1Introduction: Why Does Mentoring Matter?

The Committee's Starting Point

The committee made four presumptions based on published findings about advancing effective mentoring in STEMM. First, the committee presumed that all mentorship at its core is a reciprocal and responsive social exchange among a specified group of people with diverse individual attributes (Eby et al., 2007). Accordingly, throughout the report the committee emphasizes the relational nature of mentorship that includes contributions from both mentors and mentees involved in this social exchange. By focusing on the relationship, rather than solely on the mentee or the mentor, the committee calls out a paradigm shift in how both student agency and the burden of mentoring expectations for the participants are viewed. In general, students have a range of talents, strengths, and assets, all of which an effective mentoring relationship can capitalize on to facilitate their successful pursuit of and persistence in STEMM career pathways. Students are not merely passive recipients in mentoring but have potential agency that can contribute to effective mentoring relationships. Moreover, the burden of conforming to the expectations of a mentoring relationship does not fall only on mentees, but is developed through the establishment of a mutually agreed-upon relationship. This topic is explored further in Chapter 2.

Second, the committee presumed that promoting diversity without inclusion is not enough to create a diverse STEMM workforce (Asai, 2019; Puritty et al., 2017).¹² While promoting diversity is necessary to increase the numbers of students in STEMM from underrepresented groups, it is equally necessary to equip mentors with the skills to recognize and respond to the identity-based experiences of these students that affect their academic and career development.¹³ Therefore, the committee emphasized the importance of inclusive mentoring practices in all contexts as central to effective mentorship of all students. These topics are explored further in Chapter 3.

Third, the committee presumed that mentoring experiences occur within organizational contexts (Allen et al., 2006), including institutions of higher education, training programs, disciplinary societies, and professional meetings. Therefore, to support its call for increasing effective mentorship, the committee drew on the extensive evidence from organizational behavior research documenting how organizational structures, such as reward and leadership systems (Kerr, 1995; Pinder, 2014), and organizational culture and climate can influence the behaviors that are displayed and valued (Ostroff et al., 2013), as well as influence the way behavioral change occurs. These topics provide a focus for Chapter 7.

Finally, the committee presumed that the understanding of interpersonal interactions that has been developed through social science research can be applied to the development of effective mentoring relationships. Therefore, the committee chose six significant theories that can help frame practical questions and insights of mentorship to explore in depth in Chapter 2 and are referenced throughout the report. The committee does not expect any given reader to become an expert in the highlighted theories, but rather provides them as a primer and reference for consideration when developing a mentoring relationship or program, for inspiration when facing a potentially awkward or new situation in mentorship, or for reflection when engaging in further mentorship competency development.

The Work of the Committee

To inform the committee's deliberations, it convened 3 public workshops and held 18 listening sessions. These evidence-gathering activities were intended to supplement the intensive literature review. The committee incorporated the output of workshops and the listening sessions throughout the report and in the online guide.

The first workshop, hosted in Washington, D.C., on April 11–12, 2018, explored new knowledge, ideas, and practices in inclusive mentorship excellence and highlighted several questions posed by both practitioners and scholars of mentorship. Based on these discussions, the committee commissioned three literature reviews by outside experts: Mentoring Underrepresented Students in STEMM: A Survey and Discussion (McGee, 2018); Mentoring beyond Hierarchies: Multi-Mentor Systems and Models (Montgomery and Page, 2018); and Landscape of Assessments of Mentoring Relationship Processes in Postsecondary STEMM Contexts: A Synthesis of Validity Evidence from Mentee, Mentor, Institutional/Programmatic Perspectives (Hernandez, 2018).

The second workshop, on October 8, 2018, in Irvine, California, examined the preliminary findings of the three commissioned papers. The authors of the papers were invited to present their work, and the participants engaged the presentations from scholarly as well as lived experience perspectives. The papers were then revised based on the feedback and questions that were raised. The committee built upon the contributions of these papers throughout the report, in particular, Chapter 3 (McGee, 2018), Chapters 4 and 5 (Montgomery and Page, 2018), and Chapter 6 (Hernandez, 2018). All three papers are also produced in full as supplemental materials online.

The third and final workshop was held at Vanderbilt University on February 5, 2019. Scholars, mentors, and mentees gathered to imagine how to realize an evidence-based, online resource guide on mentorship; clarify the purpose and scope of the online resource; identify users and use cases; and define the desirable functionalities of the final product.

The 18 listening sessions were held over the course of 8 months on university campuses and at a variety of professional society meetings to gain insights on the experiences and concerns about mentorship from current students, faculty members, and administrators.¹⁴ Each session was designed to informally solicit input for the report as well as how the online guide might be engaged. For most of the sessions, after a brief overview of the committee's work, the participants were guided through three activities by one to three members of the committee: (1) they wrote down questions or ideas about theory, research, and practice of mentorship; (2) they discussed their lived mentoring experiences in small groups, focusing on what they had found to be useful for effective mentoring relationships; and (3) they described characteristics, features, and content that might be useful for the online guide. The experiences and suggestions from the participants of sessions informed the committee's approach to the materials presented, as well as the content of Chapter 7.

Sources of Evidence

The committee's task was to examine the evidence supporting effective mentoring programs and to identify the characteristics that make for an effective mentoring relationship. However, different kinds of evidence (e.g., qualitative, quantitative, narrative, anecdotal observation) differentially inform how researchers approach their studies and how practitioners (i.e., mentors and mentees) engage in their mentoring relationships.

In many cases, the practice of mentoring occurs without drawing upon theories, empirical studies, and other forms of evaluation to support such practices. Furthermore, isolating the evidence to support the effect of mentoring can prove difficult. For example, many programs incorporate mentoring, but studies of these programs have not been and may not be able to systematically evaluate specific mentoring elements apart from other program activities. There is still an absence of intentionality, both in designing programs for particular mentor and mentee functions in contexts and in evaluating these components specifically (Lunsford, 2016; Pfund, 2016).¹⁵

Different researchers and stakeholders vary in their opinions regarding the forms of evidence that are appropriate and informative when understanding mentorship. Some scholars of mentorship adhere only to evidence in the form of data that have been collected and analyzed using validated and reliable measures; other scholars observe that many deeply human and social issues require sharing of personal narratives, lived experiences, and wisdom of practice that does not adhere to the same empirical standards. The committee concluded that any well-rounded study of a complex subject such as effective mentorship would involve a range of perspectives and methods.

To prepare this report, the committee relied primarily on reports in the published literature, from both within and outside of STEMM. In addition, grey literature that focused on programs or experiences, as well as opinion pieces and editorials, were referenced to predominantly help frame issues, though not to inform conclusions or recommendations.¹⁶ The use of both opinion pieces and editorials and the grey literature was limited to domains where peer-reviewed resources and publications were not yet available to address emerging practices. For example, the grey literature was consulted in the domain of online peer mentoring network communities for scholars of color.

Integration of Mentoring Scholarship from Other Settings

Although the committee focused on mentorship in academic settings, the committee utilized work drawn from outside of academic contexts as it pertained to broader findings about mentorship in general that apply within and beyond academic settings. The committee considered academic settings as workplaces, and faculty members and researchers engaging in mentoring as employees, recognizing that organizational systems that facilitate or discourage mentoring in workplace settings in general are also applicable to workplaces of higher education or other research settings. In particular, the committee considered scholarship conducted in workplace settings outside of academia pertinent because reward structures and employee motivation in academic workplace settings contain significant parallels to other workplace settings.

A large body of literature exists on workplace mentoring, and relevant scholarship was pulled from disciplines such as organizational behavior and human resource management. Mentorship has been a key component of research on employee development in organizations ever since seminal work in 1983 posited that mentoring relationships between more-senior-level and less-senior-level employees were critical to shaping employee career development, career progression, and well-being (Kram, 1983). This early research on mentoring was included in particular because it has profoundly informed the definitions of mentoring functions, as well as virtually all models, since its publication. Since then, an expansive amount of research has contributed to an in-depth understanding of mentoring relationships in organizations, including its antecedents and consequences for both mentors and mentees (Eby et al., 2013). Research on the effects of mentorship on organizational behavior has had a strong focus on the use of quantitatively oriented, oftentimes survey-based research studies, with a particular emphasis on data collection from multiple sources and the use of objective outcome data, such as job performance (Allen et al., 2006).

Studies that have examined mentorship across contexts provide nuanced answers on the generalizability of findings to educational settings (Eby and Allen, 2008; Eby et al., 2013; Mijares et al., 2013). In general, the relationships between mentorship and key behavioral outcomes, such as performance and withdrawal behaviors, hold across sectors, but the magnitude of effect varies by domain, with mentorship in workplace and academic settings frequently being linked to stronger effects than with youth mentoring.

A comprehensive review of all mentoring across all disciplinary contexts, however, is beyond the scope of this report. The committee reviewed mentoring literature only in domains determined to be highly relevant for mentoring in academic settings, but refrained from discipline-specific mentoring work, unless it provided theoretical advances that could be generalized beyond the disciplinary or professional/educational context.

The Importance of Diversity in STEMM

Over the course of the 20th century, the United States became a global leader in science, technology, and medicine in large part because of the innovations and scientific breakthroughs produced by its highly educated and productive cadre of STEMM professionals. Maintaining that leadership role, and the resulting economic and political benefits that come with it, requires continually renewing and diversifying the talent of the U.S. STEMM enterprise. This is particularly critical today because research suggests that there are real costs to perpetuating a homogenous and White, male-dominated STEMM workforce. For example, car accidents were more likely to be deadly for women because the crash test dummies the manufacturers used in safety trials were designed to protect only an average male body type (Gendered Innovations, 2019; Kahane, 2013), and facial recognition software failed to accurately detect the faces of Black or Asian people (Klare et al., 2012; Phillips et al., 2011; Raji and Buolamwini, 2019). The faulty designs of these technologies might have been avoided if the people researching, designing, and developing the innovations better represented the diversity of the population.

Furthermore, the nation's STEMM competence, productivity, and scientific progress today relies more than ever on collaborative problem solving (Jones et al., 2008; Plume and van Weijen, 2014; Wuchty et al., 2007), and teams with diverse experiences and areas of expertise often ask different questions and tend to be more creative and innovative in how they answer those questions (Gibbs, 2014; Leung et al., 2008; Miller and Del Carmen Triana, 2009; Page, 2008; Torchia et al., 2011). More diverse research teams also, on average, produce higher-impact research (Freeman and Huang, 2014a, 2014b) and result in more effective and efficient problem solving (Higgs et al., 2005; Hong and Page, 2004; Woolley et al., 2010). In business and academia, greater diversity has been associated with improved financial performance (Carter and Wagner, 2011; Cook and Glass, 2011), reduced conflict in the workplace (Nielsen and Huse, 2010), promotion of a more equitable work environment (Flabbi et al., 2016; Pichler et al., 2018; Terjesen and Singh, 2008), and lower employee turnover (Kaplan et al., 2011; McKay et al., 2007).

In addition, the number of STEM jobs requiring a college degree or higher—93 percent of which pay better than the national average wage (Fayer et al., 2017)—is projected to grow faster than the U.S. economy as a whole. This growth will in turn create employment opportunities for Americans that must be available to and draw upon talent regardless of gender, ethnicity, sexual orientation, race, religion, family educational background, or cultures, many of whom are currently underrepresented in the current STEM workforce.

Even with understanding the benefits of diversity in STEMM, it is challenging to develop and retain the diverse talent required for the workforce. A 2012 report from the President's Council of Advisors on Science and Technology pointed out that retaining individuals from varied backgrounds in STEM fields in college and graduate school is the least expensive and fastest way to increase the number of STEM professionals needed to ensure that the nation maintains its competitive edge in the global economy (PCAST, 2012). Yet according to a 2013 study from the U.S. Department of Education's National Center for Education Statistics, 48 percent of bachelor's degree students who entered STEM fields between 2003 and 2009 had left those fields by spring 2009 (Chen and Soldner, 2013). Meanwhile, over 50 percent of those students who do complete a STEM bachelor's degree switch to jobs or graduate programs outside of STEM (National Science Board, 2018). Studies have also shown that students from UR populations in STEM—including women; students from racial and ethnic groupings such as Blacks/African Americans, Latinx, and American Indians/Alaska natives; FG students; and those from lower socioeconomic backgrounds—are more likely to move on to fields outside of STEM than students from well-represented groups (Anderson and Kim, 2006; Griffith, 2010; Hill et al., 2010; Huang et al., 2000; Kokkelenberg and Sinha, 2010; Shaw and Barbuti, 2010).

The Importance of Mentorship in Supporting Diversity

The 2013 National Center for Education Statistics study and others have identified several factors that lead undergraduate and graduate students to leave STEM fields. For example, student motivation, confidence, and beliefs about one's capacity to learn STEM subjects and be a STEM professional correlate to the high rate of attrition from STEM fields (Burtner, 2005; Huang et al., 2000; Seymour and Hewitt, 1997; Sithole et al., 2017). In the same manner, a mentee's experiences or perceptions of the institutional and workplace environments have also been linked to STEM attrition (Eagan et al., 2011).

In addition, individuals from UR groups may experience exclusionary practices in STEMM fields. Individuals in situations in which there are members of multiple social groups—the STEMM classroom or research laboratory, for example—are motivated to increase or maintain how positive and distinct their group is relative to other groups (Tajfel and Turner, 1986). This is especially true when the integrity of a social identity is threatened, such as when the composition of one's group is becoming more diverse.¹⁷ The result can be bias and discrimination, in which group boundaries are distinguished and social groups are rank-ordered and assigned differential social value (Brewer, 1979; Chen and Li, 2009).¹⁸

While individuals' motivations cannot be known, evidence on the outcomes of mentoring indicates that effective mentoring relationships can improve outcomes both for individuals' career development and for their productivity, while for institutions, effective mentorship can lead to more effective placement of graduates in the job market.¹⁹ Lamar Smith of Texas, chair of the U.S. House of Representatives Committee on Science, Space, and Technology, in the Innovations in Mentoring, Training, and Apprenticeships Act of 2018, acknowledged that “[o]utcomes show that mentoring can increase STEM student engagement and the rate of completion of STEM postsecondary degrees” (H.R. 5509, 115th Cong. [2018]). Given the economic benefits of well-trained STEMM professionals entering the workforce at a higher rate, effective mentoring can result in significant benefits to individuals, institutions, and society at large. Conversely, the economic consequences of limited access to effective mentorship may result in a lower number of STEMM professionals and more limited growth opportunities for companies and research organizations in need of STEMM workers.

Mentorship can help with workforce development by increasing access, equity, and inclusion in STEMM. While researchers and pundits alike have proposed many hypotheses to explain the underrepresentation of segments of the U.S. population in STEMM, one that has not been discussed frequently until recently is the lack of effective mentorship for those students. As the authors of a recent commentary stated, “the acknowledgment of the role and agency of the faculty in the professional development of their students is a promising step forward” (Poodry and Asai, 2018). For example, mentored students pursue graduate study more frequently than students without formal mentoring support and are more likely to be retained in STEMM (Campbell and Campbell, 2007). Indeed, effective, high-quality, and sustainable mentoring relationships for individuals across career stages can play a critical role in supporting mentee success in STEMM fields (Aikens et al., 2017; Byars-Winston et al., 2015; Dennehy and Dasgupta, 2017; Haeger and Fresquez, 2016; Lisberg and Woods, 2018; Muller et al., 2012; Pfund, 2016; Smith and George, 2019).

Not all mentoring experiences are positive, however. There are several types of negative mentoring experiences that can also affect the development of the next generation of STEMM professionals.²⁰ Negative mentoring experiences can arise from both good and ill intent, and there are some—such as abusive supervision and harassment—that qualify as detrimental research practices (NASEM, 2018d). These practices have been shown to affect not only the individuals involved, but the greater STEMM enterprise. Inadequate mentoring, advising, and career counseling has been linked to STEM attrition, particularly for UR students (Dupey et al., 2006; Sithole et al., 2017). Further research on the prevalence and impact of negative mentoring experiences may help to elucidate the most effective mechanisms for mitigating negative mentoring experiences.

Supporting Effective Mentorship to Develop Diverse Talent

Institutions can and have been supporting effective mentorship and mitigating negative mentoring experiences by developing cultures of inclusive excellence.²¹ The American Association of Colleges and Universities describes inclusive excellence as a “guiding principle for access, student success, and high-quality learning.” Creating a culture of inclusive excellence requires academic institutions to identify where student success across demographic groups is unequal, discover which of their educational practices succeed in addressing those inequities, and work intentionally to build off of those practices in a way that sustains institutional change (Williams et al., 2005). Creating a culture of inclusive excellence must also include providing access to effective mentoring for all students.

A growing body of evidence exists about how to create and sustain successful, inclusive mentoring relationships that research identifies as being important for mentors and mentees. In fact, the literature in both STEMM and non-STEMM postsecondary education and outside of postsecondary education warranted a critical review for the key evidence to guide mentoring initiatives in colleges, universities, and other educational and research settings. While most studies show the small-to-medium effect sizes for any given mentoring intervention,²² the cumulative effect can be transformative for any particular individual or institution. It is clear that mentorship is one of the pillars of effective practice in developing the diverse generation of STEMM professionals currently enrolled in undergraduate and graduate programs, especially when paired with other continuous improvement interventions, such as high-quality in-classroom pedagogy, and utilization of evidence-based teaching and training practices (Cohen, 1988, 1995; Plonsky and Oswald, 2014).