Access to this full-text is provided by Springer Nature.
Content available from Nature Communications
This content is subject to copyright. Terms and conditions apply.
Article https://doi.org/10.1038/s41467-022-34508-x
Female peer mentors early in college have
lasting positive impacts on female engineer-
ing students that persist beyond graduation
Deborah J. Wu
1
,KelseyC.Thiem
2
&NilanjanaDasgupta
3
Expanding the talent pipeline of students from underrepresented backgrounds
in STEM has been a priority in the United States for decades. However, potential
solutions to increase the number of such students in STEM academic pathways,
measured using longitudinal randomized controlled trials in real-world con-
texts, have been limited. Here, we expand on an earlier investigation that
reported results from a longitudinal field experiment in which undergraduate
female students (N= 150) interested in engineering at college entry were ran-
domly assigned a female peer mentor in engineering, a male peer mentor in
engineering, or not assigned a mentor for their first year of college. While an
earlier article presented findings from participants’first two years of college,
the current article reports the same participants’academic experiences for
each year in college through college graduation and one year post-graduation.
Compared to the male peer mentor and no mentor condition, having a female
peer mentor was associated with a significant improvement in participants’
psychological experiences in engineering, aspirations to pursue postgraduate
engineering degrees, and emotional well-being. It was also associated with
participants’success in securing engineering internships and retention in STEM
majors through college graduation. In sum, a low-cost, short peer mentoring
intervention demonstrates benefits in promoting female students’success in
engineering from college entry, through one-year post-graduation.
Student engagement, success, persistence, and career pursuit in
science, technology, engineering, and mathematics (STEM) is a
high priority in the United States, particularly given the demand for a
skilled STEM workforce and the short supply of college graduates
with skills and interest in STEM careers1,2. The shortage of skilled
STEM workers, especially in engineering, is exacerbated by the
underrepresentation of females and racial ethnic minorities3,4.Sex
and race gaps in STEM participation may also exacerbate income
inequality between the sexes, and between racial minorities and
white people, given relatively higher salaries in STEM professions
compared to many other professions5,6.
A growing body of research has been using randomized con-
trolled trials to test whether social psychological interventions can
reduce group-based disparities in students’STEM participation and
performance. Many of these interventions train individuals to cogni-
tively reappraise their experience, with an emphasis on teaching
underrepresented individuals to mentally adapt to their environment
(we call these mental reappraisal interventions)7–16. Using reappraisal
strategies such as self-affirmation, growth mindset, and emotion reg-
ulation have been successful in reducing sex, race, andclass disparities
in academic performance7–16. For instance, brief affirmation interven-
tions in which students reflect on personally held values have been
Received: 15 June 2021
Accepted: 26 October 2022
Check for updates
1
Department of Psychology, Northwestern University, Evanston, IL 60201, USA.
2
Department of Counseling Psychology, Social Psychology, and Counseling,
Ball State University, Muncie, IN 47306, USA.
3
Department of Psychological and Brain Sciences, University of Massachusetts Amherst, Amherst, MA 01003,
USA. e-mail: nd@umass.edu
Nature Communications | (2022) 13:6837 1
1234567890():,;
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
shown to decrease sex and race gaps in academic performancein both
lab and field settings7–12. Prompting students to cultivate a growth
mindset (i.e.,viewing intelligence or performance as changeable rather
than a fixed ability) improves grades among previously low-
performing students13–15. Reappraising anxiety as a positive mental
state also decreases the achievement gap in grades between high and
low-income students16. These interventions make important con-
tributions toward academic equity, but theyalso have limitations. First,
by focusing on mental reappraisal levers of change, these studies place
the responsibility on underrepresented individuals to adjust to aca-
demic institutions by changing their mindset rather than placing the
responsibility on academic institutions to change learning environ-
ments and meet the needs of diverse students. Second, several of these
studies focus on grades and test performance as measures of interest
but do not assess students’subjective experiences in academic spaces
(e.g., feelings of belonging, confidence, anxiety, motivation)8,10,12–16.
Emphasis on the subjective experiences of students from under-
represented backgrounds is critical because this may affect persis-
tence and interest even in the absence of objective performance
gaps17,18. Third, much of the research investigating mental reappraisal
interventions did not target STEM education specifically, but instead
focused on general academic performance (cumulative grade point
average), and many of these studies were conducted with adolescents
and not college students7–16.
Another body of research shineslight on the impacts of increasing
the visibility of successful own-group role models for under-
represented students in STEM. This research testing the power of role
models in STEM is consistent with the Stereotype Inoculation Model,
which posits thatexposure to successful own-group experts and peers
protects one’s mind against noxious stereotypes that cast doubt on
one’sability
18. This research shows that reading about, or interacting
with, successful own-group role models closes sex and race gaps by
increasing test performance and course grades for females and racial
ethnic minorities. Moreover, exposure to these role models also bol-
stersfeelingsofconfidence, belonging, and STEM identification17,19–25.
Collectively, this work points to evidence-based solutions that pro-
mote equity in STEM education, both in terms of objective metrics
(such as test performance and grades) and subjective metrics (con-
fidence, motivation, and belonging), thereby amplifying the impor-
tance of visible representation of underrepresented groups in STEM
environments.
Yet this research also has limitations, as these studies typically
assess how role models impact student outcomes at a single time-
point or over a short period of time (e.g., up to two weeks after
interacting with the role model), providing insufficient information
about long-term impacts17,19–25 Moreover, the majority of these studies
were laboratory experiments conducted in artificial settings rather
than naturally existing environments19–24, which means participants did
not have opportunities to form what could be considered authentic
mentoring relationships with role models. In our view, authentic long-
term mentoring relationships differ from brief exposure to, or fleeting
interactions, with role models. Authentic mentoring relationships
allow mentees to seek advice from role models, get assistance when
they struggle, and expand their academic and professional networks
through their role model or mentor’s contacts. Indeed, past research
shows that when students have authentic mentoring relationships with
faculty mentors in STEM, their academic outcomes are enhanced. For
example, engineering students who have authentic mentoring rela-
tionships with faculty members (in which faculty members provide
personalized attention and mentoring to the student), are more likely
to persist and perform well in engineering classes26.Blackstudentsat
historically Black universities are more likely to work directly with
faculty when they struggle academically27 and report higher well-
being28,29 as compared to Black students at predominantly white uni-
versities who receive less one-on-one attention from faculty30.
Collectively, extant research suggests that authentic mentoring rela-
tionships between students and mentors who share their mentees’
marginalized identity are likely to be more impactful compared to
brief exposure to such individuals29,31.
Using the Stereotype Inoculation Model18 as our theoretical fra-
mework, the current research predicts that opportunities to form
authentic mentoring relationships with role models who share one’s
marginalized identity will have substantial benefits for mentees by
reinforcing their confidence, motivation, and abilities over time. Using
a randomized controlled trial, we examine whether and how peer
mentoring relationships in engineering in the first year of college
influence female engineering students’subjective experiences and
objective academic outcomes from college entry through graduation
and up to one year post-graduation.
This current study expands upon our previously published work
(Dennehy & Dasgupta, 201732) which examined the impact of peer
mentorship on mentees’experiences in the first year of college when
mentoring was active. This earlier investigation found that being ran-
domly assigned a female peer mentor (as compared to a male mentor
or no mentor) produced academic benefits for female students in the
first year of college. Compared to their baseline measures (i.e., before
classes began), female students assigned to a male mentor or no
mentor increased in their feelings of anxiety relative to motivation in
their engineering classes, and declined in their belonging, confidence
in engineering, and aspirations to pursue post-graduate degrees in
engineering during their first year of college. In contrast, participants
assigned to female mentors showed no change in anxiety relative to
motivation, and stable belonging, confidence, and post-college
aspirations in engineering. These benefitsenduredforasecondyear
for the 52%of the sample (78 out of 150 students) that had completed
their second year of college at the time of our analysis. The remaining
48% of the sample had not completed their second year when the
original paper was published.
The present research makes three contributions beyond the ori-
ginal2017study.First,byfollowingthefullsampleoffemalepartici-
pants for 3–5 years beyond the intervention year through Bachelor’s
degree completion plus one year post-graduation, we test if the impact
of peer mentoring endures beyond Bachelor’s degree attainment long
after mentoring has ended. Second, the present research measures
additional variables that capture work experience (engineering
internships) and emotional well-being. We expect that emotional well-
being may be a key protection against attrition from STEM given past
research showing that minority status and social identity threat pre-
dicts daily experiences of emotional exhaustion and psychological
burnout among females in STEM33,34. Third, the present study provides
a more accurate measure of STEM retention than the 2017 paper, in
which retention was measured at the end of the first year of college,
well before the university deadline to make final decisions about aca-
demic majors. At most American universities, students are required to
declare their major by the end of the second year of college and they
often switch majors between the first and second year of college. By
using official university transcripts to measure students’majors at
college graduation, the current article provides a more accurate pic-
ture of how first-year peer mentorship impacts students’final deci-
sions about academic majors, which has lasting impacts on their post-
graduate career trajectory.
Our focus is on females in engineering in particular because
compared to many other STEM majors, females comprise a smaller
minority in engineering degree programs (21%)35 and the workforce
(13%)36, despite being 51% of the American population37 and 60% of
college students at American universities38. Consistent with the Ste-
reotype Inoculation Model18 and priorresearch17,19–25, we predicted that
relationships with female peer mentors in the first year of college
would act as social vaccines for young female students entering
engineering, helping them disregard negative stereotypes, preserve
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 2
Content courtesy of Springer Nature, terms of use apply. Rights reserved
their motivation in engineering courses, and increase persistence and
professional development in STEM. Importantly, we predicted these
benefits would endure through one year post-graduation, 3 to 5 years
after the conclusion of the peer mentorship intervention.
The present work also extends existing research on mentorship in
several ways. First, although mentorship is a popular intervention com-
monly discussed in academia, government, and industry39–43,manypre-
vious field-based mentorship initiatives are not controlled experiments,
making it impossible to determine whether mentorship is the causal
factor responsible for downstream outcomes40. For instance, in studies
where participants self-selected into mentorship programs, the success
of mentorship interventions could be due to pre-existing individual dif-
ferences between people who self-selected into mentoring programs and
those who did not, rather than the content of the mentorship program
itself. Further, in other studies where mentees had the freedom to
choose their own mentors, the success of mentorship interventions
could be due to the special relationship between the mentor and mentee
rather than mentorship in general. To resolve these causal inference
problems, we conducted a randomized controlled field experiment32 in
which participants were unaware that the study was related to
mentoring.
Second, our field experiment complements past research on role
models. While in theory, role model interventions should create more
supportive environments for underrepresented individuals, as noted
earlier, most prior role model experiments used artificial lab settings
(e.g., reading about successful role models) and did not create
authentic relationships19–24. In contr ast, our year-long field intervention
investigated whether and how fostering authentic relationships with
peer mentors would change female students’lived experiences in
engineering, scaffold their success during a critical transition period
into college, and sustain that success through the next 3–5years,after
their mentors have left the university. Furthermore, previous inter-
ventions addressing STEM disparities assessed impacts at a single
time-point or over a relatively short period of time8,9,16,17,19–25,providing
insufficient information about long-term impacts. Although some
research has demonstrated longer-term benefits of academic inter-
ventions (i.e., 2–3 years), these studies were not specifictoSTEM
7,11.
Finally, our research examines whether mentees gain benefits
from having a mentor who does not share their social identity by
testing whether female engineering students benefit from male men-
torship. Previous research suggests that females in STEM can benefit
from effective male mentors23,24. However, because these studies were
conducted as short-term experiments using experimenters posing as
role models, it is unclear whether these findings will generalize to real-
world academic settings. Thus, we examined the long-term impacts of
male mentorship for female students in engineering.
We conducted a field-based longitudinal randomized controlled
experiment that examined female students’subjective and objective
academic outcomes and compared the impacts of own-sex and other-
sex peer mentors to a control condition. We recruited female under-
graduates who were first-year and transfer students (N= 150) at a large
public university who planned to major in engineering. Participants
were randomly assigned a female peer mentor (N= 52), a male peer
mentor (N=51),ornomentor(N=47)for1year.Peermentorswere
junior or senior undergraduate students who volunteered to be men-
tors and were trained prior to being matched with mentees.
Researchers described the study to mentors and mentees as aiming to
identify barriers and opportunities experienced by college students in
engineering. This generic description served two purposes. First, it
ensured that participants who volunteered for this experiment were
not opting in because of an interest in mentoring. Furthermore, it kept
mentors and mentees blind to the role of sex in this study (e.g., men-
tors did not know that all mentees were female; neither mentors nor
mentees knew that mentor sex was relevant to hypotheses).
Mentor-mentee dyads met several times (median = 4 meetings)
during mentees’first year in college, after which mentoring interac-
tions ended. We surveyed mentees from college entry through
Bachelor’s degree completion to one year post-graduation. Measured
variables included students’subjective experiences in engineering
(motivation in engineering courses and confidence in overall engi-
neering abilities); participation in engineering internships that provide
technical work experience; persistence in engineering and STEM
majors; aspirations to pursue post-graduate degrees in engineering;
and emotional well-being. All reported variables except for persistence
in engineering and STEM majors were self-reported in surveys, while
major-related persistence was measured using college transcripts.
The first survey was administered before mentor assignment at the
beginning of students’first year of college, serving as a baseline.
Subsequent surveys were administered at the middle and end of the
first year of college, and then once each subsequent year through
college graduation plus one year post-college graduation. College
transcripts were also obtained with student permission to get an
objective record of participants’choice of major and grades.
Results
Data management and analysis
Results are described in two sections. First, we present results showing
outcomes for which female students in engineering in our sample
experienced significantly stronger benefits from female peer mentor-
ship than male peer mentorship or no mentorship. Second, we identify
which of these outcomes serves as a psychological mechanism that
helps explain these differences.
For continuous dependent variables measured at multiple time-
points (motivation in engineering courses, confidence in overall
engineering abilities, aspirations to pursue post-graduate degrees, and
emotional well-being), we investigated how participants’responses
changed over time through college graduation and one-year post-
graduation by coding the time-point at which each response was
obtained. Time was centered at the beginning of year 1, before mentor
assignment (baseline). Responses for subsequent time-points were
scaled in reference to the month in which the survey responses were
obtained relative to the baseline. The time variable was divided by 12,
such that slope coefficients indicate yearly change.
We estimated multilevel models using Mplus 844, utilizing the full
information maximum likelihood estimator, which is the recom-
mended practice to handle missing data45,46. Each statistical model has
two levels. Level 1 represents variables measured at multiple time-
points within participants (motivation in engineering courses, con-
fidence in overall engineering abilities, post-graduate aspirations, and
emotional well-being). Level 2 represents the mentor condition—the
independent variable that was systematically manipulated between
participants. Each model tested whether mentor condition influenced
change over time on the dependent variable. In level 1 of the model,
the dependent variable was regressed on time, creating a slope
representing change over time for that variable. Random effects were
created for the intercept (participants’baseline responses at college
entry) and slope (change over time for each participant). The intercept
and slope were allowed to covary at level 2 to control for variance
shared between the actual values on the dependent variable and its
slope or change over time. At level 2, the intercept and slope of the
dependent variable were both regressed on mentor condition. Given
that mentor condition was a multi-categorical variable, two dummy
codes were created to index the female mentor condition (1 = female
mentor, 0 = male and no mentor conditions), male mentor condition
(1 = male mentor, 0 = female and no mentor conditions), with the no
mentor condition always being the reference group. We then tested
the effect of mentor condition on longitudinal change over time on
participants’experiences in two ways: (a) by examining whether the
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
trajectory of each mentor conditionsignificantly changedfrom college
entry through graduation to one-year post-graduation; and (b) by
testing if these change trajectories significantly differed
between mentor conditions. Change trajectories and the condition
differences in change trajectories are reported using unstandardized
regression values, corresponding to the change in the variable over
one year and differences in these changes between conditions,
respectively. Two-tailed significance values arereported. Furthermore,
the multilevel models can be summarized by the following equation:
Yij =β00 +β01FemaleMentor + β02MaleMentor+ β10 Time
+β11FemaleMentor*Time + β12MaleMentor*Time + r0i
+r1iTime +eti :
ð1Þ
Post hoc Monte Carlo simulation analyses with 1000 simulations
were also conducted for each effect, indicating the proportion of
simulations in which the effect was significant47. We report power
analyses for each significant effect. Dichotomous variables (participa-
tion in engineering internships, college major at graduation) were
analyzed using chi-square tests to examine percentage differences
between mentor conditions.
Assignment to a female mentor was associated with
improvement in participants’subjective experiences in
engineering, objective choices in engineering, and emotional
well-being
Whereas female participants without mentors (B=−0.13, SE =0.05,
p=0.007, power=0.92) and participants with male mentors
(B=−0.10, SE =0.05,p= 0.025, power = 0.77) showed a decline in their
overall confidence in engineering from college entry through gra-
duation and one year post-college, those with female mentors
(B=0.03, SE =0.04, p= 0.559) maintained general confidence in
engineering across time without decline. A contrast analysis revealed
that the change trajectory of the female mentor condition was sig-
nificantly different from the change trajectories of the no mentor and
male mentor conditions (B=0.14,SE =0.06, p=0.010,power=0.94),
suggesting that female peer mentorship was associated with main-
tenance of confidence among participants across college and post-
graduation. Upon comparing the trajectories of each condition sepa-
rately, we found that participants without mentors (B=0.15,SE =0.07,
p= 0.017, power= 0.81) and participants with male mentors (B=0.13,
SE =0.06, p= 0.045, power = 0.66) reported greater decline in con-
fidence versus those with female mentors. There was no difference
between the change trajectories of the no mentor versus male mentor
conditions (B=0.03, SE =0.07, p=0.660). See Fig. 1.
Whereas participants without peer mentors (B=−0.12, SE =0.04,
p= 0.006, power = 0.92) and those assigned male mentors (B=−0.14,
SE = 0.04, p=0.001, power=0.98) showed a significant decrease in
motivation in their engineering classes from college entry through
graduation to one year post-graduation, participants assigned female
mentors showed no change in motivation (B=−0.01, SE =0.04,
p= .782). The change trajectory for female participants with female
mentors was significantly different from the change trajectories in the
other two conditions (B= 0.12, SE = 0.05, p= 0.017, power = 0.91),
suggesting thatrelationships with female mentors preserved mentees’
motivation in their engineering courses. When comparing conditions
separately, the change trajectory for participants with female mentors
versus male mentors was significantly different (B=0.13, SE =0.06,
p= 0.024, power= 0.74). The change trajectory for those with no
mentors fell in-between the other two conditions and did not sig-
nificantly differ from either mentor condition (male mentor condition:
B=0.02, SE =0.06, p= 0.746; female mentor condition: B=0.11, SE =
0.06, p= 0.064). See Fig. 2.
Whereas 61% of participants without peer mentors and 65% of
those with male mentors self-reported participating in engineering
internships while in college, the proportion of internship participation
was substantially higher (82%) for female participants with a female
mentor (see Table 1).Thedifferencebetweenthefemalementor
condition and the other two conditions was significant (χ2(1,
N=125)=4.79,p= 0.029). When comparing conditions separately, the
percentage of female participants with female mentors who had par-
ticipated in engineering internships was significantly greater than the
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Baseline Year 1 Year 2 Year 3 Year 4 Year 5
Change in Confidence
Time
No Mentor
(N = 47)
Male Mentor
(N = 51)
Female Mentor
(N = 52)
Fig. 1 | Participants’confidence in engineering from college entry through one
year post-graduation. Depicts differences between intervention conditions
(no mentor, male mentor, female mentor) in participants’average change in
confidence in their engineeringabilities over fiveyears, relativeto their confidence
levels at the beginning of college, before mentors were assigned. Error bars des-
ignate +/−1 standard error ofeach condition’s estimatedmean confidence levels at
that time-point. N= 150 participants.
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 4
Content courtesy of Springer Nature, terms of use apply. Rights reserved
percentage in the no mentor condition (χ2(1, N=82)=4.58,p= 0.032).
The percentage in the male mentor condition did not significantly
differ from the no mentor condition (χ2(1, N=81)=0.18,p=0.669) or
the female mentor condition (χ2(1, N=87)=3.12,p= 0.077).
We examined whether peer mentorship in the first year of college
affected the proportion of students who graduated with an engineer-
ing majoror any STEM major 2–4 yearslater (see Table1). Wher eas 66%
of those with no mentor and 71% of those assigned a male mentor
earned Bachelor’s degrees in engineering, 79% of participants with a
female mentor earned engineering degrees. There was not a statisti-
cally significant difference between conditions (χ2(1, N= 150) = 1.85,
p= 0.174). However, for STEM majors (specifically, degrees in physical
sciences, biological sciences, computer science, engineering, or
mathematics48)asignificant difference emerged by mentor condition:
92% of participants assigned female mentors graduated with STEM
majors compared to 78% of participants assigned male mentors and
81% without mentors (χ2(1, N= 150) = 4.09, p=0.043).Althoughhaving
female peer mentors in the first year of college did not significantly
increase engineering degree completion, it did significantly increase
STEM degree completion compared to the other two conditions. The
percentage of female STEM graduates in the female mentor condition
was greater than in the male mentor condition (χ2(1, N= 103) = 3.99,
p= 0.046). The no mentor condition did not significantly differ from
the female mentor condition (χ2(1, N= 99) = 2.84, p=0.092) or the
male mentor condition (χ2(1, N=98)=0.09, p= 0.767).
Whereas participants without mentors (B=−0.50, SE =0.14,
p<0.001,power>0.99)andthoseassignedmalementors(B=−0.44,
SE =0.13, p= 0.001, power > 0.99) showed declining interest in pur-
suing post-graduate degrees in engineering from college entry
through graduation and one year post-graduation, those assigned
female mentors did not show this decrease in intentions to pursue
post-graduate degrees in engineering (B=−0.16, SE =0.11, p=0.159).
The change trajectory for the female mentor condition was sig-
nificantly different from the no mentor and male mentor conditions
(B=0.31, SE =0.15, p= 0.036, power = 0.90), showing that having a
female peer mentor in the first year of college was associated with less
of a decline in participants’reported post-graduate aspirations in
engineering. When comparing conditions separately, the female
mentor change trajectory did not significantly differ from the no
mentor change trajectory (B=0.34, SE =0.18, p=0.056) or the male
mentor slope (B=0.28, SE =0.17, p= 0.097). The male mentor slope
did not significantly differ from the no mentor slope (B=0.06, SE =
0.18, p=0.739).SeeFig.3.
We also examined the impact of mentorship condition on emo-
tional health and well-being throughout college. Whereas participants
without mentors and those assigned male mentors declined in their
emotional well-being over time (no mentor: B=−0.32, SE =0.17,
p= 0.061, power = 0.70; male mentor: B=−0.42, SE =0.14, p= 0.002,
power = 0.87), emotional well-being for participants with a female
mentor held steady throughout college (B=0.20,SE =0.14, p=0.147).
The difference in change trajectories between the female mentor
condition and the other two conditions was significant (B=0.57,SE =
0.18, p= 0.001, power = 0.97). The female mentor change trajectory
was significantly different from the no mentor change trajectory
(B=0.52, SE =0.22, p= 0.017, power = 0.78) and the male mentor
change trajectory (B=0.62,SE =0.19,p=0.001,power=0.89).Theno
mentor and male mentor conditions did not differ (B=0.10,SE =0.22,
p=0.659).SeeFig.4.
Confidence in engineering skills mediates the effect of female
peer mentors on academic choices in engineering
Multilevel mediation analyses were conducted to test whether the
effect of female peer mentors on academic choices in engineering
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Baseline Year 1 Year 2 Year 3 Year 4 Year 5
Change in Motivation
Time
No Mentor
(N = 47)
Male Mentor
(N = 51)
Female Mentor
(N = 52)
Fig. 2 | Participants’motivation in engineering from college entry through one
year post-graduation. Depicts differences between intervention conditions (no
mentor,male mentor,female mentor)in participants’averagechange in motivation
in their engineering courses over five years, relative to their motivation levelsat the
beginning of college, before mentors were assigned. Error bars designate +/
−1 standard errorof each condition’s estimated mean motivationlevels at that time-
point. N=150 participants.
Table 1 | Percentage (by Condition) that obtained an engi-
neering internship, engineering degree, and a STEM degree
No Mentor
Condition
Male Mentor
Condition
Female Mentor
Condition
Engineering Internship 61% 65% 82%
Engineering Degree 66% 71% 79%
STEM Degree 81% 78% 92%
Table 1 depicts the percentage of participants (separated by condition) who obtained engi-
neering internships at some point in college, as well as the percentage of participants who
obtained a Bachelor’s degree in engineering in particular or STEM in general.
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
(successfully securing engineering internships, majoring in STEM at
graduation, and post-graduate aspirations in engineering) were
explained by the change trajectories of confidence in overall engi-
neering skills, motivation in engineering courses, or emotional well-
being. For the objective outcomes (engineering internship and major
at graduation), we used 2-1-2 mediation models49 with mentor condi-
tion as the independent variable that varied between participants
(level 2), change in confidence, motivation, and emotional well-being
as the mediators that varied within participants (level 1), and
participation in engineering internships or graduating with a STEM
degree as the dichotomous dependent variable that varied between
participants (level 2). The mediation model for graduate aspirations
was a 2-1-1 model because the outcome variable, graduate intentions,
also varied within participants (level 1).
In each statisticalmodel, all continuous variables (i.e., confidence,
motivation, emotional well-being, graduate aspirations) were regres-
sed on time, creating random effects for the intercepts and slopes. The
dependent variable (internship, STEM degree, and slope of graduate
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
Baseline Year 1 Year 2 Year 3 Year 4 Year 5
snoitnetnIetaudarGniegnahC
Time
No Mentor
(N = 47)
Male Mentor
(N = 51)
Female Mentor
(N = 52)
Fig. 3 | Participants’intentions to pursue graduate school in engineering
from college entry through one year post-graduation. Depicts differences
between intervention conditions (no mentor, male mentor, female mentor) in
participants’average change in aspirations to pursue graduate degrees in
engineering over five years, relative to their aspiration levels at the beginning of
college, before mentors were assigned. Error bars designate +/−1 standard error of
each condition’s estimated mean aspiration levels at that time-point. N=150
participants.
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Baseline Year 1 Year 2 Year 3 Year 4 Year 5
Change in Emotional Health
Time
No Mentor
(N = 47)
Male Mentor
(N = 51)
Female Mentor
(N = 52)
Fig. 4 | Participants’emotional health and well-being from college entry
through o ne year post-graduation. Depicts differences between intervention
conditions (no mentor, male mentor, female mentor) in participants’average
change in emotional health over five years, relative to their reported levels during
their first year of college. Error bars designate +/−1 standard error of each condi-
tion’s estimated mean emotional health levels at that time-point. N=150
participants.
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 6
Content courtesy of Springer Nature, terms of use apply. Rights reserved
aspirations) was then regressed on the mediator slopes (change in
confidence, motivation, and emotional well-being) and the mediator
slopes were regressed on the independent variable (mentor condi-
tion). All continuous variables were centered at baseline to minimize
multicollinearity and dichotomous variables were specified as cate-
gorical. In order to compare the female mentor condition with the
other two conditions, condition was coded as a dichotomous variable
(0 = no mentor and male mentor conditions and 1 = female mentor
condition). Since Mplus does not allow for bootstrapping for indirect
effects in a two-level model, each model was conducted with the Bayes
estimator with 10,000 iterations, yielding near-identical results to
bootstrapping50.
Asignificant indirect effect indicated that the effect of the female
mentor condition versus the other two conditions on participation in
engineering internships was mediated through change in confidence
(B= 0.30, 95% CI [0.04, 0.77]). Participants who had female mentors
during their first year of college exhibited less decline in confidence in
their engineering abilities compared to participants in the other two
groups (B= 0.15, 95% CI [0.03, 0.28]). A smaller decline in confidence
was associated with more success in securing engineering internships
during college (B= 2.13, 95% CI [0.65, 3.84]). There were no significant
indirect effects for motivation in engineering courses (B= 0.12, 95% CI
[−0.08, 0.47]) or emotional well-being (B=−0.01, 95% CI [−0.17, 0.13]).
Both confidence (B= 0.48, 95% CI [0.06, 1.19]) and motivation
(B= 0.39, 95% CI [0.03, 1.05]) significantly mediated the effect of the
female mentor condition (versus the other two conditions) on gradu-
ating with STEM degrees. Compared to the other two conditions,
having a female mentor in the first year of college reduced the decline
in engineering confidence (B= 0.16, 95% CI [0.03, 0.28]) and motiva-
tion in engineering courses (B= 0.15, 95% CI [0.03, 0.27]), which in turn
predicted increased likelihood of graduating with a STEM degree 3–5
years later (confidence: B= 3.27, 95% CI [1.31, 5.86]; motivation:
B= 2.80, 95% CI [0.60, 5.52]). The indirect effect for emotional well-
being was not significant (B= 0.16, 95% CI [−0.06, 0.71]).
Confidence in overall engineering abilities significantly mediated
the effect of the female mentor condition on post-graduate aspirations
in engineering (B= 0.17, 95% CI [0.02, 0.42]). Participants who had
female mentors showed less decline in confidence compared to par-
ticipants with male mentors or no mentors (B= 0.15, 95% CI [0.04,
0.25]), which in turn was associated with maintenance of self-reported
post-graduate aspirations in engineering (B= 1.18, 95% CI [0.20, 2.28]).
There were no significant indirect effects of motivation in engineering
courses (B=0.07, 95% CI [
−0.09, 0.27]) or emotional well-being
(B=0.01,95%CI[−0.05, 0.11]) on reported post-graduate aspirations in
engineering.
Discussion
Our results indicate that the benefits of same-sex peer mentoring
relationships in the first year of college endure through the 3–5year
window of female students’college experience including one year
post-graduation, long after mentorship has ended. These benefits
emerge for both subjective experiences and objective academic out-
comes. The durability of this intervention is notable, especially given
its low-cost and light-touch nature. Mentees met with their mentors
only four times on average during their first year of college.
Four findings are particularly noteworthy. First, whereas female
participants assigned a male peer mentor or no mentor significantly
declined in their confidence in overall engineering skills and their
motivation in engineering courses, being assigned a female peer
mentor was consistently associated with reduced decline in con-
fidence and motivation. Second, in comparison to the other two con-
ditions, being assigned a female peer mentor was associated with:
increased rate of participation in engineering internships during col-
lege, increased graduation rate for Bachelor’s degrees in STEM, and
maintenance of higher aspirations to pursue post-graduate degrees in
engineering. Third, female mentorship resulted in greater reported
mentee emotional well-being, whereas male mentorship or no men-
torship was associated with a decline in emotional well-being through
college and one-year post-college. Fourth, through mediation ana-
lyses, we found that stable confidence in one’s engineering ability was
the psychological mechanism among our measured outcomes that
best explained why female peer mentorship promoted success in
engineering internships, completion of STEM degrees, and pursuit of
graduate training in engineering.
In addition to steadfast confidence at the individual level, we
speculate about two additional theory-driven mechanisms—one that
operates at an interpersonal level and another at an intrapersonal level
—that may explain why female mentors were so effective. From an
interpersonal perspective, female mentors may have expanded their
mentees’access to valuable social networks to offset disparities in
professional connections that often constrain members of minority
groups51,52. Sociological research shows that for students who are
minorities in an academic discipline, opportunities to develop a tight
social network of peers who share their marginalized identity in that
target profession increases professional success. For example, Yang
and colleagues53 found that female students in MBA programs were
more successful in getting high-ranked jobs after graduation if they
had a close network of female colleagues who advised them about
where to apply, how to interview, and offered other background
information about companies. Advice from close female connections
(vs. male connections) may be especially useful because female
informal advice-givers are cognizant of barriers about which male
peers may not be aware. In Yang’s study, female students had both
males and females in their professional network, but female students
who had a close network of trusted same-sex colleagues in their field
were more successful in the job market. These social network findings
may also apply to our study.
From an intrapersonal perspective, another potential mechanism
may be thathaving a female mentor elicited a change in their mentees’
mindset during a critical transitional period of their lives. As theorized
by Walton54, interventions may have a long-term impact if they change
recursive processes early on. If an intervention is able to alter indivi-
duals’way of thinking and help them gain more constructive mindsets
regarding their learning in the beginning of a new environment, they
will be more likely to persist in their education. In contrast, if students
remain caught in cycles of psyc hological threat and poor performance,
they may be less likely to persist. Applying mindset change to our
research, the benefit of female (as compared to male) mentors may
have persisted because female students were able to develop con-
structive mindsets regarding engineering in their first year of college.
Future research should directly examine these and other potential
mechanisms driving the success of same-sex mentorship interventions
that have long-lasting effects.
We also note limitations to our study due to its modest sample
size andlack of information regarding our mentors. Females make up a
small percentage of engineering students in the United States; at the
time these data were collected, only 3.9% of female students entering
the first yearof college nationwide intended to major in engineering33.
We collected data from four incoming first-year cohorts between
2011–2014 and followed their progress across the spanof 8 years to get
a robust longitudinal sample. Never theless, we acknowledge that some
of our significant effects are underpowered and recognize the possi-
bility that some nonsignificant findings (e.g., in relation to engineering
majors at graduation) may have been significant with a larger sample
size. Future workshould seek to replicate this intervention with larger
samples. Furthermore, because we did not ask mentors about their
own engineering experiences, we were unable to test whether mentor
success was related to mentee outcomes. For example, we did not
measure whether mentors were successful in obtaining engineering
internships, thus we could not test whether mentors who had
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
engineering internship experiences were more successful in guiding
their mentees to procure internships. Future research should examine
mentor outcomes and whether the beneficial effects of female men-
torship remain after accounting for mentor success.
Future research should also compare the roles of belonging and
confidence as mediating mechanisms over a long educational time
course (note: results for belonging, anxiety, and intentions to pursue
an engineering career are reported in the Supplementary Materials
section). Our earlier article found that in the first year of college,
female engineering students’feelings of belonging in engineering
consistently mediated the effect of female peer mentors on career
aspirations in engineering32. In contrast, our present investigation
following the same female students for a longer period of time shows
that when the entire college experience is analyzed, as in the current
study, confidence in engineering abilities is a better mediator of the
positive impacts of female peer mentors on female students’academic
outcomes. These results suggest that the mechanism that predicts
mentees’persistence and success in engineering may change over time
during college. More research is needed to replicate these findings and
explore the importance of different types of psychological mechan-
isms at different developmental stages of young people’s academic
careers.
Another difference emerged in student outcomes during the first
year of college compared with a longer timeframe (across the entire
college career plus one year post-graduation). In our previous article,
we found that at the end of the first year, 100% of participants who had
been assigned a female mentor remained in engineering majors,
compared to 82% of participants assigned a male mentor and 89% of
participants assigned no mentor32. Our present research shows that
this difference was no longer significant at graduation. However, a
closer examination of the data suggests that some participants who
had female mentors switched out of engineering after their first year to
a different STEM field—specifically computer science, mathematics,
chemistry, or biology. Indeed, participants with female mentors were
more likely to graduate with STEM degrees compared to participants
in the other two conditions. Thus, it appears that having female
mentors in engineering during their first year proved beneficial for
female participants’persistence in STEM broadly, which is crucial
given the undersupply of highly skilled talent across multiple STEM
fields in the U.S. labor market and the high demand for employees with
a variety of such skills.
It is also worth noting thatfemale students’engineering and STEM
grade point average (GPA) in college did not differ between mentor
conditions (ps≥0.457), nor did their GPA trajectories significantly
change over time (ps≥0.161) or differ by mentor condition (ps≥
0.504). Thus, the benefit of female mentorship on female participants’
academic outcomes did not occur by improving mentees’grades.
Instead, our findings suggest that improving their positive subjective
experiences in engineering were more likely to be the causes of the
observed benefits.
Some important future directions include replicating these find-
ings across other STEM disciplines where females are also a minority
(e.g., computer science, physics)33 and examining mentoring inter-
ventions at later points in females’STEM careers (e.g., during graduate
or postdoctoral traineeships, and early careers in STEM). Additionally,
an important direction of future research is to investigate the gen-
eralizability of these findings to other identity groups that are under-
represented and negatively stereotyped in STEM (e.g., Black, Hispanic,
Indigenous, and working-class college students who are first in their
families to receive a Bachelor’s degree)1,2,55.
We propose that examining (and scaling up) research-driven
solutions in academic and professional institutions to support the
success of underrepresented students in STEM pathways and tracking
the long-term impacts of these solutions are important steps to
improving representation in STEM. Our findings indicate that a low-
cost, light-touch mentoring relationship with a successful own-group
peer during the transition to college yields dividends even after it has
concluded, both for subjective indicators of confidence in overall
engineering abilities, motivation in engineering courses, and emo-
tional well-being as well as objective indicators of skills and persis-
tence. Such a relational intervention, initiated in the transition to
college, constitutes an important step towards equalizing the repre-
sentation of tomorrow’sscientistsandengineers.
Methods
Participants
Our study participants (i.e., potential mentees) comprised of 150
female students majoring in engineering at the University of Massa-
chusetts Amherst. We also recruited 58 student mentors (32 females,
26 males)based on faculty recommendations.Our study was approved
by the University of Massachusetts Amherst Institutional Review
Board. Data collection took place over 8 years, from 2011–2019. Par-
ticipants were recruited during new student orientation for the
2011–12, 2012–13, 2013–14, and 2014–15 academicyears. Most students
were entering first-years (80%). The remainder were transfer students
joining in their second year (20%). During their first year in the study,
participants were randomly assigned to a female mentor (n= 52), male
mentor (n= 51), or no mentor (control condition; n= 47). Participants
were told that the study was examining factors related to success and
development in engineering majors and were unaware that the
study was related to mentoring. Participants consented to taking part
in this long-term study and also gave consent for the researchers to
access their college transcripts. Mentors were undergraduate students
in their junior or senior year in the same engineering major as their
mentee (e.g., electrical engineering, mechanical engineering) and gave
consenttobeingamentor.
At the time of the baseline survey, participants were 18.34 years of
age, on average (SD = 1.34). Participants reported their sex and race
based on National Science Foundation (NSF) demographic groupings
at the time: participants were asked to identify whether they were male
or female, and whether they identified as American Indian/Alaskan
Native, Asian or Pacific Islander, Black (not of Hispanic origin), His-
panic, white (not of Hispanic origin), multiracial, or another ethnicity.
All participants reported their sex as female. The sample was 67.3%
white,17.3%Asian,5.3%multiracial,2.7%Black,2.7%Hispanic,and2%
other ethnicity. Participants were paid $20 for the baseline survey, $30
for the second survey, and $35 for the third survey during their first
year. Participants were paid $10 for each follow-up survey in sub-
sequent years. Peer mentors were paid $100 for each mentee they had.
As some mentors had multiple mentees, we tested whether any of our
effects could be attributed to individual mentors. We found that
individual mentors did not contribute significant variance to any out-
come variable (ps≥0.251), thus, we did not control for this variable in
our analyses.
Procedure
During their first year at the university, participants completed a sur-
vey at 3 time-points during the academic year: (1) a baseline survey in
August or September prior to meeting with their mentors, (2) a mid-
year survey in January or February, and (3) an end-of-year survey in
April or May. After the first year, when the mentoring relationship had
officially ended, we asked participants to complete a follow-up survey
each year during the spring semester (February to May) or summer,
until they graduated with a Bachelor’s degree plus one year post-
graduation. The number of participants at each time point is listed in
Table 2. On average, participants completed a total of 5.10 surveys;
69% of participants (104 out of 150) completed 5 or more surveys and
22% completed 4 surveys (33 out of 150). The number of surveys
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 8
Content courtesy of Springer Nature, terms of use apply. Rights reserved
completed did not significantly differ by mentoring condition (χ2(8,
N= 150) = 7.06, p= 0.531).
Peer mentors were primarily seniors,with some juniors, who were
in good standing in their engineering major and played leadership
roles in student-run professional clubs. Before being assigned a men-
tee, mentors attended a half-day training workshop at the beginning of
the academic year. During the training, we emphasized two major
points. First, mentors were asked to reflect on their first two years in
engineering, identify any difficulties they had experienced, and the
support or information that they wished they had received during that
difficult time. After generating their personal experiences, mentors
participated in a facilitated group discussion and collectively identified
what experiences encouraged them and other students to persist in
engineering and what other difficulties discouraged students and may
lead them to drop out of engineering. The ideas generated by mentors
were summarized in the formof a mentoring guide and provided to all
mentors, together with topics of discussion to raise with their mentees
(e.g., providing advice on academic coursework; providing tutoring
help; helping mentees develop plans for college and careers including
how to search for research assistantships and internships; providing
social support; connecting mentees with other students and student
clubs). Our mentoring guide can be found at https://osf.io/968ta/.
Mentors were encouraged toshare this information with their mentees
and to be a source of support. Second, mentors were asked to meet
with their mentees once a month throughout the academic year and
engage with them on any topic in their mentoring guide based on
mentee need. On average, mentors and mentees met four times for an
hour each during mentees’first-year in college. Mentors were told that
their primary goal was to be their mentee’sfriendandeasetheir
transition into college. They were encouraged to meet over social
activities. Finally, mentors were provided the names of their mentees
and asked to reach out to their mentees to initiate contact. They were
asked to complete a brief online survey after each meeting to keep a
record of the meeting, including a summary of the topics discussed.
Measures
For our measure of confidence, participants completed two items that
measured how confident they were about their overall ability in engi-
neering, drawn from previous research17,32. These items were: “Do you
think you have a talent for engineering?”and “How confident do you
feel about your engineering ability?”They answered on scales from 1
(not at all) to 7 (very much). Responses were averaged to create an
index of confidence in engineering (αs between 0.78 to 0.92). These
items were asked at three time-points during participants’first-year in
college and subsequently once a year until one year post-graduation.
Five items were used to measure participants’self-reported
experiences of motivation in the context of their engineering cour-
ses, which were drawn from prior research29,56–59.Thefive motivation
items were: “I have the skills and abilities to be successful in my
engineering-related classes this year;”“Iwillbeabletoovercomeany
difficulties I experience in my engineering-related classes this year;”“I
have what it takes to deal with my engineering-related classes;”“Iam
prepared to deal with my engineering-related classes;”and “Ifeel
confident about my engineering related classes this year.”If partici-
pants were responding to the survey after switching majors or after
graduating from college, they answered the questions pertaining to
the most recent engineering classes they had taken (e.g., “Ihadthe
skills and abilities needed to be successful in my engineering-related
classes”). Participants reported how much they agreed with each
statement on a scale from 1 (not at all true) to 7 (very true). Responses
to these items were averaged to form an index of motivation (αs
between 0.79–0.95). These items were asked at three time-points
during participants’first-year in college and subsequently once a year
until one-year post-graduation.
At the end of the first year and in each subsequent survey in
college, participants were asked to report whether they had an engi-
neering internship in the past year. In the post-college survey, parti-
cipants reported whether they had an engineering internship during
their final year of college. Participants’responses were compiled to
create a dichotomous index of whether they had any engineering
internships while in college (0 = no engineering internship, 1 = had one
or more engineering internships). We treated this as a dichotomous
measure for two reasons. First, there was a good bit of variability in the
number of years it took participants to graduate from college
(between 3–5 years), which would alter the number of internships that
they could have had in college, with more internship possibilities for
students who took longer to graduate. Second, the number of parti-
cipants who completed the survey varied from year to year, which
meant that a non-response in a given year would count as zero
internships for that year, which might not be accurate. For both these
reasons, engineering internships was treated as a categorical variable.
Using participants’undergraduate transcripts, we created two
dichotomous variables. First, we coded whether each participant
earned an engineering degree (0= no engineering degree, 1 = earned
an engineering degree). We also coded whether participants grad-
uated with a STEM degree (0 = no STEM degree, 1 = earned a STEM
degree). This variable was created by using the U.S. Department of
Education’sdefinition of STEM48 which includes biological sciences,
physical sciences, computer sciences, engineering, mathematics and
statistics, and science technologies. We did not include social and
behavioral sciences in this definition of STEM because social and
behavioral sciences tend to have more female students at the under-
graduate level. In some social science majors, females are the majority
at the undergraduate level60.
In each survey, we asked participants how likely they were to
pursue a Master’s degree or Ph.D. in engineering on a 7-point scale
from 1 (not at all) to 7 (very much)17. These items were asked at three
time-points during participants’first-year in college and once a year in
every subsequent year until one-year post-graduation.
During the last time-point of their mentorship year (i.e., spring of
their first year), we asked participants to self-report their mental
health over the past year (“In terms of your overall mental health, how
psychologically healthy and happy did you feel during this past aca-
demic year?”) on a scale of 1 (not at all) to 7 (very healthy and happy).
This item was repeated each year until one year post-college
graduation.
Table 2 | Number of participants across timepoints for continuous variables
Year 1 (Baseline: Before
Intervention)
Year 1 (Middle of
Intervention)
Year 1 (End of
Intervention)
Year 2 Year 3 Year 4 Year 5+
Number (percent) of survey
respondents
150 (100%) 150 (100%) 150 (100%) 102 (68%) 49 (33%) 61 (41%) 103 (69%)
Number (percent) of
college transcripts collected
––––––150 (100%)
Table 2 depicts the number and percentageof participants who completed the survey each year,as well as the percentage of college transcripts with major, courses taken, and grade
information thatwere collected fromthe university registrar’soffice. Almost all participants completedtheir post-graduation survey duringyear 5 and up. A few participants (n= 6) who graduated
during year 3 completed their post-graduation survey during year 4.
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Reporting summary
Further information on research design is available in the Nature
Portfolio Reporting Summary linked to this article.
Data availability
The participant data and syntax for each model in this study have been
deposited in the Open Science Framework (OSF) repository and canbe
found at https://osf.io/968ta/. Source data are provided with
this paper.
References
1. President’s Council of Advisors on Science and Technology. Report
to the President, engage to excel: Producing one million additional
college graduates with degrees in science, technology, engineer-
ing, and mathematics. February. https://files.eric.ed.gov/fulltext/
ED541511.pdf (2012).
2. Zaza, S., Abston, K., Arik, M., Geho, P. & Sanchez, V. What CEOs have
to say: Insights on the STEM workforce. Am. Bus. Rev. 23,
136–155 (2020).
3. Xue, Y., & Larson, R. C. STEM crisis or STEM surplus? Yes and yes.
Monthly Labor Review,https://doi.org/10.21916/mlr.2015.14
(2015).
4. Ambrose, M. Panel warns US faces STEM workforce supply chal-
lenges. FYI Science Policy News from AIP, 26. Retrieved from
https://www.aip.org/fyi/2019/panel-warns-us-faces-stem-
workforce-supply-challenges (2019).
5. Allen, J. M., Muragishi, G. A., Smith, J. L., Thoman, D. B. & Brown, E.
R. To grab and to hold: Cultivating communal goals to overcome
cultural and structural barriers in first-generation college
students’science interest. Transl. Issues Psychol. Sci. 1,
331–341 (2015).
6. Phelan, J. C., Link, B. G. & Tehranifar, P. Social conditions as fun-
damental causes of health inequalities: theory, evidence, and policy
implications. J. Health Soc. Behav. 51,S28–S40 (2010).
7. Cohen, G. L., Garcia, J., Purdie-Vaughns, V., Apfel, N. & Brzustoski, P .
Recursive processes in self-affirmation: Intervening to close the
minority achievement gap. Science 324,400–403 (2009).
8. Miyake, A. et al. Reducing the gender achievement gap in college
science: A classroom study of values affirmation. Science 330,
1234–1237 (2010).
9. Shapiro, J. R., Williams, A. M. & Hambarchyan, M. Are all interven-
tions created equal? A multi-threat approach to tailoring stereotype
threat interventions. J. Personal. Soc. Psychol. 104,277–288 (2013).
10. Shnabel, N., Purdie-Vaughns, V., Cook, J. E., Garcia, J. & Cohen, G. L.
Demystifying values-affirmation interventions: Writing about social
belonging is a key to buffering against identity threat. Personal. Soc.
Psychol. Bull. 39,663–676 (2013).
11. Walton, G. M. & Cohen, G. L. A brief social-belonging intervention
improves academic and health outcomes of minority students.
Science 331,1447–1451 (2011).
12. Walton, G. M., Logel, C., Peach, J. M., Spencer, S. J. & Zanna, M. P.
Two brief interventions to mitigate a “chilly climate”transform
women’s experience, relationships, and achievement in engineer-
ing. J. Educ. Psychol. 107,468–485 (2015).
13. Paunesku, D. et al. Mind-set interventions are a scalable treatment
for academic underachievement. Psychol. Sci. 26,784–793 (2015).
14. Yeager, D. S. et al. A national experiment reveals where a growth
mindset improves achievement. Nature 573,364–369 (2019).
15. Yeager, D. S. et al. Using design thinking to improve psychological
interventions: The case of the growth mindset during the transition
to high school. J. Educ. Psychol. 108,374–391 (2016).
16. Rozek,C.S.,Ramirez,G.,Fine,R.D.&Beilock,S.L.Reducing
socioeconomic disparities in the STEM pipeline through student
emotion regulation. Proc. Natl Acad. Sci. 116,1553–1558 (2019).
17. Stout, J. G., Dasgupta, N., Hunsinger, M. & McManus, M. A. STEMing
the tide: using ingroup experts to inoculate women’s self-concept
in science, technology, engineering, and mathematics (STEM). J.
Pers.Soc.Psychol.100,255–270 (2011).
18. Dasgupta, N. Ingroup experts and peers as social vaccines who
inoculate the self-concept: The stereotype inoculation model.
Psychol. Inq. 22,231–246 (2011).
19. Marx, D. M. & Goff, P. A. Clearing the air: The effect of experimenter
race on target’s test performance and subjective experience. Br. J.
Soc. Psychol. 44, 645–657 (2005).
20. Marx, D. M. & Roman, J. S. Female role models: Protecting women’s
math test performance. Pers.Soc.Psychol.Bull.28,
1183–1193 (2002).
21. McIntyre,R.B.,Paulson,R.M.,Taylor,C.A.,Morin,A.L.&Lord,C.G.
Effects of role model deservingness on overcoming performance
deficits induced by stereotype threat. Eur. J. Soc. Psychol. 41,
301–311 (2011).
22. Ramsey,L.R.,Betz,D.E.&Sekaquaptewa,D.Theeffectsofan
academic environment intervention on science identification
amongwomeninSTEM.Soc. Psychol. Educ. 16,377–397
(2013).
23. Cheryan,S.,Siy,J.O.,Vichayapai,M.,Drury,B.J.&Kim,S.Dofemale
and male role models who embody STEM stereotypes hinder
women’s anticipated success in STEM? Soc. Psychol. Pers. Sci. 2,
656–664 (2011).
24. Cheryan, S., Drury, B. J. & Vichayapai, M. Enduring influence of
stereotypical computer science role models on women’sacademic
aspirations. Psychol. Women Q. 37,72–79 (2013).
25. Herrmann, S. D. et al. The effects of a female role model on aca-
demic performance and persistence of women in STEM courses.
Basic Appl. Soc. Psychol. 38,258–268 (2016).
26. Vogt, C. M. Faculty as a critical juncture in student retention and
performance in engineering programs. J. Eng. Educ. 97,
27–36 (2008).
27. Gasman,M.,&Nguyen,T.H.(2019).MakingBlackScientists.In
Making Black Scientists. Harvard University Press.
28. Gallup Inc. Gallup-USA Funds minority college graduates report.
Washington, DC. Retrieved from https://www.uncf.org/wp-
content/uploads/PDFs/USA_Funds_Minority_Report_GALLUP-2.
pdf (2015).
29. Golden, A. A., Bogan, Y., Brown, L., Onwukwe, O. & Stewart, S.
Faculty mentoring: Applying ecological theory to practice at his-
torically Black colleges or universities. J. Hum. Behav. Soc. Environ.
27,487–497 (2017).
30. Davis, C. S. G. & Finelli, C. J. Diversity and retention in engineering.
N. Dir. Teach. Learn. 2007,63–71 (2007).
31. Winkle-Wagner, R. & McCoy, D. L. Feeling like an “Alien”or
“Family”? Comparing students and faculty experiences of diversity
in STEM disciplines at a PWI and an HBCU. Race Ethn. Educ. 21,
593–606 (2018).
32. Dennehy, T. C. & Dasgupta, N. Female peer mentors early in
college increase women’s positive academic experiences and
retentioninengineering.Proc. Natl Acad. Sci. 114,
5964–5969 (2017).
33. Hall, W. M., Schmader, T. & Croft, E. Engineering exchanges: Daily
social identity threat predicts burnout among female engineers.
Soc. Psychological Personal. Sci. 6,528–534 (2015).
34. Jensen, L. E. & Deemer, E. D. Identity, campus climate, and burnout
among undergraduate women in STEM fields. Career Dev. Q. 67,
96–109 (2019).
35. National Science Foundation. Women, Minorities, and Persons with
Disabilities in Science and Engineering. National Center for Science
and Engineering Statistics (NCSES). https://ncses.nsf.gov/pubs/
nsf19304/digest/about-this-report (2019).
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved
36. Society of Women Engineers. SWE research update: Women in
engineering by the numbers (Nov. 2019).https://alltogether.swe.
org/2019/11/swe-research-update-women-in-engineering-by-the-
numbers-nov-2019/ (2019).
37. United States Census Bureau. People.https://www.census.gov/
quickfacts/fact/table/US/LFE046219 (2019).
38. Belkin, D. A generation of American men give up on college: ‘Ijust
feel lost’. Wall Street Journal. https://www.wsj.com/articles/
college-university-fall-higher-education-men-women-enrollment-
admissions-back-to-school-11630948233 (2021).
39. Jarrett, V ., Tchen, C. M., & White House Council on Women and Girls
(U.S.). Keeping America’s women moving forward: The key to an
economy built to last. Washington, D.C: The White House Council
on Women and Girls. https://permanent.fdlp.gov/gpo21492/
womens_report_final_for_print.pdf (2012).
40. American Association for the Advancement of Science. STEM
mentoring: Emerging strategies for inclusion. National Science
Foundation. https://www.aaas.org/sites/default/files/2019-04/19-
018%20AAAS%20STEM%20Mentoring_final_web.pdf
(2017).
41. Rabionet, S. E., Santiago, L. E. & Zorrilla, C. D. A multifaceted
mentoring model for minority researchers to address HIV health
disparities. Am. J. Public Health 99,S65–S70 (2009).
42. Woetzel,J.etal.The power of parity: How advancing women’s
equality can add $12 trillion to global growth.McKinseyGlobal
Institute. https://www.mckinsey.com/~/media/McKinsey/
Industries/Public%20and%20Social%20Sector/Our%20Insights/
How%20advancing%20womens%20equality%20can%20add%
2012%20trillion%20to%20global%20growth/MGI%20Power%20of
%20parity_Full%20report_September%202015.pdf
(2015).
43. Eby,L.T.,Allen,T.D.,Evans,S.C.,Ng,T.&DuBois,D.L.Does
mentoring matter? A multidisciplinary meta-analysis comparing
mentored and non-mentored individuals. J. Vocat. Behav. 72,
254–267 (2008).
44. Muthén, L.K. and Muthén, B.O. (1998-2017). Mplus User’sGuide.
Eighth Edition. Los Angeles, CA: Muthén & Muthén.
45. Allison, P. D. (2010). Missing data. Thousand Oaks, CA: Sage.
46. Preacher,K.J.,Wichman,A.L.,MacCallum,R.,&Briggs,N.E.
(2008). Latent growth curve modeling. Thousand Oaks, CA: Sage
Publications.
47. Muthén, L. K. & Muthén, B. O. How to use a Monte Carlo study to
decide on sample size and determine power. Struct.Equ.Model.9,
599–620 (2002).
48. U.S. Department of Education. Integrated postsecondary education
data system (IPEDS).https://nces.ed.gov/programs/raceindicators/
indicator_reg.asp (2016).
49. Preacher,K.J.,Zyphur,M.J.&Zhang, Z. A general multilevel SEM
framework for assessing multilevel mediation. Psychological
Methods 15, 209 (2010).
50. Muthén, B. O., Muthén, L. K., & Asparouhov, T. (2017). Regression
and mediation analysis using Mplus.LosAngeles,CA:Muthén&
Muthén.
51. Pedulla,D.S.&Pager,D.Raceandnetworksinthejobsearch
process. Am. Sociol. Rev. 84,983–1012 (2019).
52. Trimble,L.B.&Kmec,J.A.Theroleofsocialnetworksingettinga
job. Sociol. Compass 5,165–178 (2011).
53. Yang, Y., Chawla, N. V. & Uzzi, B. A network’s gender composition
and communication pattern predict women’s leadership success.
Proc.NatlAcad.Sci.116,2033–2038 (2019).
54. Walton, G. M. The new science of wise psychological interventions.
Curr. Dir. Psychol. Sci. 23,73–82 (2014).
55. Martin, J. P., Stefl,S.K.,Cain,L.W.&Pfirman, A. L. Understanding
first-generation undergraduate engineering students’entry and
persistence through social capital theory. Int. J. STEM Educ. 7,
1–22 (2020).
56. Beltzer,M.L.,Nock,M.K.,Peters,B.J.&Jamieson,J.P.Rethinking
butterflies: The affective, physiological, and performance effects of
reappraising arousal during social evaluation. Emotion 14,
761 (2014).
57. Harvey, A., Nathens, A. B., Bandiera, G. & LeBlanc, V. R. Threat and
challenge: cognitive appraisal and stress responses in simulated
trauma resuscitations. Med. Educ. 44,587–594 (2010).
58. Mendes, W. B., Blascovich, J., Major, B. & Seery, M. Challenge and
threat responses during downward and upward social compar-
isons. Eur. J. Soc. Psychol. 31,477–497 (2001).
59. Moore, L. J., Vine, S. J., Wilson, M. R. & Freeman, P. The effect of
challenge and threat states on performance: An examination
of potential mechanisms. Psychophysiology 49,1417–1425
(2012).
60. National Science Foundation. Women, minorities, and persons with
disabilities in science and engineering. National Center for Science
and Engineering Statistics. https://ncses.nsf.gov/pubs/nsf19304/
data (2019).
Acknowledgements
We are grateful to the women who participated in this multi-year
experiment and to the mentors who dedicated their time. We also thank
Drs. Paula Rees and Bernhard Schliemann of the College of Engineering
for their help in recruiting participants and mentors. We thank Dr. Tara
Dennehy for her role in training mentors and collecting data during the
early years of this project and Dr. Holly Laws for her advice on growth
modeling and multilevel mediation. Finally, we thank members of the
Implicit Social Cognition Laboratory for their role in data collection, data
entry, and coding, and for their suggestions on an earlier draft of this
paper. This research was supported by National Science Foundation
Grant GSE 1132651 to N.D. (PI).
Author contributions
N.D. designed research and secured grant funding to support the work.
K.C.T. collected the data. D.J.W. performed the data analysis and
wrote the first draft. K.C.T. and N.D. reviewed and revised
the paper.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41467-022-34508-x.
Correspondence and requests for materials should be addressed to
Nilanjana Dasgupta.
Peer review information Nature Communications thanks the anon-
ymous reviewer(s) for their contribution to the peer review of this
work. Peer reviewer reports are available.
Reprints and permissions information is available at
http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jur-
isdictional claims in published maps and institutional affiliations.
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate if
changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2022
Article https://doi.org/10.1038/s41467-022-34508-x
Nature Communications | (2022) 13:6837 12
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Content uploaded by Nilanjana Dasgupta
Author content
All content in this area was uploaded by Nilanjana Dasgupta on Oct 27, 2024
Content may be subject to copyright.
Content uploaded by Nilanjana Dasgupta
Author content
All content in this area was uploaded by Nilanjana Dasgupta on Nov 13, 2022
Content may be subject to copyright.
