Since their inception, fraternities have functioned as sites of exclusion in manners that have drawn criticism from individuals both in and outside of higher education. In particular, historically white fraternities mirrored the postsecondary institutions in which they emerged, preventing access to those who did not hold privileged identities (e.g., white, Protestant; Hughey, 2009). Within fraternal organizations, these philosophies took the form of clauses prohibiting people of color into their groups, as well as other minoritized communities (e.g., Jewish individuals; Barone, 2014; Gillon et al., 2019). These legacies of oppression inevitably led to the creation of organizations specifically for marginalized populations, as well as advocacy within historically white fraternities to be more inclusive in membership practices. Although fraternities eventually removed the white clauses that were previously included in their constitutions (Gillon et al., 2019), effectively opening the opportunity for people of color to join, contemporary scholars still question how embedded whiteness and racism are within these organizations (Barone, 2014).

In fact, the present moment has seen university leaders, faculty members, students’ family members, and students, including on occasions, sorority and fraternity members calling for the abolition of fraternity life due to the perpetuation of oppressive attitudes toward people of color (Lautrup, 2020), amongst other reasons (e.g., perpetuation of hazing and sexual assault; Cohen, 2021). Contemporary research, however, paints a more complex and nuanced picture of fraternal life and its influence on how students think about and interact with diverse others (Asel et al., 2015; Martin et al., 2011, 2015; Morgan et al., 2015; Porter, 2012). Namely, this body of scholarship has tracked whether participation in fraternal organizations results in members interacting with those whose identities differ than their own or if these groups lead to homophilic networks. Although this research is instrumental to understanding the potential relevance for fraternal organizations in the current period, these questions of how fraternities influence members’ openness to diversity have been further complicated by recent political, social, and public health crises (e.g., COVID-19 pandemic).

The purpose of this study was to extend research into fraternity life using a comprehensive dataset collected during the 2019–2020 and 2020–2021 academic years to address the following research question: What factors influence the development of openness to diversity and challenge (ODC) amongst fraternity members? In addition to considering students’ identity patterns (e.g., race/ethnicity, political leaning) as possible predictors, we also examined two factors germane to this population: participation in political and social activities (Pascarella, 2007) and measures of fraternal brotherhood (see McCreary & Shutts, 2015) given their potential influences on our outcome of interest. We approached this study informed by critical quantitative leanings (Hernández, 2015; Wells & Stage, 2015) given our hope to further conversations of equity in fraternity life.

We locate our rationale for this project in three phenomena. First, this study provides an opportunity to consider this question using data collected during the COVID-19 pandemic, as well as during a period reflective of heightened racial injustice, which at times, saw racialized violence occurring directly related to the pandemic; thus, any inferences drawn reflect the realities of pandemic-era student–student and student–campus relations, as well as a time in which white audiences seemingly grew more attentive to the racial realities of communities of color. Second, the longitudinal, multi-institutional nature of this inquiry allowed us to isolate the effects of campus-wide (e.g., involvement) and fraternity-specific (e.g., brotherhood) dimensions of students’ experiences on their ODC over-and-above pre-test scores (see Pascarella & Wolniak, 2004). Finally, drawing on existing quantitative (e.g., Asel et al., 2015; Martin et al., 2011, 2015) and qualitative (e.g., Joyce, 2018; Morgan et al., 2015) literature, our study expands what is known about the development of fraternity members and generates evidence that can guide decisions made at individual fraternal organizations and campuses while informing this meaningful national conversation.

Literature Review

To frame our inquiry, we examined two bodies of literature pertaining to the study’s phenomenon of interest. We first begin by describing the existing scholarship on students’ ODC before then turning our attention to research on engagement with diversity in the sorority and fraternity life context—specifically as it relates to race- and sexuality-related dynamics, given their relevance in this particular research.

Students’ Openness to Diversity and Challenge

In an ever-increasing diverse society where the inability to dialogue across difference has dire consequences (King et al., 2013), being open to diversity, and the challenges that may result from engaging across difference, is a “critical disposition” for students during their experiences in higher education and beyond (Bowman, 2014, p. 277). Scholars have found ODC to be positively associated with a number of outcomes, such as students’ interactions and engagement (Bowman, 2012, 2014), civic engagement (Denson & Bowman, 2013), as well as first-year GPA and retention from first to second year (Bowman, 2014). As such, thought leaders routinely identify it as a critical outcome of higher education (Mayhew et al., 2016; Trolian & Parker, 2022).

Interest in students’ openness to diversity has only intensified in recent years because of increased attention to, and unrest related to, social injustices (Trolian & Parker, 2022). In particular, the societal reckoning with racism that occurred in the wake of George Floyd’s murder and other attacks on Black bodies served as an impetus within higher education to invest more resources in anti-racist learning opportunities and initiatives (e.g., Watt et al., 2021). However, little is known about how, if at all, students’ openness to diversity changed after the events from the summer of 2020.

Though the development of ODC among college students is highly coveted, longitudinal change related to this outcome across students’ college experiences may be minimal (e.g., Engberg & Hurtado, 2011), or even decline during one’s time in postsecondary education (O’Neill, 2012; Shim & Perez, 2018). Identifying and exploring experiences that positively affect ODC is vital not only for the benefit of college students and their college communities, but also for our entire society as a whole.

Scholars have identified that specific curricular and extracurricular experiences may affect students’ openness to diversity. Related to the academic experience, students’ perceptions of faculty practices and faculty-student interactions (Trolian & Parker, 2022), assignments and discussions that require engagement with diverse perspectives (Shim & Perez, 2018), diversity courses (O’Neill, 2012), and collaborative learning and civic engagement (Cabrera et al., 2002; Loes et al., 2018) have been positively associated to ODC. Faculty interactions, pedagogy, and learning environments may shape students’ diversity attitudes. Like academic experiences, extracurricular experiences likely affect students’ attitudes toward diversity. Researchers have discovered that diversity experiences (Engberg, 2004), voluntary community service (O’Neill, 2012), living and learning programs and residence hall climate (Longerbeam, 2010), diversity and cultural awareness workshops (Engberg, 2004; Shim & Perez, 2018) are positively associated with ODC. Intergroup peer interactions within academic and extracurricular settings may also be particularly important in students’ development of ODC (Bowman, 2014; Shim & Perez, 2018; Whitt et al., 2001). However, not all intergroup interactions are beneficial. Findings from Shim and Perez (2018) suggested that positive interactions with diverse peers promote ODC over a student’s first year, while negative interactions lead to decreases in this outcome.

The benefits of experiences intended to enhance ODC likely vary based on students’ identities and prior experiences (Engberg & Hurtado, 2011; Pascarella et al., 1996; Shim & Perez, 2018). Scholars like Engberg (2004) have identified that white students benefit from these experiences more than students with minoritized racial or ethnic identities. In addition, Shim and Perez (2018) argued that structural diversity is needed for meaningful intergroup peer interactions.

Fraternal Organizations and Openness to Diversity

Existing scholarship shows a conflicting view on fraternity life and its influences on how students conceptualize relationships with diverse others, which many scholars use to describe interacting with groups like students from other racial backgrounds or who identify with sexualities different from one’s own (Asel et al., 2015; Bowman, 2014; Martin et al., 2011, 2015; Morgan et al., 2015; Porter, 2012). For example, Bowman (2014) asserted that the body of scholarship on fraternities and their engagement with diversity showcases that participating in these organizations does not have a relationship with members’ attitudes toward diverse others or in fact, reveals a negative association. Attempting to understand this relationship, quantitative studies using the Wabash National Study of Liberal Arts Education dataset revealed that fraternities do not have a significant impact on fraternity and sorority members’ intercultural effectiveness, a measure that describes how open students are to learning and accepting other cultures (Martin et al., 2011, 2015). However, other researchers like Asel et al. (2015) indicated that when compared to individuals not in a fraternity, first-year students who were affiliated with a fraternity or sorority were more likely to report less interactions with diverse others. Of note, Porter’s (2012) work on interactional diversity opportunities showed how institution type may in fact affect members’ abilities to engage with people different from one’s self, highlighting that those from private institutions had more interactional diversity opportunities than those who were at public colleges and universities. Therefore, it is not simply participation in fraternities that inform one’s openness to diversity, but rather, other collegiate environments are also at play.

The limited qualitative scholarship on this topic underscores that fraternity men in particular are skilled at minimizing the importance of and the need to think about race (Joyce, 2018; Joyce & Cawthorn, 2017; Morgan et al., 2015), together with targeting or rendering invisible those who identify as part of the queer community (e.g., Duran et al., 2021; Garcia & Duran, 2021; Giacalone, 2018). For instance, discussions of openness to diversity are frequently front and center during the recruitment of new members. These studies showcase that white fraternity men are prone to adopting colorblind ideologies, in which they purport to minimize prospective members’ racial identities in favor of using language of fit and tradition (Joyce, 2018; Morgan et al., 2015). Similarly, when it comes to sexuality, research has demonstrated how members may actively work against the recruitment of queer men due to fears of becoming known as a queer chapter on campus—worries that are present in fraternities broadly (Giacalone, 2018), including culturally based organizations (Garcia & Duran, 2021). As evident in Joyce and Cawthorn’s (2017) research on the group socialization of fraternity men, white fraternity men additionally struggle to determine what is racially insensitive language and behaviors. Likewise, heterosexual fraternity men are also likely to ostracize members who they perceive as not embodying heteronormative and hypermasculine ideals (Duran et al., 2021; Giacalone, 2018). Although these dynamics of marginalization are not specific to fraternal organizations, what is evident across this body of literature is that fraternities are a troublesome site when it comes to matters of diversity. Yet, we were interested in how this specific time period–given the increased attention to racial injustices and the COVID-19 pandemic—saw, and potentially furthered, these patterns.

Epistemological Foundation

Increasingly, researchers in the field of higher education have been exploring the potential for quantitative studies to advance goals of equity; namely, more and more scholars are rejecting postpositivism and embracing a critical turn in quantitative work (Hernández, 2015; Wells & Stage, 2015). Moving away from postpositivism requires quantitative researchers to intentionally design studies that critique and challenge the presence of oppressive systems—both in the phenomenon of interest under investigation, as well as in the methods utilized to make meaning of the study (Garcia et al., 2018). In particular, critically-oriented quantitative researchers must embed this epistemology in facets such as the questions they are asking, how they construct their design, and how they interpret the results (Rios-Aguilar, 2014). Scholars are often limited in engaging critical quantitative schools of thought (Hernández, 2015), especially for those utilizing already-existing data sets that may not contain variables framed in an equity-centered fashion (Oseguera & Hwang, 2014). Although the present study falls under this umbrella, we as researchers adopted a more critical epistemological leaning in the study. What this meant is that we made intentional decisions, further described below, guided by our desire to use quantitative research for liberatory aims. However, we acknowledge that we may have fallen short of these aims at times and later offer ways that future scholars can further utilize critical quantitative designs.

Conceptual Foundation

To provide a conceptual foundation for this work we took guidance from ecological perspectives, both as originally theorized (Bronfenbrenner, 1979, 1995) and as reimagined for more contemporary student development contexts (e.g., Mayhew & Rockenbach, 2021; Renn & Arnold, 2003). In these presentations, scholars have argued it is necessary to address the nested environments where students find themselves and endeavor to engage complex processes of learning and development. In this project, we used this knowledge in our mobilization of a conceptual model that accounted for institutional contexts while also including environments and experiences specific to fraternities. Importantly, in our interpretation of these nested environments, we closely considered how these settings are complicit and perpetuate oppressive systems, a viewpoint that aligns with our critical leanings. Although our model could not fully capture all aspects of the ecological conditions confronting students during this time, we were mindful to incorporate variables that accounted for the widest possible presentation of students and their institutions in context (Duran & Jones, 2019).

Notably, we found support for our longitudinal approach in the work of Renn and Arnold (2003) who detailed the distinctive importance of the chronosystem—the element of time as shaping all aspects of an ecology. Additionally, the Mayhew and Rockenbach (2021) model illustrated how previously recorded ideas about an outcome (e.g., in the form of a pre-test) can serve as an input in college impact work; this was a perspective we adopted as well, elaborated upon below in our methods. As scholars located within the very nested systems of privilege, power, and policy we hope to challenge, we now provide our own positionality as pertains to this inquiry.

Positionality

Adopting a critical quantitative orientation in this study involved the recognition that this research was not objective. Rather, critical quantitative scholars assert that who the researchers are and what their worldviews entail inevitably shapes the questions they ask, how they approach their inquiry, and how they make interpretations of the social phenomenon under investigation (Rios-Aguilar, 2014). Following this belief, we see it imperative to bring to light the positionalities of the research team and how they informed the decisions made in this study.

McCready identifies as a cisgender, straight white man who was actively involved in a historically white social fraternity as an undergraduate student. He continued his involvement in his fraternity as a graduate as an international headquarters staff member and volunteer. He served as a fraternity and sorority life professional for over a decade during his student affairs career. His identity and experiences likely afforded him access to the data for this study, because he is viewed as trustworthy by many fraternity and sorority leaders. His relationship with historically white social fraternities affected how he viewed the data, interpreted the findings, and considered the implications of the research. He acknowledges how his privilege shaped this study.

Selznick identifies as a cisgender man whose research embraces multiple epistemic approaches to studying postsecondary learning and the leadership conditions under which students can inclusively thrive. During college, he held numerous leadership positions in a co-ed residential undergraduate society at a selective private institution at which over 50% of students were involved in fraternity and sorority life. These experiences provided him with a perspective on this work which recognized the possibilities for quantitatively examining associations between collegiate experiences and the outcome using a longitudinal HLM strategy as a way to draw inferences regarding student development. This positionality also provided a lens on striving to ensure that findings from this study could potentially be used and practiced in ways that could offer opportunities for meaningful, student-centered connection between fraternities and postsecondary missions.

Duran identifies as a queer cisgender Latino man whose research largely focuses on examining how campus environments serve those with minoritized social identities. Central to this investigation, he is not affiliated with a fraternal organization, yet has served as an advisor to a culturally based fraternity and has also held the position of a fraternity House Director in the past. These professional experiences, combined with his scholarly interests, led him to be aware of the larger ecosystems in which fraternities function and also informed his criticality when it comes to their engagement with issues of diversity and equity—specifically around topics of race, gender, and sexuality. He was particularly mindful of which variables made sense to include in our models, as well as what the findings meant for practice on college campuses.

Methods

Data for the study were collected during the 2019–2020 academic year (Time 1) and 2020–2021 academic year (Time 2) through internet-based surveys distributed by Dyad Strategies, LLC (Dyad) to the active membership of three historically white college men’s social fraternities with chapters located at higher education institutions across the U.S. and Canada. More information about the three fraternities can be found in Table 1. Though response rates for the survey administrations ranged from 30.2 to 73.8%, the data set for the study included 4763 cases from individuals at 272 higher education institutions who responded during both data collection periods. We purged 929 cases through listwise deletion due to missingness not-at-random or inaccuracies (e.g., flippant responses to the race/ethnicity variable). To abide by acceptable standards for group size in multilevel modeling (Maas & Hox, 2005), we excluded 413 cases from institutions with fewer than 10 participants. The final sample included data collected from 3420 college men from 134 U.S. institutions.

Table 1 Description of historically white fraternities included in study
Full size table

Within the final sample, 0.8% identified as American Indian or Native Alaskan, 4.4% identified as Asian other than Indian, 2.3% identified as Black, 7.5% identified as Hispanic, 1.6% identified as Indian, 1.2% identified as Middle Eastern or North African, and 77.6% identified as white. Other participants identified as multiracial (3.9%), and 0.7% identified with other racial or ethnic identities. Because of the gender-exclusive policies of the three fraternities, all participants in our sample identified as men. Nearly half of participants (51.5%) identified that they worked in addition to being students, and 13.9% identified as first-generation college students (FCGS). The majority of participants (40.9%) described their political leaning as moderates, while 23.1% identified their political lean as liberal or very liberal, and 35.9% identified as conversative or very conservative. Table 2 details the demographic characteristics of the final sample.

Table 2 Descriptive statistics for individual-level categorical and ordinal variables
Full size table

The institutional sample varied by cluster size, geographic location and institutional type. The cluster sizes of the 134 institutions ranged from 10 to 145 participants (M = 25.22; SD = 19.65), with a mode of 12 participants (N = 10). Of the institutions in our sample, 26 (19.4%) are located in the Northeast, 47 (35.1%) in the South, 39 (29.1%) in the Midwest, and 22 (16.4%) in the West. The overwhelming majority of these institutions are doctoral universities (N = 101; 75.4%). Of the rest, 23 (17.3%) are master’s colleges or universities, and 10 (7.5%) are baccalaureate colleges.

Dependent Variable

The dependent variable was the ODC of participants as reported in Time 2 (see Table 3 for the descriptive statistics for the outcome and other continuous variables). ODC was measured using the seven-item scale developed by Pascarella et al. (1996) that evaluates students’ appreciation of diversity and diverse perspectives. Participants rated their agreement using a Likert format ratings scale from “strongly disagree” (1) to “strongly agree” (5) [e.g., “Contact with individuals whose backgrounds (e.g., race, national origin, sexual orientation) are different from my own is an essential part of my college education”]. Pascarella et al. (1996) found the scale had an \(\propto\)= 0.83. Scores were calculated based on the mean value of the items corresponding to the scale. In the sample, the Cronbach’s \(\propto\) = 0.934 for ODI in Time 2.

Table 3 Descriptive statistics and correlations for individual-level and institution-level continuous variables
Full size table

Individual-Level Predictors

The individual-level predictors included in analyses were students’ group-mean centered responses in Time 1 for ODC (Pascarella et al., 1996; \(\propto\) = 0.924), orientations toward political and social involvement in Time 2 (PSI) (Pascarella, 2007), and four subscales of McCreary and Schutts’s (2015) Fraternal Brotherhood Questionnaire (FBQ) from Time 2. These predictors were group-mean centered; upon creating hierarchical linear modeling (HLM) unconditional models for each, we found the intraclass correlation coefficients (ICCs) were statistically significant, p < 0.001 (see Table 3).

PSI was measured using the 11-item scale developed by Pascarella (2007) from exploratory factor analysis of 21 items from the Cooperative Institutional Research Program Freshman Survey developed by the Higher Education Research Institute at the University of California at Los Angeles. Participants reported the importance of being politically and socially involved from “Not at all important” (1) to “Essential” (5) (e.g., “becoming a community leader”). Pasceralla found the scale to have internal consistency reliabilities ranging from 0.80 to 0.83. Scores were calculated based on the mean value of the items corresponding to the scale. In the sample, the Cronbach’s \(\propto\) = 0.917.

FBQ was measured using the 21-item scale developed by McCreary and Schutts (2015). The scale includes four subscales that McCreary and Schutts argued reflect four distinct schema of fraternal brotherhood—accountability (six-items; \(\propto\) = 0.915), belonging (five-items; \(\propto\) = 0.934), shared social experience (five-items; \(\propto\) = 0.950), and solidarity (five-items; \(\propto\) = 0.776). The researchers relied on exploratory and confirmatory factor analyses to develop and validate the scale. Participants rated their agreement using a Likert format ratings scale from “strongly disagree” (1) to “strongly agree” (5) (e.g., “The first people I ask to do things with me are my fraternity brothers”). McCreary and Schutts found the scales had internal consistency reliabilities ranging from 0.71 for solidarity to 0.93 for shared social experience. Scores were calculated based on the mean value of the items corresponding to the scale.

We also included students’ self-reported racial and ethnic identities (American Indian or Native Alaskan, Asian, Black, Hispanic, Indian, Middle Eastern or North African, Multiracial, Other, or White), working during college status (held a paid job or work study role vs. did not hold a paid job or work study role), first generation college student status (either parent went to college vs. neither parent went to college), and political leaning (very liberal, liberal, moderate, conservative or very conservative) from Time 1 in our models. Because of the small number of participants who identified as Middle Eastern or North African, or American Indian or Native Alaskan, we made the decision to collapse these identities into “Other.” Unfortunately, this is a decision that many other researchers who attempt to adopt critical leanings wrestle with and we acknowledge the limitations of such a decision in our study.

We note that, consistent with our overall critically-oriented quantitative approach, we effect-coded our demographic predictors to avoid positioning any group as normative (Mayhew & Simonoff, 2015). Effect coding is a statistical practice which allows for each level (i.e., group) of a categorical variable to be entered into the regression model rather than selecting a reference category upon which to base parameter estimates. As Mayhew and Simonoff (2015) point out, effect coding “may serve as a small step toward rethinking conventions in quantitative practice and their often insidious use in privileging certain voices, experiences, and perceptions over others” (p. 174). The primary difference between traditional indicator coding and effect coding is that in effect coding parameter estimates inherently fall above or below the grand mean as opposed to reflecting an estimate relative to the reference group. As such, there will always be reported positive and negative estimates.

Institution-Level Predictors

The institution-level predictors included the aggregated group means for students’ responses about their ODC in Time 1 (Pascarella et al., 1996), orientations toward political and social involvement (Pascarella, 2007), and four subscales of McCreary and Schutts’s (2015) FBQ.

Analyses

We relied on HLM to analyze the data because the data are nested by institutions. We created an initial unconditional model to identify the amount of variance of the outcome that can be explained by the variance between institutions. After creating this model, we added individual-level predictors to the model across four steps (see Table 4). Step 1 included all of the demographic predictors (e.g., political leaning). Step 2 included the pre-test measure of Time 1 ODC. The FBQ subscales were added during step 3. PSI was added for the final individual-level step. During each step, models were specified, and predictors that were not statistically significant were removed from subsequent models. Dunnett’s test were performed to examine the statistical significance of the effect coded predictors. To further specify the final student-level model, random coefficient models were constructed to explore the variability in the student-level random slopes and the random coefficient.

Table 4 Results of multilevel modeling for openness to diversity and challenge
Full size table

Upon the final specification of the individual-level model, intercepts and slopes-as-outcomes models were constructed to identify if any institution-level variables could explain the variability in the intercept and slopes. The aggregated group mean centered predictors were added to the model to identify if any variability in openness to diversity and challenge in 2021 was due to the between-institution variability of these predictors.

Unconditional Model

The unconditional ICC for the model was 0.044, p < . 001. The reliability estimate was 0.504, indicating that there is adequate stability across the parameter estimates for each institution (Raudenbush & Bryk, 2002). The deviance for this model with two parameters was 7189.963. The design effect for the sample is 2.078. Therefore, the effective sample size is 1645.813, meaning there was sufficient statistical power to conduct HLM analyses.

Individual-Level Models

The cross-sectional one-way ANCOVA with random effects models were constructed utilizing individual-level predictors that were not allowed to vary randomly along their slopes. The preliminary models will not be reviewed thoroughly; however, the final model indicated that political leaning, ODC at Time 1, the FBQ subscales, and PSI were statistically significant predictors of the outcome. FGCS, racial and ethnic identity, and working were found not to be significant predictors of ODC (see Table 4).

In the final student-level specified model, the coefficient slopes of belonging brotherhood, ODC in Time 1 and PSI varied randomly, ps < 0.001, and these random coefficients were included from the subsequent slopes-as-outcomes models. The reliability of the intercept in the final individual-level model was sufficiently reliable at 0.777.

Slopes-as-Outcomes Models

After specifying the student-level model, institution-level predictors were added to a combined model at the intercept (see Table 4). The model was specified by comparing the deviance of the slopes-as-outcomes models with fixed effects and one with random coefficients. Because deviance of the model individual-level fixed effects (4339.789) was higher than the model with the three random coefficients (4284.132), the final model was specified to exclude the random effects. The reliability of the student-level intercept remained adequate at 0.262.

Limitations

Several important limitations confront this study that must be acknowledged prior to introducing results. First, we recognize that while response rates were relatively high, our study has no mechanism beyond reported data screening and cleaning for understanding the true motivations for student participation. Given the constructs employed on this instrument and the often-precarious contexts of its timing, it is necessary to consider the possibilities for non-response bias to exist. Second, and relatedly, though our study employs robust multilevel modeling, our findings are not necessarily generalizable to all fraternity populations. Though we are able to account for between-institution and between-subject variability, results remain to some degree sample-dependent. Third, our study did not include variables related to students’ perceptions, experiences, or emotions related to the COVID-19 pandemic or the racial injustices and social unrest that occurred during the summer of 2020. We are unable to make direct inferences related to these elements of the chronosystem. However, the data we relied on for our study were collected immediately before these events (Time 1) and during the academic year following the COVID-19 lockdown and events from the summer of 2020 (Time 2). It is more likely than not that these circumstances affected participants’ college experiences. Finally, we remind audiences that our findings are non-causal and correlative. In this regard, we envision findings serving as potential catalysts for conversations and context-specific actions on demonstrated, evolving relationships between collegiate experiences and openness to diversity and challenge among those fraternity members in our sample.

Results

Preliminary Analysis

The descriptive statistics for the variables can be found in Table 3. The mean for ODC at Time 2 was 3.973 (SD = 0.698), and the mean at Time 1 was 4.040 (SD = 0.681). While the mean declined by 0.067 and the t-test was statistically significant t(3419) = 5.123, p < 0.001, the effect size was small [d = 0.088; CI (0.054, 0.121)]. The data for ODC at Time 2 was nearly symmetrical. The skewness was − 0.281 (SE = 0.042), and the kurtosis was − 0.309 (SE = 0.084). We examined the continuous variables in the study and found that none violated the assumptions for linear regression models (Raudenbush & Bryk, 2002).

Openness to Diversity Model

Table 4 presents findings from the final specified hierarchical linear model. In the final model, the inclusion of political leaning predictors indicates ODC declines as an individual is more conservative (e.g., very conservative, b = − 0.130, p < 0.01). However, the demographic predictors only explained 1.4% of the within-institution variance in step 1. Four environmental individual-level variables included in the final model were statistically significant predictors of ODC, though the FBQ predictors of solidarity and shared social experience brotherhood were not statistically significant (see Table 4). Notably, individual-level PSI in the model (b = 0.616, p < 0.001) explained 22.7% of the with-institution variance prior to adding the institution-level predictors. As a fraternity member’s reported value of political/social activism increases, their openness to diversity and challenge was likely correspondingly increase.

Several institutional-level aggregated group-mean centered predictors were statistically significant as well (see Table 4). These findings suggest the institutional environment related to ODC in Time 1 (b = 0.243, p < 0.001), PSI (b = 0.641, p < 0.001), and conceptualizations of fraternal brotherhood for belonging (b = 0.179, p = 0.010) and shared social experience (b = − 0.151, p = 0.002) may relate to a fraternity member’s openness to diversity and challenge during the 2020–2021 academic year.

The final model explained 64.7% of the between institution variance in ODC, and 56.7% of the within institution variance. The total amount of variance explained by this model reached 57.1%.

Discussion

Results indicated that individual- and institutional-level characteristics and features of students’ experiences are associated with ODC, even after introducing a pre-test score. Building on the work of Mayhew et al. (2016), we frame these results on the premise that though the mean difference on the outcome was negligible, the ability for our model to demonstrate the influence of any feature of the college environment as influential on the outcome over-and-above pre-test provides substantial, actionable evidence of how college affects students (see also Pascarella & Wolniak, 2004). We now comment on these findings’ significance for advancing knowledge on this topic, drawing upon insights stemming from our positionalities as well as the existing literature. As scholars who have a vested interest in making fraternities a more equitable space for all, we are particularly reflective on how dynamics of privilege and marginalization manifest in these organizations.

Results provide evidence that aligns with general findings associated with college-going: students’ come to college seeking out opportunities to learn and that cocurricular experiences such as political and/or social involvement and fraternity life can be supportive of prosocial outcomes like ODC (Mayhew et al., 2016). Focusing first on the political and social involvement, we see this as evidence that students and campus ecologies that promote the frequent, productive exchange of ideas can reinforce students’ abilities to engage in discourse across forms of diversity (e.g., racial/ethnic, sexuality, ideological, cultural). From a critical lens, these interactions are particularly notable for historically white fraternities that have largely perpetuated exclusionary environments both in historical and contemporary times (Garcia & Shirley, 2019; Gillon et al., 2019), realities that we have seen higher education practitioners and members try to change.

Encouragingly for college educators, and reflecting the dedication of student affairs practitioners nationally, this large effect persisted even after a year in which many typical, in-person channels for such engagement were limited. Especially given the at-times rapidly changing nature of ecological systems across all systems, we wonder whether closer connections forged within fraternity organizations may have provided a place and space for students to make meaning of broader political and social engagement, whether occurring in connection with formal collegiate activities (e.g., classes, cocurricular programming) or elsewhere (e.g., political campaign involvement, volunteerism). Moreover, we remind that that although this variable, especially when employed in a longitudinal analysis, can yield important insights it cannot fully account for the essential consideration of “how students engage” (Tillapaugh, 2019, p. 171), opening up space for future studies to understand these nuances more.

We also saw consistently positive student- and institutional-level findings associated with perceiving the fraternity as offering a place of belonging. We interpret this finding as a fraternity-specific instantiation of other studies which consider belonging as crucial to student success (e.g., Strayhorn, 2018). It is likely that fraternities fostering belonging may be pivotal in supporting students’ development, helping them stay committed while providing a space for processing experiences taking place in and outside of coursework. Additionally, eliciting belonging may help students’ in navigating productive moments of exchange across difference; acquiring the recognition that, though disagreements are inevitable, wholesale dismissals of others based on their identities and viewpoints are not (e.g., Shapses-Wertheim, 2014).

We also comment that, at the individual level, accountability had a pronounced positive effect while at the institutional level, shared social experiences had a negative effect—all important environments that these students are situated within that influence the outcome of openness to diversity. We see these findings as being in conversation with one another with respect to how students and communities approach their understanding of fraternities. Namely: are these spaces seen as locations of shared standards and learning or, primarily, as homophilic locations that may have negative effects on promoting openness? In particular, are these students potentially replicating ideals of whiteness or heteronormativity in these environments, thus limiting their ability to engage with others who come from different backgrounds than them? Research has demonstrated how members attempt to regulate member composition based on race (Joyce, 2018) and sexuality (Garcia & Duran, 2021; Giacalone, 2018), which bolsters our assumptions.

Considering these findings collectively offers several opportunities for deeper reflection on their underlying practical significance. Consistent with our conceptual approaches to this study, these results provide additional support to previous efforts (e.g., Pascarella et al., 1996, Whitt et al., 2001) suggesting that ODC can be developed among college students in association with individual- and institutional-level features, at least among a subset of men involved in fraternities. Building on this perspective, our findings lend another voice to ongoing conversations (e.g., Garcia & Shirley, 2019) calling for a closer look at fraternities beyond homogenous and, frankly, outmoded assumptions of such activities. Calls to abolish fraternities writ large because of anti-social and adverse behaviors overlook the potential significant contributions of these organizations to foster communities that link key forms of political and social engagement to important outcomes associated with openness to diversity. Furthermore, blunt and unnanced approaches to understanding fraternities as organizations could end up providing a sincere disservice to fraternity members, including those who involve themselves in fraternity life to exercise and reinforce aspects of their own developmental journeys.

Implications for Future Research, Theory and Practice

We now offer several implications for research, theory and practice. We hope that such perspectives, in conjunction with our quantitative approach, offer scholars and practitioners opportunities to continue generating insights into fraternity life and students’ developmental experiences within these social locations.

Research and Theory

To begin, we propose that future empirical and conceptual efforts could begin to take a more critical perspective of how scholars measure and frame ODC. Though our study provides robust insights from the current measure, future work might further consider the extent to which all students are invited to both meaningfully engage in and benefit from such openness. It is also important to consider how such concepts embedded in the measure such as “taking courses that challenge my beliefs and values” are being engaged with developmentally-appropriate supports; approaching the work with a recognition that not all ‘challenge’ is inherently productive. Such a consideration becomes especially important given the reality that students’ identity as “fraternity member”' is only one of several they may hold, and that students–especially students from minoritized racial/ethnicity and/or religious identities–often rely on fraternity houses and social organizations as spaces for processing and reflection on their engagement with campus racial/ethnic and/or religious majorities (Garcia & Shirley, 2019).

Furthermore, as quantitative research continues to take up a critical epistemological orientation (Wells & Stage, 2015), there is an opportunity to produce latent variables that address issues of power and privilege more specifically. For instance, with qualitative studies on diversity in fraternity life addressing the whiteness and colorblind racism that manifests in these spaces (e.g., Joyce, 2018; Joyce & Cawthorn, 2017), as well as how heteronormativity is present (Duran et al., 2021; Giacaolone, 2018), quantitative researchers should seek to be more precise in their conceptualizations of openness to diversity by interrogating concepts like whiteness, colorblindness, and heteronormativity.

Second, we wonder if there are opportunities for further efforts directed at unpacking and more fully understanding the four components of the fraternal brotherhood questionnaire, especially belonging and accountability. As Swanson (2022) has explored, the disruptions to college life brought on by the COVID-19 pandemic and its myriad adaptations has in ways challenged current understandings of belongingness among contemporary collegians. Subsequent work could build upon our findings with a critical eye toward investigating a perhaps more layered approach to understanding belonging and accountability; one that perhaps take more of an ecological eye (see Renn & Arnold, 2003) toward locating the fraternity member within the fraternity, the fraternity within the college, the college within its community, and so on while also accounting for the role of time (i.e., the chronosystem) in developmental processes.

Reflecting on our methods and design, we sustain the urging of Mayhew and Simonoff (2015) for the greater adoption of effect coding in quantitative studies that seek to de-center majoritized identities and provide robust estimates for all groups in a nested model. Interestingly, given our experiences and positionalities, this study may provide an example of the positive synergies that can be generated when secondary data is viewed from a critical quantitative lens that extends from research question through to variable coding and finally to interpretation (see Collins, 2019). Additionally, this study highlights the substantial benefits of developing and executing studies that are both longitudinal and incorporate HLM; such a strategy continues to offer important insights into how college affects students through substantiating claims on student development in association with both experiences and environments (see Mayhew et al., 2016).

Practice

Informed by our own experiences relative to fraternities, this study offers important implications for practice toward three audiences: sorority and fraternity life (SFL) practitioners, other student affairs practitioners, and concerned stakeholders. With respect to SFL practitioners, we see substantial value in highlighting those individual and institutional-level relationships between the FBQ constructs and the outcome and, certainly, political/social involvement and the outcome. Contrary to commonly-held perceptions, belonging, accountability, and social involvement are often hallmarks of the values espoused by historical white fraternities (e.g., Pi Kappa Phi Fraternity, 2014). Yet, values statements alone are clearly not enough to promote members’ ODC. SFL practitioners should design interventions that aim to promote conceptualizations of brotherhood in historically white social fraternities that advance peer accountability and belonging. Our findings at the individual and institutional-level indicate that fostering fraternal brotherhood focused on belonging may be particularly important in shaping members’ openness to diversity. Similarly, SFL practitioners should consider efforts to foster commitment to political and social involvement among fraternity men in historically white organizations. Genuine commitment to political and social involvement may help members break from the insularity identified by Garcia and Shirley (2019), and build authentic intergroup relationships with peers within and external to the fraternity and sorority community.

Our finding that institutional climates that foster conceptualizations of brotherhood based on shared social experience is relevant to all practitioners and other concerned stakeholders. As argued by Sweeney (2014), privileged fraternity men may view the collegiate party culture framed by brotherhood based on shared social experiences as a routine aspect of the undergraduate experience. Not only may this party culture relate to members’ alcohol use and perceptions of women (Sweeney, 2014), but it may dampen their ODC. If our sample across three national fraternities is reflective of the general population of members of historically white fraternities, it should give practitioners and other stakeholders pause that an overwhelming majority of our sample identified with privileged identities (i.e., over 86.1% of the sample reported at least one parent attended college, and 77.6% identified as white). If structural diversity is paramount to students’ engagement with peers around diversity (Shim & Perez, 2018), then intentionally addressing the lack of structural diversity within historically white fraternities may affect members’ openness to diversity and other outcomes. Attending to structural diversity may take the form of having existing members undergo campus-based trainings and workshops on how race, racism, and heteronormativity plays a role in new member recruitment, in addition to receiving this messaging from their national organizations. However, this implication relies upon individuals and not structures. From a critical epistemological stance, we also implore that national organizations engage in assessments of their policies and practices regarding recruitment that in turn, would trickle down and influence fraternity chapters.

Finally, as noted by others who have examined other outcomes of fraternity men (e.g., McCready et al., 2022), our findings indicate that monolithic strategies to foster ODC among men in historically white fraternities will likely be ineffective. As such, outright bans or prohibitions may do more harm than good, as these approaches eliminate historically white fraternities that promote ODC through their brotherhood, and commitment to political and social involvement, and other fraternities that center students with minoritized identities (e.g., historically Black fraternities). More nuanced approaches are needed at the individual and institutional levels to promote fraternity members’ openness to diversity.

Conclusion

Fraternities, both historically and presently, have long implicitly or explicitly perpetuated exclusive attitudes towards minoritized groups on college campuses (Gillon et al., 2019; Hughey, 2009). As has been articulated throughout this study, our positionalities as scholars of postsecondary education provided us the tools and the space to quantitatively examine aspects of this relationship toward providing insights, albeit imperfect ones, regarding the development of fraternity men along the vital dimension of ODC. In doing so, we strove to consider how these organizations can actually be vestibules where students obtain more inclusive and equitable attitudes. Developing such openness is essential toward student and social success in the twenty-first century given the increasing imperatives of interpersonal relationships across lines of difference; the contemporary labor market, by and large, both demands and requires this level of openness (e.g., Gupta, 2022).

Our findings demonstrate that such openness can be developed among fraternity men, both in conjunction with their fraternity experiences and their incorporation of fraternity life into their overall political and social engagement. We urge consideration of this important understanding as conversations of fraternity life will undoubtedly continue over the next decade and beyond. In fact, we hope our study further nuances the roles and responsibilities of fraternities as features of university life that hold potential: to be heterogeneous or homogeneous; to serve as catalysts of belonging and accountability or purely social spaces; to be drivers or hindrances to achieving prosocial developmental benefits. In short, fraternity life can serve as places and spaces comprising developing students seeking communities, not as monolithic bad actors to be summarily abolished (c.f., Cohen, 2021; Lautrup, 2020).