Holding the Space: Individual- and Group-level Factors Predicting Member Retention in Gender-Sexuality Alliances

Abstract

Involvement in extracurricular groups is associated with positive outcomes for youth. Gender-Sexuality Alliances (GSAs) are school-based clubs that may provide benefits to sexual and gender minority (SGM) youth and their allies, yet little is known about what factors predict member retention. The current study explores individual- and group-level predictors of youth’s sustained or discontinued membership in GSAs during a school year. Participants were 410 youth (M_age = 15.56; 83% sexual minority; 57% cisgender female; 70% White) and 50 advisors in 32 GSAs purposively sampled across Massachusetts who completed surveys at the beginning and end of the school year. Sexual minority youth and youth who took on more leadership roles at the beginning of the year were less likely to have left their GSA by the end of the school year. Youth who perceived higher levels of social support from their GSA trended less likely to leave their GSA as well, although the association was not significant. Youth who engaged in more advocacy were more likely to have left. Youth in GSAs with greater structure to their meetings were less likely to discontinue their membership; specifically, having a meeting agenda was uniquely predictive of member retention. Implications for GSA inclusivity and practices to promote retention within GSAs and similar social justice-oriented clubs are discussed.

Introduction

Youth’s experiences at school occur in a broad range of contexts, as they interact with teachers, peers, and administrators in the classroom, in the halls, and in extracurricular groups. These interactions can drive students’ sense of school belonging, wellbeing, and achievement in school. Specifically, extracurricular involvement in high school may serve as a health promotive factor for youth. Extracurricular clubs are prominent in many schools and provide opportunities for youth to connect and build a sense of community with their peers (Dorsey, 2020; Martinez et al., 2016). Indeed, extracurricular involvement is associated with a myriad of positive developmental outcomes such as positive identity, commitment to learning, empowerment, and social competencies (Forneris et al., 2015; Han et al., 2017). Social justice-oriented groups in particular may serve as sources of support for marginalized youth and their allies, and their presence is associated with safer school climates and less discrimination (Marx & Kettrey, 2016). Youth who participate in these types of groups also tend to report a greater sense of empowerment and school belonging (Iwasaki, 2016; Toomey & Russell, 2013).

We focus on one such group, Gender-Sexuality Alliances (GSAs). GSAs have a growing presence in schools and are now in about 37% of U.S. secondary schools (CDC, 2019). They tend to be youth-led and adult-supported and typically meet weekly. Although GSAs are not standardized programs, they share several unified and primary aims related to providing support, sexual and gender minority (SGM)-affirming resources, and opportunities for awareness-raising and advocacy among SGM youth and their ally peers (Griffin et al., 2004; Poteat et al., 2017). Specifically, GSAs often serve as a place for members to solicit social-emotional support in the face of experiencing discrimination or other stressors, as well as to provide mutual support to their peers (Graybill et al., 2015; Griffin et al., 2004). GSAs may be among the only visibly SGM-affirming spaces in schools for youth to meet, socialize, and build a sense of community (Lee, 2002). Straight cisgender allies often are motivated to be in GSAs through their friendships with and desire to support SGM members (Scheer & Poteat, 2016). In addition, GSAs can serve as a source of information, education, and opportunities to discuss and learn about SGM identities or to access SGM-affirming resources or referrals to other SGM-affirming groups in the community (Griffin et al., 2004; Poteat et al., 2015). Finally, GSAs engage in a range of advocacy efforts to raise awareness of SGM identities and topics, and to counteract instances of discrimination and discriminatory school policies (Poteat et al., 2015; Russell et al., 2009). Of note, both SGM and heterosexual cisgender youth in schools with GSAs report less victimization and less fear for their safety (Marx & Kettrey, 2016). Given these multiple key roles that GSAs play and their potential benefits for SGM youth and their allies, they merit greater attention.

Because school climates can be heterosexist and cissexist, GSAs may be especially important for SGM youth, who experience greater victimization and lower school belonging than their heterosexual, cisgender peers (Baams et al., 2015; Davis et al., 2014). Moreover, at times SGM youth avoid other extracurricular groups due to perceived hostility (McGuire et al., 2016). GSAs may therefore be a key setting in which SGM youth can reap the same benefits that their heterosexual, cisgender peers may enjoy from other extracurricular activities. Similarly, heterosexual cisgender youth also can experience homophobic victimization from peers (Collier, et al., 2013), and thus GSAs also may be important for them.

Some attention has been given to questions related to youth’s membership in GSAs (Toomey et al., 2011). A few studies have considered why youth choose not to join GSAs, noting that SGM students indicate that fear, perceived group disorganization, and lack of consistent peer and adult relationships deter them from joining (Heck et al., 2013; McCready, 2001; McGuire et al., 2016). Others have looked to factors that characterize members who are more involved than others, including youth’s perceptions of support and the GSA’s respectful climate and meeting structure (Poteat et al., 2016).

A critical question remains unexamined: who is more likely to stay involved in their GSA, and who is more likely to discontinue their membership over time? Membership retention is a challenge that a number of GSAs can face (Poteat et al., 2015), and it differs from a struggle to recruit potential members due to their peers’ decisions not to join the GSA. Some youth who discontinue their membership may have made an active decision to stop coming to their GSA based on their personal experiences in the GSA, whereas youth who choose not to join their GSA have done so more on the basis of their perceptions or anticipations of the GSA (Heck et al., 2013; McCready, 2001; McGuire et al., 2016). Any given individual’s choice to discontinue their involvement in a group may result from negative (e.g., feeling unwelcomed), positive (e.g., feeling empowered to move into other actions outside the group), or mixed experiences. Youth retention also stands in contrast with youth engagement in that members who do not participate actively could still derive benefits from attending through a sense of community and internalizing positive messages or receiving information shared during meetings.

Member retention is also unique in its potential impact on the GSA as a group. GSAs aim to provide collective and sustained support to their members (Griffin et al., 2004; Poteat et al., 2015). Consequently, difficulty with member retention could lead to difficulties in providing a consistent space for youth to solicit support from other reliable peers in that space. Research on other types of support groups identify member turnover as a concern and a potential disruption to group functioning (Ellison & Harder, 2018). Further, advocacy efforts require sustained commitments among multiple members who may share responsibilities. Thus, member discontinuation during a school year could prove challenging for others in the GSA to maintain their momentum in working toward these larger goals. For instance, a group with more stable membership, even with some members who may be relatively less engaged than others, may have greater capacity to organize school-wide advocacy events (e.g., Day of Silence) or initiatives (e.g., securing gender-neutral bathrooms).

With these points in mind, it would be important for GSAs to understand factors that underlie the discontinued membership of certain members over the course of a school year. We consider youth’s initial experiences in their GSA at the beginning of the year and characteristics of the GSA that could predict the relative likelihood of youth discontinuing their membership by the school year’s end. We focus on youth’s initial perceptions and experiences because there is strong evidence that an individual’s initial impressions have lasting effects and influence and shape their future impressions and behavior (Harris & Garris, 2008). Thus, although youth’s experiences in their GSA could change over time, their experiences at the beginning of the school year could affect their interpretation of these future experiences (or whether they are open to future experiences) and go on to have a predictive effect in whether they remain involved through the end of the year. This information could help GSA advisors and youth leaders attend to members’ experiences within the group and structure their GSA to enhance retention.

Individual-Level Predictors of Discontinued GSA Membership

We draw from positive youth development program models and research (e.g., Catalano et al., 2004; Pearce & Larson, 2006) to consider a range of relevant predictors of youth’s likelihood of discontinued GSA membership over time, including both individual- and group-level factors. Although GSAs are not standardized programs, their aspirations and the ways in which they function are similar to other groups based on positive youth development. Some essential elements of successful youth programs include their provision of a safe and supportive space to interact, leadership opportunities for youth, sufficient organizational structure for group meetings, and involving adults as mentors (Lerner et al., 2015). Catalano and colleagues (2004) also list a number of similar characteristics that distinguish effective youth programs, which include opportunities for bonding, building competence, developing one’s identity, and fostering self-determination. That is, youth programs constitute social and personal opportunities for exploration and growth. These factors have consistently been associated with positive outcomes for youth such as civic engagement (Holt et al., 2017; Jones & Deutsch, 2011; Zeldin et al., 2013). Accordingly, we propose a model (Figure 1) that includes factors related to these prosocial features, including youth’s perceptions of social support from peers (e.g., opportunities for bonding) and taking on leadership opportunities in their GSA (e.g., building competence, self-determination). Furthermore, we introduce other factors unique to social justice-oriented groups, such as youth’s level of advocacy.

Figure 1.

Multilevel Mixed-Effect Logit Model Where (a) Individuals Factors (Level 1) and (b) GSA and Advisor Characteristics (Level 2) Predict the Likelihood of Discontinued Membership Before the End of the Year.

Social support.

We consider the dimension of social connection in the GSA by assessing youth’s initial perceptions of social-emotional support from peers in the GSA. Because SGM youth often lack support for their SGM identities in other spaces (Davis et al., 2014), and heterosexual, cisgender youth can also experience homophobic victimization or stigma (DeLay et al., 2017; Rivers, 2011), GSA members who perceive stronger social-emotional support from peers during the initial GSA meetings of the school year might be motivated to continue attending. This stands to reason, as GSAs often emphasize their role in providing social support (Griffin et al., 2004) and greater social support is associated with greater engagement during GSA meetings (Poteat et al., 2016). Thus, we expect that youth who perceive greater social support from peers in their GSA at the beginning of the year will be less likely than others to discontinue their GSA membership by the year’s end.

Leadership roles.

Youth program models also point to leadership opportunities as integral to their success (Catalano et al., 2004; Lerner et al., 2015). Leadership can include contributing to the group in a way that is meaningful and valued by others (e.g., assisting to organize the agenda of GSA meetings) and engaging in activities that are scaffolded and meant to build confidence, such as serving on a panel for an education event (Catalano et al., 2004; Oliver et al., 2006; Wood et al., 2009). Some findings suggest that leadership can promote a youth’s sense of investment in and responsibility to a group (Forgeard & Benson, 2019; Ryan & Deci, 2000; Vinoski et al., 2016). As such, youth who report taking on more leadership roles at the beginning of the year may be less likely than other youth to discontinue their GSA membership by the year’s end.

Advocacy.

Youth programs also place an emphasis on opportunities to build competence and develop a sense of self-determination. Within the context of GSAs, these opportunities may come in part through youth’s involvement in their group’s advocacy efforts, which can empower them to take collective action against discrimination and other social stressors that affect them. Advocacy is a major feature of GSAs and other social justice-oriented groups (Poteat et al., 2017), yet it is generally not a feature considered in research on traditional youth programs or extracurricular groups based on recreational or scholastic interests (Farb & Matjasko, 2012; Larson et al., 2006). However, studies suggest the potential benefits of advocacy for youth. For instance, youth’s advocacy in GSAs is associated with a greater sense of purpose, empowerment, and well-being (Iwasaki, 2016; Poteat et al., 2015; Russell et al., 2009). In this sense, youth who are more involved in advocacy in their GSA at the start of the year may derive such benefits along the way and thus be motivated to remain in the GSA. On the other hand, whereas social support is more inward-facing and focuses on individual members’ needs, advocacy is more of an outward-facing action to address broader issues external to the GSA (e.g., to counteract bias-based harassment or call for SGM-inclusive curricula; Griffin, 2004). Consequently, youth who are initially more active in advocacy might be more likely to discontinue their GSA membership to pursue these efforts beyond the GSA. We therefore consider advocacy in an exploratory manner in relation to discontinued membership, while noting that the prominence of advocacy in GSAs makes it a potential predictor of youth’s likelihood to discontinue membership.

Demographic identities.

Although it has not been a major focus in youth programs, scholars have called for greater consideration of diverse identities in the youth program literature (Williams & Deutsch, 2016). By their nature, GSAs aim to bring together youth from diverse backgrounds (Chong et al., 2019). Still, youth from some groups may feel more welcomed than others and thus may be more likely to remain members. For instance, given the unique focus of GSAs on SGM issues, SGM members may be less likely to discontinue their GSA membership than heterosexual, cisgender youth. We also consider possible racial/ethnic differences, as some research suggests that racial/ethnic minority youth perceive their GSAs to be less supportive than their White peers do, as they face extensions of racial segregation within queer programs, as well as limited awareness of racial/ethnic issues among advisors (McCready, 2001; Poteat et al., 2015). At the same time, racial/ethnic minority youth may have different experiences depending on setting characteristics (Mistry & Wu, 2010). For example, Black youth in a mostly White GSA may have different patterns of discontinued membership than Black youth in a majority Black GSA due to the salience of race, the types of support they perceive, or other factors. Therefore, we consider in an exploratory manner whether any racial/ethnic differences in youth’s likelihood of discontinued GSA membership may be conditional on the racial/ethnic make-up of the larger group.

Membership duration.

The length of time that youth have been members of their GSAs may be an important covariate when considering other contributing factors such as support and leadership. Members with longer tenure may have deeper relationships with others, as well as greater comfort taking on leadership roles. Therefore, in addition to serving as a covariate, we hypothesize that youth who have been members of their GSA for a longer period of time will be less likely to discontinue their membership by the end of the year.

Grade level.

Finally, we consider whether youth differ in their likelihood of discontinued GSA membership by grade level. Youth participation may wane as they progress through school for several reasons. First, they may gain more opportunities to engage in efforts outside of their school (Cooley, 1992). Second, they may face additional responsibilities that prevent them from further GSA involvement (e.g., completing college applications or increasingly challenging coursework).

Group-Level Predictors of Discontinued GSA Membership

We consider several characteristics of GSAs and their advisors that might predict members’ likelihood of discontinuing their GSA membership. Although positive youth development program models point to safe and supportive environments, organizational structure, and adult mentors as features of successful youth programs (Catalano et al., 2004; Lerner et al., 2015), group characteristics have received much less direct attention than individual experiences in the GSA literature (Poteat et al., 2017). There has been some descriptive work on GSA advisors and their experiences (Graybill et al., 2009; Watson et al., 2010) and how GSAs generally function (Griffin et al., 2004; Poteat et al., 2015), but there is a paucity of research that directly examines the associations between advisor characteristics or group dynamics and youth outcomes. Consequently, there remains little data upon which to base recommendations for GSA advisors’ roles or a GSA’s structure that might enhance member retention. In accordance with the broader youth program literature (Catalano et al., 2004; Lerner et al., 2015), our model includes factors related to safe and supportive environment (specifically, open and respectful climates), to program structure (in the form of organizational structure to meetings), and adult mentor features (specifically, GSA advisor self-efficacy to discuss SGM issues).

Open and respectful climate.

The emphasis of youth program models on creating a safe and welcoming environment can extend beyond an individual member’s own perceptions of support to also include a group’s overall open and respectful climate. This, too, may predict sustained membership of individuals within a GSA. Open and respectful climates are characterized as ones wherein youth can share opinions and express differences of beliefs through discussion (Gould et al., 2012; Kutsyuruba et al., 2015). They have been examined in classroom contexts and shown to be related to youth’s efficacy to effect change and social justice in their schools and in society (Godfrey & Grayman, 2014). Additionally, youth’s perceptions of a positive climate in a program are associated with their sense of belonging in the group (Byrd & Martin, 2016). Thus, we predict that youth in GSAs with open, respectful climates may be less likely to discontinue their membership.

Meeting structure.

Next, we consider whether having more structured GSA meetings protects against member discontinuation. In a review of the youth program literature, structure was identified as an integral aspect of 96% of all programs that were evaluated as effective (Catalano et al., 2004). In the context of GSAs, having a structure to meetings may include practices such as conducting check-ins at the start of meetings, having clear meeting agendas, or designating leaders or facilitators (Poteat et al., 2016). Within GSAs with greater meeting structure, advisors, youth leaders, and members may be more likely to identify members at risk of discontinuing their membership (e.g., out of perceived lack of support or invisibility in the group) and to be more responsive and provide them with needed support. Additionally, greater meeting structure may lead to more organized, focused meetings. This may also protect against discontinued membership, given that perceived disorganization is one reason that youth report not joining their school’s GSA in the first place (Heck et al., 2013). Thus, we expect that GSAs with a higher degree of meeting structure will have lower member discontinuation.

Advisor SGM self-efficacy.

Traditional positive youth development program models have pointed to the importance of adult mentors and role models for youth (Jones & Deutsch, 2011; Zeldin et al., 2013), and we therefore consider adult GSA advisors here. Importantly, though, these existing models often have not explicitly considered mentor competence in working with youth from diverse and marginalized backgrounds. In groups such as GSAs that are focused on social equity and that serve youth from marginalized populations, advisors’ preparedness to address issues of sexual orientation and gender diversity could have a significant impact on youth’s decisions to remain in the GSA. Thus, we consider advisor self-efficacy in discussing SGM-related issues in their GSAs. In GSAs, advisors could be viewed by youth as trustworthy sources of information on SGM issues or SGM-affirming resources, as well as mentors with knowledge of their unique stressors or sources of strength. Advisors who report more SGM-related competency, for example, may be able to provide unique support to youth by fostering more effective discussions on issues affecting SGM youth or by offering referrals to SGM-affirming healthcare providers or community groups. This could be important because SGM youth benefit more from LGB-specific coping strategies than from general ones (Toomey et al., 2018) and GSAs may be one of few spaces in school where SGM youth can process their experiences and needs with their peers and a trusted adult. Thus, we expect that greater advisor self-efficacy in discussing SGM issues will predict less member discontinuation in those GSAs.

Other GSA attributes as covariates.

We include GSA size and GSA meeting frequency as covariates in the model. Because providing social and emotional support is a primary aim of GSAs (Griffin et al., 2004), GSAs with more members may have greater capacity to provide that support. Similarly, GSAs that meet more regularly may provide more opportunities for members to remain engaged. Size and meeting frequency could also relate to the climate, structure, and self-efficacy of advisors, and thus are important to adjust for in our models.

Current Study

In this study, we take a multilevel approach to predict youth’s likelihood of discontinued membership in their GSA by the end of the school year based on their initial individual experiences and GSA characteristics at the beginning of the school year. At the individual level, we hypothesize that SGM youth, youth who perceive greater initial social-emotional support in the GSA, and youth who initially take on more leadership roles in the GSA will be less likely to discontinue their GSA involvement by the year’s end. We hypothesize that youth who have been members of their GSA longer and who are in higher grade levels may be more likely to discontinue their involvement in the GSA. In an exploratory manner, we consider whether youth’s race/ethnicity and level of initial advocacy predict their eventual likelihood of discontinued membership. We also consider, in an exploratory manner, the cross-level interaction between youth’s own racial/ethnic minority identity and the racial/ethnic make-up of their GSA. At the group level, we hypothesize that a more open and respectful climate, greater meeting structure, and greater advisor self-efficacy in discussing SGM-related issues will predict a lower likelihood of discontinued membership by the end of the year. Finally, we adjust for the number of GSA meetings and the number of students in the GSA.

Methods

Participants

Participants came from a larger GSA project of 580 youth and 58 adult advisors in 38 GSAs in high schools across Massachusetts (Poteat et al., 2020). Most schools included grades 9 through 12, though some were more expansive to include grades 7 through 12. The larger project was focused on youth’s experiences in their GSAs and ways in which GSAs promoted wellbeing among their members. Data were collected during the academic years of 2016-2017 and 2017-2018, with 19 GSAs participating in year 1 and 19 GSAs participating in year 2. In 32 of the 38 GSAs, advisors responded to a request for information at the year’s end about which of their participating members at the beginning of the year had discontinued their involvement in the GSA at some point during the school year prior to the year’s end. From these advisor reports, 85 youth were identified as having discontinued their GSA involvement. These 85 youth, along with the other youth in these 32 GSAs who participated at the beginning of the year (n = 492), comprised the sample for the current study. Among the 32 GSAs, 17 had one advisor and 15 had more than one (13 had two, one had three, and one had four advisors). Youth demographic data are included in Table 1.

Table 1.

Youth Demographics and Variable Descriptive Data

Variable	N (%)	M (SD)	Range
Sexual orientation
Gay or Lesbian	88 (17.9)
Bisexual	107 (21.8)
Questioning	29 (5.9)
Heterosexual	88 (17.9)
Pansexual	103 (20.9)
Asexual	17 (3.5)
Queer	21 (4.3)
Other written-in response	38 (7.7)
Not reported	1 (0.2)
Gender identity
Cisgender Male	83 (16.9)
Cisgender Female	273 (55.5)
Transgender	40 (8.1)
Genderqueer	12 (2.4)
Gender Fluid	13 (2.6)
Non-Binary	29 (5.9)
Other written-in response	40 (8.1)
Not reported	2 (0.4)
Race or ethnicity
White, non-Hispanic	338 (68.7)
Black or African American	17 (3.5)
Asian or Asian American	19 (3.9)
Latino/a/x	53 (10.7)
Biracial or Multiracial	53 (10.7)
Native American	2 (0.4)
Middle Eastern or Arab American	1 (0.2)
Other Race / Ethnicity	6 (1.2)
Not reported	3 (0.6)
Level 1 Variables
GSA Support		3.46 (0.58)	0.86 - 4.00
Advocacy		2.63 (0.96)	0.00 – 4.00
Organizing Roles		1.96 (1.34)	1.00 – 5.00
Age		15.57 (1.35)	10.00 – 18.00
Membership Duration		1.26 (1.19)	0.00 – 6.00
Grade Level		10.38 (1.25)	6.00 – 12.00
Level 2 Variables
Open, Respectful Climate		18.61 (1.64)	14.00 – 20.00
Structure		15.71 (2.43)	9.00 – 20.00
Advisor SGM Self-Efficacy		33.32 (4.35)	24.00 – 40.00
Meeting Frequency		4.61 (0.68)	3.00 – 6.00
Number of Students in GSA		18.05 (8.20)	6.50 – 35.00

Note. Level 2 GSA attributes represent the averages of advisor responses in cases where GSAs had more than one active advisor.

We considered whether youth in GSAs whose advisors responded to our inquiry about discontinued membership (32 GSAs, 492 youth) differed from those whose advisors did not respond to our inquiry and so who were not included in the analyses (6 GSAs, 88 youth). Youth in GSAs whose advisors did not respond reported slightly lower social support than youth in GSAs whose advisors did respond (Advisors did not respond: M = 3.30, SD = 0.72; Advisors responded: M = 3.47, SD = 0.58; F[1, 571] = 5.23, p = 0.02, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.009$). Youth in GSAs whose advisors responded were also more likely to identify as sexual minority (${\mathrm{\chi }}^2=8.69$, p = 0.003). No other individual differences were identified. At the group level, GSAs whose advisors did not respond had higher reported levels of structure at the beginning of the year (Advisors did not respond: M = 17.08, SD = 1.97; Advisors responded: M = 15.71, SD = 2.44; F[1, 578] = 24.74, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.041$), lower advisor SGM self-efficacy (Advisors did not respond: M = 30.65, SD = 3.64; Advisors responded: M = 33.32, SD = 4.35; F[1, 578] = 29.39, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.048$), and more members (Advisors did not respond: M = 24.09, SD = 10.96; Advisors responded: M = 18.05, SD = 8.20; F[1, 578] = 36.23, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.059$).

Procedures

We consulted with the Massachusetts Safe Schools Program for LGBTQ Students to identify GSAs in traditional public, public charter, and public vocational/technical schools across Massachusetts. We purposively sampled GSAs across Massachusetts for geographic diversity and diversity in the size and racial and socioeconomic composition of the schools. We secured permission from GSA advisors and principals to work with their GSAs. We asked youth members to participate in a study of their experiences in the GSA and informed them that their individual responses would be kept confidential. Advisors gave consent for all youth to participate, and all youth gave their assent. We used advisor adult consent over parent consent to avoid risks of inadvertently outing LGBTQ youth to their parents, a common practice among LGBTQ youth research to protect youth’s safety (Mustanski, 2011). All advisors who were recruited consented to complete their own surveys. Youth completed surveys regarding their personal experiences in their GSA and school, and advisors completed surveys regarding characteristics of the GSA as a whole. This was done in order to provide multiple-informant data and also in recognition that advisors could be a unique source of information on the GSA as a whole, given their consistent presence across GSA meetings and historical or institutional knowledge. Procedures were approved by the primary institution’s review board and by each school.

This study was conducted over a two-year period (the 2016-17 school year, and the 2017-18 school year), involving 19 GSAs in the first year and a second set of 19 GSAs the second year. This approach ensured that we could visit all participating GSAs within a close time frame each year. Toward the beginning of the school year (between mid-September and late-October), we distributed and collected surveys at a GSA meeting. The survey took 30 minutes to complete, and proctors were present to answer questions. Each participant received a $10 gift card upon completion of the survey.

Measures

Demographic information.

At the beginning of the year, youth reported their sexual orientation, gender identity, race/ethnicity, grade level, and membership duration. Table 1 shows the response options and frequencies of responses. Youth were able to write in a response if their identity was not already listed in the survey.

Perceived GSA support.

At the beginning of the year, youth reported the extent to which they perceived that their GSAs provided them social and emotional support using a 7-item subscale from the GSA Involvement scale (Poteat et al., 2016). The items were preceded by the stem, “How much do you personally feel you experience in your GSA up to this point this year” with items such as “a place of safety,” “a place where I share any concerns,” and “emotional support.” Response options ranged from 0 (not at all) to 4 (a lot). Higher average scale scores represent greater social-emotional support. The coefficient alpha reliability estimate was $\alpha =.83$.

Self-reported advocacy.

At the beginning of the year, youth reported the extent to which they personally engaged in various school- and community-based advocacy activities specifically through their GSA using another 7-item subscale from the GSA Involvement scale (Poteat et al., 2016). Example items include “organized school events to raise awareness of LGBTQ issues,” “speak out for LGBTQ issues,” “did advocacy events in the community,” and “worked with other student groups on diversity issues.” Response options ranged from 0 (not at all) to 4 (a lot). Higher average scale scores represent greater advocacy. The coefficient alpha reliability estimate was $\alpha =.88$.

Individual leadership roles.

At the beginning of the year, youth reported the extent to which they helped to organize or lead their GSA meetings, on an item, “Do you help organize or lead GSA meetings?” Response options were never, rarely, sometimes, often, and very often (scaled 1 to 5). Higher scores represent taking greater leadership responsibility.

Advisor SGM self-efficacy.

At the beginning of the year, advisors reported the extent to which they felt competent discussing LGBTQ issues using a 9-item scale. The items were preceded by the stem “How competent do you feel to do the following:” and included items such as “talk about unique experiences that LGBQ students face,” “talk about unique experiences that transgender students face,” “address issues related to sexual orientation,” and “talk about ways to support transgender students.” Response options ranged from 0 (not at all) to 5 (very). For GSAs in which there was more than one advisor, we used their average score in our analyses. The coefficient alpha reliability estimate was $\alpha =.92$.

GSA organizational structure.

At the beginning of the year, advisors reported the degree of organizational structure to their GSA meetings on a 4-item index: “we do check-ins at the start of GSA meetings,” “we follow up about things that were discussed in the last GSA meeting,” “our GSA meetings follows an agenda,” and “I or a student (or students) lead/co-lead our GSA meetings.” Response options were never, rarely, sometimes, often, and all of the time (scaled 1 to 5). These items encompass a wide variety of ways in which advisors and youth leaders of GSAs may structure their meetings, and reflect practices identified in prior descriptions of GSA meetings (Griffin et al., 2004; Poteat et al., 2015). We created a total index score with higher scores representing greater organizational structure in the GSA. For GSAs in which there was more than one advisor, we used their average score in our analyses. The coefficient alpha reliability estimate was α = .48. Despite the low reliability, we decided that the measure held face validity in reflecting common GSA organizational practices (Griffin et al., 2004; Poteat et al., 2015). Nevertheless, we ran a sensitivity analysis in which all items in this scale were considered separately so as to obtain a more detailed picture of organizational practices that could predict later discontinued membership.

Open, respectful GSA climate.

At the beginning of the year, advisors completed the 4-item Open Classroom Climate scale (Flanagan et al., 2007), which we modified to reference the GSA. Items were preceded by the stem, “Up to this point this year, in our GSA, students…” with items such as “have a voice in what happens” and “can disagree with each other, if they are respectful.” Response options ranged from 1 (strongly disagree) to 5 (strongly agree), with higher average scale scores representing a more open and respectful climate in the GSA. For GSAs in which there was more than one advisor, we used their average score in our analyses. The coefficient alpha reliability estimate was $\alpha =.66$.

GSA meeting frequency.

At the beginning of the year, advisors were asked to report how often their GSAs met. Response options were no regular meetings scheduled, once or twice a semester, once per month, every other week, once per week, and more than once per week. These responses were scaled from 1 to 6, with higher scores representing more frequent meetings. For GSAs in which there was more than one advisor, we used their average score in our analyses.

Number of GSA members.

At the beginning of the year, advisors were asked to report how many students were “regular” or “active” members in the GSA. For GSAs in which there was more than one advisor, we used their average score in our analyses.

Discontinued membership.

At the end of the school year (late May), we asked advisors to indicate which of their members from the beginning of the school year had “generally stopped coming to GSA meetings earlier in the year.” We used their responses to indicate which youth participants had discontinued their involvement in the GSA (discontinued membership = 1) and which remained regular members (discontinued membership = 0).

Analytic Approach

In preliminary analyses we considered associations among our variables. Next, we used multilevel modeling in Stata/IC 15.1 to test our models using a mixed-effect logit model, which allowed us to account for the interdependence of youth nested within GSAs and to include variables at the individual level (Level 1) and the GSA level (Level 2). Our set of variables predicted youth’s likelihood of discontinuing their GSA involvement before the year’s end.

With discontinued involvement as the dependent variable, we included the following independent variables at the individual level (Level 1): perceived initial support from the GSA, self-reported initial advocacy, leadership responsibility, sexual orientation, gender identity, race/ethnicity, membership duration, and grade level. Sexual orientation, gender identity, and race/ethnicity were included as binary variables (0 = heterosexual, 1 = sexual minority; 0 = cisgender, 1 = gender minority; 0 = White, 1 = racial/ethnic minority) due to the small representation of youth within some of the specific sexual minority, gender minority, and racial/ethnic minority groups. Continuous measures were grand-mean centered. We included the following variables at the GSA level (Level 2): open and respectful climate in the GSA, the level of organizational structure in the GSA, advisor SGM self-efficacy, GSA meeting frequency, and number of members in the GSA.

Results

First, there were significant correlations between our individual-level variables. Social support and advocacy were moderately correlated (r = .43, p < .001), and there was a small, positive correlation between leadership and social support (r = .12, p = .007), as well as between leadership and grade level (r = .28, p < .001). Membership duration was correlated with leadership (r = 0.49, p < 0.001) and grade level (r = 0.52, p < 0.001). All other correlations were not significant. In addition, we used ANOVAs to test for identity-based group differences. Sexual minority youth reported lower levels of advocacy in the GSA (M = 2.58, SD = 0.95) than heterosexual youth (M = 2.86, SD = 0.99; F[1,485] = 5.77, p = .017, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.012$). Gender minority youth also reported lower levels of advocacy in the GSA (M = 2.42, SD = 0.94) than cisgender youth (M = 2.72, SD = 0.96; F[1, 483] = 9.32, p = .002, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.019$), as well as lower social support (Gender minority youth: M = 3.38, SD = 0.62; Cisgender youth: M = 3.50, SD = 0.56; F[1, 485] = 3.95, p = .047, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.008$) and more leadership roles (Gender minority youth: M = 2.17, SD = 1.40; Cisgender youth: M = 1.88, SD = 1.31; F[1, 485] = 4.73, p = .030, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.010$). In contrast, racial/ethnic minority youth reported more advocacy in the GSA (M = 2.79, SD = 0.98) than their White peers (M = 2.56, SD = 0.95; F[1, 485] = 5.86, p = .016, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.012$).

At the group level, advisor self-efficacy in discussing SGM-related topics was moderately related to their reports of the GSA’s open and respectful climate (r = .26, p < .001), greater meeting structure (r = .32, p < .001), and the number of students in the GSA (r = .39, p < .001). Both open and respectful climate and meeting structure of the GSA were moderately associated with the number of students in the GSA (r = .43, p < .001; r = 31, p < .001).

For our primary set of analyses, we report all coefficient estimates, their standard errors, and confidence intervals for the multilevel model predicting the likelihood of discontinued GSA membership in Table 2. We also report odds ratios for individual-level factors. Contrary to our hypothesis, youth who reported higher levels of GSA support at the beginning of the year were not significantly less likely to discontinue their GSA membership before the end of the year (b = −0.550, OR = 0.58, p = .063), although the association did trend in the hypothesized direction. As hypothesized, youth who reported taking on greater leadership responsibility at the beginning of the year were less likely to discontinue their GSA membership before the end of the year (b = −0.518, OR = 0.60, p < .001). Further, youth who initially reported greater engagement in advocacy through their GSA were more likely to discontinue their membership before the end of the year (b = 0.535, OR = 1.71, p = .002). Finally, sexual minority youth were less likely than their heterosexual peers to discontinue their GSA membership (b = −0.971, OR = 0.17, p = .028). Race/ethnicity, gender identity, membership duration, and grade level were not significant predictors of discontinued GSA membership.

Table 2.

Predicting the Likelihood of Discontinued GSA Involvement

Estimated Paths	Coefficient	SE	95% CI	Odds Ratio (OR)	OR 95% CI
Level 1 Factors
GSA Support	−0.550+	0.295	(−1.128, 0.029)	0.577	(0.324, 1.029)
Advocacy	0.535**	0.172	(0.197, 0.873)	1.707	(1.218, 2.394)
Leadership	−0.518***	0.154	(−0.835, −0.200)	0.596	(0.434, 0.819)
Racial Minority	−0.056	0.321	(−0.686, 0.574)	0.946	(0.504, 1.775)
Member Duration	−0.140	0.202	(−0.536, 0.255)	0.869	(0.585, 1.291)
Sexual Minority	−0.971*	0.441	(−1.835, −0.107)	0.167	(0.160, 0.899)
Gender Minority	−0.089	0.458	(−0.987, 0.808)	0.419	(0.373, 2.244)
Grade Level	0.141	0.150	(−0.152, 0.434)	1.152	(0.859, 1.544)
Level 2 Factors
GSA Open Climate	−0.103	0.114	(−0.327, 0.121)
GSA Structure	−0.255**	0.089	(−0.428, −0.081)
Advisor SGM Self-Efficacy	−0.017	0.048	(−0.112, 0.078)
Meeting Frequency	0.662	0.486	(−0.291, 1.616)
Number of Members	0.033	0.032	(−0.029, 0.096)

Note. Values are unstandardized coefficient estimates, their standard errors (SE), and 95% confidence intervals (CI).

^***p < .001

^**p < .01

^*p < .05.

⁺p = 0.06

At the GSA level, as hypothesized, youth who attended GSAs with more meeting structure were less likely to discontinue their membership (b = −0.255, p = .004). The GSA’s open and respectful climate and advisor self-efficacy in discussing SGM issues did not predict their members’ likelihood of discontinuing their membership.

We also conducted sensitivity analyses, mindful that small group sizes could have affected some of the results of the multilevel model. GSAs ranged in size from 6 members to 33 members: 9.4% of youth were in GSAs with fewer than 10 members and 46.7% of youth were in GSAs with fewer than 15 members. However, all patterns remained the same in a model that included only youth in GSAs with more than 10 members (n = 431) and in a model that included only youth in GSAs with more than 15 members (n = 231). The replication of the results with only the larger groups suggested a degree of robustness of the findings.

Additionally, we conducted a follow-up analysis to test whether specific individual items in the measure of GSA meeting structure uniquely predicted discontinued membership. Only one item, “Our GSA meetings follow an agenda,” had a unique, negative predictive effect on discontinued membership (b = −0.731, p = .044). Youth were less likely to discontinue their membership in GSAs that more often followed an agenda (as reported by advisors).

Exploratory Analyses

In an exploratory manner, we tested whether any race/ethnicity-based differences in a youth’s likelihood of discontinued membership might be conditional on the racial/ethnic make-up of their GSA. To do so, we added a cross-level interaction between individual race/ethnicity (level 1) and the percentage of youth in the GSA who identified as a racial/ethnic minority (level 2). The interaction effect was not statistically significant (b = −0.653, p = .179). All other associations remained consistent with our main model.

Discussion

GSAs have a unique potential to benefit their SGM and ally members through their various aims, including opportunities for support and advocacy. Achieving these aims and providing optimal experiences for members, however, likely necessitates some degree of consistent, collective member involvement over time. When this consistency is maintained, GSAs are beneficial to both the school and to GSA members (Toomey et al., 2011). To this end, we identified several individual and GSA characteristics which predicted either reduced or increased likelihood of youth discontinuing their GSA membership over the course of a school year. Members who engaged in leadership activities and identified as sexual minority youth were more likely to continue membership in their GSAs, as were students whose GSAs had more organizational structure to their meetings. In contrast, youth who reported engaging in greater advocacy through their GSA at the beginning of the year were more likely to discontinue their membership. These findings have implications for GSAs and other youth programs in their aims to be maximally inclusive of their members and to engage effectively in a variety of support and advocacy efforts.

This paper provides a few unique contributions to the existing literature. First, we considered GSA involvement in terms of member retention. We conceptualized member retention as important to the GSA as an entity, which, in turn, could be beneficial to the GSA community and to the broader school community. Because it cannot be assumed that factors that promote greater engagement among members are the same as those that retain them, these findings are valuable for the purpose of informing the design and practices of GSAs. Member retention may also benefit even members with low active engagement through a sense of community and access to positive messages and information. Second, we identified specific group-level practices that may support member retention, namely methods for providing structure to the GSA. Finally, we also brought attention to the less-studied, but nevertheless integral, matter of advocacy in the GSA, which we discuss below.

Individual Support from GSA, Leadership, and Advocacy Related to Member Retention

Youth who perceived more social support from peers in the GSA were not statistically less likely to discontinue their GSA membership. It is notable, however, that the association did trend positive, and so this finding should not preclude further consideration of the role of social support within GSAs. Social support is a well-established attribute of successful youth programs and settings based on positive youth development models (Catalano et al., 2004; Holt et al., 2017). Social support is also a major aim of GSAs specifically (Griffin et al., 2004). Future research should consider specifically what forms of social support may be most valuable for youth retention and engagement. In GSAs and similar clubs which aim to bring together youth from different backgrounds, it may be important to take immediate and deliberate steps to cultivate youth’s sense of inclusion in these diverse spaces. Common practices to increase a sense of inclusion might involve sharing names and pronouns at the beginning of meetings, creating space for students to share and validate one another’s experiences, and explicitly voicing support for youth’s multiple intersecting identities (e.g., SGM identities, racial and ethnic identities, different abilities).

The opportunity to take on leadership roles has also been a marker of successful youth programs and settings (Catalano et al., 2004; Lerner et al., 2015). Our current findings further show that youth who took on more leadership responsibilities were less likely to discontinue their GSA membership. Taking on leadership roles may have promoted youth’s psychological empowerment, sense of agency, and sense of ownership and responsibility to the GSA (Poteat et al., 2016; Russell et al., 2009; Wood et al., 2009), thereby motivating them to remain members. Given that many GSAs adopt a youth-led, adult-supported approach, our findings highlight the potential benefits of youth leadership for member retention. Students may benefit from the opportunity to organize a meeting, plan an event, or select the topics for discussion at a meeting. Future research may explore more thoroughly the various leadership opportunities available within GSAs and ways in which to increase opportunities for leadership among members.

We considered youth’s advocacy in relation to discontinued membership in an exploratory fashion and found that youth who initially engaged in more advocacy through the GSA were more likely to discontinue their membership by the year’s end. This may be surprising, as advocacy is associated with a sense of purpose, empowerment, and well-being (Iwasaki, 2016; Poteat et al., 2015; Russell et al., 2009), and therefore, one might expect that these members would be more likely to remain in the GSA. Further, youth in GSAs with more experienced advisors who devoted more time to the GSA engaged in more advocacy (Poteat et al., 2020). In addition, youth’s advocacy was significantly associated with their level of support in the GSA. Given these positive associations, one explanation for these findings is that youth who felt more supported and were engaged in higher levels of advocacy might have been more inclined to extend the work they were doing in their GSAs into other settings outside of the GSA. Indeed, advocacy is an outward-facing action of GSAs (i.e., directed beyond the group to address broader issues in the school or community). However, advocacy can also be challenging and might lead some youth to feel discouraged (Pritzker et al., 2012). Youth whose GSAs perceive greater school hostility also engage in more advocacy (Poteat et al., 2015), potentially out of perceived need. This, too, could account for the predictive relationship between engaging in more advocacy and discontinued membership for some youth. Future research should explore the extent to which either of these processes may underlie these findings, as they carry different implications for how GSAs may support youth who engage in high levels of advocacy.

We also found that sexual minority youth were less likely than their heterosexual peers to discontinue their membership in the GSA before the year’s end. Because GSAs give focus to issues of sexual identity, sexual minority youth may have felt particularly supported around this identity, which they may not have experienced elsewhere in the school (Davis et al., 2014). We did not find differences in retention by gender identity or racial/ethnic identity. Nevertheless, future research should consider how other aspects of gender diverse and racial/ethnic minority youth’s experiences may differ from those of their peers.

Although there were not differences in member retention by gender or racial/ethnic identity, we did identify some differences in how youth engaged in their GSAs by identity. Sexual and gender minority youth reported engaging in less advocacy in the GSA at the beginning of the year than their heterosexual and cisgender peers. Advocacy could be challenging in the GSA context for SGM youth, who may vary in their levels of outness at their school and may have different needs or interests tied to their GSA membership. Some members may seek support from peers in a private, safe setting, whereas others may feel driven to change policies and attitudes in the school at large (Griffin et al., 2004). It may be that some youth with more marginalized identities found it more difficult to engage in outward-facing advocacy efforts in their school because it could risk outing themselves to the broader school community. Additionally, some ally members may have joined their GSAs specifically to engage in advocacy, as they may not seek the same social support as their SGM peers. Further, we found no significant interaction between individuals’ race/ethnicity and the racial/ethnic composition of the GSA in predicting the likelihood of youth discontinuing their membership. This finding could suggest that many GSAs have made efforts to cultivate an intentionally safe and intersectional atmosphere that GSAs often aim to create.

Group Attributes Related to Member Retention

Youth in GSAs with more structure to their meetings were less likely to discontinue their membership over the year. This finding ties to the broader literature on youth settings and programs that highlights the importance of sufficient structure in these spaces (Catalano et al., 2004; Lerner et al., 2015) and findings in the GSA literature that some youth choose not to join GSAs because of perceptions of the group’s disorganization (Heck et al., 2013). GSAs that had set meeting agendas, designated leaders, and check-ins at the start of meetings (which were all indicators of meeting structure in this study) may have provided scaffolding and enhanced opportunities for more youth to participate. This structure could also allow youth to revisit topics or challenges from prior meetings and provide continuity in order for youth to invest in more enduring conversations. We found in our follow-up analysis of individual items comprising our meeting structure measure that having a meeting agenda was specifically a significant unique predictor of member retention.

Contrary to our hypothesis, we did not find an open and respectful climate in the GSA or advisor self-efficacy in discussing SGM issues to be associated with member retention. However, in our preliminary analyses, open respectful climates, advisor SGM self-efficacy, and GSA meeting structure were all significantly correlated with each other. It is possible that the same advisors who felt self-efficacious in discussing SGM issues were also more skilled in fostering an open and respectful climate and scaffolding leadership and social experiences in the GSA through appropriate levels of structure (Catalano, 2004). It is possible, then, that all of these factors work in some combination with one another to cultivate an inclusive, vibrant, and productive setting wherein youth feel drawn to remain, but that meeting structure emerged as having the strongest unique predictive association of member retention in our model. Therefore, future research should continue to consider the role that climate could play in retaining youth in extracurricular groups, particularly those oriented around social justice.

Limitations and Future Directions

We note a few limitations to the current study. First, while we used data collected from two time points to consider predictive associations (youth and advisors at the beginning of the year, and advisors’ reports of member discontinuation at the year’s end), our model was correlational and thus our results cannot speak to the matter of causality. Second, due to the small representation of specific populations of racial/ethnic minority, sexual minority, and gender minority youth, we did not have the statistical power to consider differences in member retention based on youth’s more specific identities or their intersections. Future research should consider such intersectional issues, as youth with multiply marginalized identities may be more likely to experience isolation in their schools or GSAs. As another limitation, the larger project included 38 GSAs, but only advisors in 32 of these GSAs provided information on youth who had discontinued their membership in the GSA. This may have further limited our statistical power for these analyses. We also note that we focused on youth’s initial impressions of support from their GSA and their initial advocacy in the GSA after the first several meetings. Initial impressions and experiences can be powerful predictors of future experiences and behavior (Harris & Garris, 2008), thus informing our decision to focus on them. At the same time, it would have been useful to consider potential changes in youth’s experiences of support and advocacy over the ensuing months and to link these changes to their sustained or discontinued involvement in the GSA. Although we used purposive sampling techniques to enhance the representation of our sample across urban and rural schools, schools varying in racial and ethnic diversity, and schools with students from differing socioeconomic backgrounds, the study was conducted in Massachusetts, whose sociopolitical climate and education system may not be generalizable to the rest of the United States. Finally, although we were able to determine which youth ultimately discontinued their GSA involvement based on advisors’ reports, we were unable to further contextualize the circumstances under which each youth discontinued their involvement, and whether they were generally positive, negative, or mixed in nature.

We also note several strengths of our study. It is one of only a few studies that considers youth’s GSA involvement beyond cross-sectional data. We also included multi-informant data from both youth members and adult advisors, providing a more robust picture of the GSAs. We used purposive sampling to increase the representativeness of GSAs along multiple indicators. Finally, through advisor report, we were able to determine which youth participants had eventually discontinued their GSA involvement before the year’s end.

Our findings also point to several directions for future research. For instance, future research should follow youth who discontinue their GSA involvement to identify the circumstances under which they leave (e.g., to engage with other groups, not feeling welcomed or supported). Similarly, future research might consider how perceptions of group characteristics may vary between youth and the implications this may have for sustained membership. Meeting observations over time also could capture nuance in the group dynamics of GSAs that shape youth’s experiences in them and inform youth’s decisions to increase or decrease their involvement at a given point in time. Relatedly, future research might consider how youth’s GSA membership is associated with their membership in other extracurricular activities and how youth choose the activities to which they devote their time. Future research might also consider more multidimensional scales of youth leadership and advocacy to understand how youth experience different forms of leadership and advocacy. Finally, research should focus on additional individual and group factors that may enhance member retention in GSAs and similar clubs whose members come from diverse backgrounds and have varying needs and interests.

Continued and increased attention not only to why youth join GSAs but also to why they remain involved has the potential to provide vital information to GSAs on how they might best meet their members’ needs and what range of programming or services they should offer to their members. This work is of great importance for GSAs and other social justice-oriented youth settings, as these spaces provide necessary opportunities and resources to marginalized youth.

Table 3.

Specific Meeting Structure Items Predicting Discontinued Membership

Estimated Paths	Coefficient	SE	95% CI
GSA Structure
Check-ins at beginning of meetings	−0.287	0.183	(−0.645, 0.071)
Follow-up on previous discussions	0.016	0.473	(−0.912, 0.943)
Meeting agenda	−0.731*	0.363	(−1.442, −0.020)
Student or advisor leads meetings	−0.213	0.447	(−1.088, 0.663)

Note. Values are unstandardized coefficient estimates, their standard errors (SE), and 95% confidence intervals (CI). All variables at level 1 were included in this model, and no pattern or significance changed from model reported in Table 2.

^*p < .05

Significance Statement:

At the beginning of the year, youth who had more leadership roles, engaged in less advocacy, and were in GSAs with greater structure were less likely to discontinue their membership by the year’s end. Member retention is important to GSAs in their continual efforts to support SGM youth and to promote social justice in schools. Understanding factors that underlie youth’s sustained membership could signal important issues to which GSAs should attend in their efforts to cultivate and sustain their membership.