Editor’s Note: In September 2017 the Fearless Speech Index (FSI) was renamed, it is now the Campus Expression Survey (CES).
Norms about speech seem to be changing rapidly on many college campuses. Universities are offering or requiring training in recognizing “microaggressions,” and they are creating “bias response teams” to make it easy for students to report professors and fellow students who commit microaggressions. In response, many students and professors say they now feel like they are “walking on eggshells”, not just in the classroom but in informal interactions as well.
But how do we know that these changes are real? Might the stories just be a collection of anecdotes from a few disgruntled people who are over-reacting to being censured for a rude remark? Where is the data showing that students are afraid to speak their minds?
We know of no good survey to measure this phenomenon, so a group* of social scientists at Heterodox Academy created one – the Fearless Speech Index. This post explains the first draft of the survey and reports preliminary results obtained from an internet sample.
The FSI was designed to give professors and administrators a tool to assess the degree to which students feel comfortable (or reluctant) to speak up and offer their opinions in a relatively small class, with 20-30 students. We use this as the focal situation because it is the place where it is most urgent for students to participate honestly. If students are self-censoring in such class discussions, they are harming their fellow students by giving them a less interesting class and a false impression of the social consensus.
The key idea that drove the development of the FSI is that speech concerns are context dependent. You can’t just ask students for an overall rating of the speech climate on campus. You have to assess WHO is reluctant to speak about WHICH topics and WHY. The first module of the FSI therefore uses a repeating set of questions to assess students’ comfort level speaking up during class discussions about controversial issues related to
- politics, and
- a non-controversial topic, which serves as a comparison point.
The FSI is customizable; users can drop topics or add others, such as sexuality or religion.
For each topic, the student is asked an overall comfort question: “How comfortable or reluctant would you feel about speaking up and giving your views on this topic?” This is followed by questions about what consequences the student fears: “If you were to speak up and give your views on a controversial issue about [gender/race/politics] during a class discussion, how concerned would you be that the following would occur?” The six potential consequences are:
- The professor would criticize my views as offensive.
- Other students would criticize my views as offensive.
- Someone would post critical comments about my views on social media.
- Someone would file a complaint claiming that my views violated a campus harassment policy or code of conduct.
- The professor would give me a lower grade because of my views.
- The professor would say my views are wrong.
The FSI is a web-based survey. Any professor, dean, or college president can get a copy of the survey from our site (as a Word doc, or as a Surveymonkey of Qualtrics survey) and then host the survey themselves. The FSI is designed to be a tool that schools can use to obtain accurate data, which can be used to benchmark efforts to improve.
In order to obtain some data ourselves, we put the survey up at YourMorals.org, a moral psychology research site, and restricted participation to students currently enrolled in college. Ideally, we’d have data from one entire university, but this composite dataset, with students from hundreds of universities, still allows us to examine what kinds of things the FSI can reveal. We note that this dataset is in no way a representative sample of college students. The means must not be treated as accurate measures of national means. Yet when we contrast subgroups (e.g., male vs female, liberal vs. conservative), we have every reason to believe that any differences we find are likely to show up in nationally representative samples—this has been the general finding with surveys at Yourmorals.org (please see here for more information on the samples typically obtained from YourMorals; and here for more discussion about representative samples).
Initial data collection began in mid-March when we posted the survey at YourMorals.org. This preliminary report is based on data downloaded on June 19, 2017. This initial download contained a total of 796 respondents, of whom 483 indicated that they were currently enrolled in a college or university. All subsequent analyses were performed on these 483 respondents. Of these 483 respondents:
- 276 were male, 148 were female.
- 169 chose one of the 3 “liberal” response choices; 82 chose one of the 3 “conservative” response choices. The rest chose mostly chose “middle-of-the-road” (57 participants; henceforth identified as “Moderates”) or libertarian (76 participants).
- By race/ethnicity: 330 self-identified as White/Caucasian, 28 as “East or Southeast Asian, 9 as South or Southwest Asian,” 16 as Hispanic/Latino(a), and 10 as African American/Black. These small numbers in the non-White categories precluded doing any analyses by race.
Key Questions and Findings:
Q1: WHAT topics are students reluctant to speak about?
A1: Race, followed by politics and gender.
See the first row of Table 1, below. High scores show reluctance to speak. All three topics elicited scores near the midpoint of the scale (2.5 is the halfway point, neither comfortable nor reluctant). In contrast, the majority of participants (69.30%) picked “very comfortable” speaking about a non-controversial topic.
Table 1. Overall means. High scores show reluctance or concern. The highest cause of concern for each topic is shown in bold.
Q2: WHY are they afraid? What potential consequences cause them the most concern?
A2: Their fellow students.
Across all three controversial topics, students reported the highest level of concern that “other students would criticize my views as offensive.” See the bottom 6 rows of Table 1. Students reported the lowest level of concern that a comment made in class would be posted on social media. All other concerns fell in between. The values shown in Table 1 are graphed in figure 1, below, where you can see that fear of other students’ reactions is the top concern on the three controversial topics, but not on the non-controversial topic.
Figure 1: Consequences feared for each topic.
Q3: WHO is afraid to speak up?
A3: We found no gender differences in overall reluctance, but found a few differences in reasons for reluctance. On politics, the differences were much larger: moderates and especially conservatives are more reluctant to speak than are liberals.
The FSI allows us to pinpoint which groups on campus are most reluctant to speak in class. It makes it easy to identify interactions too: Who is more afraid to speak up, about WHAT, and WHY? We looked closely at gender differences, expecting to find several. We found only a few small gaps in overall reluctance, none of which were statistically significant.
Figure 2: Overall reluctance to speak, by gender and topic
When we examined the reasons for reluctance, however, several gender differences emerged showing men having greater concerns. Men were more afraid of their fellow students than were women while discussing gender. There were no sex differences in the three questions about actions that professors might take.
Figure 3: Fear of students taking offense, by gender and topic
Figure 4: Fear of social media shaming, by gender and topic
Figure 5: Fear of formal charges being filed, by gender and topic.
There are extensive ideological differences across racial, political, and gender issues.
Moderates and Conservatives reported greater reluctance to discuss racial, political, and gender issues compared to liberals. They also reported greater concern about each consequence compared to liberals. When comparing Conservatives to Liberals, all of these differences were statistically significant. This pattern of results tended to hold when comparing Moderates to Liberals, although there were some comparisons where the differences did not attain statistical significance. A subset of these graphs is presented below:
Figure 6: Overall reluctance to speak, by politics and topic.
Figure 7: Fear of students taking offense, by politics and topic
Figure 8: Fear of lower grade, by politics and topic.
The FSI is able to identify WHICH students feel intimidated speaking openly, about WHICH topics students are fearful of discussing, and WHYstudents are fearful of discussing those topics. Our preliminary results suggest that students fear discussing racial issues the most and that their greatest concern is criticism from their peers.
Males are not more reluctant to speak up in class than females, but they are more concerned about the social consequences of speech on some topics.
The largest group differences were found on politics. Conservatives were far more reluctant to speak up than liberals during class discussions related to race, politics, and gender. They were also more concerned about every negative consequence we asked about. Interestingly, moderates tended to score closer to conservatives than to liberals. Liberals expressed low levels of fear across all topics and consequences.
In conclusion, the FSI can provide a detailed map of speech concerns in a student population. It can tell you WHO is afraid to speak up, on WHICH topics, and WHY. We believe the FSI is an essential tool for professors, deans, and administrators who value free inquiry and who strive to foster open and honest class discussions among their students.
If you are currently enrolled in a college or university and are interesting in taking the FSI, please go to YourMorals.org, register for a free account, and take the survey.
If you are a professor or administrator interested in administering the FSI to a class, a department, or the entire student body please go here for more information.