You Don’t Get to Say What I Believe, I Do: Provocative Encounters as Catalysts for Self-Authored Worldview Commitments During College

ABSTRACT

This paper examines the developmental trajectories of students’ self-authored worldview commitments over four years of college and the reasons they offered for making associated developmental gains. Through a longitudinal, mixed methods approach, we unearthed four distinctive developmental trajectories of a nationally-representative cohort of 9,470 students enrolled across 122 institutions. We found that provocative encounters with worldview diversity — those that appropriately challenged students to examine their own religious commitments — served as catalysts for growth during college. The varied contexts of those provocative encounters, both inside and outside the college classroom, are highlighted in students’ own words. Implications for research are discussed.

KEYWORDS:

• Higher education
• student development
• worldview diversity

As the public continues to scrutinize the value of higher education, questions about the efficacy of college-going are on the rise (e.g., Arum & Roska, 2011, 2014). This is in no small part due to a loss of public confidence in the formerly-unquestioned role of college as an experience designed to help students develop richer, more compelling, and more complex concepts of personal purpose and meaning (see Gilligan, 1982; Kerr, 2001; Kohlberg, 1981). As a society, we have gradually stripped away the humanistic and liberal inquiry that demands intense engagement with questions of value, ethics, and ideology, and increasingly accepted the narrow, neoliberal rationalization of education for the purpose of economic advantage alone (see, Giroux & Giroux, 2006; Rhoads, 2018; Rhoades & Slaughter, 2004). In so doing, the college experience has devolved — for many — from one which a student emerges self-actualized and prepared to reshape the cultural and economic systems around them, to one in which cultural and economic systems reshape the student to their own ends.

The present study is intended to disrupt this momentum as we seek to understand how college-going helps students develop dimensions of their inner selves, said to occur at the nexus of academic interest, competency, value identification, and emotional motivation (see, Daloz Parks, 2011) — what some refer to as purpose (see, Daloz Parks, 2011), others as voice (see, Gilligan, 1982), and still others as spirituality (see, Astin et al., 2011). The purpose of this study is to examine if and how college-going helps students develop one dimension of their inner selves, what we call self-authored worldview commitment. Specifically, we address not only whether four years of college-going influences students’ self-authored worldview commitments, but how students make meaning of their college-going experiences in light of their worldview commitments. Through a mixed methods approach, we address the complications associated with answering questions about the role colleges play in shaping students’ inner lives.

Theoretical orientation and literature review

Self-authored worldview commitment (SAWC) refers to “the capacity for students to script their own religious, spiritual, and worldview narratives through repeated exposure to and thoughtful reflection on encounters with diverse others” (Mayhew et al., 2020, p. 1; see, also Mayhew & Bryant, 2013; Mayhew, Rockenbach et al., 2016). Kegan (1994) and Baxter Magolda (2008) provide the theoretical foundation that established the idea of self-authored — the process of locating and centering one’s voice as a means for making meaning of information from external authorities. Within the context of religion, spirituality, and worldview the process of self-authoring helps individuals locate their own inner voices, through careful reflection on the ideas about these three domains that are presented by — sometimes competing — external authorities and sources. Specifically, self-authoring within this context requires that students reconcile competing perspectives, carefully consider differing points of view, and integrate multiple perspectives into their own outlook on life. Ultimately, this process serves as a supportive mechanism for helping individuals make meaning of the world around them and ultimately identify their sense of purpose.

Within this context, “worldview” is defined as a guiding life philosophy — one which may be based on a particular religious tradition, spiritual orientation, nonreligious perspective, or combination thereof (Mayhew et al., 2020). Riggers-Piehl et al. (2020) noted that worldview is distinctive from both religious identity and spirituality, with religious identity consisting of outward expressions of one’s faith and non-faith-based traditions and spirituality denoting an internal pursuit for meaning in life. In this sense, worldview may more closely align with spirituality epistemically, although expressions remain important, if not critical, for authentic engagements to occur.

Commitment as a hallmark of developmental sophistication finds its origins in the work of Perry (1970), Marcia (1980, 2002), Cross and Fhagen-Smith (2005), and Fassinger (2005) who collectively suggest that learning is often expressed through a series of commitments that have been thoughtfully considered as a result of exposure to and effortful engagement with developmental crises, often harbingered by encounters with diverse ideas, perspectives, and people. Importantly, individuals may arrive at developmental commitments through reflection-based engagements like those exemplified within self-authored processes or through less thoughtful retreats to familiar beliefs and stereotypes (Mayhew & Bryant, 2013). For us, commitments achieved through self-authored processes are more developmentally sophisticated than those achieved through other means: A person who commits to a religious, secular, or spiritual identity through a self-authored process is more developmentally sophisticated than one who commits due to sociocultural pressures, for example.

Taken together, SAWC is and should be a construct considered important, if not prioritized, for higher education. Ontologically, the construct assumes that humans have the capacity to identify and establish worldview commitments based on what they observe and experience. As such, these realities take an epistemic posture of quantifiability, in that each can be measured, assessed, experienced, and, to some degree, developed. The self-authored worldview commitment process is particularly valuable due to its potential to engage individuals in multiple learning processes before committing to their worldview: These include listening to people with different religious and secular points of view; thoughtfully considering other religious and nonreligious perspectives; reconciling competing religious and nonreligious perspectives; and integrating multiple points of view into one’s worldview.

Developing SAWC is an effortful process that involves thoughtful engagement with diverse worldview perspectives (Mayhew, Rockenbach et al., 2016; Mayhew et al., 2020). As a “prime context” (p. 73) for self-authoring, college offers students opportunities to engage in careful reflection of their own worldviews based on encounters with religiously-diverse ideas, mostly presented through religious literacy courses and/or engagement with campus community members holding a variety of faith and non-faith-based positions.

To date, only three studies have examined the extent to which college-going is related to the development of SAWC. In the first cross-sectional study of its kind, Mayhew and Bryant (2013) examined the relationship between the collegiate religious, spiritual, and ideological climate and SAWC. They found that SAWC was compromised if students experienced their campus as divisive or coercive but promoted when they engaged in co-curricular religious and spiritual activities. To nuance the latter finding in another cross-sectional study, Mayhew, Rockenbach et al. (2016) examined the specific relationship between interfaith engagement and SAWC and discovered that students who participated in formal and informal interfaith activities were more likely to report higher SAWC scores. More recently, in a longitudinal study of nationally-representative sample of over 7000 first-year students enrolled in one of 122 institutions, Mayhew et al. (2020) found that students who participated in two or more formal interfaith co-curricular experiences demonstrated significantly stronger gains in SAWC development during their first year in college. Importantly, all of these studies also found that the relationship between the significant predictors (i.e., climate and interfaith experiences) and SAWC varied depending on students’ self-identified worldviews.

Across these three studies, results suggested that students’ interfaith curricular and co- curricular experiences served as catalysts for SAWC. As noted by Mayhew, Rockenbach et al. (2016), “it appears as though exposure to difference in a structured context provides the support needed to face the disequilibrium engendered by diverse worldview encounters” (p. 372). The current study extends this previous work by adopting a mixed methods approach to examining the institutional conditions and educational practices that support students’ longitudinal development of self-authored worldview commitment over their four years in college.

The purpose of the current study was to examine if and how college-going influences students’ four-year SAWC development. Employing mixed methods analyses, we illuminate the conditions and practices that help students extract meaning from their experiences with diverse worldviews and ultimately establish a more critical and thoughtful commitment to their own. It is our hope that results will provide an empirical roadmap for educators interested in the religious and spiritual development of college students.

Conceptual framework

The Interfaith Learning and Development framework (ILD; Mayhew & Rockenbach, 2021) serves as the conceptual framework for the current study. The model is informed by Hurtado et al.’s (1998) foundational work on racial climate; specifically, it expands on their conceptualization of campus climate as a multifaceted construct comprised of historical, behavioral, psychological, organizational, and structural dimensions (Hurtado et al., 1998) by adding conceptual nuance based on outcomes similar to SAWC, such as ecumenical development (Mayhew, 2012), religious commitment (Astin et al., 2011), and pluralistic orientation (Rockenbach et al., 2017). In particular, the ILD framework is comprised of the specific institutional, relational, and disciplinary contexts in which students’ interfaith interactions are embedded on the college campus.

Accordingly, as depicted in Figure 1, the ILD framework includes the numerous institutional and individual components that shape student outcomes such as SAWC. First, students’ pre-college characteristics, including both student identity variables and pre-college exposure to interfaith experiences, are included within the model, as previous work has demonstrated the ways in which pre-college experiences and specific worldview narratives have differentially shaped students’ interfaith experiences and outcomes in college (e.g., Rockenbach et al., 2015).

Figure 1. Interfaith learning and development framework.

Second, the institutional context is comprised of four components: campus conditions, organizational behaviors, campus culture, and campus climate. Campus conditions include commonly referenced institutional characteristics such as religious affiliation, Carnegie classification, selectivity, size, and geographic region. Organizational behaviors reflect the educational practices sponsored by the institution. Campus culture is intended to reflect “deeply embedded” assumptions, values, and behavioral patterns (Peterson & Spencer, 1991, p. 142), such as a shared orientation toward appreciation of others’ worldviews, pluralism, and self-authored worldview commitment. Campus climate reflects Hurtado et al.’s (1998) multidimensional perspective of the “attitudes, behaviors, and standards […] concerning the level of respect [evident] for individual needs, abilities, and potential” (Rankin & Reason, 2008, p. 264).

The relational context is comprised of five dimensions of the student experience that reflect the ways in which religion mitigates the student-environment dynamic (Mayhew & Rockenbach, 2021). Those five dimensions include the degree to which the institution provides supportive spaces for students to explore worldview difference, insensitivity toward particular worldview identifications, coercive places where students feel forced to examine or change their worldview, unproductive environments where students feel silenced by worldview-based microaggressions or overt discriminatory practice, and provocative experiences that challenge students’ worldviews.

The disciplinary context reflects differences in students’ curricular experiences based on their selected academic majors. For example, students majoring in engineering may have significantly different curricular experiences than their peers majoring in the humanities. Moreover, students’ academic majors have been shown to influence a variety of learning and development outcomes, including those related to religion and spirituality (Bryant & Astin, 2008).

Finally, student experiences during college are captured at the center of the Interfaith Learning and Development framework. As suggested by Pascarella (1984), Weidman (1989), and Tinto (1994), four separate dimensions of student experiences are considered: formal academic, informal academic, formal social, and informal social behaviors.

Methods

This mixed methods study was conducted using data from the Interfaith Diversity Experiences and Attitudes Longitudinal Survey (IDEALS), which examines the impact of campus environments and undergraduate students’ collegiate experiences on educational outcomes including SAWC, pluralism orientation, and appreciative attitudes toward individuals of other worldviews. Using an integrated mixed-methods approach known as explanatory sequential design (Creswell & Plano Clark, 2011), we designed IDEALS as a project that made meaning of the quantitative results through subsequent examination of qualitative findings. Specifically, once institutional conditions and educational practices that predicted SAWC development were identified in the quantitative analyses, qualitative cases were examined to further elucidate the relationship between SAWC and the experiences students noted as critical for its development.

Data

Data were collected via three sources. The first involved collecting survey data from students and campus stakeholders representing institutions participating in the IDEALS study. We then matched survey data to institutional records found in publicly-available sources like IPEDS, the second source of data collected for this project. Finally, we made site visits to a series of campuses selected based on analyzed information from the first two of three total rounds of data collection. In the spirit of mixed methods approaches, the quantitative data collected for this study informed every aspect of the qualitative data collection strategy, from selecting the campus sites to data collection, control, analysis and interpretation.

Quantitative survey data

The IDEALS survey, a theoretically-derived and empirically validated measure of interfaith learning and development, was administered longitudinally at 122 colleges and universities across the United States. The sample was stratified by institution type, geographic region, selectivity, and affiliation (e.g., public, religiously affiliated, private nonsectarian). The survey included items capturing the institutional conditions and educational practices that influence students’ SAWC (see, Mayhew et al., 2020).

Administration of the survey was conducted at three timepoints. Time 1 responses were collected at the outset of students’ college experience (summer or fall of 2015); Time 2 responses were collected after students’ first year in college (spring or fall of 2016); and Time 3 responses were collected near the end of students’ fourth year in college (spring of 2019). At Time 1, 20,436 students from 122 institutions responded. At Time 2 and Time 3, response rates were 43.0% (122 institutions) and 36.0% (118 institutions), respectively.

The data used in this study included responses from students who participated in IDEALS at baseline (Time 1) and at least one of the follow up surveys (Time 2 and/or Time 3), resulting in an analytic sample of 9,470 students at 122 institutions. Given the longitudinal nature of this study, students with responses to only the baseline survey (Time 1) were excluded, as there was no follow up information from which to estimate trustworthy growth trajectories. Of the 9,470 students included in the analytic sample, 36.8% (n = 3,486) responded at all three timepoints, 39.2% (n = 3,708) responded at Time 1 and Time 2, and 24.0% (n = 2,276) responded at Time 1 and Time 3. Student characteristics and their institutional affiliations are reported in Table 1.

Table 1. Analytic sample characteristics for quantitative survey (N = 9470).

		n	%
Institutional religious affiliation	Catholic	842	8.9
	Evangelical Protestant	520	5.5
	Mainline Protestant	1744	18.4
	Private — Nonsectarian	2261	23.9
	Public	4103	43.3
Institutional Carnegie classification	Research Universities	3676	38.8
	Doctoral/Research Universities	505	5.3
	Master’s Colleges and Universities	2658	28.1
	Baccalaureate Colleges	2603	27.5
	Other	28	0.3
Worldview identification	Worldview majority	4926	52.0
	Worldview minority	1460	15.4
	Nonreligious	2670	28.2
	Another worldview	153	1.6
Gender identity	Woman	6517	68.8
	Man	2816	29.7
	Another gender identity	94	1.0
Race/ethnicity	African American/Black	465	4.9
	Asian/Pacific Islander	1308	13.8
	Latino/a	694	7.3
	White	5905	62.4
	Another Race	1085	11.5
Generation status	Continuing-generation student	7197	76.0
	First-generation student	2213	23.4
Political leaning	Very conservative	285	3.0
	Conservative	1512	16.0
	Moderate	3615	38.3
	Liberal	2868	30.4
	Very liberal	1148	12.2

Percentages may not sum to 100% due to missing data.

Qualitative case study data

In addition to the quantitative survey, qualitative case studies were conducted with students, faculty, and staff at a subset of 18 participating institutions that took part in the IDEALS survey. Importantly, results from the quantitative portion of the study informed the case study site selection, data collection, and analysis. Site selection took place following the second wave of survey data collection in spring/fall 2016. A subset of IDEALS institutions was identified to reflect the full range of institutional types nationally; in particular, the subset included institutions of diverse sizes and affiliations and from all regions of the United States. Institutions were also purposefully selected, based on change scores computed from Time 1 to Time 2 on IDEALS survey outcomes regarding students’ interfaith learning and development, to ensure that varying degrees of institutional change (no change, low change, and high change) on those outcomes were represented. The institutional characteristics of the sites selected for on-campus case studies are reported in Table 2.

Table 2. Institutional characteristics for qualitative case studies (N = 18).

		n	%
Institutional religious affiliation	Catholic	4	22.2
	Evangelical Protestant	1	5.6
	Mainline Protestant	3	16.7
	Private — Nonsectarian	5	27.8
	Public	5	27.8
Institutional enrollment	More than 12,000	4	22.2
	3,000–12,000	4	22.2
	Less than 3,000	10	55.6
Geographic region	Rocky Mountains	1	5.6
	Plains	3	16.7
	Great Lakes	3	16.7
	New England	1	5.6
	Mideast	2	11.1
	Southeast	5	27.8
	Southwest	2	11.1
	Far West	1	5.6

Percentages may not sum to 100% due to missing data.

Once the 18 case study sites were identified based on institutional characteristics and computed change scores on ILD outcomes, campus contacts connected the researchers with their students, faculty, and staff on campus. Each of the campus visits included focus groups with students, semi-structured interviews with faculty and staff, and observations of campus spaces and programs. This resulted in a total of 238 student participants and 223 faculty/staff participants across the 18 institutions. The student focus group protocol was designed to elicit rich descriptions of distinct institutional contexts and students’ experiences of worldview diversity therein.

For this study, student voices were of particular interest, as we sought to more deeply examine the relevant contexts and conditions in which they engaged in development of self-authored worldviews. Thus, transcripts generated from the student focus groups at each of the 18 site visits comprised the qualitative data for examination in this study.

Measurement

For the purposes of this study, variables were included in the quantitative models to reflect the various components of the Interfaith Learning and Development framework, with SAWC as the outcome of interest.

Outcome

The outcome of interest, students’ SAWC, was measured via a theoretically-derived and empirically-validated scale comprised of four items: (1) I have thoughtfully considered other religious and nonreligious perspectives before committing to my current worldview; (2) I have had to reconcile competing religious and nonreligious perspectives before committing to my current worldview; (3) I talked and listened to people with points of view different than my own before committing to my worldview; (4) I integrated multiple points of view into my existing worldview before committing to it. Reliability of the SAWC measure was assessed for the analytic sample at all three timepoints (Time 1: α = 0.82; Time 2: α = 0.84; Time 3: α = 0.82). To aid interpretability, SAWC composite scores at Time 1 were standardized, while scores at Times 2 and 3 were standardized using the mean and standard deviation from Time 1 raw scores. As a result, scores at Time 2 and Time 3 can be interpreted as change in students’ SAWC since Time 1 in effect size units.

Model covariates

First, student-level variables were included to capture students’ pre-college characteristics and pre-college exposure to interfaith experiences. The variables included to account for students’ pre-college characteristics were gender identity, sexual orientation, race/ethnicity, worldview identification, education generation status, high school GPA, and political leaning. Following Mayhew and Simonoff (2015), multi-categorical variables were effect coded in order to compare a value for one group (e.g., students who identify as holding a majority worldview) to the overall group mean of all students. As a result, rather than testing whether each group differs significantly from an arbitrary reference group, we tested whether all groups differed significantly from the overall sample mean.

Students’ pre-college interfaith experiences reflected the extent to which students engaged in interfaith activities prior to attending college. Thus, they were measured using a single composite score that captured the number of interfaith activities in which each student participated in the 12 months before beginning college.

The interfaith learning environment was comprised of the national, institutional, relational, and disciplinary contexts in which students’ interfaith experiences were embedded. The institutional context within the interfaith learning environment was captured by a number of institution-level variables. Campus conditions included variables for institutional control (i.e., public versus private), selectivity (i.e., according to Barron’s Profiles of American Colleges, 2015), and size of undergraduate population. Organizational behaviors reflected the number of religious, spiritual, or interfaith, programs, spaces, curricular opportunities, and diversity policies provided on campus, as reported by an institutional partner. Two campus climate indices (i.e., productive climate versus nonproductive climate for worldview diversity) and three campus culture indices (i.e., the ethos of an institution regarding its commitment to appreciative attitudes toward others’ worldviews, pluralism orientation, and self-authored worldview commitment) were also computed by averaging student-level perceptions to obtain an institution-level score.

The relational context similarly included institution-level composite variables reflecting students’ perception of space for support and spiritual expression, insensitivity on campus, coercion on campus, unproductive interworldview engagement, and provocative experiences with worldview diversity.

Finally, students’ disciplinary contexts were included as a variable indicating the self-reported academic discipline/cluster in which their primary major was situated. Like other multi-categorical variables included in the model, academic major was effect coded to allow for a comparison of each cluster of academic disciplines with the overall sample mean.

Together, the aforementioned contexts form the environment in which students’ SAWC occurred during college. Interfaith engagement behaviors were disaggregated by formal (i.e., explicitly interfaith) or informal (i.e., not explicitly interfaith) and academic (i.e., curricular) or social (i.e., co-curricular). The number of formal academic, informal academic, formal social, and informal social activities in which students engaged were computed. Then, students were categorized as having participated in no activities (reference group), at least one activity, or two or more activities.

Growth mixture modeling

The quantitative analyses in this study relied on growth mixture modeling (GMM; Muthén, 2004) to evaluate and explain change over time in college students’ self-authored worldview commitment. GMM is an empirical approach to parsing out unobserved heterogeneity in change over time (Bauer, 2007). While traditional growth models assume homogeneity across individual growth trajectories (i.e., that all individuals follow similar trajectories), GMM relaxes that assumption thereby allowing for the possibility that multiple, qualitatively distinct developmental trajectories exist in the data. GMM accomplishes this by using a categorical latent variable to represent the underlying subgroups — referred to as latent classes — represented in the data. Each latent class is then characterized by its own growth trajectory. Consequently, while traditional growth models produce one average growth estimate for an entire sample, GMM produces separate growth estimates for each subgroup inferred from the data (Shiyko et al., 2012). As noted by Diallo et al. (2017), GMM is “more naturally suited” for addressing research questions “anchored in a person-centered perspective of longitudinal development” (p. 166). Thus, GMM is especially well-suited to address the complex and multifaceted nature of students’ students’ growth in self-authored worldview commitment in college (Mayhew et al., 2020).

Our approach for conducting growth mixture modeling (GMM) included multiple steps. First, the process of determining the number of latent classes present in the model — a process referred to as class enumeration — was conducted using the unconditional model (i.e., prior to including covariates in the model) to prevent overextraction of classes (Nylund-Gibson & Masyn, 2016; Tofighi & Enders, 2008). The number of classes is determined by a combination of factors including statistical fit indices, substantive theoretical justification, and interpretability of the latent classes (Muthén, 2004; Nylund et al., 2007). Several comparative indicators of global model fit were used to guide the GMM class enumeration process, including the Bayesian Information Criterion (BIC), sample-size adjusted Bayesian Information Criterion (ABIC), and Akaike Information Criterion (AIC). For all aforementioned indicators, a lower value indicates a better fitting model. The Lo et al. (2001) Likelihood Ratio Test (LMR-LRT) was also used to compare nested models. The LMR-LRT is an alternative to the likelihood ratio test, whereby a significant value suggests that an estimated model with K classes is superior to a model with K-1 classes.

Once the number of classes inferred from the data were established, growth trajectories for each subgroup were examined. Each class, or subgroup, receives its own unique estimate of mean intercept and mean slope. The mean intercept reflects students’ SAWC at Time 1, the mean slope reflects the rate and direction of change in students’ SAWC at Times 2 and 3. For both the intercept and slope, a significant value indicates that the estimate is significantly different from zero. Of particular interest in this study were significant slopes, as they serve as an indicator of whether students’ SAWC increased (or declined) in any significant ways during their time in college. For each latent class, its growth trajectory was then modeled via an equation in which Outcome = Intercept + Slope(Time).

Finally, the variables comprising the interfaith learning and development framework (Mayhew et al., 2020) were incorporated into the GMM to evaluate which variables predicted students’ class membership. Consistent with the GMM framework, rather than assuming uniform influence across the sample, the influence of covariates was modeled separately for each latent class. The ability of covariates to predict class membership in GMMs was described via multinomial logistic regression parameters (Petras & Masyn, 2010). The resulting estimates were odds ratios, which indicated the likelihood of particular individuals having membership in one class compared to a reference class. To aid interpretation, odds ratios were also converted to probabilities (i.e., the percent likelihood that an individual would be a member of a class, as compared to the reference class). Variables were determined to promote students’ SAWC if they significantly predicted students’ membership in classes with positive growth trajectories.

Given the complex nature of our dataset, two additional analytic tools were applied throughout all steps of the GMM estimation process. First, full information maximum likelihood (FIML) estimation was used to establish robust latent growth trajectories in the presence of missing data. Thus, even for students with missing data at one of the follow up timepoints, we were able to estimate change over four years of college. Additionally, design-based multilevel structural equation modeling was applied to all steps, thereby adjusting standard errors to account for complex sampling features of the survey data (i.e., the nesting of students within 122 different institutions).

Case studies

Once the qualitative focus group data from the subset of 18 institutions were collected and transcribed, teams of trained researchers coded the transcripts using a priori codes (Saldaña, 2016) derived from the IDEALS survey items. This included coding transcripts for discussion of the ILD components, including SAWC. Multiple coders reviewed each transcript in order to ensure full coverage of the focus group content. Following this initial stage of a priori coding, a total of 378 student excerpts were identified as pertaining to SAWC.

Our quantitative analyses revealed that, among the ILD components tested, provocative encounters with worldview diversity most strongly predicted positive growth in students’ SAWC. Accordingly, the 378 qualitative excerpts initially coded as pertaining to SAWC were then further reduced to include only those that also addressed provocative encounters with worldview diversity — the predictor of students’ development in SAWC that was revealed in the quantitative analyses. This reduced dataset included 22 excerpts that were both coded for SAWC and described students’ provocative encounters with worldview diversity.

Finally, it is from those final 22 excepts that we were able to extract our findings regarding the ways in which provocative encounters with worldview diversity fostered students’ development of self-authored worldview commitment. Specifically, the excerpts were examined for common themes and contexts in which those developmental experiences took place.

Limitations

Several limitations are noteworthy regarding the conceptualization and execution of this study. Conceptually, arguments have been offered to suggest that the construct of self-authorship is problematic (Bowling, 2022), as it overemphasizes the role of the individual in the process of meaning-making (Perez, 2019), deprioritizes the influence of communities of belonging on the meaning-making process (Appiah, 2005; Hofer & Sinatra, 2010; Markus & Kitayama, 1991; Pizzolato et al., 2012; Renn et al., 2003), and obfuscates the interplay between religion and culture (Edwards, 2018a, 2018b; Tisdell, 2003; Tisdell & Tolliver, 2003).

The ILD framework is relatively new, needing more scholarly scrutiny for effective use across religious, spiritual, and secular narratives. As Bowling (2022) notes, “the framework would have greater onto-epistemological flexibility” (p. 9) if it reflected the internal and external influence of religious culture (Edwards, 2018b) and placed more emphasis on the interdependence of peoples’ meaning making processes. Certainly, humans have agency to evaluate and modify their worldview commitments, but there may also be aspects of worldview enmeshed with culture in ways that render identity irreducible beyond the communal level for some students. The conceptualization of religious identity might be expanded to account for religious, secular, and spiritual identities ascribed to minoritized communities based on external markers (Small, 2020), and descriptions of religion and spirituality might refer not just to the individual and society, but also to the transcendent as understood by the student.

The ILD framework presumes and this paper argues that a person with a SAWC is more developmentally sophisticated than one who conforms to a worldview socioculturally. Embedded is an assumption that cognitively working through ideas is more sophisticated than commitments made on the basis of culture or belonging, exemplifying a Western framing of religion as belief over praxis and elevation of rationality over other ways of knowing. The ILD framework remains an important model accounting for the contextual and interactional diversity relevant to interfaith learning but could be further pushed in these directions.

With regard to the quantitative portion of this study, the variables capturing student characteristics were intended to reflect the full spectrum of student identities in their response options. However, computational restrictions in the quantitative analyses (i.e., small cell sizes resulting in model nonconvergence) required that two of those variables be collapsed into binary indicators: gender identity and sexual orientation. Specifically, gender identity was limited to reflect only respondents who identified as men and women, whereas sexual orientation was collapsed into broader categorizations of LGBQ (i.e., lesbian, gay, bisexual, queer) and heterosexual. While not ideal for providing a full picture of the relationships of nuanced gender identities and sexual orientations to SAWC, this modification allowed those important identity variables (albeit overly-simplified) to be included in the analyses.

There are also several potential limitations with regard to the qualitative portion of this study. First, campus visits for the case studies were limited in duration. As a result, the participant pool was restricted to students whose schedules were compatible with the timing of the campus visit and scheduled focus groups. Additionally, the topic of research may have appealed more to particular students than others. This may have resulted in an overrepresentation of individuals inclined toward interfaith engagement and/or limited information gathered from students of particular identity groups. Thus, while efforts were made to ensure a diverse and representative sample of student participants, findings should be interpreted with caution.

Results

Change in self-authored worldview commitment

The class enumeration process included estimating and examining fit for multiple SAWC models, ranging from one-class to five-classes. When compared to the one-class model, which assumes that SAWC development over time is homogenous across all students, the statistical fit indices (summarized in Table 3) provided consistent evidence that the multi-class models were a better fit to the data. Specifically, the two-, three-, four-, and even five-class models had lower AIC, BIC, and ABIC values; additionally, the VLMR and LMR-LRT tests were significant. As a result, it was concluded that multiple trajectories of students’ SAWC did in fact exist in the data, and those trajectories should be modeled separately (Shiyko et al., 2012). Based on the statistical fit indices, both the four-class and five-class SAWC models emerged as plausible representations of the data. As noted by Nylund et al. (2007) and others, interpretability of the resulting classes in light of existing theory should also weigh heavily in the class enumeration process. Thus, based on the interpretability of the qualitatively distinct trajectories represented by the four-class model and the sufficiently large class sizes (suggesting that the trajectories represented substantively meaningful differences in the population as opposed to sample-specific error), the 4-class model of students’ SAWC was determined to be the best-fitting model.

Table 3. Statistical fit indices for class enumeration.

Number of classes	Log likelihood	AIC	BIC	ABIC	Smallest proportion	VLMR	VLMR sig	LMR- LRT	LMR- LRT sig
1	−28,950.267	57,916.535	57,973.782	57,948.359	100.00%	-	-	-	-
2	−28,793.900	57,609.799	57,688.514	57,653.558	31.99%	−28,950.267	0.000	301.750	0.000
3	−28,720.982	57,469.964	57,570.146	57,525.656	19.63%	−28,793.900	0.000	140.713	0.000
4	−28,657.778	57,349.556	57,471.206	57,417.182	8.74%	−28,720.982	0.001	121.968	0.001
5	−28,606.454	57,252.907	57,396.025	57,332.468	6.27%	−28,657.778	0.032	99.043	0.036

AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; ABIC = sample-size adjusted Bayesian Information Criterion; VLMR = Vuong, Lo, Mendell, Rubin Likelihood Ratio Test; LMR-LRT = Lo, Mendell, Rubin Likelihood Ratio Test. For AIC, BIC, ABIC: lower value indicates a better fitting model. For VLMR and LMR-LRT: significant value suggests that estimated model with K classes is superior to model with K-1 classes.

The final 4-class model revealed 4 distinct patterns of students’ change in SAWC over time. Table 4 reports starting values (i.e., intercept) and average change over time (i.e., slope) by class. As seen in Figure 2, students in class 1 (58% of the sample) started at Time 1 with relatively high SAWC with no significant change in SAWC at subsequent timepoints; students in class 2 (9%) started at Time 1 with relatively low SAWC, as compared to some of their peers, and experienced high growth in SAWC across Times 2 and 3; students in class 3 (9%) started at Time 1 with relatively high SAWC, and experienced a significant decline in SAWC over time; students in class 4 (25%) started at Time 1 with relatively low SAWC, and experienced very small (but still significant) growth in SAWC over time.

Figure 2. Standardized SAWC growth trajectories by latent class over time.

Table 4. Standardized intercept and slope by latent class.

Class (% of sample)	Intercept (Mean)	Intercept (Sig)	Slope (Mean)	Slope (Sig)
Class 1 (58%)	0.522	0.000	−0.002	0.857
Class 2 (9%)	−1.090	0.000	0.876	0.000
Class 3 (9%)	0.427	0.000	−0.718	0.000
Class 4 (25%)	−0.843	0.000	0.069	0.007

Bold values are significant at p < .01.

Predictors of self-authored worldview commitment

Predictors were then incorporated into the 4-class model to determine which of the ILD variables were likely to promote growth in students’ self-authored worldview commitment during college. When comparing class 2 (the high growth class) with reference class 4 (students who started with SAWC similar to class 2, but experienced much less growth over time), 3 significant predictors of class membership emerged: gender (OR = 0.459, p = .034), worldview identification (OR = 0.001, p < .001), and provocative encounters with worldview diversity (OR = 2.920, p = .015). Results for all significant and nonsignificant predictors are reported in Table 5.

Table 5. Predictors of class membership for SAWC outcome.

	Est.Odds Ratio	Est. Probability	SE	Sig.
Institutional campus climate
Positive campus climate	1.853	64.9%	0.615	0.165
Negative campus climate	1.655	62.3%	0.668	0.327
Institutional campus culture
Appreciative attitudes campus culture	0.758	43.1%	0.196	0.217
Pluralism campus culture	0.795	44.3%	0.236	0.385
Self-authored worldview commitment campus culture	6.410	86.5%	3.219	0.093
Institutional campus conditions
Undergraduate enrollment	1.835	64.7%	0.989	0.398
Selectivity	0.830	45.4%	0.164	0.298
Control (1 = Private; 0 = Public)	4.712	82.5%	5.012	0.459
Institutional campus behaviors
Total religious, spiritual, or interfaith programs provided on campus	1.157	53.6%	0.305	0.608
Total religious, spiritual, or interfaith spaces provided on campus	1.169	53.9%	0.377	0.654
Total religious, spiritual, or interfaith curricular opportunities provided on campus	0.871	46.6%	0.276	0.639
Total religious, spiritual, or interfaith diversity policies provided on campus	0.811	44.8%	0.230	0.412
Student identity characteristics and pre-college behaviors
Precollege interfaith activities	1.243	55.4%	0.340	0.475
Gender (1 = Female; 0 = Male)	0.459	31.5%	0.255	0.034^*
Sexual Orientation (1 = LGBTQ+; 0 = Heterosexual)	8.821	89.8%	12.478	0.531
Generation status (1 = First-generation; 0 = Continuing generation)	0.938	48.4%	0.432	0.885
High school GPA	1.118	52.8%	0.260	0.649
Race/ethnicity
African American/Black	0.998	49.9%	0.596	0.998
Asian/Pacific Islander	1.264	55.8%	0.480	0.582
Latino	1.067	51.6%	0.957	0.944
White	0.938	48.4%	0.394	0.875
Other race	0.792	44.2%	0.531	0.695
Worldview identification
Worldview majority	8.957	90.0%	19.154	0.678
Worldview minority	8.956	90.0%	18.963	0.675
Nonreligious worldview	9.733	90.7%	20.524	0.670
Another worldview	0.001	0.1%	0.009	0.000^**
Political leaning
Very conservative	0.759	43.1%	0.963	0.802
Conservative	1.258	55.7%	0.734	0.725
Moderate	0.859	46.2%	0.333	0.671
Liberal	1.220	55.0%	0.723	0.761
Disciplinary context
Arts, Humanities, or Religion	0.681	40.5%	0.340	0.348
Social Science or Education	0.756	43.1%	0.342	0.475
Health	2.438	70.9%	1.311	0.273
Science, Math, or Engineering	1.720	63.2%	0.731	0.325
Business	0.566	36.1%	0.293	0.139
Undecided or another major	0.819	45.0%	0.242	0.455
Relational context
Space for support and spiritual expression	1.332	57.1%	0.311	0.286
Insensitivity on campus	1.156	53.6%	0.384	0.684
Coercion on campus	0.945	48.6%	0.355	0.876
Negative interworldview engagement	1.821	64.6%	0.509	0.107
Provocative encounters with worldview diversity	2.920	74.5%	0.790	0.015^*
Student interfaith behaviors
Formal academic engagement
At least one activity (no activities as reference)	1.097	52.3%	0.945	0.918
Two or more activities (no activities as reference)	1.639	62.1%	1.411	0.651
Informal academic engagement
At least one activity (no activities as reference)	1.182	54.2%	0.619	0.768
Two or more activities (no activities as reference)	1.349	57.4%	0.562	0.535
Formal social engagement
At least one activity (no activities as reference)	0.626	38.5%	0.516	0.468
Two or more activities (no activities as reference)	0.929	48.2%	0.685	0.918
Informal social engagement
At least one activity (no activities as reference)	2.554	71.9%	4.107	0.705
Two or more activities (no activities as reference)	2.655	72.6%	3.851	0.667

Reported estimates and significance pertain to predictors of membership in class 4 (high SAWC growth) with class 2 (low SAWC growth) as reference group

Probability computed as Odds Ratio/(1 + Odds Ratio)

*p < .05, ** p < .001

With regard to student identities, gender and worldview identification predicted SAWC growth over time. First, students identifying as female had only a 32% likelihood of being in the high-growth group, as compared to the low-growth group. This means that students identifying as female were less likely than their male counterparts to experience high growth in SAWC during college. Second, students identifying as holding “another worldview” had less than 1% likelihood of being in the high-growth group, as compared to the low-growth group. That is, students who identified holding a worldview other than a majority, minority, or nonreligious worldview were highly unlikely to experience high growth in SAWC.

With regard to institutional conditions and practices, students who reported experiencing more provocative encounters with worldview diversity had a 75% likelihood of being in the high-growth group, as compared to the low-growth group. Provocative encounters are those that present students with opportunities for challenge in their interactions with diverse others (Rockenbach & Mayhew, 2014). Such encounters extend beyond experiences that can be easily classified as productive or nonproductive (Winkler et al., 2020), and though they may be experienced as challenging in the moment, they frequently spur development as students make meaning of them. Thus, students who experienced encounters which challenged them to evaluate their assumptions of others and to critically reflect on their own worldviews were more likely to experience high growth in SAWC over time.

Illuminating the role of provocative encounters

Qualitative analyses illuminated the various contexts in which students’ provocative encounters with worldview diversity led to SAWC. The student focus group excerpts that were both relevant to SAWC and that reflected students’ provocative encounters with worldview diversity were examined. These 22 excerpts detailed instances in which students’ experiences with provocative encounters led them to engage in the process of developing their own “internal voice” (Baxter Magolda, 2008, p. 279) with regard to religious/spiritual identities and thus propelled development of self-authored worldviews. Those experiences occurred both inside and outside the classroom; particular contexts mentioned included intergroup dialogue, coursework (e.g., creative writing, philosophy of science, and world religion philosophy classes), peer-centered encounters, professor encounters, study abroad, and housing, among others. Additionally, the role of such provocative encounters in fostering SAWC was reported by students of all religious and nonreligious identities (e.g., agnostic, deist, Evangelical Christian, Protestant Christian, Catholic). Profiles for the institutions reflected in the excerpts below are reported in Table 6.

Table 6. Profiles for featured case study institutions.

Institution Name	Geographic Region	Religious Affiliation	Enrollment
Ohio River University	Southeast	Catholic	Less than 3,000
Upstate College	Mideast	Private — Nonsectarian	Less than 3,000
River University	Southeast	Private — Nonsectarian	3,000–12,000
Southwest University	Southwest	Evangelical Protestant	Less than 3,000
Prairie College	Plains	Catholic	Less than 3,000

Institution names are pseudonyms.

Provocative encounters inside the classroom

A number of students described the ways in which provocative experiences inside the classroom led to SAWC. For example, a Catholic man shared the ways in which he was challenged to more deeply examine and commit to his existing religious worldview as a result of his college coursework. He explained:

I’m a second-year theology and history major [and] I was Catholic when I came here […] I was strongly influenced by the theology classes that I took here that taught me to think about God in new ways, and I was able to read a lot more religious material that dived deeper into the stuff that I wanted to learn about. Now so far, I’m a bit more solid in the faith that I was raised in, and I think I understand it to a wider degree as well. (Lucas, Catholic, Ohio River University)

Another student similarly noted the ways in which his nonreligious worldview was challenged, and as a result strengthened, from engagement in his philosophy of science course:

My worldview was, and still is in some sense, very scientific, very rationalistic. I’m agnostic, I don’t put a lot of faith in faith and I put a lot stock in science. So, the philosophy of science class really pushed me on that, really introduced some doubts that I still acknowledge and still come to grips with and incorporate [into] my worldview. It very much shook my faith in science as an unbeatable, rigorous, and solid way to understand the universe. (Fitz, agnostic, Upstate College)

The process by which experiences engaging with faculty in the classroom promoted SAWC for students was further explicated by another student. She emphasized the provocative role of faculty in challenging her to critically evaluate her own commitments:

When you’re in a classroom and someone of authority is saying something, you kind of have to think about that. It’s like, in those situations, hearing beliefs that are different from mine, I’ll definitely have, like I was saying, challenge what I believe, and I have to really take that and think about that. It’s like, okay, do I believe this? If I don’t believe this, why don’t I? Things like that. (Milo, Protestant Christian, River University)

All of these encounters, occurring in a variety of classroom contexts, led students to critically reflect on, and ultimately strengthen, their existing worldviews.

Provocative encounters outside the classroom

Other students described provocative encounters outside the classroom—in housing, study abroad, and other co-curricular contexts. For example, one student described his experience working for the Chabad house on campus his freshman year. He initially sought out the opportunity as a way to better connect with his Jewish heritage, and instead found that it provided the catalyst to more deeply evaluate his self-authored beliefs. While he decided not to return to the Chabad house the following year, he described the way in which he began developing his own internal voice with regard to worldview:

I realized that [I was] forcing myself to do something or live in an image that was imposed upon me because that’s just life and that’s family and that’s how they want you to live. You can try it but it just wasn’t something for me and it gave me a kind of reality check to say, I need to stand here and think about what’s important to me and what I value and what kind of religious ideals, if any, that I value and take that into account instead of sticking myself into situations where I’m hoping that somebody will just give me all the answers and lead the way. (Adonis, deist, River University)

Other students shared the ways in which exposure to diverse worldviews among their college peers served as the catalyst for both broadening their perspective and deepening their own commitments. Regarding his experience as a Protestant man, one student explained:

I’m from here [in the southwest] and so, high school, I feel like my view of people and my mindset was very small and confined. And then when I got to college, it was just completely broadened because, like everybody has said, the interactions you’ve had with different people. And, definitely, my high school was not very Christian at all, and so having a place where there’s lots of Christians that come together, that has definitely expanded my religious outlook on life. I’ve become a lot more open minded and, also, going away from my parents’ point of view and forming my own has broadened my worldview. (Joe, Protestant Christian, Southwest University)

These narratives demonstrated the many ways in which students experienced challenging, or provocative, encounters that forced them to reexamine their existing worldview. Then, through the process of making meaning of such encounters, they “really hammered out” their own beliefs (Samantha, Catholic, Prairie College), distinct from those of their family, peers, and other external influences. Regardless of the particular context in which they occurred, those provocative encounters with worldview diversity ultimately propelled students to strengthen the commitment with which they held their existing worldviews.

Discussion

Do four years of college make a difference in the religious and spiritual lives of colleges students? For a large minority, it absolutely does. This nationally representative, mixed methods study provides the first and nuanced look at students’ self-authored worldview development from their first through fourth years in college. Although this study revealed that the development of self-authored worldview commitment occurs during four years of college, growth trajectories differed, with only 9% of the entire sample reporting steady gains from fall of 2015 to spring of 2016 to spring of 2019. This evidence substantiates self-authored worldview commitment as a developmental outcome for consideration ontologically (e.g., that is exists), epistemically (e.g., that is can be measured), and axiologically (e.g., that there is value in its study). That said, the small sample of students who demonstrated steady growth over time did so based primarily on the provocative encounters they experienced during college.

The role of provocative encounters in fostering self-authored worldview commitment was highlighted throughout the student narratives, which detailed instances where students engaged in the process of developing their own “internal voice” (Baxter Magolda, 2008) with regard to religious, spiritual, and worldview identities following a challenge to their unexamined commitments. Importantly, those experiences occurred both inside and outside the classroom and were reported by students of all religious and nonreligious identities (e.g., agnostic, deist, Evangelical Christian, Protestant Christian, Catholic). Ultimately, students shared similar stories of development, using the language of sentiment to describe moving through the process of echoing their parents’ beliefs to making them their own. As one student put it, she “really hammered out” her own beliefs, distinct from those of their family, peers, and other external influences.

Finding and centering one’s voice in the journey toward commitment is closely linked to provocative moments that disrupt assumptions and catalyze new ways of understanding one’s own worldview as well as the worldviews of others. Provocative moments spur a change- inducing disequilibrium powerful enough to transform conceptions of the self (Pizzolato, 2005). Our evidence shows that provocation is multifaceted and takes place within and beyond the classroom and formal curriculum. At the heart of such provocation are moments that challenge students to rethink their assumptions and encourage them to develop more favorable attitudes about other worldviews. Provocation can also turn a student’s gaze inward as they hear critical comments that cause them to question their own worldview, talk to others from their own worldview with whom they disagree, and have discussions that make them feel like they don’t know enough about their own worldview.

To be sure, provocative encounters have the capacity to unsettle, inspire exploration, and activate reckoning with one’s fallibility (Rine, 2012). There are critical distinctions, though, between developmentally fruitful provocation and destructive provocation that traumatizes and derails growth (Taylor & Baker, 2019). Previous findings demonstrated that students who engage in silencing, guarded, hostile, and unresolved exchanges across worldview differences do not make gains in their self-authored worldview commitments (see, Mayhew & Bryant, 2013). Results from this endeavor suggest that productive provocation, that which creates dissonance but not distress, is critical for developmental movement toward more sophisticated self-authorship, especially in the context of worldview commitments.

How might educators seeking to inspire students’ self-authoring around religion, spirituality, and worldview integrate provocative encounters into their efforts? Promisingly, our findings suggest that provocative moments are not context-specific—they happen in various spaces across campus, from the formal (classroom) to the informal (peer-to-peer socializing).

Key to conversations that are provocative are opportunities to balance outward and inward explorations of worldview. Students need to unpack and wrestle with the stereotypes and unchecked assumptions they hold regarding others’ worldviews, as well as reflect thoughtfully (perhaps courageously) on what they know and don’t know about their own worldview. As part of this process—whether it unfolds in class or in conjunction with co-curricular programming—students should engage those who share their worldview identity (to expose them to intergroup heterogeneity) and those who do not (to expose them to different approaches to meaning-making).

As this study marks an initial step toward articulating the developmental significance of provocative encounters in self-authored worldview commitment, many questions remain for future research. For example, to what extent does context matter? How do provocative encounters that surface in courses or in exchanges with faculty differ from those that take place in students’ social lives as they interact with peers and friends? How might college students who develop self-authored worldview commitments be situated to influence their post-college communities beyond college: their social networks, workplaces, and families? Hypotheses worth exploring, self-authored people may be better prepared to challenge the prejudices of their friends, encourage critical reflection and religious, secular, and spiritual autonomy/freedom among their children, and/or they may befriend/respect people on the religious margins in their communities and workplaces because they are practiced at dismantling stereotypes and assumptions. We anticipate that continued explorations of these questions and hypotheses from multiple paradigmatic frames and drawing on diverse research methodologies will be essential to advancing theoretical and empirical insights.

In addition to these mechanisms, a number of environmental conditions may be important for developmental movement to take place. Recently, the Interfaith Learning and Development framework (Mayhew & Rockenbach, 2021) has been offered as a framework for identifying the institutional conditions, educational practices, and student experiences posited for helping students grow along a number of religious and spiritual dimensions, including self-authored worldview commitment. Pertaining to the model, Mayhew and Rockenbach (2021) argue that “learning and development is most likely to occur when the culture and climate of the institution are optimized for success for students given their diverse religious, racialized, and gendered presentations and expressions” (p. 6). We can expect intersectional oppressions within systems of power (Collins, 2019; Crenshaw, 1989, 1991; Singh, 2015) to limit equitable access to religious and spiritual resources that inspire the development of self-authored worldview commitment. We need continued research, especially work framed genuinely by intersectional lenses (Harris & Patton, 2019), that shines a light on how religious and spiritual initiatives and spaces perpetuate the exclusion of students who experience multiple, interlocking forms of marginalization. Identifying, disrupting, and transforming oppressions within religious and spiritual campus life will be critical for advancing justice and inspiring student development.

From an empirical perspective, methodological nuance matters. To address the many complexities that come with the study of religion, spirituality, and worldview, scholars may need to complicate their approach to this important line of inquiry. For example, growth mixture modeling was especially appropriate for tracking student learning over three time points: In the absence of this analytic technique, findings may have been obfuscated to the degree that development may not have been detected at all. As another example, the mixed methods nature of the study enabled us to center the voices of students in making meaning of their own development — a vital methodological component for unearthing the mechanisms, language, and frameworks students use to describe their own experiences.

Speaking of students, we would be remiss without discussing the importance of collaboration in executing research like this. Not only did thousands of students give their time to help us accomplish this work, we were humbled by the hundreds of campus liaisons who helped champion this research. Of course, we also were fortunate to find foundations to support this multi-year, longitudinal, nationally-representative, mixed methods study.

Conclusion

Higher education is at a crossroads, as more citizens weigh in on its importance to American life. As educators of the whole student, higher education stakeholders play a critical role in ensuring that college remains a place where students of different religious, secular, and spiritual identities are celebrated and not demonized, embraced and not abandoned, so that college remains a place that helps students develop more enriched inner lives. Provocative encounters may be challenging, but they are perhaps the most valuable experiences of all. Our findings showed that these encounters make a powerful impression on self-authored worldview commitment because such experiences push students to reevaluate their assumptions and to critically reflect on and question their own worldviews (Mayhew, Hoggan et al., 2016). Though the process of provocation may be unsettling or even produce discomfort, students who routinely experience provocative moments develop worldview commitments that are uniquely their own and that provide a solid foundation for living authentically and engaging others productively in a diverse society.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by funders including the Andrew W. Mellon Foundation, John E. Fetzer Institute, and Julian Grace Foundation.