I Get by With a Little Help From My Friends: The Impact of a Gender-Tailored Group Mentoring Program on Peer Relationships

Abstract

Adolescents face many difficulties when transitioning to high school, particularly when it comes to their social circles and maintaining friendships. As these challenges can have long-term consequences, providing support to youth during this period is essential. The YWCA Youth Education Programs (YEP) in Vancouver, BC offer group-based, after-school mentoring programs for 7^th grade students to help them navigate the transition to high school. Survey data from 668 students (233 treatment group and 435 control group) were used to examine the impacts of the YEP on friend support. Multilevel regression models show that the YEP is effective at improving perceptions of friend support in the long term. More research is needed to explore potential gender differences in the experience of supportive friendships after youth transition to high school, as well as the impact of group-based mentoring for youth transitioning to the high school environment.

Keywords

youth friend support group mentoring program program evaluation multilevel model propensity score weights

Introduction

Throughout adolescence, building genuine and trusting friendships is vital for personal growth and development (Cantin & Boivin, 2004; Wrzus et al., 2013). Supportive peer connections are crucial for adolescents as they offer a sense of belonging, acceptance, and reassurance (Benner et al., 2017; Morin et al., 2013). However, adolescents have a greater sensitivity to risks and rewards, especially those related to social and emotional experiences (e.g., Duell et al., 2018; van Duijvenvoorde et al., 2022). Youth also tend to participate in more frequent risky behaviors when they are with their peers in comparison to how adults behave (Chein et al., 2011; Gardner & Steinberg, 2005; Lundborg, 2006). As such, adolescents’ heightened sensitivity to social influence and peer approval can shape their behavior and decision-making processes (e.g., Blakemore & Mills, 2014; Foulkes & Blakemore, 2016). Ample recent studies have highlighted the vital role that friendships play in adolescent development (e.g., Benner et al., 2017; Felmlee et al., 2018; van Rens et al., 2019).

When transitioning to high school, teenagers experience considerable disruption in their social networks and may face difficulties maintaining friendships (Ng-Knight et al., 2019). This is partly because, unlike in elementary school, high school students may not have the same class schedule as their friends. Consequently, they have less interaction with their close friends during the day and spend most of their time with a constantly changing group of unfamiliar students (Akos & Galassi, 2004; Felmlee et al., 2018). As a result, adolescents’ friendship structures shift during the transition to high school, and their central network of peers becomes smaller (Felmlee et al., 2018; Ng-Knight et al., 2019). Studies have shown that many high school freshmen struggle to adapt socially (Pratt & George, 2005; van Rens et al., 2019). Experts in the field have repeatedly emphasized the need for effective interventions to help students transition to high school (e.g., Benner & Graham, 2009; Krammer et al., 2023; Rice et al., 2011). The YWCA Youth Education Programs (YEP) offer gender-specific after-school group mentoring programs for 7^th grade students, which focus on helping adolescents navigate the upcoming transition to high school (which occurs in grade 8). The current study aims to explore the impact of participation in the YWCA YEP on students’ perceptions of friend support.

The Impact of School Transition on Adolescents’ Social Relationships

Studies have consistently found that the transition to high school can negatively impact adolescents’ social outcomes (Benner, 2011; Ng-Knight et al., 2019). After this transition, teenagers tend to feel more lonely, isolated, disconnected, and anxious in social situations (e.g., Benner et al., 2017; Felmlee et al., 2018). They also may experience a decrease in social acceptance, support from friends, and healthy peer relationships (Benner & Graham, 2009; Felmlee et al., 2018). On the other hand, studies suggest that having strong social connections throughout high school can have a beneficial effect on emotional and social well-being (Benner et al., 2017; Morin et al., 2013). In addition, having stable friendships and peer networks can lead to better academic outcomes, including increased engagement at school, success, and a reduced likelihood of dropping out (Benner et al., 2017; Carbonaro & Workman, 2013).

Gender Differences

While school transitions can have a negative impact on students regardless of their demographics (e.g., race, ethnicity, gender, socioeconomic status), the experience of transitioning to high school is not homogeneous. Some subgroups of students have a disproportionately higher risk of negative outcomes after the transition to high school, although results from empirical studies are mixed. For instance, some research suggests that females experience more disruptions to their socio-emotional well-being than their male classmates (Benner & Graham, 2009; Grills-Taquechel et al., 2010), while other findings suggest that males are not exempt from disruptions to socio-emotional well-being (Benner et al., 2017). The gender difference in social outcomes may arise due to females placing a higher value on meaningful emotional connections during adolescence (Anderson et al., 2000; Delgado et al., 2022; Rudolph & Dodson, 2022). As a result, they may face more difficulties adapting to changes in their social circles in high school. Although research has been conducted on gender differences in friendship patterns throughout adolescence (e.g., see Davis, 2019), there is a dearth of research examining the disparate social outcomes for males and females specifically after transitioning to high school.

Interventions that Support Adolescents’ Social Development

Research on positive youth development and educational transitions consistently supports providing youth with programs to help them navigate the challenges associated with the transition to high school (Bharara, 2020; Donaldson et al., 2023). In recent years, afterschool programs (ASPs) have evolved to supplement the school day by providing youth with prosocial opportunities to develop skills related to their social well-being (e.g., communication, relationships), and support youth in their social development (e.g., social and emotional learning, interpersonal skills; Durlak et al., 2010; 2011; Himmelrich, 2012; Pelcher & Rajan, 2016).

Group mentoring is a popular form of ASP; according to the 2016 Mentoring Program Survey, group-based mentoring models are more common than traditional one-to-one (1:1) mentoring programs (Garringer et al., 2017). Group mentoring models typically involve two to three mentors and five to 20 mentees (Kuperminc & Deutsh, 2021). While the programs often do not have a specific therapeutic focus, the general aim is to strengthen relationships, both interpersonal and intrapersonal (Cawood & Wood, 2014), and special attention is given to fostering supportive relationships between the mentors and mentees (Kuperminc, 2016). The group context is theorized to offer considerable benefits to youth, especially for social skills and relationships. First, because youth can engage with several mentors and mentees, they are not solely reliant on the relationship with one mentor to reap positive program benefits (Haddock et al., 2020). Through the act of interacting with other youth their age, the group context is anticipated to foster meaningful relationships, which in turn create a sense of belonging and connectedness (Skudrzyk et al., 2009). Second, scholars argue that the group context strengthens key social processes that are essential for positive youth development, such as positive interpersonal connections with peers and a sense of belonging to a group; both of which are important to overcoming common challenges that arise during adolescence (Ma & Huebner, 2008). In group mentoring, youth are often encouraged to talk about personal struggles and issues that are common during this stage of life, such as values, relationships, and boundaries. Participating in these activities promotes positive social interactions and the development of meaningful peer relationships (Kuperminc et al., 2020).

Effectiveness of Youth Mentoring Programs on Social Outcomes

A substantial body of summative research suggests that formal community-based youth mentoring programs are a promising intervention for the promotion of positive outcomes for youth (e.g., psychological, social, academic). Meta-analytic studies have examined the impact of youth mentoring programs on various outcomes and, with few exceptions, are consistent in their demonstration of conventionally small, positive, and statistically significant overall effects (i.e., Burton, 2020; Christensen et al., 2020; DuBois et al., 2011; Eby et al., 2008; Raposa et al., 2019; Tolan et al., 2014). Youth mentoring programs have demonstrated effectiveness across several social domains such as improved social competence, social skills, social support, and interpersonal relationships (Burton, 2020; Christensen et al., 2020; DuBois et al., 2011; Raposa et al., 2019).

Effects of Group-Based Mentoring

Kuperminc and Deutsh (2021) reviewed mentoring programs for youth in the U.S. and identified several positive short-term outcomes for group programs. In particular, the benefits of group mentoring were noted for social and relational outcomes such as group cohesion, a sense of belonging, and positive peer interactions, both within the youths’ mentoring group as well as in their general social networks. Two recent meta-analyses have considered the impact of program format (e.g., 1:1 mentoring versus group-based mentoring) on youth outcomes and suggest that group mentoring has a positive impact on various aspects of youth’s social development and functioning (Burton, 2020; Poon et al., 2021). However, Burton (2020) examined the combined impact of individual and group-based mentoring, while Poon and colleagues (2021) assessed the impact of group mentoring on youth in foster care. As such, the effectiveness of group-based mentoring as a standalone intervention on the general population of youth (i.e., not high-risk or vulnerable adolescents) is not well known. Additionally, it is unclear whether these positive impacts extend primarily to the youth’s mentoring group or apply to their general social context. Further, very little is known about any differential impact of group mentoring programs by gender (Liang et al., 2013).

Current Study

The purpose of the current study is to examine the impact of the YWCA YEP on youths’ perceptions of friend support. A two-group quasi-experimental design (treatment (TX) and control (CTL)) was used to assess change over three time points (pretest, short-term posttest, and long-term follow-up¹). Specifically, the study seeks to answer the following three research questions: (1) What are the short-term and long-term impacts of the YWCA YEP on students’ perceptions of friend support? (2) Do perceptions of friend support vary by gender? and (3) Is there an interaction effect of gender and YWCA YEP participation on students’ perceptions of friend support?

As one of the key objectives of the YWCA YEP is to promote the development of skills and knowledge required for developing healthy friendships, we hypothesized that the YWCA YEP would have a positive effect on participants’ perceptions of friend support. Additionally, we hypothesized that program outcomes would differ between males and females in the TX and CTL groups. As the research questions were exploratory in nature, the hypotheses were not directional. Group mentoring programs have shown positive effects for youth, but limited research exists on the long-term and gendered impacts of such interventions, particularly with respect to friendships. Additionally, although some research suggests there may be differences in the transition experiences of females and males, little research has explored the potential gender differences in social outcomes post-transition.

Method

Brief Description of the YWCA Youth Education Programs

The YEP (Welcome to My Life (WTML) and Boys 4 Real (B4R)) is a set of curriculum-based, gender-tailored, afterschool, group mentoring programs designed to support 7^th grade students with their transition to high school. The programs are owned and operated by the YWCA, which is one of the largest non-profit organizations in Vancouver, BC. In their 18 years of operation and history of reaching several thousand youth at risk of negative school and behavioral outcomes, the YWCA YEPs have never been formally evaluated.

The program runs twice a year, once in the Fall and once in the Spring, and takes place in schools with limited activities for youth. The YWCA YEPs are available to all 7^th grade students at the selected schools. The participating school districts are structured such that there is no middle or junior high school. Instead, elementary school includes kindergarten – grade 7, and high school includes grades 8 to 12. As such, students transfer to high school in 8^th grade (typically at ages 12–13 years), and multiple elementary schools feed into one high school. Participation in the YEP is voluntary (space-permitting), but priority enrollment is given to students who might have limited resources or need extra support (e.g., those from low-income families, newcomers/immigrants, those who require extra support to integrate with peers). Students from multiple classes/feeder schools participate in the program together. Between Fall 2017 and Spring 2019, the YEP enrolled a total of 458 students across metro Vancouver in their programs (249 females and 209 males). On average, 12 students were enrolled in each program. However, participant enrolment ranged from 8 to 18 students.

The nine-week program operates once a week for 2.5 hours and is led by a team of three to five volunteer mentors of the same gender as the participants. Three types of volunteers make up the leadership team: (1) High School Ambassadors, (2) Facilitators, and (3) Wisdom Champions. The High School Ambassadors are current high school students whose role is to share positive and instructive stories about their own high school experiences. The ambassadors share first-hand experiences with participants regarding their high school and help to ease any concerns or fears students may have about the transition. Typically, a YEP has one to two High School Ambassadors. The Facilitators are typically university students and lead the delivery of the program curricula and activities; one to two Facilitators lead each YEP. Last, the Wisdom Champion is an experienced, knowledgeable adult (e.g., graduate student, working professional) who has already established their career and provides insight and mentorship to the participants. The Wisdom Champion’s role is to oversee the program, share wisdom, ensure participant physical and emotional safety, and purchase the weekly group snack.

The volunteer mentors are recruited through a variety of methods, including at university volunteer fairs; advertisements on volunteer-related websites; presentations to university and high school classes; promotion by university and high school partners; and word-of-mouth referrals. None of the program facilitators/volunteer mentors are employees of the YWCA. Volunteers must complete a 1-h online training module, as well as approximately 16 hours of mandatory in-person training. The training sessions aim to prepare volunteers for their roles in leading the programs, as well as develop rapport among mentors to encourage them to work effectively as a team. The in-person training sessions are led by YWCA staff and cover topics such as facilitation skills, behavior management, diversity training, child protection policies and procedures, team building, an overview of the program modules and Community Service Project, and specialized training pertaining to each volunteer position.

Each YEP session is 150 minutes, with 100 minutes focused on the week’s module, which is devoted to a specific theme (e.g., self-esteem, friendship, internet safety). The modules are taught through interactive activities, games, group discussions, and reflection journals. The remaining 50 minutes of the session is dedicated to a Community Service Project, which brings together participants from the boys’ and girls’ programs to collaborate on a meaningful project that supports the wider community. All YEP facilitators follow the same gender-specific curriculum, which covers relevant challenges that young people face during adolescence and the transition to high school such as peer pressure, self-esteem, healthy conflict, and friendships (YWCA, 2016). It is worth noting that while males and females enrol in different YEPs according to their gender identity, the aims, objectives, and module topics of WTML and B4R are almost identical. That is, although each program is gender-tailored (i.e., each program incorporates scenarios, topics, and/or activities that might appeal/resonate more with males or females), the topics of the modules are the same across the two programs. Refer to Table 1 for a list of the YEP modules and program content.

Table 1.

YWCA Youth Education Programs Curriculum Modules.

Week	Module	Description
Welcome To My Life
1	411 on girls	Introduces the girls to the program and the main themes of healthy self-esteem, healthy relationship skills, and responsible decision-making.
2	Live your best life	Focuses on making good choices regarding healthy living, including eating well, exercising, and how a healthy body contributes to a healthy mind.
3	What matters	Helps girls identify and clarify their values and help them consider how their values can influence their decisions and behaviors.
4	That’s what friends are for	Focuses on making friendships and includes discussions on what makes a quality friend and how girls can maintain healthy, positive friendships.
5	Friend or unfriend	Addresses how girls can handle conflict, including techniques for resolving conflict in a respectful manner. Strategies for overcoming peer pressure are also included.
6	Media madness	Focuses on self-esteem and body image, and how both are impacted by media manipulation and social expectations.
7	Community service project delivery	Participants implement their community service project (if the B4R and WTML programs are active at the same school, the community service project is done as a joint activity).
8	Social media smarts	Includes a series of activities designed to teach participants how to navigate the internet and social media safely and responsibly.
9	Love who you are	Aims to help girls develop healthy self-esteem and self-worth by identifying their positive attributes.
Boys 4 Real
1	Kick-off	Introduces the program and the themes of choice, challenge, and change.
2	Live your best life	Focuses on principles of healthy living and teaches boys about making healthy choices.
3	It’s my life	Helps boys to identify their values and think about how they can stay true to their values in difficult situations.
4	RESPECT!	Focuses on how boys can develop positive relationships with adults and peers; includes topics such as empathy, communication, and trust.
5	Talk it out	Teaches strategies for resolving conflict peacefully and addresses the negative consequences of conflict and violence.
6	Analyze this	Helps boys develop critical thinking skills to enable them to analyze messages they see in the media and discuss gender stereotypes presented in the media.
7	Community service project delivery	Participants implement their community service project (if the B4R and WTML programs are active at the same school, the community service project is done as a joint activity).
8	Social media smarts	Focuses on teaching boys how to navigate the internet and social media safely and responsibly.
9	High school prep	Aims to help the boys feel ready for high school by teaching prioritization and time management strategies.

Altogether, the YEP aims to promote the development of participants’ skills and knowledge required for developing a stronger sense of self, healthier friendships and peer relationships, and better decision-making skills. YEP youth participate in group mentoring sessions that encourage them to talk about personal struggles and engage in activities that promote positive social interactions and meaningful peer relationships (Kuperminc et al., 2020). YEP also offers program lessons on how to develop better relationships with oneself and others. By participating in these sessions, it is intended that YEP youth form healthier relationships, including friendships. In particular, the program’s lessons on personal values (module 3), handling conflict with peers (module 5), and healthy/positive friendships (module 4) seek to equip YEP participants with the knowledge and skills needed to exercise boundaries with their peers, understand their personal values, and ultimately build better, healthier, and more supportive relationships with their peers.

Procedure

A prospective longitudinal two-group quasi-experimental design was used in the current study to examine the effects of the YEP on youth’s perceptions of friend support. In the current study, random assignment into the TX and CTL groups was not possible for the students and/or schools in the study due to the YWCA’s method of school selection and the voluntary nature of the program. All 7^th grade students who were recruited to participate in the study were offered the opportunity to participate in the YWCA YEP (∼2,425 students from 93 classrooms)². As such, students self-selected into the TX and CTL group; students who enrolled in the YEP were assigned to the TX group and students who did not enroll in the YEP were assigned to the CTL group. Data were collected from four cohorts of students; refer to Table 2 for survey schedule and student cohorts. Between Fall 2017 and Spring 2019, data were collected from students in 52 separate YEPs across metro Vancouver (26 WTML, 26 B4R) and their classmates who were not registered in a YEP³.

Table 2.

Overview of Study Attrition.

Wave of data collection	Cohort	N pre-test	N post-test	N follow-up
1	Fall 2017	TX: 57	TX: 51	TX: 15
1	Fall 2017	CTL: 76	CTL: 66	CTL: 15
2	Spring 2018	TX: 23	TX: 21	TX: 14
2	Spring 2018	CTL: 21	CTL: 19	CTL: 2
3	Fall 2018	TX: 118	TX: 109	TX: 37
3	Fall 2018	CTL: 261	CTL: 239	CTL: 53
4	Spring 2019	TX: 83	TX: 68	TX: 24
4	Spring 2019	CTL: 159	CTL: 137	CTL: 43
Total N by group		TX: 281	TX: 249	TX: 90
Total N by group		CTL: 517	CTL: 461	CTL: 113
Total N		798	710	203
% attrition from previous time point		/	11%	71%

Study participants (TX and CTL) were grade 7 youth from the 31 schools that offered the YWCA YEP between Fall 2017 and Spring 2019. See Table A1 in the Appendix for the survey administration schedule. All students were sent home with an information package and parent consent form and were required to return a signed parental consent form to participate in the study. Teachers from participating classes were responsible for collecting signed parental consent forms, which were collected by the research team before administration of the pre-test survey.

As shown in Table 2, the study included three surveys. All three surveys included the same set of open- and closed-ended questions, except for the demographic/background information (Survey 1 only) and program satisfaction questions (Survey 2 only). The survey questions asked students about various aspects of their school experience, such as their friendships, self-esteem/self-concept, fears/concerns about high school, and school connectedness. Survey 1 (grade 7 pre-test) and Survey 2 (grade 7 post-test) were administered during class time. Each time, a trained research assistant read aloud a standardized explanation of the study and provided instructions on how to properly fill out the survey. The surveys required approximately 20–30 minutes each to complete; following both surveys, a random draw was held in each classroom in which participants had the chance to win one of two $20 Amazon gift cards. Survey 3 (grade 8 follow-up) was administered online, and participants were contacted by both mail and e-mail (if the latter was available) six months following their transition to high school. All participants who completed Survey 3 were given a $10 e-gift card.

Permission to conduct the study was obtained from the YWCA YEP, and permission to administer the survey was received from the Superintendents of the participating school districts. Ethics approval was received from the Simon Fraser University Office of Research Ethics.

Characteristics of the Analytic Sample

Data were collected from 31 schools across metro Vancouver and baseline (pre-test) surveys were completed by 798 students.⁴ Post-test surveys were collected from 710 students (89% response rate from pre-test to post-test), and follow-up surveys from 203 students (75% attrition from pre-test to follow-up). There were some notable differences in the demographics of those who completed the follow-up survey and those who did not. Specifically, the sample of ‘completers’ had significantly more females and significantly fewer males (p < .001). Significantly more students who completed the follow-up survey reported that their parents were either married or common-law (p < .05). Additionally, in elementary school, completers had significantly fewer siblings/step-siblings, and spent significantly more day/week in extra-curricular activities (p < .05). See Table A2 in the Appendix for details. The analytic sample used in the current study includes 668 students who responded to the School Success Profile (SSP) Friend Support scale and were missing < 30% of the items on that scale.⁵ Baseline data for the (unadjusted) sample of 668 students who were included in the analysis are described below and shown in Table 3.

Table 3.

Sample Characteristics.

	Full sample (N = 668)	Intervention group (n = 233)	Control group (n = 435)
	N (%)	N (%)	N (%)
Gender
Female	351 (52.5%)	138 (59.2%)	213 (49.0%)
Male	317 (47.5%)	95 (40.8%)	222 (51.0%)
Ethnicity
Caucasian	125 (18.7%)	38 (16.3%)	87 (20.0%)
Visible minority	540 (80.8%)	193 (82.8%)	347 (79.8%)
Missing	3 (0.5%)	2 (0.9%)	1 (0.2%)
Biological parents live in the same house as you?
No	104 (15.6%)	32 (13.7%)	72 (16.6%)
Yes	556 (83.2%)	198 (85.0%)	358 (82.3%)
Missing	8 (1.20%)	3 (1.3%)	5 (1.2%)
Were you born in Canada?
No	179 (26.8%)	59 (25.3%)	120 (27.6%)
Yes	487 (72.9%)	173 (74.3%)	314 (72.2%)
Missing	2 (0.30%)	1 (0.4%)	1 (0.2%)
Are you ever afraid that someone will bully you at school?
No	364 (54.5%)	121 (51.9%)	243 (55.9%)
Yes	283 (42.4%)	105 (45.1%)	178 (40.9%)
Missing	21 (3.1%)	7 (3.0%)	14 (3.2%)
	M (SD)	M (SD)	M (SD)
Number of siblings or step-siblings	1.53 (1.20)	1.66 (1.29)	1.47 (1.15)
Number of times you have changed schools	0.84 (1.05)	0.82 (1.02)	0.85 (1.07)
Number of days per week in extra-curricular activities	2.88 (1.21)	2.85 (1.19)	2.89 (1.22)
Number of days per week in activities during school hours	0.33 (0.79)	0.49 (0.93)	0.25 (0.68)

Of the 668 students in the analytic sample, 233 participated in the YWCA YEP (treatment (TX) group) and 435 students were in the control group (CTL). Most of the study sample was female (53%), identified as a visible ethnic minority (81%) and was born in Canada (73%). On average students had one or two siblings or step-siblings (M = 1.53, SD = 1.20) and reported that both of their biological parents lived in the same house (i.e., were married/common law; 83%). In terms of the school setting, the majority of students had changed schools one time (M = 0.84; SD = 1.05). Nearly one-half of the sample said they were afraid that someone at school would bully them (42%). With respect to organized activities, the sample seldom participated in school-based activities during school hours (M = 0.33; SD = 0.79) but participated in extra-curricular activities 2–3 times per week (M = 2.88; SD = 1.21). After weighting by propensity scores, all baseline differences between the TX and CTL groups were less than 0.10 standardized differences⁶, which suggests negligible imbalance on demographic variables between the TX and CTL group (Rubin, 2001; Stuart, 2010).⁷

Measures

Dependent Variable

Perceptions of Friend Support

The School Success Profile (SSP; Bowen & Richman, 2007) is a validated questionnaire for middle and high school students that is used to assess student attitudes and beliefs about their social environment. The Friend Support scale of the SSP contains five items about the perception of friends as “trustworthy and supportive and as responsive to their needs and feelings” (Bowen & Richman, 2007, p. 10). The scale items include: (a) I can trust my friends, (b) I am able to tell my problems to my friends, (c) I feel close to my friends, (d) I can count on my friends for support, and (e) I can talk to my friends about things that bother me. Students are asked to describe the level of trust and closeness they feel toward their peers on a three-point Likert-type scale (i.e., 1 = not like me, 2 = a little like me, 3 = a lot like me). Scores are summed across the five items, with high scores indicating a greater perception of supportive friendships. In the current study, scale scores ranged from a minimum of 5 to a maximum of 15 and Cronbach’s alphas ranged between 0.80 and 0.86 across the pre-test, post-test and follow-up.⁸ In the current study, the dependent variable is set as a time series/time varying variable (Allison, 2017). As a time series variable, the dependent variable’s association with past values (e.g., the pre-test score) is accounted for in the model without adjusting for it in the model (i.e., including it as a covariate).

Independent Variables

YEP Involvement

Those who took part in the YEP (the TX group) were assigned a code of 1, and youth who did not participate in the program (the CTL group) were coded as 0.

Participant Gender

Participants were coded according to their self-reported gender (female = 1; male = 2). As only one student in the analytic sample identified as a gender other than male or female, they were not included in the analyses.

Time

Short-term outcomes assessed change in adolescents’ perceptions of friend support from pre-test (Time 1) to post-test (Time 2). In the short-term analyses, time was specified as a discrete variable. For long-term outcomes, change in the outcome was examined over three time points (Time 1 = pre-test, Time 2 = post-test, Time 3 = follow-up), and time was specified as a continuous variable to account for the unevenly spaced time periods.

Control Variables

Program Session

The YEP was implemented in schools in both the Spring and the Fall, students who participated in the YEP in the Fall were coded as 1 and students who participated in the YEP in the Spring were coded as 2.

Freedom From Anxiety

A 14-item scale (from the Piers-Harris Self-Concept Scale for Children 2 (PH2; Piers & Herzberg, 2002)) asked students to reflect on variety of specific emotions, including worry, nervousness, shyness, sadness, fear, and a general feeling of being left out of things. The dichotomized responses to each item (0 = no, 1 = yes) were summed to create an index score ranging from zero to 14. Higher scores indicate more freedom from anxiety (low feelings of anxiousness). Prior studies report good reliability (Guerin & Tatlow-Golden, 2019; Piers & Herzberg, 2002). The Cronbach’s alpha in the present study was assessed as 0.822. Anxiety was controlled for as it has been associated with peer rejection (Verduin & Kendall, 2008), fewer friends, and decreased popularity, peer acceptance, and social competence (Kingery et al., 2010; Scharfstein et al., 2011).

Analytic Approach

Data entry and coding were completed in Microsoft Excel and exported into Stata. All analyses were conducted using Stata/SE 17.0.

Missing Data Analysis and Imputation

Missing values on the validated multi-item scales was a concern, as item-nonresponse led to undercounted scale scores. Various descriptive procedures were used to investigate the amount of missingness in the dataset.⁹ Overall, the level of non-response in the multi-item scales was low; the amount of missing data on each of the scales ranged between 0% and 5%. While complete case analysis was considered (Harrell, 2001), we opted to impute missing values for the multi-item scales to preserve the sample size. Little’s test of missing completely at random (MCAR) was conducted on the full dataset and the missing data points were determined to be MCAR.

Imputing Missing Scale Items

Although rigorous imputation methods (e.g., multiple imputation) are typically recommended, Harrell (2001) suggests that single imputation is a viable option when the proportion of missing data is less than 5%. Prior to calculating total scores for the validated measures, items with missing values were imputed through random hotdeck procedures (Schonlau, 2006).¹⁰ In hotdecking, observations with missing data (recipients) are matched to similar observations that have no missing values (donors). This is achieved by selecting categorical variables that are known for both recipient and donor (e.g., demographic characteristics) as class variables, and creating a ‘donor pool’ of exact matches on the selected class variables (Andridge & Little, 2010; Lavrakas, 2008). A donor is then randomly selected from the donor pool and imputation is carried out whereby the recipient’s missing values are replaced with the selected donor’s observed values (Lavrakas, 2008).¹¹ The imputation of missing values was approached cautiously, and we opted to impute for a respondent only in cases where an observation was missing ≤ 30% of a scale’s items.¹²

Propensity Score Analysis

In the absence of random assignment, selection bias is a threat to the internal validity of observational studies because the comparison of TX and CTL group (and subsequent claims about causal inference) are complicated by selection effects (Craig, 2020). In the current study, youth voluntarily opt to participate in the YEP, leading to potential selection bias. Propensity score analysis (PSA) is a statistical analysis technique for non-experimental data that is used to estimate the effect of a treatment while reducing selection bias (Rosenbaum & Rubin, 1983). The goal of PSA is to balance the TX and CTL group on a set of observable variables that are thought to impact selection bias so that causal inference can be made with more certainty (Rosenbaum & Rubin, 1983). PSA entails the use of a propensity score, which is “the conditional probability of assignment to a particular treatment given a vector of observed covariates” (Rosenbaum & Rubin, 1983, p. 41). In other words, it is the predicted probability of group membership to either the TX or CTL group based on a respondent’s observed covariates such as demographic variables or other baseline characteristics (Baser, 2006). While a variety of techniques can be used to estimate propensity scores (see Austin, 2009; Stuart, 2010), the most commonly used approach is logistic regression (Austin, 2009). When logistic regression is used to estimate propensity scores, a binary outcome (e.g., intervention (TX = 1) or no intervention (CTL = 0)) is predicted from a set of covariates that are theorized to influence group membership (Craig, 2020). The resulting propensity score is the estimated probability that an individual will be exposed to the treatment. Propensity scores range between zero and 1; scores closer to 1 indicate a higher probability of being in the treatment group.¹³

In the current study, propensity score weighting was used to rule out any systematic pre-treatment differences between the TX and CTL groups. By weighing each individual by the inverse of their probability of receiving treatment, systematic differences in observed baseline covariates were reduced between the groups (see Table 3). We calculated inverse probability weights (IPW) for each participant. For the TX group, IPWs were calculated as 1/e (where e is the estimated propensity score) and as 1/(1−e) for the CTL group (Lunceford & Davidian, 2004). IPWs were truncated at the 1^st and 99^th percentile in order to trim weights at the extreme ends of the distribution (Chesenaye et al., 2022). To assess the comparability of youth in the TX and CTL groups after weighting, we tested balance diagnostics by comparing the standardized mean differences of covariates before and after weighting (Austin, 2009; Stuart, 2010). A visual examination of box plots and kernel density plots was also conducted to compare the distributions of continuous variables before and after weighting (Austin & Stuart, 2015; Chesenaye et al., 2022). Variables were deemed well-balanced if they had an SMD ≤ 0.10, which is conventionally considered a ‘negligible’ amount of imbalance (Rubin, 2001; Stuart, 2010). Variables that had an SMD > 0.10 after weighting were included in the models as covariates (described below).

Multilevel Modelling

Multilevel data structures, in which units of analysis are clustered within one another, are common in longitudinal school-based research (e.g., data are collected from students within schools and/or over multiple points in time; Peugh, 2010). Statistically, the clustered structure of multilevel data is problematic because it violates the assumption of independent errors; an assumption that is crucial in common data analysis models such as ordinary least squares regression and analysis of variance (Steenbergen & Jones, 2002). Ignoring the correlation between errors can result in smaller standard errors and an increased probability of Type 1 errors; both of which may result in inaccurate conclusions about treatment effects (Hair & Favero, 2019; Steenbergen & Jones, 2002). Multilevel modelling (MLM) is a regression-based approach used to model the relationship between dependent and independent variables while accounting for the correlated nature of the data. In the current study, change in the outcome measure was evaluated through three-level multilevel regression models. Specifically, the models were used to account for correlated residuals across time points (Level 1). Additionally, because students completed multiple surveys (and students’ pre-test responses are likely to be correlated with their responses at subsequent time points), the clustering of repeated measures within each student was accounted for in Level 2. Finally, the clustering of students within schools was accounted for in Level 3. Accounting for clustering at the school level is important because, for example, due to school-level factors, such as teachers, school climate, similar classmates, etc., students’ responses in School A are likely to be more similar than students’ responses at School B and School C. If this clustering is not accounted for at various levels of the model, the correlation of data points will not be captured, and the model estimates will not be accurate. An unstructured correlation structure was used to model the short-term outcomes and a first-order autoregression correlation structure was used to model the long-term outcomes (which includes three time points (pre-test, post-test, and follow-up)).

Additionally, a ‘doubly robust’ approach was used to estimate treatment effects, where IPWs were included in the MLMs to reduce systematic baseline differences between the TX and CTL groups and additional covariate adjustment was used if residual imbalance remained following propensity score weighting (Bang & Robins, 2005; Hullsiek & Louis, 2002; Nguyen et al., 2017). Two variables (‘program session’ and ‘Freedom from Anxiety’) had an SMD > 0.10 and were included in the models as covariates. Before implementing the MLM regressions, all model assumptions were checked and satisfied.

Results

Descriptive Results

Throughout the three time points, the students’ scores on the Friend Support scale remained consistent, with values ranging from 12.24 to 12.80. Refer to Table 4 for an overview of the descriptive statistics related to the Friend Support scale, segmented by time points for the full sample, the TX and CTL groups, as well as by gender.

Table 4.

Friend Support Scale Descriptive Statistics.

	Pre-test (M (SD))	Post-test (M (SD))	Follow-up (M (SD))
Full sample	12.52 (2.33)^a	12.66 (2.47)^a	12.61 (2.44)^b
TX group	12.28 (2.48)^c	12.68 (2.60)^c	12.80 (2.53)^d
CTL group	12.65 (2.25)^e	12.66 (2.40)^e	12.45 (2.37)^f
Males	12.24 (2.36)^g	12.53 (2.45)^g	12.29 (2.56)^h
Females	12.78 (2.26)ⁱ	12.78 (2.49)ⁱ	12.77 (2.37)^j

^an = 668.

^bn = 195.

^cn = 233.

^dn = 87.

^en = 435.

^fn = 108.

^gn = 317.

^hn = 65.

ⁱn = 351.

^jn = 130.

Short-Term Impact

An MLM was used to examine students’ perceptions of friend support in elementary school, as well as the short-term impact of the YEP on students’ perceptions of friend support (i.e., pre-test to post-test). No significant main effects were found for YEP involvement (B = −0.316, z = −1.34, p = .18) or time (B = −0.190, z = −1.13, p = .26). These findings indicate that the TX group did not differ from the CTL group at baseline (pre-test) with respect to perceived friend support, and youths’ perceptions of friend support did not meaningfully differ from pre-test to post-test. Interestingly, male participants reported significantly lower levels of friend support compared to their female counterparts, indicating a gender difference in the perception of friend support (B = −0.570, z = −2.45, p < .05).

Additionally, results indicate no significant two-way interaction between YEP involvement and time (B = 0.316, z = 1.02, p = .31), indicating that YEP participants did not differ in their perceptions of friend support at post-test compared to those in the CTL group. The two-way interaction of YEP involvement and gender was also not significant (B = −0.772, z = −1.52, p = .13), suggesting no significant difference in perceptions of friend support between male and female YEP participants at baseline. The three-way interaction between YEP involvement, time, and gender was also not statistically significant (B = 0.110, z = 0.19, p = .85), meaning there was no difference in perceptions of friend support over time between male and female program participants. With respect to the covariates, students who reported more freedom from anxiety (low feelings of anxiousness) had significantly higher perceptions of friend support (p < .001). The random-effects parameters indicate no variation at the primary school level above and beyond what is already controlled for in the model, and that individual-level differences have a larger effect on perceptions of friend support. See Table 5.

Table 5.

Multilevel Regression for Short-Term Outcomes (N = 668).

Dependent Variable	Coef.	Robust S.E.	z	P > \|z\|	95% C.I.
Dependent Variable	Coef.	Robust S.E.	z	P > \|z\|	Lower	Upper
YEP involvement	−0.316	.235	−1.34	0.179	−0.776	0.145
Time	−0.190	.169	−1.13	0.260	−0.522	0.141
Gender	−0.570	.233	−2.45	0.014*	−1.026	−0.114
YEP involvement*time	0.316	.311	1.02	0.310	−0.293	0.925
YEP Involvement*gender	−0.772	.509	−1.52	0.130	−1.770	0.226
Time*gender	0.376	.259	1.45	0.146	−0.131	0.884
YEP Involvementtimegender
Tx2Male	0.110	.566	0.19	0.846	−0.999	1.219
Program session	−0.141	.260	−0.54	0.588	−0.650	0.368
Freedom from Anxiety	0.123	.028	4.44	0.000***	0.069	0.178
Constant	11.874	.358	33.14	0.000***	11.17	12.58

Random-effects parameters	Estimate	Robust S.E.	95% C.I.
Random-effects parameters	Estimate	Robust S.E.	Lower	Upper
Elementary school: Identity
var(_cons)	0.374	.158	0.163	0.855
Participant ID: Identity
var(_cons)	1.671	.144	1.411	1.979
Residual: Unstructured
var(e1)	3.401	.200	3.031	3.817
var(e2)	4.274	.269	3.779	4.834
cov(e1,e2)	0.956	.194	0.576	1.335

Log pseudolikelihood = -5762.707; Wald chi2(9) = 101.28; Prob > chi2 = 0.0000.

*p < .05, **p < .01, ***p < .001.

Long-Term Impact

The long-term impact of the YEP was examined for the subset of participants who completed the follow-up survey after transitioning to high school (grade 8; N = 195 (CTL = 108¹⁴; TX = 87¹⁵). Due to the considerable attrition of the study sample on the follow-up survey, the findings with respect to the long-term impact of the YEP should be considered preliminary. Similar to the above, a multilevel regression model was used to examine the relationship between program participation, time (pre-test to follow-up), gender, and perceptions of friend support. As shown in Table 6, the analysis shows a significant main effect of gender (B = −1.56, z = −2.84, p < .01), with males in the sample reporting significantly lower levels of support from their friends than did females in the sample. This finding suggests that gender plays a crucial role in determining the overall levels of friend support reported by youth. Significant main effects were also found for YEP involvement (B = −1.072, z = −2.06, p < .05); this finding indicates that the TX group had a lower level of perceived support from their friends compared to the CTL group at baseline (pre-test). Further, the perception of friend support decreased over time across all youth, irrespective of the group they belonged to (i.e., TX or CTL; B = −0.316, z = −1.96, p = .05). This finding suggests that participants felt less supported by their friends as they transitioned to high school, regardless of whether they participated in the YEP.

Table 6.

Multilevel Regression for Long-Term Outcomes (N = 195).

Dependent Variable	Coef.	Robust S.E.	z	P > \|z\|	95% C.I.
Dependent Variable	Coef.	Robust S.E.	z	P > \|z\|	Lower	Upper
YEP involvement	−1.072	.520	−2.06	0.039*	−2.090	−0.053
Time	−0.316	.161	−1.96	0.050*	−0.632	0.001
Gender	−1.555	.548	−2.84	0.005**	−2.629	−0.482
YEP involvement*time	0.448	.153	2.93	0.003**	0.148	0.747
YEP involvementtimegender
ctl#Male	0.349	.224	1.56	0.119	−0.089	0.788
tx#Male	0.339	.210	1.61	0.106	−0.073	0.751
Program session	−0.134	.363	−0.37	0.712	−0.846	0.578
Freedom from Anxiety	0.147	.052	2.84	0.005**	0.045	0.249
Constant	12.286	.777	15.8	0.000***	10.76	13.81

Random-effects parameters	Estimate	Robust S.E.	95% C.I.
Random-effects parameters	Estimate	Robust S.E.	Lower	Upper
Elementary school: Identity
var(_cons)	0.650	.319	0.249	1.702
Participant ID: Identity
var(_cons)	1.601	.837	0.574	4.461
Residual: AR(1)
rho	0.226	.140	−0.059	0.477
var(e)	3.666	.756	2.447	5.492

Log pseudolikelihood = -2597.2; Wald chi2(8) = 33.30; Prob > chi2 = 0.0001.

*p < .05, **p < .01, ***p < .001.

There was also a significant two-way interaction between YEP involvement and time (B = 0.448, z = 2.93, p < .01), indicating that YEP participants experienced a significant increase in the perception of friend support over time, compared to their counterparts in the CTL group. This finding is particularly noteworthy as it emphasizes the positive influence of YEP involvement on adolescents’ perceptions of friend support as they transition from 7^th grade to high school. The analysis did not show any other significant interaction effects. The findings indicate that higher freedom of anxiety (low feelings of anxiousness) is associated with higher perception of friend support (p < .01). Similar to the short-term findings, the random-effects parameters suggest that there is no variation at the primary school level above and beyond what is already controlled for in the model.

Discussion

The current study aimed to assess the impact of the YEP afterschool group mentoring programs on adolescents’ perceptions of friend support. In addition, we assessed whether the impact of the YEP differs for males and females. The findings indicate that grade 7 students who participated in the YEP had significantly higher perceptions of friend support in the long term compared to those who did not participate (i.e., positive and statistically significant two-way interactions between YEP involvement and time). However, this effect was only observed in the long term (after transitioning to high school), not the short term (immediately after the program). Program effects did not vary by gender (i.e., in the TX group, there was no significant difference in perceptions of peer support between B4R and WTML participants), indicating that the YEP is equally effective at fostering perceptions of supportive friendships for both males and females. Altogether, the results of this study show beneficial effects on friendships/social skills for group mentoring programs delivered in an after-school setting.

The lack of program effects with respect to perceived friend support in the shorter term compared with the promising effects shown for friend support in the longer term were somewhat unexpected. Yet, given the YEP’s overarching goal to improve the high school transition process, these findings are encouraging. It may be that students are better able to apply the skills and knowledge learned with respect to healthy friendships in the new, unfamiliar high school setting than they are when confined to the environment of their existing grade 7 class. More specifically, as the group mentoring sessions specifically encourage the acquisition of skills to promote positive social interactions and enhance meaningful relationships with peers, students may find these skills easier to apply when meeting and interacting with new peers rather than when interacting with existing peers with whom they have relationships established prior to joining the YEP. For example, in examining the curriculum content, sessions 3, 4, and 5 focus on consideration of personal values, making and maintaining positive friendships, and learning how to handle and resolve conflict in a respectful/peaceful manner. During the adolescent years and the transition to high school, personal values become more important. Making new friends and forming friendships is a significant part of the high school experience. Students may be more likely to apply the YEP curriculum when it directly relates to their own experiences.

Although gender was not a significant predictor of the program’s impact, both regression models showed that gender had a main effect on students’ perception of friend support. Specifically, males in the full sample reported significantly lower perceptions of friend support compared to females, both at the end of grade 7 and six months after transitioning to high school. This finding is not altogether surprising, as some research has found that across the period of adolescence (but particularly in early adolescence) males perceive less support in their close friendships and report lower quality friendships than females (Davis, 2019; de Goede et al., 2009; Floody et al., 2019; Way, 2013; Way & Greene, 2006). Nevertheless, this finding is important because studies have indicated that developing and maintaining strong social connections during high school can have a significant positive impact on an individual’s overall emotional and social well-being (e.g., Benner et al., 2017; Morin et al., 2013). Furthermore, a stable network of friends is linked to better academic outcomes, including higher levels of engagement, achievement, and a lower probability of dropping out (e.g., Benner et al., 2017; Carbonaro & Workman, 2013). The possibility that males may experience more difficulties adapting to changes in their social circles around the transition to high school is worth considering. Challenges with adapting to social changes in high school could potentially result in more negative consequences, which can have a long-lasting impact on their high school experience (e.g., Benner et al., 2017; Felmlee et al., 2018). Together, these findings underscore the need for greater attention on the social well-being of high school students, particularly males, and the importance of developing strategies to help them adapt to changes in their social networks.

Overall, the results of this study enhance our understand of youth development by providing empirical evidence for gender differences and social outcomes such as perceptions of friend support. The findings indicate notable differences in how grade 7 males and females perceive friend support, emphasizing the importance of addressing the unique needs of these genders to promote supportive friendships and foster the development of knowledge and skills with respect to healthy relationships. The study supports previous research that highlights the crucial role of individual differences and contextual factors in effective mentoring practices (e.g., Liang et al., 2013). Specifically, it suggests that gender should be considered when developing strategies to promote friend support. Additionally, the study’s findings could help to inform school policies or programs that aim to improve friend support among students.

Altogether, the current study suggests that group mentoring programs in the afterschool setting, like WTML and B4R, have the potential to help youth strengthen their friendships in the longer term (see also Cawood & Wood, 2014; Kuperminc, 2016). These findings add to research on similar programs which suggest their effectiveness at supporting youth social development by promoting social and emotional learning and enhancing interpersonal skills (Durlak et al., 2010, 2011; Himmelrich, 2012; Pelcher & Rajan, 2016). This research contributes to the limited body of methodologically rigorous and longitudinal evaluations on group-based mentoring for adolescents and fills an important gap in the literature by examining group-level gender differences in youth outcomes throughout the transition to high school.

Overall, more empirical research is required to develop effective interventions that help youths transition successfully to high school (Bharara, 2020; Donaldson et al., 2023). Further, more research is needed with respect to the specific components of group mentoring programs that are linked to successful youth outcomes (e.g., program content, group format, quality or quantity of mentors). Future research should continue to examine how gender affects social outcomes during the transition to high school. Future research should also compare the effects of gender-tailored and non-gender tailored group mentoring programs to determine which approach has more beneficial outcomes for males and females. There is also a greater need for rigorous longitudinal research that follows youth for an extended period before and after the transition to high school.

Limitations

The results indicating the long-term positive impact of YEP should be interpreted with caution due to the disproportionate representation of girls in the sample. Additionally, it is noteworthy that at each time point, girls reported more friendship support than boys. This raises questions about whether the observed effects were truly due to the YEP or if they were influenced by the over-representation of girls in the long-term sample. Relatedly, another limitation of the study’s data is the high rate of non-response on the follow-up survey and the potential influence of attrition bias on the findings regarding the long-term program impact. Comparison analyses showed some significant differences in the characteristics of those who completed the follow-up survey and those who did not (i.e., gender, family structure, number of siblings, participation in extra-curricular activities, and participation in school-based activities). As participants in the follow-up assessment were more engaged in school-based and extracurricular activities, this may have influenced their reported levels of friendship support. Further bias may have occurred if participants who completed the follow-up survey were systematically different on unobserved characteristics than those who did not. For example, if youth who were struggling with adjusting to the transition to high school were less likely to complete the follow-up survey than those who were adjusting well, attrition may skew the representativeness of the sample and subsequently skew the validity of the long-term program impacts. Due to the attrition at follow-up, the findings concerning long-term program impacts should be considered preliminary and be interpreted with caution.

Second, although the YEPs are gender-based programs and WTML typically targets females while B4R typically targets males, it is important to note that youth can choose which program to enroll in based on their gender identity. As only one student in the analytic sample identified as a gender other than male or female, it was not possible to include an analysis of the group of students who did not identify as either male or female and the findings presented herein may not be generalizable to youth of all genders. Research has consistently found that non-binary and gender minority youths face higher risks of experiencing challenges with social adjustment, academic achievement, and mental well-being in school settings (e.g., Colvin et al., 2019; Durbeej et al., 2019; Kelley et al., 2022; Kosciw et al., 2018). As such, future research on the effectiveness of high school transition programs, it is important to include analyses of the experiences of sexual/gender minority youths, if possible. Relatedly, a limitation of the current design is that it did not consider the intersecting identities of its participants. Recent large scale examinations have affirmed intersectionality’s relevance in predicting youths’ general wellbeing, mental health, and vulnerability to global crises (Gottert et al., 2023; Kern et al., 2020; McIsaac et al., 2021). While there is limited evidence of its direct impact on peer relationships, there is some; for example, Exner‐Cortens and colleagues (2022) suggest that the gender and ethnic minority status of teenage males can intersect to amplify hegemonic masculinity, disproportionately negatively impacting their formation of close friendships in the transition to high school.

Next, MLMs were used to account for the clustered structure of the data. However, the school-level results of the model were not extensively examined or discussed in the paper (i.e., the research questions were not focused on how the dependent variable varied at the school-level). Since the YEP is offered across multiple schools, the success of the programs may be dependent on school-level factors, such as the amount of time teachers spend preparing their students for high school, talking about peer relationships, and so forth. Similarly, because each YEP is delivered by a different set of program facilitators, the success of the programs may be influenced by program-level factors such as the differential quality and experience of mentors. Additionally, it is possible that the program’s curriculum is implemented differently across schools (e.g., spending more time on one activity and running out of time to finish others). Low implementation fidelity, weak mentoring skills, and/or differences in the environment of each school might influence the effectiveness of the YEP on youth outcomes. The impact of these factors on program outcomes should be examined in future studies.

Finally, it is important to note that the study has an additional constraint as it remains unclear which specific friends the participants referred to while responding to the SSP Friend Support Scale (e.g., school friends, neighbourhood/non-school friends, YEP friends, friends in other extra-curricular activities). This could introduce some ambiguity into the findings, as different individuals may have different ideas of what constitutes a “friend” and what level of support they provide. To address this limitation, we recommend that future research include mechanisms that help with identifying the members of the respondents’ friendship network. For example, it may be useful to explore whether the YEP reinforces adolescent’s social circle among YEP participants and/or other friends in their social network. Gaining a deeper understanding of the specific friendships networks that the YEP helps to strengthen would be beneficial. Further understanding which friendship networks the YEP does not help to strengthen would provide an opportunity for the program to incorporate new content that could strengthen their outcomes with respect to friendships.

Conclusion

Research has shown that adolescents face many difficulties when transitioning to high school, especially in regard to their social lives and forming healthy relationships with friends. The negative effects of this experience can be long-lasting, highlighting the importance of providing support to adolescents during this time. The findings from this study indicate that group-based mentoring programs implemented in an afterschool setting show promise at improving adolescents’ perceptions of their friendships following the transition to high school. Significant gaps remain before evidence-informed lessons can be drawn about what works for group-based mentoring programs for youth and their friendships, particularly with respect to how program impacts differ for males and females. Future research should continue to explore the moderating impact of gender on social outcomes throughout adolescence (particularly during the transition to high school), as well as the effects of gender-tailored group mentoring programs.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Jessica Bouchard

Notes

Appendix

Author Biographies

Jessica Bouchard is a Research Scientist at the New Brunswick Institute for Research, Data and Training. Her work primarily focuses on examining the impact of prevention-focused programs in the fields of crime/delinquency, health, and education. Ms. Bouchard’s recent work has examined the impact of a group mentoring program on delinquency, a health coaching program on health-related outcomes, an experiential learning program on high school students’ interest in careers in the energy sector, and intimate partner violence intervention programs on attitudes and behaviors toward violence. She has also worked on meta-analyses that examine the effectiveness of aftercare programs, a sexual violence prevention campaign, and dating violence prevention programs.

Jennifer Wong is a Professor in the School of Criminology at Simon Fraser University. Her work employs applied, quantitative, and qualitative methods to study issues of delinquency/crime prevention and intervention, focusing primarily on evaluation and policy analysis in the areas of crime prevention and crime control policy. Dr. Wong’s recent work has examined gang prevention/intervention programs, intimate partner violence intervention programs, a sexual violence prevention campaign, the impact of prior deportation on recidivism among illegal immigrants, and meta-analyses on aftercare/resettlement programs, wilderness therapy programs, and dating violence prevention programs.

References

Akos

Galassi

J. P.

(2004). Middle and high school transitions as viewed by students, parents, and teachers. Professional School Counseling, 7(4), 212–221. https://www.jstor.org/stable/42732584

Allison

(2017). Longitudinal data analysis using stata. Statistical horizons. chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/. https://statisticalhorizons.com/wp-content/uploads/LDA-Stata-Sample-Materials-4.pdf

Anderson

L. W.

Jacobs

Schramm

Splittgerber

(2000). School transitions: Beginning of the end or a new beginning? International Journal of Educational Research, 33(4), 325–339. https://doi.org/10.1016/S0883-0355(00)00020-3

Andridge

Little

(2010). A review of hot deck imputation for survey non-response. International Statistics Review, 78(1), 40–64. https://doi.org/10.1111/j.1751-5823.2010.00103.x

Austin

(2009). Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Statistics in Medicine, 28(25), 3083–3107. https://doi.org/10.1002/sim.3697

Austin

Stuart

(2015). Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine, 34(28), 3661–3679. https://doi.org/10.1002/sim.6607

Bang

Robins

(2005). Doubly robust estimation in missing data and causal inference models. Biometrics, 61(4), 962–973. https://doi.org/10.1111/j.1541-0420.2005.00377.x

Baser

(2006). Too much ado about propensity score models? Comparing methods of propensity score matching. Value in Health, 9(6), 377–385. https://doi.org/10.1111/j.1524-4733.2006.00130.x

Benner

A. D.

(2011). The transition to high school: Current knowledge, future directions. Educational Psychology Review, 23(3), 299–328. https://doi.org/10.1007/s10648-011-9152-0

10.

Benner

A. D.

Boyle

A. E.

Bakhtiari

(2017). Understanding students’ transition to high school: Demographic variation and the role of supportive relationships. Journal of Youth and Adolescence, 46(10), 2129–2142. https://doi.org/10.1007/s10964-017-0716-2

11.

Benner

A. D.

Graham

(2009). The transition to high school as a developmental process among multiethnic urban youth. Child Development, 80(2), 356–376. https://doi.org/10.1111/j.1467-8624.2009.01265.x

12.

Bharara

(2020). Factors facilitating a positive transition to secondary school: A systematic literature review. International Journal of School & Educational Psychology, 8(sup1), 104–123. https://doi.org/10.1080/21683603.2019.1572552

13.

Blakemore

S.-J.

Mills

(2014). Is adolescence a sensitive period for sociocultural processing? Annual Review of Psychology, 65(1), 187–207. https://doi.org/10.1146/annurev-psych-010213-115202

14.

Bowen

Richman

(2007). The school success profile: A results-focused approach to assessment and intervention planning. The University of North Carolina at Chapel Hill, School of Social Work. https://www.academia.edu/20861169/The_School_Success_Profile_A_Results_Focused_Approach_to_Assessment_and_Intervention_Planning

15.

Burton

(2020). Cross-age peer mentoring: A meta-analysis. Dissertation. University of Massachusetts Boston.

16.

Cantin

Boivin

(2004). Change and stability in children’s social network and self-perceptions during transition from elementary to junior high school. International Journal of Behavioral Development, 28(6), 561–570. https://doi.org/10.1080/01650250444000289

17.

Carbonaro

Workman

(2013). Dropping out of high school: Effects of close and distant friendships. Social Science Research, 42(5), 1254–1268. https://doi.org/10.1016/j.ssresearch.2013.05.003

18.

Cawood

Wood

(2014). Group mentoring: The experience of adolescent mentees on probation. Social Work with Groups, 37(3), 213–229. https://doi.org/10.1080/01609513.2013.862895

19.

Cham

West

(2016). Propensity score analysis with missing data. Psychological Methods, 21(3), 427–445. https://doi.org/10.1037/met0000076

20.

Chein

Albert

O’Brien

Uckert

Steinberg

(2011). Peers increase adolescent risk taking by enhancing activity in the brain’s reward circuitry. Developmental Science, 14(2), F1–F10. https://doi.org/10.1111/j.1467-7687.2010.01035.x

21.

Chesenaye

Stel

Tripepi

Dekker

Zoccali

Jager

(2022). An introduction to inverse probability of treatment weighting in observational research. Clinical Kidney Journal, 15(1), 14–20. https://doi.org/10.1093/ckj/sfab158

22.

Christensen

K. M.

Hagler

M. A.

Stams

G.-J.

Raposa

E. B.

Burton

Rhodes

J. E.

(2020). Non-specific versus targeted approaches to youth mentoring: A follow-up meta-analysis. Journal of Youth and Adolescence, 49(5), 959–972. https://doi.org/10.1007/s10964-020-01233-x

23.

Colvin

Egan

J. E.

Coulter

(2019). School climate & sexual and gender minority adolescent mental health. Journal of Youth and Adolescence, 48(10), 1938–1951. https://doi.org/10.1007/s10964-019-01108-w

24.

Craig

(2020). Propensity score matching and generalized boosted modeling in the context of model misspecification: A simulation study. James Madison University. [Unpublished master’s thesis].

25.

Davis

(2019). Gender differences in friendship quality across adolescence Master’s thesis. University of Virginia.

26.

De Goede

Branje

Meeus

(2009). Developmental changes and gender differences in adolescents’ perceptions of friendships. Journal of Adolescence, 32(5), 1105–1123. https://doi.org/10.1016/j.adolescence.2009.03.002

27.

Delgado

Serna

Martínez

Cruise

(2022). Parental attachment and peer relationships in adolescence: A systematic review. International Journal of Environmental Research and Public Health, 19(3), 1064–1086. https://doi.org/10.3390/ijerph19031064

28.

Donaldson

Moore

Hawkins

(2023). A systematic review of school transition interventions to improve mental health and wellbeing outcomes in children and young people. School Mental Health, 15(1), 19–35. https://doi.org/10.1007/s12310-022-09539-w

29.

DuBois

D. L.

Portillo

Rhodes

J. E.

Silverthorn

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

30.

Duell

Steinberg

Icenogle

Chein

Chaudhary

Di Giunta

Dodge

K. A.

Fanti

K. A.

Lansford

J. E.

Oburu

Pastorelli

Skinner

A. T.

Sorbring

Tapanya

Uribe Tirado

L. M.

Alampay

L. P.

Al-Hassan

S. M.

Takash

H. M. S.

Bacchini

Chang

(2018). Age patterns in risk taking across the world. Journal of Youth and Adolescence, 47(5), 1052–1072. https://doi.org/10.1007/s10964-017-0752-y

31.

Durbeej

Abrahamsson

Papadopoulos

Beijer

Salari

Sarkadi

(2019). Outside the norm: Mental health, school adjustment and community engagement in non-binary youth. Scandinavian Journal of Public Health, 49(5), 529–538. https://doi.org/10.1177/1403494819890994

32.

Durlak

J. A.

Weissberg

R. P.

Dymnicki

A. B.

Taylor

R. D.

Schellinger

K. B.

(2011). The impact of enhancing students’ social and emotional learning: A meta analysis of school‐ based universal interventions. Child Development, 82(1), 405–432. https://doi.org/10.1111/j.1467-8624.2010.01564.x

33.

Durlak

J. A.

Weissberg

R. P.

Pachan

(2010). A meta-analysis of after-school programs that seek to promote personal and social skills in children and adolescents. American Journal of Community Psychology, 45(3–4), 294–309. https://doi.org/10.1007/s10464-010-9300-6

34.

Eby

L. T.

Allen

T. D.

Evans

S. C.

DuBois

D. L.

(2008). Does mentoring matter? A multidisciplinary meta-analysis comparing mentored and non-mentored individuals. Journal of Vocational Behavior, 72(2), 254–267. https://doi.org/10.1016/j.jvb.2007.04.005

35.

Exner‐Cortens

Claussen

Lewis

Orukpe

A. M.

Coupland

. (2022). Friendship quality and ethnocultural boys: An exploratory evaluation of the WiseGuyz Program. Psychology in the Schools, 59(10), 2106–2121. https://doi.org/10.1002/pits.22566

36.

Felmlee

McMillan

Inara Rodis

Osgood

D. W.

(2018). Falling behind: Lingering costs of the high school transition for youth friendships and grades. Sociology of Education, 91(2), 159–182. https://doi.org/10.1177/0038040718762136

37.

Floody

McGuire

Kuldas

Norman

(2019). Friendship quality and gender differences in association with cyberbullying involvement and psychological well-being. Frontiers in Psychology, 10(1723), 1–13. https://doi.org/10.3389/fpsyg.2019.01723

38.

Foulkes

Blakemore

S.-J.

(2016). Is there heightened sensitivity to social reward in adolescence? Current Opinion in Neurobiology, 40, 81–85. https://doi.org/10.1016/j.conb.2016.06.016

39.

Gardner

Steinberg

(2005). Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology, 41(4), 625–635. https://doi.org/10.1037/0012-1649.41.4.625

40.

Garringer

McQuillin

McDaniel

(2017). Examining youth mentoring services across America: Findings from the 2016 mentoring program survey. Mentor: The national mentoring partnership. https://www.mentoring.org/new-site/wp-content/uploads/2017/07/Mentor-Survey-Report_FINAL_small.pdf

41.

Gottert

Vieitez

Nevárez

Austrian

Muluve

Patel

S. K.

Saggurti

Bajracharya

DeMulder

Soler

Ngo

T. D.

(2023). The multidimensional, intersecting impacts of covid-19 on young people’s lives: Evidence from cross-sectional surveys in Mexico, India, and Kenya. Journal of Adolescent Health, 73(5), 820–829. https://doi.org/10.1016/j.jadohealth.2023.06.016

42.

Grills-Taquechel

Norton

Ollendick

(2010). A longitudinal examination of factors predicting anxiety during the transition to middle school. Anxiety, Stress & Coping, 23(5), 493–513. https://doi.org/10.1080/10615800903494127

43.

Guerin

Tatlow-Golden

(2019). How valid are measures of children’s self-concept/self-esteem? Factors and content validity in three widely used scales. Child Indicators Research, 12(5), 1507–1528. https://doi.org/10.1007/s12187-018-9576-x

44.

Haddock

Weiler

Lee

Henry

Lucas-Thompson

Zimmerman

Krafchick

Fredrickson

G. F.

Yetz

Youngblade

(2020). Does organizing mentor-mentee matches into small groups enhance treatment effects in a site-based mentoring program for adolescents? Results of a randomized controlled trial. Journal of Youth and Adolescents, 49(9), 1864–1882. https://doi.org/10.1007/s10964-020-01267-1

45.

Hair

J. F.

Jr. Fávero

(2019). Multilevel modeling for longitudinal data: Concepts and applications. RAUSP Management Journal, 54(4), 459–489. https://doi.org/10.1108/RAUSP-04-2019-0059

46.

Harrell

F. E.

(2001). Missing data. In Regression modeling strategies (pp. 41–52). Springer.

47.

Himmelrich

(2012). Forming relationships to benefit schooling: How students and mentors perceive benefits of afterschool mentoring. Trinity College. [Unpublished master’s thesis].

48.

Hullsiek

K. H.

Louis

(2002). Propensity score modeling strategies for the causal analysis of observational data. Biostatistics, 3(2), 179–193. https://doi.org/10.1093/biostatistics/3.2.179

49.

Kelley

Pullen Sansfaçon

Gelly

Chiniara

Chadi

(2022). School factors strongly impact transgender and non-binary youths’ well-being. Children, 9(10), 1520–1534. https://doi.org/10.3390/children9101520

50.

Kern

M. R.

Duinhof

E. L.

Walsh

S. D.

Cosma

Moreno-Maldonado

Molcho

Currie

Stevens

G. W. J. M.

(2020). Intersectionality and adolescent mental well-being: A cross-nationally comparative analysis of the interplay between immigration background, socioeconomic status and gender. Journal of Adolescent Health, 66(6S), S12–S20. https://doi.org/10.1016/j.jadohealth.2020.02.013

51.

Kingery

J. N.

Erdley

C. A.

Marshall

K. C.

Whitaker

K. G.

Reuter

T. R.

(2010). Peer experiences of anxious and socially withdrawn youth: An integrative review of the developmental and clinical literature. Clinical Child and Family Psychology Review, 13(1), 91–128. https://doi.org/10.1007/s10567-009-0063-2

52.

Kosciw

J. G.

Greytak

E. A.

Zongrone

A. D.

Clark

C. M.

Truong

(2018). The 2017 National School Climate Survey: The experiences of lesbian, gay, bisexual, transgender, and queer youth in our nation’s schools. Gay, Lesbian and straight education network (GLSEN). https://eric.ed.gov/?id=ED590243

53.

Krammer

Schrank

Pollak

Stiehl

Nater

Woodcock

(2023). Early adolescents’ perspectives on factors that facilitate and hinder friendship development with peers at the time of school transition. Journal of School Psychology, 98, 113–132. https://doi.org/10.1016/j.jsp.2023.03.001

54.

Kuperminc

Chan

Hale

Joseph

Delbasso

(2020). The role of school-based group mentoring in promoting resilience among vulnerable high school students. American Journal of Community Psychology, 65(1–2), 136–148. https://doi.org/10.1002/ajcp.12347

55.

Kuperminc

Deutsh

(2021). Group mentoring: National mentoring resource center model review. National Mentoring Resource Center. https://nationalmentoringresourcecenter.org/wp-content/uploads/2021/09/GroupMentoringReview.pdf

56.

Kuperminc

G. P.

(2016, January). Group mentoring. National mentoring resource center. Available from. https://www.nationalmentoringresourcecenter.org/index.php/whatworks-in-mentoring/model-and-population-reviews.html?xml:id=121

57.

Lavrakas

(2008). Hot deck imputation. Encyclopedia of survey research methods. Sage Publications, Inc.

58.

Liang

Bogat

Duffy

(2013). Gender in mentoring relationships. In Dubois

D. L.

Karcher

M. J.

(Eds.), Handbook of youth mentoring (pp. 159–174). Sage Publications.

59.

Lunceford

Davidian

(2004). Stratification and weighting via the propensity score in estimation of causal treatment effects: A comparative study. Statistics in Medicine, 23(19), 2937–2960. https://doi.org/10.1002/sim.1903

60.

Lundborg

(2006). Having the wrong friends? Peer effects in adolescent substance use. Journal of Health Economics, 25(2), 214–233. https://doi.org/10.1016/j.jhealeco.2005.02.001

61.

Huebner

(2008). Attachment relationships and adolescents life satisfaction: Some relationships matter more to girls than boys. Psychology in the Schools, 45(2), 177–190. https://doi.org/10.1002/pits.20288

62.

Mainzer

Apajee

Nguyen

Carlin

Lee

(2021). A comparison of multiple imputation strategies for handling missing data in multi-item scales: Guidance for longitudinal studies. Statistics in Medicine, 40(21), 4660–4674. https://doi.org/10.1002/sim.9088

63.

McIsaac

M. A.

Reaume

Phillips

S. P.

Michaelson

Steeves

Davison

C. M.

Vafaei

King

Pickett

(2021). A novel application of a data mining technique to study intersections in the social determinants of mental health among young Canadians. SSM - Population Health, 16, 100946. https://doi.org/10.1016/j.ssmph.2021.100946

64.

Morin

A. J. S.

Maiano

Marsh

H. W.

Nagengast

Janosz

(2013). School life and adolescents’ self-esteem trajectories. Child Development, 84(6), 1967–1988. https://doi.org/10.1111/cdev.12089

65.

Newman

(2014). Missing data: Five practical guidelines. Organizational Research Methods, 17(4), 372–411. https://doi.org/10.1177/1094428114548590

66.

Ng-Knight

Shelton

K. H.

Riglin

Frederickson

McManus

I. C.

Rice

(2019). Best friends forever’? Friendship stability across school transition and associations with mental health and educational attainment. British Journal of Educational Psychology, 89(4), 585–599. https://doi.org/10.1111/bjep.12246

67.

Nguyen

Collins

Spence

Daures

J.-P.

Devereaux

Landais

Le Manach

(2017). Double-adjustment in propensity score matching analysis: Choosing a threshold for considering residual imbalance. BMC Medical Research Methodology, 17(1), 78–85. https://doi.org/10.1186/s12874-017-0338-0

68.

Pelcher

Rajan

(2016). After-school program implementation in urban environments: Increasing engagement among adolescent youth. Journal of School Health, 86(8), 585–594. https://doi.org/10.1111/josh.12411

69.

Peugh

(2010). A practical guide to multilevel modeling. Journal of School Psychology, 48(1), 85–112. https://doi.org/10.1016/j.jsp.2009.09.002

70.

Piers

E. V.

Herzberg

D. S.

(2002). Piers-Harris children’s self-concept scale-Second Edition manual. Western Psychological Services.

71.

Poon

C. Y. S.

Christensen

K. M.

Rhodes

J. E.

(2021). A meta-analysis of the effects of mentoring on youth in foster care. Journal of Youth and Adolescence, 50(9), 1741–1756. https://doi.org/10.1007/s10964-021-01472-6

72.

Pratt

George

(2005). Transferring friendship: Girls’ and boys’ friendships in the transition from primary to secondary school. Children & Society, 19(1), 16–26. https://doi.org/10.1002/CHI.830

73.

Raposa

Rhodes

Stams

Card

Burton

Schwartz

Sykes

Kanchewa

Kupersmidt

Hussain

(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

74.

Rice

Frederickson

Seymour

(2011). Assessing pupil concerns about transition to secondary school. British Journal of Educational Psychology, 81(Pt 2), 244–263. https://doi.org/10.1348/000709910X519333

75.

Rombach

Gray

Jenkinson

Murray

Rivero-Arias

(2018). Multiple imputation for patient reported outcome measures in randomised controlled trials: Advantages and disadvantages of imputing at the item, subscale or composite score level. BMC Medical Research Methodology, 18(1), 87–103. https://doi.org/10.1186/s12874-018-0542-6

76.

Rosenbaum

(2010). Design of observational studies. Springer.

77.

Rosenbaum

Rubin

(1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.2307/2335942

78.

Rosenbaum

Rubin

(1984). Reducing bias in observational studies using subclassification on the propensity score. Journal of the American Statistical Association, 79(387), 516–524. https://doi.org/10.2307/2288398

79.

Rubin

D. B.

(2001). Using propensity scores to help design observational studies: Application to the tobacco litigation. Health Services & Outcomes Research Methodology, 2(3/4), 169–188. https://doi.org/10.1023/a:1020363010465

80.

Rudolph

K. D.

Dodson

J. F.

(2022). Gender differences in friendship values: Intensification at adolescence. The Journal of Early Adolescence, 42(4), 586–607. https://doi.org/10.1177/02724316211051948

81.

Scharfstein

Alfano

Beidel

Wong

(2011). Children with generalized anxiety disorder do not have peer problems, just fewer friends. Child Psychiatry and Human Development, 42(6), 712–723. https://doi.org/10.1007/s10578-011-0245-2

82.

Schonlau

(2006). Stata software package, hotdeckvar.pkg, for hotdeck imputation. https://www.schonlau.net/stata/

83.

Skudrzyk

Zera

D. A.

McMahon

Schmidt

Boyne

Spannaus

R. L.

(2009). Learning to relate: Interweaving creative approaches in group counseling with adolescents. Journal of Creativity in Mental Health, 4(3), 249–261. https://doi.org/10.1080/15401380903192762

84.

Steenbergen

Jones

(2002). Modeling multilevel data structures. American Journal of Political Science, 46(1), 218–237. https://doi.org/10.2307/3088424

85.

Stuart

(2010). Matching methods for causal inference: A review and a look forward. Statistical Science, 25(1), 1–21. https://doi.org/10.1214/09-STS313

86.

Tolan

Henry

Schoeny

Lovegrove

Nichols

(2014). Mentoring programs to affect delinquency and associated outcomes of youth at risk: A comprehensive meta-analytic review. Journal of Experimental Criminology, 10(2), 179–206. https://doi.org/10.1007/s11292-013-9181-4

87.

van Duijvenvoorde

van Hoorn

Blankenstein

(2022). Risks and rewards in adolescent decision-making. Current Opinion in Psychology, 48, 101457. https://doi.org/10.1016/j.copsyc.2022.101457

88.

van Rens

Haelermans

Groot

van den Brink

H. M.

(2019). Girls’ and boys’ perceptions of the transition from primary to secondary school. Child Indicators Research, 12(4), 1481–1506. https://doi.org/10.1007/s12187-018-9591-y

89.

Verduin

T. L.

Kendall

P. C.

(2008). Peer perceptions and liking of children with anxiety disorders. Journal of Abnormal Child Psychology, 36(4), 459–469. https://doi.org/10.1007/s10802-007-9192-6

90.

Way

(2013). Boys’ friendships during adolescence: Intimacy, desire, and loss. Journal of Research on Adolescence, 23(2), 201–213. https://doi.org/10.1111/jora.12047

91.

Way

Greene

M. L.

(2006). Trajectories of perceived friendship quality during adolescence: The patterns and contextual predictors. Journal of Research on Adolescence, 16(2), 293–320. https://doi.org/10.1111/j.1532-7795.2006.00133.x

92.

Wrzus

Hanel

Wagner

Neyer

F. J.

(2013). Social network changes and life events across the life span: A meta-analysis. Psychological Bulletin, 139(1), 53–80. https://doi.org/10.1037/a0028601

93.

YWCA Metro Vancouver . (2016). Facilitator manual: Welcome to my life/that’s just me/boys 4 real. YWC Metro Vancouver.