View in EDS HTML Full Text PDF Full Text

The Effectiveness of Interventions for Youth That Activate the Social Network: A Meta-Analytic Study

Saved in:
Bibliographic Details
Title: The Effectiveness of Interventions for Youth That Activate the Social Network: A Meta-Analytic Study
Language: English
Authors: Natasha Koper, Roos M. van der Heijden, Sophie Donk, Thao Kieu, Hanneke E. Creemers, Levi van Dam, Susan Branje (ORCID 0000-0002-9999-5313), Geert Jan J. M. Stams
Source: Applied Developmental Science. 2025 29(3):195-219.
Availability: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
Peer Reviewed: Y
Page Count: 25
Publication Date: 2025
Document Type: Journal Articles
Information Analyses
Descriptors: Literature Reviews, Youth, Young Adults, Intervention, Social Networks, Social Support Groups, Program Descriptions, Program Effectiveness, Educational Research, Individual Characteristics, Evaluation Methods
DOI: 10.1080/10888691.2024.2317714
ISSN: 1088-8691
1532-480X
Abstract: This meta-analysis aimed to examine the effectiveness of interventions for youth that activate the social network for improving youth outcomes (e.g. psychological problems, child safety). A literature search yielded 37 studies with 35 independent samples (N = 712,269) of youth aged 0-26 years (M = 7.20), and 409 effect sizes. A three-level meta-analysis controlling for the dependency among effect sizes within studies showed no overall effect of interventions activating the social network (d = 0.11, p = 0.241). Yet, moderator analyses revealed positive effects for youth-initiated mentoring interventions (d = 0.46), youth deciding who to involve (d = 0.52), interventions that involve only one person (d = 0.56), European samples (d = 0.40), interventions targeting youth with mental health needs (d = 0.75), data retrieved through questionnaires (d = 0.10) and official records (d = 0.14), assessments completed by professionals (d = 0.34) or parents (d = 0.17), and outcomes that were corrected for pretest differences between conditions (d = 0.27). This meta-analysis demonstrates that social network activation matters for intervention effectiveness under specific conditions.
Abstractor: As Provided
Entry Date: 2026
Accession Number: EJ1502924
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwGfvp0MncV806PEfILxaOgCAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDEaIKJIFN2XK4QIS-gIBEICBmvMK6h_SD1iPH0g9mJTKj5Po2qQtUJsFd6Hss2-vj1kXVY2SrOvJmWeBlss63A2dgOE6oB7VkQakm7PQfykH4xtJXR2mE7rqyTJiIeYLELbpoVUpwrOFb-Dr6wPmBVUlOHYrLXbmjsjXlCiTxrF85iTDS9FGWSj7FApnVm927SH-rYGcQeXWStAL24F6BRL0Z_DUZXZLzPy95NU=
Text:
  Availability: 1
  Value: <anid>AN0186283553;7lf01jul.25;2025Jul02.03:21;v2.2.500</anid> <title id="AN0186283553-1">The effectiveness of interventions for youth that activate the social network: a meta-analytic study </title> <p>This meta-analysis aimed to examine the effectiveness of interventions for youth that activate the social network for improving youth outcomes (e.g. psychological problems, child safety). A literature search yielded 37 studies with 35 independent samples (N = 712,269) of youth aged 0-26 years (M = 7.20), and 409 effect sizes. A three-level meta-analysis controlling for the dependency among effect sizes within studies showed no overall effect of interventions activating the social network (d = 0.11, p =.241). Yet, moderator analyses revealed positive effects for youth-initiated mentoring interventions (d = 0.46), youth deciding who to involve (d = 0.52), interventions that involve only one person (d = 0.56), European samples (d = 0.40), interventions targeting youth with mental health needs (d = 0.75), data retrieved through questionnaires (d = 0.10) and official records (d = 0.14), assessments completed by professionals (d = 0.34) or parents (d = 0.17), and outcomes that were corrected for pretest differences between conditions (d = 0.27). This meta-analysis demonstrates that social network activation matters for intervention effectiveness under specific conditions.</p> <p>Keywords: meta-analysis; social network activation; intervention effectiveness; youth outcomes</p> <p>Childhood adversity and mental health problems are associated with various adverse outcomes, including substance abuse, low quality of life and suicidality during adulthood (e.g. Brown et al., [<reflink idref="bib13" id="ref1">13</reflink>]; Lee et al., [<reflink idref="bib59" id="ref2">59</reflink>]; McLaughlin et al., [<reflink idref="bib67" id="ref3">67</reflink>]; Simon et al., [<reflink idref="bib94" id="ref4">94</reflink>]). Psychological care is considered the primary resource to help prevent and reduce many of these mental health problems in youth (Weisz et al., [<reflink idref="bib119" id="ref5">119</reflink>]). Although youth (mental health) care in general has beneficial effects, not all interventions are effective (Howick et al., [<reflink idref="bib44" id="ref6">44</reflink>]; Weisz et al., [<reflink idref="bib118" id="ref7">118</reflink>]) and thus, interventions or the conditions under which they are provided should be improved (Weisz et al., [<reflink idref="bib120" id="ref8">120</reflink>], [<reflink idref="bib118" id="ref9">118</reflink>]). One of the avenues to achieve this is by making use of the youth's and their primary caregivers' informal social network, because support offered by social network members is associated with resilience and positive youth development (Sarason & Sarason, [<reflink idref="bib88" id="ref10">88</reflink>]; Southwick et al., [<reflink idref="bib95" id="ref11">95</reflink>]; Ungar, [<reflink idref="bib106" id="ref12">106</reflink>]; van Dam, Smit, et al., [<reflink idref="bib111" id="ref13">111</reflink>]). The informal social network includes organically formed relationships outside the nuclear family, for instance with extended family, friends, peers and neighbors. From here onwards, we will refer to this informal support system as "social network".</p> <hd id="AN0186283553-2">Activating the social network in interventions</hd> <p>Activating the social network by engaging social network members and utilizing their resources in youth interventions is increasingly integrated in youth interventions, because of its promise to enhance effectiveness while reducing the involvement and time investment of professionals. Social network activation is defined as giving social network members an active role in the provided care that often comes with responsibilities such as participating in decision making, attending treatment sessions, and regular meetings with the youth or the youth's parents at which the intervention is targeted. Through this active role, social network members know which changes the youth or parent is making, and could therefore help adjusting behavior accordingly, and help sustaining this behavior after the intervention ends. A meta-analysis indicated that interventions for promoting family wellness that focused on building parents' social networks were more effective than interventions without such element (MacLeod & Nelson, [<reflink idref="bib63" id="ref14">63</reflink>]), and recent (meta-analytic) reviews on youth-initiated mentoring (YIM), an intervention program in which one member from the social network is activated, also found positive effects (Dantzer & Perry, [<reflink idref="bib18" id="ref15">18</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref16">108</reflink>]). However, other overview studies failed to demonstrate the positive effects of specific programs that activate the social network, that is, family-group decision making (FGDM; Dijkstra et al., [<reflink idref="bib23" id="ref17">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref18">66</reflink>]) and the multi-family program Families and Schools Together (FAST; Valentine et al., [<reflink idref="bib107" id="ref19">107</reflink>]).</p> <p>Activating the social network is increasingly popular in youth interventions, yet the evidence on the effectiveness of specific interventions in extant reviews is unequivocal and no recent systematic quantitative review (i.e. meta-analysis) exists that determines the overall evidence base of interventions that activate the social network in order to prevent or reduce (mental health) problems in youth. Therefore, this meta-analysis examined the effectiveness of various youth interventions that activate the social network compared to forms of interventions that do not involve the social network, and factors that might moderate their effectiveness.</p> <p>Several theories can explain why activating the social network in interventions may be a promising approach to promote positive youth outcomes, including theories on self-determination (Ryan & Deci, [<reflink idref="bib85" id="ref20">85</reflink>]), social support (e.g. Ashida et al., [<reflink idref="bib4" id="ref21">4</reflink>]; Chan et al., [<reflink idref="bib14" id="ref22">14</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref23">108</reflink>]), and social influence (Berkowitz, [<reflink idref="bib9" id="ref24">9</reflink>]; Telzer et al., [<reflink idref="bib103" id="ref25">103</reflink>]). First, social network activation in interventions is thought to contribute to the fulfillment of the basic needs for self-determination (autonomy, competence, and relatedness; Ryan & Deci, [<reflink idref="bib85" id="ref26">85</reflink>]), which may be expected to promote motivation and subsequently the effectiveness of the intervention (Krause, [<reflink idref="bib57" id="ref27">57</reflink>]; van der Helm et al., [<reflink idref="bib112" id="ref28">112</reflink>]). Activation of the social network facilitates shared decision-making about self-concordant treatment goals (Bartelink et al., [<reflink idref="bib8" id="ref29">8</reflink>]) and how to achieve these goals from the perspective of <emph>autonomy</emph>, it fosters self-experienced <emph>competence</emph> by emphasizing what youth can achieve by themselves with their own social environment, and strengthens feelings of <emph>relatedness</emph> with their network (Ashida et al., [<reflink idref="bib4" id="ref30">4</reflink>]; Chan et al., [<reflink idref="bib14" id="ref31">14</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref32">108</reflink>]). Second, activating the social network in interventions is assumed to result in increased social support and higher quality relationships (e.g. Ashida et al., [<reflink idref="bib4" id="ref33">4</reflink>]; Chan et al., [<reflink idref="bib14" id="ref34">14</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref35">108</reflink>]). As these relationships are likely to endure after the intervention ends (Zimmerman, [<reflink idref="bib124" id="ref36">124</reflink>]), this may yield a large cost benefit, as short-term positive effects from the intervention could be maintained or even increased over time due to social support. Support is thought to enhance resilience, or the ability to recover from adversity (Ungar, [<reflink idref="bib106" id="ref37">106</reflink>]) by acting as a protective factor during times of crisis (Ozbay et al., [<reflink idref="bib75" id="ref38">75</reflink>]; Southwick et al., [<reflink idref="bib95" id="ref39">95</reflink>]; Ungar, [<reflink idref="bib106" id="ref40">106</reflink>]). If intervention effects and resilience indeed improve as a result of social support, this potentially leads to less future care use. Third, social influence is a broad concept that encompasses various theories that concern the way people impact one another, leading to changes in attitudes, beliefs, or behaviors. For example, social norms theory predicts that social network activation in interventions can correct misperceptions in social norms. That is, most youth overestimate unhealthy behaviors (e.g. alcohol use) of their peers. By providing accurate normative feedback on unhealthy behaviors, healthier norms are created, which will have a beneficial effect on most youth (Berkowitz, [<reflink idref="bib9" id="ref41">9</reflink>]; Telzer et al., [<reflink idref="bib103" id="ref42">103</reflink>]).</p> <p>In summary, interventions that activate the youth's social network may promote treatment motivation and foster higher quality relationships and social support, which can help alleviate problems. These interventions can be differentiated from treatments that target social network problems, such as multisystemic therapy or multidimensional family therapy, by their focus on <emph>utilizing</emph> the resources in existing social support networks outside the nuclear family rather than improving the social network by reducing interaction problems (Baldwin et al., [<reflink idref="bib7" id="ref43">7</reflink>]). Thus, this meta-analysis focused on the effectiveness of youth interventions in which the social network is utilized through its strength rather than the network being the target of the treatment itself.</p> <hd id="AN0186283553-3">Moderators of intervention effects</hd> <p>In addition to assessing the overall effectiveness of social network engagement in youth interventions, the aim of this study was to identify who benefits most from these interventions and under which circumstances, by examining moderators of intervention effects. Four categories of moderators were examined: program characteristics, sample characteristics, outcome characteristics, and study quality characteristics.</p> <hd id="AN0186283553-4">Program characteristics</hd> <p>Studying program characteristics, such as intervention type, duration and method of involvement of the social network, as moderators of effectiveness, can help understand which intervention types are most effective.</p> <p>Programs vary in who and how many people from the social network are involved, and who chooses to involve whom. That is, interventions vary in who is involved, for example family members (Dijkstra et al., [<reflink idref="bib23" id="ref44">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref45">66</reflink>]), or school staff and peers (Valentine et al., [<reflink idref="bib107" id="ref46">107</reflink>]). Some interventions allow anyone to be involved regardless of their role or type of relationship, such as in youth-initiated mentoring (Dantzer & Perry, [<reflink idref="bib18" id="ref47">18</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref48">108</reflink>]) and some family group decision-making programs (Dijkstra et al., [<reflink idref="bib23" id="ref49">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref50">66</reflink>]). Moreover, child-centered interventions may promote youth to have close contact with one person (Dantzer & Perry, [<reflink idref="bib18" id="ref51">18</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref52">108</reflink>]), whereas family-centered interventions may include components in which different parties with an interest in the youth's wellbeing are gathered without the youth necessarily having close contact with the people engaged (Dijkstra et al., [<reflink idref="bib23" id="ref53">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref54">66</reflink>]). Thus, the number of people involved varies greatly, ranging from one person in youth-initiated mentoring (van Dam et al., [<reflink idref="bib108" id="ref55">108</reflink>]) to a maximum of around 30 people in some family group decision-making programs (Dijkstra et al., [<reflink idref="bib23" id="ref56">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref57">66</reflink>]). We expected that it is more effective to involve one person rather than several people, based on the mechanism of diffusion of responsibility (Darley & Latane, [<reflink idref="bib19" id="ref58">19</reflink>]; Fischer et al., [<reflink idref="bib33" id="ref59">33</reflink>]). That is, if one person holds responsibility, the task will be taken more seriously, whereas if multiple people are involved, they may believe that others from the group will take responsibility, which can result in a situation where no one takes action, because everyone is waiting for someone else to act. Thus, we expected that if one or relatively few people are involved, they are more likely to act (Darley & Latane, [<reflink idref="bib19" id="ref60">19</reflink>]; Fischer et al., [<reflink idref="bib33" id="ref61">33</reflink>]), thus making these programs more effective than interventions involving more people.</p> <p>Furthermore, we expected that if youth themselves make the decision on whom to involve, such as in youth-initiated mentoring, this increases their sense of autonomy, which subsequently stimulates their treatment motivation, which thus contributes to more effective interventions (Ryan & Deci, [<reflink idref="bib85" id="ref62">85</reflink>]). Therefore, we expected that treatment motivation and consequently intervention effectiveness is lower in interventions such as multi-family or family group decision-making programs, in which youth cannot choose who to involve, for example because their parents makes this decision (Dijkstra et al., [<reflink idref="bib23" id="ref63">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref64">66</reflink>]), or people become involved because they belong to the same social group (e.g. classmates, Valentine et al., [<reflink idref="bib107" id="ref65">107</reflink>]).</p> <hd id="AN0186283553-5">Sample characteristics</hd> <p>By examining sample characteristics such as age, socioeconomic status (SES), gender, and region as moderators, we gain insight in for whom interventions with a social network component work best. First, interventions with a social network component may be more effective for adolescents than for children, because youth become more autonomous and create relationships independently from their parents during adolescence (Spear & Kulbok, [<reflink idref="bib96" id="ref66">96</reflink>]). Perhaps youth benefit more from interventions if they autonomously involve individuals from their social network.</p> <p>Second, social network involvement in interventions may be more effective depending on the youth's SES background. That is, low SES families may be less likely to find and involve someone from the network who is well able to assist or engage in an intervention for youth (Schwartz & Rhodes, [<reflink idref="bib90" id="ref67">90</reflink>]), as they have less social support and social capital (American Psychological Association, [<reflink idref="bib3" id="ref68">3</reflink>]; Leventhal & Brooks-Gunn, [<reflink idref="bib61" id="ref69">61</reflink>]). This might result in a higher effectiveness for high and average SES youth than for low SES youth. Alternatively, low SES youth may have more to gain from interventions as they experience more problems (Mclaughlin et al., [<reflink idref="bib67" id="ref70">67</reflink>]), thus resulting in greater effects compared to high SES youth.</p> <p>Third, findings related to the moderating impact of gender were inconsistent in other meta-analysis on specific intervention programs. That is, in a meta-analysis on YIM, larger effect sizes were found in predominantly female samples than in samples with predominantly males or with equal gender distribution (van Dam et al., [<reflink idref="bib108" id="ref71">108</reflink>]), whereas other mentoring meta-analyses demonstrated larger effects in programs serving more male youth (DuBois et al., [<reflink idref="bib24" id="ref72">24</reflink>]; Raposa et al., [<reflink idref="bib79" id="ref73">79</reflink>]).</p> <p>Fourth, the region has been found to be non-significant in meta-analyses on YIM (van Dam et al., [<reflink idref="bib108" id="ref74">108</reflink>]) and family-group decision-making (Dijkstra et al., [<reflink idref="bib23" id="ref75">23</reflink>]) but has still been added to control for a possible impact of the region.</p> <hd id="AN0186283553-6">Assessment characteristics</hd> <p>Intervention effectiveness may be influenced by outcome or assessment characteristics, such as type of assessment and informant. For example, assessment types could influence intervention effectiveness due to response bias (Shadish et al., [<reflink idref="bib91" id="ref76">91</reflink>]). That is, if a researcher is present during data collection, such as in interviews, people are less likely to disclose socially undesirable information and report higher levels of program satisfaction to accommodate the researcher (Ford et al., [<reflink idref="bib34" id="ref77">34</reflink>]; Gnambs & Kaspar, [<reflink idref="bib35" id="ref78">35</reflink>]). Hence, we expected that data collected through interviews yield larger effect sizes than questionnaire data and official reports. Intervention effects could also vary depending on the information source or informant, as reports from youth, parents, professionals, and school staff concern different contexts, perspectives, and frames of reference. For example, professionals and parents may overestimate positive intervention effects as they are both (often) highly involved in the intervention, and therefore positive outcomes may feel more rewarding. On the contrary, due to the tendency of youth (Breuk et al., [<reflink idref="bib12" id="ref79">12</reflink>]) and school staff (Stams et al., [<reflink idref="bib97" id="ref80">97</reflink>]) to give a favorable presentation of themselves or their students, they potentially underestimate problems at pretest, which unduly limits the chance of finding intervention effects.</p> <hd id="AN0186283553-7">Study quality characteristics</hd> <p>Intervention effectiveness may vary depending on characteristics that make up the quality of a study, such as study design, type of control group, and sample size, as these factors have consistently shown to affect meta-analytic results (Cheung & Slavin, [<reflink idref="bib15" id="ref81">15</reflink>]). For example, studies with a more robust design like a randomized-controlled trial (RCT) may yield smaller effect sizes than quasi-experimental studies (Farrington, [<reflink idref="bib30" id="ref82">30</reflink>]), as groups are generally less comparable in quasi-experimental studies (Shadish et al., [<reflink idref="bib91" id="ref83">91</reflink>]). Furthermore, quasi-experimental studies can be retrospective, and may only include participants who completed the intervention, which is a select group. In comparison, RCTs generally include all participants who started the intervention, including those who did not complete it, and thus include participants in the evaluation who likely did not benefit (Farrington, [<reflink idref="bib30" id="ref84">30</reflink>]). Thus, we expected to find larger effect sizes in studies using the completer approach compared to the intention-to-treat approach.</p> <p>Additionally, we expected that studies in which the control group receives no intervention (waiting list) yield larger effect sizes than studies in which the control group receives care as usual (CAU). That is, participants receiving another form of intervention are more likely to improve during the study than untreated participants (Shadish et al., [<reflink idref="bib91" id="ref85">91</reflink>]), possibly due to the effect of common therapeutic factors (e.g. Marcus et al., [<reflink idref="bib64" id="ref86">64</reflink>]). Thus, we expected to find larger effect sizes if the control group received no treatment, than in CAU control groups.</p> <p>Finally, we expected that the quality of older studies is lower compared to more recent studies due to statistical and methodological advances in social science research over the last decades. Additionally, studies that have been published in journals with high impact factors may report larger effects than unpublished studies due to publication bias (Cheung & Slavin, [<reflink idref="bib15" id="ref87">15</reflink>]; Saha et al., [<reflink idref="bib87" id="ref88">87</reflink>]).</p> <hd id="AN0186283553-8">Current study</hd> <p>In this study, we aimed to gain knowledge on the effectiveness of youth care interventions with a social network component, and on the impact of potential moderators of effectiveness. For this purpose, a literature search was conducted to trace relevant studies. A three-level meta-analytic design was used, allowing the extraction of multiple effect sizes from individual primary studies. In this way, a comprehensive understanding of the effectiveness of social network involvement in interventions for youth is provided.</p> <hd id="AN0186283553-9">Materials and methods</hd> <p></p> <hd id="AN0186283553-10">Study selection</hd> <p>The literature search included English-language peer-reviewed studies and (unpublished) dissertations and reports about interventions for youth in which the social network is involved. A number of criteria were specified to determine whether studies could be included in this meta-analysis. First, interventions should be focused on youth (0–26 years) to be included. That is, interventions that do not have youth as (primary) client (e.g. parent training programs) were not included. Second, interventions had to include at least one component of social network activation. This component should be an essential and integral part of the intervention, hence, the activation of the social network should not be optional or recommended. The social network should be involved to provide youth and/or parents with support, and interventions that aim to change dysfunctional dynamics in the social network (such as family therapy) were excluded. Likewise, interventions that are led by the social network (e.g. peer-led or teacher-led) were also excluded. Moreover, interventions that only engaged the immediate family (parents, siblings, and other members of the household), and no other members of the social network, were not considered for this study. Third, only controlled studies were included. Although RCTs can be regarded as the "golden standard" in effectiveness studies (Farrington, [<reflink idref="bib30" id="ref89">30</reflink>]), we decided to include any experimental study with a control group to increase the power, enhance the generalizability of the results, and to reduce the risk of missing relevant results. Fourth, studies had to report on youth outcomes, such as social and emotional competencies, psychological or behavioral functioning, and child safety. Sixth and finally, studies had to report or provide statistics suitable for meta-analyses, or sufficient statistical information that is required for calculating an effect size manually (e.g. proportions or mean scores and standard deviations).</p> <p>We conducted a literature search through Ovid in electronic databases ERIC and PsycInfo using a search string with five elements: (<reflink idref="bib1" id="ref90">1</reflink>) social network component; (<reflink idref="bib2" id="ref91">2</reflink>) intervention; (<reflink idref="bib3" id="ref92">3</reflink>) target group; (<reflink idref="bib4" id="ref93">4</reflink>) research design; and (<reflink idref="bib5" id="ref94">5</reflink>) youth outcome. The complete search strategy is shown in Appendix A. This search strategy resulted in 6303 records. Next, we identified more potentially relevant primary studies using snowballing. We inspected reference lists of meta-analyses and systematic reviews (Allan et al., [<reflink idref="bib2" id="ref95">2</reflink>]; Dijkstra et al., [<reflink idref="bib23" id="ref96">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref97">66</reflink>]; Valentine et al., [<reflink idref="bib107" id="ref98">107</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref99">108</reflink>]) and studies found through the electronic search. These additional sources resulted in 73 records. After deduplicating, these search strategies resulted in a list of 5979 studies, which were screened by the first, second and third authors in Rayyan (Ouzzani et al., [<reflink idref="bib74" id="ref100">74</reflink>]) according to the aforementioned inclusion criteria. Studies were most frequently excluded because the interventions were not youth-focused and/or did not include a social network component. We excluded 13 studies because full texts were unavailable. A total of 37 studies, with 35 independent samples, and 409 effect sizes met the inclusion criteria. The study selection process is presented in a flowchart in Figure 1. Table 1 shows the characteristics of the included studies.</p> <p>PHOTO (COLOR): Figure 1. Flowchart of the Study Selection Process. * The most common reasons for exclusion were: (<reflink idref="bib1" id="ref101">1</reflink>) intervention has no social network component; (<reflink idref="bib2" id="ref102">2</reflink>) study is not an empirical study (e.g., review, meta-analysis, or publication without empirical data); and (<reflink idref="bib3" id="ref103">3</reflink>) wrong study design (e.g., not an intervention study, not a controlled study).</p> <p>Table 1. Characteristics of included studies.</p> <p> <ephtml> <table><thead><tr><td>Study</td><td>Intervention</td><td>Design</td><td><italic>n</italic></td><td>No. of outcomes</td><td>No. of ES</td></tr></thead><tbody valign="top"><tr><td>Baffour (<xref ref-type="bibr" rid="bibr6">2006</xref>)</td><td>FGDM</td><td>RCT</td><td>292</td><td>1</td><td>1</td></tr><tr><td>Berzin (<xref ref-type="bibr" rid="bibr10">2006</xref>)</td><td>FGDM</td><td>QE</td><td>327</td><td>3</td><td>4</td></tr><tr><td>de Vries et al. (<xref ref-type="bibr" rid="bibr21">2017</xref>)</td><td>NP</td><td>RCT</td><td>101</td><td>11</td><td>22</td></tr><tr><td>de Vries et al. (<xref ref-type="bibr" rid="bibr20">2018</xref>)</td><td>NP</td><td>RCT</td><td>101</td><td>9</td><td>18</td></tr><tr><td>Dijkstra et al. (<xref ref-type="bibr" rid="bibr22">2019</xref>)</td><td>FGDM</td><td>RCT</td><td>328</td><td>7</td><td>22</td></tr><tr><td>Edwards et al. (<xref ref-type="bibr" rid="bibr28">2006</xref>)</td><td>FGDM</td><td>QE</td><td>680</td><td>1</td><td>3</td></tr><tr><td>Greeson and Thompson (<xref ref-type="bibr" rid="bibr36">2017</xref>)</td><td>YIM</td><td>RCT</td><td>17</td><td>11</td><td>11</td></tr><tr><td>Hauken et al. (<xref ref-type="bibr" rid="bibr40">2018</xref>)</td><td>CPP</td><td>RCT</td><td>35</td><td>2</td><td>2</td></tr><tr><td>Hollinshead et al. (<xref ref-type="bibr" rid="bibr43">2017</xref>)</td><td>FGDM</td><td>RCT</td><td>503</td><td>3</td><td>3</td></tr><tr><td>James et al. (<xref ref-type="bibr" rid="bibr46">2016</xref>)</td><td>NP</td><td>RCT</td><td>127</td><td>21</td><td>21</td></tr><tr><td>Jeong et al. (<xref ref-type="bibr" rid="bibr47">2012</xref>)</td><td>FGDM</td><td>RCT</td><td>782</td><td>2</td><td>2</td></tr><tr><td>King et al. (<xref ref-type="bibr" rid="bibr50">2006</xref>)</td><td>YIM</td><td>RCT</td><td>197</td><td>17</td><td>28</td></tr><tr><td>King et al. (<xref ref-type="bibr" rid="bibr49">2009</xref>)</td><td>YIM</td><td>RCT</td><td>448</td><td>15</td><td>17</td></tr><tr><td>King et al. (<xref ref-type="bibr" rid="bibr48">2019</xref>)</td><td>YIM</td><td>RCT</td><td>448</td><td>2</td><td>2</td></tr><tr><td>Knox et al. (<xref ref-type="bibr" rid="bibr52">2011</xref>)</td><td>FAST</td><td>QE</td><td>282</td><td>4</td><td>10</td></tr><tr><td>Kratochwill et al. (<xref ref-type="bibr" rid="bibr55">2004</xref>)</td><td>FAST</td><td>RCT</td><td>100</td><td>5</td><td>20</td></tr><tr><td>Kratochwill et al. (<xref ref-type="bibr" rid="bibr56">2009</xref>)</td><td>FAST</td><td>RCT</td><td>134</td><td>5</td><td>20</td></tr><tr><td>Lambert et al. (<xref ref-type="bibr" rid="bibr58">2017</xref>)</td><td>FGDM</td><td>QE</td><td>613,180</td><td>1</td><td>1</td></tr><tr><td>Leon et al. (<xref ref-type="bibr" rid="bibr60">2016</xref>)</td><td>FF</td><td>QE</td><td>458</td><td>4</td><td>4</td></tr><tr><td>McGarrell & Hipple (<xref ref-type="bibr" rid="bibr65">2007</xref>)</td><td>FGDM</td><td>RCT</td><td>782</td><td>2</td><td>9</td></tr><tr><td>Millenky et al. (<xref ref-type="bibr" rid="bibr69">2010</xref>)</td><td>NGYCP</td><td>RCT</td><td>3,074</td><td>55</td><td>55</td></tr><tr><td>Millenky et al. (<xref ref-type="bibr" rid="bibr70">2014</xref>)</td><td>NGYCP</td><td>RCT</td><td>1,173</td><td>12</td><td>12</td></tr><tr><td>Onrust et al. (<xref ref-type="bibr" rid="bibr73">2015</xref>)</td><td>FGDM</td><td>QE</td><td>124</td><td>1</td><td>1</td></tr><tr><td>Pennell et al. (<xref ref-type="bibr" rid="bibr76">2010</xref>)</td><td>FGDM</td><td>QE</td><td>649</td><td>1</td><td>1</td></tr><tr><td>Reekers et al. (<xref ref-type="bibr" rid="bibr80">2018</xref>)</td><td>SoS</td><td>QE</td><td>38</td><td>5</td><td>5</td></tr><tr><td>Rushovich et al. (<xref ref-type="bibr" rid="bibr84">2021</xref>)</td><td>FGDM</td><td>RCT</td><td>1,423</td><td>5</td><td>13</td></tr><tr><td>Sheets et al. (<xref ref-type="bibr" rid="bibr93">2009</xref>)</td><td>FGDM</td><td>QE</td><td>4,066</td><td>1</td><td>1</td></tr><tr><td>Sundell and Vinnerljung (<xref ref-type="bibr" rid="bibr100">2004</xref>)</td><td>FGDM</td><td>QE</td><td>239</td><td>8</td><td>8</td></tr><tr><td>Teal (<xref ref-type="bibr" rid="bibr102">2013</xref>)</td><td>FGDM</td><td>QE</td><td>755</td><td>2</td><td>2</td></tr><tr><td>Tijms et al. (<xref ref-type="bibr" rid="bibr104">2018</xref>)</td><td>BBC</td><td>RCT</td><td>90</td><td>1</td><td>2</td></tr><tr><td>Turley et al. (<xref ref-type="bibr" rid="bibr105">2017</xref>)</td><td>FAST</td><td>RCT</td><td>3,084</td><td>2</td><td>4</td></tr><tr><td>van Dam et al. (<xref ref-type="bibr" rid="bibr110">2017</xref>)</td><td>YIM</td><td>QE</td><td>200</td><td>1</td><td>1</td></tr><tr><td>van Dam, Klein Schaarsberg, et al. (<xref ref-type="bibr" rid="bibr109">2018</xref>)</td><td>YIM</td><td>QE</td><td>42</td><td>3</td><td>3</td></tr><tr><td>Vandivere et al. (<xref ref-type="bibr" rid="bibr114">2017</xref>)</td><td>FF</td><td>RCT</td><td>568</td><td>38</td><td>63</td></tr><tr><td>Wang et al. (<xref ref-type="bibr" rid="bibr116">2012</xref>)</td><td>FGDM</td><td>QE</td><td>80,690</td><td>1</td><td>1</td></tr><tr><td>Weigensberg et al. (<xref ref-type="bibr" rid="bibr117">2009</xref>)</td><td>FGDM</td><td>QE</td><td>650</td><td>6</td><td>6</td></tr><tr><td>Wingrove and Weisz (<xref ref-type="bibr" rid="bibr121">2005</xref>)</td><td>FGDM</td><td>QE</td><td>66</td><td>2</td><td>2</td></tr></tbody></table> </ephtml> </p> <p>1 <emph>Note.</emph> ES: effect size; FGDM: family group decision making; NGYCP: National Guard Youth Challenge program; NP: New Perspectives; YIM: youth-initiated mentoring; CPP: Cancer PEPSONE Program; FAST: Families and Schools Together; FF: Family Finding; SoS: Signs of Safety; BBC: Bibliotherapeutic Bookclub; QE: quasi-experimental trial; RCT: randomized-controlled trial.</p> <hd id="AN0186283553-11">Coding studies and potential moderators</hd> <p>To code relevant study characteristics, a coding instrument was developed based on the coding instruments of Dijkstra et al. ([<reflink idref="bib23" id="ref104">23</reflink>]) and van Dam et al. ([<reflink idref="bib108" id="ref105">108</reflink>]). We also included additional relevant codes during the screening and coding processes based on the scope of our meta-analysis and the available information in the included studies. The potentially relevant moderators were clustered into four categories: program characteristics, sample characteristics, outcome characteristics, and study quality characteristics. These are common categories in meta-analyses in general and in this specific field, which facilitates comparisons between the results of different meta-analyses. The codebook can be found in Appendix B.</p> <p>Coding was done by the first, second, third, and fourth author with continuous support from the last author. To ensure consistency in coding, the last author trained the four coders during a training phase, in which all coders coded the same study, and differences in coding were discussed and resolved until absolute agreement was reached. During the coding process there were regular meetings with the team in which ambiguities were discussed. After all studies had been coded, the first author checked about 25% of the codes from the second, third, and fourth author. Inconsistencies appeared very rarely, and were discussed with the coder concerned and the last author until an agreement was reached.</p> <hd id="AN0186283553-12">Program characteristics</hd> <p>Intervention program characteristics were coded by the nature of the intervention (universal preventive, selective preventive, indicative preventive, curative), program context (child welfare, mental health care, law enforcement, or community/school), program duration (in sessions and in months), whether the intervention was a youth-initiated mentoring, family group decision-making or multi-family program (yes, no), whether youth themselves decide on who to involve (yes, no), the types of people that are involved (extended family, school, peers, and neighborhood), the number of people that are involved (one person, multiple people), and the percentage of participants from the experimental group in which social network engagement was successful, which served as a measure of fidelity.</p> <hd id="AN0186283553-13">Sample characteristics</hd> <p>The following sample characteristics were coded: continent, level of risk, special populations (juvenile offenders, mental health needs), gender, ethnicity, age, SES, and family composition. The continent was coded as the continent from which the sample originated. Level of risk was coded as low risk (normative samples) or high risk (samples with apparent preexisting risk factors). Additionally, two special populations were identified, that is, juvenile offenders and youth with mental health needs (yes, no). Gender was operationalized as the percentage of boys in the sample, and age as the mean age. We coded the percentage of ethnic minorities in the sample as a measure of ethnicity. SES was coded as predominantly low, or predominantly average or high based on the education, job, and income of youth and/or their parents. We calculated the percentage of intact families as a measure of family composition.</p> <hd id="AN0186283553-14">Assessment characteristics</hd> <p>The following assessment characteristics were coded: outcome domain (first narrow domains were coded, e.g. <emph>aggression</emph>, which were then clustered into broad domains, e.g. <emph>externalizing problems</emph>, to ensure robust moderation analyses; 1 = academic/work functioning: school, work, executive functioning; 2 = externalizing problems: aggression, delinquency, conduct problems; 3 = family functioning/child unsafety: child unsafety, child maltreatment, out of home placements, service use; 4 = physical health; 5 = psychological problems: internalizing problems, substance use, unhealthy coping; and 6 = social: social support, social skills), assessment type (questionnaire, interview, official record, observation), information source (youth, parents, school, combination, staff, official record), time of assessment (post-test, follow-up), and the number of weeks after the intervention ended. For each outcome, we coded the <emph>n</emph> per group and the corresponding effect size.</p> <hd id="AN0186283553-15">Study quality characteristics</hd> <p>The following study quality characteristics were coded: year of publication, whether a study was peer-reviewed (yes, no), journal impact factor, Q-rank, sample size, non-response, study design (RCT or quasi-experimental, and retrospective or prospective), whether the effect size is corrected for pretest levels (yes, no), intention-to-treat (yes, no), and control condition (CAU, no care).</p> <hd id="AN0186283553-16">Calculation and analysis of effect sizes</hd> <p>Cohen's <emph>d</emph> was calculated to establish the effectiveness of social network involvement in youth interventions based on differences between youth receiving interventions containing a social network component and youth receiving regular care or no care. Following the criteria of Cohen ([<reflink idref="bib17" id="ref106">17</reflink>]), an effect size of <emph>d</emph> = 0.20 was considered small, an effect size of <emph>d</emph> = 0.50 was considered medium and an effect size of <emph>d</emph> = 0.80 was considered large. In most cases, Cohen's <emph>d</emph> was computed based on means, standard deviations, <emph>t, F</emph>, <sups>2</sups> or a one-tailed <emph>p</emph>-value using formulas of Lipsey et al. ([<reflink idref="bib62" id="ref107">62</reflink>]).</p> <p>All coded data and calculated effect sizes were entered in SPSS version 26. Before the analyses were performed, categorical variables were recoded into dummy variables for each category of a variable, and continuous variables were centered around their mean.</p> <p>To examine the overall effect of interventions involving the social network, a multi-level meta-analysis was performed as most of the included studies assessed more than one outcome. To account for the dependency of the effect sizes from the same study a three-level meta-analytic approach was used. This approach results in a better estimation of effects and more statistical power compared to a traditional meta-analytic approach (Bijleveld & Commandeur, [<reflink idref="bib11" id="ref108">11</reflink>]). An advantage is that multiple variables can be tested as potential moderators of the overall effect (Assink & Wibbelink, [<reflink idref="bib5" id="ref109">5</reflink>]).</p> <p>Three sources of variance are modeled in a three-level meta-analysis: (<reflink idref="bib1" id="ref110">1</reflink>) the sampling variance of the observed effect sizes, (<reflink idref="bib2" id="ref111">2</reflink>) the variance between effect sizes obtained from the same study, and (<reflink idref="bib3" id="ref112">3</reflink>) the variance between studies. To determine whether the variance on the second (within-study) and/or third (between-study) level was significant, two one-sided log-likelihood-ratio tests were conducted (Assink & Wibbelink, [<reflink idref="bib5" id="ref113">5</reflink>]). Significant variance at level two or three indicates heterogeneity in the effect size distribution, meaning that the overall mean effect size is not a correct estimate of a common effect size. In such cases, moderator analyses can be performed in an attempt to explain within-study and/or between-study heterogeneity in effect sizes.</p> <p>All analyses were conducted in <emph>R</emph> (version 4.0.5, R Core Team, [<reflink idref="bib78" id="ref114">78</reflink>]) using the metafor package (Viechtbauer, [<reflink idref="bib115" id="ref115">115</reflink>]), and the syntax by Assink and Wibbelink ([<reflink idref="bib5" id="ref116">5</reflink>]). All model parameters were estimated using the restricted maximum likelihood estimate in random effects meta-analytic models, using a 5% significance level. Knapp and Hartung ([<reflink idref="bib51" id="ref117">51</reflink>]) was used in testing the significance of individual regression coefficients, implying that the significance of coefficients was tested using the <emph>t</emph>- and <emph>F</emph>-distributions rather than the <emph>z-</emph>distribution. Moderator analyses were conducted to test the moderating effect of the selected sample characteristics on the overall effect of the interventions. Finally, a multiple moderator model was tested, including all significant moderators to examine the unique impact of each moderator. The final data set and <emph>R</emph> scripts can be found on Open Science Framework (OSF): https://osf.io/9vzg2/.</p> <hd id="AN0186283553-17">Publication bias</hd> <p>A common problem in performing meta-analyses is that studies may not have been published because of non-significant or unfavorable findings, the so-called "file drawer problem", resulting in publication bias (Rosenthal, [<reflink idref="bib82" id="ref118">82</reflink>]). We obtained unpublished material as best as possible, which is the simplest solution to publication bias (Mullen, [<reflink idref="bib71" id="ref119">71</reflink>]). To examine the potential presence of publication bias, we applied three methods. First, we used Egger et al. ([<reflink idref="bib29" id="ref120">29</reflink>]) regression test to test publication bias. Following Fernández-Castilla et al. ([<reflink idref="bib32" id="ref121">32</reflink>]), an adapted version of Egger's test was used, accounting for the dependency of effect sizes, to test the association between the effect size and the standard error. The standard error of the effect size was included as a moderator in the regression model.</p> <p>Second, we used an extension of the funnel plot test for use in three-level meta-analyses (Fernández-Castilla et al., [<reflink idref="bib32" id="ref122">32</reflink>]), formally testing funnel plot asymmetry. Following the guidelines by Fernández-Castilla et al. ([<reflink idref="bib31" id="ref123">31</reflink>]), we depict both funnel plots of effect sizes and plots of study effects. Effect sizes missing in the lower-left part of the funnel plot indicates publication bias. In a funnel plot of study effects, separate random-effects meta-analyses are conducted on each study, resulting in a dot based on the sample size and the number of effect sizes within the study.</p> <p>Third, we performed the trim-and fill-method (Duval & Tweedie, [<reflink idref="bib26" id="ref124">26</reflink>], [<reflink idref="bib27" id="ref125">27</reflink>]), testing whether effect sizes are missing on the left side of the distribution, indicating publication bias. Previous simulation studies have shown that effect size estimates based on the imputation of effect sizes after the trim-and-fill procedure may not be accurate (Fernández-Castilla et al., [<reflink idref="bib32" id="ref126">32</reflink>]; Peters et al., [<reflink idref="bib77" id="ref127">77</reflink>]). Therefore, we used the trim-and-fill procedure as outlined by Fernández-Castilla et al. ([<reflink idref="bib32" id="ref128">32</reflink>]), which estimates the number of effect sizes imputed at the right side or left side of the distribution to examine whether the overall effect size estimates were sensitive to the potential presence of publication bias. Fernández-Castilla et al. ([<reflink idref="bib32" id="ref129">32</reflink>]) have proposed a method in which the estimated number of effect sizes on the left side of the funnel plot distribution is related to a cutoff value of the estimator of the trim-and-fill method, based on the population effect size and power (number of effect sizes). If the number of imputed studies exceeds the cutoff value, this may be indicative of publication bias.</p> <hd id="AN0186283553-18">Results</hd> <p></p> <hd id="AN0186283553-19">Descriptive characteristics</hd> <p>The current meta-analysis consisted of 37 studies, with 35 independent samples, and 409 effect sizes. The samples consisted of 712,269 participants in total. There were two studies with extremely large samples (<emph>n</emph> > 5,000), namely Wang et al. ([<reflink idref="bib116" id="ref130">116</reflink>]) <emph>n</emph> = 80,690, and Lambert et al. ([<reflink idref="bib58" id="ref131">58</reflink>]) <emph>n</emph> = 613,180. Samples consisted of youth aged 0-26 years (<emph>M</emph> = 7.20), and 52.7% boys on average. Most samples came from North-America (71.4%), and some from Europe (28.6%). Almost all samples included at-risk youth populations (94.1%), such as youth with mental health needs (17.1%) and juvenile offenders (11.4%). Studies were published between 2004 and 2021. Most studies had an RCT (<emph>k</emph> = 21, 56.8%) as the study design, others had a quasi-experimental design (<emph>k</emph> = 16, 43.2%). Intention-to-treat was the most common analytic design (<emph>k</emph> = 16, 43.2%), other studies were analyzed using completer-analysis (<emph>k</emph> = 14, 37.8%), or did not mention the analysis strategy (<emph>k</emph> = 7, 18.9%). The studies examined various program types involving the social network, such as youth-initiated mentoring, family group decision-making, and multi-family programs (see Table 2).</p> <p>Table 2. Number of studies and effect sizes per intervention type.</p> <p> <ephtml> <table><thead><tr><td /><td>YIM</td><td>FGDM</td><td>Multi-family</td><td>Other</td><td>Total</td></tr></thead><tbody valign="top"><tr><td># Studies (%)</td><td>11 (29.7)</td><td>20 (54.1)</td><td char=".">4 (10.8)</td><td>2 (5.4)</td><td char=".">37</td></tr><tr><td># ES (%)</td><td>190 (46.5)</td><td>147 (35.9)</td><td char=".">69 (16.9)</td><td>3 (0.7)</td><td char=".">409</td></tr></tbody></table> </ephtml> </p> <p>2 <emph>Note</emph>. # studies: number of studies; # ES: number of effect sizes; YIM: youth-initiated mentoring; FGDM: family group decision making.</p> <hd id="AN0186283553-20">Overall effect and heterogeneity in effect sizes</hd> <p>Overall, interventions with social network involvement for youth were not significantly more effective than interventions without such component, <emph>d</emph> = 0.11, 95% CI [-0.07; 0.29], <emph>t</emph>(<reflink idref="bib408" id="ref132">408</reflink>) = 1.17, <emph>p =</emph>.241 (see Table 3 for detailed model results and Appendix C for the forest plot). The one-sided log-likelihood ratio tests showed that significant variance was present both at level 2 and level 3 of the meta-analytic model, <emph>χ</emph><sups>2</sups>(<reflink idref="bib2" id="ref133">2</reflink>) = 2922.30, <emph>p</emph> <.001, and <emph>χ</emph><sups>2</sups>(<reflink idref="bib2" id="ref134">2</reflink>) = 84.12, <emph>p</emph> <.001, respectively. Of the total variance, 23.2% was distributed at the within-study level (level 2), and 75.2% at the between-study level (level 3). Random sampling error accounted for 1.6% of the total variance.</p> <p>Table 3. Overall effect of interventions involving the social network on youth outcomes.</p> <p> <ephtml> <table><thead><tr><td><italic>K</italic></td><td><italic>#ES</italic></td><td>Mean <italic>d</italic> (<italic>SE</italic>)</td><td>95% <italic>CI</italic></td><td><italic>p</italic></td><td>σ<sup>2</sup><sub>level 2</sub></td><td>σ<sup>2</sup><sub>level 3</sub></td><td>% σ<sup>2</sup><sub>level 1</sub></td><td>% σ<sup>2</sup><sub>level 2</sub></td><td>% σ<sup>2</sup><sub>level 3</sub></td></tr></thead><tbody valign="top"><tr><td>37</td><td>409</td><td>0.11 (0.09)</td><td char=".">−0.07, 0.29</td><td>.241</td><td char=".">0.08</td><td char=".">0.25</td><td char=".">1.61</td><td char=".">23.20</td><td char=".">75.19</td></tr></tbody></table> </ephtml> </p> <p>3 <emph>Note. k</emph>: number of studies; <emph>#ES</emph>: number of effect sizes; Mean <emph>d</emph>: mean effect size (Cohen's <emph>d</emph>); <emph>SE</emph>: standard error; <emph>CI</emph>: confidence interval; σ<sups>2</sups><subs>level 2</subs>: variance between effect sizes extracted from the same study; σ<sups>2</sups><subs>level 3</subs>: variance between studies; % σ<sups>2</sups>: percentage of variance distributed.</p> <hd id="AN0186283553-21">Publication bias</hd> <p>The significant Egger's test (<emph>b</emph> = −3.57, <emph>z</emph> = −4.58, <emph>p</emph> <.001) and the visual inspection of the funnel plots (Figures 2 and 3) showed some indication of publication bias. However, publication bias is not confirmed in the formal test of funnel plot asymmetry (<emph>b</emph> = 0.00, <emph>z</emph> = −0.92, <emph>p</emph> =.356). Moreover, further funnel plot analysis shows that there are no effect sizes that could or should be imputed on the left or right side of the funnel to restore symmetry. Since publication bias was unlikely, imputation of effect sizes was not necessary.</p> <p>Graph: Figure 2. Funnel Plot of Effect Sizes.</p> <p>Graph: Figure 3. Funnel Plot of Study Effects.</p> <hd id="AN0186283553-22">Sensitivity analyses</hd> <p>We conducted two types of sensitivity analyses to check the robustness of the overall effect. First, we winsorized the effect sizes for seven outlying effect sizes (one positive and six negative) that exceeded a <emph>Z</emph>-score of 3.3 (Tabachnick & Fidell, [<reflink idref="bib101" id="ref135">101</reflink>]). The analysis including the winsorized effect sizes produced an overall effect of <emph>d</emph> = 0.03, 95% CI [−0.05; 0.11], <emph>t</emph>(<reflink idref="bib408" id="ref136">408</reflink>) = 0.73, <emph>p =</emph>.467. This effect size is somewhat below our initial estimated overall effect (Δ<emph>d</emph> = −0.08), but it lies within the confidence interval of the initial effect size, indicating that the outliers did not significantly influence our results.</p> <p>Second, we performed a sensitivity analysis excluding two studies with extremely large samples (<emph>n</emph> > 5000), namely Wang et al. ([<reflink idref="bib116" id="ref137">116</reflink>]) <emph>n</emph> = 80,690, and Lambert et al. ([<reflink idref="bib58" id="ref138">58</reflink>]) <emph>n</emph> = 613,180. The analysis excluding these studies produced an overall effect of <emph>d</emph> = 0.13, 95% CI [−0.06; 0.31], <emph>t</emph>(<reflink idref="bib406" id="ref139">406</reflink>) = 1.30, <emph>p =</emph>.194. This effect size is very similar to our initial estimated overall effect (Δ<emph>d</emph> = 0.02) and lies within its confidence interval, indicating that the two large samples did not significantly influence our results. Therefore, we continued with the original, untransformed data including all studies in the following moderator analyses.</p> <hd id="AN0186283553-23">Moderator analysis bivariate models</hd> <p></p> <hd id="AN0186283553-24">Program characteristics</hd> <p>Three program characteristics influenced the effectiveness of interventions in which the social network is activated. First, intervention type was a significant moderator. That is, interventions involving a form of youth-initiated mentoring (<emph>d</emph> = 0.46) yielded small to medium intervention effects (Cohen, [<reflink idref="bib17" id="ref140">17</reflink>]), whereas interventions without mentoring (<emph>d</emph> = −0.03) were not effective, <emph>F</emph>(<reflink idref="bib1" id="ref141">1</reflink>, 407) = 6.51, <emph>p</emph> =.011. Post-hoc analyses revealed that the effect sizes remained similar when excluding studies with children with a mean age below 11, which was the minimum age in the included studies on youth-initiated mentoring. That is, <emph>d</emph> = 0.50 for mentoring interventions, and <emph>d</emph> = −0.02 for interventions without mentoring, <emph>F</emph>(<reflink idref="bib1" id="ref142">1</reflink>, 355) = 3.66, <emph>p</emph> =.056. Yet, the <emph>p</emph>-value demonstrated that the effect was no longer statistically significant. There was no effect for family group decision-making (<emph>p</emph> =.134) and multi-family interventions (<emph>p</emph> =.488). Second, interventions in which youth decide themselves who to involve (<emph>d</emph> = 0.52) were more effective than interventions in which other parties decide (<emph>d</emph> = −0.03), <emph>F</emph>(<reflink idref="bib1" id="ref143">1</reflink>, 379) = 7.11, <emph>p</emph> =.008. Post-hoc analyses demonstrated that even after excluding studies involving children with a mean age below 11, where decision-making themselves may not be possible, the effect sizes remained similar and statistically significant. That is, <emph>d</emph> = 0.56 for intervention in which youth decide, and <emph>d</emph> = −0.02 for interventions in which other parties decide, <emph>F</emph>(<reflink idref="bib1" id="ref144">1</reflink>, 327) = 4.02, <emph>p</emph> =.046. Third, interventions involving just one person from the social network (<emph>d</emph> = 0.56) was more effective than interventions involving multiple people (<emph>d</emph> = −0.02), <emph>F</emph>(<reflink idref="bib1" id="ref145">1</reflink>, 407) = 7.60, <emph>p</emph> =.006. The effects of interventions in which youth decide and interventions involving just one person were medium in size (Cohen, [<reflink idref="bib17" id="ref146">17</reflink>]).</p> <p>The remaining program characteristics did not moderate the intervention effects, that is, intervention nature (<emph>p</emph> =.822), context (<emph>p</emph> =.228), the number of sessions (<emph>p</emph> =.636), duration (in months, <emph>p</emph> =.787), who is involved from the social network (<emph>p</emph>s ≥.446), and the percentage of participants from the experimental group in which social network engagement was successful (<emph>p</emph> =.246).</p> <hd id="AN0186283553-25">Sample characteristics</hd> <p>Only two sample characteristics significantly moderated the effectiveness of interventions that activate the social network. First, the continent from which samples originated moderated intervention effects, <emph>F</emph>(<reflink idref="bib1" id="ref147">1</reflink>, 407) = 4.34, <emph>p</emph> =.038. That is, intervention effects were small (Cohen, [<reflink idref="bib17" id="ref148">17</reflink>]) in European samples (<emph>d</emph> = 0.40), whereas there were no intervention effects in North-American samples (<emph>d</emph> = −0.01). Second, while the overall risk level of the sample did not moderate intervention effects (<emph>p</emph> =.718), effects were significantly moderated by whether a sample included youth with mental health needs. That is, interventions were more effective for youth with mental health needs, than those without, <emph>F</emph>(<reflink idref="bib1" id="ref149">1</reflink>, 407) = 14.01, <emph>p</emph> <.001. The effect sizes indicated that interventions for youth with mental health needs (<emph>d</emph> = 0.75) were medium to large (Cohen, [<reflink idref="bib17" id="ref150">17</reflink>]), whereas interventions for other youth was not effective (<emph>d</emph> = −0.03). The effectiveness did not vary for samples with and without juvenile offenders, <emph>p</emph> =.679.</p> <p>Other sample characteristics did not moderate the effectiveness, that is, gender (<emph>p</emph> =.974), age (<emph>p</emph> =.151), SES (<emph>p</emph> =.381), ethnicity (<emph>p</emph> =.180), and family composition (<emph>p</emph> =.235).</p> <hd id="AN0186283553-26">Assessment characteristics</hd> <p>Assessment type and information source were significant moderators of intervention effectiveness. First, intervention effects were smaller if data was retrieved through interviews (<emph>d</emph> = −0.08), than through questionnaires (<emph>d</emph> = 0.10) and official records (<emph>d</emph> = 0.14), <emph>F</emph>(<reflink idref="bib2" id="ref151">2</reflink>, 406) = 5.04, <emph>p</emph> =.007. Yet, all effects were (very) small (Cohen, [<reflink idref="bib17" id="ref152">17</reflink>]). Second, information source was significant, indicating that intervention effects were larger if assessments were completed by professionals (<emph>d</emph> = 0.34) or parents (<emph>d</emph> = 0.17), or if data was retrieved from official records (<emph>d</emph> = 0.15), compared to assessments completed by youth (<emph>d</emph> = −0.06), school staff (<emph>d</emph> = 0.06), or a combination of information sources (<emph>d</emph> = 0.00), <emph>F</emph>(<reflink idref="bib5" id="ref153">5</reflink>, 403) = 4.06, <emph>p</emph> =.001. The effects were small for assessments completed by professionals, and very small for parent-reported data and official records (Cohen, [<reflink idref="bib17" id="ref154">17</reflink>]).</p> <p>Outcome domain (<emph>p</emph> =.793), timing of assessment (<emph>p</emph> =.770), and the number of weeks after ending the intervention (<emph>p</emph> =.565) did not significantly moderate intervention effectiveness.</p> <hd id="AN0186283553-27">Study quality characteristics</hd> <p>Effect sizes were significantly larger if the effect size was corrected for differences between the intervention and control conditions at pretest, <emph>F</emph>(<reflink idref="bib1" id="ref155">1</reflink>, 406) = 10.72, <emph>p</emph> =.001. That is, effect sizes were non-significant if they were not corrected (<emph>d</emph> = 0.03), and small in size if they were corrected for pretest differences (<emph>d</emph> = 0.27), which points to a self-selection effect.</p> <p>The other study quality characteristics did not moderate intervention effectiveness, that is, publication year (<emph>p</emph> =.098), peer-reviewed studies (<emph>p</emph> =.243), impact factor (<emph>p</emph> =.497), Q-rank (<emph>p</emph> =.987), sample size (<emph>p</emph> =.357), non-response (<emph>p</emph> =.482), study design (<emph>p</emph> =.098 for RCT vs. quasi-experimental, and <emph>p</emph> =.213 for retrospective vs. prospective), intention-to-treat as analytic method (<emph>p</emph> =.358), and control condition (<emph>p</emph> =.246). See Table 4 for the results of all moderator analyses.</p> <p>Table 4. Results of the moderator analyses.</p> <p> <ephtml> <table><thead><tr><td>Moderator</td><td># samples</td><td># ES</td><td>β<sub>0</sub>/Mean <italic>d</italic> (95% <italic>CI</italic>) </td><td>β<sub>1</sub> (95% <italic>CI</italic>) </td><td><italic>F (df</italic><sub>1</sub><italic>, df</italic><sub>2</sub><italic>)</italic></td><td><italic>p</italic></td></tr></thead><tbody valign="top"><tr><td><bold>Program characteristics</bold></td><td /><td char="." /><td char="." /><td char="." /><td char="." /><td char="." /></tr><tr><td>Program nature</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.31 (3, 405)</td><td char=".">.822</td></tr><tr><td> Universal preventive</td><td>3</td><td char=".">64</td><td char=".">−0.14 (−0.76, 0.47)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Selective preventive</td><td>3</td><td char=".">47</td><td char=".">0.10 (−0.53, 0.74)</td><td char=".">0.25 (−0.64, 1.13)</td><td char="." /><td char="." /></tr><tr><td> Indicative preventive</td><td>1</td><td char=".">1</td><td char=".">0.40 (−0.85, 1.65)</td><td char=".">0.55 (−0.85, 1.94)</td><td char="." /><td char="." /></tr><tr><td> Curative</td><td>28</td><td char=".">297</td><td char=".">0.13 (−0.08, 0.34)</td><td char=".">0.28 (−0.37, 0.92)</td><td char="." /><td char="." /></tr><tr><td>Program context</td><td /><td char="." /><td char="." /><td char="." /><td char=".">1.45 (3, 403)</td><td char=".">.228</td></tr><tr><td> Child welfare</td><td>18</td><td char=".">144</td><td char=".">−0.02 (−0.26, 0.23)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Mental health care</td><td>10</td><td char=".">183</td><td char=".">0.39 (0.07, 0.71) *</td><td char=".">0.41 (0.01, 0.81) *</td><td char="." /><td char="." /></tr><tr><td> Law enforcement</td><td>1</td><td char=".">10</td><td char=".">0.25 (−0.21, 0.71)</td><td char=".">0.27 (−0.25, 0.79)</td><td char="." /><td char="." /></tr><tr><td> Community / school</td><td>5</td><td char=".">70</td><td char=".">0.01 (−0.45, 0.46)</td><td char=".">0.03 (−0.49, 0.54)</td><td char="." /><td char="." /></tr><tr><td>Program duration (sessions)</td><td>7</td><td char=".">47</td><td char=".">0.06 (−0.12, 0.24)</td><td char=".">0.10 (−0.03, 0.05)</td><td char=".">0.23 (1, 45)</td><td char=".">.636</td></tr><tr><td>Program duration (months)</td><td>13</td><td char=".">232</td><td char=".">0.04 (−0.13, 0.20)</td><td char=".">−0.00 (−0.03, 0.03)</td><td char=".">0.07 (1, 230)</td><td char=".">.787</td></tr><tr><td>YIM program</td><td /><td char="." /><td char="." /><td char="." /><td char=".">6.51 (1, 407)</td><td char=".">.011</td></tr><tr><td> Not YIM</td><td>26</td><td char=".">219</td><td char=".">−0.03 (−0.22, 0.17)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> YIM program</td><td>9</td><td char=".">190</td><td char=".">0.46 (0.14, 0.77) **</td><td char=".">0.48 (0.11, 0.86) *</td><td char="." /><td char="." /></tr><tr><td>FGDM program</td><td /><td char="." /><td char="." /><td char="." /><td char=".">2.25 (1, 407)</td><td char=".">.134</td></tr><tr><td> Not FGDM</td><td>17</td><td char=".">329</td><td char=".">0.24 (−0.01, 0.49)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> FGDM program</td><td>18</td><td char=".">80</td><td char=".">−0.03 (−0.28, 0.22)</td><td char=".">−0.27 (−0.62, 0.08)</td><td char="." /><td char="." /></tr><tr><td>Multi-family program</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.48 (1, 407)</td><td char=".">.488</td></tr><tr><td> Not multi-family</td><td>31</td><td char=".">340</td><td char=".">0.13 (−0.06, 0.33)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Multi-family program</td><td>4</td><td char=".">69</td><td char=".">−0.06 (−0.57, 0.45)</td><td char=".">−0.19 (−0.74, 0.35)</td><td char="." /><td char="." /></tr><tr><td>Youth decides</td><td /><td char="." /><td char="." /><td char="." /><td char=".">7.11 (1, 379)</td><td char=".">.008</td></tr><tr><td> Youth do not decide</td><td>26</td><td char=".">219</td><td char=".">−0.03 (−0.23, 0.17)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Youth decide</td><td>8</td><td char=".">162</td><td char=".">0.52 (0.17, 0.86) **</td><td char=".">0.54 (0.14, 0.94) **</td><td char="." /><td char="." /></tr><tr><td>Extended family involved</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.20 (1, 407)</td><td char=".">.655</td></tr><tr><td> Extended family not involved</td><td>5</td><td char=".">70</td><td char=".">0.01 (−0.47, 0.48)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Extended family involved</td><td>30</td><td char=".">339</td><td char=".">0.13 (−0.07, 0.32)</td><td char=".">0.12 (−0.40, 0.63)</td><td char="." /><td char="." /></tr><tr><td>School involved</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.11 (1, 407)</td><td char=".">.745</td></tr><tr><td> School not involved</td><td>5</td><td char=".">36</td><td char=".">0.04 (−0.44, 0.51)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> School involved</td><td>30</td><td char=".">373</td><td char=".">0.12 (−0.08, 0.32)</td><td char=".">0.09 (−0.43, 0.60)</td><td char="." /><td char="." /></tr><tr><td>Peers involved</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.58 (1, 407)</td><td char=".">.446</td></tr><tr><td> Peers not involved</td><td>8</td><td char=".">177</td><td char=".">−0.02 (−0.39, 0.35)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Peers involved</td><td>27</td><td char=".">232</td><td char=".">0.15 (−0.06, 0.36)</td><td char=".">0.16 (−0.26, 0.59)</td><td char="." /><td char="." /></tr><tr><td>Neighborhood involved</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.02 (1, 407)</td><td char=".">.897</td></tr><tr><td> Neighborhood not involved</td><td>5</td><td char=".">87</td><td char=".">0.08 (−0.39, 0.55)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Neighborhood involved</td><td>30</td><td char=".">322</td><td char=".">0.11 (−0.09, 0.31)</td><td char=".">0.03 (−0.48, 0.56)</td><td char="." /><td char="." /></tr><tr><td>Number of people involved</td><td /><td char="." /><td char="." /><td char="." /><td char=".">7.60 (1, 407)</td><td char=".">.006</td></tr><tr><td> Multiple people</td><td>28</td><td char=".">280</td><td char=".">−0.01 (−0.20, 0.17)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> One person</td><td>7</td><td char=".">129</td><td char=".">0.56 (0.19, 0.92) **</td><td char=".">0.57 (0.16, 0.98) **</td><td char="." /><td char="." /></tr><tr><td> % successful involvement</td><td>19</td><td char=".">260</td><td char=".">0.16 (−0.21, 0.54)</td><td char=".">0.76 (−0.53, 2.06)</td><td char=".">1.35 (1, 258)</td><td char=".">.246</td></tr><tr><td><bold>Sample characteristics</bold></td><td /><td char="." /><td char="." /><td char="." /><td char="." /><td char="." /></tr><tr><td>Continent</td><td /><td char="." /><td char="." /><td char="." /><td char=".">4.34 (1, 407)</td><td char=".">.038</td></tr><tr><td> North-America</td><td>25</td><td char=".">305</td><td char=".">−0.01 (−0.21, 0.20)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Europe</td><td>10</td><td char=".">104</td><td char=".">0.40 (0.07, 0.73) *</td><td char=".">0.41 (0.02, 0.80) *</td><td char="." /><td char="." /></tr><tr><td>Level of risk</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.13 (1, 407)</td><td char=".">.718</td></tr><tr><td> Low risk</td><td>3</td><td char=".">10</td><td char=".">0.00 (−0.63, 0.63)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> High risk</td><td>32</td><td char=".">399</td><td char=".">0.12 (−0.07, 0.31)</td><td char=".">0.12 (−0.54, 0.78)</td><td char="." /><td char="." /></tr><tr><td>Juvenile offenders</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.17 (1, 407)</td><td char=".">.679</td></tr><tr><td> No juvenile offenders</td><td>31</td><td char=".">336</td><td char=".">0.12 (−0.07, 0.32)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Juvenile offenders</td><td>4</td><td char=".">73</td><td char=".">0.00 (−0.52, 0.53)</td><td char=".">−0.12</td><td char="." /><td char="." /></tr><tr><td>Mental health needs</td><td /><td char="." /><td char="." /><td char="." /><td char=".">14.01 (1, 407)</td><td char="."><.001</td></tr><tr><td> No mental health needs</td><td>29</td><td char=".">357</td><td char=".">−0.03 (−0.19, 0.14)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Mental health needs</td><td>6</td><td char=".">52</td><td char=".">0.75 (0.37, 1.12) ***</td><td char=".">0.78 (0.37, 1.18) ***</td><td char="." /><td char="." /></tr><tr><td>Gender (% boys)</td><td>25</td><td char=".">305</td><td char=".">0.19 (−0.08, 0.47)</td><td char=".">−0.01 (−0.49, 0.48)</td><td char=".">0.00 (1, 303)</td><td char=".">.974</td></tr><tr><td>Ethnicity (% ethnic minority)</td><td>24</td><td char=".">325</td><td char=".">0.03 (−0.06, 0.12)</td><td char=".">−0.21 (−0.51, 0.10)</td><td char=".">1.80 (1, 323)</td><td char=".">.180</td></tr><tr><td>Age</td><td>24</td><td char=".">192</td><td char=".">0.33 (0.01, 0.65) *</td><td char=".">0.04 (−0.02, 0.10)</td><td char=".">2.08 (1, 190)</td><td char=".">.151</td></tr><tr><td>SES</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.77 (1, 166)</td><td char=".">.381</td></tr><tr><td> Average or high SES</td><td>4</td><td char=".">33</td><td char=".">0.14 (−0.13, 0.40)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Low SES</td><td>10</td><td char=".">135</td><td char=".">0.00 (−0.17, 0.16)</td><td char=".">−0.14 (−0.45, 0.17)</td><td char="." /><td char="." /></tr><tr><td>Family composition (% intact)</td><td>5</td><td char=".">123</td><td char=".">0.03 (−0.10, 0.15)</td><td char=".">0.54 (−0.36, 1.44)</td><td char=".">1.42 (1, 121)</td><td char=".">.235</td></tr><tr><td>Assessment characteristics</td><td /><td char="." /><td char="." /><td char="." /><td char="." /><td char="." /></tr><tr><td>Outcome domain</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.48 (5, 403)</td><td char=".">.793</td></tr><tr><td> Academic/work functioning</td><td>7</td><td char=".">32</td><td char=".">0.17 (−0.05, 0.38)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Externalizing problems</td><td>12</td><td char=".">74</td><td char=".">0.14 (−0.07, 0.34)</td><td char=".">−0.03 (−0.18, 0.12)</td><td char="." /><td char="." /></tr><tr><td> Family functioning/child safety</td><td>19</td><td char=".">119</td><td char=".">0.12 (−0.08, 0.31)</td><td char=".">−0.05 (−0.21, 0.10)</td><td char="." /><td char="." /></tr><tr><td> Physical health</td><td>1</td><td char=".">6</td><td char=".">0.08 (−0.22, 0.38)</td><td char=".">−0.09 (−0.34, 0.17)</td><td char="." /><td char="." /></tr><tr><td> Psychological problems</td><td>16</td><td char=".">128</td><td char=".">0.08 (−0.12, 0.27)</td><td char=".">−0.09 (−0.23, 0.05)</td><td char="." /><td char="." /></tr><tr><td> Social</td><td>9</td><td char=".">50</td><td char=".">0.11 (−0.10, 0.31)</td><td char=".">−0.06 (−0.21, 0.08)</td><td char="." /><td char="." /></tr><tr><td>Assessment type</td><td /><td char="." /><td char="." /><td char="." /><td char=".">5.03 (2, 406)</td><td char=".">.007</td></tr><tr><td> Questionnaire</td><td>16</td><td char=".">258</td><td char=".">0.10 (−0.10, 0.29)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Interview</td><td>4</td><td char=".">60</td><td char=".">−0.08 (−0.30, 0.14)</td><td char=".">−0.18 (−0.34, −0.02) *</td><td char="." /><td char="." /></tr><tr><td> Official record</td><td>22</td><td char=".">91</td><td char=".">0.15 (−0.05, 0.33)</td><td char=".">0.05 (−0.08, 0.18)</td><td char="." /><td char="." /></tr><tr><td>Information source</td><td /><td char="." /><td char="." /><td char="." /><td char=".">4.06 (5, 403)</td><td char=".">.001</td></tr><tr><td> Youth</td><td>10</td><td char=".">213</td><td char=".">−0.06 (−0.26, 0.14)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Parents</td><td>8</td><td char=".">44</td><td char=".">0.17 (−0.06, 0.40)</td><td char=".">0.23 (0.06, 0.41) *</td><td char="." /><td char="." /></tr><tr><td> School staff</td><td>3</td><td char=".">38</td><td char=".">0.06 (−0.22, 0.33)</td><td char=".">0.12 (−0.12, 0.36)</td><td char="." /><td char="." /></tr><tr><td> Professionals</td><td>5</td><td char=".">20</td><td char=".">0.34 (0.07, 0.60) *</td><td char=".">0.40 (0.16, 0.64) **</td><td char="." /><td char="." /></tr><tr><td> Combination</td><td>3</td><td char=".">5</td><td char=".">0.01 (−0.36, 0.37)</td><td char=".">0.07 (−0.26, 0.39)</td><td char="." /><td char="." /></tr><tr><td> Official record</td><td>21</td><td char=".">89</td><td char=".">0.15 (−0.04, 0.34)</td><td char=".">0.21 (0.09, 0.33) ***</td><td char="." /><td char="." /></tr><tr><td>Time of assessment</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.09 (1, 403)</td><td char=".">.770</td></tr><tr><td> Post-test</td><td>25</td><td char=".">193</td><td char=".">0.10 (−0.09, 0.29)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Follow-up</td><td>17</td><td char=".">212</td><td char=".">0.11 (−0.08, 0.31)</td><td char=".">0.01 (−0.08, 0.11)</td><td char="." /><td char="." /></tr><tr><td>Number of weeks after the intervention ended</td><td>32</td><td char=".">330</td><td char=".">0.14 (−0.08, 0.35)</td><td char=".">0.00 (−0.00, 0.00)</td><td char=".">0.33 (1, 328)</td><td char=".">.565</td></tr><tr><td>Study quality characteristics</td><td /><td char="." /><td char="." /><td char="." /><td char="." /><td char="." /></tr><tr><td> Year of publication</td><td>35</td><td char=".">409</td><td char=".">0.11 (−0.07, 0.28)</td><td char=".">0.02 (−0.00, 0.05)</td><td char=".">2.74 (1, 407)</td><td char=".">.098</td></tr><tr><td>Peer-reviewed</td><td /><td char="." /><td char="." /><td char="." /><td char=".">1.37 (1, 407)</td><td char=".">.243</td></tr><tr><td> Not peer-reviewed</td><td>5</td><td char=".">65</td><td char=".">0.02 (−0.21, 0.25)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Peer-reviewed</td><td>30</td><td char=".">344</td><td char=".">0.12 (−0.06, 0.30)</td><td char=".">0.10 (−0.07, 0.27)</td><td char="." /><td char="." /></tr><tr><td>Journal impact factor</td><td>30</td><td char=".">344</td><td char=".">0.11 (−0.09, 0.31)</td><td char=".">0.02 (−0.03, 0.06)</td><td char=".">0.46 (1, 342)</td><td char=".">.497</td></tr><tr><td>Q-rank</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.00 (1, 341)</td><td char=".">.987</td></tr><tr><td> Q1</td><td>14</td><td char=".">217</td><td char=".">0.01 (−0.12, 0.13)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Q2</td><td>15</td><td char=".">126</td><td char=".">0.01 (−0.17, 0.17)</td><td char=".">0.00 (−0.11, 0.12)</td><td char="." /><td char="." /></tr><tr><td>Sample size</td><td>35</td><td char=".">409</td><td char=".">0.12 (−0.06, 0.30)</td><td char=".">−0.00 (−0.00, 0.00)</td><td char=".">0.85 (1, 407)</td><td char=".">.357</td></tr><tr><td>% of non-response</td><td>17</td><td char=".">304</td><td char=".">0.00 (−0.09, 0.09)</td><td char=".">0.09 (−0.16, 0.34)</td><td char=".">0.50 (1, 302)</td><td char=".">.482</td></tr><tr><td>Study design (trial)</td><td /><td char="." /><td char="." /><td char="." /><td char=".">2.75 (1, 407)</td><td char=".">.098</td></tr><tr><td> Quasi-experimental</td><td>16</td><td char=".">41</td><td char=".">0.28 (0.01, 0.54) *</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Randomized-controlled trial</td><td>19</td><td char=".">368</td><td char=".">−0.02 (−0.25, 0.21)</td><td char=".">−0.30 (−0.64, 0.06)</td><td char="." /><td char="." /></tr><tr><td>Study design (prospective)</td><td /><td char="." /><td char="." /><td char="." /><td char=".">1.56 (1, 405)</td><td char=".">.213</td></tr><tr><td> Prospective</td><td>25</td><td char=".">391</td><td char=".">0.07 (−0.12, 0.26)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Retrospective</td><td>9</td><td char=".">16</td><td char=".">0.27 (−0.04, 0.56)</td><td char=".">0.20 (−0.11, 0.51)</td><td char="." /><td char="." /></tr><tr><td>Effect size corrected for pretest differences</td><td /><td char="." /><td char="." /><td char="." /><td char=".">10.72 (1, 406)</td><td char=".">.001</td></tr><tr><td> Not corrected</td><td>22</td><td char=".">232</td><td char=".">0.03 (−0.16, 0.22)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Corrected</td><td>13</td><td char=".">176</td><td char=".">0.27 (0.06, 0.47) *</td><td char=".">0.24 (0.10, 0.38) **</td><td char="." /><td char="." /></tr><tr><td>Analysis method</td><td /><td char="." /><td char="." /><td char="." /><td char=".">0.85 (1, 326)</td><td char=".">.358</td></tr><tr><td> Completer</td><td>15</td><td char=".">87</td><td char=".">0.08 (−0.15, 0.32)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Intention-to-treat</td><td>13</td><td char=".">241</td><td char=".">0.17 (−0.06, 0.41)</td><td char=".">0.09 (−0.11, 0.29)</td><td char="." /><td char="." /></tr><tr><td>Control condition</td><td /><td char="." /><td char="." /><td char="." /><td char=".">1.35 (1, 407)</td><td char=".">.246</td></tr><tr><td> No care</td><td>5</td><td char=".">93</td><td char=".">0.02 (−0.21, 0.25)</td><td char="." /><td char="." /><td char="." /></tr><tr><td> Care as usual</td><td>30</td><td char=".">316</td><td char=".">0.12 (−0.06, 0.30)</td><td char=".">0.10 (−0.07, 0.27)</td><td char="." /><td char="." /></tr></tbody></table> </ephtml> </p> <ulist> <item>4 <emph>Note</emph>. # samples: number of samples; # ES: number of effect sizes; β<subs>0</subs>: intercept; Mean <emph>d</emph>: mean effect size (Cohen's <emph>d</emph>); <emph>CI</emph>: Confidence Interval; β<subs>1</subs>: Regression coefficient; <emph>F: F</emph>-statistic (omnibus test); <emph>df</emph>: degrees of freedom; <emph>p: p</emph>-value of the omnibus test. The number of samples and effect sizes do not always add up to the total number of samples and effect sizes, due to missing information in the included studies.</item> <item>5 *<emph>p</emph> <.05; **<emph>p</emph> <.01; ***<emph>p</emph> <.001.</item> </ulist> <hd id="AN0186283553-28">Multiple moderator model</hd> <p>We tested a multiple moderator model to determine the unique contribution of the significant moderators in the bivariate models. We excluded the variable official record as an information source, as it overlaps with the variable official record as an assessment type.</p> <p>The omnibus test was significant, indicating that the moderators significantly explained the heterogeneity in intervention effects, <emph>F</emph>(<reflink idref="bib12" id="ref156">12</reflink>, 367) = 4.46, <emph>p</emph> <.001. Yet, there was still significant residual heterogeneity, <emph>p</emph> <.001. Only one individual moderator was significant, that is, whether effect sizes were corrected for pretest differences between conditions, <emph>p</emph> =.020. None of the other moderators were significant in this model, <emph>p</emph>s ≥.059. See Table 5 for the results of the multiple moderator model.</p> <p>Table 5. Results of the multiple moderator model.</p> <p> <ephtml> <table><thead><tr><td>Moderator variables</td><td>β (<italic>SE</italic>)</td><td>95% <italic>CI</italic></td><td><italic>t</italic>-statistic</td></tr></thead><tbody valign="top"><tr><td>Intercept</td><td char=".">−0.10 (0.10)</td><td char=".">−0.30, 0.10</td><td char=".">−0.97</td></tr><tr><td>YIM program</td><td char=".">−0.03 (0.36)</td><td char=".">−0.73, 0.68</td><td char=".">−0.07</td></tr><tr><td>One person involved</td><td char=".">0.35 (0.47)</td><td char=".">−0.57, 1.26</td><td char=".">0.75</td></tr><tr><td>Country Europe</td><td char=".">0.20 (0.21)</td><td char=".">−0.21, 0.62</td><td char=".">0.96</td></tr><tr><td>Mental health needs</td><td char=".">0.68 (0.36)</td><td char=".">−0.03, 1.38</td><td char=".">1.89</td></tr><tr><td>Assessment type: Questionnaire</td><td char=".">0.06 (0.32)</td><td char=".">−0.58, 0.69</td><td char=".">0.18</td></tr><tr><td>Assessment type: Interview</td><td char=".">−0.11 (0.33)</td><td char=".">--0.76, 0.53</td><td char=".">−0.35</td></tr><tr><td>Information source: Youth</td><td char=".">−0.25 (0.32)</td><td char=".">−0.88, 0.38</td><td char=".">−0.78</td></tr><tr><td>Information source: Parent</td><td char=".">−0.09 (0.33)</td><td char=".">−0.74, 0.55</td><td char=".">−0.28</td></tr><tr><td>Information source: School staff</td><td char=".">−0.13 (0.34)</td><td char=".">−0.80, 0.53</td><td char=".">−0.40</td></tr><tr><td>Information source: Professionals</td><td char=".">−0.06 (0.34)</td><td char=".">−0.73, 0.60</td><td char=".">−0.18</td></tr><tr><td>Information source: Combination</td><td char=".">−0.25 (0.29)</td><td char=".">−0.81, 0.31</td><td char=".">−0.88</td></tr><tr><td>Effect size corrected for pretest differences</td><td char=".">0.22 (0.09) *</td><td char=".">0.04, 0.40</td><td char=".">2.34</td></tr><tr><td><italic>F</italic> (<italic>df</italic><sub>1</sub>, <italic>df</italic><sub>2</sub>)</td><td>4.45 (12, 367) ***</td><td /></tr><tr><td>σ<sup>2</sup><sub>level 2</sub></td><td char=".">0.07***</td><td char="." /><td char="." /></tr><tr><td>σ<sup>2</sup><sub>level 3</sub></td><td char=".">0.17***</td><td char="." /><td char="." /></tr></tbody></table> </ephtml> </p> <ulist> <item>6 <emph>Note.</emph> β: estimated regression coefficient; <emph>SE</emph>: standard error; <emph>CI</emph>: confidence interval; <emph>F:</emph><emph>F</emph>-statistic (omnibus test); <emph>df</emph>: degrees of freedom; σ<sups>2</sups><subs>level 2</subs>: variance between effect sizes extracted from the same study; σ<sups>2</sups><subs>level 3</subs>: variance between studies. Two variables were ignored in the analysis, i.e. <emph>youth decides</emph> and <emph>assessment type: official record</emph>, therefore, the results of these variables are not presented.</item> <item>7 *<emph>p</emph> <.05; **<emph>p</emph> <.01; ***<emph>p</emph> <.001.</item> </ulist> <hd id="AN0186283553-29">Discussion</hd> <p>The aim of this three-level meta-analysis was to gain insight in the overall effectiveness of youth interventions that activate the social network, and the impact of potential moderating variables. This study is the first comprehensive meta-analysis including all types of youth interventions that activate the social network to improve youth outcomes, i.e. academic and work functioning, externalizing problems, psychological problems, family functioning and child safety, and social functioning. This meta-analysis represents a synthesis of 37 studies with 35 independent samples and 409 effect sizes. Overall, we found that interventions in which the social network was activated were not more effective than interventions without this component. Moderator analyses revealed eight significant moderators of intervention effectiveness, yielding larger effects for youth-initiated mentoring interventions, youth deciding who to involve, interventions that involve only one person from the social network, European samples, interventions targeting youth with mental health needs, data retrieved through questionnaires and official records, assessments completed by professionals or parents, and outcomes that were corrected for pretest differences between conditions</p> <p>Although we expected larger effects for interventions with a social network component, the lack of an overall effect for youth interventions that activate the social network is in line with various previous meta-analyses that found that such interventions yielded either small (Dantzer & Perry, [<reflink idref="bib18" id="ref157">18</reflink>]; MacLeod & Nelson, [<reflink idref="bib63" id="ref158">63</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref159">108</reflink>]) or no effects (Dijkstra et al., [<reflink idref="bib23" id="ref160">23</reflink>]; McGinn et al., [<reflink idref="bib66" id="ref161">66</reflink>]; Valentine et al., [<reflink idref="bib107" id="ref162">107</reflink>]). Yet, despite the lack of an overall effect, we found several conditions under which interventions with social network components were effective, suggesting that the effectiveness of such interventions depends on the way the intervention is implemented, the target group at which the intervention is directed, and the way the effects are measured.</p> <hd id="AN0186283553-30">Moderators of intervention effects</hd> <p>Moderator analyses revealed three intervention characteristics that moderated intervention effects. Similar to recent meta-analyses on youth-initiated mentoring (Dantzer & Perry, [<reflink idref="bib18" id="ref163">18</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref164">108</reflink>]), we found that youth-initiated mentoring interventions had a small to medium effect, whereas other forms of interventions were not effective in promoting positive youth outcomes. Youth-initiated mentoring is a child-centered program, and promotes youth to have close contact with one person (van Dam et al., [<reflink idref="bib108" id="ref165">108</reflink>]). This method of social network activation can stimulate the youth's feelings of relatedness (Ryan & Deci, [<reflink idref="bib86" id="ref166">86</reflink>]) and social resourcefulness (Schwartz et al., [<reflink idref="bib89" id="ref167">89</reflink>]), which, in turn, are related to positive intervention outcomes (Chan et al., [<reflink idref="bib14" id="ref168">14</reflink>]; Rhodes et al., [<reflink idref="bib81" id="ref169">81</reflink>]).</p> <p>The finding that youth-initiated mentoring interventions were more effective than other interventions activating the social network may also be explained by two other moderators of intervention effects, that is, whether youth decide who to involve, and how many people are involved in the intervention. Intervention effects were larger for interventions in which youth decide who to involve, and for interventions in which just one person was involved, which are both the case in youth-initiated mentoring and less likely in other forms of interventions. Autonomy may explain why interventions in which youth decide themselves who to involve were more effective than interventions in which other parties decide. Youth are more likely to experience autonomy if they are given the freedom to decide who to involve in the intervention. Autonomy is linked to increased treatment motivation, which subsequently contributes to better intervention effects (Ryan & Deci, [<reflink idref="bib85" id="ref170">85</reflink>]). Additionally, the influence of the social network may also be greater if youth can decide who te involve. That is, in the realm of social influence, the impact of others can vary based on factors such as proximity, similarity, and relationship closeness (Cialdini, [<reflink idref="bib16" id="ref171">16</reflink>]). Youth are most likely to choose someone to with whom they have a strong bond and who is similar (Koper et al., [<reflink idref="bib53" id="ref172">53</reflink>]), therefore enhancing the potential of positive social influence. In most intervention types included in this study the parents—not the youth—choose who to involve, which may explain why we only found intervention effects on youth outcomes for youth-initiated mentoring programs.</p> <p>Youth-initiated mentoring interventions involve just one person, which was found to be more effective than involving multiple people in social network interventions. This may be explained through the mechanism of diffusion of responsibility (Darley & Latane, [<reflink idref="bib19" id="ref173">19</reflink>]; Fischer et al., [<reflink idref="bib33" id="ref174">33</reflink>]). Research indeed indicated that youth-initiated mentors aim to be pro-active in their role, and aim to act in the best interest of the youth (Koper et al., [<reflink idref="bib53" id="ref175">53</reflink>]). Another possible explanation for the effectiveness of interventions that involve just one person is the concept of collective intelligence, which is the general ability of a group of people to perform well in different tasks (Woolley et al., [<reflink idref="bib123" id="ref176">123</reflink>]), and has been found to be present in small groups of two to five people (Woolley et al., [<reflink idref="bib123" id="ref177">123</reflink>], [<reflink idref="bib122" id="ref178">122</reflink>]).</p> <p>We found that intervention effects were larger in certain samples. Intervention effects were small but significant in European samples, whereas there were no significant intervention effects in North-American samples. This is surprising given that related meta-analyses (Dijkstra et al., [<reflink idref="bib23" id="ref179">23</reflink>]; van Dam et al., [<reflink idref="bib108" id="ref180">108</reflink>]) did not find differences in effectiveness between continents, and therefore warrants caution in interpretation of this effect. The difference could <emph>potentially</emph> be explained by social processes which can vary between regions due to a range of factors, including cultural, historical, and institutional influences (e.g. Hall, [<reflink idref="bib37" id="ref181">37</reflink>]; Hofstede, [<reflink idref="bib42" id="ref182">42</reflink>]), and differences in living conditions because of relatively high poverty rates in the United States compared to many European countries (e.g. OECD, [<reflink idref="bib72" id="ref183">72</reflink>]). Future research should focus on pinpointing in what (cultural) environment interventions with a social network component work best.</p> <p>Interventions yielded medium to large effects in samples of youth with mental health needs, whereas interventions were not effective in other samples. This finding is in line with other meta-analyses demonstrating that intervention effects are generally larger if problems are more severe at the start of the program (e.g. Stice et al., [<reflink idref="bib99" id="ref184">99</reflink>]; van Loon et al., [<reflink idref="bib113" id="ref185">113</reflink>]). However, our results indicated that this did not hold for the level of risk in general: Neither overall risk level nor the nature of the intervention (i.e. preventative vs. curative interventions) moderated effectiveness. Moreover, intervention effects did not vary depending on whether juvenile offenders were present in the sample. Thus, interventions in which the social network is activated seem especially beneficial for youth with mental health needs, as opposed to at-risk populations in general. There are two possible explanations for this finding, namely the importance of social connectedness for mental health, and overlap with other moderators. Research has repeatedly confirmed the importance of social support and connectedness for mental health (Harandi et al., [<reflink idref="bib38" id="ref186">38</reflink>]), including youth mental health (e.g. DuBois & Silverthorn, [<reflink idref="bib25" id="ref187">25</reflink>]; Sterrett et al., [<reflink idref="bib98" id="ref188">98</reflink>]; van Dam, Smit, et al., [<reflink idref="bib111" id="ref189">111</reflink>]). Thus, it can be expected that youth with mental health needs in particular benefit from the social support offered in interventions with a social network component. Alternatively, the effect could be explained by the overlap with another moderator. That is, all but one study in samples with youth with mental health needs examined the effectiveness of youth-initiated mentoring interventions, which proved to be the only effective intervention type in this meta-analysis.</p> <p>Our moderator analyses revealed several assessment characteristics that impacted the effectiveness of interventions with a social network component. First, results indicated that both assessment type and information source were moderators of intervention effects. Effects were smaller if data were retrieved through interviews than through questionnaires and official records, which was contrary to our hypothesis. Yet, all effects were (very) small. Second, intervention effects were larger if assessments were completed by professionals or parents, or if data were retrieved from official records, compared to assessments completed by youth, school staff, or a combination of information sources. The effects were small but significant for assessments completed by professionals, and very small for parent-reported data and official records. These differences may be explained by youth wanting to give a favorable presentation of themselves (Breuk et al., [<reflink idref="bib12" id="ref190">12</reflink>]), and school staff showing a tendency to give a favorable presentation of their students with psychological or behavioral problems in order to avoid negative labeling (Stams et al., [<reflink idref="bib97" id="ref191">97</reflink>]), thereby underestimating problems at pretest assessments, which unduly reduces the chance of finding intervention effects at post- or follow-up assessments. On the contrary, helping professionals and parents may overestimate positive intervention effects, because positive treatment outcomes may serve positive and rewarding professional self-evaluations, and give parents feelings of hope and expectancy, and motivation to continue their efforts and intervention.</p> <p>Finally, intervention effects were small but significant if effect sizes were corrected for pretest differences between the intervention and control conditions, whereas there was no effect if the effect size was not corrected. This may seem contradictory, since we could expect that more robust designs in which effect sizes are corrected yield smaller effects (Farrington, [<reflink idref="bib30" id="ref192">30</reflink>]). Yet, Shadish et al. ([<reflink idref="bib92" id="ref193">92</reflink>]) argue that selection effects that can be present in non-randomized designs (Hariton & Locascio, [<reflink idref="bib39" id="ref194">39</reflink>]) may explain why effect sizes were larger if outcomes were corrected for pretest differences. That is, if youth in the intervention condition experience more problems before starting care than youth in control conditions, and these differences are not corrected in effect size estimations, this causes bias and underestimation of effects (Hariton & Locascio, [<reflink idref="bib39" id="ref195">39</reflink>]).</p> <hd id="AN0186283553-31">Strengths, limitations, and future directions</hd> <p>Our meta-analysis has several noteworthy strengths. First, applying a multi-level approach to meta-analysis is a strong method using all available data while dealing with dependency of effect sizes (Assink & Wibbelink, [<reflink idref="bib5" id="ref196">5</reflink>]). Effect sizes that stem from the same study sample are likely dependent, as they are obtained in a similar context and with the same study procedures. Ignoring the correlation between effect sizes may lead to flawed inferences due to underestimation of standard errors, which in turn increases the likelihood of false positives (Hedges & Olkin, [<reflink idref="bib41" id="ref197">41</reflink>]). Three-level models properly accounted for the dependency among effect sizes within studies (Assink & Wibbelink, [<reflink idref="bib5" id="ref198">5</reflink>]). Additionally, we used publication bias tests appropriate for multi-level meta-analyses (Fernández-Castilla et al., [<reflink idref="bib31" id="ref199">31</reflink>], [<reflink idref="bib32" id="ref200">32</reflink>]). Second, the high number of included studies increases the statistical power of this meta-analysis. Hox et al. ([<reflink idref="bib45" id="ref201">45</reflink>]) recommends to include at least 20 studies, which is amply exceeded with an inclusion of 37 studies in this meta-analysis.</p> <p>This meta-analysis also has several noteworthy limitations. First, some studies lacked information on program, sample, and assessment characteristics, including the sample's SES, program duration and program fidelity. This limited the robustness of some moderation analyses and potentially resulted in bias if missing values were not distributed randomly over intervention types (Rubin, [<reflink idref="bib83" id="ref202">83</reflink>]). It is important that future studies investigating intervention effects report sufficient information about the program, sample, and study, in order to further determine what works for whom and under what circumstances (Kraemer et al., [<reflink idref="bib54" id="ref203">54</reflink>]).</p> <p>Second, as some outcomes were only examined in a few studies (e.g. executive functioning, aggression, and substance use), we clustered these with related outcomes in broad domains to ensure robust moderation analyses. This may have resulted in more heterogeneity in some domains compared to others, and has limited our knowledge on which specific outcomes can be significantly improved by social network activation in interventions.</p> <p>Third, it was not possible to meaningfully test for moderation by developmental stage, as many studies included youth in different developmental stages (e.g. programs offered to families with infants, children, and adolescents). Especially FGDM is often employed across a broad age range of the children because its focus is on the family unit. FGDM is designed to address family dynamics and decision-making processes within the family unit. As a result, the developmental stage of the children in the family system does not matter for the implementation of the intervention. Thus, although moderation analyses showed that intervention effectiveness did not vary by the sample's mean age, we were unable to determine the potential moderation effect of the developmental stage.</p> <p>Fourth, detailed information on the control groups was lacking in many studies. That is, studies reported if the control groups received CAU or no care, but it was often unclear what CAU entailed, how CAU differed from the intervention condition, and how high treatment fidelity was. This limits our certainty that the CAU control group interventions did not activate the social network, which could potentially result in biased outcomes. However, we found no moderating effect of the type of control condition (CAU vs. no care), giving us confidence in the accuracy of our results.</p> <p>Fifth and finally, although almost half of the included studies included follow-up assessments, not many followed up on youth for longer than two years (i.e. 17 effect sizes). By including long-term assessments, potential sleeper effects (i.e. effects that increase further over time) could be observed. Long-term assessments are especially relevant for interventions in which the social network is activated, as it is hypothesized that the permanence of support from social network members after the intervention has ended (Zimmerman, [<reflink idref="bib124" id="ref204">124</reflink>]) could enhance resilience (Ozbay et al., [<reflink idref="bib75" id="ref205">75</reflink>]; Southwick et al., [<reflink idref="bib95" id="ref206">95</reflink>]; Ungar, [<reflink idref="bib106" id="ref207">106</reflink>]) and thereby help maintain or even enhance the positive short-term effects of the intervention. Therefore, future studies should include more follow-up assessments to examine the effects of interventions that activate the social network over longer periods of time.</p> <hd id="AN0186283553-32">Conclusion</hd> <p>This meta-analysis showed that, overall, youth interventions that activate the social network do not outperform care as usual in improving youth outcomes. However, youth-initiated mentoring interventions, interventions in which youth decide who to involve, and interventions involving only one person from the social network, showed positive outcomes. Additionally, interventions that activate the social network were more effective in European samples and youth with mental health needs, as well as in studies in which data were collected through questionnaires or official records, or in which assessments were completed by professionals or parents. Thus, the way in which the social network is activated in interventions matters for its effectiveness. According to our findings, youth-initiated mentoring seems the most promising method of social network activation in interventions to promote positive youth outcomes. It seems that interventions that aim to promote positive youth outcomes by activating the social network should do so by giving youth autonomy to select who to involve, and by involving only one person to enhance intervention effects.</p> <hd id="AN0186283553-33">Acknowledgements</hd> <p>We thank dr. Jesse Roest for his help with performing the publication bias analyses.</p> <hd id="AN0186283553-34">Authors contributions</hd> <p>NK, RvdH, SD, HC, LvD, SB and GJS contributed to the design of the study. NK, RvdH and SD performed the literature search and selected studies for inclusion. NK, RvdH, SD and TK were responsible for coding. GJS supervised the processes of the literature, selection and coding. NK wrote the manuscript in close collaboration with all authors. All authors read and approved the final manuscript.</p> <hd id="AN0186283553-35">Disclosure statement</hd> <p>NK and LvD are employed by the YIM Foundation to conduct this research project. LvD is involved in the development and implementation of youth-initiated mentoring in the Netherlands. The other authors declare that they have no competing interests.</p> <hd id="AN0186283553-36">Data availability statement</hd> <p>The final data set and R scripts can be found on OpenScience Framework (OSF): https://osf.io/9vzg2/.</p> <hd id="AN0186283553-37">Appendix A:</hd> <p></p> <hd id="AN0186283553-38">Search strategy</hd> <p>Literature search performed on June 29, 2022 through Ovid electronic databases ERIC and PsycINFO.</p> <hd id="AN0186283553-39">Social network</hd> <p>social network or social networks or communit* or support network or support networks or natural mentor* or informal mentor* or natural youth mentor* or informal youth mentor* or naturally acquired mentor* or naturally occurring mentor* or community mentor* or non-parental adult* or nonparental adult* or peer leader* or school-based mentor* or informal connection* or informal network or informal networks or YIM or youth initiated mentor* or youth-initiated mentor* or youth nominated support team or youth-nominated support team or family group or family team meeting* or family decision making or family decision-making or team decision making or team decisionmaking or family-to-family or family to family or family unity meeting or family team meeting or family meeting or FGC or FACT or flexible assertive community treatment or assertive community treatment or SNAP or "Stop Now And Plan" or National Guard Youth Challenge Program or New Perspectives or family finding or assertive continuing care or family critical time intervention or social capital intervention</p> <hd id="AN0186283553-40">Intervention</hd> <p>intervention* or treatment* or program* or therap* or care or project evaluation</p> <hd id="AN0186283553-41">Target group</hd> <p>newborn* or new-born* or infan* or baby* or babies or toddler* or child* or kid or kids or prepubescen* or prepuberty* or preadolesc* or pubescen* or puberty or teen* or adolesc* or juvenile* or under ag* or underag* or youth* or girl* or boy*</p> <hd id="AN0186283553-42">Research design</hd> <p>RCT* or randomized controlled trial* or randomized-controlled trial* or randomized controlled trial* or randomized-controlled trial* or randomized design or randomized design or experiment* or control group* or control condition* or comparison group or trial* or randomly assigned or random assignment or intent-to-treat*</p> <hd id="AN0186283553-43">Youth outcome element</hd> <p>internal* or anxi* or depress* or stress or external* or aggress* or delinq* or crime* or criminal* or recidiv* or "substance use" or substance abuse or "drug use" or drug abuse or "alcohol use" or alcohol abuse or out-of-home or out of home or wellbeing or well-being or resilien* or school function* or school drop-out or school drop out or academic achievement or truancy or educational outcome* or youth care or child care or youth welfare or child welfare or youth protect* or child protect* or maltreatment or child abuse or mental disorder* or psychological disorder* or psychiatric disorder* or mental illness* or personality disorder* or ADHD or mood disorder* or eating disorder* or symptom* or self-harm* or selfharm* or self harm* or self-injury or self-mutilation or suicid*</p> <hd id="AN0186283553-44">Appendix B</hd> <p></p> <hd id="AN0186283553-45">Codebook</hd> <p>Appendix B</p> <p> <ephtml> <table><thead><tr><td>Category</td><td>Variable</td><td>Label</td></tr></thead><tbody valign="top"><tr><td>Study information </td><td>articleID </td><td>article or report ID</td></tr><tr><td>studyID </td><td>study or sample ID</td></tr><tr><td>effectsizeID </td><td>effect size ID</td></tr><tr><td>Program characteristics </td><td>program_nature</td><td>nature of the program: universal prev. (1), selective prev. (2), indicative prev. (3), or curative (4)</td></tr><tr><td>program_context</td><td>context of the program: child welfare (1), mental health care (2), law enforcement (3), or community/school (4)</td></tr><tr><td>YIM_program</td><td>whether the program includes YIM: no (0), yes (1)</td></tr><tr><td>FGDM_program</td><td>whether the program includes FGDM: no (0), yes (1)</td></tr><tr><td>multifamily_program</td><td>whether the program includes multi-family sessions: no (0), yes (1)</td></tr><tr><td>program_duration </td><td>the duration of the program in months</td></tr><tr><td>program_sessions</td><td>number of sessions in the program</td></tr><tr><td>youth_decides </td><td>whether youth decide who to involve from the social network: no (0), yes (1)</td></tr><tr><td>sn_extendedfamily </td><td>which people from the social network are engaged? extended family (e.g. aunts, uncles, grandparents) : no (0), yes (1)</td></tr><tr><td>sn_school </td><td>which people from the social network are engaged? professionals from school: no (0), yes (1)</td></tr><tr><td>sn_peers </td><td>which people from the social network are engaged? peers or friends: no (0), yes (1)</td></tr><tr><td>sn_neighborhood </td><td>which people from the social network are engaged? neighborhood: no (0), yes (1)</td></tr><tr><td>sn_number</td><td>how many people from the social network are engaged? one person (0) or multiple people (1). If one or more people can be engaged, code as "1". </td></tr><tr><td>sn_executed</td><td>percentage of participants in the experimental group in which engagement of the social network was successful and/or executed as intended</td></tr><tr><td>Sample characteristics</td><td>continent </td><td>continent in which the study took place: North-America (0) or Europe (1)</td></tr><tr><td>risk_level </td><td>low risk (0) or high risk (1) </td></tr><tr><td>juvenileoffenders</td><td>special population: juvenile offender</td></tr><tr><td>mentalhealthneeds</td><td>special population: mental health needs </td></tr><tr><td>perc_male </td><td>percentage male </td></tr><tr><td>ethnic_min </td><td>percentage ethnic minority </td></tr><tr><td>age_mean </td><td>mean age </td></tr><tr><td>SES </td><td>socioeconomic status (SES) of youth and/or parents (scored based on education, job, and income): low (1) or average or high (2)</td></tr><tr><td>fam_intact</td><td>Percentage intact families</td></tr><tr><td>Assessment characteristics</td><td>domain_broad </td><td>outcome domain: 1 = academic/work, 2 = externalizing, 3 = family functioning‎/child safety, 4 = physical health, 5 = psychological, 6 = social </td></tr><tr><td>assessment_type </td><td>asessment type: questionnaire (0), interview (1), or official record (2) </td></tr><tr><td>information_source </td><td>informant: youth (0), parent(s) (1), school (2), combination (3), professional (e.g. psychologist, social worker) (4), official record (5) </td></tr><tr><td>assessment_timing</td><td>post-test (first assessment after intervention; 0), or follow-up (assessments after post-test; 1)</td></tr><tr><td>assessment_weeks</td><td>number of weeks after ending the program</td></tr><tr><td>Study quality characteristics</td><td>Year</td><td>year of publication</td></tr><tr><td>peer_reviewed </td><td>not peer-reviewed (0) or peer-reviewed (1) </td></tr><tr><td>impact_factor </td><td>impact factor as mentioned on journal website</td></tr><tr><td>Q-rank </td><td>Q-rank (2021), 1 = Q1, 2 = Q2, 3 = Q3, 4 = Q4 </td></tr><tr><td>N_exp</td><td>number of participants in experimental group </td></tr><tr><td>N_ctrl</td><td>number of participants in control group </td></tr><tr><td>N_total</td><td>total number of participants (at pretest)</td></tr><tr><td>non-response </td><td>percentage of participants who have not completed the study</td></tr><tr><td>design_RCT</td><td>study design: quasi-experimental (0) or randomized (RCT) (1) </td></tr><tr><td>design_prospective</td><td>study design: retrospective (0) or prospective (1) </td></tr><tr><td>Category</td><td>Variable</td><td>Label</td></tr><tr><td>intention_to_treat </td><td>analyses including drop-outs (intention-to-treat; 1) or excluding drop-outs (completer; 0) (drop-outs = did not complete the program)</td></tr><tr><td>control_condition </td><td>condition of the control group: no care/waitlist (0) or care as usual (CAU; 1) </td></tr><tr><td>Effect sizes </td><td>ES </td><td>effect size </td></tr><tr><td>ES_corrected </td><td>is the variable "effect size" corrected for the effect size at pretest? no (0), yes (1)</td></tr><tr><td>stderr </td><td>standard error</td></tr><tr><td>v </td><td>Variance</td></tr></tbody></table> </ephtml> </p> <hd id="AN0186283553-46">Appendix C</hd> <p></p> <hd id="AN0186283553-47">Forest plot</hd> <p>Graph: Figure C1. Forest plot of effect sizes per independent sample.</p> <ref id="AN0186283553-48"> <title> References </title> <blist> <bibl id="bib1" idref="ref90" type="bt">1</bibl> <bibtext> *References marked with an asterisk indicate studies included in the meta-analysis.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref91" type="bt">2</bibl> <bibtext> Allan, H., Rauktis, M. E., Pennell, J., Merkel-Holguin, L., & Crampton, D. (2021). Family meetings as a system reform to address racial disproportionality and disparities. In Child maltreatment: Contemporary issues in research and policy (Vol. 11, pp. 309 – 338). Springer. https://doi.org/10.1007/978-3-030-54314-3_16</bibtext> </blist> <blist> <bibl id="bib3" idref="ref68" type="bt">3</bibl> <bibtext> American Psychological Association. (2006). Report of the APA Task Force on Socioeconomic Status. https://<ulink href="http://www.apa.org/pi/ses/resources/publications/task-force-2006">www.apa.org/pi/ses/resources/publications/task-force-2006</ulink>. pdf</bibtext> </blist> <blist> <bibl id="bib4" idref="ref21" type="bt">4</bibl> <bibtext> Ashida, S., Sewell, D. K., Schafer, E. J., Schroer, A., & Friberg, J. (2018). Social network members who engage in activities with older adults: Do they bring more social benefits than other members? Ageing and Society, 2018 (5), 1 – 20. https://doi.org/10.1017/S0144686X17001490</bibtext> </blist> <blist> <bibl id="bib5" idref="ref94" type="bt">5</bibl> <bibtext> Assink, M., & Wibbelink, C. J. M. (2016). Fitting three-level meta-analytic models in R: A step-by-step tutorial. The Quantitative Methods for Psychology, 12 (3), 154 – 174. https://doi.org/10.20982/tqmp.12.3.p154</bibtext> </blist> <blist> <bibl id="bib6" type="bt">6</bibl> <bibtext> * Baffour, T. D. (2006). Ethnic and gender differences in offending patterns: Examining family group conferencing interventions among at-risk adolescents. Child and Adolescent Social Work Journal, 23 (5-6), 557 – 578. https://doi.org/10.1007/s10560-006-0075-4</bibtext> </blist> <blist> <bibl id="bib7" idref="ref43" type="bt">7</bibl> <bibtext> Baldwin, S. A., Christian, S., Berkeljon, A., & Shadish, W. R. (2012). The effects of family therapies for adolescent delinquency and substance abuse: A meta-analysis. Journal of Marital and Family Therapy, 38 (1), 281 – 304. https://doi.org/10.1111/J.1752-0606.2011.00248.X</bibtext> </blist> <blist> <bibl id="bib8" idref="ref29" type="bt">8</bibl> <bibtext> Bartelink, C., van Yperen, T. A., & ten Berge, I. J. (2015). Deciding on child maltreatment: A literature review on methods that improve decision-making. Child Abuse & Neglect, 49, 142 – 153. https://doi.org/10.1016/j.chiabu.2015.07.002</bibtext> </blist> <blist> <bibl id="bib9" idref="ref24" type="bt">9</bibl> <bibtext> Berkowitz, A. (2005). An overview of the social norms approach. In L. C. Lederman, L. Stewart, F. Goodhart, & L. Laitman (Eds.), Changing the culture of college drinking: A socially situated prevention campaign. Hampton Press.</bibtext> </blist> <blist> <bibtext> * Berzin, S. C. (2006). Using sibling data to understand the impact of family group decision-making on child welfare outcomes. Children and Youth Services Review, 28 (12), 1449 – 1458. https://doi.org/10.1016/j.childyouth.2006.03.003</bibtext> </blist> <blist> <bibtext> Bijleveld, C. C. J. H., & Commandeur, J. J. F. (2009). Multivariate analyse. Een inleiding voor criminologen en andere sociale wetenschappers. [Multivariate analysis. An introduction for criminologists and other social scientists.] (2nd ed.). Boom Juridische Uitgevers.</bibtext> </blist> <blist> <bibtext> Breuk, R. E., Clauser, C. A. C., Stams, G. J. J. M., Slot, N. W., & Doreleijers, T. A. H. (2007). The validity of questionnaire self-report of psychopathology and parent–child relationship quality in juvenile delinquents with psychiatric disorders. Journal of Adolescence, 30 (5), 761 – 771. https://doi.org/10.1016/J.ADOLESCENCE.2006.10.003</bibtext> </blist> <blist> <bibtext> Brown, J., Cohen, P., Johnson, J. G., & Smailes, E. M. (1999). Childhood abuse and neglect: Specificity of effects on adolescent and young adult depression and suicidality. Journal of the American Academy of Child and Adolescent Psychiatry, 38 (12), 1490 – 1496. https://doi.org/10.1097/00004583-199912000-00009</bibtext> </blist> <blist> <bibtext> Chan, C. S., Rhodes, J. E., Howard, W. J., Lowe, S. R., Schwartz, S. E. O., & Herrera, C. (2013). Pathways of influence in school-based mentoring: The mediating role of parent and teacher relationships. Journal of School Psychology, 51 (1), 129 – 142. https://doi.org/10.1016/j.jsp.2012.10.001</bibtext> </blist> <blist> <bibtext> Cheung, A. C. K., & Slavin, R. E. (2016). How methodological features affect effect sizes in education. Educational Researcher, 45 (5), 283 – 292. https://doi.org/10.3102/0013189X16656615</bibtext> </blist> <blist> <bibtext> Cialdini, R. B. (1993). Influence: Science and practice. (3rd ed.). HarperCollins College Publishers.</bibtext> </blist> <blist> <bibtext> Cohen, J. (1988). Statistical power analysis for the behavioral sciences. In Computers, environment and urban systems (2nd ed.). Lawrence Erlbaum Associates.</bibtext> </blist> <blist> <bibtext> Dantzer, B., & Perry, N. E. (2022). Advancing and mobilizing knowledge about youth-initiated mentoring through community-based participatory research: A scoping review. Youth, 2 (4), 587 – 609. https://doi.org/10.3390/youth2040042</bibtext> </blist> <blist> <bibtext> Darley, J. M., & Latane, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8 (4), 377 – 383. https://doi.org/10.1037/H0025589</bibtext> </blist> <blist> <bibtext> * de Vries, S. L. A., Hoeve, M., Asscher, J. J., & Stams, G. J. J. M. (2018). The long-term effects of the youth crime prevention program "New Perspectives" on delinquency and recidivism. International Journal of Offender Therapy and Comparative Criminology, 62 (12), 3639 – 3661. https://doi.org/10.1177/0306624X17751161</bibtext> </blist> <blist> <bibtext> * de Vries, S. L. A., Hoeve, M., Wibbelink, C. J. M., Asscher, J. J., & Stams, G. J. J. M. (2017). A randomized controlled trial of the effectiveness of the youth crime prevention program 'New Perspectives' (NP): Post-treatment changes and moderator effects. Children and Youth Services Review, 82, 413 – 426. https://doi.org/10.1016/j.childyouth.2017.10.011</bibtext> </blist> <blist> <bibtext> * Dijkstra, S., Asscher, J. J., Deković, M., Stams, G. J. J. M., & Creemers, H. E. (2019). A randomized controlled trial on the effectiveness of family group conferencing in child welfare: Effectiveness, moderators, and level of FGC completion. Child Maltreatment, 24 (2), 137 – 151. https://doi.org/10.1177/1077559518808221</bibtext> </blist> <blist> <bibtext> Dijkstra, S., Creemers, H. E., Asscher, J. J., Deković, M., & Stams, G. J. J. M. (2016). The effectiveness of family group conferencing in youth care: A meta-analysis. Child Abuse & Neglect, 62, 100 – 110. https://doi.org/10.1016/J.CHIABU.2016.10.017</bibtext> </blist> <blist> <bibtext> DuBois, D. L., Portillo, N., Rhodes, J. E., Silverthorn, N., & Valentine, J. C. (2011). How effective are mentoring programs for youth? A systematic assessment of the evidence. Psychological Science in the Public Interest, 12 (2), 57 – 91. https://doi.org/10.1177/1529100611414806</bibtext> </blist> <blist> <bibtext> DuBois, D. L., & Silverthorn, N. (2005). Natural mentoring relationships and adolescent health: Evidence from a national study. American Journal of Public Health, 95 (3), 518 – 524. https://doi.org/10.2105/AJPH.2003.031476</bibtext> </blist> <blist> <bibtext> Duval, S., & Tweedie, R. (2000a). A nonparametric "trim and fill" method of accounting for publication bias in meta-analysis. Journal of the American Statistical Association, 95 (449), 89 – 98. https://doi.org/10.1080/01621459.2000.10473905</bibtext> </blist> <blist> <bibtext> Duval, S., & Tweedie, R. (2000b). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56 (2), 455 – 463. https://doi.org/10.1111/j.0006-341X.2000.00455.x</bibtext> </blist> <blist> <bibtext> * Edwards, M., Tinworth, K., Burford, G., & Pennell, J. (2006). Family team meeting (FTM) process, outcome, and impact evaluation: Phase II report.</bibtext> </blist> <blist> <bibtext> Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315 (7109), 629 – 634. https://doi.org/10.1136/bmj.315.7109.629</bibtext> </blist> <blist> <bibtext> Farrington, D. P. (2003). Methodological quality standards for evaluation research. In Annals of the American academy of political and social science (Vol. 587). Sage. https://doi.org/10.1177/0002716202250789</bibtext> </blist> <blist> <bibtext> Fernández-Castilla, B., Declercq, L., Jamshidi, L., Beretvas, S. N., Onghena, P., & van den Noortgate, W. (2020). Visual representations of meta-analyses of multiple outcomes: Extensions to forest plots, funnel plots, and caterpillar plots. Methodology, 16 (4), 299 – 315. https://doi.org/10.5964/meth.4013</bibtext> </blist> <blist> <bibtext> Fernández-Castilla, B., Declercq, L., Jamshidi, L., Beretvas, S. N., Onghena, P., & Van den Noortgate, W. (2021). Detecting selection bias in meta-analyses with multiple outcomes: A simulation study. The Journal of Experimental Education, 89 (1), 125 – 144. https://doi.org/10.1080/00220973.2019.1582470</bibtext> </blist> <blist> <bibtext> Fischer, P., Krueger, J. I., Greitemeyer, T., Vogrincic, C., Kastenmüller, A., Frey, D., Heene, M., Wicher, M., & Kainbacher, M. (2011). The bystander-effect: A meta-analytic review on bystander intervention in dangerous and non-dangerous emergencies. Psychological Bulletin, 137 (4), 517 – 537. https://doi.org/10.1037/A0023304</bibtext> </blist> <blist> <bibtext> Ford, R. C., Bach, S. A., & Fottler, M. D. (1997). Methods of measuring patient satisfaction in health care organizations. Health Care Management Review, 22 (2), 74 – 89. https://doi.org/10.1097/00004010-199704000-00009</bibtext> </blist> <blist> <bibtext> Gnambs, T., & Kaspar, K. (2014). Disclosure of sensitive behaviors across self-administered survey modes: A meta-analysis. Behavior Research Methods, 47 (4), 1237 – 1259. https://doi.org/10.3758/S13428-014-0533-4/FIGURES/3</bibtext> </blist> <blist> <bibtext> * Greeson, J. K. P., & Thompson, A. E. (2017). Development, feasibility, and piloting of a novel natural mentoring intervention for older youth in foster care. Journal of Social Service Research, 43 (2), 205 – 222. https://doi.org/10.1080/01488376.2016.1248269</bibtext> </blist> <blist> <bibtext> Hall, E. T. (1976). Beyond culture. Anchor Press/Double day.</bibtext> </blist> <blist> <bibtext> Harandi, T. F., Taghinasab, M. M., & Nayeri, T. D. (2017). The correlation of social support with mental health: A meta-analysis. Electronic Physician, 9 (9), 5212 – 5222. https://doi.org/10.19082/5212</bibtext> </blist> <blist> <bibtext> Hariton, E., & Locascio, J. J. (2018). Randomised controlled trials—the gold standard for effectiveness research. BJOG, 125 (13), 1716 – 1716. https://doi.org/10.1111/1471-0528.15199</bibtext> </blist> <blist> <bibtext> * Hauken, M. A., Pereira, M., & Senneseth, M. (2018). The effects on children's anxiety and quality of life of a psychoeducational program for families living with parental cancer and their network. Cancer Nursing, 41 (6), 473 – 483. https://doi.org/10.1097/NCC.0000000000000529</bibtext> </blist> <blist> <bibtext> Hedges, L. V., & Olkin, I. (2014). Statistical methods for meta-analysis. Academic Press.</bibtext> </blist> <blist> <bibtext> Hofstede, G. (1980). Culture's consequences: Differences in work-related values. SAGE Publications.</bibtext> </blist> <blist> <bibtext> * Hollinshead, D. M., Corwin, T. W., Maher, E. J., Merkel-Holguin, L., Allan, H., & Fluke, J. D. (2017). Effectiveness of family group conferencing in preventing repeat referrals to child protective services and out-of-home placements. Child Abuse & Neglect, 69, 285 – 294. https://doi.org/10.1016/j.chiabu.2017.04.022</bibtext> </blist> <blist> <bibtext> Howick, J., Koletsi, D., Ioannidis, J. P. A., Madigan, C., Pandis, N., Loef, M., Walach, H., Sauer, S., Kleijnen, J., Seehra, J., Johnson, T., & Schmidt, S. (2022). Most healthcare interventions tested in Cochrane reviews are not effective according to high quality evidence: A meta-epidemiological study. Journal of Clinical Epidemiology, 148, 160 – 169. https://doi.org/10.2139/ssrn.4031358</bibtext> </blist> <blist> <bibtext> Hox, J. J., Moerbeek, M., & van de Schoot, R. (2017). Multilevel analysis: Techniques and applications. Routledge. https://doi.org/10.4324/9780203852279</bibtext> </blist> <blist> <bibtext> * James, C., Asscher, J. J., Stams, G. J. J. M., & Van Der Laan, P. H. (2016). The effectiveness of aftercare for juvenile and young adult offenders. International Journal of Offender Therapy and Comparative Criminology, 60 (10), 1159 – 1184. https://doi.org/10.1177/0306624X15576884</bibtext> </blist> <blist> <bibtext> * Jeong, S., McGarrell, E. F., & Hipple, N. K. (2012). Long-term impact of family group conferences on re-offending: The Indianapolis restorative justice experiment. Journal of Experimental Criminology, 8 (4), 369 – 385. https://doi.org/10.1007/s11292-012-9158-8</bibtext> </blist> <blist> <bibtext> * King, C. A., Arango, A., Kramer, A., Busby, D., Czyz, E., Foster, C. E., & Gillespie, B. W. (2019). Association of the youth-nominated support team intervention for suicidal adolescents with 11- to 14-year mortality outcomes: Secondary analysis of a randomized clinical trial. JAMA Psychiatry, 76 (5), 492 – 498. https://doi.org/10.1001/jamapsychiatry.2018.4358</bibtext> </blist> <blist> <bibtext> * King, C. A., Klaus, N., Kramer, A., Venkataraman, S., Quinlan, P., & Gillespie, B. (2009). The Youth-nominated Support Team-version II for suicidal adolescents: A randomized controlled intervention trial. Journal of Consulting and Clinical Psychology, 77 (5), 880 – 893. https://doi.org/10.1037/a0016552</bibtext> </blist> <blist> <bibtext> * King, C. A., Kramer, A., Preuss, L., Kerr, D. C. R., Weisse, L., & Venkataraman, S. (2006). Youth-nominated support team for suicidal adolescents (version 1): A randomized controlled trial. Journal of Consulting and Clinical Psychology, 74 (1), 199 – 206. https://doi.org/10.1037/0022-006X.74.1.199</bibtext> </blist> <blist> <bibtext> Knapp, G., & Hartung, J. (2003). Improved tests for a random effects meta-regression with a single covariate. Statistics in Medicine, 22 (17), 2693 – 2710. https://doi.org/10.1002/sim.1482</bibtext> </blist> <blist> <bibtext> * Knox, L., Guerra, N. G., Williams, K. R., & Toro, R. (2011). Preventing children's aggression in immigrant Latino families: A mixed methods evaluation of the Families and Schools Together program. American Journal of Community Psychology, 48 (1-2), 65 – 76. https://doi.org/10.1007/s10464-010-9411-0</bibtext> </blist> <blist> <bibtext> Koper, N., Creemers, H. E., van Dam, L., Stams, G. J. J. M., & Branje, S. (2021). Needs of youth and parents from multi-problem families in the search for youth-initiated mentors. Youth & Society, 55 (3), 420 – 446. https://doi.org/10.1177/0044118X211058217</bibtext> </blist> <blist> <bibtext> Kraemer, H. C., Wilson, G. T., Fairburn, C. G., & Agras, W. S. (2002). Mediators and moderators of treatment effects in randomized clinical trials. Archives of General Psychiatry, 59 (10), 877 – 883. <ulink href="http://www.ncbi.nlm.nih.gov/pubmed/12365874">http://www.ncbi.nlm.nih.gov/pubmed/12365874</ulink> https://doi.org/10.1001/archpsyc.59.10.877</bibtext> </blist> <blist> <bibtext> * Kratochwill, T. R., McDonald, L., Levin, J. R., Bear-Tibbetts, H. Y., & Demaray, M. K. (2004). Families and schools together: An experimental analysis of a parent-mediated multi-family group program for American Indian children. Journal of School Psychology, 42 (5), 359 – 383. https://doi.org/10.1016/j.jsp.2004.08.001</bibtext> </blist> <blist> <bibtext> * Kratochwill, T. R., McDonald, L., Levin, J. R., Scalia, P. A., & Coover, G. (2009). Families and Schools together: An experimental study of multi-family support groups for children at risk. Journal of School Psychology, 47 (4), 245 – 265. https://doi.org/10.1016/j.jsp.2009.03.001</bibtext> </blist> <blist> <bibtext> Krause, M. S. (1966). A cognitive theory of motivation for treatment. The Journal of General Psychology, 75, 9 – 19. https://doi.org/10.1080/00221309.1966.9710345</bibtext> </blist> <blist> <bibtext> * Lambert, M. C., Johnson, L. E., & Wang, E. W. (2017). The impact of family group decision-making on preventing removals. Children and Youth Services Review, 78, 89 – 92. https://doi.org/10.1016/j.childyouth.2017.05.005</bibtext> </blist> <blist> <bibtext> Lee, S. S., Humphreys, K. L., Flory, K., Liu, R., & Glass, K. (2011). Prospective association of childhood attention-deficit/hyperactivity disorder (ADHD) and substance use and abuse/dependence: A meta-analytic review. Clinical Psychology Review, 31 (3), 328 – 341. https://doi.org/10.1016/J.CPR.2011.01.006</bibtext> </blist> <blist> <bibtext> * Leon, S. C., Saucedo, D. J., & Jachymiak, K. (2016). Keeping it in the family: The impact of a Family Finding intervention on placement, permanency, and well-being outcomes. Children and Youth Services Review, 70, 163 – 170. https://doi.org/10.1016/j.childyouth.2016.09.020</bibtext> </blist> <blist> <bibtext> Leventhal, T., & Brooks-Gunn, J. (2000). The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126 (2), 309 – 337. https://doi.org/10.1037/0033-2909.126.2.309</bibtext> </blist> <blist> <bibtext> Lipsey, M. W., Wilson., & D. B., Lipsey. (2001). Practical meta-analysis: Applied social research methods series (Vol. 49). Sage.</bibtext> </blist> <blist> <bibtext> MacLeod, J., & Nelson, G. (2000). Programs for the promotion of family wellness and the prevention of child maltreatment: A meta-analytic review. Child Abuse & Neglect, 24 (9), 1127 – 1149. https://doi.org/10.1016/S0145-2134(00)00178-2</bibtext> </blist> <blist> <bibtext> Marcus, D. K., O'Connell, D., Norris, A. L., & Sawaqdeh, A. (2014). Is the Dodo bird endangered in the 21st century? A meta-analysis of treatment comparison studies. In Clinical Psychology Review, 34 (7), 519 – 530. https://doi.org/10.1016/j.cpr.2014.08.001</bibtext> </blist> <blist> <bibtext> * McGarrell, E. F., & Hipple, N. K. (2007). Family group conferencing and re-offending among first-time juvenile offenders: The Indianapolis experiment. Justice Quarterly, 24 (2), 221 – 246. https://doi.org/10.1080/07418820701294789</bibtext> </blist> <blist> <bibtext> McGinn, T., Best, P., Wilson, J., Chereni, A., Kamndaya, M., & Shlonsky, A. (2020). Family group decision-making for children at risk of abuse or neglect: A systematic review. Campbell Systematic Reviews, 16 (3), e1088. https://doi.org/10.1002/CL2.1088</bibtext> </blist> <blist> <bibtext> Mclaughlin, K. A., Costello, E. J., Leblanc, W., Sampson, N. A., & Kessler, R. C. (2012). Socioeconomic status and adolescent mental disorders. American Journal of Public Health, 102 (9), 1742 – 1750. https://doi.org/10.2105/AJPH.2011.300477</bibtext> </blist> <blist> <bibtext> McLaughlin, K. A., Green, J. G., Gruber, M. J., Sampson, N. A., Zaslavsky, A. M., & Kessler, R. C. (2012). Childhood adversities and first onset of psychiatric disorders in a national sample of US adolescents. Archives of General Psychiatry, 69 (11), 1151 – 1160. https://doi.org/10.1001/archgenpsychiatry.2011.2277</bibtext> </blist> <blist> <bibtext> * Millenky, M., Bloom, D., & Dillon, C. (2010). Making the transition: Interim results of the National Guard Youth ChalleNGe Evaluation.</bibtext> </blist> <blist> <bibtext> * Millenky, M., Schwartz, S. E. O., & Rhodes, J. E. (2014). Supporting the transition to adulthood among high school dropouts: An impact study of the National Guard Youth ChalleNGe program. Prevention Science, 15 (4), 448 – 459. https://doi.org/10.1007/s11121-013-0388-4</bibtext> </blist> <blist> <bibtext> Mullen, B. (2013). Advanced basic meta-analysis. Lawrence Erlbaum Associates. https://doi.org/10.4324/9780203771938</bibtext> </blist> <blist> <bibtext> OECD. (2022). Poverty rate (Indicator). OECD.</bibtext> </blist> <blist> <bibtext> * Onrust, S. A., Romijn, G., & De Beer, Y. (2015). Family group conferences within the integrated care system for young people with ID: A controlled study of effects and costs. BMC Health Services Research, 15 (1), 392. https://doi.org/10.1186/s12913-015-1062-2</bibtext> </blist> <blist> <bibtext> Ouzzani, M., Hammady, H., Fedorowicz, Z., & Elmagarmid, A. (2016). Rayyan-a web and mobile app for systematic reviews. Systematic Reviews, 5 (1), 210. https://doi.org/10.1186/s13643-016-0384-4</bibtext> </blist> <blist> <bibtext> Ozbay, F., Johnson, D. C., Dimoulas, E., C. A., Morgan, I., Charney, D., & Southwick, S. (2007). Social support and resilience to stress: From neurobiology to clinical practice. Psychiatry, 4 (5), 35. /pmc/articles/PMC2921311/</bibtext> </blist> <blist> <bibtext> * Pennell, J., Edwards, M., & Burford, G. (2010). Expedited family group engagement and child permanency. Children and Youth Services Review, 32 (7), 1012 – 1019. https://doi.org/10.1016/j.childyouth.2010.03.029</bibtext> </blist> <blist> <bibtext> Peters, J. L., Sutton, A. J., Jones, D. R., Abrams, K. R., & Rushton, L. (2007). Performance of the trim and fill method in the presence of publication bias and between-study heterogeneity. Statistics in Medicine, 26 (25), 4544 – 4562. https://doi.org/10.1002/sim.2889</bibtext> </blist> <blist> <bibtext> R Core Team. (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing.</bibtext> </blist> <blist> <bibtext> Raposa, E. B., Rhodes, J. E., Stams, G. J. J. M., Card, N., Burton, S., Schwartz, S. E. O., Sykes, L. A. Y., Kanchewa, S. S., Kupersmidt, J., & Hussain, S. (2019). The effects of youth mentoring programs: A meta-analysis of outcome studies. Journal of Youth and Adolescence, 48 (3), 423 – 443. https://doi.org/10.1007/s10964-019-00982-8</bibtext> </blist> <blist> <bibtext> * Reekers, S. E., Dijkstra, S., Stams, G. J. J. M., Asscher, J. J., & Creemers, H. E. (2018). Signs of effectiveness of signs of safety? – A pilot study. Children and Youth Services Review, 91, 177 – 184. https://doi.org/10.1016/j.childyouth.2018.06.012</bibtext> </blist> <blist> <bibtext> Rhodes, J. E., Reddy, R., & Grossman, J. B. (2005). The protective influence of mentoring on adolescents' substance use: Direct and indirect pathways. Applied Developmental Science, 9 (1), 31 – 47. https://doi.org/10.1207/s1532480xads0901_4</bibtext> </blist> <blist> <bibtext> Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86 (3), 638 – 641. https://doi.org/10.1037/0033-2909.86.3.638</bibtext> </blist> <blist> <bibtext> Rubin, D. B. (1976). Inference and missing data. Biometrika, 63 (3), 581 – 592. https://doi.org/10.1093/biomet/63.3.581</bibtext> </blist> <blist> <bibtext> * Rushovich, B., Hebert, A., Crampton, D., & Malm, K. (2021). Results from a randomized controlled trial of team decision-making in child welfare. Children and Youth Services Review, 131, 106263. https://doi.org/10.1016/j.childyouth.2021.106263</bibtext> </blist> <blist> <bibtext> Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. The American Psychologist, 55 (1), 68 – 78. https://doi.org/10.1037/0003-066X.55.1.68</bibtext> </blist> <blist> <bibtext> Ryan, R. M., & Deci, E. L. (2008). A self-determination theory approach to psychotherapy: The motivational basis for effective change. Canadian Psychology/Psychologie Canadienne, 49 (3), 186 – 193. https://doi.org/10.1037/a0012753</bibtext> </blist> <blist> <bibtext> Saha, S., Saint, S., & Christakis, D. A. (2003). Impact factor: A valid measure of journal quality? Journal of the Medical Library Association, 91 (1), 42 – 46.</bibtext> </blist> <blist> <bibtext> Sarason, I. G., & Sarason, B. R. (2009). Social support: Mapping the construct. Journal of Social and Personal Relationships, 26 (1), 113 – 120. https://doi.org/10.1177/0265407509105526</bibtext> </blist> <blist> <bibtext> Schwartz, S. E. O., Kanchewa, S. S., Rhodes, J. E., Gowdy, G., Stark, A. M., Horn, J. P., Parnes, M., & Spencer, R. (2017). "I'm having a little struggle with this, can you help me out?": Examining impacts and processes of a social capital intervention for first-generation college students. American Journal of Community Psychology, 61 (1-2), 166 – 178. https://doi.org/10.1002/ajcp.12206</bibtext> </blist> <blist> <bibtext> Schwartz, S. E. O., & Rhodes, J. E. (2016). From treatment to empowerment: New approaches to youth mentoring. American Journal of Community Psychology, 58 (1-2), 150 – 157. https://doi.org/10.1002/AJCP.12070</bibtext> </blist> <blist> <bibtext> Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental & quasi-experimental designs for generalized causal inference. Houghton Mifflin.</bibtext> </blist> <blist> <bibtext> Shadish, W. R., Navarro, A. M., Matt, G. E., & Phillips, G. (2000). The effects of psychological therapies under clinically representative conditions: A meta-analysis. Psychological Bulletin, 126 (4), 512 – 529. https://doi.org/10.1037/0033-2909.126.4.512</bibtext> </blist> <blist> <bibtext> * Sheets, J., Wittenstrom, K., Fong, R., James, J., Tecci, M., Baumann, D. J., & Rodriguez, C. (2009). Evidence-based practice in family group decision-making for Anglo, African American and Hispanic families. Children and Youth Services Review, 31 (11), 1187 – 1191. https://doi.org/10.1016/j.childyouth.2009.08.003</bibtext> </blist> <blist> <bibtext> Simon, N. M., Herlands, N. N., Marks, E. H., Mancini, C., Letamendi, A., Li, Z., Pollack, M. H., Van Ameringen, M., & Stein, M. B. (2009). Childhood maltreatment linked to greater symptom severity and poorer quality of life and function in social anxiety disorder. Depression and Anxiety, 26 (11), 1027 – 1032. https://doi.org/10.1002/DA.20604</bibtext> </blist> <blist> <bibtext> Southwick, S. M., Morgan, C. A., Vythilingam, M., & Charney, D. (2006). Mentors enhance resilience in at-risk children and adolescents. Psychoanalytic Inquiry, 26 (4), 577 – 584. https://doi.org/10.1080/07351690701310631</bibtext> </blist> <blist> <bibtext> Spear, H. J., & Kulbok, P. (2004). Autonomy and adolescence: A concept analysis. Public Health Nursing, 21 (2), 144 – 152. https://doi.org/10.1111/J.0737-1209.2004.021208.X</bibtext> </blist> <blist> <bibtext> Stams, G. J. J. M., Juffer, F., Rispens, J., & Hoksbergen, R. A. C. (2000). The development and adjustment of 7-year-old children adopted in infancy. Journal of Child Psychology and Psychiatry, 41 (8), 1025 – 1037. https://doi.org/10.1017/S0021963099006344</bibtext> </blist> <blist> <bibtext> Sterrett, E. M., Jones, D. J., McKee, L. G., & Kincaid, C. (2011). Supportive non-parental adults and adolescent psychosocial functioning: Using social support as a theoretical framework. American Journal of Community Psychology, 48 (3-4), 284 – 295. https://doi.org/10.1007/s10464-011-9429-y</bibtext> </blist> <blist> <bibtext> Stice, E., Shaw, H., Bohon, C., Marti, C. N., & Rohde, P. (2009). A meta-analytic review of depression prevention programs for children and adolescents: Factors that predict magnitude of intervention effects. Journal of Consulting and Clinical Psychology, 77 (3), 486 – 503. https://doi.org/10.1037/a0015168</bibtext> </blist> <blist> <bibtext> * Sundell, K., & Vinnerljung, B. (2004). Outcomes of family group conferencing in Sweden: A 3-year follow-up. Child Abuse & Neglect, 28 (3), 267 – 287. https://doi.org/10.1016/j.chiabu.2003.09.018</bibtext> </blist> <blist> <bibtext> Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics: Data handling in science and technology. (6th ed.). Pearson Education, Inc.</bibtext> </blist> <blist> <bibtext> * Teal, B. L. (2013). The effectiveness of team decision making in children's division on child welfare outcomes. Capella University.</bibtext> </blist> <blist> <bibtext> Telzer, E. H., van Hoorn, J., Rogers, C. R., & Do, K. T. (2018). Social influence on positive youth development: A developmental neuroscience perspective. Advances in Child Development and Behavior, 54, 215 – 258. https://doi.org/10.1016/bs.acdb.2017.10.003</bibtext> </blist> <blist> <bibtext> * Tijms, J., Stoop, M. A., & Polleck, J. N. (2018). Bibliotherapeutic book club intervention to promote reading skills and social–emotional competencies in low SES community-based high schools: A randomised controlled trial. Journal of Research in Reading, 41 (3), 525 – 545. https://doi.org/10.1111/1467-9817.12123</bibtext> </blist> <blist> <bibtext> * Turley, R. N. L., Gamoran, A., McCarty, A. T., & Fish, R. (2017). Reducing children's behavior problems through social capital: A causal assessment. Social Science Research, 61, 206 – 217. https://doi.org/10.1016/j.ssresearch.2016.06.015</bibtext> </blist> <blist> <bibtext> Ungar, M. (2011). Community resilience for youth and families: Facilitative physical and social capital in contexts of adversity. Children and Youth Services Review, 33 (9), 1742 – 1748. https://doi.org/10.1016/j.childyouth.2011.04.027</bibtext> </blist> <blist> <bibtext> Valentine, J. C., Leach, S. M., Fowler, A. P., Stojda, D. K., & Macdonald, G. (2019). Families and schools together (FAST) for improving outcomes for children and their families. The Cochrane Database of Systematic Reviews, 7 (7), CD012760. https://doi.org/10.1002/14651858.CD012760.pub2</bibtext> </blist> <blist> <bibtext> van Dam, L., Blom, D., Kara, E., Assink, M., Stams, G. J. J. M., Schwartz, S. E. O., & Rhodes, J. E. (2020). Youth initiated mentoring: A meta-analytic study of a hybrid approach to youth mentoring. Journal of Youth and Adolescence, 50 (2), 219 – 230. https://doi.org/10.1007/s10964-020-01336-5</bibtext> </blist> <blist> <bibtext> * van Dam, L., Klein Schaarsberg, R., Wissink, I. B., Zwaanswijk, M., Stams, G. J. J. M., & Schwartz, S. E. O. (2018). Youth initiated mentoring: A mixed-methods follow-up study among youth at-risk for out-of-home placement.</bibtext> </blist> <blist> <bibtext> * van Dam, L., Neels, S., de Winter, M., Branje, S., Wijsbroek, S., Hutschemaekers, G., Dekker, A., Sekreve, A., Zwaanswijk, M., Wissink, I. B., & Stams, G. J. J. M. (2017). Youth initiated mentors: Do they offer an alternative for out-of-home placement in youth care? The British Journal of Social Work, 47 (6), 1764 – 1780. https://doi.org/10.1093/bjsw/bcx092</bibtext> </blist> <blist> <bibtext> van Dam, L., Smit, D., Wildschut, B., Branje, S., Rhodes, J. E., Assink, M., & Stams, G. J. J. M. (2018). Does natural mentoring matter? A multilevel meta-analysis on the association between natural mentoring and youth outcomes. American Journal of Community Psychology, 62 (1-2), 203 – 220. https://doi.org/10.1002/ajcp.12248</bibtext> </blist> <blist> <bibtext> van der Helm, G. H. P., Kuiper, C. H. Z., & Stams, G. J. J. M. (2018). Group climate and treatment motivation in secure residential and forensic youth care from the perspective of self determination theory. Children and Youth Services Review, 93, 339 – 344. https://doi.org/10.1016/j.childyouth.2018.07.028</bibtext> </blist> <blist> <bibtext> van Loon, A. W. G., Creemers, H. E., Beumer, W. Y., Okorn, A., Vogelaar, S., Saab, N., Miers, A. C., Westenberg, P. M., & Asscher, J. J. (2020). Can schools reduce adolescent psychological stress? A multilevel meta-analysis of the effectiveness of school-based intervention programs. Journal of Youth and Adolescence, 49 (6), 1127 – 1145. https://doi.org/10.1007/s10964-020-01201-5</bibtext> </blist> <blist> <bibtext> * Vandivere, S., Malm, K. E., Allen, T. J., Williams, S. C., & McKlindon, A. (2017). A randomized controlled trial of Family Finding: A relative search and engagement intervention for youth lingering in foster care. Evaluation Review, 41 (6), 542 – 567. https://doi.org/10.1177/0193841X17689971</bibtext> </blist> <blist> <bibtext> Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36 (3), 1 – 48. https://doi.org/10.18637/jss.v036.i03</bibtext> </blist> <blist> <bibtext> * Wang, E. W., Lambert, M. C., Johnson, L. E., Boudreau, B., Breidenbach, R., & Baumann, D. (2012). Expediting permanent placement from foster care systems: The role of family group decision-making. Children and Youth Services Review, 34 (4), 845 – 850. https://doi.org/10.1016/j.childyouth.2012.01.015</bibtext> </blist> <blist> <bibtext> * Weigensberg, E. C., Barth, R. P., & Guo, S. (2009). Family group decision making: A propensity score analysis to evaluate child and family services at baseline and after 36-months. Children and Youth Services Review, 31 (3), 383 – 390. https://doi.org/10.1016/j.childyouth.2008.09.001</bibtext> </blist> <blist> <bibtext> Weisz, J. R., Kuppens, S., Ng, M. Y., Eckshtain, D., Ugueto, A. M., Vaughn-Coaxum, R., Jensen-Doss, A., Hawley, K. M., Krumholz Marchette, L. S., Chu, B. C., Weersing, V. R., & Fordwood, S. R. (2017). What five decades of research tells us about the effects of youth psychological therapy: A multilevel meta-analysis and implications for science and practice. The American Psychologist, 72 (2), 79 – 117. https://doi.org/10.1037/a0040360</bibtext> </blist> <blist> <bibtext> Weisz, J. R., Sandler, I. N., Durlak, J. A., & Anton, B. S. (2005). Promoting and protecting youth mental health through evidence-based prevention and treatment. American Psychologist, 60 (6), 628 – 648. https://doi.org/10.1037/0003-066X.60.6.628</bibtext> </blist> <blist> <bibtext> Weisz, J. R., Ugueto, A. M., Cheron, D. M., & Herren, J. (2013). Evidence-based youth psychotherapy in the mental health ecosystem. Journal of Clinical Child and Adolescent Psychology, 42 (2), 274 – 286. https://doi.org/10.1080/15374416.2013.764824</bibtext> </blist> <blist> <bibtext> * Wingrove, T. A., & Weisz, V. (2005). An evaluation of expedited family group conferences [Paper presentation]. American Law-Psychology Society Annual Conference,.</bibtext> </blist> <blist> <bibtext> Woolley, A. W., Aggarwal, I., & Malone, T. W. (2015). Collective intelligence and group performance. Current Directions in Psychological Science, 24 (6), 420 – 424. https://doi.org/10.1177/0963721415599543</bibtext> </blist> <blist> <bibtext> Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. Science (New York, N.Y.), 330 (6004), 686 – 688. https://doi.org/10.1126/science.1193147</bibtext> </blist> <blist> <bibtext> Zimmerman, F. J. (2005). Social and economic determinants of disparities in professional help-seeking for child mental health problems: Evidence from a national sample. Health Services Research, 40 (5 Pt 1), 1514 – 1533. https://doi.org/10.1111/J.1475-6773.2005.00411.X</bibtext> </blist> </ref> <aug> <p>By Natasha Koper; Roos M. van der Heijden; Sophie Donk; Thao Kieu; Hanneke E. Creemers; Levi van Dam; Susan Branje and Geert Jan J. M. Stams</p> <p>Reported by Author; Author; Author; Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib13" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib59" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib67" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib94" firstref="ref4"></nolink> <nolink nlid="nl5" bibid="bib119" firstref="ref5"></nolink> <nolink nlid="nl6" bibid="bib44" firstref="ref6"></nolink> <nolink nlid="nl7" bibid="bib118" firstref="ref7"></nolink> <nolink nlid="nl8" bibid="bib120" firstref="ref8"></nolink> <nolink nlid="nl9" bibid="bib88" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib95" firstref="ref11"></nolink> <nolink nlid="nl11" bibid="bib106" firstref="ref12"></nolink> <nolink nlid="nl12" bibid="bib111" firstref="ref13"></nolink> <nolink nlid="nl13" bibid="bib63" firstref="ref14"></nolink> <nolink nlid="nl14" bibid="bib18" firstref="ref15"></nolink> <nolink nlid="nl15" bibid="bib108" firstref="ref16"></nolink> <nolink nlid="nl16" bibid="bib23" firstref="ref17"></nolink> <nolink nlid="nl17" bibid="bib66" firstref="ref18"></nolink> <nolink nlid="nl18" bibid="bib107" firstref="ref19"></nolink> <nolink nlid="nl19" bibid="bib85" firstref="ref20"></nolink> <nolink nlid="nl20" bibid="bib14" firstref="ref22"></nolink> <nolink nlid="nl21" bibid="bib103" firstref="ref25"></nolink> <nolink nlid="nl22" bibid="bib57" firstref="ref27"></nolink> <nolink nlid="nl23" bibid="bib112" firstref="ref28"></nolink> <nolink nlid="nl24" bibid="bib124" firstref="ref36"></nolink> <nolink nlid="nl25" bibid="bib75" firstref="ref38"></nolink> <nolink nlid="nl26" bibid="bib19" firstref="ref58"></nolink> <nolink nlid="nl27" bibid="bib33" firstref="ref59"></nolink> <nolink nlid="nl28" bibid="bib96" firstref="ref66"></nolink> <nolink nlid="nl29" bibid="bib90" firstref="ref67"></nolink> <nolink nlid="nl30" bibid="bib61" firstref="ref69"></nolink> <nolink nlid="nl31" bibid="bib24" firstref="ref72"></nolink> <nolink nlid="nl32" bibid="bib79" firstref="ref73"></nolink> <nolink nlid="nl33" bibid="bib91" firstref="ref76"></nolink> <nolink nlid="nl34" bibid="bib34" firstref="ref77"></nolink> <nolink nlid="nl35" bibid="bib35" firstref="ref78"></nolink> <nolink nlid="nl36" bibid="bib12" firstref="ref79"></nolink> <nolink nlid="nl37" bibid="bib97" firstref="ref80"></nolink> <nolink nlid="nl38" bibid="bib15" firstref="ref81"></nolink> <nolink nlid="nl39" bibid="bib30" firstref="ref82"></nolink> <nolink nlid="nl40" bibid="bib64" firstref="ref86"></nolink> <nolink nlid="nl41" bibid="bib87" firstref="ref88"></nolink> <nolink nlid="nl42" bibid="bib74" firstref="ref100"></nolink> <nolink nlid="nl43" bibid="bib17" firstref="ref106"></nolink> <nolink nlid="nl44" bibid="bib62" firstref="ref107"></nolink> <nolink nlid="nl45" bibid="bib11" firstref="ref108"></nolink> <nolink nlid="nl46" bibid="bib78" firstref="ref114"></nolink> <nolink nlid="nl47" bibid="bib115" firstref="ref115"></nolink> <nolink nlid="nl48" bibid="bib51" firstref="ref117"></nolink> <nolink nlid="nl49" bibid="bib82" firstref="ref118"></nolink> <nolink nlid="nl50" bibid="bib71" firstref="ref119"></nolink> <nolink nlid="nl51" bibid="bib29" firstref="ref120"></nolink> <nolink nlid="nl52" bibid="bib32" firstref="ref121"></nolink> <nolink nlid="nl53" bibid="bib31" firstref="ref123"></nolink> <nolink nlid="nl54" bibid="bib26" firstref="ref124"></nolink> <nolink nlid="nl55" bibid="bib27" firstref="ref125"></nolink> <nolink nlid="nl56" bibid="bib77" firstref="ref127"></nolink> <nolink nlid="nl57" bibid="bib116" firstref="ref130"></nolink> <nolink nlid="nl58" bibid="bib58" firstref="ref131"></nolink> <nolink nlid="nl59" bibid="bib408" firstref="ref132"></nolink> <nolink nlid="nl60" bibid="bib101" firstref="ref135"></nolink> <nolink nlid="nl61" bibid="bib406" firstref="ref139"></nolink> <nolink nlid="nl62" bibid="bib86" firstref="ref166"></nolink> <nolink nlid="nl63" bibid="bib89" firstref="ref167"></nolink> <nolink nlid="nl64" bibid="bib81" firstref="ref169"></nolink> <nolink nlid="nl65" bibid="bib16" firstref="ref171"></nolink> <nolink nlid="nl66" bibid="bib53" firstref="ref172"></nolink> <nolink nlid="nl67" bibid="bib123" firstref="ref176"></nolink> <nolink nlid="nl68" bibid="bib122" firstref="ref178"></nolink> <nolink nlid="nl69" bibid="bib37" firstref="ref181"></nolink> <nolink nlid="nl70" bibid="bib42" firstref="ref182"></nolink> <nolink nlid="nl71" bibid="bib72" firstref="ref183"></nolink> <nolink nlid="nl72" bibid="bib99" firstref="ref184"></nolink> <nolink nlid="nl73" bibid="bib113" firstref="ref185"></nolink> <nolink nlid="nl74" bibid="bib38" firstref="ref186"></nolink> <nolink nlid="nl75" bibid="bib25" firstref="ref187"></nolink> <nolink nlid="nl76" bibid="bib98" firstref="ref188"></nolink> <nolink nlid="nl77" bibid="bib92" firstref="ref193"></nolink> <nolink nlid="nl78" bibid="bib39" firstref="ref194"></nolink> <nolink nlid="nl79" bibid="bib41" firstref="ref197"></nolink> <nolink nlid="nl80" bibid="bib45" firstref="ref201"></nolink> <nolink nlid="nl81" bibid="bib83" firstref="ref202"></nolink> <nolink nlid="nl82" bibid="bib54" firstref="ref203"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1502924
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: The Effectiveness of Interventions for Youth That Activate the Social Network: A Meta-Analytic Study
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Natasha+Koper%22">Natasha Koper</searchLink><br /><searchLink fieldCode="AR" term="%22Roos+M%2E+van+der+Heijden%22">Roos M. van der Heijden</searchLink><br /><searchLink fieldCode="AR" term="%22Sophie+Donk%22">Sophie Donk</searchLink><br /><searchLink fieldCode="AR" term="%22Thao+Kieu%22">Thao Kieu</searchLink><br /><searchLink fieldCode="AR" term="%22Hanneke+E%2E+Creemers%22">Hanneke E. Creemers</searchLink><br /><searchLink fieldCode="AR" term="%22Levi+van+Dam%22">Levi van Dam</searchLink><br /><searchLink fieldCode="AR" term="%22Susan+Branje%22">Susan Branje</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-9999-5313">0000-0002-9999-5313</externalLink>)<br /><searchLink fieldCode="AR" term="%22Geert+Jan+J%2E+M%2E+Stams%22">Geert Jan J. M. Stams</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Applied+Developmental+Science%22"><i>Applied Developmental Science</i></searchLink>. 2025 29(3):195-219.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 25
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2025
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Information Analyses
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Literature+Reviews%22">Literature Reviews</searchLink><br /><searchLink fieldCode="DE" term="%22Youth%22">Youth</searchLink><br /><searchLink fieldCode="DE" term="%22Young+Adults%22">Young Adults</searchLink><br /><searchLink fieldCode="DE" term="%22Intervention%22">Intervention</searchLink><br /><searchLink fieldCode="DE" term="%22Social+Networks%22">Social Networks</searchLink><br /><searchLink fieldCode="DE" term="%22Social+Support+Groups%22">Social Support Groups</searchLink><br /><searchLink fieldCode="DE" term="%22Program+Descriptions%22">Program Descriptions</searchLink><br /><searchLink fieldCode="DE" term="%22Program+Effectiveness%22">Program Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Research%22">Educational Research</searchLink><br /><searchLink fieldCode="DE" term="%22Individual+Characteristics%22">Individual Characteristics</searchLink><br /><searchLink fieldCode="DE" term="%22Evaluation+Methods%22">Evaluation Methods</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1080/10888691.2024.2317714
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1088-8691<br />1532-480X
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: This meta-analysis aimed to examine the effectiveness of interventions for youth that activate the social network for improving youth outcomes (e.g. psychological problems, child safety). A literature search yielded 37 studies with 35 independent samples (N = 712,269) of youth aged 0-26 years (M = 7.20), and 409 effect sizes. A three-level meta-analysis controlling for the dependency among effect sizes within studies showed no overall effect of interventions activating the social network (d = 0.11, p = 0.241). Yet, moderator analyses revealed positive effects for youth-initiated mentoring interventions (d = 0.46), youth deciding who to involve (d = 0.52), interventions that involve only one person (d = 0.56), European samples (d = 0.40), interventions targeting youth with mental health needs (d = 0.75), data retrieved through questionnaires (d = 0.10) and official records (d = 0.14), assessments completed by professionals (d = 0.34) or parents (d = 0.17), and outcomes that were corrected for pretest differences between conditions (d = 0.27). This meta-analysis demonstrates that social network activation matters for intervention effectiveness under specific conditions.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2026
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1502924
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1502924
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1080/10888691.2024.2317714
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 25
        StartPage: 195
    Subjects:
      – SubjectFull: Literature Reviews
        Type: general
      – SubjectFull: Youth
        Type: general
      – SubjectFull: Young Adults
        Type: general
      – SubjectFull: Intervention
        Type: general
      – SubjectFull: Social Networks
        Type: general
      – SubjectFull: Social Support Groups
        Type: general
      – SubjectFull: Program Descriptions
        Type: general
      – SubjectFull: Program Effectiveness
        Type: general
      – SubjectFull: Educational Research
        Type: general
      – SubjectFull: Individual Characteristics
        Type: general
      – SubjectFull: Evaluation Methods
        Type: general
    Titles:
      – TitleFull: The Effectiveness of Interventions for Youth That Activate the Social Network: A Meta-Analytic Study
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Natasha Koper
      – PersonEntity:
          Name:
            NameFull: Roos M. van der Heijden
      – PersonEntity:
          Name:
            NameFull: Sophie Donk
      – PersonEntity:
          Name:
            NameFull: Thao Kieu
      – PersonEntity:
          Name:
            NameFull: Hanneke E. Creemers
      – PersonEntity:
          Name:
            NameFull: Levi van Dam
      – PersonEntity:
          Name:
            NameFull: Susan Branje
      – PersonEntity:
          Name:
            NameFull: Geert Jan J. M. Stams
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 1088-8691
            – Type: issn-electronic
              Value: 1532-480X
          Numbering:
            – Type: volume
              Value: 29
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Applied Developmental Science
              Type: main
ResultId 1