Exploring the Intersectional Identities of Youth Mentees and Their Mentors in Shaping the Quality and Duration of Their Relationships

Abstract

Mentoring programs enjoy considerable empirical support as a strategy for supporting the positive development of youth, particularly those with marginalized identities (e.g., minoritized racial identities) and/or from economically disadvantaged backgrounds. There tends to be considerable variability, however, in both the quality of the mentor–mentee relationships established through programs and whether these relationships are sustained for intended periods of time. Research indicates these features of relationships can be significant influences on the outcomes of participating youth, thus making it imperative to better understand the factors that may contribute to mentoring relationship quality and duration. To expand inquiry in this area, the current study draws on an intersectionality framework and the statistical technique of classification and regression tree (CART) analysis to explore the potentially interacting role of multiple sociodemographic aspects of both the mentee and mentor—specifically age, gender, race/ethnicity, and economic status—in shaping the quality and duration of mentoring relationships (n = 806) established through the Big Brothers Big Sisters community-based mentoring program. Findings suggest that mentor age, mentor income, mentee family income, and mentee race have complex associations with indicators of mentoring relationship quality.

1. Introduction

Community-based mentoring programs that utilize volunteer mentors are popular interventions that have demonstrated the potential to promote a variety of positive outcomes for young people (DuBois & Karcher, 2014; DuBois et al., 2011; Raposa et al., 2019b). Several features of the mentoring relationships established within these programs, such as feelings of closeness and relationship duration, have been found to be important in predicting the outcomes of the participating young people (Grossman et al., 2012; Grossman & Rhodes, 2002; Raposa et al., 2019b). Various studies provide evidence suggesting that mentor and mentee demographic factors, such as race/ethnicity, gender, and age, may be associated with the quality and length of mentoring relationships (e.g., Kupersmidt et al., 2017; Raposa et al., 2019b). Despite a growing body of theory and research to suggest the value of considering how multiple aspects of identity (e.g., gender and race/ethnicity; Crenshaw, 1990; Velez & Spencer, 2018) may interact with one another to influence social experiences, few prior studies of mentoring relationship development have addressed this possibility.

As the population in the United States becomes more racially, ethnically, culturally and economically diverse (Mather et al., 2019), it is necessary to investigate how young people’s multiple identities influence how they navigate their environment and how environmental contexts impact their development (see Ruck et al., 2021; Velez & Spencer, 2018). Thus, considering the multiple intersecting identities of young people could provide important insights for mentoring programs. Therefore, in this study, we examine whether and how multiple aspects of the identities of both the mentee and mentor (i.e., age, gender, race, ethnicity, class) may combine to predict relationship quality and duration within a community-based mentoring program. The study is situated in community-based Big Brothers Big Sisters programs in which adult volunteers are paired with youth for one-to-one mentoring and the mentoring interactions and activities are able to occur in varied community settings, as informed by the preferences of the youth and always conditional on parental approval. These programs represent only one of many potential configurations and approaches to mentoring (Keller et al., 2026). Illustratively, programs for younger youth may employ older young persons as near-peer mentors and those for older youth may incorporate activities that focus on supporting transitions to independence and adulthood (for an overview of different types of mentoring programs, see Garringer et al., 2017).

1.1. Relationship Quality and Duration in Mentoring Programs

An important consideration for the quality of a mentoring relationship is how the mentor not only forms a strong personal connection with the mentee but also employs the adult role, with inherently greater experience and capacity, to support the young person’s positive growth and development (Keller & Pryce, 2010; Lyons et al., 2019). For example, researchers have assessed relationship quality in terms of perceived support provided by the mentor to facilitate the young person’s ability cope with various situations in their life (DeWit et al., 2016), perceived feelings of growth experienced by young people in the mentoring relationship (Liang et al., 2013), the mentor providing a young person-centric approach (Werntz et al., 2023), and the young person feeling safe and supported in the relationship (Spencer, 2007). Likewise, feelings of emotional connection (i.e., closeness, perceived support) have been examined as an indicator of mentoring relationship quality and have been highlighted as significant by several researchers (Grossman & Rhodes, 2002; Nakkula & Harris, 2014; Renick Thomson & Zand, 2010; Spencer et al., 2016).

Match length (i.e., how long the relationship lasts) has also been identified as an important predictor of outcomes for young people. Within program models expecting a long-term commitment (e.g., many community-based mentoring programs such as Big Brothers Big Sisters [BBBS] require a minimum commitment of 12 months), research indicates relationships that last longer tend to produce better outcomes and that early-terminating relationships may even cause harm to the young person (Grossman et al., 2012; Grossman & Rhodes, 2002; Herrera et al., 2007; Spencer, 2006). Moreover, mentors who center the relationship around young people’s interests and exhibit flexibility tend to have longer-lasting relationships (Deutsch & Spencer, 2009; Keller & Pryce, 2010; Morrow & Styles, 1995; Renick Thomson & Zand, 2010). These constructs, although conceptually distinct, are also associated with one another, as interpersonal interactions inside a match can influence match length (Spencer et al., 2020). Taken together, evidence points to a variety of ways that mentor and mentee experiences within the relationship are associated with its length and quality, which in turn have been linked to more positive outcomes for young people (Holt et al., 2008; Liao & Sánchez, 2019).

1.2. Mentee and Mentor Identities

Identities held by mentors and young people could shape how they experience and persist with a formal mentoring relationship. In regard to young people, individual characteristics (i.e., gender, age, race) have been linked to the overall quality of mentoring relationships and also premature relationship endings (Liang et al., 2014; McQuillin & Lyons, 2021; Raposa et al., 2016; Schwartz et al., 2011; Spencer et al., 2018). For example, a study of a national database of mentoring programs (92% of these programs featured one-to-one mentoring relationships) revealed that older adolescents in mentoring programs tended to report more challenges in establishing close relationships with mentors than younger children (Kupersmidt et al., 2017). Grossman and Rhodes (2002) similarly found that Big Brothers Big Sisters program mentoring relationships involving older youth were shorter in duration. Parra et al. (2002), in a study of mentoring relationships in this program, did not find significant associations of mentoring relationship experiences or duration with mentee age. In regard to gender, research suggests that mentoring relationships involving females who report lower quality relationships with their guardian/parent are more likely to end prematurely (Grossman & Rhodes, 2002; Spencer et al., 2018). Additionally, previous research has illustrated that social class and family context are important factors in mentoring relationship quality as well (Spencer et al., 2022; Herrera et al., 2013; Weiler et al., 2019). Considering the many demographic identities of youth in mentoring programs is essential because programs that are not mindful of how these various identities influence interactions in the mentoring relationship can run the risk of reinforcing or creating oppressive experiences like those youth may encounter in their everyday lives due to racism, sexism, or classism.

Mentor identity may also impact the formation and maintenance of mentoring relationships (Sánchez et al., 2014). Miranda-Díaz et al. (2020) found that demographic characteristics of the mentor (e.g., race and gender) were related to the mentor’s motivations to volunteer as a mentor and their overall ethnocultural empathy. Parra et al. (2002) found that youth mentees with older mentors reported deriving significantly fewer benefits from the relationships. With regard to potential implications of how the mentor’s identity interacts with the mentee’s identity during the mentoring relationship, Raposa et al. (2019a) found that a shared dislike of activities between the mentor and the young person, as well as shared racial and ethnic identity, predicted longer mentoring relationships. This highlights that the nuanced similarities and differences could impact the quality of the mentoring relationship in a variety of ways (Jones et al., 2023). Despite the acknowledgement of the importance of different aspects of both the mentor’s and young person’s identity in a mentoring relationship, much of the existing research focuses on single dimensions of identity. However, as discussed below, this approach overlooks the potential importance of how multiple aspects of the mentor or mentee’s identity may combine to shape relationship outcomes.

1.3. Considering Intersectional Identities of Mentees and Mentors

Intersectionality, a theoretical framework developed by scholar Kimberlé Crenshaw, provides a lens to conceptualize how an individual’s multiple identities (e.g., gender, race, class, sexuality, ability status) interact to create unique experiences and multiple forms of discrimination and oppression (Crenshaw, 1990). Applying intersectionality in the context of youth mentoring, a young person’s intersecting social identities may influence not only their individual interactions with their mentor but broader systems impacting the young person (e.g., school, legal, and peers groups) may create unique experiences of oppression and privilege that dictate their needs within the mentoring program that then, in turn, can impact the quality of the mentoring relationship (Sánchez et al., 2014; Weiston-Serdan, 2017). For example, a young person who identifies as a non-Hispanic White heterosexual girl from a working-class family might have different structural barriers they have to face, as well as different needs and strengths, than a Latino heterosexual boy from a middle-class family, which impact their respective mentoring relationships due to differences in cultural beliefs, economic access, and other factors. Intersectionality provides a framework that could help mentors understand the unique needs, barriers, and strengths of the mentees they serve, as different intersecting identities may bring about different needs, structural hurdles that can impede the youth’s development and nuanced interactions important to consider in the mentoring process (Gowdy et al., 2024). Further, considering intersectionality in the context of mentoring relationships can help identify the nuanced dynamics in the mentor–mentee pairing (e.g., power dynamics, inequality, discrimination, and structural barriers) that impact the interactions and takeaways in the mentoring relationship (Albright et al., 2017; Jones et al., 2023).

1.4. The Current Study

Despite evidence of mentee and mentor identity predicting the strength of the mentoring relationship, there is a lack of research examining how the multiple identities participants possess potentially influence relationship duration and other indicators of the quality of the mentoring relationship (e.g., coping skills, personal growth, young person-centered approach, and feelings of security). Moreover, scholars have called for more quantitative applications of intersectionality (Bauer et al., 2022). Specifically, scholars have encouraged more research to include multiple facets of identity simultaneously to better capture the complexities of how identities shape interactions. In this study, identity refers to a feature about an individual that they might identify as (e.g., gender) or a social category with which an individual is eligible to be a member (e.g., ethnic/racial identity or socioeconomic status; Chandra, 2006; McCormick-Huhn et al., 2019). The purpose of this study is to explore intersectional aspects of identities of the young person and mentor as potentially distinct and interacting predictors of different aspects of a strong relationship between the young person and their mentor. Grounded in the intersectionality framework, we did not want to pre-determine which aspects of identity may uniquely interact to predict relationship quality. Therefore, this study utilized classification and regression tree (CART) analysis to explore multiple predictors and interactions at the same time without specifying them in advance (King & Resick, 2014). CART analysis is utilized to distinguish what potential intersecting identities (e.g., race, gender, income) are associated with the length of relationships and overall relationship quality among young people engaged in a community-based mentoring program. Specifically, we seek to answer the following research questions:

• What aspects of mentee identity (i.e., gender, race, class) intersect to predict mentee-reported relationship quality and duration (i.e., mentor support)?
• What aspects of mentor identity intersect to predict mentee-reported relationship quality and duration?
• What aspects of mentor and mentee identity interact to predict mentee-reported relationship quality and duration?

2. Materials and Methods

This study utilizes data from a large randomized comparative effectiveness trial involving 10 Big Brothers Big Sisters community-based mentoring agencies across nine U.S. states (DuBois & Keller, 2017). In this study, pairs of mentees and mentors (i.e., matches) were randomly assigned to either receive mentoring as usual or mentoring plus an intervention in which mentors were provided additional training to promote positive youth development through their match activities with the mentee. Ultimately, mentors and mentees were matched, 406 in the control condition and 400 in the intervention. Measures were administered to participants at study baseline (Time 1; prior to random assignment and mentoring relationship initiation) and 15 months after the beginning of the young person’s match (Time 2). Follow-up data were able to be collected from three-quarters of young people in the total sample (75.2%; N = 606). As the purpose of the present study is not to test the intervention itself, but to examine differences in match quality and duration based on intersectional identities, the intervention is not described in detail. For more information on the intervention and design for the larger study (see DuBois & Keller, 2017; Keller & DuBois, 2021).

The average age for mentees in the analytic sample was 14 years (M = 13.85, SD = 1.56), ranging between 11 and 18 years of age. Mentor age ranged between 18 and 66 years old (M = 30.62, SD = 9.48). Mentors were not asked about their household income as children; however, an item was included in the survey that asked whether mentors experienced poverty as a child, and 12% of mentors indicated that they did. Additional descriptive information for mentors and mentees is available in

Table A1. Description information for mentor and mentee identity measures.

	Mentors	Mentees
Gender
Female	63.4%	61.9%
Male	36.6%	38.0%
Race
Black/African American	14.5%	50.5%
White	58.3%	9.7%
Hispanic	2.9%	27.9%
More than One Race	19.4%	9.4%
Other Race	4.1%	2.5%
Income—Mentee *
Less than $10,000		11.3%
$10,000 to $14,999		5.0%
$15,000 to $19,999		3.6%
$20,000 to $24,999		5.8%
$25,000 to $29,999		3.5%
$30,000 to $34,999		3.6%
$35,000 and over		5.2%
Income—Mentor *
Under $20,000	11.4%
$20,000 to $49,999	31.0%
$50,000 to $69,999	22.1%
$70,000 to $99,999	14.6%
$100,000 or more	19.1%

Note. * Income categories differed between mentee and mentor surveys.

.

2.1. Measures

2.1.1. Identity

Aspects of both mentor and mentee identity were measured from demographic variables, specifically gender, race, and income. Table A1 provides descriptive statistics on these variables for the sample.

Gender. Mentors and mentees were asked to identify as either male or female.

Ethnicity/Race. Mentors and mentees were asked to report on their ethnic background (“Hispanic/Latino” or “Not Hispanic/Latino”) as well as their race. Racial categories were collapsed into four categories, including White, African American/Black, more than one race, or Other, to account for small sample sizes within groups.

Income. Mentee’s parents/guardians were asked to report on their household’s total combined family income over the last year. Options ranged from “$0–$10,000” to “$80,001 and more.” Mentors were asked to report on their income over the last year, with options ranging from “under $20,000” to “$100,000 or more.”

2.1.2. Mentor Experiences of Poverty

Mentors were asked to report on whether they experienced a number of adverse childhood experiences while growing up. This included indicating whether they experienced “poverty” while growing up.

2.1.3. Relationship Quality

The young person’s reported relationship quality was assessed through multiple measures during the Time 2 survey. Several measures were either established or adapted for the purpose of this study; this process has been discussed in more detail in prior research (see DuBois & Keller, 2017; Keller & DuBois, 2021).

Table A2. Reliability estimates, descriptive statistics, and correlations between measures of relationship quality.

	α	M	SD	1	2	3	4	5	6
1. Coping	0.91	4.28	1.04
2. Thriving	0.96	3.28	0.83	0.74
3. Growth Focus	0.89	3.20	0.72	0.70	0.80
4. Youth Center	0.89	3.49	0.65	0.67	0.78	0.78
5. Safe Haven	0.88	3.35	0.80	0.75	0.75	0.82	0.77
6. Match Length *	----	11.68	4.51	0.27	0.32	0.30	0.31	0.29

* Match Length is measured in months.

includes reliability estimates, descriptive statistics, and intercorrelations for the measures.

Mentor Support of Coping. Several items from the Youth Strength of Relationship scale (Rhodes et al., 2017) were used to measure the extent to which young people perceived their mentors to provide them with support for coping. This scale included three items (e.g., “When something is bugging me, my Big (i.e., mentor) listens while I talk about it”). All items were responded to on a five-point scale from never true to always true.

Mentor Support for Thriving. A measure was developed to assess the young person’s perceptions of the extent to which their mentor provided support to help them thrive. The measure included 8 items (e.g., “My Big helped me explore what my talents, interests, or hobbies are.”). All items used a four-point scale from not at all to a lot.

Growth-Focused Relationship. A growth-focused scale was developed to assess the extent to which young people perceived their mentor to provide support for their goals and development. This scale includes six items, such as “My Big helps me set and reach goals,” which were rated on a 4-point scale ranging from very false to very true.

Youth-Centered Relationship. A scale assessing young people’s perceptions that their mentor considers their interests was administered (Herrera et al., 2013). This scale included five items, such as “My Big and I do things I really want to do.” Responses were on a 4-point scale from very false to very true.

Safe Haven. Seven items from the Cassidy and Woodhouse Secure Base Scale (Cassidy & Woodhouse, 2003) were utilized to measure whether the mentor served as a secure base for the mentee. This scale included items such as “My Big is someone I can count on when I need help.” Young people reported agreement using a 4-point scale from very false to very true.

Match Length. Match start and end dates were recorded for all participants and were used to calculate the total match length. Time 2 data were collected after matches had been together for 15 months, and, therefore, the maximum match length for this study is 15 months. For any matches that were still active at the time of Time 2 data collection, match length was censored at 15 months even though the match lasted longer (i.e., researchers were unable to obtain specific information on how long the match actually lasted).

2.2. Data Analytic Plan

The CART approach evaluates the relative importance of multiple predictors to develop distinct subgroups of participants with similar experiences on the outcome of interest. CART employs an iterative data segmentation procedure to identify the variables and splitting values that maximize the prediction of the outcome. Within this approach, recursive partitioning is applied to predict an outcome based on a set of predictors. At each step, the algorithm searches over all predictors included in the model, and split points are developed to divide the data into subgroups based on the likelihood for each outcome. This process is repeated until a decision tree is developed with predictor values that are associated with different outcome levels (Krzywinski & Altman, 2017; Loh, 2011). Stopping parameters were set as a minimum of 50 cases within a child node and a maximum tree depth of 5. For this study, dependent variables included five measures of mentee-reported quality of relationship and one measure of match length. A separate CART analysis was conducted for each of the six relationship quality outcomes. Predictors included aspects of both mentor and mentee identity: age, gender, race/ethnicity, and income. Characteristics of the mentor and mentee were included as candidate predictors. The self-reported measure of mentor experience of early poverty was also included, as was the treatment condition (mentoring as usual vs. participation in enhancement intervention). CART analysis was conducted using SPSS (version 27.0; Cronk, 2024).

3. Results

3.1. Primary Analyses

The CART analysis generated no decision tree for two of the relationship quality indicators. Specifically, the results revealed that no aspect of mentor or mentee identity predicted mentor support for coping or a youth-centered relationship as reported by the mentee.

3.2. Mentor Support for Thriving

The regression tree predicting mentor support for thriving resulted in four nodes (see Figure 1). Mentor age was the first splitting variable, with the split point distinguishing the two groups at a mentor age of 22.5 years. On the left side of the tree, the split resulted in a terminal node of 74 young people whose mentors were 22.5 years or younger and reported relatively lower mentor support for thriving compared to all other groups (M = 3.02). The split toward the right side of the tree resulted in a group of 517 young people with mentors over the age of 22.5 years old, with relatively higher mentor support for thriving (M = 3.32). This subset of 517 young people was further segmented by mentor age at a split point of 25.5 years of age. A terminal node of 141 young people with mentors between 22.5 and 25.5 years of age reported the highest support for thriving (M = 3.50) compared to all other groups.

3.3. Growth-Focused Relationship

The analysis yielded 10 nodes defining distinct groups with different mean scores on the growth-focused relationships scale (see Figure 2). The lowest scoring group was distinguished by just one factor; it included young people with mentors under 22.5 years old (M = 2.96). Mentees who identified as black and had mentors between the ages of 22.5 and 25.5 years old reported the highest average on the growth-focused relationships scale (Node 5, M = 3.50). For young people who identified as Black, being assigned a mentor over the age of 25.5 years old was related to relatively higher averages when the mentee was 13.5 years old or younger (M = 3.37) compared to those who were over 13.5 (M = 3.16). For young people who did not identify as Black, mentees assigned to a mentor earning over $69,999 annually scored relatively higher on average (M = 3.28) than those with mentors making less than $69,999 annually (M = 3.05).

3.4. Safe Haven

Nentor income emerged as the only factor in the regression tree predicting safe haven (see Figure 3). Young people assigned to a mentor who earned under $49,999 (M = 3.24) had a relatively lower mean score on average than young people assigned to a mentor who earned $49,999 or more (M = 3.43).

3.5. Match Length

Mentor income, mentor age, type of intervention (mentoring plus enhancements or mentoring as usual), and mentee income emerged as factors in the regression tree predicting match length (see Figure 4). Mentees assigned mentors earning over $99,999 annually reported the longest average match length of any group (M = 11.70 months). The second longest average match length was reported by mentees assigned mentors who earned less than $99,999 annually, were between the ages of 23.5 and 38.5 years old, were assigned mentoring plus enhancements, and had a family income of equal to or greater than $20,000. Mentees assigned mentors who earned less than $99,999 and who were over the age of 38.5 years old, on average, reported the shortest match length (M = 9.06 months). Mentees assigned a mentor who was 23.5 years old or younger reported relatively shorter match lengths. For example, mentees assigned a mentor 23.5 years old or younger reported, on average, a match length of 9.08 months, whereas mentees assigned a mentor over 23.5 years old reported an average match length of 10.65 months, even though both groups were assigned mentors who earned less than $99,999 annually and were assigned to mentoring plus enhancements (Nodes 7 and 8).

4. Discussion

The purpose of this study was to examine how different aspects of mentee and mentor identities intersect to predict relationship quality in their mentoring relationship. Interestingly, some indicators of relationship quality, namely mentor support for coping and growth-focused activities, were not associated with any of the mentee or mentor factors included in the analysis. However, the results for other indicators of relationship quality suggest associations with certain combinations of both mentee and mentor factors. Specifically for the mentor, the results indicate that age and income of the mentor were predictors of the mentor’s support for thriving, growth-focused relationships, young people feeling secure, and the length of the relationship. For the mentee, race, age, and household income were predictors of growth-focused relationships and the length of the relationship.

The results, revealing that young people with mentors older than 25 reported higher mentor support for thriving and growth-focused relationships, may speak to the time and resources volunteers have available to invest in the relationship when in higher income brackets (see Niebuur et al., 2018). Findings for a growth-focused relationship, indicating that Black young people with mentors between the ages of 22.5 and 25.5 reported higher relationship quality compared to young people from other racial/ethnic groups, may illustrate a difference in mentoring approaches between younger mentors and older mentors when mentoring certain demographics of young people. Previous research on the impact of mentor age on program outcomes has been mixed, as some research indicates that high school and college-age mentors tend to be less effective than older mentors (Herrera et al., 2008; Grossman et al., 2012), and other research has illustrated that the age of the mentor is not a significant factor in program outcomes (Raposa et al., 2019b). Perhaps these mentors who are exiting emerging adulthood are able to provide a unique type of goal-oriented support that older mentors are not. Researchers have emphasized the importance of utilizing an empowerment approach with Black young people (Gaylord-Harden et al., 2018; Travis & Leech, 2014) and addressing identity development in the lives of Black young people, specifically racial identity development (Cross et al., 2017; Gaylord-Harden et al., 2007; Neblett et al., 2012; Sánchez et al., 2018; Umaña-Taylor et al., 2014). More research is needed to explore potential explanations of which racial/ethnic groups may be more receptive to specific types of mentoring that support different constructs of relationship quality they find important in their lives.

There were several instances where higher mentor income was associated with a higher quality of relationship (greater than $49,999 for safe haven, $69,999 for growth focus, $99,999 for match length). The previous literature has indicated high income volunteer mentors tend to stay in relationships longer than volunteer mentors who make less money (Grossman & Rhodes, 2002). This may be attributed to the job flexibility, education level, and/or social support mentors with a higher income have access to during the mentoring relationship. It is also important to highlight that the variables focused on the young person’s coping and centering the young person’s interests did not result in a tree being created in the analysis. This might speak to the emphasis placed on specific outcomes of the mentoring programs that participated in this study.

Results indicate that young people in families that earned more than $20,000 annually had matches that lasted on average over two months longer than matches of young people living in families that reported earning less than this amount. Given that the expected match length in Big Brothers Big Sisters is 12 months, there are a few possible program and relationship reasons for this difference. First, it is possible that matches closing earlier is due to classism experienced by low-income families in the mentoring relationship. A recent qualitative study conducted by Spencer and colleagues (Spencer et al., 2022) found that mentors often held judgmental views of the low-income families they volunteered with, which resulted in them blaming parents and not providing grace to different life circumstances related to the social class of the family. Although we do not know who requested to end the match or why the match closed, it is possible that negative ideas about families led to the match ending before fulfilling the 12-month commitment. Moreover, these findings could also highlight the structural barriers low-income families have to deal with to participate in youth development programs. For example, a lack of financial resources could leave low-income individuals with less time to participate in the youth mentoring relationships than their peers with higher incomes. Furthermore, there may be practical barriers to regular contact for mentoring activities associated with financial distress, such as housing instability, disconnected utilities, inconsistent child care, lack of transportation, etc. This is important to consider as youth development literature has underscored that low-income families participate in youth development programs for less time than their peers despite the benefits of program involvement (Heath et al., 2022; Theokas & Bloch, 2006; Wimer et al., 2006). Lastly, the findings highlighted the fact that mentor factors were more prominent in the different domains of relationship quality. This may suggest that mentor circumstances could have a greater bearing on the relationship than the mentee and would be consistent with the idea that mentors have greater responsibility in setting the tone of the relationship (Goldner & Ben-Eliyahu, 2021).

Overall, these findings add to the previous youth mentoring literature that has underlined factors that impact the strength of the relationship by utilizing an intersectional approach to examine how these factors, in combination with one another, relate to youth mentoring relationship quality. Specifically, the findings further reinforce how the combination of income, age, and race is associated with various aspects of the youth mentoring relationship (e.g., strength of relationship and match length). Previous research also illustrates that volunteer mentors with higher incomes and ages were associated with longer relationships (Grossman & Rhodes, 2002; Herrera et al., 2011; Rhodes et al., 2017). Specifically, Grossman and Rhodes (2002) found relationships that lasted longer than those of mentors who reported lower incomes, and mentors who were between the ages of 26 and 30 were less likely to terminate their matches when compared to mentors 25 and younger, as well as mentors who were between 26 and 30 and also married (Grossman & Rhodes, 2002). Our findings complement and add to this research by underscoring that the combination of different factors (e.g., mentor age and the income of both the mentor and the young person’s family, along with the racial identity of the mentee) is a pivotal factor that may play specific roles in different aspects of the youth mentoring relationship.

4.1. Strengths and Limitations

This study responds to calls from other youth development and mentoring scholars to integrate critical frameworks and intersectionality when assessing youth mentoring programs on the lives of young people (Albright et al., 2017; Sánchez et al., 2018, 2021; Weiston-Serdan, 2017). Further, this study utilizes an exploratory quantitative method to examine the intersections of identity that are related to relationship quality, an essential component of an impactful youth mentoring experience. Finally, this study adds to the youth mentoring and broader youth development literature by highlighting how multiple intersecting identities (i.e., age and income) are associated with aspects of mentor support and relationship length.

While this study has numerous strengths, it is not without limitations. First, the model fit was not as high as we would like to see, indicating that there is some room for improvement and a need for further studies. Further, the exploratory nature of the findings underscores the importance of replication and validation in other research. More research is needed to capture the complexity of this topic, as demonstrated by the vastly different results in the different regression trees despite generally very high correlations among all the indicators other than duration. Second, we were not able to look at all racial/ethnic groups separately due to sample size limitations. Additionally, not all identities that are important in the young person’s life may be represented; for example, individuals may have had their gender identity missing because their specific identity was not represented (e.g., nonbinary). Moreover, aspects of one’s identity are more nuanced than demonstrated in this paper—age, geographical location, sexual orientation, religion, etc., are all salient within contexts, and these are not represented within this paper. Further, two people of the same racial identity may have completely different perceptions of how their racial identity impacts their lives. Methodologically, it should be taken into account when considering findings that they are based on only approximately three-quarters of the sample for whom follow-up data were able to be collected. Consequently, although a concerted effort was made to obtain follow-up data for all youth, regardless of the status of their mentoring relationship, it may be the case that those for whom these data were able to be collected tended to be in stronger relationships and, thus, the results reported here are biased to some degree in that respect. Additionally, access to programs such as BBBS is nested within communities and thus it is important to keep in mind that access is also impacted by what is available, how it is marketed, and who has the capacity to participate in these types of programs. Finally, these findings may not be widely generalizable to other types of mentoring programs, countries, and cultures where social constructs like race, ethnicity, and social class may have other meanings.

4.2. Future Directions

Additional research that utilizes an intersectionality framework to view different elements of youth mentoring relationships is important in order to understand how these relationships function differently and have different impacts on people based on the many identities they navigate society with on a daily basis. Specifically, more attention is needed to the within-group differences in youth who participate in youth mentoring programs such as BBBS (Williams & Deutsch, 2016). Further, investigating these within-group differences could lead to a more in-depth matching process between the mentee and mentor, which, in turn, could influence the longevity and relationship quality of the match. Moreover, additional research should explore the potential barriers that prevent families from lower-income backgrounds from maintaining mentor relationships and what can be done to support the longevity of mentoring relationships for the demographic of young people engaged in mentoring programs. Lastly, findings from this study illustrate the importance of mentoring programs to consider innovative ways to support mentors in specific income brackets so they are able to provide the best mentoring experience possible to their mentees. Specifically, mentoring programs can provide resources to both mentees and mentors to mitigate barriers that may interfere with them getting the most out of the mentoring relationship.

5. Conclusions

Overall, this study is another step in integrating critical frameworks such as intersectionality into research centering youth development programs, specifically youth mentoring programs. The research questions focused on what aspects of the mentee’s and mentor’s identity (i.e., gender, race, class) intersected to predict mentee-reported relationship quality and duration. Findings illustrated that the mentor’s age and income were predictors of their support for thriving, growth-focused relationships, young people feeling secure, and the length of the mentoring relationship. For the mentee, race, age, and household income were predictors of growth-focused relationships and the length of the relationship. This project provides preliminary evidence on the importance of considering multiple aspects of the mentor’s identity simultaneously, specifically race, age and income, when assessing aspects of relationship quality in a youth mentoring relationship. More research utilizing intersectionality and other critical frameworks is needed to examine the nuanced aspects of a youth mentoring relationship and its unique impact on diverse groups of young people.