A Brief Group Social-Belonging Intervention to Improve Mental-Health and Academic Outcomes in BIPOC and First-Generation-to-College Students

Abstract

Despite greater emphasis on diversity and inclusion on college and university campuses, inequities persist. Awareness of structural and social threats to success can lead students from underrepresented identities to question whether they will fully belong at a given institution, which jeopardizes their psychological well-being and academic performance. This study tested a brief social-belonging intervention, delivered in a group format, that emphasized that first-year challenges are normative and that, over time, students develop relationships that deepen their sense of belonging. Participants (N = 122) who reported poorer belonging at baseline experienced greater depressive symptoms, greater worry, and worse psychological well-being over the 14-month follow-up period. The intervention significantly reduced risk for major depression during the first 2 years of college and specifically reduced risk for participants experiencing more discrimination. Hypotheses that the intervention would improve psychosocial or academic outcomes specifically for Black, Indigenous, and people of color and first-generation-to-college students were not supported.

Keywords

belonging intervention depression college mental health discrimination

Despite greater emphasis on diversity and inclusion on college and university campuses, inequities persist for matriculated Black, Indigenous, and people of color (BIPOC) and first-generation-to-college students (i.e., individuals who are the first generation in their immediate family to attend a 4-year college or university). These students may experience greater social stress, greater sense of isolation within the campus community, and poorer academic outcomes compared with continuing-generation White students (Espenshade & Radford, 2009; Nunn, 2021; Steele, 1997). BIPOC and first-generation students are often numerically underrepresented at selective colleges and universities and may encounter negative assumptions about their academic preparation or abilities in addition to other microaggressions (Espenshade & Radford, 2009; Harwood et al., 2012; Steele, 1997). Awareness of these structural and social threats to success can lead students to question whether they will fully belong at a given institution and whether initial difficulties during the transition to college predict stable challenges over the next 4 years (Walton & Cohen, 2007, 2011). Offidani-Bertrand and colleagues (2022) reported that BIPOC students transitioning from racially diverse high schools to predominantly White collegiate institutions may particularly be vulnerable to questioning whether they belong at the college and then disengaging from social or academic life.

Questions and concerns about belonging, termed “belonging uncertainty” by Walton and Cohen (2007), can have a potent impact on well-being and academic performance. For example, Yeager et al. (2016) found that belonging uncertainty uniquely predicted first-year dropout even when accounting for high school grade point average (GPA) and SAT scores. As Brady and colleagues (2020) noted, belonging uncertainty may be rooted in an awareness of exclusion and stereotyping, but these worries also subtly influence perception, stress appraisals, and behaviors that contribute to racial disparities in academic achievement and well-being. At higher levels of belonging uncertainty, one may perceive commonplace collegiate events such as receiving a poor grade on an exam or feeling out of place at an event as evidence that one does not belong at the institution (Murphy et al., 2020). These interpretations and associated low mood promote academic and/or social disengagement, which constrain life satisfaction, mental health, and GPA.

Fortunately, several brief social-belonging interventions have successfully challenged the assumption that initial difficulties with belonging portend long-term alienation. These interventions use statements from older college students that normalize academic and social challenges and emphasize how one’s sense of belonging expands over time. Walton and Cohen (2007) framed belonging uncertainty as a hypothesis, rather than a fixed belief, that guides interpretation; thus, belonging uncertainty is malleable in the face of contradictory evidence. Students who become more optimistic about their belonging may then engage more in ways that reinforce belongingness by broadening and strengthening relationships with peers and faculty.

Social-Belonging Interventions

Walton and Cohen (2007, 2011) designed an innovative one-session intervention to dispel the belief that only certain students experience difficulty with the transition to college. Black and White participants were presented with quantitative and qualitative evidence that social adversity is common and transient and that adversity may be due to the challenge of transitioning to a new social environment rather than to fixed qualities of the student. Participants then became advocates of that message by writing about their own experience as an example of this growth process. The intervention specifically benefited Black students: By students’ senior year, a racial academic-achievement gap was cut by 79%, and the gap in subjective happiness disappeared. In a long-term follow-up 7 to 11 years after the one-session intervention, Black students from the intervention condition reported greater life satisfaction than participants in the control condition (Brady et al., 2020). The authors asserted that this brief intervention “lifted a psychological obstacle—persistent group-based worry about belonging, rooted in awareness of social disadvantage—to allow students to develop, on their own, authentic relationships of significance that, in many cases, lasted well past college graduation” (Brady et al., 2020).

Yeager et al. (2016) extended this research to online preparatory interventions delivered before students began their first year of college. The interventions again imparted the message that many students at first feel that they do not belong in college but find their comfort over time. Across three experiments, socially disadvantaged students who received the intervention had greater college attendance and persistence rates and higher first-year GPA compared with control participants. Socially advantaged participants were not helped or harmed by the interventions.

Like BIPOC students, first-generation-to-college students experience group-based belonging uncertainty, being more prone than continuing-generation students to feel excluded and to have trouble finding their place on campus, particularly within selective institutions (Nunn, 2021; Ostrove & Long, 2007). Continuing-generation students are more likely to experience social belonging early within the first year and consistently (Nunn, 2021). To address identity-based belonging uncertainty and the related social-class achievement gap, Stephens et al. (2014) designed a brief intervention that employed a difference-education framework that aimed to destigmatize difference rather than destigmatizing challenge, as was intended in the social-belonging interventions. Participants attended a panel discussion that emphasized, again through students’ personal stories, how difference matters and can be used to overcome background-specific obstacles to success. Although the authors found an end-of-year gap of 0.30 grade points between first-generation and continuing-generation students in the control condition, grades between first-generation and continuing-generation students in the difference-education condition did not differ significantly. An online adaptation of the difference-education intervention similarly closed the generation-status gap; first-generation students who received the intervention had higher cumulative GPA than first-generation control participants at the end of their second year of college (Townsend et al., 2019). These findings further underscore the promise of brief interventions that convey that students from different backgrounds belong at and can thrive at a given college.

Although previous social-belonging interventions have focused on academic achievement and general psychological well-being, the transition to college also coincides with increasing vulnerability to clinical depression. At least 20% of young adults have experienced major depression by age 18, and there is further risk for new onset and recurrences over the college years (Rohde et al., 2013). The notable incidence rate of depression during emerging adulthood makes prevention efforts an important research aim. Belonging uncertainty is associated with poorer mental health and depression more specifically (Dutcher et al., 2022; Gopalan & Brady, 2020). Freire and Hurd (2023) found that among minoritized students at a predominantly White institution, discrimination experiences lowered sense of belonging, which later predicted higher levels of depressive symptoms. Across three studies, Dutcher et al. (2022) found that lower daily sense of belonging predicted higher end-of-term depression as much as 4 months in advance. These findings held when controlling for loneliness and social interactions, suggesting the unique importance of sense of belonging to mental health. Dutcher and colleagues suggested that universities interested in reducing student depression should offer early interventions to enhance feelings of belonging.

The previous social-belonging interventions applied psychological principles to remove cognitive barriers toward academic success and life satisfaction. We expect that reducing concerns about belonging could prevent negative mood spirals toward clinical depression. One prior study found that an individually delivered social-belonging intervention reduced fluctuations in depressive symptoms over 1 semester (Marksteiner et al., 2019). However, depressive symptoms were assessed through only a brief self-report measure. To our knowledge, this is the first social-belonging-intervention investigation to examine mental health via a “gold-standard” diagnostic interview as a primary outcome.

Intervention Design

The social-belonging-intervention strategies effectively build on robust psychological science: They counter pluralistic ignorance that most students struggle in the transition to college; they challenge fixed mindsets, emphasizing instead that one’s sense of belonging remains malleable throughout college; and they exploit the “saying is believing effect” to internalize the intervention’s message (Aronson et al., 2002; Higgins & Rholes, 1978). Student success may be held back by continuous concerns about belonging. Activating a process that reduces these daily concerns could substantially reduce inequality in outcomes (Yeager & Walton, 2011). Walton and Cohen (2011) argued that their brief intervention begins to alter how participants interpret daily social adversity. Initial change starts a virtuous, recursive process, with socially disadvantaged students becoming less sensitive to daily interactions as a signal of persistent, poor fit within the community.

In the present study, we capitalized on previous mixed-methods research at the study site to develop a brief group intervention. Rather than delivering the intervention individually as in prior investigations, we tested a group format, led by older peers, that could be implemented more efficiently as institutional practice. In addition, discussing the intervention’s themes with fellow student-participants could more directly address pluralistic ignorance that it is quite common to struggle with transitions and to question belonging (Binning et al., 2020). Similar to the previous brief interventions, psychosocial outcomes of loneliness, happiness, and life satisfaction were evaluated, as was GPA as an academic outcome. Depression and clinical worry were also assessed by diagnostic interview and self-report measures.

Hypotheses

Our primary interests were the mental health and psychological well-being of students who face social disadvantages in higher education, defined as BIPOC students and first-generation students of any racial identities. We hypothesized that socially disadvantaged students assigned to the group intervention would experience fewer depressive episodes over the follow-up period than socially disadvantaged students in the control condition. We also hypothesized that socially disadvantaged students assigned to the group intervention would have better mental-health outcomes (less depression and less worry) 8 and 14 months after the intervention than socially disadvantaged students in the control condition.

We hypothesized that socially disadvantaged students assigned to the group intervention would have better psychosocial outcomes (lower belonging uncertainty, lower loneliness, greater happiness, and greater satisfaction with life) 8 and 14 months after the intervention than socially disadvantaged students in the control condition. We hypothesized that socially disadvantaged students assigned to the group intervention would have higher GPA in the fall and spring semesters of their sophomore year than socially disadvantaged students in the control condition. Continuing-generation White students were expected to benefit less from the intervention because these identities afford more of a presumption of belonging on a college campus. We did not expect significant differences on any of these outcomes between socially advantaged students in the intervention versus control conditions.

Race and first-generation status are key demographic variables that can affect sense of belonging at a predominantly White institution, but they are far from the only identity factors that affect belonging. To be inclusive of all who may question their belonging, we also examined the above hypotheses with baseline belonging uncertainty substituted for identity group in the analyses. We similarly hypothesized that participants who experienced greater uncertainty about their belonging and received the intervention would have better mental-health outcomes, better psychosocial outcomes, and higher GPA than participants in the control condition who experienced greater belonging uncertainty. Notable positive outcomes in this project would encourage colleges to implement similar group interventions that normalize initial social adversity during the transition to college.

Transparency and Openness

Our hypotheses were preregistered on OSF following data collection but before we scored data or conducted any analyses: https://osf.io/zr97y. Intervention materials, including the protocol for group leaders and the discussion handout, are available at https://osf.io/8up67. Self-report measure instructions and response anchors are available at https://osf.io/8up67; full texts of the self-report measures were not posted because of copyright restrictions. We report how we determined our sample size, all data exclusions, and all manipulations. All study procedures were approved by the college’s institutional review board (2017-107).

Method

Participants

Participants were 122 undergraduate students (29.5% men, 0.8% nonbinary/fluid/genderqueer, 0.8% not reported, 68.9% women) at a liberal arts college in the northeastern United States. The only eligibility criterion was being a first-year student at the college. Age at enrollment ranged from 18 to 21 (M = 18.48 years, SD = 0.58). Participants reported the following racial identities: Asian (21.3%), Black (9.8%), Latinx/Hispanic (9.0%), Middle Eastern/North African (0.8%), Native American/Alaska Native (0.8%), Pacific Islander/Native Hawaiian (0.8%), multiracial (4.1%), and White (53.3%); 22.1% identified as first-generation-to-college students, defined as being among the first generation in their family to attend a 4-year college or university. Thirteen participants (10.7%) were international students. Sixty-two participants were BIPOC and/or first-generation students, classified in this study as “socially disadvantaged” because these identity groups face social and structural disadvantages in higher education. Sixty participants were White continuing-generation-to-college students, classified as “socially advantaged.”

The target sample size was 120 comprising 60 first-generation-to-college students and/or BIPOC students and 60 continuing-generation and White students. This sample size was consistent with the primary motivating study (N = 92; Walton & Cohen, 2011) and was also the maximum number feasible for the research team to enroll within each 2-month recruitment window. This target sample size could be limited in power to detect moderator interactions with intervention effects, but recruitment was limited by feasibility for the research team and the context.

Email addresses for all first-year students were provided by the college’s dean of studies, and a random sample, stratified by identity group, was invited by email to join the study. Rather than describing the intervention condition as such, participants were informed that a subset of the sample would attend an additional focus-group session. This framing avoids the stigma of being identified as in need of intervention by asking participants for their help rather than offering them help (Walton & Cohen, 2011). Participants first expressed interest by replying to the invitation, a follow-up description was provided, and if they remained interested, the first baseline session was scheduled. Overall, 13.6% of invited students enrolled in the study. BIPOC and first-generation students were significantly more likely to accept the invitation (19.6%) compared with continuing-generation White students (10.4%), χ²(1) = 14.88, p < .001. Random assignment and retention are summarized in Figure 1.

Fig. 1.

Study CONSORT diagram.

Participants were reimbursed $30 for the baseline sessions, whether or not they attended the intervention session. Alternatively, participants could earn two psychology research credits and $10.¹ Participants were reimbursed $20 or 1.5 psychology research credits for each follow-up session.

Procedure

All study procedures were approved by the college’s institutional review board. Participants were recruited in two cohorts, from January to March 2018 and 2019. At Session 1, participants provided their informed consent and then completed each self-report measure through the Qualtrics data-collection website. Participants also provided demographic information, adapted from Suyemoto et al. (as cited in Wadsworth et al., 2016). After Session 1, participants were randomly assigned to the intervention or control condition; random assignment was stratified by first-generation versus continuing-generation status, racial identity (BIPOC vs. White student), and gender. Approximately 1 week after Session 1, participants returned for a semistructured diagnostic interview. The Sunday evening following Session 2, participants assigned to the intervention condition attended the 90-min session. Participants assigned to the control condition completed the same assessment schedule but did not attend a group session. Using a naturalistic control directly tested the benefits of this brief intervention compared with campus life as usual.

Participants returned for two follow-up assessments approximately 8 months and 14 months after the baseline sessions (September/October and March/April of their sophomore year). These Time 2 and Time 3 follow-up sessions began with participants completing the same self-report outcome measures. Participants then completed a diagnostic interview to assess for depressive episodes and anxiety disorders over the follow-up period. The follow-up interviewer was blind to participants’ experimental condition. Follow-up data collection ended in May 2020. The coronavirus disease 2019 (COVID-19) pandemic emerged as final follow-up assessments were occurring with the second cohort. Of the participants who completed the Time 3 follow-up session, 32% (n = 33) provided final data after the COVID-19 pandemic affected the study site. All assessment procedures remained the same, except diagnostic interviews were conducted via a telehealth website.

Intervention condition

The social-belonging intervention was closely adapted from Walton and Cohen’s (2007, 2011; Walton et al., 2017) model. However, rather than delivering the intervention individually, this project tested a group format that facilitated discussion between participants and could be implemented more efficiently as institutional practice to reduce disparities. The intervention aimed to communicate three themes: (a) It is normal to worry whether you belong at a new school and to face challenges in the transition, (b) experiencing challenges does not mean that you do not belong, and (c) with time and effort, you can develop a strong sense of belonging at the college. As Yeager and Walton (2011) noted, for a wise intervention to succeed, theoretical expertise must be combined with contextual expertise; every intervention must be customized to its specific context. The group-intervention content was constructed around qualitative and quantitative data collected in a previous college-specific study of social belonging (N = 55). Previous participants wrote about difficulties they experienced in their transition to college and how their sense of belonging improved with time. Eleven of these writing samples were selected and edited to include in the intervention discussion. The writers’ class year and hometown were altered to protect privacy, emphasize diverse backgrounds, and correspond with the regional accents of student volunteers who audio recorded the edited statements.

Two undergraduate (peer) research assistants led the group discussion for three to six participants per week.² After an introductory discussion about college life, the intervention participants reviewed graphs depicting how social belonging and social well-being improve over time. Next, the group leaders facilitated a discussion of factors that promote and prevent social belonging, structured around the 11 audio-recorded statements. Group leaders emphasized that the recordings represented a diverse group of students. All of the recorded statements emphasized that first-year challenges were transient and social belonging improved over time. For example, one statement read,

Feeling the pressures to fit into a social group rather than being a floater made it harder at the beginning when all the freshmen were grouping off. I wanted to be in so many groups at once but felt like I was being forced to choose just one. I got involved in more clubs which definitely helped my transition because they made me feel like I belong, but I was not forced to hang out with them all the time.

Finally, to exploit the saying-is-believing effect and internalize the intervention’s themes (Higgins & Rholes, 1978), participants wrote a brief essay that discussed how sense of belonging improves, drawing from their own experiences. Participants were given the option to type, audio record, or video record the final version of their essay, with the explanation that their comments might be used in future materials to offer incoming students accurate expectations of the transition into college. The group intervention session lasted approximately 1.5 hr.³

Measures

Diagnostic interview

The Structured Clinical Interview for DSM-5 Research Version (SCID) is a commonly used semistructured interview that assesses current and lifetime diagnoses of psychological disorders (First et al., 2015). The SCID was administered at baseline to assess for current major depression, bipolar I disorder, bipolar II disorder, panic disorder, agoraphobia, social anxiety disorder, and generalized anxiety disorder and previous major depression during high school. Follow-up SCIDs assessed for new major depressive episodes (MDEs) over the follow-up period and the course of panic disorder, social anxiety disorder, and generalized anxiety disorder. In addition to MDEs, the SCID was used to assess for depressive episodes with insufficient symptoms, referred to below as “subclinical depressive episodes,” in which having two to four MDE symptoms were present and were associated with clinically significant distress or impairment. Diagnostic interviews were completed by PhD- and LCSW-level researchers. Twenty percent of baseline and follow-up interviews were coded for interrater reliability and indicated strong agreement (baseline MDE: κ = .91; subclinical depressive episodes: κ = .83; follow-up MDE: κ = .81; subclinical depressive episodes: κ = .86).

Self-report measures

Depression severity

The Beck Depression Inventory–II (BDI-II; Beck et al., 1996) is a widely used 21-item measure of the severity of depression symptoms over the previous 2 weeks. Items (e.g., sadness) are rated by selecting one of four statements of increasing severity, and items are scored from 0 to 3. The scale had excellent internal consistency in this study (baseline: Cronbach’s α = .91; Time 2: Cronbach’s α = .92; Time 3: Cronbach’s α = .92) and has demonstrated robust construct validity (Erford et al., 2016).

Worry

The Penn State Worry Questionnaire (Meyer et al., 1990) is a widely used 16-item measure of the frequency and intensity of worry. Participants rated statements (e.g., “I know I should not worry about things, but I just cannot help it”) on a 5-point scale from 1 (not at all typical of me) to 5 (very typical of me). The scale had excellent internal consistency in this sample (baseline: Cronbach’s α = .94; Time 2: Cronbach’s α = .94; Time 3: Cronbach’s α = .95) and has strong convergent and discriminant validity (Brown et al., 1992; Meyer et al., 1990).

Belonging uncertainty

Participants’ level of uncertainty about their social belonging at the college was assessed using a two-item scale (Walton & Cohen, 2007). Each item (“Sometimes I feel that I belong at [college name], and sometimes I feel that I don’t belong at [college name]” and “When something bad happens, I feel that maybe I don’t belong at [college name]”) was rated on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). The scale had acceptable internal consistency in this sample (baseline: Spearman-Brown coefficient = 0.69; Time 2: Spearman-Brown coefficient = 0.80; Time 3: Spearman-Brown coefficient = 0.76).

Loneliness

The UCLA Loneliness Scale, Version 3 (Russell, 1996) is a 20-item measure of participants’ sense of loneliness. Participants rated how often they felt an aspect of loneliness (e.g., “How often do you feel that there is no one you can turn to?”) on a 4-point scale ranging from 1 (never) to 4 (always). The scale had excellent internal consistency in this sample (baseline: Cronbach’s α = .93; Time 2: Cronbach’s α = .93; Time 3: Cronbach’s α = .94). Previous research has demonstrated evidence for its discriminant and construct validity (Russell, 1996).

Happiness

The Subjective Happiness Scale (Lyubomirsky & Lepper, 1999) is a four-item measure of dispositional happiness. Participants rated statements on a 7-point scale in which the highest anchors indicate greater happiness (e.g., 1 = a very unhappy person, 7 = a very happy person). The scale had excellent internal consistency in this sample (baseline: Cronbach’s α = .91; Time 2: Cronbach’s α = .91; Time 3: Cronbach’s α = .91). Construct-validation studies have demonstrated its convergent and discriminant validity (Lyubomirsky & Lepper, 1999).

Life satisfaction

The Satisfaction With Life Scale (Diener et al., 1985) is a five-item measure of global life satisfaction. Participants rated statements (e.g., “So far I have gotten the important things I want in life”) on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). The scale had good internal consistency in this sample (baseline: Cronbach’s α = .87; Time 2: Cronbach’s α = .87; Time 3: Cronbach’s α = .83). Prior studies established the validity of the scale (Diener et al., 1985).

Everyday discrimination

The Everyday Discrimination Scale (Williams et al., 1997) is a 10-item measure of experiences with discrimination in participants’ day-to-day life. Participants rated the frequency with which they experienced discriminatory behaviors (e.g., “You are treated with less respect than other people are”) on a 6-point scale ranging from 0 (never) to 5 (almost every day). The measure’s instructions do not direct respondents to consider a specific context or time frame, only their “day-to-day life.” Thus, participants’ responses were likely informed by their experiences on campus over the previous 5 to 7 months but could have also reflected interactions in the surrounding community, online, or elsewhere. The scale had good internal consistency in this sample (baseline: Cronbach’s α = .85). The measure has been validated among diverse racial and cultural groups in the United States (Kim et al., 2014). This measure was not listed in the study’s preregistration and was included for post hoc exploratory analyses.

GPA

All participants consented for the college registrar to provide academic transcripts to the principal investigator. Noncumulative semester GPA was measured on a 4.0 scale, where A+ = 4.3, A = 4.0, A– = 3.7, B+ = 3.3, and so on. The COVID-19 pandemic affected the final GPA time point (sophomore year spring GPA) for the second cohort of participants (n = 59). After most students returned home and all classes shifted to remote learning in March 2020, the college allowed students to take any course as satisfactory/unsatisfactory, which would not contribute to the semester GPA.

Preregistered data-analysis plan

The analysis plan was preregistered on OSF following data collection but before scoring data or conducting any analyses. Results were analyzed on an intent-to-treat basis; thus, group participants were included whether or not they attended the intervention group.⁴ The effect of the intervention on new depressive episodes was evaluated through logistic regression in IBM SPSS (Version 25). Condition and identity group (socially advantaged vs. socially disadvantaged) and their interaction were examined as (a) predictors of having a new MDE over the 14-month follow-up period and (b) predictors of having a new major or subclinical depressive episode over the 14-month follow-up period. The effect of the intervention on mental health was also evaluated using 2 × 2 × 3 mixed-model analyses of variance (ANOVAs). Condition and identity group were the between-subjects independent variables; time (before intervention, 8-month follow-up, and 14-month follow-up) was the within-subjects variable. Self-report data on current depressive symptoms and worry were examined as dependent variables.

The effect of the intervention on psychosocial outcomes was evaluated using 2 × 2 × 3 mixed-model ANOVAs. Condition and identity group were the between-subjects independent variables; time (before intervention, 8-month follow-up, and 14-month follow-up) was the within-subjects variable. Self-report data on belonging uncertainty, loneliness, subjective happiness, and life satisfaction were examined as dependent variables. The effect of the intervention on noncumulative GPA was evaluated using a 2 × 2 × 4 mixed-model ANOVA. Condition and identity group were the between-subjects independent variables; time (first-year fall, first-year spring, sophomore fall, and sophomore spring) was the within-subjects variable.

Race and first-generation status, which defined the socially advantaged versus socially disadvantaged variable, can affect sense of belonging at a predominantly White institution, but they are not the only identity factors that affect belonging. To be inclusive of all who may question their belonging, the analyses described above were also conducted with baseline belonging uncertainty substituted for identity group. Linear mixed-effects modeling examined condition, centered belonging uncertainty, and their interaction as predictors of self-report mental-health, psychosocial, and academic outcomes. These findings did not differ meaningfully from the mixed-model ANOVA results and therefore are provided in the Supplemental Material available online.

The preregistered analysis plan stated that all analyses would include gender and current therapy at the time of the intervention as covariates and baseline variables that were not equally distributed via random assignment. Gender was coded as 0 = female, 1 = other gender identities. Therapy at baseline was coded as 1, no therapy = 0.

Post hoc analytic adjustment because of COVID-19

The COVID-19 pandemic affected the final GPA time point for the second cohort of participants; there was a significant difference between the first and second cohorts’ sophomore spring GPA, t = 5.39, p < .001, d = 1.00. Therefore, cohort (0 = enrolled in 2018, 1 = enrolled in 2019) was included as a covariate in GPA analyses. This was a deviation from the study’s preregistration. Analyses including only the first three GPA time points were not meaningfully different (see the Supplemental Material).

Results

Because of outliers, a 95% winsorization was performed on the BDI-II at baseline and first-year fall GPA, first-year spring GPA, sophomore-year spring GPA, and age at Time 2. Raw means of the self-report measures, unadjusted for covariates, are presented in Table 1. All dependent variables were screened for skewness and kurtosis. Square-root transformations were used to address positive skewness in current depressive symptoms and negative skewness in life-satisfaction scores. Log transformations addressed substantial negative skewness in GPA each semester.⁵ Correlations for the baseline self-report measures are presented in Table 2.

Table 1.

Raw Means and Standard Deviations of Scores on Self-Report Measures

	Group intervention		Control
Assessment time point	Socially disadvantagedn = 32	Socially advantagedn = 29	Socially disadvantagedn = 30	Socially advantagedn = 31
Beck Depression Inventory
Baseline^a	10.84 (8.00)	10.07 (9.22)	8.80 (7.59)	7.68 (5.94)
Time 2^a	8.89 (7.51)	6.57 (6.69)	9.15 (8.56)	8.75 (7.68)
Time 3	10.07 (8.26)	8.48 (8.30)	10.80 (10.67)	9.96 (7.75)
Penn State Worry Questionnaire
Baseline	51.53 (14.07)	53.59 (15.34)	53.53 (13.46)	53.74 (14.73)
Time 2	46.75 (14.41)	49.91 (16.06)	55.33 (12.29)	54.07 (14.11)
Time 3	48.48 (13.33)	51.68 (15.72)	53.68 (14.25)	53.50 (15.10)
Belonging uncertainty
Baseline	4.75 (1.40)	4.83 (1.53)	4.67 (1.56)	3.77 (1.54)
Time 2	4.29 (1.47)	4.46 (1.40)	4.26 (1.79)	3.80 (1.54)
Time 3	4.39 (1.44)	4.30 (1.53)	4.20 (1.76)	4.10 (1.56)
UCLA Loneliness Scale
Baseline	44.06 (9.45)	44.41 (10.73)	43.70 (9.04)	40.65 (9.26)
Time 2	42.43 (8.89)	42.61 (11.33)	44.19 (10.21)	38.64 (8.08)
Time 3	44.30 (7.24)	41.75 (11.53)	43.64 (10.78)	40.19 (9.80)
Subjective Happiness Scale
Baseline	4.55 (1.56)	4.27 (1.59)	4.72 (1.35)	4.90 (1.32)
Time 2	4.73 (1.44)	4.47 (1.60)	4.73 (1.39)	5.20 (1.16)
Time 3	4.79 (1.37)	4.41 (1.47)	4.66 (1.28)	4.92 (1.40)
Satisfaction With Life Scale
Baseline	22.59 (6.47)	25.79 (5.83)	23.60 (7.71)	26.97 (5.09)
Time 2	23.93 (6.33)	24.96 (7.18)	23.52 (7.31)	27.43 (5.63)
Time 3	23.56 (4.97)	26.24 (6.12)	23.36 (6.95)	26.31 (5.32)
Everyday Discrimination Scale
Baseline	13.56 (9.33)	12.07 (6.97)	11.07 (6.51)	7.97 (4.80)

Note: Socially disadvantaged = Black, Indigenous, and people of color and/or first-generation-to-college students; socially advantaged = White continuing-generation students; Time 2 = 8-month follow-up; Time 3 = 14-month follow-up.

A 95% winsorization was performed to reduce the influence of outliers.

Table 2.

Pearson Correlation Coefficients for Baseline Self-Report Measures

Variable	1	2	3	4	5	6
1. Beck Depression Inventory^a,b	—
2. Penn State Worry Questionnaire	.54***	—
3. Belonging uncertainty	.42***	.36***	—
4. UCLA Loneliness Scale	.58***	.38***	.44***	—
5. Subjective Happiness Scale	–.61***	–.43***	–.41***	–.67***	—
6. Satisfaction With Life Scale^b	–.60***	–.30**	–.36***	–.66***	.53***	—
7. Everyday Discrimination Scale	.37***	.11	.29**	.28**	–.27**	–.37***

A 95% winsorization was performed to reduce the influence of outliers.

Square-root transformation was used to address skewness.

p < .01. ***p < .001.

Baseline group differences and retention over follow-up

Twenty-three participants (18.9%) reported current participation in therapy at baseline. Current therapy did not differ between conditions, χ²(1) = 0.48, p = .487. Participants randomly assigned to the intervention condition were significantly older, on average, than control participants, t(120) = 2.02, p = .045. Following the preregistered analysis plan, age was included as a covariate in all analyses. There were no significant differences between the intervention and control groups on baseline depressive symptoms, worry, loneliness, subjective happiness, satisfaction with life, or first-year fall GPA (all ps > .060). Participants randomly assigned to the intervention reported greater uncertainty in their belonging at baseline, t(120) = 2.08, p = .040, compared with control participants. Following the preregistered analysis plan, centered baseline belonging uncertainty was included as a covariate in all analyses unless belonging uncertainty was the dependent variable.

Of the 61 participants assigned to the intervention, 52 (85.2%) attended the group session. One hundred and six participants (86.9%) completed the 8-month follow-up assessment, and 103 participants (84.4%) completed the final, 14-month assessment. Study attrition did not differ between conditions, χ²(1) = 0.06, p = .803, or between identity groups, χ²(1) = 0.03, p = .863. Study attrition was not predicted by gender, first-year fall GPA, or baseline worry, belonging uncertainty, or subjective happiness (all ps > .067). Participants who did not complete the follow-up period were more likely to be in therapy at baseline, χ²(1) = 4.76, p = .029, and to report greater depressive symptoms, χ²(1) = 5.57, p = .018; greater loneliness, χ²(1) = 4.19, p = .041; and less life satisfaction, χ²(1) = 6.02, p = .014, at baseline.

Depressive episodes

Two participants met criteria for a current MDE at baseline, and 30 other participants met criteria for a past MDE during high school. Twenty-four participants (19.7%) reported a new MDE over the 14-month follow-up period. Logistic regression was performed to examine the effect of the intervention, identity group, and their interaction on the likelihood that participants would experience a new MDE. To be consistent with the repeated measures analyses, baseline depressive symptoms was included as a covariate in addition to gender, concurrent therapy at baseline, age, and baseline belonging uncertainty.⁶ The logistic regression model was statistically significant, χ²(8) = 24.65, p = .002. Baseline depressive symptoms was the only significant covariate, χ²(1) = 9.94, p = .002, and greater symptoms predicted future depression (all other ps > .321). The Condition × Identity Group interaction was not significant, χ²(1) = 0.37, p = .542. Condition and identity group were not statistically significant predictors in the first model; condition: χ²(1) = 3.59, p = .058; identity group: χ²(1) = 1.04, p = .308. However, when the nonsignificant interaction term was removed in a post hoc analysis, condition was a unique predictor of MDEs, χ²(1) = 4.62, p = .032 (see Fig. 2). Eight participants (13.1%) randomly assigned to the intervention condition reported a new depressive episode compared with 16 participants (26.2%) assigned to the control condition. The odds ratio, adjusted for the covariates, of 0.28 (95% confidence interval [CI] = [0.09, 0.89]) indicated that intervention participants had a 72% lower likelihood of experiencing major depression over the follow-up period.

Fig. 2.

Percentage of participants reporting new depressive episodes over the 14-month follow-up period. N = 105 regarding new major depressive episodes, and N = 107 regarding new major or subclinical depressive episodes. Two participants reported a new major depressive episode and two others reported a new subclinical depressive episode at the 8-month follow-up assessment before dropping out of follow-up.

An additional 13 participants (10.7%) reported a new subclinical depressive episode without meeting criteria for an MDE over the follow-up period. The logistic regression model was significant when predicting having an MDE or subclinical depressive episode, χ²(8) = 20.67, p = .008. The Condition × Identity Group interaction was not significant, χ²(1) = 0.23, p = .632. Condition and identity group were not statistically significant predictors; condition: χ²(1) = 2.37, p = .124; identity group: χ²(1) = 0.60, p = .440. When the nonsignificant interaction term was removed, condition remained nonsignificant as a predictor of any depressive episode, χ²(1) = 2.96, p = .085 (see Fig. 2). However, the odds ratio, adjusted for the covariates, of 0.43 (95% CI = [0.17, 1.12]) indicated that intervention participants had a 57% lower likelihood of experiencing either an MDE or subclinical depressive episode by the end of their sophomore year.

Because racial-identity categories and first-generation status do not encompass all who question their belonging in college, baseline belonging uncertainty was examined in exploratory analyses as a possible moderator of the intervention’s impact on depression using the PROCESS macro for IBM SPSS (Version 3.5.3; Hayes, 2018), PROCESS Model 1.⁷ Baseline belonging uncertainty was not a significant moderator of the intervention effect on MDE, χ²(1) = 0.13, p = .717, or any new depression, χ²(1) = 0.08, p = .784. However, this may have been complicated by the fact that participants randomly assigned to the intervention by chance reported greater uncertainty in their belonging at baseline. To extend the question, we considered if another, related variable could be a possible moderator of the intervention’s effect. “Everyday” discrimination experiences were conceptualized as a more specific type of threat to belonging and thus were examined in post hoc exploratory analyses. At baseline, the difference between socially disadvantaged and advantaged participants in everyday discrimination experiences was not significant, t(120) = 1.83, p = .070.

Baseline reports of discrimination experiences was not a significant moderator of the intervention effect on major depression, χ²(1) = 3.83, p = .050. However, the Johnson-Neyman technique was used to probe the interaction and identify the region(s) of significance. The intervention effect transitioned to significant at −0.63 below the mean, indicating that the intervention significantly reduced risk for MDE in participants just below the mean and for all above the mean on everyday discrimination experiences (46.7% of the sample). Discrimination scores did significantly moderate the intervention effect on any new depression (MDE or subclinical depression), χ²(1) = 4.34, p = .037. The Johnson-Neyman technique indicated that the intervention effect transitioned to significant at 0.45 above the mean for everyday discrimination experiences; the intervention reduced risk for MDE or subclinical depression in participants above this value, participants with greater discrimination scores (43.0% of the sample).

Within-subjects effects in self-reported mental-health and psychosocial outcomes

Overall, self-report mental-health outcomes did not change over time, the intervention condition did not affect their change over time, and these outcomes did not change over time differently between identity groups. Contrary to our hypotheses, the mixed-model ANOVAs did not reveal statistically significant Time × Condition × Identity Group interactions on current depressive symptoms, F(2, 184) = 0.003, p = .997, η_p² < .001, or worry, F(2, 184) = 0.17, p = .847, η_p² = .002. Depressive symptoms did not significantly change between assessment points, F(2, 184) = 2.42, p = .091, η_p² = .03, nor did worry, F(2, 184) = 0.88, p = .416, η_p² = .009. Change in current depressive symptoms was not affected by condition, F(2, 184) = 1.79, p = .171, η_p² = .02, and did not differ by identity group, F(2, 184) = 0.18, p = .840, η_p² = .002. Change in worry was not affected by condition, F(2, 184) = 1.05, p = .352, η_p² = .01, and did not differ by identity group, F(2, 184) = 0.23, p = .796, η_p² = .002.

Overall, psychosocial outcomes did not change over time, the intervention condition did not affect their change over time, and these outcomes did not change over time differently between identity groups. Contrary to our hypotheses, the mixed-model ANOVAs did not reveal statistically significant Time × Condition × Identity Group interactions on the psychosocial outcomes: belonging uncertainty, F(2, 186) = 0.55, p = .577, η_p² = 0.006; loneliness,⁸ F(1.86, 168.90) = 1.41, p = .246, η_p² = .02; subjective happiness, F(2, 184) = 0.58, p = .564, η_p² = .006; or life satisfaction, F(2, 184) = 1.73, p = .181, η_p² = .02.

Belonging uncertainty did not change significantly over time, F(2, 186) = 1.25, p = .288, η_p² = .01. Change in belonging uncertainty was not affected by condition, F(2, 186) = 0.10, p = .905, η_p² = .001, and did not differ by identity group, F(2, 186) = 0.67, p = .515, η_p² = .007. Loneliness did not change significantly over time, F(1.86, 168.90) = 0.002, p = .998, η_p² < .001. Change in loneliness was not affected by condition, F(1.86, 168.90) = 0.35, p = .688, η_p² = .004, and did not differ by identity group, F(1.86, 168.90) = 0.48, p = .606, η_p² = .005. Subjective happiness did not change significantly over time, F(2, 184) = 0.82, p = .442, η_p² = .009. Change in happiness was not affected by condition, F(2, 184) = 0.98, p = .377, η_p² = .01, and did not differ by identity group, F(2, 184) = 0.18, p = .837, η_p² = .002. Life satisfaction did not change significantly over time, F(2, 184) = 1.30, p = .276, η_p² = .01. Change in life satisfaction was not affected by condition, F(2, 184) = 0.11, p = .899, η_p² = .001, and did not differ by identity group, F(2, 184) = 0.40, p = .672, η_p² = .004.

Between-subjects effects on mental-health and psychosocial outcomes

Across the mixed-model ANOVAs, there were consistent between-subjects effects of the covariate belonging uncertainty, which highlights the importance of this construct in college environments and the need for social-belonging interventions. Every mental-health and psychosocial self-report outcome was negatively predicted by questioning one’s belonging. Participants with less certainty in their belonging reported higher depressive symptoms, F(1, 92) = 7.11, p = .009, η_p² = .07, and greater worry, F(1, 92) = 16.87, p < .001, η_p² = .16, across the study. Participants with less certainty in their belonging also reported greater loneliness, F(1, 91) = 15.95, p < .001, η_p² = .15; less subjective happiness, F(1, 92) = 13.84, p < .001, η_p² = .13; and less life satisfaction, F(1, 92) = 10.16, p = .002, η_p² = .10.

In addition, participants of historically disadvantaged identities reported less life satisfaction, F(1, 92) = 4.32, p = .041, η_p² = .05. Female participants were less lonely than participants of other gender identities, F(1, 91) = 8.57, p = .004, η_p² = .09. Finally, younger participants reported higher depressive symptoms, F(1, 92) = 6.95, p = .010, η_p² = .07, and greater worry, F(1, 92) = 6.77, p = .011, η_p² = .07. Younger participants also reported greater loneliness, F(1, 91) = 8.08, p = .006, η_p² = .08, and less life satisfaction, F(1, 92) = 6.44, p = .013, η_p² = .07. No other between-subjects effects on mental-health or psychosocial self-report outcomes were significant (ps > .072).

Academic outcomes

Within-subjects effects

The mixed-model ANOVAs did not reveal a statistically significant Time × Condition × Identity Group interaction on noncumulative GPA, F(2.79, 284.05) = 0.26, p = .842, η_p² = .003. GPA did not change significantly over time, F(2.79, 284.05) = 1.73, p = .165, η_p² = .02. Change in grades was not affected by condition, F(2.79, 284.05) = 1.37, p = .252, η_p² = .01, and did not differ by identity group, F(2.79, 284.05) = 2.39, p = .074, η_p² = .02. There was a significant Cohort × Time interaction, F(2.79, 284.05) = 9.86, p < .001, η_p² = .09. For the latter cohort that was affected by COVID-19 grade policies, sophomore spring GPA was higher than other semesters.

Between-subjects effects

Participants of historically advantaged identities had higher GPAs, F(1, 102) = 8.91, p = .004, η_p² = .08. There was a significant between-subjects effect of cohort. The latter cohort that was affected by COVID-19 grade policies had higher GPAs, F(1, 102) = 4.59, p = .035, η_p² = .04. No other between-subjects effects were significant (ps > .170).

Discussion

The purpose of the present study was to develop and test a brief social-belonging intervention intended to reduce psychological and academic disparities. Rather than delivering the intervention individually as in prior investigations (e.g., Walton & Cohen, 2011), we tested a single-session group format that could be implemented more efficiently as institutional practice. Through the presentation and discussion of college-specific student statements, the intervention emphasized that first-year challenges with belonging are normative—regardless of socioeconomic class, race, or gender—and that over time, students develop relationships and networks that expand their sense of academic and social belonging.

Longitudinal self-report assessments in this study confirmed that belonging uncertainty is a significant barrier to mental health and well-being in college students. Participants who reported poorer belonging at baseline experienced greater depressive symptoms, greater worry, greater loneliness, less happiness, and less life satisfaction over the follow-up period. This highlights the importance of sense of belonging in college environments and the potential impact of belonging-focused interventions.

This study was the first to examine major depression as a primary outcome of a brief social-belonging intervention, and it uncovered promising findings. Note that although there were no clinical inclusion criteria for the study, 26.2% of the participants already had a history of major depression, experienced during high school and/or the fall or winter of their first year of college. This rate again underscores the concerning prevalence of adolescent depression and the importance of efforts to prevent depression onset and recurrence (Rohde et al., 2013). Nearly 20% of the sample reported a new MDE over the 14-month follow-up period that ended in spring of their sophomore year. This brief social-belonging intervention significantly reduced risk for major depression; 26.2% of control participants reported an MDE compared with 13.1% of participants assigned to the intervention condition. In addition, the intervention specifically reduced risk for depression among participants with greater discrimination experiences. This intervention may have disrupted the indirect effect reported by Freire and Hurd (2023) in which discrimination experiences predict greater depression via lowered sense of belonging. Self-report depressive symptoms did not demonstrate the same intervention effect, but this can be explained by a difference in time frame. The BDI-II (Beck et al., 1996) assessed only depressive symptoms experienced over the most recent 2 weeks, whereas the diagnostic interviews covered depressive symptoms experienced over the full 14-month follow-up window. Thus, information from the diagnostic interviews offered a more complete picture of the course of participants’ mood and the benefits of the intervention to their mental health.

Categories of racial identity and first-generation status do not fully capture individual differences in day-to-day experience, which became important in identifying who benefited from this intervention (see Walton et al., 2023). It is not evident that this intervention directly altered group-based worry about belonging as conceptualized by Walton and colleagues (Brady et al., 2020; Walton & Cohen, 2011). However, it may have altered how participants with a variety of discriminatory experiences interpreted difficulties. The group intervention may lift some self-blame for social challenges when participants learn from student statements and through discussion with other participants that many other first-year students similarly quietly struggle with belonging. Reducing ruminative self-blame and global attributions may then prevent a slide toward clinical depression. Other investigations have demonstrated that social-belonging interventions increase engagement, particularly with peers and mentors (Binning et al., 2020; Brady et al., 2020). Mayo and Le (2023) reported that among college students, perceived discrimination lowered mentoring support (and academic self-concept) and, in turn, lowered mental health. By framing initial challenges as common and temporary, this intervention may have lifted a psychological barrier to engagement and social connection, thereby having a preventive effect consistent with the behavioral-activation model of depression (Martell et al., 2022).

The present research extends the outcomes potentially improved by a brief belonging intervention, but the findings do not replicate those of prior investigations in which BIPOC or first-generation students uniquely benefited from a social-belonging or difference-education intervention (Brady et al., 2020; Stephens et al., 2014; Townsend et al., 2019; Walton & Cohen, 2011; Yeager et al., 2016). This style of intervention appears to have beneficial but heterogeneous effects, working differently in different contexts (Walton et al., 2023). There were no significant differences between socially disadvantaged control and intervention participants, and this social-belonging intervention did not lead to statistically significant improvement in psychosocial or academic outcomes in general. Note that this study was not designed as a replication study testing whether the previously reported effect was “real,” but instead, it tested whether the effect persists when delivered in a group-discussion format.

There are several reasons why the group intervention may not have provided the same benefit to BIPOC and first-generation students as previous interventions. On the one hand, differences in the study context, positive or negative, may have affected the outcome. It appears that Black participants in Walton and Cohen’s (2011) intervention started at higher belonging uncertainty and lower subjective happiness than the socially disadvantaged participants in this study. Belonging uncertainty was not significantly different between socially disadvantaged and advantaged participants at baseline or over time; thus, there may have been less of a belonging gap to close in this context than in prior investigations. Future investigations should continue to evaluate constructs beyond categorical group identity, such as discrimination or harassment experiences, as possible predictors of belonging concerns.

On the other hand, one cannot rule out that the study site did not support a more adaptive mindset about belonging for socially disadvantaged students. Social-belonging interventions intend to offer a more adaptive hypothesis that challenges are transient and that belonging improves with time. The effects will not take hold if the institutional environment does not offer the necessary resources and opportunities for all students to develop positive relationships and a general sense of belonging (Brady et al., 2020; Walton et al., 2023).

In addition, there are methodological differences with prior studies. Other investigations had an active control condition, such as providing information that one’s familiarity with the college’s physical environment or one’s sociopolitical attitudes change over time in college; the present study employed a naturalistic control condition (Stephens et al., 2014; Walton & Cohen, 2011). Finally, it is possible that the group format made the intervention more overt. Walton and Cohen (2011) suggested that conscious awareness of a brief intervention’s intent may cause resistance or reactance to the message in some participants or may convey a discouraging message that the attendees are in need of help. Nevertheless, this social-belonging intervention delivered via structured group discussion significantly reduced risk for MDEs.

The present findings should be considered in light of some limitations. This group-intervention study had a relatively small sample size compared with what is feasible via online interventions (e.g., Yeager et al., 2016). The sample size limited power to detect statistically significant intervention effects on psychosocial and academic outcomes because effect sizes for the intervention and its interaction with identity group were small (η_p²s range = .001–.02). On the other hand, the impact of the intervention may have been partially obscured given that 14.8% of participants assigned to the intervention did not attend the group session. These findings from a small liberal arts college might not replicate in larger university or nonresidential settings. It is important to extend investigations of college belonging and mental health to a wide range of institutions, from selective colleges to broad-access universities (Murphy et al., 2020). Finally, these findings may not generalize over time. National awareness and discussion of structural racism and inclusion and equity issues on campuses shifted in the year following data collection for this project. One hopes that this will continue to encourage institutional-level attention to disparities in belonging, but this change will not happen at the same pace across institutions.

This study contributes to the expanding literature on social-belonging interventions, finding that a brief, structured group discussion reduced risk of depression through the first 2 years of college. Further research is needed to illuminate the process by which the intervention benefited participants. For example, did participants in the intervention condition expect some challenges in reaching a sense of belonging and therefore personalized everyday problems less, allowing for greater hope and less negative affect? The intervention’s normalization of transitional challenges may lead to greater expectations for difficulties, but expectations that are accompanied by beliefs in malleability and growth. Future studies should examine changes in malleability beliefs, perceived stress, and stress appraisals as possible mediators of the intervention’s effects.

Future research on campus belonging should also distinguish academic belonging from social belonging and distinguish each from campus-community belonging (Nunn, 2021). For example, previous focus groups at this study site indicated that academic belonging was strong, whereas social and campus-community belonging were more variable. Future interventions could first identify the form of belonging of greatest concern at a specific site. At some institutions, a social-belonging intervention may be effective in encouraging participants to develop personal relationships, but it may not be able to alter broader campus-community belonging. Finally, it is important to note that individual-level intervention cannot substitute for institutional-level endeavors to create contexts in which all students feel welcomed and valued to develop a deeper sense of belonging. Acknowledging the regularity of microaggressions and identifying structural inequalities that affect marginalized students will also affect sense of belonging (Harwood et al., 2012). Students who were able to externalize discriminatory experiences as evidence of persisting inequalities on campus were less likely to question their own belonging and less likely to question their ability to succeed (Offidani-Bertrand et al., 2022).

In summary, this study did not extend previous findings that a belonging intervention specifically benefits BIPOC and first-generation-to-college students’ psychosocial well-being and academic performance. Instead, a brief, structured group discussion of social belonging significantly reduced risk for depression during the first 2 years of college and specifically reduced risk for students experiencing more discrimination.

Supplemental Material

sj-docx-1-cpx-10.1177_21677026231220060 – Supplemental material for A Brief Group Social-Belonging Intervention to Improve Mental-Health and Academic Outcomes in BIPOC and First-Generation-to-College Students

Supplemental material, sj-docx-1-cpx-10.1177_21677026231220060 for A Brief Group Social-Belonging Intervention to Improve Mental-Health and Academic Outcomes in BIPOC and First-Generation-to-College Students by Erin S. Sheets and Denise Young in Clinical Psychological Science

Footnotes

Acknowledgements

We express our appreciation to James Scott, PhD, for statistical consultation and to Mahal Alvarez-Backus, Cari Daniels, Celine El-Abboud, Margaret Hall, Nathan Huebschmann, Robbi Melvin, S. J. Sahagun, Sonia Tremblay, and Yanlin Zhao for contributing to intervention development, leading the intervention groups, and assisting in data collection.

Transparency

Action Editor: Kelsie T. Forbush

Editor: Jennifer L. Tackett

Author Contributions

Erin S. Sheets: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Visualization; Writing – original draft; Writing – review & editing.

Denise Young: Data curation; Investigation; Resources; Writing – review & editing.

ORCID iD

Erin S. Sheets

Supplemental Material

Additional supporting information can be found at

Notes

References

Aronson

Fried

C. B.

Good

(2002). Reducing the effects of stereotype threat on African American college students by shaping theories of intelligence. Journal of Experimental Social Psychology, 38, 113–125. https://doi.org/10.1006/jesp.2001.1491

Beck

A. T.

Steer

R. A.

Brown

G. K.

(1996). Beck Depression Inventory (2nd ed.). The Psychological Corporation.

Binning

K. R.

Kaufmann

McGreevy

E. M.

Fotuhi

Chen

Marshman

Kalender

Z. Y.

Limeri

L. B.

Betancur

Singh

(2020). Changing social contexts to foster equity in college science courses: An ecological-belonging intervention. Psychological Science, 31(9), 1059–1070. https://doi.org/10.1177/0956797620929984

Brady

S. T.

Cohen

G. L.

Jarvis

S. N.

Walton

G. M.

(2020). A brief social-belonging intervention in college improves adult outcomes for Black Americans. Science Advances, 6(18), Article eaay3689. https://doi.org/10.1126/sciadv.aay3689

Brown

T. A.

Antony

M. M.

Barlow

D. H.

(1992). Psychometric properties of the Penn State Worry Questionnaire in a clinical anxiety disorders sample. Behaviour Research and Therapy, 30(1), 33–37. https://doi.org/10.1016/0005-7967(92)90093-V

Diener

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75. https://doi.org/10.1207/s15327752jpa4901_13

Dutcher

J. M.

Lederman

Jain

Price

Kumar

Villalba

D. K.

Tumminia

M. J.

Doryab

Creswell

K. G.

Riskin

Sefdigar

Seo

Mankoff

Cohen

Dey

Creswell

J. D.

(2022). Lack of belonging predicts depressive symptomatology in college students. Psychological Science, 33(7), 1048–1067. https://doi.org/10.1177/09567976211073135

Erford

B. T.

Johnson

Bardoshi

(2016). Meta-analysis of the English version of the Beck Depression Inventory–Second Edition. Measurement and Evaluation in Counseling and Development, 49, 3–33. https://doi.org/10.1177/0748175615596783

Espenshade

T. J.

Radford

A. W.

(2009). No longer separate, not yet equal: Race and class in elite college admission and campus life. Princeton University Press.

10.

First

M. B.

Williams

J. B. W.

Karg

R. S.

Spitzer

R. L.

(2015). Structured clinical interview for DSM-5-research version. American Psychiatric Association.

11.

Freire

D. S.

Hurd

N. M.

(2023). Discrimination and mental health outcomes among underrepresented college students: The role of sense of belonging at predominantly White institutions. Emerging Adulthood, 11(3), 654–668. https://doi.org/10.1177/21676968231166967

12.

Gopalan

Brady

S. T.

(2020). College students’ sense of belonging: A national perspective. Educational Researcher, 49(2), 134–137. https://doi.org/10.3102/0013189X19897622

13.

Harwood

S. A.

Huntt

M. B.

Mendenhall

Lewis

J. A.

(2012). Racial microaggressions in the residence halls: Experiences of students of color at a predominantly White university. Journal of Diversity in Higher Education, 5(3), 159–173. https://doi.org/10.1037/a0028956

14.

Hayes

A. F.

(2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). The Guilford Press.

15.

Higgins

E. T.

Rholes

W. S.

(1978). “Saying is believing”: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14, 363–378. https://10.1016/0022-1031(78)90032-X

16.

Kim

Sellbom

Ford

K. L.

(2014). Race/ethnicity and measurement equivalence of the Everyday Discrimination Scale. Psychological Assessment, 26, 892–900. https://doi.org/10.1037/a0036431

17.

Lyubomirsky

Lepper

H. S.

(1999). A measure of subjective happiness: Preliminary reliabiltiy and construct validity. Social Indicators Research, 46, 137–155. https://doi.org/10.1023/A:1006824100041

18.

Marksteiner

Janke

Dickhäuser

(2019). Effects of a brief psychological intervention on students’ sense of belonging and educational outcomes: The role of students’ migration and educational background. Journal of School Psychology, 75, 41–57. https://doi.org/10.1016/j.jsp.2019.06.002

19.

Martell

C. R.

Dimidjian

Herman-Dunn

(2022). Behavioral activation for depression: A clinician’s guide (2nd ed.). The Guilford Press.

20.

Mayo

(2023). Perceived discrimination and mental health in college students: A serial indirect effects model of mentoring support and academic self-concept. Journal of American College Health, 71(4), 1184–1195. https://doi.org/10.1080/07448481.2021.1924725

21.

Meyer

T. J.

Miller

M. L.

Metzger

R. L.

Borkovec

T. D.

(1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. https://doi.org/10.1016/0005-7967(90)90135-6

22.

Murphy

M. C.

Gopalan

Carter

E. R.

Emerson

K. T. U.

Bottoms

B. L.

Walton

G. M.

(2020). A customized belonging intervention improves retention of socially disadvantaged students at a broad-access university. Science Advances, 6(29), Article eaba4677. https://doi.org/10.1126/sciadv.aba4677

23.

Nunn

L. M.

(2021). College belonging: How first-year and first-generation students navigate campus life. Rutgers.

24.

Offidani-Bertrand

Velez

Benz

Keels

(2022). “I wasn’t expecting it”: High school experiences and navigating belonging in the transition to college. Emerging Adulthood, 10(1), 212–224. https://doi.org/10.1177/2167696819882117

25.

Ostrove

J. M.

Long

S. M.

(2007). Social class and belonging: Implications for college adjustment. Review of Higher Education, 30, 363–389. https://doi.org/10.1353/rhe.2007.0028

26.

Rohde

Lewinsohn

P. M.

Klein

D. N.

Seeley

J. R.

Gau

J. M.

(2013). Key characteristics of major depressive disorder occurring in childhood, adolescence, emerging adulthood, and adulthood. Clinical Psychological Science, 1(1), 41–53. https://doi.org/10.1177/2167702612457599

27.

Russell

D. W.

(1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66(1), 20–40. https://doi.org/10.1207/s15327752jpa6601_2

28.

Steele

C. M.

(1997). A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist, 52(6), 613–629. https://doi.org/10.1037//0003-066x.52.6.613

29.

Stephens

N. M.

Hamedani

M. G.

Destin

(2014). Closing the social-class achievement gap: A difference-education intervention improves first-generation students’ academic performance and all students’ college transition. Psychological Science, 25, 943–953. https://doi.org/10.1177/0956797613518349

30.

Townsend

S. S. M.

Stephens

N. M.

Smallets

Hamedani

M. G.

(2019). Empowerment through difference: An online difference-education intervention closes the social class achievement gap. Personality and Social Psychology Bulletin, 45(7), 1068–1083. https://doi.org/10.1177/0146167218804548

31.

Wadsworth

L. P.

Morgan

L. P.

Hayes-Skelton

S. A.

Roemer

Suyemoto

K. L.

(2016). Ways to boost your research rigor through increasing your cultural competence. The Behavior Therapist, 39, 83–89.

32.

Walton

G. M.

Cohen

G. L.

(2007). A question of belonging: Race, social fit, and achievement. Journal of Personality and Social Psychology, 92(1), 82–96. https://doi.org/10.1037/0022-3514.92.1.82

33.

Walton

G. M.

Cohen

G. L.

(2011). A brief social-belonging intervention improves academic and health outcomes of minority students. Science, 331(6023), 1447–1451. https://10.1126/science.1198364

34.

Walton

G. M.

Murphy

M. C.

Logel

Yeager

D. S.

, & The College Transition Collaborative. (2017). The social-belonging intervention: A guide for use and customization. http://gregorywalton-stanford.weebly.com/

35.

Walton

G. M.

Murphy

M. C.

Logel

Yeager

D. S.

Goyer

J. P.

Brady

S. T.

Emerson

K. T. U.

Paunesku

Fotuhi

Blodorn

Boucher

K. L.

Carter

E. R.

Gopalan

Henderson

Kroeper

K. M.

Murdock-Perriera

L. A.

Reeves

S. L.

Ablorh

T. T.

Ansari

. . . Krol

(2023). Where and with whom does a brief social-belonging intervention promote progress in college? Science, 380(6644), 499–505. https://doi.org/10.1126/science.ade4420

36.

Williams

D. R.

Jackson

J. S.

Anderson

N. B.

(1997). Racial differences in physical and mental health: Socio-economic status, stress and discrimination. Journal of Health Psychology, 2, 335–351. https://doi.org/10.1177/135910539700200305

37.

Yeager

D. S.

Walton

G. M.

(2011). Social-psychological interventions in education: They’re not magic. Review of Educational Research, 81(2), 267–301. https://doi.org/10.3102/0034654311405999

38.

Yeager

D. S.

Walton

G. M.

Brady

S. T.

Akcinar

E. N.

Paunesku

Keane

Kamentz

Ritter

Duckworth

A. L.

Urstein

Gomez

E. M.

Markus

H. R.

Cohen

G. L.

Dweck

C. S.

(2016). Teaching a lay theory before college narrows achievement gaps at scale. Proceedings of the National Academy of Sciences, USA, 113(24), E3341–E3348. https://doi.org/10.1073/pnas.1524360113

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB