta-analysis.

Self-esteem, which is most commonly assessed by the 10-item Rosenberg Self-Esteem scale (Rosenberg, 1989). Some research has delineated between global self-esteem (e.g., measured by the Rosenberg scale) and specific self-esteem (e.g., measured by the area-specific Hare Self-Esteem Scale), but no specific self-esteem outcomes were measured by this review’s included studies.

Secondary outcomes were academic performance, antisocial behavior, and the following additional developmental assets: social supports and connections, social skills, emotional intelligence, and coping and problem-solving skills. The review’s protocol listed prominent example measures for each outcome. Standardized and nonstandardized measures were accepted for primary and secondary outcomes, including self-reports, third-party or researcher observations, interviews, or official records.

Assessment of Study Quality

Study quality was systematically appraised using criteria developed in a previously published systematic review (Zief, Lauver, & Maynard, 2006). These criteria use forty-one characteristics of study design and reporting to appraise studies according to four core standards:

Evidence that there was no significant control group contamination;

Evidence that there was no significant overall study attrition or differential attrition that would bias results;

The use of appropriate statistical measures for analyses; and

The measurement of primary outcomes at follow-up for all available sample members, thereby meeting the qualification of “intention-to treat,” not “treatment-on-treated,” analysis.

The full appraisal checklist is included in Morton and Montgomery (2011). The checklist was used to guide discussion and meaningful analyses of study quality but was not used to determine inclusion in the review nor to score or rank studies.

Electronic Searches

A systematic literature search was conducted to identify published and unpublished studies meeting the review’s inclusion criteria. Twelve major electronic databases were searched: Applied Social Science Index and Abstracts, Australian Educational Index, British Educational Index, CINAHL, Cochrane Library (CENTRAL), Dissertation and Theses Abstracts, EMBASE, ERIC, Medline, PsycInfo, Social Service Abstracts, and Sociological Abstracts. Additional searches were conducted of the 10 institutional web-based publication databases from around the world. Further details on the databases and dates searched are available elsewhere (Morton & Montgomery, 2011).

Search Terms

The following search terms were used for each database, with Booelean operators, wildcard symbols, and field indexes modified depending on the requirements of the specific database.

Population:

(young OR youth* OR child* OR teen* OR adolescen* OR minors OR school ADJ student* OR boy* OR girl* OR NEETs OR NEET OR 14-19).ab,ti.

AND

Intervention:

(pyd OR cyd OR empowerment OR youth ADJ engag* OR volunteerism OR volunteering OR youth ADJ advocacy OR youth ADJ activism OR youth ADJ development OR youth ADJ leader* OR youth ADJ inclusion OR community ADJ service OR after ADJ school OR afterschool OR youth ADJ1 decision-making OR youth ADJ driven OR youth ADJ run OR youth ADJ adult ADJ partnership* OR youth/adult ADJ partnership* OR youth-adult ADJ partnership* OR youth ADJ action OR youth ADJ1 involvement OR youth ADJ participation OR young ADJ people* ADJ participation OR youth ADJ led OR peer ADJ education OR peer ADJ led OR peer ADJ participation OR youth ADJ voice OR service ADJ learning OR youth ADJ council* OR teen ADJ council* OR non-formal ADJ education OR nonformal ADJ education OR informal ADJ education OR teen ADJ cent* OR youth ADJ cent* OR participatory ADJ research).ab,ti.

AND

Methods:

(control* OR random* OR trial* OR effectiveness OR efficacy OR compar* OR clinical* OR experiment* OR impact ADJ evaluation OR impact ADJ study OR impact ADJ assessment OR outcome ADJ evaluation OR outcome ADJ study OR outcome ADJ assessment).af.

Other Resources

In order to explore potentially eligible studies among unpublished as well as published literature, institutions and individuals regarded as professional leaders in the area of youth development and research were contacted directly and asked for any leads on specific studies, or relevant databases. Seventy professionals from over 50 governmental, charitable, and research institutions were contacted, but none of these yielded any additional eligible studies for this review.

Results

A total of 7,985 citations were retrieved from electronic databases and 804 more from institutional web-based databases, yielding a total of 8,789 citations. Sixty-eight studies were retrieved for full-text inspection by both reviewers using a predetermined screening guide (Morton & Montgomery, 2011). Three studies met all of the review’s inclusion criteria; 62 studies were excluded for not meeting inclusion criteria, and 3 studies were excluded due to insufficient data or intervention details. Citations and reasons for exclusion for these 65 studies are detailed in Morton and Montgomery (2011). Figure 1 provides a flow diagram of the review’s search and selection process.

Figure 1.

Review flow diagram.

Description of Studies

The three included studies were evaluations of the Youth Action Research Project (YARP) in the United States by Berg, Coman, and Schensul (2009); the Youth Leadership Program (YLP) in the United States by Olson-Merichko (2006); and the Questscope Non-Formal Education (QS NFE) program in Jordan by Morton and Montgomery (2012).

Two of the included studies used randomized controlled trials but had small sample sizes (N = 127 and N = 40) and were described as pilot studies (Morton & Montgomery, 2012; Olson-Merichko, 2006). Review of the Olson-Merichko (2006) study, which was an unpublished dissertation, was limited to details and data provided in the dissertation, as the author is deceased. The third included study was a peer-reviewed quasi-experimental evaluation with a matched comparison design (N = 316; Berg, Coman, & Schensul, 2009). Further details and data were required to adequately review the study, and these were supplied by the study’s primary author in response to a request from the reviewers.

None of the three studies used a comparison group with an intervention of comparable exposure to assess the impact empowerment-based components. Berg and colleagues (2009) compared the intervention group to a group of young people participating in other summer employment programs, but the nature and dosage of these programs were not assessed. In Morton and Montgomery’s (2012) study, youth randomly assigned to the waitlist control condition were offered a basic biweekly recreational activity that did not use empowerment-based methodology and was conducted at a lesser dosage than the YEP. Olson-Merichko’s (2006) study involved a no-intervention control.

The study samples had similar mean ages (ranging from 15.2 to 16.0) but were otherwise notably heterogeneous in terms of urban versus rural settings, cultural contexts, ethnicity, gender makeup, and life circumstances (e.g., in-school vs. out-of-school). Morton and Montgomery (2012) and Olson-Merichko (2006) conducted posttests at 4 months. Berg and colleagues (2009) conducted a data collection at 3 months that only captured a summer training institute component of the intervention and a 12-month posttest that captured empowerment-based components (youth-led projects). All references to Berg and colleagues’ findings in this review refer to findings from the 12-month posttest unless otherwise indicated.

Two of the evaluated programs, YARP and YLP were based on participatory action research models in the United States (Berg et al., 2009; Olson-Merichko, 2006). A substantial amount of youth empowerment literature in recent years has focused on participatory research as a means for engaging young people in programs, schools, and communities (Kirby, 2004; Ozer et al., 2008; Suleiman, Soleimanpour, & London, 2006; Worrall, 2000). The third program, QS NFE, involved an empowerment-based nonformal education model for out-of-school youth in Jordan (Morton & Montgomery, 2012). Nonformal education is often associated with empowerment approaches to working with marginalized populations through participatory learning (Castelloe & Watson, 1999; Moulton, 1997), which comprised a central aspect of the QS NFE theory of change.

The three programs differed in structure, context, and content; basic characteristics are outlined in Table 1. . QS NFE is a 24-month intervention, but only the first 4 months of the program were evaluated. YARP is a 10.5-month intervention, but data collections were conducted at 3, 6, and 12 months. The YLP is a 4-month intervention with a post-test at the end. All three interventions involved weekly programming with minimum exposure ranging from 2 to 4 hours per week for most portions of the interventions.

Table 1.

Program Characteristics

Program name	Evaluators	Location	Description	Duration and evaluation follow-ups	Weekly exposure	Youth empowerment components
Program name	Evaluators	Location	Description	Duration and evaluation follow-ups	Weekly exposure	Decision making	Adult involvement	Skill building
Youth Action Research Project (YARP)	Berg, Coman, and Schensul (2009)	United States (urban)	Participatory action research and community advocacy with training in one community site	11 months: 7-week summer institute, 8-month projects (3-, 6-, 12-month follow-ups)	20 hours per week (summer), 4 hours per week (school year)	Youth-initiated shared decisions with adults; Youth designed, led and implemented projects	Researchers and trained project staff offering youth research training, supporting youth-led projects, facilitating discussions	7-week institute with manualized training curriculum to develop participatory research skills; participatory research activities
Youth Leadership Program (YLP)	Olson-Merichko (2006)	United States (rural)	Participatory research with training in one school site	4 months (4-month follow-up)	2–3 hours per week	Youth-initiated shared decisions with adults; Youth designed, led and implemented projects	Lead researcher offering training to youth leaders and supporting youth-led projects	25-hour leadership training curriculum for youth team leaders only (not for study participants); participatory research activities
Questscope Non-Formal Education (QS NFE)	Morton and Montgomery (2012)	Jordan (urban)	Participatory nonformal education in multiple school and community sites	24 months: three 8-month cycles (4-month follow-up)	4-10 hours per week	Adult-initiated, shared decisions with youth; Youth chose learning topics and activities, with some leadership and project planning	Retrained teachers facilitating youth participation in determining program activities and topics	24-hour leadership and research training curriculum for youth on advisory council only (not for study participants); participatory social, cultural, and educational activities

Of the three programs, the YARP intervention evaluated by Berg and colleagues represented the most thorough application of the youth empowerment process theory of change. YARP was the only program to include all of the following components: an intensive and manualized 7-week capacity-building institute, constant adult facilitator presence, emphasis on youth–adult partnerships in both short-term decisions and ongoing youth-led projects, emphasis on collective action, and 10-month intervention exposure.

There were several other notable differences between the three programs and evaluations included in this review. Youths in YLP and QS NFE did not receive financial compensation for their participation while YARP participants were employed and financially compensated. Youth participation in YLP and YARP was centered on explicit social action projects involving school or community advocacy while these were not formal aspects of the QS NFE intervention model. YARP and YLP were both interventions designed and implemented by the respective researchers. QS NFE is an ongoing program jointly led by a nongovernmental organization, Questscope, and the Jordanian Ministry of Education; the study authors were unaffiliated with the program design and implementation.

Youth Participation in Decision-Making

The YLP and YARP interventions were developed specifically for the purpose of testing the impacts of youth empowerment. As such, these programs facilitated not only a high level of youth involvement in program decision making but also in initiating and planning activities and projects. Youth participants were responsible for designing and implementing research projects, activities, and meetings, with adult facilitators acting as supporters and guides throughout the implementation process. In terms of Hart’s Ladder of Children’s Participation (Hart, 1992), empowerment generally ranged from “adult-initiated, shared decisions with children” with respect to program initiation and training implementation, to “child-initiated, shared decisions with adults” with respect to the research and community action activities.

The QS NFE program used participatory methodology that involved youth in determining learning topics and social or recreational activities and stipulated regular involvement of young people in daily program decision making and co-learning relationships between youth and adults. While some QS NFE sites included longer term youth leadership roles and youth-led planning, the intervention differed from YARP and YLP in not formally requiring this level of participation. In terms of Hart’s Ladder, QS NFE involved adult-initiated programming with shared decisions with adolescents. The associated process study suggests that in sites with lower implementation fidelity, QS NFE programming slipped toward “consulted and informed” levels of youth participation.

All three studies included qualitative research to study intervention implementation. Only Morton and Montgomery (2012) also used a quantitative instrument to measure the extent to which participants felt empowered by the program process.

Adult Involvement

All three studies described adults involved in the programs as “facilitators” whose roles involved facilitating youth participation and contribution rather than controlling or directing youth involvement. The YLP’s adult involvement consisted primarily of the lead researcher training youth leaders and supporting youth-led projects. In YARP, researchers and trained project staff both led youth research training, supported youth-led projects, and facilitated reflective discussions. In contrast, QS NFE identified and retrained teachers from the formal education system to serve as facilitators for the program outside of school hours.

The task of implementing quality youth empowerment processes can require special skills and competencies. This is especially true where YEPs are delivered to marginalized youth who have had few previous empowering experiences. Youth empowerment literature has emphasized the importance of adult training to help staff or volunteers developing facilitation and youth development skills for empowerment processes and overcome any conflicting deficit-based inclinations toward youth (Jennings et al., 2006). To this end, facilitators in YARP and QS NFE received structured training specifically in empowerment methodologies prior to intervention implementation. The YLP did not include an adult training component in empowerment methodology, but youth team leaders who led project teams participated in 25 hours of leadership skills training.

The YLP gave a particularly high level of autonomy to participants. Youth “team leaders” were selected and trained to lead youth groups, which then met without adults to discuss and implement projects. QS NFE involved a higher level of adult control in program planning and implementation, partly in order to meet educational curriculum standards established by adult officials. Among the three programs, YARP most closely approximated Wong and colleagues’ (2010) “pluralistic” standard of youth–adult shared control, though all three programs stressed shared control to varying degrees.

Skill building

YARP provided a distinct training component that engaged all program youth in a 7-week, 20-hour per week summer institute focused on developing participatory research skills. YLP involved 25 hours of leadership training for youth team leaders. QS NFE delivered 24 hours of leadership and research training for youth on the program advisory council, but these youths were not included in the RCT samples.

All three intervention descriptions stressed the development of social and leadership skills through social activities and empowerment-based methodologies, but only YARP included a defined a curriculum for these. In both YARP and YLP, skill building primarily constituted of involvement in ongoing team research project activities. QS NFE used educational games and cultural, vocational, or recreational activities that varied according to program site and youth preferences. Unlike the other two programs, QS NFE did not involve youths in ongoing team-based projects.

YEP literature suggests an important role for structured, ongoing training components designed to prepare young people for meaningful participation in empowering program processes, specific tasks (e.g., participatory research, photography, etc.), and broader civic engagement (Jennings et al., 2006; Wilson et al., 2006). The absence of such components in QS NFE and YLP may have limited these programs’ ability to develop young people’s skills and fully engage them in participatory opportunities.

Risk of Bias in Included Studies

Allocation

The Morton and Montgomery (2012) and Olson-Merichko (2006) randomized studies of QS NFE and YLP (respectively) reported comparable groups at baseline on demographic and dependent variables, indicating successful randomization. Dependent variables in both studies included a wide range of outcome measures. Demographic variables included gender, age, and working status in both studies. Olson-Merichko (2006) reported similarity on several additional variables (e.g., race, parent’s education, and household income). Berg and colleagues (2009) used a quasi-experimental design to create two groups matched on demographic variables. However, there were significant baseline differences between the groups on several intended outcome measures. Berg and colleagues attempted to statistically adjust for these baseline differences by treating outcome measures as covariates in their analyses. However, this method is problematic because the variance in an observed dependent variable accounted for by a covariate is not likely to be independent of the variance in the dependent variable accounted for by between-group variance (Miller & Chapman, 2001). There is thus a risk that Berg and colleagues’ findings were affected by selection bias.

Blinding

No blinding to trial arm membership was reported for randomization or assessment for any of the included studies for allocation or outcomes assessment. Morton and Montgomery (2012) utilized computer-based randomization and data collection programs to minimize opportunities for bias in the absence of blinding.

Attrition and Missing Data

Morton and Montgomery (2012) and Olson-Merichko (2006) had low attrition rates (6.3% and 2.5%, respectively). The latter did not use intention-to-treat (ITT) analysis, but the low attrition makes it unlikely that ITT analyses would have made any difference to the results. Berg and colleagues (2009) reported an attrition rate of 26% (author contact). While Berg and colleagues found no significant differences on demographic variables between completers and noncompleters, there is a real risk that noncompleters differed from completers in their response to the intervention or on other unobserved characteristics. Berg and colleagues’ study was therefore not included with the other two studies in this review’s meta-analysis.

Selective Reporting

Berg and colleagues (2009) only reported statistics for outcomes with statistically significant intervention effects in the published paper but made full data for all other outcomes readily available. Olson-Merichko (2006) did not report statistics for antisocial behavior outcomes, which did not show statistically significant intervention effects. Only Morton and Montgomery (2012) published a protocol prior to trial commencement prospectively stating outcomes to be measured.

Other Potential Sources of Bias

There was a risk of potential contamination in two of the three studies. In Olson-Merchiko’s study, the risk of contamination arose from intervention and control participants attending the same school throughout the intervention. Berg and colleagues explicitly report that control participants took part in alternative summer employment programs that may have integrated empowerment-based programming components. Unfortunately, despite these risks, neither Berg and colleagues (2009) nor Olson-Merichko (2006) reported using any measures taken to assess whether contamination had occurred.

The risk of contamination was far lower in Morton and Montgomery’s (2012) QS NFE evaluation, where the QS NFE target population (out-of-school youths) was spread across several communities and siblings were block randomized. In addition, the risk of contamination was assessed by documenting control group participants’ activities, and no evidence of contamination was found.

Effects of Intervention

None of the included studies could provide evidence about whether YEPs impact self-esteem. Olson-Merichko (2006) and Morton and Montgomery (2012) did not include outcome measures for self-esteem. Berg and colleagues (2009) did so, but treated self-esteem as a covariate rather than an outcome in their analyses because of significant baseline differences between the intervention and control groups.

All three studies included measures of self-efficacy outcomes. Olson-Merichko (2006) and Morton and Montgomery (2012) used the same 10-item measure for general self-efficacy (Schwarzer & Jerusalem, 1995). Berg and colleagues (2009) only included task-specific self-efficacy measures for drug avoidance and sexual behavior. None of the three studies found statistically significant differences in self-esteem outcomes between intervention and control group participants.

General self-efficacy was the only primary outcome measured by more than one included evaluation and was therefore the only outcome that was meta-analyzed. Meta-analysis of data from Morton and Montgomery (2012) and Olson-Merichko (2006) did not show a combined intervention effect on self-efficacy (z = 1.21 95% CI [−0.12, 0.49]). Meta-analysis data and a forest plot using a random-effects model are given in Figure 2.

Figure 2.

Meta-analysis of general self-efficacy results.

Given the limited sample sizes and includable studies, the results of this meta-analysis should not be interpreted as an authoritative statement on the effects of YEPs. The results simply reflect a small number of impact studies’ aggregate effects on self-efficacy. While some secondary outcome areas were assessed by more than one included study (social supports, social skills, and problem areas), these were not meta-analyzed because of the level of heterogeneity among constructs captured by the different measures.

Independent results for all of the review’s primary and secondary outcomes measured by the included studies are displayed in Tables 2 and 3 (adaptations of table templates used by the Zief et al (2006) review). Berg and colleagues and Olson-Merichko conducted one-tailed analyses to test for intervention effects. Because this review accommodates the possibility of unintended adverse outcomes, significance levels were recalculated using two-tailed tests. Although the tables present relatively little data given the small number of includable studies, future updates of the review could follow the template with additional material.

Table 2.

Independent Study Results for Primary Outcomes

Outcome	Number of studies (Combined N)	Number of measured outcomes	Outcomes favoring intervention		Outcomes favoring comparison		Null effects (p >.10)
Outcome	Number of studies (Combined N)	Number of measured outcomes	p ≤ .05	.05 < p ≤ .10	p ≤ .05	.05 < p ≤ .10	Number of studies	%
General self-efficacy^a	2 (n = 167)	2	0	0	0	0	2	100
Specific self-efficacy^b	1 (n = 316)	2	0	0	0	0	2	100
Self-esteem^c	1 (n = 316)	1	0	0	0	0	1	100

^a Olson-Merichko (2006) and Morton and Montgomery (2012)

^b Berg et al. (2009): drug prevention and sexual behavior. ^c Berg et al. (2009) (had baseline differences, reported here but otherwise not treated as an outcome in the primary study or by this review).

*p < .10. **p < .05. ***p < .01 intervention effect.

Table 3.

Independent Study Results for Secondary Outcomes

Outcome	Number of studies (Combined N)	Number of measured outcomes	Outcomes favoring intervention		Outcomes favoring comparison		Null effects (p >.10)
Outcome	Number of studies (Combined N)	Number of measured outcomes	p ≤ .05	.05 < p ≤ .10	p ≤ .05	.05 < p ≤ .10	Number of studies	%
Social supports and connections^a	2 (n = 443)	5	0	0	0	0	4	80
Social skills^b	3 (n = 483)	5	1	0	0	0	2	66
Emotional intelligence	0	0	0	0	0	0	0	n/a
Coping and problem-solving skills^c	1 (n = 40)	1	1	0	0	0	0	0
Civic engagement	0	0	0	0	0	0	0	n/a
Academic performance	0	0	0	0	0	0	0	n/a
Problem behavior^d	3 (n = 483)	8	3	1	0	0	4	50

^a Berg, Coman, and Schensul (2009): social connectedness and school bonding (latter had baseline differences, not treated as outcome); Morton and Montgomery (2012): social supports of friends, social supports of family, and adult connectedness. ^b Berg et al. (2009): social assertiveness skills and social skills; Olson-Merichko (2006): team skills**; Morton and Montgomery (2012): social skills and prosocial attitude. ^c Olson-Merichko (2006): proactive coping**. ^d Berg et al (2009): alcohol use*, marijuana use**, had sex, and number of sex partners**; Olson-Merichko (2006): delinquency, drug use, and alcohol use; Morton and Montgomery (2012): conduct problems**.

*p < .10. **p < .05. ***p < .01 intervention effect.

Discussion and Applications to Practice

A thorough and systematic search of the international literature found that few evaluations met this review’s inclusion criteria. Of the three evaluations that did meet inclusion criteria, none demonstrated statistically significant intervention effects on this review’s primary outcomes: self-efficacy and self-esteem. All three included studies assessed self-efficacy, but none showed significant intervention effects compared to control groups at posttest. Differences were insignificant for both general (Morton & Montgomery, 2012; Olson-Merichko, 2006) and task-specific (Berg et al., 2009) measures of self-efficacy. Results for general self-efficacy from the two experimental evaluations were combinable and meta-analyzed; even with the increased power yielded from combined samples, no significant intervention effects were found.

None of the three included studies measured self-esteem outcomes. Therefore, despite the considerable amount of literature and institutions promoting YEPs as a means to develop positive attitudes and behaviors in youths, this review concludes that there is insufficient empirical evidence to support the claim.

All three included studies measured multiple secondary outcomes. While some statistically significant intervention effects were found, in each of the studies these were outnumbered by null effects. On the other hand, no study showed evidence of harm nor were there any studies that failed to produce at least one statistically significant, positive intervention effect for the review’s secondary outcomes—including social skills (team skills), coping skills (proactive coping), and problem behaviors (conduct problems, marijuana use, and number of sex partners). Given the large number of outcomes measured by each study, however, some of these significant outcomes are likely to have occurred by chance alone (Feise, 2002).

Overall Completeness and Applicability of the Evidence

In general, the review demonstrates a paucity of evidence from high-quality impact studies of YEPs outside of formal education contexts. The lack of intervention effects on the primary outcomes reported by included studies might be attributable to several factors. For example, low attendance rates or inadequate dosage could have diluted program effects. In the QS NFE Program in Jordan, for instance, 52% of study participants attended less than the minimum amount expected (2 days per week). Intervention duration in the included studies ranged from 4 to 10 months. Longer program durations and better implementation strategies to increase participation might provide the levels of program exposure that could be needed to achieve measurable impacts, particularly on adolescents considered at higher risk.

Alternatively, it is conceivable that, given the disadvantaged backgrounds of many young people in the trial samples, more targeted, evidence-based therapeutic programs could be needed prior to or along with YEPs for the programming to have an impact. In other words, if adverse, traumatic experiences, such as exposure to neglect, abuse, or household dysfunction, have significantly impaired young people’s social–emotional functioning, they may require more intensive interventions to improve basic functioning in order to fully benefit from YEPs, which emphasize broader developmental goals (Silverman et al., 2008). However, it is also possible that YEPs in the forms represented by included studies are largely ineffective in changing social and emotional youth outcomes. None of these possible explanations, however, can be settled without a more complete evidence base.

Quality of the Evidence

Only three studies met all of the review’s inclusion criteria; these used mixed levels of methodological quality and relatively small sample sizes. The two randomized controlled trials (Morton & Montgomery, 2012; Olson-Merichko, 2006) had fairly short posttest periods at 4-month follow-up, and only Morton and Montgomery’s had gone through peer review. The third study evaluated a more intensive youth empowerment model, with longer participatory experiences and more structured leadership training (Berg et al., 2009). The posttest period was also longer, with data collected for up to 12 months. Nonetheless, the nonrandomized design was more susceptible to bias.

Only one of the three studies (Morton & Montgomery, 2012) had published a protocol prior to recruitment. Neither of the other two studies reported data for outcomes that yielded nonsignificant findings. Although evaluators were willingly provided such unreported data when contacted by reviewers, this form of reporting bias can skew the public’s understanding of the full effects of an intervention (Smyth et al., 2011).

Potential Biases in the Review Process

Youth development programs frequently aim to effect changes across multiple behavioral, social, and attitudinal outcomes. Self-esteem and self-efficacy were selected as primary outcomes for this review because they are emphasized in the youth empowerment literature. However, such emphasis could be seen as a form of bias resulting in the review’s conclusions being restricted to narrow set out of measures.

A broader set of secondary outcomes was included and reviewed to mitigate this risk. Additionally, no study was excluded based on its outcomes measures, and the search strategy did not include terms related to outcomes. Consequently, although the review focused its analyses and discussion on those outcomes most strongly associated with youth empowerment, the authors’ selection of primary and secondary outcomes did not constrain the number of studies that were included in this review. The primary reason for the small number of includable studies was the absence of studies meeting both the review’s intervention criteria (youth programs outside of formal education that regularly involved participants in program decision-making processes) and study methods criteria (experimental or controlled quasi-experimental designs).

The search for gray literature through professional outreach was limited to contacts in the authors’ networks or those that the authors could identify from the literature. It is possible that other groups have conducted includable but unpublished studies. Additionally, the inclusion of “youth empowerment programs” is susceptible to interpretation because of challenges in determining what constitutes youth empowerment. Nevertheless, the authors attempted to establish objective and inclusive intervention inclusion criteria based on regular participation of youths in program decision making. Moreover, both authors independently screened the studies, and there were no disagreements related to program definitions.

Implications for Practice

The review demonstrates an insufficient evidence-base for YEPs’ impact on self-efficacy and self-esteem. As such, the authors are unable to offer definitive conclusions about the impacts of YEPs. While the few includable studies do not show positive intervention effects on these primary outcomes, there is also no evidence of harm from their outcomes data. There is limited evidence for intervention effects on the review’s secondary outcomes (e.g., social skills and antisocial behavior), suggesting a potentially important role for youth empowerment in changing these outcomes. However, further research is needed because the significance of findings with regard to secondary outcomes was inconsistent across the three studies.

The reviewers could not identify sufficient data to assess the extent to which differences or similarities between YEPs, and their target populations, help to explain outcomes. As expected, there was heterogeneity between YEPs evaluated by the included studies with respect to program activities, the extent of youth participation in program decision making, the nature of adults’ roles in the program, the amount of program exposure, and the characteristics of the study sample. As required, all three programs shared a commitment to regular involvement of young people in program decision making as an aspect of their intervention experience, samples of similar age groups, a supportive adult presence, and asset-building activities intended to build on young people’s strengths.

It remains to be seen whether different degrees of youth participation in program leadership and decision making explain different levels of program effects on youth development outcomes, and for different populations of youth. Exploratory analysis conducted by Morton and Montgomery (2012) did indicate a positive correlation between program centers in which young people reported higher experiences of empowerment and social–emotional outcomes, which suggests some promise for youth empowerment programming if implemented with high fidelity. These findings, however, are highly tentative without further investigation with more robust designs and larger sample sizes to assess the effects of implementation quality in YEPs. Moreover, youth empowerment has varying intervention implications across different cultural contexts; these too should be explored in future studies.

In order to better understand the effects of YEPs in the context of different implementation and cultural circumstances, it is essential that future research integrates carefully coordinated mixed methods that assess process and implementation factors with experimental and quasi-experimental designs, as the Medical Research Council’s Framework for Developing and Evaluating Complex Interventions recommends (Craig et al., 2008). Further, for future research to be most useful, and for programs to increase their likelihood of demonstrating positive impacts, YEP developers, including youth and adult stakeholders, should be careful to articulate very clear theories of change with which YEP models are designed, implemented, and evaluated (Hardeman et al., 2005; Weiss, 1995).

Conclusion

This is the first known review of the effects of YEPs. The review findings agree with previously stated expectations in the literature that very little evidence on youth empowerment interventions has been generated through high-quality impact evaluation (Crowley & Skeels, 2010; Gray & Hayes, 2008; Zeldin et al., 2000). This review has confirmed the validity of this concern by establishing that few high-quality evaluations have been conducted. Given the shortage of rigorous impact evaluation in this area of intervention, this review’s findings should stimulate further research investment and action to improve understanding of whether, how, and in what contexts YEPs impact social–emotional well-being of adolescents.

Footnotes

Authors’ Notes

Expert support was provided by the Centre for Evidence-Based Intervention, University of Oxford. The authors have no vested interest in the outcomes of this review nor any incentive to represent findings in a biased manner. Both authors jointly authored an experimental study of a youth empowerment program in Jordan, which is included in this review.

Acknowledgements

The inspiration for and development of this review are due to many colleagues, both researchers and practitioners. In particular, we thank our colleagues affiliated with the Centre for Evidence-Based Intervention at the University of Oxford for their methodological and theoretical guidance. Many thanks, also, to the professional contacts who provided resources and support throughout the review process.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Editor’s Note

This article was invited and accepted by the Editor. It is based upon a systematic review published in the Campbell Collaboration Library, found at .

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Akers

R. L.

Krohn

Lanza-Kaduce

Radosevich

(1979). Social learning and deviant behavior: A specific test of a general theory. American Sociological Review, 44, 636–655.

Anderson

K. S.

Sandmann

(2009). Toward a model of empowering practices in youth-adult partnerships. Journal of Extension, 47, 1–8.

Aspy

Oman

Vesely

McLeroy

Rodine

Marshall

(2004). Adolescent violence: The protective effects of youth assets. Journal of Counseling & Development, 82, 268–276.

Bandura

(1993). Perceived self-efficacy in cognitive development and functioning. Educational Psychologist, 28, 117–148.

Berg

Coman

Schensul

(2009). Youth action research for prevention: A multi-level intervention designed to increase efficacy and empowerment among urban youth. American Journal of Community Psychology, 43, 345–359. doi:10.1007/s10464-009-9231-2

Cargo

Grams

Ottoson

Ward

Green

(2003). Empowerment as fostering positive youth development and citizenship. American Journal of Health Behavior, 27, S66–S79.

Castelloe

Watson

(1999). Participatory education as a community practice method. Journal of Community Practice, 6, 71–89.

Craig

Dieppe

Macintyre

Michie

Nazareth

Petticrew

(2008). Developing and evaluating complex interventions: The new Medical Research Council guidance. British Medical Journal, 337, 979–983. doi:10.1136/bmj.a1655

Crowley

Skeels

(2010). Getting the measure of children and young people's participation. In Percy-Smith

Thomas

(Eds.), A handbook of children and young people's participation: Perspectives from theory and practice (pp. 184–192). Abingdon, UK: Routledge.

10.

Cunha

Heckman

Lochner

Masterov

(2006). Interpreting the evidence on life cycle skill formation. In Hanushek

Welch

(Eds.), Handbook of the economics of education (Vol. I, pp. 697–812). Amsterdam, The Netherlands: Elsevier.

11.

Dishion

T. J.

McCord

Poulin

(1999). When interventions harm: Peer groups and problem behavior. American Psychologist, 54, 755–764. doi:10.1037/0003-066x.54.9.755

12.

Dumont

Provost

M. A.

(1999). Resilience in adolescents: Protective role of social support, coping strategies, self-esteem, and social activities on experience of stress and depression. Journal of Youth and Adolescence, 28, 343–363. doi:10.1023/a:1021637011732

13.

Feise

R. J.

(2002). Do multiple outcome measures require p-value adjustment? BMC Medical Research Methodology, 2, 8.

14.

Flannery

D. J.

Singer

M. I.

Wester

(2001). Violence exposure, psychological trauma, and suicide risk in a community sample of dangerously violent adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 40, 435–442. doi:10.1097/00004583-200104000-00012

15.

Gray

Hayes

(2008). Understanding the state of knowledge on youth engagement financing and sustainability. Washington, DC: The Finance Project.

16.

Hardeman

Sutton

Griffin

Johnston

White

Wareham

N. J.

Kinmonth

A. L.

(2005). A causal modelling approach to the development of theory-based behaviour change programmes for trial evaluation. Health Education Research, 20, 676–687. doi:10.1093/her/cyh022

17.

Hart

(1992). Children's participation from tokenism to citizenship. Florence, Italy: UNICEF Innocenti Research Centre.

18.

Jennings

Parra-Medina

Messias

McLoughlin

(2006). Toward a critical social theory of youth empowerment. Journal of Community Practice, 14, 31–35.

19.

Jessor

Van Den Bos

Vanderryn

Costa

F. M.

Turbin

M. S.

(1995). Protective factors in adolescent problem behavior: Moderator effects and developmental change. Developmental Psychology, 31, 923–933. doi:10.1037/0012-1649.31.6.923

20.

Kia-Keating

Dowdy

Morgan

M. L.

Noam

G. G.

(2011). Protecting and promoting: An integrative conceptual model for healthy development of adolescents. Journal of Adolescent Health, 48, 220–228. doi:10.1016/j.jadohealth.2010.08.006

21.

Kirby

(2004). A guide to actively involving young people in research: For researchers, research commissioners, and managers. London, England: INVOLVE Support Unit.

22.

Loeber

(1990). Development and risk factors of juvenile antisocial behavior and delinquency. Clinical Psychology Review, 10, 1–41. doi:10.1016/0272-7358(90)90105-j

23.

Miller

G. A.

Chapman

J. P.

(2001). Misunderstanding analysis of covariance. Journal of Abnormal Psychology, 110, 40–48.

24.

Morton

Montgomery

(2011). Youth empowerment programs for improving self-efficacy and self-esteem of adolescents. Campbell Systematic Reviews. Retrieved from http://campbellcollaboration.org/lib/project/114/

25.

Morton

Montgomery

(2012). Empowerment-based non-formal education for Arab youth: A pilot randomized trial. Children and Youth Services Review, 34, 417–425. doi:10.1016/j.childyouth.2011.11.013

26.

Moulton

(1997). Formal and nonformal education and empowered behavior: A review of the research literature. Washington, DC: Support for Analysis and Research in Africa (SARA), Academy for Educational Development.

27.

Oliver

K. G.

Collin

Burns

Nicholas

(2006). Building resilience in young people through meaningful participation. Australian e-Journal for the Advancement of Mental Health, 5, 1–7.

28.

Olson-Merichko

J. A.

(2006). Youth leadership program evaluation: Implications for implementation (Dissertation). Indiana University of Pennsylvania, Indiana, PA.

29.

Oman

R. F.

Vesely

Aspy

C. B.

McLeroy

K. R.

Rodine

Marshall

(2004). The potential protective effect of youth assets on adolescent alcohol and drug use. American Journal of Public Health, 94, 1425–1430. doi:10.2105/ajph.94.8.1425

30.

Ozer

Cantor

Cruz

Fox

Hubbard

Moret

(2008). The diffusion of youth-led participatory research in urban schools: The role of the prevention support system in implementation and sustainability. American Journal of Community Psychology, 41, 278–289.

31.

Rosenberg

(1989). Society and the adolescent self-image. Middletown, CT: Wesleyan University Press.

32.

Roth

J. L.

Brooks-Gunn

(2003). What is a youth development program? Identification of defining principles. In Jacobs

Wertlieb

Lerner

R. M.

(Eds.), Enhancing the life chances of youth and families: Contributions of programs, policies, and service systems (Vol. 2, pp. 197–223). Thousand Oaks, CA: Sage.

33.

Rutter

(1987). Psychosocial resilience and protective mechanisms. American Journal of Orthopsychiatry, 57, 316–331. doi:10.1111/j.1939-0025.1987.tb03541.x

34.

Schwarzer

Jerusalem

(1995). Generalized Self-Efficacy scale. In Weinman

Wright

Johnston

(Eds.), Measures in health psychology: A user's portfolio. Causal and control beliefs (pp. 35–37). Windsor, UK: Nfer-Nelson.

35.

Scott

Knapp

Henderson

Maughan

(2001). Financial cost of social exclusion: follow up study of antisocial children into. British Medical Journal, 323, 191–196. doi:10.1136/bmj.323.7306.191

36.

Silverman

W. K.

Ortiz

C. D.

Viswesvaran

Burns

B. J.

Kolko

D. J.

Putnam

F. W.

Amaya-Jackson

(2008). Evidence-based psychosocial treatments for children and adolescents exposed to traumatic events. Journal of Clinical Child & Adolescent Psychology, 37, 156–183. doi:10.1080/15374410701818293

37.

Sinclair

(2000). Young people's participation research practice briefings. Staffordshire, England: Making Research Count, Keele University.

38.

Smyth

R. M. D.

Kirkham

J. J.

Jacoby

Altman

D. G.

Gamble

Williamson

P. R.

(2011). Frequency and reasons for outcome reporting bias in clinical trials: Interviews with trialists. British Medical Journal, 342, 7153–7165 doi:10.1136/bmj.c7153

39.

Suleiman

A. B.

Soleimanpour

London

(2006). Youth action for health through youth-led research. Journal of Community Practice, 14, 125–145.

40.

The Nuffield Foundation. (2009). Time trends in adolescent well-being: Update 2009 Changing Adolescence Programme. London, England: The Nuffield Foundation.

41.

UNFPA. (2008). Supporting adolescents and youth. Retrieved February 1, 2010, from http://www.unfpa.org/adolescents/overview.htm

42.

Weiss

C. H.

(1995). Nothing as practical as good theory: Exploring theory-based evaluation for comprehensive community initiatives for children and families. In Connell

J. P.

Kubisch

A. C.

Schorr

L. B.

Weiss

C. H.

(Eds.), New approaches to evaluating community initiatives: Concepts, methods, and contexts (pp. 65–92). Washington, DC: Aspen Institute.

43.

Wilson

Minkler

Dasho

Carrillo

Wallerstein

Garcia

(2006). Training students as facilitators in the youth empowerment strategies (YES!) Project. Journal of Community Practice, 14, 201–217.

44.

Wong

Zimmerman

Parker

(2010). A typology of youth participation and empowerment for child and adolescent health promotion. American Journal of Community Psychology, 46, 100–114. doi:10.1007/s10464-010-9330-0

45.

World Bank. (2003). Adolescent nutrition. Retrieved February 1, 2010, from http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTHEALTHNUTRITIONANDPOPULATION/EXTPHAAG/0,,contentMDK:20587649~menuPK:1314796~pagePK:64229817~piPK:64229743~theSitePK:672263,00.html

46.

World Health Organization. (2009). Adolescent health and development. Retrieved February 1, 2010, from http://www.searo.who.int/EN/Section13/Section1245_4980.htm

47.

Worrall

(2000). Young people as researchers: A learning and resource pack. London, England: Save the Children UK and Joseph Rowntree Foundation.

48.

Zeldin

McDaniel

Topitzes

Callvert

(2000). Youth in decision making: A study on the impacts of youth on adults and organizations. Washington, DC: National 4H Council.

49.

Zief

Lauver

Maynard

(2006). Impacts of after-school programs on student outcomes: A systematic review for the Campbell Collaboration. Retrieved from http://campbellcollaboration.org/lib/project/12/

50.

Zimmerman

(2000). Empowerment theory: Psychological, organizational and community levels of analysis. In Rappaport

Seidman

(Eds.), Handbook of community psychology (pp. 49–64). New York, NY: Kluwer Academic/Plenum Publishers.