Effectiveness of a Bystander Intervention Training Program to Interrupt Gender and Race Bias Among U.S and Foreign Faculty

Abstract

This study evaluates the effectiveness of a Bystander Leadership™ Program (BLP) designed to mitigate gender and racial biases among faculty at a research university. The BLP workshop, grounded in comprehensive local research, provided faculty with practical intervention skills to address observed biases in academic settings. A quasi-experimental design was employed, involving tenured and tenure-track faculty (N = 253) from STEM and Social and Behavioral Science (SBS) departments. Participants were surveyed before and three months after the workshop (Intervention Group) or three months apart (Comparison Group), assessing five variables: notice bias, self-efficacy, perceived benefits of action, perceived risks of action, and actual intervention in bias incidents. The findings reveal that the Intervention Group exhibited significant increases in self-efficacy and actual interventions from pretest to posttest, unlike the Comparison Group, which showed no changes. Notably, foreign-born faculty initially had lower notice bias scores compared to U.S.-born faculty. Post-intervention, notice bias significantly increased among foreign-born faculty in the Intervention Group. Furthermore, a 60% faculty participation rate was identified as the threshold for perceiving departmental endorsement of inclusive practices. These results underscore the BLP’s potential to enhance faculty’s confidence and proactive engagement in addressing bias, particularly among foreign-born faculty, and highlight the importance of broad participation for fostering inclusive academic environments.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10755-024-09765-x.

Many higher education institutions have attempted to increase the gender and racial diversity of faculty to bring the institutions in line with their stated goals of being merit-based and inclusive (Stewart & Valian, 2018). One such long-standing effort to diversify the faculty has been pursued since 2000 by the federally funded National Science Foundation ADVANCE program, with the specific aim of increasing the number of women faculty within science, technology, engineering and math (STEM) fields (Laursen & Austin, 2020).

Successful diversity, equity and inclusion (DEI) programs frequently have focused on overcoming two forces that sustain the interests of dominant groups: institutional structures and homophily within groups. Formal and informal institutionalized structures perpetuate the exclusion of women and underrepresented minorities, such as the tendency to hire faculty from a very small number of elite university doctoral programs across all disciplines (Wapman, et al., 2022). Norms established by the dominant homogeneous group that become institutionalized as indicators of success, such as what areas of study within a discipline are significant, may marginalize “outsiders” or minority group members if they pursue new scholarly topics (Settles et al., 2022).

Homophily refers to the tendency to seek out and bond with others who are like us (McPherson et al., 2001). The result is that people’s personal networks tend to be homogeneous regarding many sociodemographic, behavioral, and intrapersonal characteristics. Homophily limits people's social worlds in a way that has powerful implications for the information they receive, the attitudes they form, and the interactions they experience. Homophily in race and ethnicity creates the strongest divides in our personal environments, with homophily of age, religion, education, occupation, and gender contributing to further divisions. Homophily of nationality and language also divides groups. The tendency for groups to form around social identities such as gender, race, or nationality, often leads to segregation in academia among research groups and in other social activities that limits cross group contact (Chesler et al., 2005).

If hired, candidates with social identities different from most of the department, may still be excluded from departmental subgroups, contributing to their “outsider” status or a sense of “not belonging” that can affect career success (Settles et al., 2022; Zambrana et al., 2017). The challenge of inclusion, then, is to identify and change behaviors, practices, or structures that exclude or devalue the individuals that the institution seeks, such as historically underrepresented faculty (i.e., women and U.S. underrepresented minority groups or URMs). Departmental colleagues are an important group for intervention because departmental faculty serve as peers, evaluators, mentors, and on hiring committees, etc. Thus, individual faculty can play an important role in increasing inclusion at both the individual and departmental level. To do so, they must be cognizant of their own and other’s interpersonal interactions and notice the impact of often unintentionally exclusionary or biased behaviors or institutional structures, values, or policies. Furthermore, they must be prepared and willing to act to address these issues. Stewart and Valian (2018) emphasize that:

“everyone can do something [to support diversity and inclusion], no matter what position they occupy in the university. Everyone can identify an area that needs improvement. Everyone can work effectively to realize and maintain those improvements. Improvements that become part of the way a department, school, or institution operates are improvements that may last….changes that are part of a team are more likely to endure than changes as a result of one person’s action. The more people involved in making a change, the more likely it is that the change will stick” (p. xvii).

In the research reported here, we describe the results of an innovative research project funded by the National Science Foundation ADVANCE Institutional Transformation grant program (HRD-1629889) to increase the number of women and minority faculty in Science, Technology, Engineering and Mathematics (STEM), and the Social and Behavioral Sciences (SBS). The research illustrates the effectiveness of a bystander intervention training program to notice and interrupt biased interactions among faculty and to influence departmental faculty to support the professional development of all faculty, regardless of their backgrounds, race, gender or ethnicity.

Bystander intervention is a broad term that describes the actions (or inactions) in a potentially urgent situation when the bystander is present but not initially involved (Mazar, 2019). A bystander’s choices may include stepping into the situation physically or verbally, involving someone else (such as an authority figure), or doing nothing. Intervention training prepares the bystander to first, notice bias, and then act to interrupt the event and its consequences in a constructive manner.

The institutional study site was a Carnegie-classified very high research university in the southeast with 55,000 students, of which sixty-four percent were Hispanic1¹ and twelve percent were Black. In 2016, when the study began, seventeen percent of the STEM tenure-track faculty were women and eleven percent of STEM faculty were faculty of color. An assessment of the demographics of the STEM faculty revealed that a large proportion of engineers were foreign-born faculty (72%), including Asian (47%), Middle Eastern (18%), and Eastern European (7%). Further observation made us aware that many research groups shared a non-English language (e.g., Farsi, Mandarin). The pattern of foreign-born faculty recruiting same-nationality doctoral students and new faculty hires appeared to be an unstudied barrier to diversifying faculty and research teams (Rose & Farhangi, 2016). Thus, a variable that was of particular interest here was the effectiveness of diversity training with foreign-born faulty.

In sum, three research questions guided our work. First (Q1), compared to a group of faculty that did not participate in intervention training, would the intervention result in a significant increase in notice bias, self-efficacy (i.e., confidence in intervening); assessments of the benefit of taking action (or reduction in the risk), and take action as assessed three-months after the workshop? Second (Q2), would the intervention result in an increase in notice bias among foreign-born faculty? Third (Q3), what proportion of faculty participation within a department was necessary to affect departmental practices?

Bystander Intervention Training

A major focus of bystander Intervention training has been to use it to lower interpersonal and sexual violence victimization rates (Coker et al., 2016; Senn & Forrest, 2016). Coker et al. (2016) compared interpersonal violence victimization rates at three public university campuses. One of the campuses implemented a violence prevention bystander intervention training for students; students at the other two campuses had no training. The campus with the Intervention training had a 17% lower interpersonal violence rate than the comparison campuses. Senn & Forrest (2016) found a significantly lower incidence of rape, attempted rape, and nonconsensual sexual contact among a group of college students that attended a sexual assault bystander intervention workshop compared to a control group of students that did not attend.

Racial bias also has been successfully reduced using bystander intervention training (Nelson et al., 2011), as has school-based bullying (Polanin et al., 2012). Training that includes an experiential, embodied behavioral change approach is particularly impactful: repeated practice sessions throughout the day help participants integrate new concepts and move from cognition to action by experimenting with different ways to take action (Scully & Rowe, 2009). Bystander intervention training also supports organizational change by engaging numerous individuals in a similar behavioral change process at the same time (Collins et al., 2021).

According to Latané and Darley (1969), bystander intervention requires choosing a single course of action through a complex set of possible actions. The first step is for the bystander to notice that an event is causing or will cause harm and interpret the interaction as requiring action. Then, the bystander must take responsibility to act (i.e., lead). Here the self-efficacy or experience and skill of the bystander is important. Self-efficacy is defined as the “belief in one’s capabilities to organize and execute the course of action required to produce given attainments” (Bandura & Wessels, 1979, p. 3). Increased self-efficacy is an important outcome in many intervention trainings (Chen et al., 2021). Having a range of behaviors from which to choose increases confidence to act. Next, the bystander must decide what action to take, which involves a benefit- risk analysis. Last, the bystander must take action (Latané & Darley, 1969).

Our first research question (Q1) concerned the impact of an intervention training program, Bystander Leadership™, on faculty’s ability to notice instances (of gender or race) bias, to increase self-efficacy (confidence in intervening), to “lead” by taking responsibility to intervene, to choose an action that increases the benefit (or reduces the risk) of intervening (benefit-risk analysis), and to take action to intervene when compared to a group of faculty that did not attend the workshop.

Patricia Devine and colleagues pioneered the use of a bias habit intervention training program to break the race bias habit among students (Devine et al., 2012). In their twelve-week longitudinal study, 91 students who received the intervention showed increased self-efficacy after the intervention and were more likely to predict that they would intervene in such instances. Carnes, et al. (2015) applied the breaking the bias model to gender bias using a large sample of 2,290 faculty from 92 departments at one university. Results indicated that their intervention program significantly increased gender equity promoting behaviors at three months post-intervention (Carnes et al., 2015). This intervention was the first of its kind: a randomized controlled study intended to change faculty behavior and improve department climate in ways that would promote gender equity. The intervention was a 2.5 hour interactive workshop that aimed to increase awareness [i.e.,notice bias] of gender bias in academic medicine, science and engineering and to promote motivation, self-efficacy, and positive outcome expectations (i.e., increased benefits or lowered risks) for consistently acting in ways that supported gender equity (Carnes, et al., 2015). Effective habit-breaking interventions must both increase the notice[ing] of problematic behavior and motivate people to learn and deliberately practice new behaviors until they become habitual.

Bystander intervention training also has been used successfully to reduce microaggressions among faculty. Haynes-Baratz and her colleagues (2021) developed a two-hour bystander training workshop to promote intervention in response to microaggressions that may occur in faculty interactions and recruited twenty-one percent of all full-time faculty in their institution as participants. Post workshop, participants endorsed higher levels of likelihood of bystander intervention and higher levels of self-efficacy to deal with microaggression than at the pretest; scores remained significantly greater at 12-weeks post-training, as well.

For the current study, we selected five measures used by Carnes et al. (2015) to assess the impact of the Bystander intervention training on faulty, because Carnes also was an intervention study aimed at reducing bias using a population of full-time faculty at a research institution, similar to the current study. The included measures that were relevant to our training were those used by Carnes et al. (2015) that also mapped loosely onto four of the steps of bystander intervention (Latané & Darley, 1969) (i.e., notice, interpret, lead, decide[what to do] and act) and included notice bias (e.g., noticing and interpreting the situation as requiring action), self-efficacy (i.e., leadership/confidence to take action), decide what to do (e; g;, based on an expected positive benefit or a reduced risk of intervening), and take action to intervene in instances of bias (i.e., act). These variables were assessed in a pretest before the workshop and a posttest given three months after the workshop.

Foreign-Born Faculty

Our rationale for the study of foreign-born faculty was based on previous research on workplace friendship, networks, group norms, and in-group and out-group relations (Brown & Zagefka, 2011; Hortulanus et al., 2006; Merritt et al., 2010; Ragins & Gonzalez, 2003; Vaccaro, 2012). We anticipated that foreign-born men faculty might less often notice bias or be less concerned than U.S. men faculty about gender and race bias as experienced by women and underrepresented U.S. minority faculty, reflecting a lack of familiarity with U.S. racial groups or a preference for birthplace gender and racial values. Notice what is or will cause harm (e.g., bias) was identified by Latané and Darley (1969) as the critical first step to bystander intervention. Corte and Amrein-Beardsley (2023) reported that an educational intervention they did to increase participants’ ability to notice racial bias was an important precursor to motivating faculty participants to “work to mitigate [its] negative effects” (p. 143).

At our study site, foreign faculty represented about 62% of the tenure-ladder STEM faculty. The STEM foreign faculty were mostly Asian, Middle Eastern, or Eastern European men (Rose & Farhangi, 2016). This pattern suggested that it was important to develop ways to educate foreign-born faculty about bias and disadvantage accruing to historically underrepresented groups in the U.S. (i.e., women and URMs). Foreign-born faculty may not be familiar with the history of, or had much interaction with, U.S. women scientists or historically underrepresented U.S. Black or Hispanic people. Thus, they may not have a cultural context to notice bias, i.e., to be aware of what these groups regard as bias or microaggressions. In our preliminary interviews, we became aware that many international faculty lacked knowledge of the history of U.S. marginalized groups and lacked firsthand experience interacting with women or URMs as professional peers. As a result, they may be unaware of what constitutes an offensive interaction, such as gender-offensive sexual innuendos or physical touch (Gahungu, 2011) or racist “jokes”. Also, their knowledge of racial groups may have been drawn from stereotypes rooted in U. S. socialization sources such as television or movies. Little is known, however, about what stereotypes foreign-born faculty may hold about U.S. women and underrepresented minorities. Some research indicates that members of different racial and ethnic groups within the U.S. (i.e., African Americans, Latinos, Asian Americans, and Whites) share similar perceptions of where groups are positioned along the dimensions of perceived inferiority and foreignness (Zou & Cheryan, 2017) suggesting that negative stereotypes of URMs are quite robust and may be likely to affect international scholars as well.

Diversity programs in higher education institutions implicitly rely on the premise that faculty in the U.S. are likely to value and be influenced by arguments and evidence that encourage fairness and equity (Stewart & Valian, 2018). It is not known if this is the case with foreign-born faculty. The current study explored the impact of bystander intervention training on foreign-born as well as U.S. born faculty.

The concept of intersectionality was our vehicle for addressing the notice bias step within the training. Intersectionality refers to the interconnections of race, class, and gender as they apply to a given individual or group, and the overlapping and interdependent systems of discrimination or disadvantage they create (Cole, 2009; Shields, 2008). An understanding this concept might prove beneficial to increasing foreign-born faculty’s understanding of the gender and race issues that women, Black, Hispanic and other minority faculty face. Having the shared identity of being a professor might lead faculty to assume that their colleagues are more alike in their social identities, privileges, and responsibilities than they are different, e.g., perhaps assuming that other faculty also had backgrounds that economically benefited them as students. Perspective-taking exercises have been shown to increase participants’ willingness to engage with “unalike” others (Wang et al., 2014). Thus, part of our training was aimed at helping participants realize that the social identities of other faculty might vary significantly from their own social identities and become more aware of the possible consequences for faculty members with multiple intersectional ties or social identities, e.g., being Black and a woman (Womack, et al., 2020).

Although many foreign-born faculty indicate being lonely in departments dominated by Anglo faculty (Collins, 2008), as Gahungu (2011) stated:

“They [foreign-born faculty] may not have experienced the hardships that domestic minority groups [in the U.S.] have faced year after year. They were not viewed as racially different in their countries of origin and, from their formative years on, made to feel that they were second class citizens who did not belong in higher education or in leadership positions” (p. 7)

In the present study, we explored (Q2) if the Intervention would increase notice bias for foreign-born faculty. We anticipated that U.S. faculty would be more familiar with the concept of bias and perhaps more often notice bias in an interaction than would international faculty.

Tipping Point

Our third research question was about what proportion of faculty participation within a department represented the “tipping point” that would result in a perceived change within the department towards more inclusive standards or practices. Theories of the “critical mass” necessary for change to occur set participation ranging from 25 to 60% as necessary for organizational change. The theory of diffusion of innovation posited that an adoption/participation level by forty percent of a group is the critical mass that is necessary for social change to occur (Grey, 2006). Abrahamson and Rosenkopf (1993) argued that most group members are prompted to adopt the innovation when the participation level is between 40 to 60 percent. At this point, the social system is expected to change, impacting individual attitudes, social climate, and institutional policy (Centola, 2013).

In sum, Q3 asked what level of proportion of departmental faculty participation in the Bystander training was necessary for participants to report an increase in in their department's adoption of more inclusive standards or practices.

Method

Design Overview

Participants were tenure-line faculty in twelve STEM and six SBS departments. The Intervention group was surveyed within one week before the intervention and at three months post-intervention. The Comparison group completed two surveys three months apart.

Participants

All tenure-line faculty in STEM and SBS departments (N = 406) that were not part of the facilitation team were invited to participate in the study. The STEM faculty population (N = 274) came from five science departments, including biology, chemistry, earth & environment, mathematics & statistics, and physics, and seven engineering departments, including biomedical, civil & environmental, computer & information sciences, construction management, electrical & computer, engineering education, and mechanical & materials. The SBS faculty population (N = 132) represented six departments, including criminology & criminal justice, economics, global & sociocultural studies, history, politics & international relations, and psychology.

The final sample (N = 253) was recruited over a two-year period and included: (1) an Intervention group (N = 173) that completed the Bystander Leadership™ workshop and two surveys (pre- and three months post-workshop), and (2) a Comparison group (N = 80) that completed two surveys three months apart. A majority of the Comparison group (95%) participated in Fall 2018, simultaneously with the start of the Bystander workshop offerings.

Participants were recruited by email over a two-year period from Fall 2018 to Spring 2020. The flowchart in Fig. 1 shows how the Intervention, Comparison and Non-participant groups were formed. Faculty were given the choice of volunteering to participate in the study as follows: (1) attend a bystander intervention workshop and complete two surveys, a pre- and a posttest, one week before and three months after the workshop (Intervention group; N = 173), or (2) complete two surveys three months apart (Comparison group; N = 80). The surveys were completed online using Qualtrics survey software. The remainder of the faculty who did not enroll in the study were considered part of the Non-Participant group (N = 153).

Fig. 1.

Recruitment steps to form the intervention, comparison and non-participant groups

Recruitment was done in waves in fall and spring semesters to enable teaching schedule changes to free up new faculty to enroll for the all-day workshop. The first wave of participants (both Intervention and 95% of Comparison group participants) completed the pretest in September 2018 or October 2018, with posttests in November 2018 and early January 2019, respectively. Wave 2 was recruited and completed the pretest in January 2019 and the posttest in March 2019, and included the remaining 5% of Comparison group participants. We continued to recruit the Intervention group through Wave 3 and 4 (Fall 2019 and Spring 2020). The final workshops were completed before the March 2020 COVID shutdown. Posttests for the final Spring 2020 workshop were completed in early May 2020 using Qualtrics.

All survey data responses were anonymous and acrostic codes generated by the respondents were used to connect the surveys longitudinally. All surveys were able to be connected longitudinally. The university Institutional Review Board (IRB) reviewed and approved this study. Participants provided written informed consent for the surveys and workshop intervention participation. Non-Participant group demographic information was collected from institutional records.

Intervention group participants were compensated $250 for attending the workshop and provided with a $25 gift card for completing each of the two surveys. Comparison group participants were compensated with a $25 gift card for completing each of the two surveys.

Intervention

The intervention was a one-day seven-hour interactive workshop. Each workshop was held in-person on campus. Seventeen workshops were held over the two years, each with a maximum capacity of 24 participants. A team of 8–10 faculty facilitators led each workshop. The details of the development of the intervention program and training of the facilitators can be found in Rose, Wood, & Farhangi, (2022).

In depth local research was conducted to develop the case materials for the workshop, including individual interviews and focus groups with forty STEM and SBS tenure-line faculty from across 18 departments at the same university. Interviews included eighteen faculty (five women, thirteen men). Four focus groups with 22 faculty also were conducted, including two groups with a total of nine women and two with a total of thirteen men. The interviews and focus groups explored faculty members’ personal experiences in their departments. The recorded material was transcribed, and case studies were derived, modified, and anonymized based on actual events that participants had experienced at the university (Rose, et al., 2022). The case studies focused on the issues of faculty hiring, mentoring, tenure and promotion processes, and doctoral student selection and education. The resulting twenty case studies were used to pilot test components of the workshop or included in the final workshop materials.

The lead facilitator began the workshop by explaining the importance of an inclusive workplace to attracting and retaining underrepresented faculty, as well as to enhancing creativity, productivity, and satisfaction for all faculty. Next, the five bystander intervention steps (Ashburn-Nardo et al., 2008; Latané & Darley, 1969) were explained. The first two steps, notice and interpret, we labeled as the insight components. The last three steps, lead, decide and act, we categorized as action components. Each step was presented sequentially throughout the day. Concepts relevant to each step were explained and illustrated using experiential exercises, interactive theater performances by professional actors, case study discussions, or role-plays with participant-enacted behavioral interventions. The components were arranged to begin with observational learning (e.g., the professional theater component) and proceed through each step to more active involvement and embodied learning (for a detailed explanation, see Table 1 and Authors (2022).

Table 1.

Bystander leadership™ workshop agenda and components

Time	Agenda	Component	Intervention Step	Content	Format
8:30am-9:30am	1. Introductions & intersectionality exercise		1. Notice	Intersectionality exercise	Experiential
9:30-10am	Unconscious bias-IAT	Insight	1. Notice	Implicit bias	Didactic
10–11:15am	Interactive theater		1. Notice & 2. Interpret	Interactive theater with professional actors	Observational, experiential
11:15–12:45 pm	Interventions table SeeTable 2		1. Notice & 2. Interpret	Self- and peer group exploration	Experiential
11:45–12:30 pm	Third year review skit		1. Notice & 2. Interpret	Enacted by Facilitation Team	Observational, Experiential
12:20-1 pm		Lunch
1–1:45 pm	Doctoral student selection skit and readers'theater	Action	3. Lead (take responsibility)	Enacted by participants with interventions	Embodied behavior
2:45–3:45 pm	Two skits with role play		4. Decide on action	Participant-enacted readers’ theater with interventions	Embodied behavior
1:45–2:45 pm	Case studies with role play		4. Decide & 5. Act	Participant-enacted	Embodied behavior
2:45–3:45 pm	Two skits with role play		4. Decide & 5. Act	Participant-enacted	Embodied behavior
3:45–4:30 pm	Discussion and evaluation				Conversational

Modified from Authors (2022)

The insight component focused on exploring the naturally occurring individual differences within the workshop peer group to encourage participants to engage in Step 1, to notice the dynamics of the interaction being observed and how others might experience it, and Step 2, to interpret the interaction as signaling an occurrence of gender, race, or other bias, as appropriate. Three concepts were introduced to explain the insight component, including intersectionality, implicit bias, and intent versus impact. Participation in the group exercises for the insight component was aimed at changing social norms in two ways: 1) by illustrating that one’s social position shapes what one knows and notices and 2) by revealing that underlying differences among faculty could contribute to different interpretations of events regardless of shared peer status as a faculty member. Further description of the insight component is available in Authors (2022). Taking the third step, to lead, involves taking responsibility to intervene. The expectation was set in the workshop that all participants would intervene. All components provided practice with intervening. Participants were free to opt-out of intervening, but only a few chose to opt-out early in the day and by the end of the day, all participated in intervening. Fourth, deciding what to do engages a bystander in a benefit-risk analysis that often depends on the person’s intervention skill set. Having a range of practiced intervention responses in one’s behavioral repertoire helps to reduce the perceived risks associated with intervening. Therefore, a typology of bystander intervention behaviors proposed by Bowes-Sperry and O’Leary-Kelly (2005, p. 290) and adapted by Ashburn-Nardo et al. (2008) (See Table 2) was modified to apply to events pertaining to gender and race bias and used to illustrate a range of response options available to them. The typology categorizes intervention along two dimensions: (1) immediacy of the intervention (immediate or delayed) and (2) level of involvement (high or low). Steps 3, 4, and 5 were then discussed and practiced using the response options available to them.

Table 2.

Bystander intervention strategies

Immediate – Low Involvement	Immediate – High Involvement
1. Question/interrupt	1. Invoke the offender’s egalitarian self- concept
“Do you feel that way about that group in general or are you just speaking of one person?”	“I always thought of you as open-minded.”
“I am not sure that I heard you correctly. Could you explain that again?”	“You’ve always said that it is important to support all the students on our team so how does you plan work with that?”
2. Support the person being targeted during the situation	2. Focus on how the offender’s action makes observers feel
“I think XX has been making a real contribution to the team.”	“I’m uncomfortable when you say things like that.”
“YY has really done a terrific job on this project.” Elaborate	“I am disappointed by this conversation.”
	3. Avoid self-righteousness
3. Express disapproval nonverbally	“With all the stereotypes we encounter,I can see how you might say that, but we should all try harder to avoid thinking that way.”
Eye roll	4. Appeal to norms that contradict the offender
Groan	“That isn’t in line with how we do things on our team.”
Shrug	5. Express disapproval/disagree
4. Deflecting/changing the subject	“There isn’t any evidence that women can/t succeed in our field.”
“Let me introduce you to…”	6. Suggest new procedures or policies
“Hey, have you heard this funny story?”	“Let’s change our procedures so this doesn’t happen again.”
“What about that [any sport] game last night?”	Suggest to a colleague that you work together to change the procedures
Delayed – Low Involvement	Delayed – High Involvement
1. Privately support the person being targeted	1. Privately speak to the offender/perpetrator
“Would you like to take a walk to talk about what just happened? You have my support.”	“I was surprised by what you were saying about [women, racial minorities, LGBT people]. Could you tell that was making the new student in our lab very uncomfortable? How would you feel in her situation?”
2. Arrange an educational program	2. Report/appeal to a higher authority
“Let’s ask the ADEPT team to do a special presentation for our advisors.”	“As our faculty advisor, can you intervene in this situation?”
3. Foster networks or coalitions	3. Change practices or policies
“We are going to start a Women in STEM monthly lunch.”	“Let’s request that there be a student representative on the Faculty Assembly.”

Modified from the framework and examples developed by Bowes-Sperry & O’Leary-Kelly (2005) and Ashburn-Nardo and et al., (2008)

Measures

We used five measures from Carnes et al. (2015) to study the effects of the Bystander Leadership™ intervention workshop on notice bias; self-efficacy (confidence to act); perceived benefit or risk of taking action, and take action in response to racial or gender bias. Table 3 indicates how the outcome variables were created.

Table 3.

Category and description of outcome variables

Outcome Variable	Description1
Notice Bias	I notice when others exhibit bias towards indivituals from racial/ethnic minority groups/women (2 items; a = 0.923)
Self-efficacy	I am confident that I can enact equity promoting acts, (4 items; a = 0.85)
Risk1	It would be risky for me to promote equity in my department, (4 items; a = 0.90)
Benefit	I would benefit from promoting equity in my department, (4 items; a = 0.89)
Take action	I engage in this action on a regular basis (4 items; a = 0.85)

¹All items used a 7-point Likert scale

²Reverse scored

Notice bias was assessed using the mean ratings of two questions that were combined because they were highly correlated: “I notice when others exhibit bias towards individuals from racial/ethnic minority groups,” and “I notice when others exhibit bias towards women.”

Four scenarios were presented that reflected common bias situations drawn from our microclimate study (see Table 4). The situations were ones that were directly addressed as intervention opportunities in the workshop. Participants answered four questions about each situation: 1) their self-efficacy to take action, b (i.e., their confidence in their ability to act, 2) the level of benefit or risk associated with taking action; and 5) take action (i.e., if the participants regularly acted to intervene). Each measure was averaged across the four situations. The risk item responses were reverse scored for consistency of interpretation.

Table 4.

Examples of a biased situations and response measures

Situation	Measures in response to each situation*
1.Speak about equity and diversity in my workplace to my colleagues	1. I am confident I can do this
2.Challenge a hiring decision if I think it has been influenced by stereotypes	2. I would benefit from doing this
3.Challenge a tenure/promotion decision if I think it has been influenced by stereotypes	3. t would be risky for me to do this.**
4. Act to intervene	4. I engage this action on a regular basis

*7-point Likert scales, 7 = strongly agree

**reverse scored

Carnes, et al. (2015). The effect of an intervention to break the gender bias habit for faculty at one institution

Academic Medicine, 90(2), 221–230. https://doi.org/10.1097/acm.0000000000000552

The last question (Q3) was focused on what proportion of faculty participation within a department was sufficient to influence the use of more inclusive practices or procedures. Intervention departments were classified as having high participation (> 60% of faculty participating); medium participation ( 40-to 60%); or low participation (< 40%). The Intervention group was asked about their experiences of engaging in bystander interventions in the three months after the workshop and their opinion about the changes in their own or departmental practices, if any. The participants also were given the opportunity to provide written comments explaining their responses to these questions. Response options for the post-intervention survey questions were “Yes” (coded as 1) or “No” (coded as 0):

1. Have you talked about the Bystander Leadership™ Workshop with your colleagues since the workshop?

2. In your opinion, has the workshop had any effect on your own or your department's practices (like the department's diversity plan)?

The participants were also given the opportunity to provide free-response comments in addition to their Yes/No answers; their responses were coded and analyzed as described below.

Results

Since recruitment was voluntary and not randomized, Chi-square tests of association were conducted to assess whether the Intervention, Comparison, and non-participant groups differed in terms of the proportional representation among the three groups in terms of gender, race, rank or birthplace (e.g.,. foreign-born vs U.S.-born). No significant differences in representation were found for any of the variables. Table 5 compares the demographics of the Intervention, Comparison, and Non-participant faculty groups (Intervention N = 173, Comparison N = 80, Non-Participant N = 153) and discipline subgroups (Science, Engineering and SBS). Four Comparison group participants later attended the workshop and were reclassified as being in the Intervention group. Proportions for Science and Engineering are shown separately because of the much higher percentage of foreign-born faculty in Engineering. Women were 30% of the Intervention group, 27% of the Comparison Group and 21% of Non-Participants. A majority of faculty identified as White across all three groups (58% in intervention, 56% in comparison, and 59% on-Participant), followed by Asian (29% Intervention; 30% Comparison, and 23% Non-Participant). See Table 5.

Table 5.

Demographics of faculty by group

	Intervention (N = 173)				Comparison (N = 80)				Non-Participant (N = 153)
	Sciencea N (%)	Engb N (%)	SBSc N (%)	Total N (%)	Science N (%)	ENG N (%)	SBS N (%)	Total N (%)	Science N (%)	ENG N (%)	SBS N (%)	Total N (%)
Gender
Male	44 (25)	50(29)	27(16)	121(70)	28 (35)	14 (18)	16 (20)	58(72)	47 (31)	34 (22)	39 (25)	120(78)
Female	21 (12)	15 (9)	16 (9)	52(30)	6(8)	4(5)	12(15)	22(27	8(5)	3(2)	22(14)	33(21)
Race/Ethnicity
ASIAN	15 (9)	31(18)	4(2)	50(29)	12(15)	11(14)	1(1)	24(30)	13(8)	14(0)	7(5)	34(22)
BLACK	3(2)	4(2)	1(1)	8(15)	1(1)	0	3(4)	4(5)	3(2)	2(1)	4(3)	9(6)
Multi-racial		1(1)		1(< 1)
White (non- Hispanic)	41(24)	24(14)	35(20)	100(58)	19(24)	5(6)	21(26)	44(55)	33(22)	20(13)	37(24)	90(59)
Hispanic	6(3)	5(3)	3(2)	14(8)	2(3)	2(3)	4(5)	8(11)	7(5)	0	13(8)	20(13)
Birthplace
US	32(18)	12(7)	30(17)	74(43)	11(14)	2(3)	25(31)	38(47)	30(20)	11(7)	37(24)	78(51)
Foreign- born	33(19)	53(31)	13(8)	99(58)	23(29)	16(20)	3(4)	42(53)	25(16)	26(17)	24(16)	75(49)
Rank
Assistant	6(3)	22(13)	7(4)	35(20)	5(6)	2(3)	5(6)	12(15)	9(6)	11(7)	17(11)	37(24)
Associate	21(12)	22(13)	22(13)	65(38)	13(16)	7(9)	11(14)	31(39)	18(12)	12(8)	28(18)	58(38)
Professor	38(22)	21(12)	14(8)	73(42)	16(20)	9(11)	12(15)	37(46)	28(18)	14(9)	16(10)	58(38)

^aScience departments; ^bEngineering departments; ^cSocial and Behavioral Science departments ^dUnderrepresented Minorities: NSF definition includes Blacks or African Americans, Hispanics or Latinos, & American Indians or Alaska Natives

We used faculty’s birthplace to classify faculty as foreign-born or U.S. born. Foreign-born status did not appear to affect the participation rate of faculty in either the Comparison, Intervention or non-participant group. As shown in Table 5, a majority (58%) of the Intervention group faculty was foreign-born, as was 53% and 49% for the Comparison and Non-Participant groups respectively. In terms of rank, Assistant professors were 20% of the Intervention group, 15% of the Comparison and 24% of the Non-Participant group. The percentage of the Associate professors was 38% in the Intervention group, while it was 39% and 38% for the Comparison and Non-Participant groups, respectively. Last, Full professors comprised 42% of the Intervention, 46% of the Comparison and 38% of the Non-Participant group.

Intervention versus Comparison Group

Descriptive statistics for all pre- and posttest measures are presented for Intervention and Comparison groups in Table 6. We used listwise deletion to deal with missing data because less than 10 percent of responses were missing (i.e., skipped question) for any one item. A MANOVA indicated that at baseline/pretest, the Intervention and Comparison group means did not differ significantly for any variable. See means and SDs in Table 6.

Table 6.

Anova results comparing baseline/pretest responses for the two groups

	Intervention (N = 173)	Comparison (N = 80)	F(df)	p value
	M(SD)	M(SD)
Notice bias	5.39(1.48)	5.57(1.53)	0.02(1,245)	0.91
Self-efficacy	3.38(1.17)	5.48(1.00)	0.32(1,233)	0.58
Benefit	4.86(1.40)	4.54(1.59)	0.24(1,233)	0.12
Risk1	3.49(1.59)	3.66(1.74)	0.56(1,233)	0.46
Take action	3.91(1.37)	3.91(1.24)	0.00(1,234)	0.99

All items, 7-point Likert scale, 7 = very much

¹Reverse scored

To test for differences between the Intervention and Comparison groups, we compared the pre- to posttest means for the two groups using a 2 × 2 (group x time) MANOVA for the five dependent variables in SAS and JAMOVI. See Table 7. A linear regression model analysis with time as a within-subjects factor, and group as a two-level between-subjects factor (Comparison and Intervention) was then conducted. We used Bonferroni-adjusted differences of least square means to determine any significant interactions. A significant interaction between time and group was observed for two measures: self-efficacy and take action increased significantly among the Intervention group after the workshop (See Table 7). No significant effects were found for the benefit or risk measures. No significant change occurred over time for the Comparison group on any of the measures. Means, SDs, and post hoc analyses are presented in Table 8.

Table 7.

The model fit measures and the significance test results for group x time interaction

Overall Model Test
Measure	R	R2	F	df1	df2	p (overall model)	p (group x time)
Notice bias	0.39	0.15	9.31	9	468	0.001	0.36
Self-effficacy	0.31	0.09	5.85	8	453	0.001	0.01*
Benefit	0.29	0.09	5.29	8	448	0.001	0.37
Risk	0.48	0.23	1.60	8	451	0.001	0.96
Take action	0.32	0.10	6.21	8	452	0.001	0.01*

*Shows significant p for the interaction

gender, race, ethnicity, birthplace and rank were included as covariates in the analysis

Table 8.

Means and SDs for the five response measures by group and time

	Group
Measure	Intervention		Comparison
	Pre	Post	Pre	Post
	M(SD)	M(SD)	M(SD)	M(SD)
Notice bias	5.39 (1.48)	5.93 (1.04)	5.57 (1.53)	5.36 (1.53)
Self-Efficacy	5.39(1.17)a	5.96(0.92)b	5.48(1.00)a	5.50(1.24)a
Benefit	4.86(1.40)	5.15(1.48)	4.45 (1.60)	4.56(1.52)
Riskc	3.49(1.59)	3.47(1.64)	3.66(1.74)	3.72(1.63)
Action taken	3.91(1.74)a	4.48(1.35)b	3.91(1.24)a	3.80(1.35)a

gender, race, ethnicity, birthplace, and rank were included as covariates in these analyses

All items, 7-point Likert scale, 7- very much

^a,bMeans with different superscripts differ significantly, main effect of group, p_Tukey's < 0.001

^cReverse-scored

Figure 2 visually illustrates the marginal means of self-efficacy and their confidence level based on the model we used. It shows that self-efficacy significantly increased for the Intervention group from pretest (M = 5.39 (SD- = 1.17)) to posttest (M = 5.96, SD = 0.92). p < 0.001). There was no significant change over time in self-efficacy scores for the Comparison group (M = 5.48, SD = 1.0 and M = 5.50, SD = 1.24, respectively).

Fig. 2.

Self-efficacy means and confidence intervals for intervention and comparison groups at pretest and post test. Note: Posttest means for the two groups differed significantly, p < 0.0001

Likewise, Fig. 3 visually illustrates that take action significantly increased for the Intervention group from pretest (M = 3.91, SD = 1.74) to posttest (M = 4.48, SD = 1.35, p < 0.0009), but that no change occurred for the Comparison group (M = 3.91, SD = 1.24; M = 3.80, SD = 1.350).

Fig. 3.

Take action means and confidence intervals for intervention and comparison groups at pretest and post test. Note: Posttest means for the two groups differed significantly, p < 0.0001

Foreign-born Faculty

Foreign-born faculty, like U.S. born faculty that attended the workshop, had significant pre- to posttest gains in self-efficacy and take action. See means and descriptive plots in the Appendix.

To address the question (Q2) asking if foreign-born faculty differed from U.S. born faculty or if the Intervention would significantly increase notice bias, we used a three-way mixed MANOVA with five dependent variables. A significant effect of birthplace was found for one variable, notice bias, JAMOVI ($F_{\text{15,440}}$= 10.94, p = 0.001). Univariate analyses revealed that at the pretest, notice bias scores for Foreign-born faculty in both Intervention and Comparison groups were significantly lower (M = 5.07, SD = 1.6 and M = 4.83, SD = 1.64, respectively) than those of U.S. born faculty in the intervention or comparison group $(M=$ 5.79, SD = 1.12, M = 5.95, SD = 1.26, respectively, $p_{\mathit{tukey}}<.001)$. See means and SDs in Table 9. No effect of birthplace was found for the other variables.

Table 9.

Notice bias means and SDs by group, birthplace and time

	Intervention Group				Comparison Group
	U.S-born		Foreign -born		U.S.-born		Foreign-born
Time	M	(SD)	M	(SD)	M	(SD)	M	(SD)
Pretest	5.79a	1.12	5.07b	1.65	5.95a	1.26	4.83b	1.64
post-test	6.04a	1.03	5.84a	1.05	6.17a	0.99	4.97b	1.76

^abMeans with different superscripts differ significantly, Tukey's test, p < 0.005

Figure 4 further visually illustrates that, at posttest, notice bias was significantly higher for foreign- born faculty in the Intervention group (M = 5.84, SD = 1.05) than for foreign-born faculty in the comparison group (M = 4.97, SD = 1.76) (p < 0.001).

Fig. 4.

Notice Bias Means and Confidence Intervals for U.S. Born and Foreign-born Faculty by Group and Time. Posttest means for foreign-born faculty in the Intervention group differed significantly from those for the Comparison group, p < 0.001

Tipping Point

Next, we looked at what proportion of participation by departmental faculty in the Intervention workshop was the “tipping point” or critical mass necessary to lead to increased departmental engagement in equity and inclusion issues. To evaluate if we could detect any collective change in department’s faculty’s behavior, we analyzed responses to the question, “In your opinion, has the workshop had any effect on your own or your department's practices (like the department's Diversity plan)yes/no)” which was included only in the Intervention group’s posttest. The Comparison group was not asked this question because they had not attended the workshop. We defined a low engagement department as one in which less than 40 percent of invited faculty participated in the intervention group. High engagement included departments with more than 60 percent participation in the workshop. Medium engagement was defined as being between 40 to 60 percent participation.

A one-way Welch’s ANOVA test showed a significant difference in self-reports of climate change for different levels of departmental participation ($F_{\text{2,62}}$= 3.28, p = 0.05). See Table 10.

Table 10.

Means and SDs for the measure “the effect on self or department practices” by proportion of departmental faculty participating

Engagement Level	N of departments	N of responses	Mean	SD
High (> 60% faculty participation)	3	25	0.84a	0.4
Medium (40–59% participation)	6	95	0.72ab	0.5
Low < 40% participation	9	41	0.56b	0.5

The item was coded 1 for YES and 0 for NO response

^abMeans with different superscripts differ significantly, p < 0.05

Post hoc Tukey’s HSD test for multiple comparisons found that High Engagement departments (i.e., 60 percent or more faculty participation) significantly more often reported that the Bystander Leadership™ workshop changed their own or their department’s practices than did the Low Engagement departments (Mean/High = 0.84, Mena/Low = 0.56, respectively; Tukey’s_p < 0.05); the medium engagement departments (40 to 60 percent participation) did not differ significantly from either the high or low engagement departments.

To explore what specific changes had resulted from 60 percent of more participation, a qualitative analysis of open-ended responses about the effect of the workshop on departmental practices was conducted. The responses were coded by two raters into eight categories (96% interrater reliability); examples are given in Table 11. Most Intervention participants (N = 126; 73%) provided written text for analysis and indicated the workshop had encouraged the use of inclusive departmental practices.

Table 11.

Participants’ perception of areas of personal or departmental change (N = 126)

Coding Category	Example quote	N (%)
Mentioned the importance of a specific part of the workshop, e.g., implicit bias, role play; practice; said it prepared me to take action	" It made me aware of how to … solve [the problems]."	71 (57)
	"Discussion of skits and how we all approached situations, and our viewpoints."
	"The interactive live enacting of us with stage actors to advise on communication issues was a great learning experience to see how each faculty thinks so differently."
Applied it to the departmental diversity plan, or to graduate student admissions, or to teaching	"[Discussed it] when discussing the [departmental] diversity plan."	53 (46)
	"Shared some of the methods for addressing issues of diversity in my department."
Had an impact on me personally, made me aware of my own biases or my own situation	"With respect to my own practices, I have become more aware of things that I say/think about saying, that may not be taken the way that I am thinking."	51 (44)
	"I am able to better distinguish subtle discriminative comments and speak up during various occasions, and the credit goes to Bystander Workshop."
Became more aware of bias of department or colleagues	"We actively look for language that supports stereotypes or that might send the wrong message."	27 (21)
	"it has made faculty more aware."
Provided a common framework on the value of diversity and fairness; set new standards	"It reinforced my own values and showed that first of all those are shared by others and emphasized by the university, and secondly, that exercising recognition and intervention is acceptable and encouraged."	25 (20)
	"Colleagues who attended the workshop are more active in departmental meetings than before."
Indicated that it influenced hiring or search processes	" I think a decent number of us have taken it, and I feel like everyone was very conscious [of it] during recent hiring and admissions decisions."	23 (18)
	"I am definitely more conscious about concepts of bias when discussing for example, the departmental hiring plan."

The table shows examples of the codes and the comments that faculty gave about their perception of changes within themselves or their department. Most participants highly recommended the workshop to colleagues and indicated that it prepared them to act. For example, one woman SBS faculty member commented:

“The workshop definitely changed my behavior. I feel more comfortable speaking up now if I see something may be wrong (although I also had a positive tenure vote which makes me more comfortable as well). I served on a search committee and there was a female applicant that was really competitive, and she had nearly all first author publications. One male committee member suggested that maybe she was not a good collaborator since her pubs were mostly first author. I made the point that if this had been a male candidate, this would not have been said.”

The most often action taken was to apply what was learned in the workshop to the development of a diversity and inclusion plan for their department, and usually specifically to ensure that objective evaluation practices were used for hiring and tenure decisions or graduate admissions.

At the end of the workshop, participants were asked one thing that was learned during the day was most important to them. The overwhelming response was “the actual practice of intervening.” Many said that the intervention they performed was not what they imagined they would do, which speaks to the importance of enacting the intervention verbally and nonverbally. Another highlight frequently mentioned was how supported individuals felt by allies (e.g., other participants) that would back their intervention efforts. Allies are members of a group that take a proactive stance against the discriminatory actions of perpetrators (Sue et al., 2019). Allies can play a critical role in creating a bias-free and inclusive workplace (Ashburn-Nardo et al., 2019; Erskine & Bilimoria, 2019). In fact, some participants spontaneously induced other participants to be an ally by asking them to accompanying them when responding to a (confederate) “perpetrator” in a role-play interaction.

Discussion

The findings of the present study indicate that the Bystander Leadership™ intervention workshop significantly enhanced faculty member’s capacity to intervene in situations of gender or racial bias. These results provide evidence that the intervention was effective at the three-month post-test as reflected by the sustained increase in self-efficacy beliefs and actual interventions. Behavioral change theories emphasize self-efficacy as the foundation of change, and the numerous practice opportunities in our workshop appear to have bolstered self-efficacy. Additionally, our findings showed that a minimum of 60 percent participation in the workshop was the critical mass necessary to have a positive effect on inclusive departmental practices.

The findings suggest that interactive engagement in bystander intervention training is an effective way to influence faculty awareness, confidence, and behaviors to intervene to reduce gender and race bias. This finding contributes to the recent surge of interest in using bystander intervention techniques to reduce microaggressions and to increase faculty members’ effectiveness to be allies to women and historically underrepresented faculty in academe (e.g., Haynes-Baratz et al., 2021).

The intervention did not appear to influence the benefit or risk measures. Latané and Darley (1969) noted that it is unlikely for interventions to yield a benefit. No significant change was found for these measures. One possible reason for this is that 90 percent of the participants were tenured faculty who perhaps perceived that there was little risk for intervening, given the job security afforded by tenure. Another possible reason is that the response options allowed for a delayed response to observed incidents, such as reporting the event to a higher authority after the event or proposing a change in department policies or practices. These delayed options were possibly seen as low risk strategies that allowed even timid faculty to find a way to intervene.

These results also speak to the level of intervention necessary to influence behavioral change. Greater than 60 percent participation was necessary to catalyze departmental change using this training. This suggests that change efforts must be deep and broad to influence change.

Institutional transformation requires multiple levels of focus, but change is driven by individuals (Nonaka, 1994; Rogers, 1995). Consistent with the tenants of institutional change, we found that an intervention that taught individual faculty members to notice bias and practice new skills served to increase their self-efficacy and intervention actions.

The consensus among our team was that the most effective part of the training was the numerous opportunities to practice intervening. The frequent practice sessions gave participants a chance to observe many interventions being done by their colleagues. Even though the situations that were used were “staged” with the facilitators acting as targets or perpetrators of bias, many Ps reported being nervous to interact with a perpetrator/colleague and address their behavior directly or indirectly, even in a role play situation. Participants reported that the practice was very empowering to them.

There seems to be a strong ethos in academe not to confront or criticize one’s peers or ‘blow the whistle’ on a colleague’s behavior. The practice sessions seemed to break through the “mind your own business” mindset by giving participants permission to act. We think this mindset is related to the issue of why Notice Bias did not increase among the U.S. faculty. U.S. faculty seemed able to quickly identify what was ‘wrong’ with the scenarios we used. But they did not know what to do or perhaps did not see it as their responsibility to do anything. The workshop gave them the tools to intervene and stressed the need for action.

Limitations

This study has several limitations. It is limited to a single institution; however, it included faculty from a wide variety of science, engineering, and social science disciplines. The participation level was high, with 43% of those contacted attending the full day workshop, which is a high level of commitment from busy faculty. The use of a quasi-experimental rather than randomized experimental design might pose threats to the validity of the study. The Comparison and Intervention group were subject to selection bias. Given the seven-hour-plus commitment required to attend the workshop and complete assessments, random assignment to groups was not possible. However, there were no differences between the intervention and control groups in terms of demographics or baseline measures. Additionally, we did not control for possible confounding variables, such as the bystander participant’s previous history with being a target of bias incidents or the length of time that foreign-born faculty had been in the U.S. Another limitation was that the posttest for the Comparison group did not include the question about how the workshop affected departmental changes, since they had not attended the workshop.

Directions for Future Research

Future research on the long term effects of bystander training are warranted. We had a hint that faculty continued to be interested in the BLP more than a year after attending. We offered two Bystander Booster sessions by zoom in April, 2021 (during COVID) that was attended by seventy faculty who had previously attended the workshop. Furthermore, after the study reported here was completed, we continued to offer the workshops from Fall 2021 to Fall 2023. An additional 216 faculty participated during that time (post-COVID), without monetary compensation for participants, suggesting that the program was popular due to word-of-mouth. Further study post-COVID may be warranted to understand if findings may differ in the current context. We were unable to do further follow-up because state laws that were passed in December 2023 did not allow state universities to continue to do diversity training. The unit that housed the Bystander Leadership™ Program was permanently closed in January 2024.

The results suggests that further research on foreign-born faculty regarding their commitment to and understanding of diversity, equity and inclusion as practiced in U.S. higher education is a potentially important area of investigation. Our evidence showed that foreign-born faculty overall were less likely to notice bias than were U.S. born faculty, both before and after the intervention, but that the intervention did improve their ability to notice bias. There might be additional differences between foreign-born and US-born faculty that have not yet been explored. This is an important point because the number of international faculty in STEM is quite high. By 2013, the professoriate was less dominated by White U.S.-born males, with almost a quarter of faculty in colleges and universities identified as foreign-born (Theobald, 2013). In 2020, international students earned 57% of all U.S. STEM doctorate degrees (National Center for Education Statistics, 2023). Therefore, the foreign-born faculty member’s commitment to inclusive excellence is an important component of organizational change.

When studying foreign-born faculty, it is also important to consider nationality and experience differences among them. The term Asian encompasses many nationalities, including Chinese, Indian, Pakastani, Korean, Japanese, and Malaysian that do not have a singular cultural identity or hold common beliefs about race and gender. Amount of exposure to U.S. culture or diversity training experience are other variables of possible interest. Many foreign-born faculty become U.S. citizens and may become versatile in understanding these aspects of U.S. culture.

This work lends itself to the development of practitioner-oriented materials. Team members have utilized the bystander method to improve undergraduate advising (Secules et.al., 2024), as means for thinking about equity in higher education classrooms (Secules, 2024), as a way to reduce the marginalization of minority students in engineering teams (Dickerson et al., 2024), and to educate same-nationality graduate student research teams (e.g., Asian, Iranian) about inclusive excellence practices (Rose et al., 2021).

Conclusion

The present study provides empirical support for the Bystander Leadership™ intervention model to interrupt bias, which emphasizes intervening either in the immediate situation or at a later time. The intervention was found to change faculty behavior and improve department climate in ways that would promote gender and race equity. Participants appreciated the 2 × 2 response matrix (Bowes-Sperry & O’Leary-Kelly (2005), which identified immediate or delayed high involvement and low involvement interventions. The range of options it provided enabled every participant to find an intervention that they perceived would be effective as well as low risk, thereby significantly increasing their confidence to act. Repeated exposure to how others responded to a specific situation also increased participant’s confidence in their ability to intervene, as well as illustrated the effectiveness of having more than one person intervene in the same situation.

Within organizations, bystander intervention can diminish workplace bias and increase inclusion (Jennings et al. (2024). This study adds support to Griffith. Malone, & Shea’s (2021) conclusion that bystander intervention is “perhaps the most empirically supported method of addressing undesirable interpersonal behavior” (p. 273). The findings reported here provide evidence that the Bystander Leadership™ Program was a useful strategy for developing a sense of collective responsibility for promoting inclusive excellence among faculty. Foreign-born faculty may themselves benefit from bystander intervention training that addresses exclusion or bias towards international faculty. Mamiseishvili and Lee (2018) indicated that international faculty’s job satisfaction and belongingness might be enhanced if their interest in pursuing research agendas outside of the dominant disciplinary methods or content areas was supported.

One of the principle contributions of this research is to offer an intervention that fosters a sense of collective responsibility for promoting inclusive practices among faculty. By having multiple members of departments attend the workshop, a cohort of allies was developed that could support interventions made by individuals. As Collins et al. (2021) contended, “bystander intervention and ally development are necessary tools to implement social justice initiatives that can address issues at both the interpersonal and systemic levels.” (p. 486). Bystander training shows everyone how they can participate in this work.

Supplementary Information

Below is the link to the electronic supplementary material.

Biographies

Suzanna M. Rose

, PhD, is emerita professor of psychology and was the founding Associate Provost of the Office to Advance Women, Equity & Diversity at Florida International University (FIU) in Miami. Rose was the lead investigator for FIU’s NSF ADVANCE Institutional Transformation grant, 2016-2023. A specific focus of the grant concerned how to engage international faculty in ADVANCE work. Rose has published extensively on issues related to gender, race, and sexual orientation, including professional networks, career development, friendship, and personal relationships. She has consulted with universities both nationally and internationally concerning strategies for advancing women faculty in science and engineering.

Sanaz Farhangi

, PhD, is a codesign researcher in Leanlab Education, a nonprofit organization that elevates the voices of teachers and students in designing educational technology. She is trained as an engineer and has worked as a science teacher, and these experiences have led her to wonder about how people engage with science and engineering and why the opportunity for this engagement is not as accessible for all learners. Her research interests are in theoretical foundations of learning sciences, developing teaching methods for promoting theoretical-systematic thinking in science learners, and exploring the underrepresentation of women and racial minorities in STEM.

Dr. Michelle Hospital

, PhD, is a Research Associate Professor in the Department of Biostatistics at Florida International University and Associate Director of FIU’s Community-Based Research Institute (CBRI). A Developmental Psychologist and Licensed Mental Health Counselor, she has over two decades of experience designing, implementing, and evaluating community-engaged prevention and intervention research primarily focused on reducing risk behaviors and promoting well-being among diverse underserved populations. Dr. Hospital has served as an investigator on numerous extramural grants from organizations such as the NIH, NSF, SAMHSA, and private foundations.

Dr. Kirsten E. Wood

is an Associate Professor of History at Florida International University, specializing in the social, economic, and business history of the early American republic. She earned her Ph.D. from the University of Pennsylvania in 1998. Her works include Accommodating the Republic: Taverns in the Early United States (UNC Press, 2023), Masterful Women: Slaveholding Widows from the American Revolution through the Civil War (UNC Press, 2004), and numerous articles and essays. A recipient of the Southern Historical Association’s Francis B. Simkins Prize, Dr. Wood also engages in institutional change-making, focusing on diversity and equity within universities and scholarly organizations.

Author Contributions

All authors contributed to the study conception and design. Material preparation was performed by Suzanna M. Rose, Sanaz Farhangi, and Kirsten Wood. Data collection and analysis were performed by Sanaz Farhangi and Michelle Hospital. The first draft of the manuscript was written by Suzanna M., Rose and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability

The participants of this study did not give written consent for their data to be shared publicly, so due to the sensitive nature of the research, supporting data is not available.

Declarations

Ethics Approval

The questionnaire and methodology for this study was approved by the Human Research Ethics committee of the Florida International University (IRB-18–0282).

Informed Consent

Informed consent was obtained from all participants.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose. The authors have no competing interests to declare related to the content of this article.