Regulating campus speech: Student ideology and support for free speech in a conservative university system

ABSTRACT

What are undergraduate students’ preferences and knowledge regarding free speech on college campuses? Using an original survey of college students in a conservative Upper-Midwestern state, we gather responses across six key areas of speech controversy. Student ideology predictably and significantly divides student support for speech regulation. Compared to liberals, conservative students are more supportive of free expression in the context of potential bias incidents, such as hate speech. Meanwhile, liberal students are more supportive where traditional values are potentially challenged, such as unpatriotic speech. We further find that what students believe about the legal ability to punish or prevent speech on campus mostly aligns with their preferences. However, regardless of ideology, our respondents also show broad support for hate speech regulation, and most believe incorrectly that public institutions are legally empowered to punish or prevent it. The results of this analysis are informative for the growing debate about campus life involving students, faculty, administrators, and spectators alike.

KEYWORDS:

• Free speech
• public opinion
• ideology
• hate speech
• higher education policy

Disclosure statement

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

Notes

¹ Democratic students were most likely to support disinviting a racist speaker (52% support) on a list of 25 groups and individuals. For Republican students, a racist speaker was the third highest response (27%; FIRE, 2017).

² Moreover, many of the poll questions about the Stop WOKE Act emphasized the portion of the bill that dealt with K-12 education. The Spectrum News/Siena College Poll (2022), for example, asked, “Do you support or oppose the ‘Stop the Wrongs to our Kids and Employees act,” known as the WOKE Act, which includes prohibiting teaching critical race theory in K-12 public schools, and prohibits mandatory critical race theory training for employees?”

³ See Appendix B for the survey items and the complete list of speech types addressed.

⁴ Chong (2006, p. 43) found that among conservative 18- to 23-year-old college students, support for “racist teachers” was lower compared to support for an “anti-religion teacher,” “communist teacher,” “homosexual teacher,” and a “militarist teacher.”

⁵ In separate modeling, we conducted our analysis among “non-traditional” age undergraduate and graduate student respondents. The sign and significance of the primary independent variable, ideology, remains largely unchanged except for a non-significant effect relating to preferences on speech restrictions in concordance with university values. Several control variables also report different effects in terms of sign and significance. Ultimately, given that the discourse does center around “traditional” college students we choose to focus our analysis on this segment of our respondents.

⁶ Institutions of higher education, like all actors constrained by court rulings, may act contrary to legal precedent (Boldt & Gizzi, 2018; Canon & Johnson, 1984; Rosenberg, 2008). The willingness of a given university to do so may rely on several factors, e.g., the likelihood of ultimate success in litigation; pressure from constituencies; the larger political context of the state; the resources of the institution to litigate, settle, or pay for damages that may be awarded; and the strength of precedents involved (Galanter, 1974; Hansford & Spriggs, 2006). Given the strong historical consensus in the judiciary behind the framework of free speech law governing campuses (Chemerinsky & Gillman, 2018, p. 113), we anticipate that few colleges or universities will accept the risks of defying the courts in this space.

⁷ Importantly, this hypothesis is consistent with the proposition that legal information may be rarer than other types of political information. Specifically, ideology will influence which information a person retains when they encounter information about legal precedents in these areas.

⁸ In separate modeling conducted among “non-traditional” age undergraduate and graduate student respondents, the sign and significance of the primary independent variable, ideology, remains largely unchanged except for a non-significant effect relating to the respondents’ knowledge of whether universities can regulate triggering speech. Several control variables also report different effects in terms of sign and significance.

⁹ College students do appear to share the broader population’s preference for living amongst likeminded individuals (Ulbig, 2020, p. 157; for more discussion of partisan geographic sorting in the general population, also see Sussell, 2013; Tam Cho et al., 2013); consequently, we might expect conservative students to opt in to these more conservative college environments (and liberals tending to opt out). At the same time, it seems plausible that conservatives who opt into a traditional college experience, even in a relatively conservative environment, are fundamentally different from conservatives more broadly, particularly in terms of their openness to diverse views.

¹⁰ The UMUS headcount by race does not distinguish between undergraduate and graduate students, and we suspect graduate students are more likely to identify as non-White. Thus, we suspect that the difference in racial and ethnic diversity is due to our focus on undergraduate students.

¹¹ We excluded 17 respondents who self-identified as gender non-binary from our analysis due to a lack of variation on the dependent variable and significance in our models. Specifically, regarding punishing or preventing unpatriotic speech, all 17 respondents did not support their institutions’ doing so. As a result, no estimates were generated for the non-binary category for that model since the dependent variable does not change. Further, across all models, the non-binary respondent category was not statistically significantly different from students who identified as male or students who identified as female.

¹² UMUS data on enrollees’ gender and parents’ education level are unavailable.