Mentors’ Motivation to Address Race/Ethnicity in Research Mentoring Relationships

Abstract

The purpose of this study was to investigate the motivation of research mentors to address race/ethnicity in their research mentoring relationships, using self-determination theory as a conceptual framework. Mentors from STEM fields primarily in the biological sciences (N = 115) were asked to report their level of motivation and the reasons behind their motivation to address the role of race and ethnicity in their mentoring relationships. Mentors’ responses were coded using a qualitative approach, and results were examined by mentors’ degree of motivation, previous experience with mentoring trainees from different racial/ethnic groups, and mentor race/ethnicity. Extrinsic motivation and amotivation were the most frequently assigned codes to mentors’ responses. Implications of these findings for mentor practices, higher education initiatives, and for diversifying the STEM workforce are discussed.

Mentors' Motivation to Address Race/Ethnicity in Research Mentoring Relationships

The future of science is dependent upon a strong and demographically diverse workforce. Despite efforts to recruit and retain scientists from historically underrepresented (HU) groups, the number of women and individuals from racial/ethnic groups pursuing careers in science, technology, engineering, or mathematics (STEM) fields has not kept pace with the changing demographics of the United States population (Valantine & Collins, 2015). Unwelcoming institutional climate (Johnson, 2007; President’s Council of Advisors on Science and Technology [PCAST], 2012), fear of confirming stereotypes (i.e., stereotype threat; Steele 1997), and bias (Valantine & Collins, 2015) can each contribute to the unequal participation of individuals from HU groups in STEM. Experiences of racial stigma in the research environment (e.g., labs and research groups), along with other psychological and contextual factors, may influence success and persistence in the sciences (e.g., applying for a PhD degree; pursuing a career in science) for trainees from HU groups (Carlone & Johnson, 2007; Chemers, Zurbriggen, Syed, Goza, & Bearman, 2011; Estrada-Hollenbeck et al., 2011; Hurtado, Cabrera, Lin, Arellano, & Espinosa, 2009; Johnson, Brown, Carlone, & Cuevas, 2011). Yet research mentors of HU trainees may be unaware of these factors and their potential impact on trainees’ research experiences.

Research mentors can both positively and negatively influence the ability of trainees to successfully cope with the psychological and environmental challenges that they may encounter in their research training (Byars-Winston, Branchaw, Pfund, Leverett, & Newton, 2015; Thiry & Laursen, 2011). Empirical evidence has documented a positive correlation between effective mentoring relationships and HU trainees’ confidence in their ability to successfully conduct research, a key predictor of academic persistence in STEM (Byars-Winston et al. 2015). Among the factors that are posited to play an important role in broadening participation in STEM are several that relate to culturally responsive mentoring, that is, mentors’ acknowledgement that the research mentoring relationship occurs within the cultural contexts of mentors and their trainees (Byars-Winston, 2017; Pfund, Byars-Winston, Branchaw, Hurtado, & Eagan, 2016).

Research mentors in training programs often do not acknowledge that racial or ethnic-related identity factors can have an impact on a trainee’s research experience (Davidson & Foster-Johnson, 2001; Prunuske, Wilson, Walls, & Clarke, 2013). This may be due to views that STEM fields are “objective” and thus cultural diversity factors like race and ethnicity are considered to be irrelevant to the technical aspects of research that are often conducted in labs or field studies, and in research groups (Carlone & Johnson, 2007; Calabrese Barton, 1998; Conefrey, 2001; Foor, Walden, & Trytten, 2007; Ong, 2001). Despite the persistent beliefs of some researchers that STEM fields are “colorblind” (McCoy, Winkle-Wagner & Luedke, 2015; Prunuske et al., 2013), trainees from HU groups may want to address matters of racial and ethnic diversity in their research experiences and mentoring relationships, and some studies indicate that doing so can directly benefit their academic/career outcomes (Chan, 2008; Haeger & Fresquez, 2016; Johnson, 2007). The tension between the desire of trainees to discuss their lived experiences and the reluctance of some research mentors to acknowledge the relevance of such experiences in a research training context can result in a dilemma: do mentors address racial/ethnic factors in research experience and mentoring relationship, or do they ignore these factors and focus on the science? This dilemma may be especially salient with respect to race and ethnicity, a topic with which many individuals in the sciences have difficulty engaging (Colón Ramos & Quiñones-Hinojosa, 2016).

The purpose of this study was to investigate the motivation of research mentors to address race and ethnicity in their research mentoring relationships. We also investigated whether mentors’ motivation differed with respect to cross-cultural mentoring relationships or by the racial or ethnic identity of mentors. Using a mixed-methods approach, we used self-determination theory (Deci & Ryan, 1985; Ryan & Deci, 2000) as a conceptual model of motivation for framing our research questions and analyzing our data.

Theoretical Framework

What motivates human behavior is complex. Numerous theories from the field of psychology posit that motivational drive results from some combination of individuals’ desire to avoid negative or unpleasant experiences and fulfill intrinsic needs for personal growth and achievement. Self-determination theory (Deci & Ryan, 1985; Ryan & Deci, 2000) provides one such framework through which researchers can explore different types of motivation and their effects on interpersonal interactions.

According to self-determination theory, humans have a fundamental need for competence, autonomy, and relatedness, which work together to influence individual motivation (Deci & Ryan, 2000). Competence refers to individuals’ feelings of efficacy with respect to a given task; autonomy is the extent to which individuals feel a sense of control and agency over their environment and their actions; and relatedness is the sense of belonging or community one feels. Within self-determination theory, organismic integration theory offers a spectrum of motivation ranging from amotivation to full integration of socially ascribed values or goals (Deci & Ryan, 2000; Ryan & Deci, 2000). Varying levels of motivation within this spectrum are characterized by increasing levels of competence, autonomy, relatedness, and self-determination. Amotivation represents the absence of motivation to act. This could be due to a failure to see the relevance or value in an activity, lack of competence, or feelings of futility (Ryan & Deci, 2000). Extrinsic motivation refers to external forces that drive individuals to act. Deci and Ryan (1985) posited that four different types of extrinsic motivation exist. Externally regulated actions are completed by individuals due to an external demand, reward or punishment. This is the most controlled type of extrinsic motivation because an individual’s behavior is dictated entirely by external forces. Introjected regulation is a slightly less controlled and slightly more internalized form of extrinsic motivation that drives behavior based on self-esteem and individuals’ concerns about how they might be perceived. Identification and Integration each represent more autonomous types of extrinsic motivation. Identification may be characterized by a person who is motivated by the value or importance of doing something. Integration is achieved when an individual fully integrates an external regulation into his/her belief system or worldview. Intrinsic motivation is the most internally driven form of motivation and represents the inherent interest and curiosity that drives individuals to explore and learn. This is also theorized to be the most adaptive and sustainable form of motivation (Deci & Ryan, 1985; Ryan & Deci, 2000). Intrinsic motivation is fundamentally different from integrated motivation in that the latter still refers to an external force that is driving the behavior. In other words, individuals recognize the value and importance of the behavior and found it to be acceptable and desirable. Individuals who are intrinsically motivated, on the other hand, do not have to be convinced that the behavior is desirable, nor do they act due to an external reward or punishment (Ryan & Deci, 2000). Although identified, integrated, and intrinsic forms of motivation are distinct from one another, researchers have studied them together as a composite of autonomous motivation (Deci & Ryan, 2000).

Other psychologists have used self-determination theory to explore the motivations of individuals to behave in a non-biased manner. In the field of social psychology, self-determination theory has been used as a conceptual framework to explore the internal (i.e., identified or internalized) and external (i.e., external or introjected regulation) motivations to respond without prejudice. In their review, Butz and Plant (2009) used self-determination theory to present four different profiles representing varying levels of internal and external motivation to respond without prejudice: Effective (Internalized), Dedicated (Identified), Compliant (External/Introjected Regulation); and Unmotivated (Amotivation). LeGault, Green-Demers, Grant, and Chung (2007) found that individuals who reported more self-determined motivation to reduce prejudice (i.e., intrinsic, integrated, or identified motivation) are associated with significantly lower levels of self-reported racial prejudice and implicit and explicit racial bias compared to individuals who reported less self-determined motivation (i.e., external or introjected regulation; amotivation). Moreover, endorsing externally regulated reasons for reducing prejudice can backfire, leading individuals to exhibit more prejudice than if no intervention had been conducted at all. In other words, a more externally controlled stance towards reducing prejudice can lead individuals to actively defy regulations against prejudicial behaviors (LeGault, Gutsell, & Inzlicht, 2011; see also Plant, Devine & Peruche, 2010).

Although self-determination theory has been used to study individuals’ motivation to reduce prejudice, few researchers have applied it within the context of mentoring relationships. One exception was the qualitative study conducted by Janssen, van Vurren and deJong (2014), in which they interviewed individuals who served as informal mentors to business employees in the Netherlands. Janssen et al. (2014) asked mentors why they supported their trainees’ professional development and used self-determination theory as a framework to code mentors’ responses. Their results suggest that this framework can provide valuable insight into the role that motivation plays in mentoring relationships by illustrating the range of reasons that mentors might engage in mentoring trainees. Janssen and colleagues encouraged future researchers to “examine how motives for specific mentoring functions can vary within and across developmental relationships” (p. 273). To answer this call, we examined what motivates mentors to address the role of cultural diversity in their mentoring relationships.

Using self-determination theory as a framework, we investigated mentors’ level of motivation to address the role that race or ethnicity play in their research mentoring relationships. Our research was guided by the following questions:

1. What motivates mentors to address the role of race/ethnicity in their research mentoring relationships?

2. How does the type of motivation differ:

   1. Based upon the degree of self-reported motivation?
   2. In diverse mentoring relationships?
   3. In mentors who identify with a historically underrepresented race/ethnicity?

Method

Research Design

We chose a mixed-methods approach in designing our study. Mixed methods studies involve collecting and analyzing a combination of quantitative and qualitative data. Mentors’ responses were analyzed using a concurrent triangulation design, which is a mixed-methods approach to data analysis (Creswell, Plano Clark, Gutmann & Hanson, 2003). In this approach, both qualitative and quantitative methods are used to explore the research question as a means of capitalizing on the strengths of each approach while also minimizing the weakness of using a single-method approach (Creswell et al. 2003). In using both qualitative and quantitative approaches, we were able to provide additional evidence of the validity of our findings. We accomplished this by examining the quantitative representation of our codes as a representation of the spectrum of motivation for mentors to address this topic and by corroborating those qualitative codes with mentors’ quantitative reports of their level of motivation. We expected that mentors who reported higher levels of motivation on our quantitative measure would report more internalized (i.e., identified, integrated, or intrinsic) motivation for addressing the role of race/ethnicity in their mentoring relationships, whereas mentors expressing lower levels of motivation on our quantitative measure would more often report amotivated, externally regulated, or introjected responses.

Participants

Participants were 152 mentors from STEM fields--mostly in the biological sciences, who were a subset of a larger sample of mentors throughout the United States (N = 301) that completed a survey between Fall 2015 and Spring 2016 that pilot tested a quantitative measure of cultural diversity awareness in research mentoring relationships. Mentors were recruited via an email to Research Experiences for Undergraduates (REU) site directors and via snowball sampling techniques.

The sample contained slightly more women than men (55% and 45%, respectively), with less than 1% reporting another gender identity. The majority of mentors self-identified as White (88%), 6% self-identified as Hispanic or Latino(a), 4% self-identified as Asian, 2% self-identified as multiracial/multiethnic, and 2% self-identified as Black/African American; 4% chose not to report their race or ethnicity. Sixty percent (60%) of participants were faculty members, 25% were graduate students or postdoctoral researchers, and the remaining 16% held other positions. The distribution of race/ethnicity and gender in this sample is generally reflective of the demographic distribution of faculty members in the biological sciences, as indicated by comparing our sample demographics to those of the top 100 Biological Sciences Departments (Nelson & Brammer, 2010). Individuals who identified as women or HU are overrepresented in our sample. Fifty-five percent (55%) of our sample identified as women, whereas, among the top 100 Biological Sciences departments in 2007, 24% of the faculty were female. Ten percent (10%) of individuals in our sample self-identified as HU compared to 4.1% of Biological Sciences faculty. This overrepresentation of women and individuals from HU groups is likely due to the presence of early career mentors (i.e. graduate students and postdoctoral researchers) in our sample, where women and HU individuals are represented in larger numbers (Nelson & Brammer, 2010).

Procedure

This study was reviewed and approved by the University’s Institutional Review Board as part of a larger study measuring cultural diversity awareness in research mentoring relationships (protocol #2015–1086). Mentors were asked to report their level of motivation to engage in six tasks that are commonly associated with culturally responsive mentoring as part of a survey. These items were originally included as a subscale of the cultural diversity awareness measure mentioned above, but were ultimately removed during the pilot testing process due to the psychometric properties of the items in this subscale and the conceptual overlap with another subscale in the measure. Mentors could select any option on a Likert-type scale ranging from 1 (extremely unmotivated) to 6 (extremely motivated). For each item, mentors were then asked to provide a response as to why they were motivated (or unmotivated) to complete each task. For this paper, we focus on the 115 mentors who indicated their level of motivation to address the role of race/ethnicity in the mentoring relationship. The mean response to this item was 4.25, SD = 1.38. We chose to focus on this item because we felt it most broadly focused on behaviors to address differences in the mentoring relationship as opposed other items which focused on a specific action or self-awareness (e.g., “ask questions about a racial/ethnic experience when I do not understand”; “provide opportunities for mentees to talk about their racial/ethnic identity as it relates to their research experience”).

In addition to these questions, mentors were also asked if they ever had a trainee whose racial/ethnic background was different than their own. Mentors were given the option to respond Yes, No, or I don’t know. Ninety-four percent (94%) of mentors reported that they had had a trainee whose racial/ethnic background was different than their own. Finally, we asked mentors to report their race/ethnicity, which we dichotomized into mentors from historically underrepresented groups (i.e., HUs) and well-represented groups (i.e., non-HUs) in the sciences. The former category consisted of members who identified as Black, Hispanic, and Bicultural; the latter consisted of mentors who identified as White and Asian.

Analytical Approach

Open-ended responses were coded using a deductive (i.e., theory driven) approach. We elected to examine these questions using qualitative methodology. Although the questions asked in the survey provided a level of mentors’ motivation, this measure was limited in its capability to capture the type of motivation that compelled mentors to act (or not) to address race and ethnicity in their research mentoring relationships. As noted above, quantitative measures of motivation have used SDT to examine similar constructs (e.g., motivation to reduce prejudice; LeGault et al. 2007). However, with this being the first study to examine the motivation to address race within the context of research mentoring relationships, we felt that examining mentors’ open-ended responses using the SDT framework provided a foundation from which future researchers can begin to develop more multidimensional quantitative measures of motivation. Using the spectrum of motivation outlined in self-determination theory, we identified three “parent” codes, which represented the three types of motivation articulated by Deci and Ryan (2000): amotivation, extrinsic motivation, and intrinsic motivation. Within each parent code, “child” codes were created to represent the different subcategories of motivation described in self-determination theory. Four child codes were added under the amotivation parent code to further differentiate the reasons why an individual might be amotivated: lack of relevance/value; lack of competence; lack of desired outcome, and not wanting to put forth the effort. Three child codes were added under the extrinsic motivation parent code to reflect the continuum of extrinsic motivation as described in self-determination theory: external regulation, introjected regulation, and identification/integration. The codes for identification and integration were collapsed into one code due to the difficulty of distinguishing between responses from mentors that reflected identification or integration. A response reflecting integration would indicate an assimilation of the value into the mentor’s identity on an ongoing basis; something that was difficult to determine given the cross-sectional nature of the data. To keep our coding process as objective as possible, we presented identification and integration together as one code representing the autonomous forms of extrinsic motivation, consistent with other composite variables of autonomous motivation that have been used in prior research (Ryan & Deci, 2000). The parent code intrinsic motivation did not have any child codes assigned, as different types of intrinsic motivation are not articulated within self-determination theory.

Two additional parent codes emerged inductively as part of the coding process. These codes captured responses that did not fit into the a priori theoretical framework and coding scheme. Uncategorized was created to capture responses unrelated to the question asked, such as when a mentor wrote text that was not directly answering the question. In the cases where mentors referred to previous responses (n = 5), we confirmed that previous responses directly addressed the question of why they would be motivated to address the role of race/ethnicity in their research mentoring relationships and then proceeded to code those responses. Other was the second parent code that emerged inductively. Two child codes were developed under this parent code. Strategy was created to capture responses that were relevant in some way to culturally responsive mentoring, but instead of describing why mentors were more or less motivated to bring up race and ethnicity in their mentoring relationships, these responses described how mentors approached or avoided the topic. Responses that did not fit into strategy were coded as other. Our final coding framework (presented in Table 1) consisted of 5 parent codes and their respective child codes.

Table 1.

Coding Categories and Example Responses

Code	N (%) of Codes Assigned	Example
Amotivation	35 (30%)
Not Relevant/Lacking Value	25 (22%)	I don’t believe race/ethnicity should play a role in mentoring
Lacking Competence	3 (3%)	I would not feel qualified
Not Leading to Desired Outcome	5 (4%)	I don’t think a mentee who is a person of color would appreciate me drawing attention to the fact that I am white and they are not.
Effort	2 (2%)	I prefer to focus on research and how to be successful by overcoming obstacles of all kinds (my ability to provide strategies encompasses race/ethnicity).
Extrinsic Motivation	59 (51%)
External Regulation	11 (10%)	I do not want to bring race/ethnicity into the relationship as a pertinent factor unless the mentee does first.
Introjected Regulation	6 (5%)	I want my mentees to know that I don’t view myself as superior to them in any way, just that I’m further along in the path they are taking and therefore I have experiences I can share with them so they too can succeed.
Identification/Integration	47 (41%)	I want to provide all my mentees with a chance to thrive, and part of doing that is knowing how my privilege as a white person might be affecting the way I interact with my mentees who come from different backgrounds.
Intrinsic Motivation	0 (0%)
		I am keenly interested in cultural diversity1.
Other Codes	18 (16%)
Other	11 (10%)	Would value best-practices advice on this issue.
Strategy	7 (6%)	I judge when it is and it is not appropriate to bring it up.
Uncategorized	3 (3%)	.

Note. N = 115. Child code percentages may not add up to parent code percentages due to discordant codes at the child code level.

¹This example was taken from another open-ended prompt in the same survey and is included for illustrative purposes.

Coding and Analyses

The open-ended data were first analyzed qualitatively by a team of coders using deductive coding. The coding team consisted of two White females and one Latino male. A fourth team member who identified as a White female contributed to discussions of coding definitions and theory and consulted with the coding team when coding disagreements arose. The team discussed the coding definitions and examples and their relationship to self-determination theory over the course of several months to ensure that all team members felt comfortable with the coding scheme and the process of determining appropriate codes for each response. Several practice rounds of coding with coding definitions and example responses were used to ensure that definitions were clear and interpreted similarly among the coding team. Discussions among the coding team led to continual refinement of the coding definitions and organization of the parent and child codes. Based on these discussions, we determined that a two-stage coding process, where responses were first coded using parent codes and then further coded using the child codes, would be the best analytical approach. Prior to analysis, mentors’ career stage, race, and gender were removed from the dataset that was used by the coding team. The first round of coding using the finalized coding scheme was conducted using the five parent codes and yielded an initial coder agreement of 81% over 115 responses. The coders then met to discuss any discordant codes until 100% agreement was achieved.

The second round of coding consisted of the coding team using child codes under each of the parent codes to further investigate the type of motivation represented in each mentor’s response. The initial coder agreement for the second round of coding was 80%; coders subsequently discussed discordant codes until 97% agreement was reached; consensus was unable to be reached on four responses. For those four responses, the different codes assigned by team members were each included in the final tabulations of the number of codes assigned.

Three instances of disagreement stemmed from responses that we felt could be coded as either identified/integrated or as introjected regulation. For example, a mentor who identified as White and Female indicated that she was extremely motivated to address the role of race and ethnicity in the mentoring relationship and offered up this reason why: “The first time a doctoral student accused me of racism really got my attention! The student objected to using statistical examples in which [African Americans] do worse [than] other groups. It was a real eye opener for me.”

One coder noted that this response reflected that the mentor was motivated to address the role of race and ethnicity to avoid being called racist. The coder recommended that this response be coded as introjected regulation because it reflected motivation as a result of avoiding embarrassment or shame. Other coders on the team felt that this anecdote served as the turning point for this mentor, helping her to realize the importance of taking a culturally responsive stance toward her research. As such, they recommended coding this response as identification/integration.

Disagreements like this one led to numerous discussions throughout our coding process. Although our team often came to consensus, there were times, like the example presented above, where consensus could not be reached. These instances of disagreement were raised by all members of the coding team. These disparate coding perspectives mirror the differing perspectives that mentors and trainees must contend with in the research experience. In this case, we determined it was best to openly acknowledge that differences in team members’ lived experiences, privilege, and identities likely contributed to the different coding assignments. Once the data were coded, we next examined the data quantitatively by calculating the number and percentage of codes assigned for the full sample, by level of motivation and by the differences in racial/ethnicity identity reported by mentors.

Positionality of research and coding team.

We wish to acknowledge that each of the authors brings both shared and unique perspectives to this study. As members of a larger research team studying the research mentoring relationships of mentors and trainees from historically underrepresented racial and ethnic groups, we all share a common commitment to critically examining how personal and professional identities intersect. Our approach to studying mentoring relationships is rooted in Critical Race Theory and Culturally Relevant/Sustaining Pedagogy (Ladson-Billings, 1995; Ladson-Billings & Tate, 1995; Ladson-Billings, 2014). That is, we contend that racial identity and racism perpetuate inequities in many aspects of life, including research training, and that those with the power to cultivate the next generation of scholars (e.g., mentors) must critically examine how their identities and the identities of their trainees influence the training environment. In this spirit, we acknowledge that each of us came to this project with prior experiences, privilege, and bias that inevitably shaped the way in which we interpreted this data. During the coding process and at research team meetings we allowed time to share how our perspectives informed our coding decisions and interpretations of the coding definitions and to thoughtfully challenge one another when discrepant coding occurred.

Results

In this section we present the pattern of data that emerged from our coding process for each of our research questions along with illustrative quotes from the mentors. Using the analytical approach described above, we examined the quantitative and qualitative responses of 115 research mentors. We next describe the results for each of our research questions.

RQ1: What Motivates Mentors to Address the Role of Race/Ethnicity in Their Research Mentoring Relationships ?

To answer our first research question, we examined the data patterns for the full sample. The percentage of responses assigned to each parent and child code, along with example responses, appears in Table 1. Below we describe in further detail the responses that reflected mentors who were motivated to address race, mentors who were not motivated to address race, and mentors whose responses did not immediately fit into our coding framework.

Motivated mentors.

Extrinsic motivation was the most frequently assigned parent code (51% of all responses). Within this parent code, the top child code assigned reflected identified or integrated stances towards addressing the role of race and ethnicity in the research mentoring relationship (41%). Many mentors whose responses were coded as extrinsic motivation shared answers that reflected an identified or integrated motivation to address race in the mentoring relationship, such as this response provided by a Male and White-identifying tenured faculty member: “Diversity differences in mentors and trainees is a critical point for understanding the outcome from mentoring.” Another powerful example from a mentor who identified as Hispanic and Male described how he integrates cultural identity into his mentoring relationships:

I tell the students my history and the way I interpret the value of science to allow them to have the conversation for themselves. I do this with white students and students of color. This allows them to explore topics they wonder about and acknowledge that identity is important for why they pursue their journey and [how] we must interact with others.

By modeling this approach, this mentor is showing his trainees that contemplating how your racial/ethnicity identity informs your research and career pathway can ultimately lead to a deeper understanding of their research interests and foster a respect for the individualized pathway that each scientist takes toward his/her goals. Such examples reflect mentors who recognize that one’s cultural and scientific identity are not necessarily mutually exclusive (Bumpus, 2015).

Among the responses that were coded as external regulation (10%), a few mentors noted that they would only be inclined to address the role of race or ethnicity in the mentoring relationship if the trainee brought it up first, as stated by this mentor who identified as Female and Hispanic; “I think that addressing the topic of race in the mentoring relationship should be up to the student, but if they feel like they need to address than [sic] I would have no problem discussing.” This response reflects an externally regulated motivation to discuss race or ethnicity in the mentoring relationship due to the mentor’s deferral to an outside cue (i.e., the trainee) to regulate their action or behavior. Within self-determination theory, this is considered to be the most controlled form of extrinsic motivation.

The least frequently assigned child code within extrinsic motivation was introjected regulation (5%). These mentors’ responses reflected concerns about appearing a certain way to trainees or others, as reflected in this response from a White and Male-identified mentor:

I want my trainees to know that I don’t view myself as superior to them in any way, just that I’m further along in the path they are taking and therefore I have experiences I can share with them so they too can succeed.

In this example, the mentor emphasizes first and foremost that he does not want to appear superior to his trainees, but wishes to support them by serving as a model of success. The emphasis of this mentor on the importance of not appearing superior indicates that he is particularly mindful of how he is perceived by his trainees, and his actions are controlled by their perception. No responses reflected an intrinsic motivation to address race or ethnicity in the research mentoring relationship. In other words, mentors in this sample did not provide responses that reflected inherent interest or curiosity in broaching the topic of race or ethnicity with their trainees.

Unmotivated mentors.

Amotivation was the next most frequently assigned parent code (30% of responses). This code reflected a mentor’s lack of intention to address the role of race/ethnicity in the mentoring relationship. Within amotivation, the highest percentage (22%) of child codes assigned related to responses indicating that race/ethnicity was not relevant to the mentoring relationship as expressed by this mentor who identified as a White Male: “The best part about science is that it doesn’t care about race or ethnicity (some scientists care, but that’s immaterial). My aim is to mentor students at their ability level, but without directly acknowledging race/ethnicity explicitly.”

A few mentors (3%) cited a lack of competence (e.g., “I would not feel qualified”) as the reason why they would not address the role of race/ethnicity in the research mentoring relationship. Others (2%) expressed a preference for focusing their effort on other aspects of the mentoring relationship, like research. Finally, a few mentors (4%) believed that talking about race/ethnicity would not lead to a desired outcome, as evidenced by this response from a White Male mentor: “I feel like if I bring up the possibility of bringing up race/ethnicity it could create an issue where there was none before.”

Other responses.

The third most frequently assigned parent code contained responses that did not directly address why mentors felt more or less motivated to address race and ethnicity in their mentoring relationships (i.e., other; 16%). A closer look at the responses in this category revealed that they primarily fell into two areas; those that conveyed motivation to broach the topic but a lack of confidence in their ability to do so successfully, and those that offered examples of strategies that mentors employed to approach or avoid the topic. Mentors whose responses were coded as other (10%) often conveyed interest, but a lack of certainty on how best to bring it up, as illustrated by a White Female mentor who indicated that she would “Like to know what to do if such questions come up.” Mentors also shared their strategies for broaching or for avoiding the role of race/ethnicity in mentoring relationships (i.e. strategy; 6%). To talk about race, a mentor who identified as Hispanic Female noted “I judge when it is and it is not appropriate to bring it up.” To avoid race, a White Female mentor said that “I treat everyone the same.” Although these responses do not directly address why some mentors address or avoid the topic of race/ethnicity in their mentoring relationships, such responses still offer an insight into where mentors fall along the motivation spectrum. The former response highlights a measured approach to talking about race/ethnicity, as the mentor recognizes that there are times that may be better suited to talk about race and ethnicity than others. The latter response infers that this mentor takes a one-size fits all approach to working with her trainees. This suggests to us that this mentor may not see the relevance of racial/ethnic identity to her research mentoring relationships.

Summary.

The results above reveal that the majority of mentors reported extrinsically motivated reasons for why they address the role of race and ethnicity in their mentoring relationships, followed closely by responses from mentors that conveyed a lack of motivation to broach this topic in their mentoring relationships. The examples provide some insight into how these different types of motivation manifest themselves among mentors who are considering whether to address race in their mentoring relationships. We next examined the data patterns by self-reported degree of motivation (RQ2a), whether mentors had mentored a trainee whose race/ethnicity was different from their own (RQ2b), and the racial/ethnic identity of the mentor (RQ2c) to determine if any descriptive differences emerged in the data.

RQ2(a): Differences in Type of Motivation by Degree of Self-Reported Motivation

Table 2 provides the frequency of codes assigned for each possible response to the question “How motivated are you to address the role of race/ethnicity in the mentoring relationship?” We found that amotivation was the most frequently assigned code for responses from individuals who indicated that they were unmotivated on some level to address the role of race/ethnicity in the mentoring relationship. All individuals who responded that they were extremely unmotivated to address the role of race/ethnicity in the mentoring relationship provided responses to the open-ended prompt that conveyed a belief that race/ethnicity was not relevant to the mentoring relationship. Most individuals (73%) who responded that they were moderately unmotivated to address the role of race/ethnicity provided a similar response. Many of the mentors who indicated that they were only somewhat unmotivated to address the role of race/ethnicity in their mentoring relationships conveyed a lack of relevance in their responses (41%), but several expressed that it was the trainee’s responsibility to bring it up (i.e., external regulation; 18%).

Table 2.

Motivation Codes Assigned by Level of Motivation

Code Assigned		Unmotivated			Motivated
	Full Sample (N = 114)	Extremely Unmotivated (n = 4)	Moderately Unmotivated (n = 11)	Somewhat Unmotivated (n =22)	Somewhat Motivated (n = 26)	Moderately Motivated (n = 24)	Extremely Motivated (n = 27)
Amotivation	35 (31%)	4 (100%)	11 (100%)	13 (59%)	6 (23%)	0 (0%)	1 (4%)
Not Relevant/Lacking Value	25 (22%)	4 (100%)	8 (73%)	9 (41%)	3 (12%)	0 (0%)	1 (4%)
Lacking Competence	3 (3%)	0 (0%)	1 (9%)	1 (5%)	1 (4%)	0 (0%)	0 (0%)
Not Leading to Desired Outcome	5 (4%)	0 (0%)	2 (18%)	3 (14%)	0 (0%)	0 (0%)	0 (0%)
Effort	2 (2%)	0 (0%)	0 (0%)	0 (0%)	2 (8%)	0 (0%)	0 (0%)
Extrinsic Motivation	59 (51%)	0 (0%)	0 (0%)	5 (23%)	12 (46%)	18 (75%)	24 (89%)
External Regulation	11 (10%)	0 (0%)	0 (0%)	4 (18%)	4 (15%)	2 (8%)	1 (4%)
Introjected Regulation	6 (5%)	0 (0%)	0 (0%)	0 (0%)	1 (4%)	3 (13%)	2 (7%)
Identification/Integration	47 (41%)	0 (0%)	0 (0%)	1 (5%)	7 (27%)	16 (67%)	23 (85%)
Intrinsic Motivation	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Other/Strategy	17 (15%)	0 (0%)	0 (0%)	3 (14%)	6 (23%)	6 (25%)	2 (7%)
Other	10 (8%)	0 (0%)	0 (0%)	2 (9%)	2 (8%)	5 (21%)	1 (4%)
Strategy	7 (6%)	0 (0%)	0 (0%)	1 (5%)	4 (15%)	1 (4%)	1 (4%)
Uncategorized	3 (3%)	0 (0%)	0 (0%)	1 (5%)	2 (8%)	0 (0%)	0 (0%)

Note. One individual did not include their reported level of motivation and as a result was not included in this table. Percentage of codes represents the ratio of the number of codes assigned in a designated coding category for a given group (numerator) to the total number of individuals, excluding those with missing or uncategorized responses. Percentages were rounded to the nearest whole number, so percentages may add up to over 100 percent in some cases.

Extrinsic motivation was the most frequently assigned parent code for mentors who indicated that they were either somewhat (46%), moderately (75%), or extremely motivated (89%) to address the role of race/ethnicity in the mentoring relationship. The percentage of responses that were coded as identification/integration increased as individuals reported higher levels of motivation; 27%, 67%, and 85% of individuals reporting that they were somewhat, moderately, or extremely motivated, respectively had responses that were coded as identification/integration.

For a few individuals, responses to the Likert-question did not match up with the code assigned to their response. Six mentors (23%) who indicated that they were somewhat motivated to address the role of race/ethnicity in the mentoring relationship provided reasons that were coded as amotivation. One White Male mentor who indicated that he was extremely motivated provided this response: “Would help me being [sic] a better mentor, although I don’t think race/ethnicity should matter in science - I think everybody is equal.” This response was coded as amotivation, because although the mentor seemed to convey that it might make him a better mentor, he ultimately believed that race/ethnicity had no place in science (i.e., it was not relevant).

RQ2(b): Differences by Experience Mentoring Trainees from Different Racial/Ethnic Groups

Table 3 presents frequency of codes assigned based upon whether a mentor has had a trainee whose race/ethnicity was different than his or her own. It should be noted that very few mentors in our sample (n = 4) had not mentored a trainee from a different racial/ethnic background. One mentor indicated that they did not know if they had mentored a trainee whose race/ethnicity was different than their own. The majority of individuals in each group (i.e., those who had mentored an individual of a different race/ethnicity and those who had not) had responses that were coded as extrinsic motivation. Among mentors who had mentored trainees of a different race/ethnicity, 51% provided responses that were coded as extrinsic motivation, with most responses aligning with an identified or integrated motivation to address the role of race/ethnicity in the mentoring relationship. Thirty-one percent (31%) provided responses that were coded as amotivation, with the majority of those responses referring to a lack of relevance.

Table 3.

Motivation Codes Assigned by Whether Mentor Had Worked with a Mentee of a Different Racial/Ethnic Background

Code Assigned	Full Sample (N = 115)	Yes (n = 110)	No (n = 4)	I don’t know (n = 1)
Amotivation	35 (30%)	34 (31%)	0 (0%)	1 (100%)
Not Relevant/Lacking Value	25 (22%)	24 (22%)	0 (0%)	1 (1%)
Lacking Competence	3 (3%)	3 (3%)	0 (0%)	0 (0%)
Not Leading to Desired Outcome	5 (4%)	5 (5%)	0 (0%)	0 (0%)
Effort	2 (2%)	2 (2%)	0 (0%)	0 (0%)
Extrinsic Motivation	59 (51%)	56 (51%)	3 (75%)	0 (0%)
External Regulation	11 (10%)	11 (10%)	0 (0%)	0 (0%)
Introjected Regulation	6 (5%)	5 (5%)	1 (25%)	0 (0%)
Identification/Integration	47 (41%)	45 (41%)	2 (50%)	0 (0%)
Intrinsic Motivation	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Other/Strategy	18 (16%)	18 (16%)	0 (0%)	0 (0%)
Other	11 (10%)	11 (10%)	0 (0%)	0 (0%)
Strategy	7 (6%)	7 (6%)	0 (0%)	0 (0%)
Uncategorized	3 (3%)	2 (2%)	1 (25%)	0 (0%)

Note. Percentage of codes represents the ratio of the number of codes assigned in a designated coding category for a given group (numerator) to the total number of individuals, excluding those with missing or uncategorized responses. Percentages were rounded to the nearest whole number, so percentages may add up to over 100 percent in some cases.

RQ2(c): Differences by Mentor’s Self-Reported Race/Ethnicity

Finally, the frequency of codes assigned based on the race/ethnicity of the mentor (i.e., whether they belonged to a historically underrepresented group in the sciences) was examined. Table 4 presents the coding frequencies for each of these groups. Both groups had the majority of their responses coded as extrinsic motivation; a higher percentage of White and Asian (i.e., historically well-represented) mentors had responses that were coded as amotivation (32%) compared to mentors from HU groups (15%). A higher percentage of responses from mentors from HU racial/ethnic groups (62%) were coded as extrinsic motivation compared to their colleagues from historically well-represented groups (46%).

Table 4.

Motivation Codes Assigned by Historically Underrepresented (HU) Status of Mentor

Code Assigned	Full Sample (N = 114)	HU (n = 13)	Not HU (n = 101)
Amotivation	34 (30%)	2 (15%)	32 (32%)
Not Relevant/Lacking Value	24 (21%)	2 (15%)	22 (22%)
Lacking Competence	3 (3%)	0 (0%)	3 (3%)
Not Leading to Desired Outcome	5 (4%)	0 (0%)	5 (5%)
Effort	2 (2%)	0 (0%)	2 (2%)
Extrinsic Motivation	59 (51%)	8 (62%)	51 (51%)
External Regulation	11 (10%)	3 (23%)	8 (8%)
Introjected Regulation	6 (5%)	0 (0%)	6 (6%)
Identification/Integration	47 (41%)	5 (39%)	42 (42%)
Intrinsic Motivation	0 (0%)	0 (0%)	0 (0%)
Other/Strategy	18 (16%)	2 (15%)	16 (14%)
Other	11 (10%)	1 (8%)	10 (10%)
Strategy	7 (6%)	1 (8%)	6 (6%)
Uncategorized	3 (3%)	1 (8%)	2 (2%)

Note. One individual did not report his/her race/ethnicity and was excluded from this analysis. White and Asian mentors were included in the Not HU category; individuals identifying as Black, Hispanic, or Bicultural were included in the HU category. Percentage of codes represents the ratio of the number of codes assigned in a designated coding category for a given group (numerator) to the total number of individuals within that group. Percentages were rounded to the nearest whole number, so percentages may add up to over 100 percent in some cases.

Discussion and Implications

Self-determination theory (Ryan & Deci, 2000) posits that supporting competence, relatedness, and autonomy are central to promoting sustained motivation and engagement. Within self-determination theory, organismic integration theory provides a useful framework for understanding the range of motivation that mentors have for addressing race and ethnicity in their research mentoring relationships.

Mentoring plays an important part in the professional development of trainees in STEM. In particular, the mentoring received by trainees from HU racial/ethnic groups is often cited as a pivotal factor in trainees’ self-beliefs, persistence, and academic success in STEM (Byars-Winston et al. 2015; Griffin, Pérez II, Holmes, & Mayo, 2010; McCoy, Luedke, & Winkle-Wagner, 2017). Following the work of other scholars who have investigated cross-cultural mentoring from the perspective of mentors (e.g., McCoy, Winkle-Wagner, & Luedke, 2015), we sought to better understand the reasons why mentors choose to discuss or avoid racial/ethnic matters as they relate to the research mentoring relationship.

Using mentors’ responses to a Likert-type item in combination with their open-ended responses, we examined the different types of motivation by degree of motivation, by whether the mentor had a cross-racial/ethnic mentoring relationship, and by the mentor’s race/ethnicity. Most mentors in our sample aligned either with an integrated/identified type of motivation who see importance and value in addressing the role of race/ethnicity in the mentoring relationship or an amotivated stance who believe these factors are not relevant to the research mentoring relationship. Nevertheless, with the exception of intrinsic motivation, we discovered responses from mentors that represented the full spectrum of motivation outlined within self-determination theory. Below, we discuss our key findings, highlight the significance of this research for efforts to diversify science, and consider the implications of our findings for mentoring and future research directions.

Mentors Address Race/Ethnicity for Many Reasons

The highest percentage of codes assigned in our sample were within the extrinsic motivation parent code. Mentors who conveyed an identified or integrated type of motivation for addressing race/ethnicity comprised the largest percentage of child codes assigned in the full sample. Identification/Integration was also the most frequently assigned child code among mentors who indicated that they were somewhat, moderately, or extremely motivated, and regardless of whether or not the mentor had mentored a trainee of a different race/ethnicity or if they themselves were a member of a HU racial/ethnic group. Mentors who professed an identified or integrated motivation to address race and ethnicity in their research mentoring relationships are more likely to be engaged in discussing this topic and to have a general feeling of competence in their capabilities to effectively broach the subject with their trainees. They are also more likely to be more autonomous and agentic in their actions, not waiting for others to bring up the subject of race or ethnicity first.

Whereas mentors aligned with an integrated/identified type of motivation see importance and value in addressing the role of race/ethnicity in the mentoring relationship, mentors who endorse introjected or externally regulated forms of motivation may feel that they are competent in their capability to discuss race and ethnicity with their trainees, but might have varying reasons for doing so. In our sample, mentors who express introjected forms of regulation are more likely to talk about race and ethnicity if it makes them look good or to avoid feelings of guilt or anxiety. These mentors might raise the topic themselves, but struggle internally with the reasons behind their actions. For example a mentor might bring up the subject in response to recent events, statements, or mandates from their institution that make them feel guilty for not already addressing race or ethnicity in their research mentoring relationships. Many of the mentors who expressed externally regulated motivation for addressing race and ethnicity in their mentoring relationships were likely to talk about race and ethnicity only if the trainee brought it up first. Mentors who defer to the trainee to bring up this topic may be well-intentioned; after all, why would they want to put their trainee in a possibly awkward position if race/ethnicity has not been raised by the trainee as important or an issue? Nevertheless, externally regulated motivations may in fact have an adverse effect, making others perceive individuals as biased or prejudiced (LeGault et al. 2011). In other words, the silence from mentors on the topic of race/ethnicity in the research mentoring relationship could be perceived by the trainee as ignoring an important aspect of his or her experience or social identity. The power differential often present in research mentoring relationships may cause some trainees to be hesitant to bring up the topic, even if they want to (Byars-Winston, Leverett et al., 2017). This puts the pressure on trainees, who are often concerned with pleasing their mentor and doing what it takes to be successful in their profession, to bring up the subject. A key component of successful research mentoring relationships is the trust built between mentor and trainee; if mentors wait until a trainee brings up the issue, they may miss out on an opportunity to build trust within the relationship, operating with a genuine appreciation and knowledge of others’ experience. This may be particularly important in cross-cultural mentoring relationships (Bailey & Cervero, 2004). Mentors who acknowledge that trainees are conducting research in a cultural context can open the door for HU trainees to feel welcomed and valued (Bultsma, 2011).

Some Mentors Are Not Motivated to Address Race/Ethnicity

The prevalence of amotivation-oriented responses in our sample suggests that many mentors still do not see the relevance of race/ethnicity to research mentoring. Many mentors who provided amotivated responses promoted a racial colorblind approach by emphasizing the point that race and ethnicity have nothing to do with mentoring within STEM. For example, a tenured faculty member who identified himself as White noted “I firmly believe that a strong mentoring relationship transcends race/ethnicity.” Regardless of the good intentions that this mentor may have, responses like this one leave the trainee responsible for interpreting the often ambiguous silence around race and ethnicity, especially as the word transcend has nuanced meanings that might be interpreted differently by different individuals. Trainees might take this to mean that their mentor will go above and beyond superficial conversations about race to foster a deep meaningful relationship; alternatively, trainees may assume that their mentor finds discussions about racial and ethnic identity negatively impact an otherwise good mentoring relationship, making it uncomfortable to broach the subject.

Recent findings suggest that this approach is not beneficial to interpersonal relationships (McCoy et al. 2015; Prunuske et al. 2013). Adopting a racial colorblind approach when interacting with individuals from HU racial and ethnic groups can result in additional cognitive strain (Holoien & Shelton, 2012), leave students feeling disempowered, and can promote a meritocratic philosophy (i.e., everyone has an equal chance of being successful if they just work hard enough; McCoy et al. 2015). Moreover, numerous studies have shown that implicit race-based biases and assumptions persist, and can be further exacerbated in the face of color-blindness (Plaut, 2010). Though there are trainees who may not want to discuss race/ethnicity nor address them in their mentoring relationships, some do (Acosta & Ackerman-Barger, 2017; Byars-Winston, Leverett et al. 2017). In the absence of these conversations, important pillars of mentoring, such as trust and communication, can break down.

A higher percentage of mentors from racial and ethnic groups that have been historically well-represented in STEM (i.e., mentors who identified as White or Asian) provided responses that were coded as amotivated compared to their HU colleagues. It bears noting that, although Asian mentors and trainees are well-represented in STEM, they may face different but nonetheless challenging stereotypes and bias in the STEM environment (e.g., the model minority myth; McGee, Thakore & LaBlance, 2017). Due to the low number of mentors who identified as Asian in our sample (n = 5), we were unable to examine responses from Asian mentors separately. However, alternative sampling methods focused on obtaining a representative sample of mentors from diverse racial and ethnic backgrounds could reveal more nuanced differences in motivation among mentors of different races and ethnicities and would be an important contribution to the literature.

Can Mentors be Intrinsically Motivated to Address Race/Ethnicity?

It is important to note that no responses to the open-ended prompt were coded as intrinsic motivation. There are several possible explanations for this. First this difference could be due to the wording of the open-ended prompts. For example, one survey prompt asked mentors how motivated they are to ask questions about a racial/ethnic experience, which implies that the mentor has some curiosity or interest, whereas the impetus for addressing the role of race or ethnicity in the mentoring relationship may be more externally imposed on the mentor. Another possibility could be that it is difficult for individuals to be intrinsically motivated to address race due to cultural norms in the United States that promote a racial color-blind approach (Plaut, 2010). Mentors may also receive opposing external messages on the value or importance of race/ethnicity in research mentoring relationships and therefore have to make a decision whether discussing these factors is appropriate and/or desirable. Thus, individuals would be most likely to internalize the importance and value of addressing race/ethnicity into their worldview through increased self-awareness and self-education as opposed to an inherent curiosity or need. Our results partially answer the call by Janssen et al. (2014) to explore mentors’ motives for one specific task within the mentoring relationship, which is addressing the role of race and ethnicity. We encourage other researchers to use this framework to explore mentors’ motivation for other tasks to deepen our understanding of the specific ways in which intrinsic motivation might contribute to the research mentoring relationship.

We also discovered some responses that did not fall within the a priori codes based on self-determination theory. Among these responses were individuals who indicated that they were motivated to address the topic of race/ethnicity, but lacked the competence or confidence to do so. This hesitance to have these conversations is often accompanied by a fear of saying or doing the wrong thing (Byars-Winston, Leverett et al., 2017). This suggests that mentors could benefit from opportunities to reflect upon their broaching style toward engaging with sensitive or difficult topics (Carroll & Barnes, 2015). Further, they may benefit from practicing broaching strategies in low-risk, role playing exercises, and thereby build their confidence to successfully engage in conversations about race/ethnicity with their trainees.

Higher Levels of Motivation Align with Extrinsic Motivation to Address Race/Ethnicity

For the most part, mentors who indicated a higher degree of motivation on the Likert-type scale had responses that were coded as extrinsically motivated, with higher percentages of responses coded as identification/integration among mentors who responded that they were either moderately or extremely motivated to address the role of race/ethnicity in their mentoring relationships. This suggests that mentors with comparatively higher levels of motivation are more likely to express reasons that are more autonomous and self-directed. In other words, mentors who are more motivated appear to have internalized, at least to some extent, the idea that race/ethnicity can play a role in a trainee’s professional development.

A few individuals who reported being somewhat, moderately, or extremely motivated on the Likert-type scale to address the role of race/ethnicity provided responses were coded as amotivation. There are a few possible explanations for this apparent contradiction. One possibility is that individuals who responded to the Likert-type questions were not closely attending to the values that they were endorsing. It is also possible that mentors felt that a motivated quantitative response was more socially desirable, or that they interpreted the Likert-type question in a way different than was intended (i.e., mentors were motivated to avoid the topic of race/ethnicity). Regardless of how we interpret the quantitative responses of these participants, the open-ended responses suggest that these mentors believe that racial/ethnic matters have no relevance to the mentoring relationship.

Implications for Mentoring

Across the spectrum of motivation, we found instances where mentors expressed wanting to make STEM a more welcoming place for HU trainees. Despite the good intentions of mentors who want to ignore race and focus on the research, or wait for the mentee to bring up the subject to avoid embarrassment or feelings of discomfort, these tactics can in fact backfire, making mentees feel unwelcome or unsure as to whether they can integrate their personal identity with their professional identity. Mentors at all ends of the motivation spectrum may benefit their mentees by moving towards a more internalized motivation to address race and ethnicity and contribute to a research and institutional environment that fosters their competence, autonomy, and relatedness.

Some mentors expressed trepidation that addressing race/ethnicity may make trainees feel singled out or marginalized or were skeptical about the possible connection of racial or ethnic identity to the success of trainees in STEM. Regardless of the reasons behind their motivation, mentors who have up until this point only talked about the research with their trainees might find the prospect of discussing something as personal as social identity daunting. For mentors who do not see the relevance of race/ethnicity to mentoring or who lack confidence to broach the topic with their trainees, evidence-based mentor training interventions (e.g., Byars-Winston, Womack et al. in press; Carnes et al. 2015) may be an effective way for mentors to learn about the science that supports the connection between cultural diversity and the research experience. Mentor training that raises the cultural awareness of mentors (e.g., Byars-Winston, Womack et al. in press) could increase both the knowledge of how addressing diversity matters may benefit all trainees in STEM and provide them opportunities to actually practice culturally responsive strategies that can effectively support all trainees, especially those from HU groups. Opportunities to build competence could be particularly useful for those mentors who were amotivated or relied upon more controlled forms of motivation (e.g., external regulation, introjected regulation; Deci & Ryan, 2000). Evidence-based mentor trainings provide “low stakes” opportunities for mentors to try out different broaching styles with the help of a skilled facilitator, which may allay some of the anxiety that mentors have about unintentionally offending their trainees. Likewise, trainees may also potentially benefit from opportunities to learn how to broach sensitive topics, like race, with their mentors. This includes trainees from HU as well as well-represented groups, as both need preparation to work and conduct research in culturally diverse contexts and research teams. Future interventions and trainings that explore the role of cultural identity in mentoring relationships would be well-served to invite mentors to examine their own motivations for addressing (or not addressing) this topic as a means of raising cultural awareness and promoting honest conversations about identity as it relates to trainees’ research experiences. Such trainings can not only support mentors’ competence for effectively broaching the subject with their trainees, but also can help mentors understand the importance and value in creating opportunities for trainees to talk about their identity. This can help create a sense of agency in mentors (i.e., autonomy) that will encourage them to further internalize these values and integrate them with their other values and beliefs (e.g., wanting to be a good mentor and to effectively train the next generation of researchers).

Because mentors cannot know whether or not racial/ethnic matters are important to their trainees or are of interest to them to discuss, we encourage mentors to consider having general conversations early in mentoring relationships that signal that racial/ethnic topics are acceptable to discuss. Mentors might consider including items in their office or research space that represent their social identity (e.g. racial/ethnic heritage, family photos, posters or items that represent places, people, and ideas important to you); such “artifacts” can indicate to trainees that you are open to bringing your identity into the research space and that they are welcome to do so as well. Mentors might also consider asking their trainees why they chose their major or career, or disclosing why mentors themselves chose a career path or instances where they encountered difficulties. Such approaches open the doors to conversations about the different contextual factors, including social identity that can influence career decision-making and the learning experience. In other words, we encourage mentors to lead discussions by example by being prepared and willing to share as appropriate. To help build confidence and readiness to engage in culturally responsive mentoring, individual faculty members, graduate students, and postdoctoral researchers mentoring trainees may want to take advantage of synchronously and asynchronously available mentor training opportunities available through organizations like the National Research Mentoring Network (NRMN), which offers training, resources, and community for research mentors committed to improving research mentoring in STEM.

In addition to implications for individual mentor practice, our findings point to the need for commitment from institutions to build mentors’ capacity to thoughtfully engage in race/ethnicity salient discussions. An institutional commitment to broadening diversity in STEM necessitates increasing the racial/ethnic diversity of individuals in these disciplines by increasing the capacity of individuals to make STEM disciplines a more welcoming and inclusive space. To this end, institutions of higher education can support research mentor training offered as professional development alongside research experiences for trainees (Pfund, Maidl Pribbenow, Branchaw, Miller Lauffer & Handelsman, 2006). Mentor training curriculum, such as Entering Mentoring (Pfund, Branchaw, & Handelsman, 2015) is widely available and shown to be effective in improving research mentoring relationships in the sciences (Pfund et al. 2006). Established at the University of Wisconsin-Madison and used at colleges and universities nationwide, this curriculum focuses on competencies of effective mentoring relationships such as aligning expectations, promoting professional development, maintaining effective communication, addressing equity and inclusion, assessing understanding, fostering independence, cultivating ethical behavior, and articulating a mentoring plan. Program directors may do well to provide professional development opportunities using these evidence-based trainings as part of preparation for mentors working with students. By creating a departmental and institutional culture that promotes and actively engages in these discussions, addressing race and identity becomes less of an exception and more of a norm. In so doing, departments and institutions can foster a sense of belonging and connectedness (i.e. relatedness) around the common goal of broadening participation in STEM.

Limitations

There are some limitations to this study. The data for this study came from a convenience sample of research mentors. Though the demographic composition of our sample is similar to the demographic composition of biological sciences faculty, we acknowledge that this data may not be representative of the entire population of research mentors in the biological sciences. Future studies with random sampling methods should be conducted to further test the generalizability of these findings. There are also some limitations to this study with respect to the wording of items and response options. The response options to our question, “how motivated are you to address the role of race/ethnicity in the mentoring relationship?” (i.e., extremely unmotivated to extremely motivated) may have made it difficult for individuals to provide an amotivated response. Changing the response scale for these items to better represent the spectrum of motivation outlined by Deci and Ryan (e.g., not at all motivated to extremely motivated) may yield additional coding results of the open-ended responses. In addition, the brief nature of the open-ended prompt included in this study did not allow us to fully capture the context that informed mentors’ degree and type of motivation. This also contributed in part to our difficulty in distinguishing between identified and integrated forms of motivation in mentors’ responses. In depth-qualitative interviews with mentors to ascertain mentors’ level of exposure to cultural diversity, namely race/ethnicity, in research mentoring and their own level of racial/ethnic cultural awareness would provide more context to interpret the results reported in this study. In our study, we had predominantly White and slightly more female early career mentors. The limited number of mentors from HU racial and ethnic groups made it difficult to further examine the differences in motivation to address race and ethnicity among mentors from different racial and ethnic groups. Our descriptive findings are an important first step, but a sample with more mentors from HU groups would permit researchers to better understand how the racial and ethnic identity of mentors factors into motivation to broach the topic of race and ethnicity with trainees. This would also allow researchers to determine whether our findings can be replicated with a more diverse sample of mentors. Finally, we acknowledge that the social identities and positionality of the individual coders as well as the collective positionality of the coding team inevitably influenced the interpretation of mentors’ responses. We encourage scholars interested in furthering this area of research to build awareness-raising and reflection into their training of and discussions among coders so that power, privilege, and bias can be named early in the coding process; and to foster an open and productive dialogue among team members. In doing so, researchers can “walk the talk” of what it means to address issues of race/ethnicity in the context of a research group in a way that enriches both the individuals and the research itself.

Future Directions

This study is one of the first to investigate the motivation of research mentors to address the role that race and ethnicity can play in the mentoring relationship and in their trainee’s experiences as a researcher. Through our qualitative inquiry into research mentors’ relevant motivations, we addressed the recommendations of researchers who have promoted qualitative methodologies as a means of obtaining a deeper understanding of the rationale and strategies mentors use to approach or avoid the topic of race/ethnicity (Griffin et al. 2010). We believe that the findings presented in this paper can assist mentors in identifying their approach to addressing race and ethnicity in the research mentoring relationship and determining whether that approach is conducive to broadening participation in STEM. The results of this study revealed different types of motivation that compel mentors to approach or avoid the topics of race and demonstrated that this area of research investigating the role that motivation plays in successful culturally-responsive mentoring relationships needs further study. By identifying what motivates participants to address or avoid the topic of race/ethnicity in research mentoring relationships, we have taken an important step towards understanding the various stances mentors take with their trainees from HU groups. In our own work, we intend to explore how mentors’ motivation to engage in specific culturally-responsive practices can impact the mentoring relationship as well as a trainee’s research experience. Based on the mixed-methods findings presented in this study, a quantitative measure of motivation to address the role of race and ethnicity in research mentoring relationships could be developed and validated. Such a measure would help researchers to better understand the motivational profiles of mentors in STEM and the effect mentors’ motivation to address race and ethnicity can have on the beliefs and outcomes of HU trainees. Understanding the trainee’s motivation to engage in similar behaviors could also provide insight into how motivational dispositions might differ across identity and training/career stage. A study examining differences in motivation to address cultural identity in mentor/trainee pairs could be particularly useful in learning how a match (or mismatch) in degree and type of motivation contributes to the quality and outcomes of research mentoring relationships.

Conclusion

The results from this study provide evidence that more work is needed to convince some mentors of the value of culturally responsive mentoring (Byars-Winston, 2017) and to build the confidence and competence of STEM mentors to effectively support a more diversified workforce. As trainees from HU groups continue to face barriers to pursuing a career in science (Valantine & Collins, 2015), it is important to heed the mounting qualitative and quantitative evidence that race/ethnicity does matter in science (Byars-Winston, Leverett et al. 2017, Carlone & Johnson, 2007 Estrada-Hollenbeck et al., 2011; Hurtado et al. 2009; Johnson et al. 2011; Chemers et al. 2011) and in the research mentoring relationship.