Teaching assistants’ contributions to creating inclusive and equitable learning spaces in engineering

ABSTRACT

Sustainable engineering education requires us to address discrimination and inequity of experience, thereby contributing to a renewal of the engineering profession that reflects a rich diversity of experiences and perspectives. Many students experience discriminatory events in their classrooms, however the presence of teaching assistants in relatively high numbers means that they could provide both proactive support and reactive intervention to support inclusion and combat discrimination. This paper examines teaching assistants’ (TAs) contributions, employing observations from students (3988), student teaching assistants (277) and doctoral teaching assistants (65). We found that (i) students overwhelmingly reported respectful interactions with their TAs, (ii) while both student assistants and doctoral assistants see inclusion as part of their role, student assistants expressed more confidence in their ability to foster inclusion. Our analysis indicates that TAs are a constructive force to build inclusive engineering classrooms, informing our recommendations for improvingcampus culture.

KEYWORDS:

• Diversity, equity & inclusion
• teaching assistants
• sustainable education
• inclusive classrooms
• engineering education

Introduction

Addressing discrimination and inequity of experience is an integral part of the sustainable development agenda (United Nations [UN] 2015). The sustainable development goals (SDGs) explicitly address inclusivity and equity in education (goal 4), the reduction of inequalities in the world (goal 10) and the promotion of accountable and inclusive institutions at all levels (goal 16). This requires that we consider the barriers that discrimination and harassment in higher education (including engineering education) create for sustainability due to their effect on both student access and success (Bondestam and Lundqvist 2020). The presence of diverse and inclusive perspectives is in itself an asset for sustainable development (Lucena and Schneider 2008; The United Nations Educational, Scientific and Cultural Organization [UNESCO] 2021). In light of the key role that engineers play in developing and shaping the systems and tools that underpin many of the other SDGs, it is particularly relevant to consider issues of equity and discrimination in engineering education. Therefore, tackling discrimination in engineering education has the double effect of both addressing inequity directly in students’ educational experiences and in supporting the development of engineering systems and products that incorporate broader perspectives and which respond to a greater diversity of experiences and contexts.

Engineering has typically been an exclusive discipline and issues with recruitment of diverse students (Barnard et al. 2012; Chubin, May, and Babco 2005; Sagebiel and Dahmen 2006) as well as issues with retention of diversity in engineering persist. A recent report from the United States has highlighted the stark differences in degrees granted in engineering (National Center for Science and Engineering Statistics [NCSES] 2023). Across educational levels, women were awarded far fewer engineering degrees as compared to men (24% of bachelor's degrees, 27% of master’s degrees, 25% of doctoral degrees). Minorities and people with disabilities were underrepresented in the total number of engineering degrees awarded (both in terms of their representation in the general population and total degrees awarded in scientific fields). Several studies have shown links between adverse educational experiences in engineering and undesirable student learning outcomes (Aeby et al. 2019; Bahnson et al. 2022; Dececchi, Timperon, and Dececchi 1998; Geisinger and Raman 2013; Sagebiel and Dahmen 2006; Swim et al. 2003), including minority students leaving engineering entirely (Male et al. 2018; Seymour and Hunter 2019). This shutting down or shutting out of diverse voices is detrimental to sustainability (Ihsen and Buschmeyer 2007; Male et al. 2018) and the integrated, interdisciplinary, contextualised, and systemic thinking it requires (Kolmos, Hadgraft, and Holgaard 2016).

Studies over recent decades have shown that a high degree of these discriminatory events occur as part of students’ in-class experiences (Alpay et al. 2010; Bahnson et al. 2022; Barnard et al. 2012; Geisinger and Raman 2013; National Academies of Sciences, Engineering, and Medicine [NASEM] 2018b; Seymour and Hunter 2019), and can be most severe for students in their first-year of engineering education (Dececchi, Timperon, and Dececchi 1998). Observations made at our institution support these prior findings. The Culture of Respect Survey (Le Duc 2022; Tormey 2021), an internal institutional questionnaire that explored the social well-being of students and employees, provides insight into the experience of belonging, discrimination and harassment experienced within an engineering school. Students were the group most likely to have experienced or witnessed inappropriate or derogatory comments (29% of entire survey population where about half of respondents were students), been targeted by discrimination during their work or studies (12%), been a target or victim of violence or psychological harassment (17%), and experienced unwanted physical contact in the last 5 years during their work or studies (12%). These injurious experiences are directly linked to students’ life on campus and are not disconnected from their engineering studies: for students who signalled being a target of inappropriate remarks, 79% of these remarks took place in a classroom. Similarly, classrooms were the second most common location where sexual harassment or violence was witnessed.

Experiencing, or even witnessing, such events leads to increased feelings of alienation and decreased mental health among students (Nora and Cabrera 1996; Woodford et al. 2014). This is especially true for minority students, who are able to perceive nuances of discrimination to a higher degree than their non-minority peers (Bahnson et al. 2022; Cabrera and Nora 1994; Marsden et al. 2016; McGee and Martin 2011; Poyrazli and Lopez 2007). These are multifaceted problems which arise from different types of interaction dynamics (including frequency and power structure) that occur in the classroom: student-teacher, student-teaching assistant, student-student. In this study we focus on one dimension of this issue: teaching assistant-student interactions in the first-year introductory courses at a large European engineering school.

Teaching assistants are essential for many teaching activities in undergraduate classrooms, especially in courses with large numbers of students. It is therefore reassuring that students have a positive opinion of teaching assistants (Kendall and Schussler 2012) and the effect that teaching assistants have on their learning (Clements et al. 2022; Santhanam and Codner 2012). Teaching assistants have a beneficial influence on students’ attitudes towards learning (Talbot et al. 2015) and their learning outcomes (Alzen, Langdon, and Otero 2018; Crowe, Ceresola, and Silva 2014; Knight et al. 2015; Sellami et al. 2017), especially by facilitating active learning pedagogies (Barrasso and Spilios 2021; Fingerson and Culley 2001; Ruder and Stanford 2018). Of particular relevance for the current study, teaching assistants can have an important effect on reducing inequities in student academic performances (Preszler 2009; Van Dusen, White, and Roualdes 2016; Van Dusen and Nissen 2020) and on increasing student belonging (Clements et al. 2022; Perlmutter, Salac, and Ko 2023).

A constraint to teaching assistants’ capacity to contribute to inclusive classrooms is that novice instructors can struggle to identify conditions and relevant information that can lead to discriminatory and hostile situations (Wolff et al. 2016; 2017). They tend to focus on reactionary rather than proactive strategies to manage such situations (Reupert and Woodcock 2010). Thus, while teaching assistants appear well positioned to model inclusion (de Lima, Isaac, and Kovacs 2023) through their direct, frequent interactions with students, they may not feel ready or indeed be effective for inclusion. This study investigates the perceptions of undergraduate and graduate assistants, as well as students, with respect to teaching assistants’ impact on respectful interactions in engineering classrooms.

Research questions

In light of the Culture of Respect survey and a campus-wide project to integrate sustainability across the curriculum, we wanted to know more about teaching assistants’ (both student and doctoral) perception of their role and their self-reported contributions with respect to creating inclusive, equitable learning situations in first year courses. Accordingly, we formulated the following questions to investigate how teaching assistants support the participation of all students both during their studies and going forward as professional engineers:

• Do teaching assistants perceive fostering inclusion and equity as part of their role?

• Do teaching assistants feel able to act to support inclusion and equity for students?

• Are students satisfied with teaching assistants’ contributions to classroom climate?

Methodology

Settings & participants

This study was conducted at a large, research-intensive Swiss engineering university, where efforts to improve learning outcomes for first-year students include a training program for teaching assistants that provided ideal context for this study. We consider both types of teaching assistants in our institution: student assistants (SA) who are pursuing their own Bachelor or Masters degrees and doctoral assistants (DA) who are pursuing a PhD. We use the term teaching assistant to refer to both groups together.

To examine how teaching assistants perceive their role in creating inclusive and equitable learning spaces in first year courses, and how students perceive classroom climate as co-created by these teaching assistants, we collected and analysed 3 sets of data¹ in addition to using 2 previously published institutional reports. The Culture of Respect survey (Le Duc 2022; Tormey 2021) conducted in spring 2021 established a benchmark for understanding how campus is experienced by community members including harassment, violence, and discrimination (2512 responses from across the campus community with 1187 responses from students) and from a survey of doctoral candidates (Tormey 2019) that included their teaching experiences and teaching self-efficacy (1043 respondents from 2179 doctoral candidates). Since 2022, all members of the institution are encouraged to participate in a 45-minute online training session ‘Promoting Respect’.

The data sets specific to this study are summarised in Table 1, and relate to three distinct but related populations:

• Student teaching assistants (SAs) in first year courses: To maintain the institutionally desired assistant to student ratio in first year courses, the institution employs students in their second year and above to work as assistants. Instructors of the individual courses select their own assistants, but the grade obtained in the course, overall grade, and involvement as a student are common criteria for selection. SAs are expected to interact synchronously and asynchronously with the students, but do not grade their work. SAs receive mandatory, financially compensated, pedagogical training spaced over the semester. Inclusive pedagogy was not present in the workshop descriptions and was, at the time of this study, only addressed on a single slide encouraging SAs to ensure all students benefit from a good learning environment and do not encounter discriminatory behaviour. For this study we administered one questionnaire to the SAs before the start of the training, and another one after they had completed all five hours of the training. Participating SAs appear once in the data set. NOTE: In some institutions, student assistants are called undergraduate teaching assistants (UTA). We do not use this term because our SAs includes both Bachelors and Masters students.

• Doctoral teaching assistants (DAs): Doctoral students at our institution are engaged in teaching; a 2019 survey found that almost 90% had taught in the past year (Tormey 2019). DAs are pursuing PhDs in a wide range of engineering fields in our institution. Those assigned to assist with first year Maths, Physics, and programming courses can find the material quite removed from their current studies. In addition to the duties described in the SA section, many DAs also grade students’ exams. Throughout the 2022–2023 academic year, DAs had the opportunity to participate in any of a series of short non-mandatory pedagogical training workshops about teaching in exercise sessions, labs, project or lecture format courses (Tormey, Hardebolle, and Isaac 2020). In addition to being at a more advanced level of study, this population also differs from the SAs in that participation in the workshops is based on the individual motivation of the doctoral student and is typically not actively encouraged by their supervisors. Inclusive pedagogy was not present in the workshop descriptions or workshop material, so there is no inherent predisposition of this self-selected group to be unrepresentative of the larger PhD population. For this study, DAs responded to questionnaires administered at the end of these workshops and it is possible for a DA to appear more than once in the dataset. NOTE: In some institutions, DA would be referred to as graduate teaching assistants (GTA). We do not use this term because it implies both Masters and Doctoral candidates.

• First-year students: For this study, data was collected from students enrolled in first-year Maths, Physics, and programming courses that have regular timetabled exercise-sessions, facilitated by the previously mentioned teaching assistants, where students practise applying course concepts and solving assigned problem sets.

Table 1. Data collection from student assistants, doctoral assistants, and students in our study.

Population	Institutional population in fall 2022 (approx.)	Data collection points	Data collection mode	Observations in our sample
Student assistants (SAs) in first year maths, physics and programming service courses (pursuing Bachelor or Masters degrees)	600	2022 fall; 2 time points	Paper	277
Doctoral assistants (DAs, pursuing PhD)	2000	2022; single time point	Paper + online	65
First year students	2000	2021 fall,	Online	2912
		2022 fall	Online	1076

Data collection

Information about the study was sent digitally to all the teaching assistants prior to the trainings and was additionally provided as a hard copy on the day of the training. Students’ feedback was collected as part of institutional efforts to assess instructional quality. While students and teaching assistants were invited to complete the survey, there were no incentives to respond nor adverse consequences for not responding. All data was anonymous at the point of collection.

To examine how teaching assistants perceive their role around creating inclusive and equitable learning spaces, we identified 4 actions that assistants could use to promote inclusion. As shown in Table 2, we asked teaching assistants to respond on a Likert scale both for how ‘important’ these 4 actions are to their role as assistants and also how ‘ready and able’ they felt to enact them. Additionally, student assistants (SAs) were asked to self-report instances of actually taking action to increase inclusion or address discrimination during the term and the time they spent on specific tasks that built inclusive classrooms (e.g. spontaneously starting interactions, managing disruptions). Confidential codes were employed to enable the matching of SAs’ responses from November 2022 with their September data. Finally, to examine how students perceived teaching assistants’ contributions to classroom climate, we asked first year students to respond to a prompt about their perception of assistants being respectful (fall 2021 and fall 2022).

Table 2. Legend and items for student assistant and doctoral assistant about inclusive teaching.

Prefixes	IMP	As a teaching assistant, I think it is important for me to:
	ABLE	As a teaching assistant, I feel able and ready to:
Items	Ensure	Ensure each student feels respected and safe during class time.
	Be aware	Be aware of discriminatory or biased behaviour happening in the classroom.
	Take action	Take action when a student is the target of discriminatory or biased behaviour during class time.
	Demonstrate concern	Demonstrate concern for each student’s academic and personal well-being.
Example codes composed of prefix + item	IMP-Ensure	As a teaching assistant, I think it is important for me to ensure each student feels respected and safe during class time.
	ABLE-Take action	As a teaching assistant, I feel able and ready to take action when a student is the target of discriminatory or biased behaviour during class time.

The questionnaires for teaching assistants (SAs + DAs) asked ‘If you have any comments about inclusion or discrimination in exercise sessions, we would like to hear them’ as an open-ended question. While first-year students were not asked a specific open-ended question about inclusion or discrimination, their questionnaires concluded with an opportunity to provide any comments they had about their experiences in exercise sessions.

Data analysis

Most of the data used in this study is quantitative data, allowing us to use statistical tests to look for patterns in teaching assistants’ self-reported attitudes and behaviour as well as with the experiences reported by students. Due to the non-normal distribution of our paired data, we predominantly used the non-parametric Wilcoxon signed rank test for our ordinal Likert data and Chi squared tests for categorical data. We analysed the responses to open-ended questions using qualitative content analysis methodology (Schreier 2014).

None of the authors of this paper were involved in teaching the first-year courses, or in the design or delivery of the ‘Promoting Respect’ training. Two of the authors were involved in the design and facilitation of the pedagogical training workshops for the SAs and DAs.

Results

To keep the paper succinct, we limit the results reported in this section to those directly relevant to our analysis (such as effect sizes and significance) and present the full test results in the Supplementary Information section. We start with the SA results, then DA, and conclude with students’ experiences.

Student assistants (SAs)

In September 2022, student assistants (SAs) completed the first questionnaire as they arrived for the training workshop. This questionnaire sought their perception of the importance and their readiness for 4 aspects relevant to inclusion. Student assistants’ responses on the 5-point Likert agreement scale ascribed high importance to each of the 4 elements (Figure 1; mean 4.44–4.84; SD .53 to .78; N of responses = 258-259) but reported lower readiness and ability to enact the 4 elements (mean 4.14–4.59; SD .69-.95; N of responses = 256-257; complete descriptive statistics in Table S1). The statement ‘Demonstrate concern for each student’s academic and personal well-being’ was given the lowest importance and ‘Ensure each student feels respected and safe during class time’ was ascribed the highest importance. SAs reported feeling most able to ‘Ensure students felt respected’ and least able to ‘Take action when a student is the target of discriminatory or biased behaviour during class time’. Wilcoxon signed rank test between SAs’ IMP and ABLE versions for each item revealed a statistically significant difference (p < .001) of moderate effect size (.34 < r < .51) with the exception of ‘demonstrate concern’ (Table S2). This difference shows that SAs consistently expressed higher IMP than ABLE values.

Figure 1. Student assistants’ perception of their role for inclusion (pre-survey). Refer to Table 2 for full text of the prompts.

Set of 4 paired stacked bar-plots showing most student assistants strongly agree that in their role as teaching assistants it is important for them to enact practices relevant to inclusion, but that they feel slightly less prepared to do so. N of responses between 256 and 259 per item.

We tested the interactions within the group of IMP items and within the group of ABLE items with Wilcoxon signed rank tests. Significant (p < .001) and moderate (.31 < r < .49) interactions between several of pairs of IMP items (Table S3) and ABLE items (Table S4) means that, in general, SAs’ responses to the IMP + ABLE version of the same item are more different than the differences across their response to all 4 items in either group. Put another way, we found that the responses to the four ‘I think it is important for me’ (IMP) items given by each SA were very similar to each other and different from the responses given by the same SA for the four ‘I feel ready and able to’ (ABLE) items. This suggests that IMP is one ‘factor’/construct and ABLE is another construct. Unfortunately, skewness and kurtosis of the data set rendered the application of factor analysis methods, to assess structure among the items, inappropriate.

In November 2022 (two months after the first questionnaire), we asked the same student assistants to complete a questionnaire about their behaviour as assistants during the semester. Of these respondents 59% (N of paired responses = 240) reported that they had taken actions to support inclusion (Figure 2(a)). While only 17% of assistants expressed disagreement with this item, over 20% of respondents (58 SAs) replied ‘no opinion’ to this item. This is much higher than 1.7% ‘no opinion’ for the 24 other items on the questionnaire (not included in the current analysis) that had ‘no opinion’ check boxes, which may indicate SAs’ uncertainty about how to characterise ‘specific actions to increase students’ inclusion and reduce potential discrimination’. It could also reflect a reluctance on the part of SAs to report the socially undesirable condition of not having taken any action to promote inclusion, which is at odds with the importance and readiness that SAs indicated in the first questionnaire.

Figure 2. Student assistants self-reported instances of (a) taking specific inclusive actions, and (b) participating in ‘Promoting Respect’ training.

(a) Pie chart showing that more than half of the student assistants stated that they had taken specific actions aimed at increasing inclusion and reducing discrimination (N of responses = 240). (b) Pie chart showing that 57% of the student assistants had taken either the Promoting Respect training or some similar training (N of responses = 238).

In the November questionnaire, we also asked student assistants if they had participated in the ‘Promoting Respect’ training which they were strongly encouraged to attend (Figure 2(b)). Between the SAs who said they had followed this training (n = 105) and those who did another training (n = 31), more than half of SAs had received some training about bias and discrimination. Chi squared test of the association between SAs’ reported actions (reverting to parametric tests since they are more powerful and this data is normal) to increase inclusion and participation in the ‘Promoting Respect’ training found a substantive relationship of moderate effect size, with those taking action more likely to have attended (Cramer’s V .256, p < .001, Table S5).

In the November questionnaire, SAs reported the time that they had devoted to specific actions during the exercise sessions. Despite managing classroom disruptions being a common concern of new instructors (Figure 3), almost 70% of assistants ‘never’ spent time on it during the semester and only 11% of assistants reported spending 5 or more minutes per session on classroom management. Of the three activities, managing disruption occupied the least time for SAs and talking to students about science and their studies was afforded the most time. Spearman correlations of the association between SAs who reported they ‘took actions to increase inclusion’ and time spent in the final weeks of the semester ‘spontaneously starting interactions’, ‘talking to students about their studies, EPFL, and science/engineering in general’ and ‘managing disruptions’ found no significant associations (Table S6). As these 3 actions would be among those likely to increase inclusion, it appears that acting to increase inclusion in both proactive and reactive ways does not represent a significant amount of SAs’ time during exercise sessions, or SAs identify something outside these 3 actions as their main contribution to inclusion.

Figure 3. Student assistants’ November report of their time managing disruptions and interacting with students.

Three treemaps exploring how much time student assistants spent on specific activities. 70% of student assistants did not spend any time on ‘Managing students’ disruptive behaviour’, while 11% spent 5 or more minutes on it. 62% spent no time or less than 5 minutes on ‘Talking to students about their studies’, while 12% spent more than 15 minutes on it. 61% spent no time or less than 5 minutes on ‘spontaneously starting interactions’, while 13% spent more than 15 minutes on it. N of responses between 232 and 239 per item.

We also looked for associations between SAs’ responses to the September survey items related to the importance and their readiness to address inclusion and their November responses about their actions during the preceding weeks (Table S7). Fisher exact tests of association between SAs who did/ did not take ‘action specifically aimed at increasing students’ inclusion and reducing potential discrimination’ and their responses to IMP + ABLE items from September found statistically significant (.05 < p < .003) but weak relationships (.1 < Somer’s d < .2) with all 4 IMP and 3 ABLE items (Table S8). This means that SAs who ascribed higher importance or reported greater readiness in Sept were only slightly more likely to enact inclusive teaching during the semester.

There were several open-ended questions posed to SA in the November questionnaire, including ‘What obstacles to being an excellent teaching assistant have you encountered?’, ‘What impact do you, as a teaching assistant, have on students and on student learning?’ and ‘If you have any comments about inclusion or discrimination in exercise sessions, we would like to hear them’. While these items predominantly relate to the pedagogical aspects of the training and will be reported elsewhere, there are a few responses relevant to this study.

While the majority of SAs’ comments (n = 218) on the obstacles they had encountered focused on pedagogy, we found 18 comments that mentioned ideas related to building inclusive classrooms, such as "Student having a lot of difficulty no matter how I (or other teaching assistants) explain, he gets frustrated and gets ‘angry’ at us". and "To know how to manage student’s disruptive behaviour". This low number is coherent with the low frequency of problematic incidents reported in the quantitative data and with SAs’ relatively high self-assessment of their ability to manage classroom situations.

When describing their perceived impact on students and student learning we saw a similar pattern in SAs’ responses. Most of the ideas (n = 214) focussed on pedagogical aspects, however 26 SAs mentioned that they had helped build healthy learning environments or developed good relationships with their students. Representative comments include "I think I motivated some students and established a healthy environment" and "Quite good, I developed a nice relationship and I feel they are not afraid of asking silly questions (supposing they exist)".

Only 16 SAs responded to the open-ended prompt about inclusion, mostly to report not witnessing any acts of discrimination or harassment. One SA wrote about the concrete action that they had taken to prevent it, stating "I emphasised on it in our session that we don’t support and tolerate these actions". One SA pushed back against efforts to be more inclusive, writing "There are no problems of inclusion or discrimination, nobody should bring any saying they want to fix them".

Doctoral assistants (DAs)

On completion of the optional pedagogical development workshops, doctoral teaching assistants (DAs) were asked to complete a questionnaire that included the same items from the SA questionnaire about their perception of the importance and their readiness for 4 aspects relevant to inclusion. DAs’ responses on the 5-point Likert agreement scale ascribed high importance to each of the 4 elements (mean 4.95–4.57; SD .21 to .68; N of responses = 64–65) but reported lower readiness and ability to enact the 4 elements (mean 4.46–4.12; SD .66–.94; N 64–65) (Figure 4; Table S9). The statement ‘Demonstrate concern for each student’s academic and personal well-being’ was given the lowest importance and the other 3 statements were all evaluated to be of very high importance. We observe the same pattern as the SA, with DAs reporting feeling most able to ‘Ensure students felt respected’ and least able to ‘Take action when a student is the target of discriminatory or biased behaviour during class time’. We used Wilcoxon signed rank tests, since the 4 IMP items were not normally distributed, to assess the difference between the importance and readiness DAs reported for each item. The statistically significant (p < .002) and large effect size (.60 < r < .67) for all items except ‘demonstrate concern' which has a moderate effect size (r = .38; Table S10) indicate that DAs rate the importance of these inclusive actions appreciably higher than their ability to enact the actions. To take a specific example, DAs’ responses about the importance of ‘Ensure each student feels respected and safe during class time’ were consistently and significantly higher than their reported ability to enact this action.

Figure 4. Doctoral assistants’ perception of their role for inclusion. Refer to Table 2 for full text of the prompts.

Set of 4 paired stacked bar-plots showing most doctoral assistants strongly agree that in their role as teaching assistants it is important for them to enact practices relevant to inclusion, but that they feel markedly less prepared to do so. N of responses between 64 and 65 per item.

To understand if the DAs had participated in any trainings which would increase their ability to perceive issues or to enact practices relevant to inclusion, we also asked the doctoral assistants if they had participated in the ‘Promoting Respect’ training offered within the institution, and less than a quarter of the DAs reported having done so (Figure 5). Given the interest DAs expressed in contributing to a good class climate, this may indicate that information about these trainings is not reaching doctoral candidates.

Figure 5. Doctoral assistants’ reported training in inclusion.

Pie chart showing that 75% of doctoral assistants had not taken Promoting Respect training or some similar training. N of responses = 65.

Wilcoxon signed rank tests within the 4 IMP items (Table S11) reveal significant (p < .002) and moderate (.39 < r < .47) interactions between ‘demonstrate concern’ and each of the other 3 actions. This means that, in general, DAs’ IMP + ABLE responses to the same items are more different than their responses across all 4 IMP items. We conducted a similar analysis within the 4 ABLE items (Table S12) found only 2 significant (p < .01) and moderate (.30 < r < .39) interactions. This smaller DA sample does not contradict the observation with SA data that suggests that ABLE is one ‘factor’/construct and IMP is another construct.

The overall picture created by this data is that DAs overwhelmingly see inclusion as an integral part of their role however report markedly lower expression of ability/readiness to fulfil it. As with student assistants, DAs scored ‘demonstrate concern’ lower on importance than other items, however roughly the same in terms of ability/readiness. The difference in DAs perception of the importance versus their readiness of the 4 actions is aptly demonstrated by ‘taking action when a student was a target’ which had only 2 neutral/disagree responses but 16 combined across the same categories for ‘ready/able’.

Of the 11 responses from DAs to the open-ended prompt, the most commonly expressed sentiment was need for additional training as illustrated by the comment "Sometimes there is a discrepancy between what we would like to do (ex: reduce inequalities) and what we can do within the context of a classroom. More guidance on how to behave in case of discrimination would be appreciated". A few responses revealed that not all DAs see inclusion as being part of their role, as illustrated by the comment "I don’t feel it’s my place to intervene with intra class discrimination". Comments from 2 DAs expressed that they were concerned about discrimination they had encountered in graduate school although they did not perceive students might also be targets, as exemplified in the comment "Bachelor/Master’s student discrimination can happen but I don’t see it that often in EPFL settings. However, as a PhD student, I want to say please also Iook at this as discrimination happen at that level".

Students

It is relevant to note that no distinction was made between ‘student assistants’ and ‘doctoral assistants’ when asking students about their perceptions, however student assistants greatly outnumber doctoral assistants in first year courses.

Students overwhelmingly reported that their assistants behaved in ways that promoted mutual respect (Figure 6). Chi squared comparison of students’ perception in fall 2021 and fall 2022 found a very slight yet statistically relevant improvement in respectful behaviour (Cramer’s V .07; p < .001). Given the overwhelmingly positive data from 2021, we cannot expect a large shift in the following year, so this significant but weak relationship is coherent with a favourable trend towards more inclusive classrooms. Chi squared analysis with assistants’ behaviour and department, language of instruction (English, French), and gender of teacher, found no statistically valid correlations. The effect of class size was moderate (Cramer’s V .145; p = .008) with fewer issues in smaller classes, which is interesting as institutional guidelines set a consistent assistant to student ratio for courses. In a finer analysis of the problematic areas (Table 3), we see that 0.2% of students characterised the lack of respect as highly problematic (‘strongly disagree’, 7/3971 student responses) and that issues were concentrated in specific courses (Cramer’s V .12; p = .01). The number of courses where assistants were perceived as problematic dropped between 2021 and 2022 although the percentage of respondents expressing negative opinions increased. Given the massive drop in response rate between 2021 and 2022, which we attribute to student fatigue with online questionnaires, the 1% increase in negative experiences is not concerning. Indeed, the overall trend is more respect from assistants. Finally, chi squared comparisons of the 2021 edition of each course to its 2022 edition found a small effect size indicating that the 2021 course atmosphere had some correlation with the atmosphere in 2022 (Cramer’s V .12; p = .011, Table S13).

Figure 6. Students’ perception of respectful behaviour from assistants in first year courses.

Stacked graph showing that more than 90% of the student responses indicated that teaching assistants behaved in a way that promotes mutual respect. N of responses = 3971.

Table 3. Distribution of problematic behaviour from student assistants in first year courses.

	2021 (Numbers)		2022 (Numbers)		Combined (Percentages)
	Courses (N = 38)	Responses (N = 2912)	Courses (N = 38)	Responses (N = 1076)	Courses (N = 76)	Responses (N = 3989)
Strongly disagree	5	5	2	2	9%	0.2%
Disagree	14	22	15	17	38%	1.0%
Total (n)	17	27	16	19
Total (%)	45%	0.9%	42%	1.8%	43%	1.2%

Number and percentages of negative responses to the question ‘The assistants behave in a way that promotes mutual respect’.

Of the almost 600 comments provided by the students to the open-ended prompt about their experiences in their exercise sessions, only 39 were relevant to inclusion, discrimination and class climate. Of these the majority (n = 31) were very positive about the teaching assistants’ behaviours, describing the teaching assistants being friendly, courteous, caring, kind, patient, and respectful. An example of such a comment (translated from the original French) "Very numerous assistants, available, courteous, competent and caring: by far the most pleasant exo sessions I've had this semester!" Nine comments mentioned disrespectful behaviour from teaching assistants, including behaviour that was demotivating or condescending as illustrated by the comment "One assistant is quite aggressive, which doesn’t really motivate the students to ask questions".

Discussion

Our study shows that teaching assistants see inclusive actions as an important part of their roles. While student teaching assistants (SAs) were more confident than doctoral teaching assistants (DAs) about their ability to act to promote inclusion (with more than half of the SAs reporting that they took specific actions to increase inclusion), both groups rated the importance of most items moderately higher than their capacity to enact them. Although their responses indicate that taking action related to discriminatory behaviour was more difficult in comparison to other 3 actions, they were nevertheless quite positive about their ability and readiness to take action. Previous research has also shown that teaching assistants are thinking about inclusive learning practices, including using active learning strategies and being empathetic (de Lima, Isaac, and Kovacs 2023), and that they consider increasing student belonging to be a part of their role (Perlmutter, Salac, and Ko 2023). This shows that student assistants can play an important role in developing inclusive educational practices – which in turn have been shown to improve student learning outcomes (Hardebolle et al. 2022; Preszler 2009; Theobald et al. 2020).

The overall positive opinion of students about their teaching assistants indicates that most students are satisfied with teaching assistants’ behaviours in promoting an inclusive and respectful classroom environment. While the course size should not affect the assistant to student ratio, smaller classes were found to have moderately fewer issues with disrespectful teaching assistants. This is important as positive relationships with teaching assistants have been documented to contribute to student satisfaction (Talbot et al. 2015), student learning outcomes and academic performance (Alzen, Langdon, and Otero 2018; Barrasso and Spilios 2021; Knight et al. 2015; Sellami et al. 2017), reducing inequalities in academic performance (Preszler 2009; Van Dusen, White, and Roualdes 2016; Van Dusen and Nissen 2020), and enhancing belonging (Clements et al. 2022; Perlmutter, Salac, and Ko 2023). The importance of considering attributes like ‘empathetic towards struggling students’ and ‘active listener’ when hiring teaching assistants was shown to result in students expressing higher opinions of teaching assistants who were hired in a process that specifically prioritised these traits (among others) (Leyzberg, Lumbroso, and Moretti 2017).

Yet, even with the overall positive picture for this study, the Culture of Respect Survey at the same institution identified classrooms as a main location of problematic behaviour with half the students expecting that by-standers would not intervene to support them. Our results show that it indeed seems difficult for teaching assistants to take action and respond to discriminatory behaviour. While not the focus of our study, elucidation on any particular contextual reasons that give rise to lack of confidence that witnesses will defend targets of discrimination in the institution would be useful. With only 9 out of 600 student comments referring to discriminatory teaching assistants’ practices, it appears that cases of teaching assistants disrespectful behaviour towards students are not common. However, while it might be rare that a teaching assistant acts in a disrespectful manner, even the subtlest signs (microaggressions) of discrimination or disrespect can have a negative impact on students’ learning and persistence in the field of engineering (Bahnson et al. 2022; Dececchi, Timperon, and Dececchi 1998; Marsden et al. 2016).

Prior research in our institution (Isaac, Kotluk, and Tormey 2023) suggests that engineering students working in teams found it easier to apply proactive strategies to reduce opportunities for bias and discrimination than reactive strategies to respond to problematic behaviour. This is coherent with the assistants in the current study expressing lower ability to ‘take action’ than the 3 other inclusive practices. Furthermore, this is consistent with previous work that found that while most people believe they will stand up to prejudice, less than half of people will actually speak up (i.e. take action) when confronted with discriminatory actions in practice (Swim and Hyers 1999). In terms of SAs’ reflections on activities they spend most time on, the high number of SAs reporting that they ‘never’ managed students’ disruptive behaviour could arise both from SAs not perceiving the disruptive behaviour as such, or from having trouble responding to such behaviour. Finally, the opinions of the students that were assisted by teaching assistants in our study is quite clearly positive, with less than 10% of students reporting that assistants did not behave in a way that promoted mutual respect. While these findings support the opinion of high value and importance for a respectful culture among the SAs and DAs, it is still unknown whether and to what extent students feel supported by assistants when they are exposed to or are a victim of discriminatory behaviour. A further caveat to our rosy interpretation is that the current relatively low diversity in engineering classes (NCSES 2023) could mean that many students are not readily able to identify problematic situations and interactions. For instance, a study at a large German engineering school found female students were considerably more likely to have identified instances of gender-based discrimination occurring on campus than male students (Marsden et al. 2016). Thus, more complete demographic information about the teaching assistant population at our institution would be a priority before concluding that teaching assistants contribute a safe space with their behaviour.

A large proportion of the discrimination and harassment that students face during their university studies happens in classrooms, and most of it is perpetrated by other students (Le Duc 2022; Tormey 2021). In our institution, SAs are typically present in a relatively high ratio with respect to first year students. Our study has found these SAs consider inclusion to be a significant part of their role, with more than ¾ reporting that they feel able to take action and a similar percentage reporting that they took action during the semester to increase inclusion in the classroom. While this sounds like an ideal set up for SAs to act to increase inclusion, it is at odds with reports of instances of harassment and discrimination in classrooms reported in the campus survey. The disparity that we observe obliges us to reflect on the way the assistants have perceived the question itself and what they define as action against discriminatory behaviour, as well as what they perceive as being discriminatory. The comparatively high percentage of ‘no opinion’ responses from SA to the item on having taken action underlines a potential point of uncertainty with the wording of our question. The Culture of Respect survey had very specific and clearly described items of discrimination and bias (including abuse, verbal, physical and sexual violence, and different forms of discrimination based on race, gender, etc.). In our survey we ask a more general question, hence there is an information gap in terms of how assistants understood these general terms while answering the questions. However, we have also seen from the responses to open ended questions that SAs and DAs can be both resistant to the idea that discrimination occurs in their institution and fail to perceive its presence. That novice teachers find it difficult to identify behaviours and conditions that can be discriminatory and/or lead to discrimination and harassment (Wolff et al. 2016; 2017) can propagate a culture of minimising or denying the discrimination, thereby further impeding the development of inclusive classrooms.

Additionally, we have identified a methodological limitation. Responses, particularly from SAs, to several survey items skewed heavily towards ‘strongly agree’, indicating that the items were ‘too easy’. This tendency also reflects the implicit ‘correct’ and socially desirable response of saying that these inclusion actions are important. The result is that some variation within our population may have been obscured and therefore prevented exploratory factor analysis of the underlying structure within survey items. Future work should either use a wider Likert scale (i.e. 7-point), reformulate the survey items to make them more selective, and separate data collection from the training to reduce potential pressure to respond to meet their perception of the trainers’ expectations. Testing the tool qualitatively, or undertaking a complementary qualitative approach, could support better understanding of the data. Finally, our data explored only one dimension of student interactions in the classroom, that between students and teaching assistants. The two other dimensions, teacher-student and student-student, should be addressed in future work. This latter pair is potentially more problematic, as the Culture of Respect survey showed that students were most likely to be perpetrators of discrimination, sexual violence or harassment.

Conclusions & recommendations

The Culture of Respect survey highlighted the scale of the problem we face in reaching the sustainable development goals related to equity in engineering schools and provided us with a valuable baseline to structure and measure future work. The current study of teaching assistants’ perceptions provides a more positive picture, highlighting that student assistants and doctoral assistants see addressing these issues as integral to their work and that they mostly feel able to enact practices to increase inclusion. Additionally, responses from first year students show that these teaching assistants do create a climate that is conducive to learning and respect. However, although it is a minority, we must not ignore the non-inclusive perspectives of some of the teaching assistants, as well as the discrimination and harassment faced by some of the students. As it is impossible to predict when a problematic situation will occur, we recommend ensuring that assistants both have training on how to respond and also that the institution clearly communicates to assistants that taking action is an integral part of their responsibilities in supporting the campus learning climate. Together, this should increase the probability that problematic situations are handled competently when they occur. As the overall situation with respect to teaching assistants is relatively positive, proposing more hours of training or other initiatives focusing on assistants would likely consume considerable resources without providing major improvements. Accordingly, we conclude by reviewing recommendations that are not limited to teaching assistants.

While there are multiple resources that provide recommendations on how to prevent and address discrimination and harassment in engineering (NASEM 2018a; 2022), based on our study we highlight three:

• Supporting those who experience discrimination and harassment

  Targets of discrimination and harassment often do not know about the services available to them or find them too rigid to meet their needs. It is therefore important to develop systems that are target-centric (and not institution-centric) and to increase awareness and reduce barriers for people accessing them.

  Our study showed that doctoral teaching assistants strongly see building inclusive classrooms as part of their role and therefore may be prepared to develop their capacity to act as ‘first responders’ in situations where discrimination occurs. This would empower them not only to act as allies, but also to orient students towards help and resources.

• Improve transparency and accountability

  Two common reasons why targets do not report incidents are not knowing how they will be treated and not believing that the perpetrator will be subject to consequences. It is therefore important that institutions are transparent about what happens after reporting (including addressing any expectations from the target and provisions to ensure that there is no retaliation). It is also important to maintain a publicly available (sufficiently general to be anonymous) record to provide accountability about the incidents reported and the consequences faced by the perpetrators.

  We also strongly recommend that institutions consider conducting and sharing a survey similar to the Culture of Respect survey. While an institution may have concerns about the potential adverse image that might arise as a consequence, solid baseline data and understanding of each institution's individual issues is fundamental to implementing and evaluating initiatives designed to address equity and sustainability on campus, as well as modelling transparency and accountability.

• Provide training that is relevant and practical

  In our study, we saw that the doctoral assistants wanted additional training and therefore targeted communication about existing opportunities would likely have good results. DAs have a unique positionality in the classroom, with sufficient status to intervene as ‘first responders’ in cases of biased interactions yet small hierarchical difference to act as near role models for constructive and inclusive behaviours. Such training would empower them to provide supportive proximity to targets, lower the barriers to reporting and access to support, as well as provide early intervention to prevent escalation of harassment and discrimination. As both SAs and DAs consider inclusivity to be a part of their role, initiatives that encourage them to contribute more to inclusion takes advantage of an already open door rather than knocking on closed doors.

  Another strategy is to include respect and proactive strategies to address adverse situations into other training offered by the institution. This serves two main purposes: (i) to increase the coherence of the message at an institutional level and emphasise the connection and relevance of respect and safety in all aspects of academic and personal life (transformative change – Sterling 2004; Tien, Namasivayam, and Ponniah 2019), and (ii) ensure that such training is not only being accessed by people who are already interested and invested (avoid preaching to the choir).

  It is, however, important to be attentive and iterative when implementing training. Previous research has shown that pedagogical trainings (such as those for SAs and DAs in this study) encourage teaching assistants prioritise inclusive student-centered practices such as active learning (Isaac and de Lima 2022; Ruder and Stanford 2018; White et al. 2012). However, they can inadvertently reduce the importance assistants afford to other ideas, such as being empathetic to students (de Lima, Isaac, and Kovacs 2023). The workshop described by Isaac, Kotluk, and Tormey (2023) for educating engineering students to manage bias and discrimination in their project teams is an excellent example of how to support students’ to work together more equitably, as their follow up survey with participants found they had actually employed the strategies.

A sustainable education is one that is ‘inclusive and equitable’ (UN 2015) and includes aspects both on the product/provision (teaching about sustainability) and on the process/paradigm (teaching sustainably) (Sterling 2004); the former is incomplete without the latter. Sustainability requires integrated, interdisciplinary, contextualised, and systemic thinking (Kolmos, Hadgraft, and Holgaard 2016; Sterling 2004). It requires increased diversity in voices, so engineering programs must act to address issues contributing to the well-documented under-representation of some genders, sexual orientations, and ethnic groups in the engineering profession and in engineering education (Barnard et al. 2012; Lichtenstein et al. 2014; Meadows et al. 2015; True-Funk et al. 2021). We are pleased to share the rich perspective on inclusion in first year engineering classrooms provided by student assistants, doctoral assistants and students in this article.

Building sustainability into engineering education requires a transformative change (Sterling 2004; Tien, Namasivayam, and Ponniah 2019), but will eventually help to enhance sustainable practices among students and ultimately improve the capacity of graduating engineers to work sustainably and contribute to realising the UN SDGs (UN 2015). It is therefore imperative that engineering programmes assess the impact of discrimination and harassment, and evaluate the results of programs targeted at reducing these issues, as an integral part of their pursuit of sustainable education and ensuring the renewal of the engineering profession that reflects a rich diversity of experiences and perspectives.

Acknowledgements

The authors would like to thank Dr. Ilya Eigenbrot for helping to collect the data, and Dr. Roland Tormey & Dr. Ingrid Le Duc for insightful feedback, and the teaching assistants and students for responding to the surveys.

Author contribution

Joelyn de Lima: Conceptualisation, Methodology, Investigation, Formal analysis-qualitative, Data Curation, Writing – Original Draft, Writing – Review & Editing, Visualisation, Supervision. Siara Isaac: Conceptualisation, Methodology, Investigation, Formal analysis – quantitative, Data curation, Writing – Original draft, Writing – Review & Editing, Visualisation. Helena Kovacs: Methodology, Writing-Original Draft, Writing – Review & Editing, Project administration.

Disclosure statement

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

Additional information

Notes on contributors

Joelyn de Lima

Joelyn de Lima is a discipline-based education researcher. She is a scientist & pedagogical advisor at the Teaching Support Center at the Ecole Polytechnique Fédérale de Lausanne. Currently her research and practice are focused on increasing inclusivity in higher education. Her background has given her a unique blend of perspectives – in terms of culture (She has lived, worked, and taught on 3 continents), theoretical grounding (natural sciences and education), and practice (research & teaching, formal & informal education).

Siara Isaac

Siara Isaac is a researcher and project manager at the Centre for Learning Sciences at the Ecole Polytechnique Fédérale de Lausanne. Her recent work, including her 2022 book (Facilitating Experiential Learning in Higher Education – Teaching and Supervising in Labs, Fieldwork, Studios, and Projects) focuses on how students develop transversal skills such as teamwork and risk assessment in science and engineering contexts. Siara has previously worked as teaching advisor and has taught in Canada, China, France and Switzerland. She holds a M.Sc. in Chemistry from McGill University and a PhD in Educational Research from Lancaster University.

Helena Kovacs

Helena Kovacs is a project manager for transversal skills at the Transversal Skills and Career Centre at EPFL, where she is in charge of developing, implementing and assessing the framework for future skill sets of students and graduates. Her previous work revolved around exploring changes in teaching methods and approaches, as well as on teaching and learning transversal skills. This, together with continuous involvement in research, supports her current focus on translating knowledge and evidence related to educating holistic engineers and equipping students for the uncertain futures. She holds a BSc degree in Community Youth Work, and a MA in Lifelong Learning: Policy and Management. Her doctoral work examined teacher learning in innovative learning environments in the context of educational reforms.

Notes

1 Institutional research ethics board approval was accorded for all surveys.