Validation of a youth social capital scale Informed by positive youth development

Abstract

Informed by a positive youth development perspective, the current study develops and validates a brief social capital scale for adolescents and emerging adults from racially and ethnically diverse backgrounds. We first detail the development of the scale and examine its factor structure using exploratory factor analysis. Second, using a new and independent sample, we use confirmatory factor analysis to replicate and confirm the factor structure of the scale. Finally, we conduct measurement invariance by gender, age, and race-ethnicity to determine if the scale held similar meaning for participants from different groups. Findings support a multidimensional scale of youth social capital that captures three important elements: network strength, network diversity, and network mobilization. The creation of this scale expands possibilities for researchers and practitioners by providing a comprehensive and actionable measure for empirically assessing youth social capital, which can be leveraged to support positive youth development.

Social capital is a transformative force in the lives of youth, exerting a profound impact on their educational and career trajectories (Dika & Singh, 2002; Mishra, 2020). Strong social networks can equip and empower youth by serving as a conduit to invaluable information about career opportunities including job referrals, job openings, and introductions and access to influential hiring managers (Brown et al., 2016; Hernandez-Gantes et al., 2018). Moreover, social capital has been linked with aiding young adults in securing employment, especially during times of elevated rates of unemployment (Kramarz & Skans, 2014; Hällsten et al., 2017). Social capital can also play a pivotal role in facilitating the pursuit of educational aspirations (Mishra, 2020; Stephan, 2013). Supportive relationships with family members, peers, educators, and both formal and informal mentors, for example, can extend support to youth by offering guidance, assisting with college preparation and enrollment, and helping to navigate the complexities of the educational system (Stephan, 2013).

Every young person possesses social capital, yet social capital theory posits that the diversity and strength of these connections, and young people’s ability to leverage them while pursuing life goals varies, often as a result of long-standing systemic forces such as racism and discrimination (Dill & Ozer, 2019). This structural variability in who has access to opportunity and power contributes to individual-level variation often observed in education and employment outcomes, variation that has little to do with individual-level talents or motivation (Lin, 1999; Stanton-Salazar, 2011). Due to these structural barriers, social capital may be a particularly valuable asset for young people from minoritized and/or historically-marginalized communities. However, some scholars have also noted the potential negative consequences of social capital, sometimes referred to as the ‘dark side’ of social capital (Baycan & Öner, 2023), including the inequality of resources between social groups (Lin, 1999) and social exclusion (Portes, 2009), all of which may contribute to the preservation of inequitable systems and institutions. At the same time, the potential for positive outcomes associated with youth social capital has led many youth-serving organizations to look for strategies to support young people’s social capital development, particularly among youth who come from historically-marginalized backgrounds.

Despite the potential of social capital to increase postsecondary educational and employment opportunities for youth from minoritized and/or historically-marginalized backgrounds, there remains a notable lack of conceptual clarity around its definition (Bankston & Zhou, 2002; Jackson, 2019; Scales et al., 2020). Jackson (2019) characterizes the literature in this way: The literature on social capital is “sprawling and inconsistent in its multitude of alternative definitions and uses of the term” (p. 316). Some scholars consider social capital an individual-level resource for personal benefits (e.g. Rothon et al., 2012), and others attribute social capital to the group, neighborhood, or community context with benefits being experienced collectively (e.g. De Clercq et al., 2012). Similarly, some have characterized it as resources to enable attainment of majority-culture outcomes (e.g. Coleman, 1988), while others have conceptualized social capital as the resources needed to effect change not just in opportunity but in the goals and outcomes that are considered socially valuable (e.g. Dill & Ozer, 2019).

Part of the challenge of conceptualizing a shared definition of social capital is due to its multifaceted origins, spanning a multitude of disciplines (e.g. sociology, economics). Most prominent scholars of social capital identify access to resources through social relationships, as we also do, as the essence of social capital (Bourdieu, 1986; Coleman, 1988; Lin, 1999), but some scholars have examined social capital at the individual level, whereas others have described it as a feature of neighborhoods or communities (i.e. community level). For example, taking an individual-level perspective, Kim and Schneider (2005) described social capital as “realized through the social ties that connect individuals to resources” (p. 1197). In contrast, De Clercq et al. (2012), suggested that social capital is not an individual asset, but is rather a “collective characteristic of places arising from people’s shared experiences. It can be defined as the quantity and quality of social relationships such as formal and informal social connections as well as norms of reciprocity and trust that exist in a place or community” (p. 203). Notably, other scholars have asserted that social capital is both a collective and individual asset in which resources are embedded within a social structure, but are also accessed and mobilized by individuals through “purposive actions” (Lin, 1999; p. 35).

These varied perspectives have led, understandably, to dozens of instruments having been developed for measuring social capital, most intended for adults (reviewed in Scales et al., 2020), and measuring different aspects of social capital aligned with those differing concepts. In an extensive review of measures intended for use with youth, Ahlborg et al. (2022) identified only 20 instruments worldwide, only five of which were developed with U.S. samples. Among those measures designed for U.S. youth, three of the five were focused on health (e.g. sexual and reproductive health, substance use), leaving only two instruments intended to be more general assessments of the multiple dimensions of social capital. Only one of those two studies presented confirmatory factor analysis evidence of their measure’s validity, but they also cautioned their sample size was not large (n = 140, see Ryan & Junker, 2019).

Thus, in part due to the absence of a shared and precise definition, the positive youth development field and spaces that serve young people (e.g. schools, youth-serving organizations) have few options when looking for a brief, yet validated scale to capture youth reports of the social capital they experience, especially in racially-ethnically and socioeconomically diverse samples of U.S. youth. The absence of a brief and validated scale to assess youth social capital that is also rooted in practitioner perspectives and therefore most actionable to practitioners limits the ability of researchers and program staff to gain and act on a nuanced understanding of the diverse spectrum of social capital among youth. Therefore, the current study attempts to address this gap by developing and validating a scale to measure youth social capital among diverse U.S. youth, with significant input from youth and practitioners.

Conceptualizing youth social capital using a positive youth development perspective

The theoretical principles of Positive Youth Development (PYD: Benson et al., 2006; Lerner, 2006) provide a different framework for conceptualizing social capital than has been typical in the social capital literature. For example, among the relatively small number of studies that have examined youth social capital, young people are often viewed as passive recipients of social capital rather than, as PYD articulates, active agents capable of activating and mobilizing resources in pursuit of their goals (Leonard, 2008; Varga & Zaff, 2018). PYD offers a strengths-based perspective that views all youth as having the potential to thrive and reach their full potential. In this view, youth are not seen as problems to solve but as resources to develop (Pittman et al., 2001). In particular, PYD takes a whole-ecology perspective in which each setting youth inhabit has a contribution to make to their development, and in which community peers and adults, not only professionals such as teachers and youth program leaders, have a key role to play in youth positive development (Benson et al., 2006). Notably, PYD frameworks have for decades elevated this importance of what the social capital literature has called “weak ties” that often provide the external and internal “developmental assets” youth need for positive development, such as youth being given opportunities to contribute, or developing a sense of self-efficacy (Benson et al., 2011). In social capital parlance, such relationship-based developmental assets reflect the bridging and linking social capital (connections across groups that have more diverse memberships and/or power and status differences; Szreter & Woolcock, 2004) youth need to move beyond the opportunities provided through their more common bonding social capital strong ties with people like themselves (e.g. family members, friends; Flanagan et al., 2014). Thus, PYD offers a well-aligned perspective for conceptualizing the full complexity of youth social capital development. This potential is realized when youth are embedded within relationship- and nutrient-rich environments and contexts (Benson et al., 2011)—also known as a supportive youth system (Lerner et al., 2016). Within this system, relationships are the bedrock or the active ingredient that support adolescent development (Li & Julian, 2012).

Despite relationships being recognized as a critical component across most youth development theories and social capital theories, few researchers have articulated the most effective ways to mobilize these relationships within youth systems, let alone attempt to measure them so that they can be better understood and strengthened. However, there have been examples of more expanded definitions and frameworks of youth social capital that align with a positive youth development perspective. The Webs of Support Framework (Varga & Zaff, 2018), for example, integrates research on relationships, social support, networks, and social capital to describe a constellation of relationships in which resources and opportunities are leveraged and mobilized to promote positive developmental outcomes across contexts (e.g. schools, families). Similarly, the Community Cultural Wealth Framework highlights assets and resources that young people of color can acquire within their communities while navigating social systems (Yosso, 2005). In another example, Dill and Ozer (2019) distinguished “social support” as a form of capital that can be useful under certain circumstances such as helping youth get by or deal with daily stressors, and “social leverage,” which can help youth get ahead and access connections and valuable resources that lead to social and economic mobility. This form of social capital can be used to change systems to be more equitable by “fostering political consciousness and positive racial identity to analyze, respond to, and address issues such as racism, police violence, and neighborhood violence in their schools and communities” (Dill & Ozer, 2019, p. 3). Other scholars have also highlighted forms of critical social capital, which emphasize youth agency, critical consciousness, and empowerment among Black youth (Ginwright, 2007). These examples of social capital all view youth as actors in facilitating change, while also acknowledging structural constraints. Informed by these conceptualizations, we have synthesized numerous definitions and perspectives on social capital to define youth social capital in a way we believe accurately summarizes the literature: The resources that arise from a web of relationships that young people can mobilize and activate as they pursue their life goals (Scales et al., 2020).

Measuring youth social capital

Given the variability in how youth social capital has been defined, studies have also measured it in myriad ways including the use of proxy indicators, name and resource generators, and the quality of dyadic relationships. First, many studies do not measure the social capital of youth directly but instead try to capture indicators of capital within a particular context such as the family, school, or community (Ferguson, 2006). For example, three large scale reviews of studies examining the impact of social capital on children and adolescent wellbeing have synthesized the ways that social capital is operationalized across these contexts (Ahlborg et al., 2022; Ferguson, 2006; McPherson et al., 2013). Each study has found that the indicators used to assess social capital within family settings include factors such as family structure, the quality of parent-child relationships, caregiver involvement in school, and parental monitoring, while indicators commonly used to assess social capital within community settings include social support, civic engagement, trust and safety, quality of schools, and religiosity (Ahlborg et al., 2022; Ferguson, 2006; McPherson et al., 2013). While these indicators have demonstrated positive associations with important youth development outcomes (Dika & Singh, 2002; Ryan & Ream, 2016), these proxy measures do not fully reflect the multidimensional nature of social capital across an array of contexts and settings, nor the role of youth as active agents in utilizing these resources.

Second, at times social capital theory has also been treated as synonymous with social network theory (Ryan & Junker, 2019). This has resulted in some scholars measuring youth social capital using structural properties of networks, such as size or density (Van der Gaag & Snijders, 2004). Name and resource generators are data collection tools that are often used to map the structure of a young person’s network. Youth may be asked to name the number of people with whom they share a particular type of social relationship. Once a list of names is produced, participants are presented with a series of follow-up questions (e.g. type of relationship, degree of closeness, occupation of person) that gathers information about each of the people identified within the network (Lin, 2008). While these tools can be useful for capturing the structure of a young person’s network, they might not capture other important elements of an individual’s social capital such as access to resources and young people’s ability to mobilize and activate their network of support. These approaches are also often limited to examining a network of individuals within a specific setting (e.g. peers within a classroom), and thus may not capture relationships and resources that contribute to a young person’s social capital across multiple settings. Finally, these tools are often time intensive to administer and may require personal identification of individuals within a young person’s network, which limits anonymity and makes these tools particularly difficult for schools and other youth-serving organizations to use.

Finally, other studies have empirically examined youth social capital by assessing the resources provided by and/or the quality of youth-adult relationships within a single context (e.g. school, mentoring program). For example, relationships that provide bonding forms of capital have been shown to be instrumental in providing emotional and other forms of support (Ciarrochi et al., 2017). Additionally, studies have examined what has been referred to as “institutional agents’' or adults who hold a position of power and/or status and are therefore able to provide access to resources and opportunities that may not otherwise be available (Stanton-Salazar, 2011; p. 1067). These institutional agents may provide bridging and/or linking forms of social capital and take the form of high-quality relationships with teachers, mentors, and others—all of which have been shown to contribute to youth social capital and positive youth development outcomes (Boat et al., 2022; Hurd et al., 2014; Raposa et al., 2018; Scales et al., 2022).

Notably, one study examined the differential impact of both strong-tie or close-knit natural mentoring relationships that youth have frequent interactions with such as family members and friends (bonding social capital) and weak-tie natural mentors, individuals that youth have more socially distant relationships with, such as teachers, coaches, and employers (bridging and linking social capital) and found both were associated with positive youth outcomes, but contributed to different types of outcomes (Hagler & Rhodes, 2018). Strong-tie mentoring relationships were associated with more close friendships, whereas weak-tie mentoring relationships were associated with higher educational attainment and greater likelihood of volunteering (Hagler & Rhodes, 2018). This suggests that both weak-ties and strong-ties can support and contribute to youth social capital, but may do so in different ways.

The current study

Given the potential for social capital to enhance the education and career opportunities for young people from minoritized and/or marginalized communities, it is imperative for practitioners and researchers to find useful ways to assess if they are making strides in strengthening and supporting youth social capital. Research suggests that there is no agreed-upon best method or any single instrument that captures all aspects of social capital, and that can be easily implemented within youth settings. Indeed, in their comprehensive review Ahlborg et al. (2022) concluded that assessing the validity of social capital measures is difficult “due to there not being any golden standard for the measurement of social capital” (p. 15602). As a result, there are only a relatively small number of social capital scales targeting youth, only a handful are developed for samples of U.S. youth, and that focus on general measures of social capital (see Ahlborg et al., 2022).

To address this need, we aimed to capture the multidimensional nature of social capital among youth through a brief and practitioner-friendly scale. We focus on operationalizing and measuring youths’ individual social capital, emphasizing the significance of personal connections in shaping individual life trajectories. Informed by a PYD perspective and past social capital theory and empirical research, we attempted to capture several elements of social capital including the network structure (access to both weak and strong ties), access to valuable and varied resources via these relationships, and youth’s activation and mobilization of their network. We specifically focused on relationships and resources that youth are using as they work toward education and career goals, as social capital has consistently been linked to these important developmental outcomes (Dika & Singh, 2002; Mishra, 2020). Moreover, we developed our instrument with input from practitioners and racially and ethnically diverse youth who were participating in education and workforce development programs with an explicit focus on supporting youth social capital development to increase access to valuable resources, connections, and opportunities for youth from historically-marginalized communities.

The current study is composed of two sequential studies and details the development and validation of a social capital scale for adolescents (ages 14-17) and emerging adults (ages 18-25) called the Youth Network of Support Scale. While items used in the EFA were theoretically grounded, we were less certain about the factorial validity of any potential sub-constructs of a multidimensional social capital measure. Therefore, in Study 1, we describe the development of the scale and first explore its factor structure using exploratory factor analysis (EFA; Matsunaga, 2010). In Study 2, we use confirmatory factor analyses (CFA) to attempt to replicate the factor structure identified via the EFA analysis in a new and independent sample of participants. All research materials and procedures were reviewed and approved by an independent Institutional Review Board (FWA00021831).

Study 1: Methods

Participants

Participants from the first study included 888 young people ages 14-25 who were recruited from six youth and young adult-serving education and workforce development programs. All programs were located in urban regions of the United States. Two programs were national postsecondary education support programs that provided first-generation college students and students from historically-marginalized communities with near-peer coaches who provided ongoing support and connections to on-campus resources. Both of these programs primarily operated in major urban cities on college campuses within New York, Chicago, Newark, and the Bay Area. The other four programs were workforce development programs that operated in New York and/or the Bay Area. Two programs primarily served first-generation college graduates as they navigated the job search process; one served high school students as they identified their sense of purpose and set future career goals; and one served young adults as they built skills to pursue career advancement opportunities. All four of these programs served youth and young adults from historically-marginalized backgrounds and provided support through peer and near-peer mentoring.

Programs were selected to participate in Study 1 to support the development and validation of practitioner-friendly measures of social capital and related constructs. The six partner programs were recruited based on the following criteria (1) a shared mission to enhance education and/or career outcomes for adolescents and emerging adults by strengthening program participant’s social capital, (2) predominantly served young people from minoritized and/or historically-marginalized communities, and (3) designed to intentionally connect program participants with others including peers, near-peers (i.e. individuals who serve as mentors/coaches and provide ongoing education and/or career goal support), program staff, and other community members.

More than half of the 888 participants identified as female (71.5%), 27.5% identified as male, 0.7% identified as non-binary, and 0.3% preferred to self-describe (e.g. gender fluid). Less than 1% of the sample (0.3%) identified as transgender. Age ranged from 14–25 years (M = 19.82; SD = 2.08). Roughly a third (34.6%) of the sample identified as Hispanic/Latina/o/x, 29.6% identified as Black/African American, 18.8% identified as Asian/Pacific Islander, 10.4% identified as Multiracial, 5.3% identified as White, 0.2% identified as Native American or Alaskan Native, and 1.2% identified as another race.

Procedures

Item development and content and face validity

We used a mixed methods process in designing the Youth Network of Support Scale, which closely aligns with recommendations outlined by Boateng et al., 2018. We first defined youth social capital and generated initial items using both deductive (i.e. based on prior literature) and inductive approaches (i.e. based on qualitative findings). A literature review helped us (1) identify gaps in the scientific literature on the social capital development of youth and young adults, particularly among young people of color and from low-income backgrounds, (2) establish a theoretical foundation for the conceptualization of social capital and how youth and young adult-serving organizations support social capital development, and (3) identify how social capital has been previously measured and assess whether there were existing social capital instruments from which potential items may be adapted or borrowed (Scales et al., 2020). We also conducted 18 focus groups with the six partner programs (involving 33 current program participants, 17 alumni, and 24 staff members) to learn how they conceptualized “social capital,” and to better understand how young people experience relationships and social capital within their respective programs. The resulting qualitative data were analyzed and key themes were identified and used to directly inform item generation (see Boat et al., 2020 for a full description of research methodology and findings). Drawing on the literature review and the themes that surfaced during the focus groups, initial survey items were created, without limiting the number of items. Scales consisted mostly of original items, with a handful of adapted existing measures (e.g. Search Institute’s Developmental Relationships 360 measure).

Expert reviews

Face and content validity of measures were assessed through expert reviews with staff at all six partner organizations (n = 9) and with researchers (n = 4) with content expertise in social capital. Through an online survey, reviewers were asked to indicate whether items reflected the overarching construct and were asked several open-ended questions regarding the clarity of items and response options, whether important dimensions of social capital were left out, and whether there is anything they would change. Staff from partner organizations were also asked about whether they felt the items would provide useful information that could inform practices or program improvements.

Cognitive interviews

Survey items were pretested through semi-structured cognitive interviews with program participants (n = 10) from each partner organization until saturation (i.e. youth participants noted the same feedback) was reached. Items were administered and participants were asked to verbalize their mental process as they read the item and responded to it (Boateng et al., 2018). This was followed up with probing questions to understand what participants think the question means (e.g. “In your own words, what do you think this question is asking?”), how they selected their answer (e.g. “How did you choose your answer?”), and comprehension (e.g. “what does the word X mean to you?”). Based on learnings from expert reviews and cognitive interviews, items were then revised and reduced to increase conceptual clarity, limit overlap, eliminate subpar items, and minimize survey fatigue.

Preliminary data collection

The final items for the Youth Network of Support Scale are provided in Table 2. Youth were provided with the following instructions: “The next set of questions asks about your network. By network, we mean the people in your life both within and outside of [Program/Organization Name] who can help you achieve your education or career goals.” Youth were then instructed to respond to the following prompt, “how much do you disagree or agree with each statement?” All items were rated on a 5-point scale ranging from Strongly Disagree (1) to Strongly Agree (5).

Table 1. Exploratory factor analysis with one to four factors.

Number of Factors	Eigenvalue	χ^2 (df)	RMSEA	CFI	SRMR
1	5.984	649.42 (77)***	.101	.807	.079
2	1.574	277.50 (64)***	.068	.928	.042
3	1.110	104.43 (52)***	.037	.982	.021
4	0.736	65.17 (41)**	.028	.992	.016

Note. *** = p < .001; ** = p < .01; * = p < .05.

Table 2. Youth network of support scale: exploratory factor analysis.

Factor Name and Items	Factor Loadings
Factor 1. Network Diversity	1	2	3
I have people in my network…
with many different careers or career interests	.547*	.205*	.002
from many different cultures or racial/ethnic backgrounds	.752*	−0.006	.019
from many different economic backgrounds	.621*	.007	.109
Factor 2. Network Strength
I have people in my network…
that I can trust to help me pursue my education or career goals	.004	.878*	−0.133*
that introduce me to others who can help me reach my education or career goals	.142*	.685*	.015
who I am close to that help me pursue my education or career goals	.018	.799*	.009
who I am less close to but who are influential in helping me reach my education or career goals	.095	.541*	.053
who help me when they say they are going to help me	−0.028	.754*	−0.016
from when I was younger that I can ask for support, when needed	−0.135	.526*	.136*
with different skills that will be useful to me as I pursue my goals	.321*	.419*	−0.027
Factor 3. Network Mobilization
When working toward my education or career goals, I ask for help when I need it	.000	−0.025	.627*
I go out of my way to meet new people in order to reach my education or career goals	−0.144	.007	.919*
I form strong relationships with people who are useful for helping me reach my education or career goals	.044	−0.018	.774*
I have already sought out people who can help me pursue my education or career goals	.095	.028	.597*

Notes. *** = p < .001; ** = p < .01; * = p < .05.

The survey was administered by all six partner organizations. Partners invited all current program participants to take the online survey over a 2-week period between January and March 2021. Program participants completed the survey on computers or tablets using a web-based survey that was hosted via a secure data collection platform. Program staff helped facilitate program participants’ access to the online survey and were available to answer clarifying questions using standardized administration procedures. The survey took participants roughly 10–15 min to complete. All participants had the opportunity to enter a raffle for one of several $50 e-gift cards as a thank you for their participation.

Analytical plan

We first conducted Kasier-Meyer-Olkin (KMO) and Barlett’s test of sphericity in order to confirm significant correlations among proposed items and to examine the proportion of common variance between items. Exploratory Factor Analysis (EFA) factors were extracted using Maximum Likelihood with Robust Standard Errors (MLR) estimation. The size of the interfactor correlations and magnitudes of cross-loadings were considered when determining the most appropriate rotation method (Sass & Schmitt, 2010). While a simple factor solution, where each item loads on to only one factor, is the most parsimonious, most measures and instruments have a complex factor structure (i.e. items may load on to more than one factor). Because we expected factors to be correlated, we conducted an EFA using an oblique rotation method known as Geomin rotation (i.e. the default rotation method) in Mplus version 8.7 (Muthén & Muthén, 1998–2017). Model fit was assessed using Comparative Fit Index (CFI), Standardized Root Mean Square Residual (SRMR), and Root Mean Square Error of Approximation (RMSEA). Common guidelines were used to assess goodness-of-fit: CFI > .95, SRMR < .06, and RMSEA < .08 (McDonald & Ho, 2002).

An EFA factor solution was obtained using the following criteria: Kaiser’s criterion (retaining factors with eigenvalues greater than one), the interpretability of obtained factor solutions, and model fit indices. The resulting EFA factor solution was then used to inform the retention and removal of any items. Items that exceeded a loading of .40 and above and without significant cross-loadings onto other factors were retained. Missing data across items ranged from 17.8% to 33.1%. Little’s MCAR test suggests that the data are not missing completely at random (χ² = 255.50, df = 212, p < .05). Full information maximum likelihood estimation (which uses all available information) was used to handle missing data.

Study 1: Results

Bartlett’s test of sphericity was significant, χ² = 5,595.83, df = 105, p < .001, and KMO was .94, and all items were significantly correlated at or above .30 with at least one other proposed scale item. These findings indicate that data were satisfactory for factor analysis. In comparing one- through four-factor EFA models, a final scale of three factors and 14 items emerged. Overall, the three-factor model was determined to be the final EFA model because it yielded an interpretable factor structure, met Kaiser’s criterion (eigenvalue greater than one), and was one of the best fitting models (see Table 1).

The solution was composed of three conceptually meaningful factors reflective of underlying constructs. The first factor consisted of three items that measure the diversity of a young person’s network. The items asked if youth had relationships within their network that had different racial-ethnic and economic backgrounds, as well as if they had relationships with individuals with different careers and career interests. This factor is aligned with previous research that has found network diversity and access to bridging and linking forms of capital to be an important component of social capital (Engbers et al., 2017; Jackson, 2019). The second factor consists of seven items that assess the degree to which a young person has a strong network of relationships that can provide access to valuable resources (i.e. network strength). Indicators include having individuals who youth can go to for help, trust, and are influential in providing valuable resources and connections. The items capture both weak-tie and strong-tie relationships that can provide bonding, bridging, and linking forms of capital (Granovetter, 1973; Lin, 2008). The third factor consists of four items that measure the degree to which a young person actively builds strong relationships and uses the relationships and the resources they have in pursuit of their education and career goals (i.e. network mobilization). This factor is consistent with PYD research that demonstrates the important role of youth agency (Varga & Zaff, 2018) in the activation and mobilization of resources. Table 2 shows the loadings of these items onto the three factors in the EFA.

Study 2: Methods

Participants

Participants from the second study included 702 young people ages 14-25 who were recruited from an additional six youth and young adult-serving education and workforce development programs. Four programs operated nationally in large urban cities across the United States. Two of these four programs served high school students in under-resourced schools, and the other two programs served both high school and college students from historically-marginalized backgrounds. A fifth program operated in the Twin Cities region to support career pathways for both high school and college students in underserved communities. The sixth program served young adults without a high school diploma in Philadelphia by providing the opportunity and resources needed to earn a diploma and additional work readiness skills.

Programs were selected to participate in Study 2 as part of a larger study focused on understanding how youth and young-adult organizations support young people’s social capital development. Similar to Study 1, partner programs were recruited based on the following criteria (1) a shared mission to enhance education and/or career outcomes for adolescents and emerging adults by strengthening program participant’s social capital, (2) predominantly served young people from minoritized and/or historically marginalized communities, and (3) designed to intentionally connect program participants with others including peers, near-peers (i.e. individuals who serve as mentors/coaches and provide ongoing education and/or career goal support), program staff, and other community members.

Of the 702 young people included in Study 2, more than half identified as female (70.7%), 26.3% identified as male, 2.7% identified as non-binary, and 0.3% preferred to self-describe (e.g. gender fluid). Age ranged from 14–25 years (M = 19.22; SD = 2.49). Roughly a third (30.8%) of the sample identified as Hispanic/Latina/o/x, 23.8% identified as Black/African American, 31.9% identified as Asian/Pacific Islander, 9.2% identified as Multiracial, 2.9% identified as White, 0.4% identified as Native American or Alaskan Native, and 1.0% identified as another race.

Procedure

The scale from Study 1 was administered in an online survey. Partners invited all participants to take the survey over a 2-month period between January and March 2023. The purpose of the survey was to better understand how partner organizations and their program staff were supporting program participant’s social capital development, and to inform the development of tools and resources that program staff could use to support young people’s social capital growth. Program participants completed the survey on computers or tablets using a web-based survey that was hosted via a secure data collection platform. Program participants received a link to the online survey staff sent out via email. All participants had the opportunity to receive a $10 e-gift card for their participation.

Analytical plan

In Study 2, we conducted a confirmatory factor analysis (CFA) on the same set of items in Study 1 to further validate the construct validity of the proposed factor structure (e.g. Worthington & Whittaker, 2006). The CFA model provides a more rigorous test of the underlying factor structure, and can be used to evaluate how well each item in the scale loads onto a single factor (Kline, 2023). MLR estimation was used, and the fixed factor method was used to scale the latent constructs. Model fit was assessed using CFI, SRMR, and RMSEA. Common guidelines were used to assess goodness-of-fit: CFI > .90, SRMR < .06, and RMSEA < .08 (McDonald & Ho, 2002). Missing data on items ranged from 10.1% to 11.5%. Little’s MCAR test showed that the data were missing at random (χ² = 138.46, df = 162, p < .05).

Convergent and concurrent validity were tested using latent factor correlations between the proposed social capital scale and previously validated measures of developmental relationships and youth progress toward education and career goals. Developmental relationships were assessed with a five-item scale informed by Search Institute’s Developmental Relationships Framework (Pekel et al., 2018), which has previously demonstrated strong psychometrics and has been used to assess developmental relationships in other studies (e.g. see Boat et al., 2021; Scales et al., 2022). Example items include “my peers help me accomplish tasks or goals,” “my peers listen to my ideas and take them seriously,” and “my peers introduce me to new experiences or opportunities.” Progress toward education and career goals was assessed with a three-item scale (Boat et al., 2021). Items include “I have made a plan to reach my education or career goals,” “I am making progress toward my education or career goals,” and “I have already taken important steps toward pursuing my education or career goals.” Items across both scales were assessed on a 5-point scale ranging from Strongly Disagree (1) to Strongly Agree (5). Both the developmental relationships scale (α = .92) and the progress toward education and career goals scale were internally consistent (α = .86).

A positive correlation between the proposed social capital scale and developmental relationships with peers and program staff as well as with progress toward education and career goals was hypothesized based on past studies (Boat et al., 2021; Ferguson, 2006; Scales et al., 2022). It might appear that the developmental relationships measure is not appropriate for assessing the convergent validity of our social capital measure, because “social capital” does not appear in the measure’s name. However, the developmental relationships measure goes beyond more general assessments of high-quality relationships to assess additional relational qualities and resources beyond care and closeness to include providing support and advocating for youth, giving youth decision-making power, and connecting them with people and other resources that can help youth pursue their goals, all of which are important aspects of social capital.

Measurement invariance tests by gender, age, and racial-ethnic group were also conducted to determine if items held similar meaning for participants from different groups using a series of nested models (Little, 2013). Nested multi-group CFAs were run to examine configural (shown by equivalent factor structure), metric (shown by equivalent factor loadings), and scalar invariance (shown by equivalent item intercepts) across subgroups. The configural model with no imposed constraints is compared to the metric model with factor loadings constrained to be equal across groups, and the metric model is then compared to the scalar model with both the factor loadings and the item intercepts constrained to be equal across groups. In larger sample sizes, the χ² difference test statistic has been shown to be overly sensitive to differences between the actual and modeled covariance matrices (Cheung & Rensvold, 2002). Therefore, model invariance was assessed using the following alternative criteria: Δ CFI < .010 and Δ RMSEA < .015 (Cheung & Rensvold, 2002). Little to no changes in these metrics demonstrate no significant decreases in model fit, providing evidence for measurement equality across groups (Little, 2013).

Study 2: Results

Confirmatory factor analysis (CFA)

The three-factor model identified via EFA in Study 1 was cross-validated by a CFA model in Study 2. Model fit indices indicated the three-factor solution was a good fit to the data: SRMR = .039, RMSEA = .053, and CFI = .960. Overall the scale demonstrated a moderate to strong fit to the data, with the lowest factor loading being 0.63 and the majority loading above 0.70. Table 3 provides standardized factor loadings, standard errors, and R² values for the final CFA model. Correlations among the three latent constructs were statistically significant and positively correlated with one another, such that network diversity was positively correlated with network strength (r = .826, p < .001) and network mobilization (r = .718, p < .001). Network strength was also positively correlated with network mobilization (r = .777, p < .001). The three subscales of the Youth Network of Support Scale were also internally consistent, demonstrating Cronbach’s alpha estimates of .82 (network diversity), .91 (network strength), and .84 (network mobilization). The Cronbach’s alpha for the full scale was .93.

Table 3. Youth network of support scale: confirmatory factor analysis.

Latent Variable and Indicators	β	SE	R^2
Factor 1. Network Diversity (α = .82)
I have people in my network…
with many different careers or career interests	.774	.027	.599
from many different cultures or racial/ethnic backgrounds	.747	.033	.558
from many different economic backgrounds	.783	.029	.613
Factor 2. Network Strength (α = .91)
I have people in my network…
that I can trust to help me pursue my education or career goals	.819	.022	.670
that introduce me to others who can help me reach my education or career goals	.835	.017	.697
who I am close to that help me pursue my education or career goals	.813	.021	.661
who I am less close to but who are influential in helping me reach my education or career goals	.727	.025	.529
who help me when they say they are going to help me	.783	.022	.613
from when I was younger that I can ask for support, when needed	.631	.033	.398
with different skills that will be useful to me as I pursue my goals	.735	.025	.540
Factor 3. Network Mobilization (α = .84)
When working toward my education or career goals, I ask for help when I need it	.730	.032	.532
I go out of my way to meet new people in order to reach my education or career goals	.797	.021	.636
I form strong relationships with people who are useful for helping me reach my education or career goals	.800	.020	.640
I have already sought out people who can help me pursue my education or career goals	.664	.031	.442

Notes. All factor loadings were significant at p < .001; Standardized coefficients are provided.

Convergent and concurrent validity

Convergent validity was examined to evaluate whether the Youth Network of Support Scale was correlated with the conceptually related construct of youth-staff and peer-to-peer developmental relationships. Latent factor correlations showed that the full Youth Network of Support Scale was positively correlated with stronger youth-staff developmental relationships (r = .667, p < .001) as well as its core dimensions—network strength (r = .660, p < .001), network diversity (r = .573, p < .001), and network mobilization (r = .559, p < .001). Similar findings were also found when examining peer-to-peer developmental relationships with the full Youth Network of Support Scale (r = .692, p < .001) and its core dimensions—network strength (r = .669, p < .001), network diversity (r = .586, p < .001), and network mobilization (r = .646, p < .001).

Concurrent validity was examined to evaluate whether the Youth Network of Support scale was correlated with progress toward education and career goals. Latent factor correlations showed that the full Youth Network of Support Scale was positively correlated with greater progress toward education and career goals (r = .716, p < .001) as well as its core dimensions—network strength (r = .671, p < .001), network diversity (r = .618, p < .001), and network mobilization (r = .700, p < .001).

Measurement invariance

To examine whether the three-factor Youth Network of Support Scale possessed measurement invariance across gender, age, and race/ethnicity groups, we conducted nested multigroup measurement invariance tests. Measurement invariance testing was conducted with two gender groups: cis-gender girls and cis-gender boys. Due to the small sample size among youth who identified as White, Multiracial, and Native American/American Indian, these groups were combined in order to conduct invariance testing. Therefore, invariance testing was done with four racial-ethnic groups: Asian/Pacific Islander, Black/African American, Hispanic/Latina/o/x, and Other. Measurement invariance testing was conducted for two age groups: adolescence (14-17) and emerging adults (18-25). Configural, metric, and scalar invariance were all satisfied (see Table 4). All models provided good fit to the data and the changes in CFI and RMSEA did not exceed the evaluation criteria (i.e. Δ CFI < 0.01 and Δ RMSEA < .015). These results show that the scale is invariant by gender, racial-ethnic background, and age.

Table 4. Goodness-of-fit indicators for measurement invariance by subgroup.

Subgroup	Model	χ^2 (df)	RMSEA	Δ RMSEA	CFI	Δ CFI	SRMR	Δ SRMR	Decision
Gender	Configural	318.57 (148)	.062	–	.947	–	.046	–	Accept
	Metric	332.28 (159)	.060	.002	.947	.000	.058	.012	Accept
	Scalar	343.22 (170)	.058	.002	.947	.000	.058	.000	Accept
Racial-Ethnic Background	Configural	459.50 (296)	.060	–	.953	–	.053	–	Accept
	Metric	492.34 (329)	.057	.003	.953	.000	.070	.017	Accept
	Scalar	534.35 (362)	.055	.002	.950	.003	.074	.004	Accept
Age	Configural	285.85 (148)	.054	–	.958	–	.043	–	Accept
	Metric	301.89 (159)	.053	.001	.957	.001	.054	.011	Accept
	Scalar	323.70 (170)	.053	.000	.953	.004	.055	.001	Accept

Notes. *** = p < .001; ** = p < .01; * = p < .05. Due to the small sample size among youth who identified as White, Multiracial, Native American/American Indian, and Other race, these groups were combined in order to conduct invariance testing.

Discussion

Within the extant literature, social capital has been recognized as an invaluable asset for young people as they pursue education and career goals and forge pathways toward economic and social mobility (Dika & Singh, 2002; Mishra, 2020). Nevertheless, the PYD field has lacked a consensus on how to conceptualize youth social capital and has therefore produced limited options for validly and reliably measuring youth social capital at the individual level. This study addresses this need by developing and validating a brief social capital scale in partnership with 12 education and workforce development programs designed for adolescents and young adults from minoritized and historically-marginalized backgrounds within the United States. The scale provides actionable insights that both scholars and practitioners can use to unpack the factors that underpin individual youth social capital and its influence on youth outcomes, thereby broadening our understanding of its impact on the lives of youth.

The study reveals that youth social capital at the individual level is best understood as a multidimensional construct. Relatively consistent with our theorizing, the factors reflected three distinct components of social capital: network strength, network diversity, and network mobilization. Although all of these dimensions of social capital have been discussed and assessed to some degree in previous literature (Engbers et al., 2017; Lin, 2008), they have not been used together to fully capture the multidimensional nature of social capital. Moreover, this may be one of the first social capital measures informed by PYD frameworks to incorporate elements of youth agency by including items that assess youth’s ability to activate and mobilize their own social resources in pursuit of their goals. Additionally, the relatively high reliability of the full scale and its three separate dimensions supports its use in future research for examining the average as well as the unique and additive effects of each dimension on youth’s developmental trajectories and pathway toward thriving. Finally, unlike any youth social capital measure we are aware of, this study has demonstrated the invariance of our measure across age, gender identity, and racial-ethnic identity, making it especially useful for assessing social capital in more diverse samples of youth.

Prior research has suggested a strong relationship between youth social capital and the strength and quality of dyadic relationships (Boat et al., 2021; Scales et al., 2022), and education and career outcomes (Dika & Singh, 2002; Mishra, 2020). It is theorized that strong relationships are at the core of and are foundational to young people’s social capital (Varga & Zaff, 2018). It is often through these strong developmental relationships that youth gain access to weaker-ties and/or connections who have access to additional resources and opportunities that may not otherwise be available (Scales et al., 2020). In fact, high-quality relationships have often been used as a proxy or an indicator of youth social capital (Ferguson, 2006). In line with this prior research, we found significant positive correlations between the measurement of developmental relationships with peers and program staff and the measure of youth social capital. Moreover, we also theorized that our Youth Network of Support Scale would be positively correlated with youth’s progress toward education and career goals. Our hypothesis was confirmed. Both the full scale and each of its dimensions were positively correlated with greater progress toward education and career goals. These findings provide preliminary evidence of convergent and concurrent validity.

Unlike any of the youth social capital measures examined in the Ahlborg et al. (2022) systematic review, the Youth Network of Support Scale also shows strong measurement invariance, demonstrating invariance by racial-ethnic identity, gender, and age using two independent samples that included a large racially and ethnically diverse population of adolescents and young adults. These findings support the use of the scale across a range of groups. Yet, it is important to note that the sample size of some racial and ethnic groups were smaller and thus needed to be combined in order to assess measurement invariance. Additionally, our sample was primarily composed of youth and young adults of color who were participants in education support and workforce development programs. It will be important for future research to continue to assess measurement invariance across a large range of youth with different identities and life experiences.

Because the Youth Network of Support Scale captures multiple dimensions of social capital, it may be useful for helping researchers better understand how youth social capital results in positive youth developmental outcomes. For example, while both network strength and diversity have been identified as components of social capital, little is known if either of these aspects of social capital are more strongly associated with specific developmental outcomes (e.g. educational attainment) than others across both contexts and time. Furthermore, the current Youth Network of Support Scale includes a measure of network mobilization, which may be used to better understand how social capital results in positive outcomes. Therefore, it will be important for future research to examine how these three components work in concert to influence positive youth trajectories.

This measure also holds promise for practitioners conducting program evaluations or undergoing organizational improvement efforts. The scale was developed alongside education and workforce support programs that serve adolescents and young adults who are pursuing education and career goals. The majority of these programs were interested in improving their programming to support the social capital development of the young people they served, and looking for a measure that could provide actionable knowledge to both make program improvements and better understand where youth need additional support. The fact that the items were developed with the significant input from youth, program staff, and alumni in these programs adds to the measure’s content validity and face validity, and provides strong prima facie evidence for the actionability of the measure, because in the measure development process the program partners advised the researchers as to which items would be most practically useful to them. While it will be important to validate the scale in other contexts and settings, we believe that this brief self-report scale is useful for assessing group-level trends, and that the insights that are gleaned from its use can help guide practitioners in how to better support young people’s network strength, diversity, and mobilization.

Limitations and future directions

The current study had several limitations. While the current study provides some support for the scale’s convergent and concurrent validity, it will be valuable for future research to continue to test convergent validity with other previously established social capital scales. Moreover, it will be important to assess the predictive validity of the scale using longitudinal data to see if the scale is positively associated with theorized outcomes such as educational attainment and/or employment. The scale was developed and validated with youth-serving organizations focused on the educational and workforce development of adolescents and emerging adults from minoritized and/or historically-marginalized communities, therefore it may not be generalizable across all youth populations or settings. While the scale was intentionally designed not to be specific to one context (e.g. schools, family), it will be important for future studies to examine the validity of the scale in other such contexts. Qualitative work may also be helpful to determine what other dimensions might be added to the current scale, as well as the utility of the scale for schools and other youth-serving organizations in informing and improving their impact. Additionally, because of the small sample size among youth who identified as White, Multiracial, and Native American/American Indian, these groups were combined in order to conduct invariance testing. This group is not interpretable due to the differences in identity and life experiences among these youth. Future research should continue to assess how well the scale operates across youth with various racial-ethnic identities. Finally, although the items in the network diversity subscale do effectively assess youths’ perception of the racial, ethnic, and socioeconomic diversity of their networks, none of the items currently addresses in an explicit way how much youth believe they are experiencing social capital that helps them contribute to social change, what Dill and Ozer (2019) called “critical social capital,” that advances racial, ethnic, and socioeconomic equity in educational and occupational opportunities. Further research with the Youth Network of Support Scale should explore how much youths’ experience of social capital is helping equip them for such civic roles in the promotion of equity, and whether this individual youth social capital can contribute to collective efficacy and greater community-level forms of social capital.

Conclusion

The current study developed and validated a Youth Network of Support Scale that can be used by both scholars and practitioners who are committed to supporting the well-being of young people. Positive features of the instrument that likely render it being considered practical and actionable include its brevity (i.e. exerting less burden on users), voice of users in its development, and evidence of validity and invariance across race, gender identity, and age. If the Youth Network of Support Scale continues to operate well among other contexts and with additional samples, many possibilities for future research will open. This research has the potential to advance our understanding of youth social capital, lead to a greater consensus on the conceptualization of individual youth social capital, and strengthen how the field of PYD can be leveraged to help us better understand the role of youth agency in activating and mobilizing their network or web of support. The resulting insights will hopefully lead to more effective and impactful practice and expanded opportunities for all youth to pursue their life goals.

Disclosure statement

The authors have no competing interests to declare

Data availability statement

The data that support the findings of this study are available from Search Institute upon reasonable request.

Additional information

Funding

This work was supported, in whole or in part, by the Bill & Melinda Gates Foundation [INV-033362]. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission.