Understanding the values and limitations of student evaluations of instruction (SEIs) or student evaluations of teaching (SETs) is critical to the long-term success of higher education. SEIs have been used since at least the 1920s (Remmers and Brandenburg Reference Remmers and Brandenburg1927) as a major criterion in hiring, firing, promotion, and tenure decisions (Seldin Reference Seldin1993). These evaluations also can indirectly impact an instructor’s career by influencing factors such as course enrollment (Yu, Mincieli, and Zipser Reference Yu, Mincieli and Zipser2021).
However, the appropriateness of SEIs in measuring teaching effectiveness has been called into question for decades. Whereas a meta-analysis found a moderate-sized correlation (r = 0.31) between SETs and student achievement (Wright and Jenkins-Guarnieri Reference Wright and Jenkins-Guarnieri2012), SETs can be influenced by factors other than teaching quality (Henson and Scharfe Reference Henson and Scharfe2011; Langbein Reference Langbein1994; Stroebe Reference Stroebe2020; Youmans and Jee Reference Youmans and Jee2007). For example, correlations have been found between SEIs and the perceived easiness of a course, instructor physical attractiveness, course subject, and instructor gender (Rosen Reference Rosen2018). Additionally, SEIs are inconsistent and better at providing information about student evaluators than the instructors (Clayson Reference Clayson2018). Moreover, there are measurement issues regarding which students are motivated to complete SEIs (Hoel and Dahl Reference Hoel and Dahl2019), the likelihood of misleading conclusions due to acquiescence (Valencia Reference Valencia2019), small sample sizes (Holland Reference Holland2019), and students not always responding honestly to SEI items (McClain, Gulbis, and Hays Reference McClain, Gulbis and Hays2018). Even under the best-case scenario regarding bias and lack of reliability, SEIs alone are unlikely to provide high-quality information about instruction (Esarey and Valdes Reference Esarey and Valdes2020)—and even after implementation of evidence-based teaching practices, SEIs may not change significantly (Stewart, Speldewinde, and Ford 2018). SEIs also may include inappropriate comments, which can harm the well-being of instructors (Lakeman et al. Reference Lakeman, Coutts, Hutchinson, Lee, Massey, Nasrawi and Fielden2021).
One potential reason that nonteaching factors may influence SEIs is that there may be a relationship between these factors and perceived teacher friendliness and availability. The concept of immediacy refers to the degree of perceived physical or psychological closeness between people in a relationship (Mehrabian Reference Mehrabian1966, Reference Mehrabian1971). Applied to the classroom, teacher immediacy refers to the physical or psychological closeness between teachers and their students (Frymier Reference Frymier, Hattie and Anderman2013). Teachers can engage in various verbal behaviors (e.g., use personal examples, encourage students to ask questions, and refer to the class as “our” class) and nonverbal behaviors (e.g., gesture while talking, move around the class, and maintain eye contact) to increase immediacy (Christophel Reference Christophel1990; Meyerberg and Legg Reference Meyerberg, Legg, Jhangiani, Troisi, Fleck, Legg and Hussey2015). One study found a significant correlation (r = 0.54) between ratings of teacher immediacy and teaching evaluations (Moore et al. Reference Moore, Masterson, Christophel and Shea1996).
Some of the factors that influence perceptions of teacher immediacy are commonly considered skills for public speaking (e.g., use of personal examples and making eye contact), which could explain why there is a relationship between teacher immediacy and SEIs. However, other factors relevant to immediacy that are not associated with public-speaking skills could impact SEIs. Two of these factors are an instructor’s name preference (i.e., being on a first-name basis with students versus using one’s professional title, such as “Dr.” last name) and the instructor’s demographics (i.e., race and gender). Thus, it is important to continue building on the current research that examines the influences on and the usefulness of SEIs.
STUDY 1
Instructors often decide how they would like undergraduate students to refer to them; some use their profession title (e.g., “Dr.” or “Professor”), whereas others are on a first-name basis with their students. People can treat others differently depending on their name (Watson, Appiah, Thornton Reference Watson, Appiah and Thornton2011); however, little research has examined the honorifics of the name. One study found that graduate students rated faculty members who were addressed by their first name as more approachable and helpful than those who were addressed by their formal title (McDowell and Westman Reference McDowell and Westman2005). Another study found that female professors (but not male professors) who use the title “Dr.” are perceived to be less accessible (Takiff, Sanchez, and Stewart Reference Takiff, Sanchez and Stewart2001). That study found evidence for this effect in two different ways. First, when students reported how they addressed their instructors, they more often referred to their male professors with their professional title. In their follow-up study, Takiff, Sanchez, and Stewart (Reference Takiff, Sanchez and Stewart2001) manipulated the instructor’s name and gender by having students read a hypothetical transcript of a class. In the transcript, hypothetical students in the class referred to the instructor by either their first name (i.e., “Richard” for male and “Sharon” for female) or as “Professor Parks.” Whereas female professors who went by their professional title were perceived as equal in status to male professors who went by their professional title, the former were rated lower in accessibility. These results are concerning for female professors who may feel the need to choose between being perceived as having high status (i.e., using their formal title) or being accessible (i.e., using their first name), both of which are important for their career.
Although the Takiff, Sanchez, and Stewart (Reference Takiff, Sanchez and Stewart2001) study provided interesting and concerning data, there is a need to examine this effect on SEIs from actual courses. Based on the concept of teacher immediacy and the likelihood that students may feel closer to instructors with whom they are on a first-name basis—according to Takiff, Sanchez, and Stewart (Reference Takiff, Sanchez and Stewart2001), they may be seen as more accessible—I predicted that student evaluations for a female instructor would be higher when the instructor introduces herself using her first name:
Hypothesis H1: An instructor will receive higher SEI scores for courses in which she expresses a preference for using her first name rather than her professional title.
METHODS
The following methods were used to examine the relationship between an instructor’s name preference and SEIs.
Participants
SEI data are deidentified, aggregated, and publicly available online; therefore, a representative from the Institutional Review Board (IRB) confirmed that a human-subjects review was not necessary. SEI data were examined from 16 sections of a course in persuasive communication (Foster Reference Foster2022).
Measures
Students were asked to evaluate three dimensions of teaching on a five-point scale: (1) instructor’s preparedness, organization of material, and clarity of presentation (i.e., well organized, instructor well prepared, communicated subject matter clearly); (2) rapport and instructor commitment (i.e., instructor interested in teaching, instructor interested in helping students, created learning atmosphere); and (3) students’ sense of their own learning (i.e., intellectually stimulating, encouraged independent thinking, learned greatly from instructor), as well as an overall rating. However, the nine items for the different dimensions of teaching loaded onto one factor (Cronbach α = 0.97; McDonald ω = 0.90), which was highly correlated with the overall rating (rs = 0.82). Therefore, the overall ratings were analyzed.
Procedure
Between 2015 and 2020, the same course was taught 16 times by the same instructor. For 10 classes, the instructor expressed a preference for going by her first name and for six classes (table 1), she expressed a preference for going by “Dr.” last name. Sometimes there were two sections of the course in the same semester, one of which was the first-name class and the other was the “Dr.”-last-name class. Because these courses were taught back to back, the order of which class (i.e., earlier or later) was first name or “Dr.” last name was counterbalanced. All classes were taught in person, and the instructor’s name preference was on the syllabus, mentioned on the first day of class, and reinforced with a name tent visible throughout the lecture that listed either her first name or “Dr.” last name. Other than the difference in name preference, the classes were designed to be equivalent.
Table 1 Overall Evaluation for Classes with Instructor Using First or Last Name
Note: SEI=student evaluation of instruction.
RESULTS AND DISCUSSION
The mean SEI response rate across the classes was 88%. This high response rate likely was due to all course sections being offered extra credit if the class reached an 80% or higher completion rate. As expected, SEIs were higher in classes in which the instructor expressed a preference to be called by her first name (M = 4.76, SD = 0.06) than in classes in which she expressed a preference to be called “Dr.” followed by her last name [(M = 4.37, SD = 0.18), F(1, 14) = 38.34, p<0.0001, d = 3.54]. Similar results were obtained when the courses were weighted by the class size [Ms = 4.76 and 4.37, F(1, 14) = 37.39, p<0.0001] and number of responses [Ms = 4.76 and 4.37, F(1, 14) = 35.89, p<0.0001]. Similar results also were obtained when the response rate was used as a covariate in the model [adjusted Ms = 4.76 and 4.37, F(1, 13) = 27.83, p<0.0001].
As expected, teaching evaluations were higher in classes in which the instructor expressed a preference to be called by her first name (M = 4.76, SD = 0.06) than in classes in which she expressed a preference to be called “Dr.” followed by her last name [(M = 4.37, SD = 0.18), F(1, 14) = 38.34, p<0.0001, d = 3.54].
The findings are consistent with previous research on teacher–student rapport (Richmond et al. Reference Richmond, Berglund, Epelbaum and Klein2015) and the concept of teacher immediacy (Moore et al. Reference Moore, Masterson, Christophel and Shea1996), suggesting that students feel closer to instructors who prefer to be called by their first name. From this study, however, it is important that there was a significant difference in SEI scores, not perceived closeness.
One limitation is that the instructor for all courses was a white female. It is important to determine whether these findings replicated for other instructors, especially given the complex relationship between an instructor’s gender and SEIs (Rosen Reference Rosen2018; Wong and Bouchard Reference Wong and Bouchard2021). Previous research suggests that female professors (but not male professors) who use the title “Dr.” are perceived to be less accessible (Takiff, Sanchez, and Stewart Reference Takiff, Sanchez and Stewart2001). Additionally, minority instructors have been found to receive lower evaluations than white instructors (Carle Reference Carle2009).
STUDY 2
The results of Study 1 provide data from an actual course that are consistent with previous research (Takiff, Sanchez, and Stewart Reference Takiff, Sanchez and Stewart2001) that used a hypothetical course. However, it is unclear whether these results would be generalizable to other instructors. Additionally, other demographic characteristics of the instructor (e.g., gender and race) could interact with name preference to influence SEIs. Thus, an experimental study explored these additional demographics, which acknowledged that experimental manipulations could provide insight into real-world phenomena but do not replicate them.Footnote 1
Based on the results of Study 1, I predicted more favorable teacher ratings for instructors who expressed a preference for using their first name (i.e., “Brian” for males and “Rachel” for females) than for instructors who expressed a preference for using “Dr. Moore.” These names were chosen because they were in the top 16 most common names from 1980,Footnote 2 making them plausible names for an instructor.
Hypothesis H1: Instructors who introduce themselves using their first name will have higher evaluations than instructors who introduce themselves using their professional title.
Additionally, gender may influence SEIs and may interact with name preference (Takiff, Sanchez, and Stewart Reference Takiff, Sanchez and Stewart2001). Thus, I predicted less-favorable SEIs for female instructors (but not for male instructors) who expressed a preference for using “Dr. Moore.”
Hypothesis H2: Female (but not male) instructors who introduce themselves using their first name will have higher evaluations than instructors who introduce themselves using their professional title.
Additionally, the instructor’s race may influence SEIs (Carle Reference Carle2009) and may interact with name preference or gender. For example, one study found that female minority instructors were rated lowest on student evaluations (Chavez and Mitchell Reference Chavez and Kristina2020). Thus, I predicted that Black female instructors who preferred to use “Dr. Moore” would receive the lowest teacher ratings and white female instructors who preferred to use their first name would receive the highest teacher ratings.
Hypothesis H3: The lowest evaluations will be for instructors who are Black and female and who introduce themselves using their professional title.
I also sought a better understanding of why these demographics and preferences may influence SEIs. Using a mediation model, I predicted that a name preference for “Dr. Moore” would decrease teacher ratings through a decrease in perceived teacher immediacy—in this case, defined by teachers who are perceived as being more “friendly,” “engaging,” “fun,” “nurturing,” and “caring”—especially for female instructors. Previous research has shown that students expect female instructors to be more caring (Andersen and Miller Reference Andersen and Miller1997; Langbein Reference Langbein1994), and female instructors might be perceived as less caring if they request students to use the title “Dr.” when referring to them.
Hypothesis H4: Instructors who introduce themselves using their first name will have higher evaluations because of increased perceptions of being friendly and closer to students and also depending on the instructor’s race and gender.
Furthermore, I expected that this mediation might depend on individual student characteristics (e.g., political ideology). However, based on previous research (Mosso et al. Reference Mosso, Briante, Aiello and Russo2013; Ratliff et al. Reference Ratliff, Redford, Conway and Smith2019), I expected this result to be limited to students who identify as conservative and Republican.
Hypothesis H5: Conservative students will rate Black and female instructors lower on teaching evaluations when the instructors introduce themselves by their professional title rather than by their first name.
METHODS
The following methods were used to examine the relationship between an instructor’s demographics and SEIs.
Participants
Study 2 was approved by the university’s IRB (Protocol 2021B0044). A power analysis revealed that 512 participants (i.e., 64 in each of the eight conditions) were needed to detect a medium-sized effect (d = 0.5) at the 0.05 significance level (two-sided), given that power = 0.80 (Cohen Reference Cohen1988). The participants included 648 undergraduate students who received course credit for their voluntary participation. Those who failed the two attention-check items were excluded (N = 150), resulting in a total of 498 valid responses. Thus, the study was slightly underpowered. There were 331 females, 165 males, and two nonbinary participants: 69.5% white, 10.0% Black, 14.5% Asian, and 6.0% Hispanic. Most participants identified as being affiliated with the Democratic Party (45%), followed by Independents (30%) and Republicans (17%); 8% declined to respond. They ranged in age from 18 to 63 (M = 20.30, SD = 3.59).
Measures
Using exploratory factor analysis, indices were created, as follows:
-
• Four items measured the likelihood of registering for the course (i.e., the scale with the following stems: “How likely are you to” “register for this class,” “recommend this class to a friend,” “tell a friend about this class,” and “want to learn more about this class”) (Cronbach α = 0.88; McDonald ω = 0.88).
-
• Six items measured evaluations of the course (i.e., whether they thought it was useful, interesting, engaging, fun, intellectually stimulating, and worthwhile) (Cronbach α = 0.87; McDonald ω = 0.87).
-
• Eight items measured support for traditionally liberal social and political issues including gender equality, racial equality, women’s rights, LGBTQ+ rights, Black Lives Matter, universal healthcare, free college, and the environmental movement (Cronbach α = 0.89; McDonald ω = 0.89).
-
• Three items measured support for traditionally conservative social and political issues including the men’s rights movement, All Lives Matter, and Blue Lives Matter (Cronbach α = 0.83; McDonald ω = 0.85).
-
• Five items measured instructor immediacy (i.e., whether the instructor was friendly, engaging, fun, nurturing, and caring) (Cronbach α = 0.89; McDonald ω = 0.89).
-
• Nine items measured SEIs as in Studies 1 and 2 (Cronbach α = 0.90; McDonald ω = 0.90).
Procedures
Similar as in the Takiff, Sanchez, and Stewart study (Reference Takiff, Sanchez and Stewart2001), students were told that they would be introduced to a new course to gauge their interest. However, whereas the Takiff Sanchez, and Stewart study had participants read transcripts of the course and manipulated whether a student in the transcript referred to the instructor by their first name or title, this manipulation had the instructor directly share their name preference with students in a welcome video.
The welcome video was between 28 and 31 seconds long and showed the instructor facing the camera from the shoulders up with no background objects (i.e., a plain, neutral background). The script was the same for all instructors except for their name preference, which was either “Dr. Moore,” “Brian” for male instructors, or “Rachel” for female instructors. Specifically, they started by stating their full name and then their preference; for example, “My full name is Dr. Brian Moore, but I prefer to go by Brian” or “My full name is Dr. Rachel Moore, but I prefer to go by Dr. Moore.” After viewing the welcome video, students read the course syllabus (which differed only in the name of the instructor) and answered survey questions.
To increase generalizability (Wells and Windschitl Reference Wells and Windschitl1999), I used two examples of each instructor’s demographic—that is, there were two white males, two Black males, two white females, and two Black females who created welcome videos, introducing themselves with a preference for either their first name or their professional title. Thus, the study design was a two (instructor name preference: first or last), by two (instructor race: white or Black), by two (instructor gender: female or male) design.
There was no significant SEI difference between the two examples of white females (p = 0.70), the two examples of Black females (p = 0.90), the two examples of Black males (p = 0.69), or the two examples of white males (p = 0.11). Therefore, the data from the two examples of each demographic category were combined.
RESULTS AND DISCUSSION
As expected in Hypothesis H1, SEIs were higher for instructors who preferred going by their first name (M = 4.25, SD = 0.57) than for instructors who preferred going by “Dr. Moore” [(M = 4.11, SD = 0.64), F(1, 493) = 7.11, p = 0.008, d = 0.25]. However, the gender-by-name-preference interaction expected in Hypothesis H2 was nonsignificant (p = 0.32). Thus, the name-preference effect held for both male and female instructors.
Regarding Hypothesis H3, there was a significant three-way interaction among instructor gender, race, and name preference [F(1, 493) = 3.79, p = 0.05]. I probed the interaction by examining each group separately. Black men who introduced themselves as “Brian” had higher SEIs (M = 4.30, SD = 0.81) than Black men who introduced themselves as “Dr. Moore” [(M = 3.95, SD = 0.87), t(492) = 2.67, p = 0.03, d = 0.42]. Name preference did not influence teaching ratings for Black women (p = 0.64), white men (p = 0.73), or white women (p = 0.13).
Black men who introduced themselves as “Brian” had higher SEIs (M = 4.30, SD = 0.81) than Black men who introduced themselves as “Dr. Moore” [(M = 3.95, SD = 0.87), t(492) = 2.67, p = 0.03, d = 0.42].
For explanatory mechanisms, I predicted that name preference for “Dr. Moore” would decrease teacher ratings due to a decrease in perceived immediacy. This hypothesis was partially supported because the mediation model was significant only when the instructor was a Black male. Using Hayes’s (Reference Hayes2018) PROCESS model 4, I found that name preference was predictive of perceived instructor immediacy (β = –0.25, p = 0.04) and that perceived immediacy, in turn, was predictive of SEI ratings (β = 0.56, p<0.001). For the overall model, whereas the direct effect of name preference on SEI ratings was not significant (p = 0.28), the indirect effect through perceived teacher immediacy was significant (β = -–0.14, 95% CI = –0.27, –0.0098). Because the preference for “Dr. Moore” was coded as “1” and the preference for “Brian” was coded as “0,” those who preferred “Dr. Moore” had lower perceived immediacy and therefore lower SEI scores (correlations are shown in table 2).
Table 2 Correlations from Study 2
Notes: *Correlation is significant at the 0.05 level (two-tailed); **correlation is significant at the 0.01 level (two-tailed).
I used Hayes’s (Reference Hayes2018) PROCESS model 7 to examine potential boundary effects for Hypothesis H4 using moderated mediation, but none of these analyses were significant. Thus, the evidence for bias against Black male instructors who expressed a preference for the title “Dr.” remained regardless of social and political factors.
GENERAL DISCUSSION
The literature on the efficacy of SEIs is complicated because some studies show no correlation between instructor demographics such as gender and race (Park and Dooris Reference Park and Dooris2020). However, this current study, aligned with other previous research (Mitchell and Martin Reference Mitchell and Martin2018; Murray et al. Reference Murray, Boothby, Zhao, Minik, Berube, Larivière and Sugimoto2020), provides evidence that nonteaching factors can influence SEIs. When research yields conflicting results, it often is the case that there are boundary conditions to an effect. Examining possible boundary effects and adding to the existing body of literature, courses in which the instructor preferred her first name rated her higher (i.e., never lower than 4.70 on a 5.00 scale) than courses in which the same instructor preferred her last name proceeded by “Dr.” (i.e., never higher than 4.52 on a 5.00 scale).
There are, of course, alternative explanations for the discrepancy between teaching skills and SEIs. For example, it is possible that instructors with a more negative attitude toward SEIs also may receive lower SEIs (Carlozzi Reference Carlozzi2018). In this case, however, the instructor’s SEIs from Studies 1 and 2 were influenced by nonteaching factors despite the SEIs being at or above average for similar courses taught by other instructors.
One notable difference between the actual SEIs in Study 1 and the experiment in Study 2 is that the white female instructors’ name preference in the experiment did not predict SEI scores (i.e., significant prediction was only for Black male instructors). Although further research is necessary to determine the reason for this discrepancy, one possibility is that the current study was slightly underpowered, resulting in the effects of name preference on SEI scores for white female instructors approaching but not reaching significance (p = 0.13). In post hoc analysis, I found that although the immediacy scale was not predicted as a whole (i.e., indexed by averaging) by name preference for white females, the individual item “friendly” was higher for white females who preferred the use of their first name (M = 4.09, SD = 0.723) rather than their professional title [(M = 3.75, SD = 1.03), t(119) = 2.14, p = 0.017]. Additionally, white females who went by “Dr. Moore” were considered more “strict” (M = 2.98, SD = 1.04) than those who went by “Rachel” [(M = 2.67, SD = 0.911), t(121) = –1.77, p = 0.079)]. This led me to consider that it is possible that one approximately 30-second introduction video did not have as strong an impact on participants in the experiment as a semester-long course with an instructor. This demonstrates the importance of examining real-world data in addition to experimental studies (see also Feldman Reference Feldman1993). It also raises the question: If the name preference in a 30-second video significantly influenced SETs for Black male instructors in a hypothetical course, what might it do in an actual course?
If the name preference in a 30-second video significantly influenced teacher evaluations for Black male instructors in a hypothetical course, what might it do in an actual course?
CONCLUSION
As in previous research, SEIs continue to be influenced by factors other than teaching quality. In particular, factors relating to perceived closeness to the instructor (i.e., rating the instructor as “friendly,” “engaging,” “fun,” “nurturing,” and “caring”) appear to influence evaluations, especially for Black male instructors. It is concerning to note that although the SEI scores were relatively high for Black male instructors who went by their first name, their SEI scores decreased dramatically when they went by “Dr. Moore.” This result was consistent even among self-identifying liberal students who were supportive of social issues such as the Black Lives Matter movement. This highlights the need to learn more about how instructor demographics can have a (perhaps largely unconscious) role in SEIs, which can impact their career in important ways. This research also provides evidence to support the need for additional methods to evaluate teaching abilities that are less susceptible to bias than SEIs (McCarthy, Niederjohn, and Bosack Reference McCarthy, Niederjohn and Bosack2011).
ACKNOWLEDGMENTS
I thank Brad J. Bushman for his helpful comments on an earlier draft of this article.
CONFLICTS OF INTEREST
The author declares that there are no ethical issues or conflicts of interest in this research.
Open access
