Inspiration, inoculation, and introductions are all critical to successful mentorship for undergraduate women pursuing geoscience careers

Persistent underrepresentation and marginalization of women and other groups in the geosciences remains despite decades of effort. Prior work identifies mentoring and role modeling as promising mechanisms to improve recruitment, retention, and persistence in STEM, yet multi-component programs can be resource-intensive and difficult to scale, and the unique contribution of each component is not well isolated. The authors build on the PROGRESS program, which integrates exposure to diverse female geoscience role models, training on coping with bias and building mentoring networks, and follow-up mentoring. This study asks: Which components—Inspiration (exposure to role models), Inoculation (skills/training to grow networks and cope with bias), and Introduction (direct introduction to a local female geoscientist)—are necessary and sufficient to increase undergraduate women’s role modeling and mentoring experiences, strengthen science identity and coping skills, and ultimately enhance intentions to persist in geoscience careers? Addressing this question is important for designing scalable, transferable, and effective mentorship interventions that can broaden participation in geoscience.

The paper synthesizes research showing that role models and mentors can enhance students’ identification with science, sense of belonging, and persistence, especially for women and other underrepresented groups in male-dominated fields. Theory and evidence support benefits of multiple-mentor networks over reliance on a single mentor, including improved academic and career outcomes. Role modeling interventions often focus on exposure to inspirational biographies or panels without facilitating sustained relationships, while mentoring studies emphasize tangible relational support without isolating the unique benefits of role model inspiration. The stereotype inoculation model and related social identity frameworks suggest that exposure to counter-stereotypical exemplars and training on coping with bias can protect against negative stereotypes. However, prior multi-component programs like PROGRESS have been difficult to decompose, limiting clarity on which elements drive outcomes and how to scale them efficiently. This study aims to distinguish the effects of inspiration, inoculation, and introductions, and to test whether building multiple-mentor networks is a key pathway to enhanced science identity, coping skills, and persistence intentions.

Design: Randomized experiment with blocked randomization within each participating university. Participants were assigned to one of three workshop conditions that varied in resource intensity and content: (1) Inspiration: exposure to geoscience careers via interactive panels with diverse female geoscience role models; (2) Inspiration & Inoculation: Inspiration plus training on developing and maintaining mentoring relationships and inoculation content addressing implicit bias, stereotype threat, and coping strategies; (3) Inspiration, Inoculation, & Introduction: the prior two components plus an email-facilitated introduction to a local female geoscience mentor for a low-intensity, initial meeting (e.g., 1 hour).

Participants and setting: Undergraduate women (≥18 years) currently majoring or intending to major in geoscience-related fields from 10 U.S. universities spanning public research-intensive PWIs, two large minority-serving institutions (including an HSI), one research-intensive HBCU, a mid-sized teaching-intensive PWI, and a small private teaching-intensive PWI. Workshops were 1-day events held locally in spring 2019. Transportation was provided. Of 187 attendees, analyses focus on N=158 who completed both pre- and post-workshop surveys (n=85 attended a complete 1-day workshop session).

Procedures and intervention content: All conditions included modules introducing Earth systems and environmental sciences, panels on career pathways featuring diverse female geoscientists, sessions on mentoring expectations, career goals, resilience toolkits, and professional correspondence with mentors. The inoculation content included interactive discussion and exercises illustrating implicit bias and stereotype threat (e.g., a mock Implicit Association Test) and practical coping strategies. The Introduction condition paired students with local potential mentors based on shared interests via survey, with expectations set for at least one meeting.

Measures (pre- and post-):

• Role models: Number of female STEM career role models identified (0–3), subsequently dichotomized for analyses (one vs none; multiple [≥2] vs fewer).
• Mentor network size: Number of mentors (0–3), dichotomized (one vs none; multiple [≥2] vs fewer).
• Quality of mentorship support: Aggregated ratings across mentors on support, encouragement, career development, and network-related mentoring (higher scores indicate higher support).
• Science identity: Multi-item scale (1–7) assessing identification with being a scientist (e.g., “I have come to think of myself as a scientist”).
• Coping skills: Subscales for active coping, emotional support seeking, and informational support seeking, adapted from the Brief COPE.
• Persistence intentions: Three-item scale (1–7) measuring intentions to pursue scientific graduate study and careers; averaged to form a composite.

Analysis plan: Path analysis within a structural equation modeling framework (Mplus v8) with maximum likelihood estimation and robust bootstrapped standard errors to estimate direct and indirect effects of workshop condition on outcomes via role modeling and mentoring variables, controlling for pretest levels of outcomes. Role model and mentor counts were analyzed as binary indicators (one vs none, multiple vs fewer) to facilitate mediation testing. Model fit assessed via χ², CFI, RMSEA, and SRMR, with conventional cutoffs for good fit. Indirect effects were evaluated with bias-corrected bootstrapped 95% confidence intervals. All control variables were centered.

• Descriptive changes: Women in the Inspiration group showed significant gains in identifying multiple female STEM career role models from pretest to posttest (+2.72; Wilcoxon Signed Rank Test Z = 2.68, p = 0.007). The Inspiration & Inoculation workshop did not yield significant pre-to-post changes in role model identification (Z = 0.47, p = 0.64; Z = 0.33, p = 0.74 for examined contrasts).
• Model fit: The path model fit the data well, χ²(46) = 83.05, p = 0.557; CFI = 1.00; RMSEA = 0.00 (90% CI [0.00, 0.04]); SRMR = 0.066.
• Role models vs mentoring: Compared to the Inspiration & Inoculation group, participants in the Inspiration group were more likely to identify multiple female STEM career role models. However, participants in the Inspiration, Inoculation, & Introduction group were significantly more likely to report having multiple mentors than those in the Inspiration group. The quality of mentorship support did not differ across groups.
• Mediators to motivation and coping: Having multiple mentors had positive direct effects on posttest science identity, active coping, emotional support seeking, and informational support seeking, controlling for pretest.
• Persistence intentions: Only posttest science identity had a significant positive direct effect on posttest scientific career persistence intentions, controlling for pretest intentions.
• Indirect effects: Being in the Inspiration, Inoculation, & Introduction group indirectly increased science identity, strengthened coping skills, and ultimately increased intentions to persist in science through the development of multiple-mentor networks, relative to the Inspiration group.
• Suppression effect: A negative direct effect of identifying multiple female role models on posttest science identity emerged but was interpreted as a statistical suppression artifact, as role model identification was uncorrelated with science identity but positively correlated with having multiple mentors.

The study disentangles the distinct roles of inspiration (exposure to female role models), inoculation (skills and strategies to build networks and cope with stereotypes), and introductions (facilitated mentor connections). A workshop emphasizing inspiration alone effectively helps undergraduates identify multiple female role models but does not, by itself, expand their mentoring networks. The addition of inoculation content plus a direct introduction to a local mentor is necessary to catalyze growth in multiple mentoring relationships. Multiple-mentor networks, in turn, bolster science identity and coping skills (active, emotional, informational), with science identity emerging as the primary proximal predictor of intentions to persist in scientific careers. These findings clarify mechanisms suggested by mentoring theory: inspiration primes identification and belonging, inoculation equips students to navigate barriers, and introductions translate readiness into actual network expansion. By isolating components, the research shows that a low-intensity, scalable combination—especially including introductions—can reproduce outcomes usually attributed to more intensive, multi-component programs.

This randomized experiment identifies the critical combination of elements for effective mentorship programs aimed at undergraduate women in the geosciences. Inspiration through exposure to diverse female role models is necessary but insufficient for network growth; adding inoculation content and a facilitated introduction to a local mentor is essential to increase multiple-mentor networks, which enhance science identity and coping skills and, ultimately, intentions to persist. The study demonstrates that benefits of more intensive programs can be replicated in a compressed 1-day workshop with follow-up mentoring opportunities and simple, email-facilitated introductions. These components are readily adoptable by departments and professional societies seeking scalable diversity and inclusion interventions. Future research should examine generalization across intersecting identities, test mentor gender matching effects, measure role models and mentors without caps to allow continuous modeling, and evaluate long-term behavioral outcomes (e.g., sustained mentoring relationships, degree completion, and career entry) beyond intentions.

• Sample composition and generalizability: Despite inclusive recruitment across 10 institutions and holding half the workshops at a minority-serving institution, most participants were White, cisgender women, limiting tests of intersectional effects and generalizability to broader underrepresented groups.
• Geographic and institutional scope: Participants were from two U.S. regions and 4-year institutions; results may not generalize to other regions or 2-year institutions.
• Measurement constraints: Counts of role models and mentors were capped at three and then dichotomized for analysis, potentially reducing variability and obscuring dose–response relationships; future work should use uncapped, continuous measures.
• Intervention uptake: Even with facilitated introductions, the overall rate of mentorship formation was relatively low, and many underrepresented women were hesitant to initially engage with mentors.
• Mentor matching: The study only introduced students to female mentors; it remains unknown whether similar benefits would occur with male mentors or mixed-gender networks.
• Design scope: The study assessed short-term post-workshop outcomes; long-term behavioral persistence was not measured within this experiment.
• Potential suppression effects: A negative direct effect of role model identification on science identity likely reflects statistical suppression, indicating caution in interpreting specific direct paths.