The university bundle: Unpacking the sources of undergraduate moral socialization

ABSTRACT

Recent evidence suggests that higher education promotes moral attitudes typical of the progressive left. What aspects of the university experience contribute to this moral change? We conduct an exploratory analysis unpacking how curricular content and peer networks—two aspects of the ‘bundle’ of social influences that occur in university settings—might affect moral attitudes. Using two waves of data from students at a Canadian university (n = 232), we find some evidence that exposure to content related to social justice and involvement in left-leaning university peer circles can promote more individualistic forms of morality over ‘binding’ moral concerns for traditional social order, and heighten a more absolutist endorsement of social justice. Taken together, the university experience appears to be morally formative, but not uniformly so: moral change is shaped by a combination of factors implicating both formal and informal aspects of university life which students experience at varying rates.

KEYWORDS:

• Moral attitudes
• higher education
• socialization

For some time now, sociologists have approached higher education primarily through the prism of stratification, exploring its role in reproducing social inequality (Domina et al., 2017; Hout, 2012). More recent work breaks from this course by returning inquiry to higher education as a moral institution, focusing on expressive dimensions, and particularly student development (Baker, 2014; Broćić & Miles, 2021; Guhin & Klett, 2022; Monaghan, 2020). Monaghan (2020) evocatively likens the university to a Cathedral, where faculty serve as secularized priests conferring credentials that bestow moral virtue upon their holders. Normative commitments endorsed by universities and internalized by students are often described as universalistic, combining cosmopolitanism with scientism, and championing meritocratic worth (Baker, 2014; Monaghan, 2020). But some scholars warn that an ideological undercurrent risks compromising these commitments. Horowitz et al. (2018), for instance, argue that some academics in the social sciences are engaged in a ‘sacred mission’ wedded to progressive orthodoxy where pursuit of social justice advocacy overrides the integrity of scientific inquiry and ostracises dissenting viewpoints (c.f., Smith, 2014). The moral project of higher education, seen this way, becomes a politicized one.

These perceptions form the backdrop of growing animus towards higher education (Mitchell, 2019). A recent Gallup poll finds that American confidence in higher education has reached a historic low, dropping 21% points since 2015 (Brenan, 2023). While confidence has declined across the board, the greatest drop is found among the political right where perceptions that universities are bastions of liberal ‘indoctrination’ have become an important concern (Newport & Busteed, 2017). Popular accounts (mostly on the right) attribute the rise of progressive moralism in public culture to influence trickling down from university professors (Pluckrose and Lindsay 2020; Rufo, 2023; Shapiro, 2010). While these critiques are not new, recent years have seen them gain wider resonance through frames such as ‘cancel culture’, ‘wokeness’, and the like (Vogels, 2022; Zúquete, 2023). What is more, they appear to be gaining traction outside the American context, crystallizing into a broader critique of the ‘new class’ composed of intellectuals, experts, and bureaucrats across Western democracies (Finlayson, 2021; Lind, 2020).

Scholars weighing in on these debates partly assuage concerns by minimizing the role of college in effecting student change, arguing that ideological leanings mostly crystallize prior to enrollment (C. Campbell & Horowitz, 2016; Elchardus & Spruyt, 2009; Gross, 2013). However, recent research indicates that many students’ moral beliefs shift in a progressive direction during university, suggesting that higher education is indeed an important site for moral change (Broćić & Miles, 2021). Be that as it may, it remains unclear which aspects of the collegiate experience are responsible for producing change and in what ways. Existing research points in different directions, with some accounts emphasizing formal sources such as administration, faculty, and curriculum, while others point to the more informal networks making up student subcultures (Broćić & Miles, 2021; Logan et al., 2021; Rauf, 2021; Woessner & Kelly-Woessner, 2020). Determining how—and how much—different aspects of the ‘university bundle’ shift students’ moral beliefs has direct implications for partisan debates, suggesting a need to give these questions careful empirical scrutiny. Our study aims to bring greater clarity to these debates by conducting an exploratory analysis of undergraduate socialization with an eye to identifying sources of change and cataloguing their relative importance.

Moral dimensions of change

Scholars approach moral development in higher education in different ways, variously examining moral reasoning, altruism, and pro-social behaviour among other dimensions (Mayhew et al., 2016). While many of these conceptions are regarded as unambiguously ‘moral’, we adopt a more pluralistic approach characteristic of research on moral culture where differences in what is regarded as right and wrong become a focal point of concern (Vaisey & Miles, 2014). This approach has been influential for research on the moral dimensions of political conflict, where scholars evaluate the ‘moral divide’ hypothesis positing that political conflict rests on deeper disagreements over ‘first principle’ beliefs pertaining to proper conceptions of right and wrong (Haidt, 2012; Miles & Vaisey, 2015). Moral differences arguably raise the stakes of political conflict, making division more acrimonious, opposition more intractable, and feeding into a wider ‘culture war’ (Hunter, 1991; Skitka et al., 2021).

Moral Foundations Theory (MFT) provides an influential account in this line of work, advancing a pluralistic theory of morality that differentiates individualizing from binding moral foundations, both of which ‘suppress or regulate selfishness and make cooperative social life possible’, but in different ways (Haidt, 2012, p. 360). Individualizing foundations refer to moral concerns for ensuring fair and equal outcomes, as well as harm-reduction for vulnerable populations, exhibiting universalistic criteria of judgment where protection of individuals is prioritized. By contrast, binding foundations are more particularistic, oriented to protection of traditional social order, and comprised of moral concern for group loyalty, authority, and sanctity. A large body of research, mostly using data from the United States, documents variation in the relative importance assigned to these foundations along political ideology: liberals, and those on the political left, tend to prioritize individualizing foundations over binding foundations, while conservatives, and those on the political right, are more likely to endorse both more equally (Jesse et al., 2009; Miles & Vaisey, 2015). These different moral grammars arguably drive affective polarization (Jung & Clifford, forthcoming) and help explain opposing positions on ‘culture war’ issues concerning abortion, immigration, or same-sex marriage net of partisanship (Koleva et al., 2012), making their relation to higher education important to consider. It is in these cultural rather than economic terms that we understand moral profiles typical of those self-identifying as ‘liberals’ or the ‘left’, i.e., prioritizing the pursuit of greater care, harm-reduction, and fairness for individuals, over concerns for preserving moral norms prescribed by traditional social institutions. While this is mostly consistent with the unidimensional correlation of moral foundations with both ‘liberal-conservative’ and ‘left-right’ spectrums found in North American contexts, it is important to note that there are exceptions to this patterning both within and outside this context (see Kivikangas et al., 2021).¹

The moral dimensions of political conflict over higher education are not exhausted by changing content of moral concerns, however, but can implicate forms of orthodoxy as well. Our previous work found that while many students become more ‘liberal’ over the course of their studies, increasingly prioritizing individualistic over binding concerns, they also expressed changing forms of commitment, departing from typical relativism by becoming more assured that their views reflected the definite moral truth (Broćić & Miles, 2021). While these findings align with accounts of an increasingly absolutist endorsement of social justice on campuses, our coarse measurements rendered any such conclusion speculative (Zúquete, 2023). We follow up on these initial findings by considering ‘left-wing authoritarianism’ (LWA) – a highly-validated construct of authoritarian tendencies on the left, capturing moral absolutism applied to social justice concerns, where conservatives are regarded as inherently immoral, and aggression is directed towards established traditional power structures (Costello et al., 2020).

LWA usefully captures some of the trends toward more rigid, morally absolute views observed by scholars and social commentators on university campuses, particularly in the humanities, arts, and social sciences (Zúquete, 2023). For example, Horowitz et al. (2018) observed that some sociologists expressed disgust, anger, and incredulity when confronted with narratives contrary to their preferred progressive explanations for phenomena such as poverty or gendered occupational choices, a reaction characteristic of LWA’s aversion to dissenting views. Others have argued that social scientists endorse a set of ‘sacred values’ that delineate what sorts of explanations about social groups are permissible (Clark & Winegard, 2020; Honeycutt & Jussim, 2020). In line with LWA, sacred values lend themselves to absolutism and often evoke spirited defense when threatened, which may help to explain willingness to discriminate against those who oppose their moral/political views (Honeycutt & Freberg, 2017; Horowitz et al., 2018; Inbar & Lammers, 2012; Peters et al., 2020). By capturing left-leaning variants of moral absolutism, LWA provides a useful instrument for empirically scrutinizing claims about college campuses (Zúquete, 2023). We therefore supplement our analysis of moral foundations with an examination of LWA. The two are not synonymous, as one can adopt more liberal/left moral concerns, without becoming more absolutist in these commitments.

Higher education and moral change

What is the role of higher education in effecting change in these moralized attitudes? While research on moral foundations, and especially LWA, remains scarce, we can turn to the wider literature on college effects for insight. That the highly educated hold liberal cultural attitudes is one of social science’s most well-documented findings (Gross, 2013; Ladd & Martin Lipset, 1976; Newcomb, 1943). However, the reason for this relationship continues to be intensely scrutinized. Some scholars point to the role of self-selection, arguing that if the college-educated are more liberal, it is because they were more liberal to begin with (C. Campbell & Horowitz, 2016; Elchardus & Spruyt, 2009; Gross, 2013). In this view, any change that occurs during university is minimal, and the collegiate experience is mostly an occasion for self-verification of earlier proclivities. This selection model is supported by evidence from within-sibling models that find that higher education’s effects on political outcomes are significantly reduced when controlling for unobserved familial confounders (C. Campbell & Horowitz, 2016; Simon, 2022).

However, evidence of some direct causal effects found in this work suggests that socialization is not to be discounted entirely. Further, recent research finds evidence of sizable within-person attitudinal change associated with higher education (Broćić & Miles, 2021; Scott, 2022). These findings are consistent with scholarship documenting the importance of the transition from adolescence to young adulthood as an impressionable period that is formative for beliefs (Lersch, 2023), as well as recent qualitative work on the significance students attribute to higher education in their moral development (Monaghan, 2020).

If higher education does meaningfully influence moral attitudes, what is the scope of change? Scholars have long identified a moral project in undergraduate education that arguably affects students broadly. Higher education is a gateway into the ranks of those who wield disproportionate influence in public life. The overarching task of undergraduate socialization thus involves preparing an educated citizenry distinguished by their competence, intelligence, and ethical commitments commensurate with a leadership role (Talcott & Platt, 1973). According to Monaghan (2020, p. 183), credentials confer those who wield them with moral virtue—‘to be educated is to acquire morally correct knowledge: the “truth” of a tolerant cosmopolitanism.’ While discussions of education have ranged widely, several research traditions have suggested this ‘truth’ promoted by universities is the virtue of rational inquiry and universalism (Baker, 2014; Monaghan, 2020; Talcott & Platt, 1973). Practically, this means that multiple features of undergraduate education should converge to encourage students to move from particularistic commitments rooted in traditionalism towards a more cosmopolitan universalism. Consistent with this, some studies find that higher education encourages individuals to shift from moral judgments based on personal interest and conformity to social norms and adopt more universal principles of justice (Broćić & Miles, 2021; Mayhew et al., 2016; Parker et al., 2016).

However, while higher education might have formative experiences that are common across the board, the university experience is not monolithic. Rather, the university contains a rich social ecology consisting of varied influences arising from multiple sources—such as faculty, curriculum, administrators, or peers—which could work individually or collectively to effect change (Talcott & Platt, 1973). Thus, the university is a ‘bundle’ of different actors and demands that intersect in ways that are not entirely congruous. Methodologically, this requires moving beyond the coarse measures of educational attainment sometimes used in prior work (C. Campbell & Horowitz, 2016; Scott, 2022). Capturing the effects of higher education requires tracking these sources and determining how different aspects contribute—or fail to contribute—to change.

Below we consider how different aspects of the university bundle might produce change in these moral dimensions, focusing on variation across fields of study and peer networks before briefly addressing the role of exposure to peripheral intellectual life in online spaces.

Fields of study

Variation in socializing experiences is perhaps most decidedly shaped by field of study. Ladd and Martin Lipset (1976) conceived of fields of study as subcultures containing their own norms that students became socialized into. While ideological leanings of different disciplines have changed historically (Ladd & Martin Lipset, 1976, pp. 73–4), survey research since the 1960s documents that those in the social sciences hold more critical attitudes towards existing social norms and arrangement and are more likely to support social change than those in the natural sciences, business, or agriculture (Broćić & Miles, 2021; Stubager, 2008; Werfhorst & Graaf, 2004).

Fields of study can be expected to socialize students primarily through exposure to subject matter and relations with faculty. The humanities, arts, and social sciences address questions related to social problems, and the human condition more generally which can have a moral valence (Parker et al., 2016; Stubager, 2008). Curricular content that is more technical and tangentially related to moral concerns, as in STEM degrees, may be less likely to elicit moral change. Morally-valanced topics can be addressed in a variety of ways with potentially different implications for moral change. For instance, a wide range of viewpoints might be presented, or only a subset of viewpoints. Likewise, instructors might be careful to avoid expressing their own views, or they might tacitly (or explicitly) endorse particular moral positions. To the extent that disciplines in the humanities, arts, and social sciences disproportionately expose students to perspectives or communicate moral messaging in line with social justice concerns, students may shift their beliefs away from traditional moral positions towards more liberal sensibilities. Influence may be especially effective when students perceive instructors to be reliable sources they can identify with (Dey, 1996).

But if students majoring in the humanities, arts, and social sciences may be expected to adopt more liberal moral concerns, whether they become more orthodox in these commitments is less clear. A long literature associates these degrees with anti-authoritarian attitudes that cut against traditional orthodoxies (Peterson & Lane, 2001; Roccato, 2008; Scott, 2022). However, it is unclear whether a similar effect for left-wing authoritarianism can be observed. To the extent that professors in these fields pursue a ‘sacred mission’ committed to social justice, we may expect these degrees to heighten moral absolutism characteristic of LWA. Scholars may be more likely to selectively present left-leaning views at the expense of alternative views, and/or convey their own moral beliefs to students, either in class or in personal interactions.² This may especially apply to fields with a strong tradition of social justice activist-scholarship, such as race, gender, and sexualities studies (Horowitz et al., 2018), and hence students in those fields might be especially likely to increasingly endorse LWA.

Peers

While the foregoing discussion points to more top-down sources of change, social influence might also come from peer networks. Scholars point to the importance of peers in creating group obligations that inform how identities and beliefs are internalized (Peter & Stets, 2009; Stryker & Burke, 2000). Horowitz et al. (2017), for instance, find that role expectations set by friends are critical for how activist identities become internalized among college students. Recent research also points to the importance of homophilous peer networks on college campuses for ideological affiliations—change can track in both liberal and conservative directions based on the affiliations of the network (Binder & Wood, 2013; Logan et al., 2021; Rauf, 2021). To the extent that moral attitudes express group affiliations, this work suggests peer networks could contribute to moral change that shifts in both liberal and conservative directions depending on group composition.

Peer networks may also uniquely relate to left-wing authoritarianism. Talcott and Platt (1973), for instance, identified student solidarities as seedbeds for radical left-wing moralism. Student movements from the 1960s embraced absolutist ‘anti-establishment’ politics which conflicted with the pluralistic inclinations of campus faculty and administration (see also Erikson, 1994). It is unclear whether contemporary left-leaning student subcultures will experience a similar tension today, given a more left-leaning faculty and administration (Eagan et al., 2015). Some accounts suggest that progressive orthodoxy increasingly informs the official culture of university and even aspects of its corporate administration (Campbell & Manning, 2018; Paul, 2022). The relative influence of peer networks on encouraging absolutist left-leaning views thus remains an open question.

Online experiences

Moral change might also occur in spaces peripheral to campus life. Studies find that adolescents and young adults increasingly receive their primary exposure to political and cultural life in the form of memes, satire, and political ‘infotainment’ that they access online (Jakubowski, 2021; Nadeem, 2020). Online spaces are also common sites for ‘culture war’ conflict, particularly its more radical expressions (Finlayson et al., 2022; Nagle, 2017). This raises the possibility that moral change during university might have little to do with experiences on campus, but instead reflects students’ online activities. Of course, online effects might reflect the influence of higher education to the extent that university experiences with peers, staff, or course content encourage students to engage online (e.g., by cultivating political interest, Le & Nguyen, 2021). However, some studies indicate that online engagement is a predictor of educational attainment, suggesting that influence could run in the opposite direction (Bowden & Doughney, 2012). Rather than attempt to tease out the relationship of higher education to online engagement, we restrict our focus to field of study and peer effects and treat online experiences as a confound that we control for in our analyses.

Current study

Recent evidence that higher education can affect moral change net of selection raises questions over what aspects of the collegiate experience are responsible for effecting moral change. In this study, we explore two broad facets of college life: influence from fields of study—a term we use to include both the types of courses students take and the content of those courses—and influence from peers. Scholars have discussed the importance of these features, but their analyses leave outstanding questions about their relative importance and interrelations.

We begin to explore these questions using two waves of data from a large university in Canada. These data contain measures of both moral attitudes and various aspects of the collegiate experience, making them well suited to explore the mechanisms of moral change at universities.

Methods

The first wave consists of pooled samples collected between 2019–2021. Students taking an introductory sociology class were invited to participate in a survey and were rewarded one bonus percentage point towards their course grade for their participation. 455 students participated during the Fall 2019 semester, 281 during the Winter 2020, 230 during Fall 2020, and 221 during the Winter 2021 semester, for a total sample of 1,187. We recruited participants into our wave two survey via email. Participation was incentivized with a pre-paid $5.00 Amazon gift card, and by entering those who completed the survey into a raffle for one of thirty $100 Amazon gift cards. After an initial email and several follow ups, we received 243 usable surveys, for a 20% response rate. Of these, an additional four cases were removed for unduly influencing results, and seven for completing less than 33% of the wave 2 survey, meaning they did not answer the questions related to moral attitudes.³

Measures

We measured moral concern using the 30-item Moral Foundations Questionnaire (MFQ-30). The MFQ-30 measures five aspects of morality that can be grouped into individualizing and binding foundations described earlier. The Cronbach alpha scores for binding foundations were 0.84 for wave 1, and 0.87 for wave 2. For individualizing foundations they were 0.81 at wave 1 and 0.75 at wave 2. For most analyses, we used the difference between the two (individualizing—binding), which can be interpreted as the extent to which respondents prioritized individualizing concerns over binding concerns. This reflects both the practical necessity of maximizing statistical power due to the small sample size (discussed further below), as well as the theoretical expectation that university moves students closer to a moral profile typical of liberals, in which case the key distinction is the relative emphasis placed on each type of moral foundation.

Wave 2 also contains Costello et al.’s (2020) 13-item left-wing authoritarianism scale (LWA). LWA consists of three factor components 1) anti-hierarchical aggression reflecting the desire to overthrow established hierarchies and punish those in power; 2) top-down censorship which reflects the willingness to wield group authority to suppress dissent; and 3) anti-conventionalism, which reflects an absolutist approach to social justice, and intolerance of conservative views that are ideologically divergent. To simplify the presentation of results and avoid measuring constructs with too few items, we included all items in a single scale. The Cronbach alpha score for LWA in our sample was 0.78.

The second major substantive focus of our survey involves a series of questions related to the different agents of socialization described above. This includes questions related to 1) fields of study, in particular course content and faculty relations; 2) student peer groups; and 3) online engagement. These variables were only measured at wave 2.

We grouped the majors students reported into three broad fields of study: business, STEM, and the humanities, arts, and social sciences (HASS). Majors were coded into only one category. However, individual students could have more than one major and hence belong to multiple fields of study. Appendix B in the supplemental online material shows which majors were coded into which field of study categories.

We capture experiences in fields of study in a few ways. First, we create an indicator for whether students reported ever taking a ‘gender/sexuality studies’ or a ‘race/ethnic studies’ course, as these courses might be particularly likely to endorse moral beliefs consistent with social justice activism. We also presented a list of ideas related to left-wing social justice positions scholars associate with the ‘sacred mission’ in academia, and alternative ideas more common on political center or on the right, and asked students to reported whether they had ever heard an instructor discuss each one during class (Horowitz et al., 2018; Smith, 2014). The full list can be found as part of the survey questionnaire in Appendix A in the supplemental online material. Example claims from the left include ‘capitalism is a patriarchal system’ and ‘society is rife with systemic racism’. Examples from the right/center are ‘political correctness is detrimental to the quality of public discourse’ and ‘immigrant assimilation is important for a well-functioning society’. Given our theoretical interest in assessing claims about how professors endorsing a ‘sacred mission’ dedicated to social justice affects students, we distinguish exposure to left-wing ideas consistent with this platform from right/center ideas that break from it. Claims were averaged to form scales representing exposure to left-wing and right/center claims (α = 0.78 and 0.59, respectively), with 0 representing no exposure to those claims, and 1 representing high exposure.⁴

To capture peer group effects, respondents were asked ‘to what extent do you and those friends you met at university agree on political matters?’ with responses ranging from 0 = not at all to 4 = entirely. If respondents reported that they and their friends agreed ‘a great deal’ or ‘entirely’ on political matters, we coded them as belonging to a politically homogenous peer group (vs. not belonging to that group). Groups were coded as left or right-leaning based on whether respondents reported (at wave 2) that they are on the left or right side of the political spectrum. We refer to these groups as respondents having ‘left’ and ‘right’ university peers. Equivalent measures were created with reference to friends met outside of the university, to serve as controls.

Online engagement variables were included as controls in some analyses. These were how often respondents reported that politics comes up on social media (0 = never to 4 = very often) and for those who reported that politics does come up, a variable capturing the type of political content that the respondent is most likely to read (1 = left-leaning, 2 = an even mixture of left-leaning and right-leaning, 3 = right-leaning). Our analysis treated the type of political content as a binary variable capturing whether or not respondents were most likely to read left-leaning content online (0 = read right-leaning/even mixture of left-leaning and right-leaning, 1= left-leaning content) given its theoretical importance as a potential confound of other left-leaning sources of influence in universities, as well as the low number of respondents reporting that they mostly read ‘right-leaning’ content.

The full survey questionnaire can be found in Appendix A in the supplemental online material.

Analyses

All analyses are exploratory. We opted to perform exploratory analyses given both the relatively large number of possible effects that we wished to examine, and the small sample size.

Given the evidence of selection effects in past work, we begin by examining evidence for change in our sample, focusing on changes in individualizing and binding foundations between waves 1 and 2, as well as students’ self-reports of moral change. We then move to exploring the mechanisms that might explain moral change by assessing how moral foundations and LWA at wave 2 are patterned by experiences during higher education, specifically by politically homogenous peer groups and fields of study. Our analysis of moral foundations focuses on the relative priority placed on individualizing and binding foundations. Analyses of change in moral foundations and LWA used linear regression with robust standard errors and were estimated with full-information maximum likelihood to help adjust for missing data (Enders, 2022). Multi-equation (path) models were estimated simultaneously.

An important consideration for our analyses was the small sample size. Given so few cases, we suspected that power would be an issue, so we avoided gratuitously adding controls variables to our models ‘just in case’. Instead, we used a data-driven selection procedure which interested readers can reproduce using our study analysis files. We selected variables from wave 1 that we believed could select a person either into the wave 2 sample, a politically homogenous friend group, or into a HASS major—the set of majors identified in past work as being most strongly linked to moral change (Broćić & Miles, 2021). We then pared down the list using both forward and backward stepwise regression. Of the remaining set, we eliminated variables that did not predict our outcome variables. This left us with the wave 1 difference between individualizing and binding foundations as a control for analyses of both field of study effects and peer group effects. Gender was also retained as a control for peer effects (0 = male, 1 = female).

For related reasons, our analysis reports results that reach marginal significance (p < 0.10). We suspect that the size of the standard errors will be inflated by the small sample size, raising concerns over reporting ‘false negative’ results. By noting marginally significant results, we aim to mitigate this risk and highlight potentially meaningful patterns that strict adherence to the arbitrary cut-off value of p < 0.05 could obscure (Hackshaw & Kirkwood, 2011). However, we encourage caution when interpreting marginally significant results, and recognize that they do not provide strong evidence for observed relationships.

Results

We begin by describing patterns of change in moral foundations between wave 1 and 2, as well as self-reported moral change (measured at wave 2). Changes in LWA could not be assessed, as LWA was only measured at wave 2.

Figure 1 plots the change in individualizing and binding foundations, as well as changes in how much respondents prioritize individualizing over binding foundations. The results lend initial support to the claim that student moral beliefs change during their studies. Interestingly, non-directional t-tests suggest that both individualizing (−0.10, p = 0.019) and binding foundations (−0.21, p < 0.001) decrease among students. However, binding foundations decrease more, meaning that the moral priority students assign to individualizing over binding foundations increases during their studies (0.11, p = 0.037). This general shift among students is consistent with accounts arguing that the highly educated adopt more cosmopolitan moral sensibilities that increasingly shun traditionalism over the course of their studies.

Figure 1. Within-person change in moral foundations (0–5) between wave 1 and 2.

Self-reports of moral change among students offer additional evidence that higher education can shift moral beliefs. During the wave 2 survey, students were asked how much their university experience had changed their beliefs about what is morally right or wrong. Their responses are shown in Figure 2. Figure 2 indicates that 49% of students report that their university experience changed their beliefs either somewhat or very much, and only 12% indicate that their moral beliefs did not change at all. While we should be cautious at accepting these self-reports fully at face-value, it is still noteworthy that respondents believe that their moral beliefs shifted given that morality tends to be a central component of many individuals’ sense of self, and stable self-related identifications resist change (Peter & Stets, 2009; Strohminger & Nichols, 2014).

Figure 2. Cross-tabulated responses to ‘how much has your university experience changed your beliefs about what is morally right and wrong, if at all?’.

We turn now to an analysis of the pathways through which universities might shift moralized attitudes. Given the anticipated issues with statistical power, we first present analyses of peer and curriculum effects separately before presenting a joint model.

We begin with peer effects. Results are displayed in Table 1. Politically left-of-center respondents with likeminded university peers prioritize individualizing morality over binding morality to a greater extent than other students (b=0.36, p = 0.006). Likewise, these students report higher levels of LWA compared to their peers (b=0.30, p = 0.029). Effects for left-leaning non-university peers are also sizable but are not significant for either outcome. Unexpectedly, results for right-leaning university peer groups resemble those from left peer groups. However, these results are based on very few cases (n = 10), so we refrain from interpreting them. Overall, the results shown in Table 1 suggest one way that universities might influence moralized attitudes is through social ties to politically homogenous peer groups.

Table 1. Linear regressions of moral foundations and LWA on Peer Groups.

	Individualizing vs Binding Foundations			Left-Wing Authoritarianism
	B	SE	p	B	SE	p
Left-university peers	0.36**	0.13	0.006	0.30*	0.14	0.029
Right-university peers	0.30	0.19	0.119	0.44	0.41	0.290
Left non-university peers	0.26^†	0.14	0.074	0.25^†	0.13	0.063
Right non-university peers	0.12	0.19	0.540	0.11	0.23	0.637
Constant	0.29	0.12	0.022	2.47	0.14	0.000

N = 232. ***p < 0.001; **p < 0.01; *p < 0.05; ^†p < 0.10. Controls for individualizing vs. binding morality at wave 1 and gender not shown.

Turning now to field of study effects, the key finding from past work is that majoring in the humanities, arts, and/or social sciences (HASS) shifts students away from traditionalist moral conceptions. Our results support the claim that these majors have an effect but suggest that it is indirect. Figure 3 displays a path diagram of how curricular content affects change through direct and indirect pathways. Perhaps unsurprisingly, HASS majors are more likely to have ever taken a gender/sexuality or race/ethnic studies course (b=0.22, p = 0.004). These courses, in turn, increase exposure to hearing instructors discuss left-wing social justice claims by about 20% points. There is a small direct effect of being in a HASS degree program—about 8% points—indicating that these types of claims are discussed in other HASS courses as well. It is exposure to these ideas rather than enrollment in particular courses or majors per se that then has a direct effect on beliefs. As seen in panel A, those exposed to left-wing ideas related to social justice are then more likely to prioritize individualizing over binding morality (b = 0.70, p = 0.001). Notably, exposure to right-centrist ideas has the opposite effect and it is roughly equivalent in magnitude (b = −0.73, p = 0.008). This suggests that an education that exposes students to a range of perspectives is less likely to systematically shift students’ moral attitudes in a particular direction (though it might well shift individual students significantly, but in various directions).

Figure 3. Path diagrams plotting direct and indirect effects of Curricular Content on individualizing vs. Binding foundations (panel A), and left-wing authoritarianism (panel B). N = 232; † p < .10; *p < .05; **p < .01; ***p < .001. Controls for individualizing vs. binding morality at wave 1 are not shown. All sub-models within each path model were estimated simultaneously using full-information maximum likelihood estimation.

The story is similar for LWA, though the estimates are less precise, as evidenced by the larger standard errors. As seen in panel B, exposure to left-wing social justice ideas predicts higher levels of LWA (b = 0.67, p = 0.013), while exposure to right-centrist ideas might have the opposite effect (b = −0.50, p = 0.074). Further, it is possible that having taken a gender or race course has a direct effect—the effect is modestly sized, but only marginally significant (b = 0.19, p = 0.082).

To this point, we have examined university peer and course effects separately. We now include them in the same model to control for shared variation and compare effect sizes. We also add controls for exposure to political content on social media to address the possibility that it is online engagement that leads to moral change, rather than processes more directly tied to higher education. Results are displayed in Table 2.⁵ Turning first to individualizing vs. binding morality, we see that results are largely consistent with the results from the previous models. Exposure to left and right-centrist ideas both have significant effects in the same direction as observed previously, though the magnitude of these effects is smaller (b = 0.49, p = 0.02 and b = −0.63, p = 0.019, respectively). Likewise, associating with left university peers predicts prioritizing individualizing over binding morality, though the effect is now smaller and only marginally significant (b = 0.24, p = 0.07). Results for LWA are also largely consistent with previous results, though effects generally shrink and become either marginally significant (exposure to left-wing ideas, b = 0.55, p = 0.062) or non-significant (everything else, p > 0.10).

Table 2. Linear regressions of moral foundations and LWA on university bundle.

	Left-wing Ideas				Right-center Ideas				Enrolled in Race/Ethnicity or Gender/Sexuality Course				Individualizing vs. Binding Foundations				Left-wing Authoritarianism
	B	SE	z	p	B	SE	z	p	B	SE	z	p	B	SE	z	p	B	SE	z	p
Left-wing Ideas													0.49*	−0.21	2.33	0.020	0.55^†	−0.3	1.87	0.062
Right-center Ideas													−0.63*	−0.27	−2.35	0.019	−0.35	−0.28	−1.23	0.220
Enrolled in Race/Ethnicity or Gender/Sexuality Course	0.20***	0.04	5.32	0.000	0.02	0.03	0.59	0.553					0.03	−0.1	0.25	0.802	0.12	−0.12	0.98	0.325
Humanities, Arts, or Social Science Major	0.08*	0.04	2.08	0.037	−0.03	0.04	−0.80	0.425	0.21**	0.08	2.84	0.005	−0.12	−0.11	−1.12	0.263	0.01	−0.12	0.11	0.910
Left University Peers													0.24†	−0.13	1.81	0.070	0.17	−0.13	1.35	0.176
Right University Peers													0.35†	−0.19	1.84	0.066	0.38	−0.41	0.93	0.351
Left non-university Peers													0.12	−0.15	0.79	0.427	0.16	−0.12	1.31	0.191
Right non-university Peers													0.1	−0.18	0.55	0.580	0.14	−0.22	0.64	0.525
Controls
Left-leaning political content online													0.26*	−0.12	2.15	0.031	0.26*	−0.12	2.11	0.035
Read politics online													0.03	−0.05	0.66	0.506	−0.08	−0.07	−1.23	0.217
Individualizing vs. Binding Foundations (wave 1)	0.05†	0.02	1.94	0.053	−0.02	0.02	−1.17	0.242	0.11*	0.05	2.25	0.024	0.63***	−0.07	8.83	0.000	0.08	−0.08	0.99	0.324
Female													−0.07	−0.11	−0.59	0.558	0.15	−0.16	0.93	0.351
Constant	0.30***	0.05	6.53	0.000	0.43***	0.05	9.35	0.000	0.26**	0.09	3.05	0.002	0.35	−0.22	1.63	0.102	2.58***	−0.23	11.05	0.000
R^2	0.21				0.01				0.07				0.54				0.19

Overall R²  = 0.59; N = 232. ***p < 0.001; **p < 0.01; *p < 0.05; ^†p < 0.10.

Interpreting these results is complicated by the low power of our analyses. While the magnitude of many point estimates shrink, as we would expect when controlling for shared variation, it strikes us that they generally remain closer to the original effect sizes than to zero. This hints that many of the estimates we observed in the individual models for field of study and peer effects might be genuine, if inflated, and could be detected in the joint model given a larger sample with greater power. For instance, the alternative possibility that peer effects are spurious when accounting for field of study seems unlikely given these results. Rather, the findings appear more consistent with the notion that these different aspects of the university bundle can each contribute to moral change.⁶

To give a better sense for the magnitude of moral change, Table 3 shows predicted values of individualizing/binding foundations and LWA for students based on varying exposure to mechanisms associated with the university bundle. Predictions were made using the joint model (Table 2). All values not specified in the table were set to their sample means. Table 3 reveals that, compared to a hypothetical student who is average on all model variables, field of study effects (i.e., being at the 75^th percentile in exposure to left-wing ideas and taking a race/gender course) predicts that a student will place greater emphasis on individualizing over binding foundations, and express stronger endorsement of LWA. In contrast, a student who is exposed to right-center ideas places slightly less priority on individualizing vs. binding foundations but does not differ in LWA. When peer effects are examined alone, the tendency is toward higher levels of both individualizing vs. binding morality and LWA, but effects are not significant (again, likely due to low power). The magnitudes of leftward field of study and peer effects are comparable for moral foundations (0.13 and 0.16 above a sample averaged prediction), while field of study effects are somewhat higher for LWA (0.20 vs. 0.12). This suggests that both official and unofficial facets of the university experience contribute in non-trivial ways to explaining why higher education can change moral attitudes. When leftward field effects are combined with left university peer effects, students show the highest levels of prioritizing individualizing morality (0.30 above a sample averaged prediction), and the highest levels of LWA (0.31). Change therefore varies by exposure to different social mechanisms which can have cumulative effects.

Table 3. Predicted values of moral foundations and left-wing Authoritarianism Across the University Bundle.

	Race/Gender Course	Left-wing ideas	Right-center ideas	Left university peers	Predicted individualizing vs. binding priority	Predicted LWA
Comparison	Sample average	Sample average	Sample average	Sample average	1.56	3.13
Field of study effects (left)	Yes	75^th percentile	Sample average	Sample average	1.69, p = 0.015	3.33, p = 0.028
Field of study effects (right-center)	No	Sample average	75^th percentile	Sample average	1.45, p = 0.019	3.08, p = 0.220
Peer effects	Sample average	Sample average	Sample average	Yes	1.72, p = 0.070	3.25, p = 0.176
Field and peer effects (left)	Yes	75^th percentile	Sample average	Yes	1.86, p = 0.005	3.44, p = 0.008

p-values are from a χ² test of the difference between the listed prediction and the comparison prediction. All variables not shown were set to their sample averages.

To better understand the substantive implications of these predicted values, we compare how contrasting university trajectories relate to relevant mean-differences. For instance, students exposed to left-leaning social justice curricular content, with left-leaning university peer affiliations (row 5 in Table 3) score 0.25 higher on left-wing authoritarianism than those exposed to right-center ideas (row 3 in Table 3). This is slightly smaller, but comparable, to the mean-difference between those who self-identify on the left vs. right side of the political ideological spectrum in our sample (0.30 on a 7-point scale). By contrast, these university trajectories develop moral foundations differences that are about one third of the size of left vs. right differences in the sample (0.23 vs. 0.68). We can also compare these within-university differences to those across higher education by drawing on nationally-representative data from the National Study of Youth and Religion which captures moral foundations of Americans of a similar age during 2013 (see Broćić & Miles, 2021).⁷ In the NSYR sample, those who enrolled in HASS majors are 0.42 higher in the priority they assign individualizing over binding morality than those who never enrolled in university. Taken together, these comparisons suggest that higher education does not exert an overwhelming influence on moral attitudes, but that its effects can nonetheless be substantial.

We would likely observe greater range of within-university differences if our sample included a wider cross-section of university students. Because we sampled from introductory sociology classes, our sample consists of a disproportionately high number of students in HASS majors (71%). This might explain why the moral profile of the sample skews towards a moral profile typical of those on the left. Despite notable differences among students, individualizing vs. binding morality scores are generally high in the sample. Consider that the average individualizing vs. binding foundations of HASS majors in the NSYR sample (0.88) lands closer to the bottom quartile of our sample (0.78) than it does to our sample average (1.56). The average student in our sample also lands at the midpoint of the LWA scale (3.05). While the trajectories we observe may therefore contribute to divergences, it is important to note that differentiation is overwhelmingly within a left-leaning moral milieu.

Discussion and concluding remarks

Growing backlash against higher education typically rests on the assumption that universities promote political and moral attitudes typical of the left. While some scholars attribute differences between those who do and do not attend college partly to selection (C. Campbell & Horowitz, 2016; Elchardus & Spruyt, 2009; Gross, 2013), others argue that the collegiate experience can produce real change (Broćić & Miles, 2021; Scott, 2022). Drawing on two wave data from a Canadian university, we find that most students report some level of moral change and that, on average, students develop moral profiles prioritizing individualizing over binding moral foundations, bringing them closer to the cosmopolitan universalism typical of the left. But why does this change occur? Our study begins to unpack the ‘university bundle’—the different social mechanisms present in higher education that might account for change. In particular, we find that moral change is tied to influences from both fields of study and peers.

Curricular content appears significant: students exposed to greater amounts of social justice content by their professors are more likely to prioritize individualizing over binding sensibilities. Exposure to this material is concentrated among majors in the humanities, arts, and social sciences, and especially those taking courses related to race and gender. Exposure to ideas more typical of the center/right can have the opposite effect, though our data suggest that encountering these positions is rarer in practice which might explain why, overall, students place growing importance on harm-reduction and fairness for individuals over binding moral concerns for traditional social order.

While our results support the idea that students may develop more cosmopolitan sensibilities, they do not necessarily suggest that students become more ‘liberal’ minded. Rather, exposure to social justice content can increase endorsement of left-wing authoritarianism, a construct that captures the aggressive, absolutist aspects of some varieties of social justice. In this way, our findings qualify research positing that humanities, arts, and social sciences dampen authoritarianism (c.f. Peterson & Lane, 2001; Roccato, 2008; Scott, 2022). This may apply to right-wing authoritarianism, but not necessarily extend to an alternative variant of authoritarianism applied to left-wing social justice. While we cannot clearly establish this as a causal relationship given that LWA was only measured at wave 2, these findings suggest a worthy direction for future research.

Our findings also indicate that peer networks contribute to moral change. Students who report that most (or all) of their close friends at university are on the political left are more likely to prioritize individualizing over binding moral attitudes, and to express higher levels of LWA. However, when controlling for other social mechanisms these effects become marginally significant for moral foundations and non-significant for LWA. We believe that this likely reflects the low statistical power of the joint model rather than the absence of effects, as effect sizes remain non-negligible. Replicating these analyses with a larger sample should be a priority in future work.

Critiques of higher education’s moral influence often center on the belief that change occurs from the top-down, with left-wing professors and administrators endorsing social justice as the correct moral truth of an enlightened society. Our findings are consistent with this narrative to a degree, but also suggest that informal peer influence might be just as important in explaining why many students seem to emerge from higher education with moral attitudes typical of the left. While not the focus of our study, we also note that our models in Table 2 suggest that those who reported reading mostly left-leaning online content were more likely to both prioritize individualizing over binding morals and more strongly endorse LWA. How online and university experiences interrelate remains unclear, but the upshot of these findings is that strong claims of leftist indoctrination are overstated. Evidence to date suggests a more nuanced picture. Some of the effect of higher education is spurious due to selection into universities, while the rest arises from a combination of multiple factors, only some of which are directly tied to the official structures of the university.

In conclusion, our study advances understanding of education and moral socialization by identifying both field of study and peer influence as mechanisms that might explain the effects of higher education on moral attitudes, and by showing that both of these aspects of college life can make substantive contributions to moral change (Broćić & Miles, 2021). While these findings largely correspond with expectations derived from earlier work, the fact remains that our analyses were exploratory and likely underpowered, so our conclusions remain tentative. Additionally, our results are based on a sample from a single Canadian university. We therefore recommend that our analyses be replicated in a larger sample that includes students from a wider variety of majors and universities to determine whether the patterns we found are robust. Attention could also be given to exploring how higher education intersects with the online spheres increasingly affecting moral development among youth, and to determining if the pathways identified in this study are causal (e.g., by estimating fixed effects models, especially on multiple waves of data that allow for cross-lagged effects; Allison et al., 2017).

Disclosure statement

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

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03057240.2024.2349332

Additional information

Notes on contributors

Miloš Broćić

Miloš Broćić is an Assistant Professor of Sociology at McGill University. His research lies at the intersection of political sociology, culture, and sociological theory, with a particular focus on the social formation of moral orientations and its relation to political conflict.

Andrew Miles

Andrew Miles is an Associate Professor of Sociology at the University of Toronto. His work lies at the intersection of the sociology of culture, social psychology, and moral and cognitive psychology. His research focuses on the social development of different moral cultures, the role moral constructs like values and identities play in predicting behavior, and the effects of moral behavior on emotions. He also studies how cognitive processes affect action and has an abiding love for learning and teaching quantitative methods.

Notes

1. In the American context, Kivikangas et al. (2021) note that the association of moral foundations to political orientation is weaker for Black Americans, while Iyer et al. (2012) highlight libertarians whose moral sensibilities also break from the standard left-right dichotomy. Post-communist countries also depart from typical patterning, such as in Latvia where the relative weighing of individualizing over binding foundations is stronger on the right compared to the left (Kivikangas et al., 2021).

2. This would also be true of scholars high in RWA, though given the overwhelming majority of those in the humanities, arts, and social sciences are on the political left, LWA is more likely to be relevant.

3. Originally 15 respondents selected ‘other’ as their political party. These respondents had a large impact on some of the selection models used to determine the control variables used in analyses, so they were dropped as unlikely to be capturing replicable patterns. Of this 15, only 4 participated in wave 2.

4. Technically, the average captures what proportion of the 10 left-wing and 10 right/center claims respondents heard discussed in class. Because the claims do not represent all possible claims, the measure is best interpreted as a level of exposure to types of ideas.

5. The models meet the assumption for linear regression. Visual inspection of scatterplots supports linearity between moral foundations, LWA, and each independent variable. Breusch-Pagan tests fail to reject the null hypothesis that variance is homogenous. Variance inflation factor values were below 10 for all predictors, suggesting no problematic multicollinearity. Q-Q plots and the Shapiro-Wil test suggest that the residuals are approximately normally distributed.

6. It is also noteworthy that results from the joint model remain comparable to original effects even after controlling for online engagement. Indeed, our results indicate that this may be another important source of influence for students. Consuming left-leaning political content is a significant predictor of both prioritizing individualizing over binding foundations (b = 0.26p = 0.031), and LWA (b = 0.26, p = 0.035). Existing work gives us reasons to believe that higher education and online political consumption are interrelated, so it is not entirely clear whether this effect can be interpreted as entirely separate from the university bundle, hence why we treat it primarily as a control. Be that as it may, the finding raises the possibility that moral change during higher education is not confined to top-down sources of influence, or interactions taking place on campuses, but may result from peripheral sources as well.

7. Data is drawn from wave four of National Study of Youth and Religion (2014) and examines respondents 20–25 years of age (n = 1098) comparing those who reported being enrolled in HASS Majors vs. those who report no college experience. See Broćić and Miles (2021) for coding details.