Content uploaded by Mark J Graham
All content in this area was uploaded by Mark J Graham on Feb 05, 2015
Content may be subject to copyright.
A 2012 report by the President’s Coun-
cil of Advisors on Science and Tech-
nology (PCAST) predicts that the
U.S. workforce will suffer a deﬁ cit of one
million college graduates in science, technol-
ogy, engineering, and mathematics (STEM)
over the next decade ( 1). The report calls for
addressing the shortfall by increasing reten-
tion of college students in STEM. But many
academic leaders have not responded aggres-
sively to workforce needs by implementing
measures that increase retention. Some of
this nonaction is likely due to lack of knowl-
edge about proven retention strategies.
Here, we introduce a “persistence frame-
work” that integrates evidence from psychol-
ogy and education research into a guide for
launching and evaluating initiatives aimed
at increasing persistence of interested col-
lege students ( 2– 5). We emphasize “persis-
tence,” which focuses on student agency ( 6),
rather than on the institutional perspective of
“retention,” although the intended outcomes
are the same. Most of the research we review
was conducted in the United States, but the
problem is not unique to one country, and the
solutions are probably universal.
Persistence of more top students would
address the projected STEM workforce def-
icit, while building a deeper, broader talent
pool ( 1). Less than half of the three mil-
lion students who enter U.S. colleges yearly
intending to major in a STEM ﬁ eld persist in
STEM until graduation ( 1). The exit rate is
especially high for the so-called “underrep-
resented majority,” women, racial, and eth-
nic minorities who are underrepresented in
STEM majors but collectively make up 68%
of college students in the United States ( 7).
For example, African-American students
who intend to major in STEM switch to a
non-STEM ﬁ eld before graduation twice as
often as white students ( 8). Such stark statis-
tics invite a hard look at research and prac-
tice that bear on retention.
The concept of persistence originates in
social and cognitive psychology as one man-
ifestation of motivation ( 6). In education,
motivation is viewed as a driver of student
engagement. Among the important con-
structs underlying motivation is the powerful
inﬂ uence of conﬁ dence (i.e., self-efﬁ cacy),
which is a requirement for persistence ( 9).
Therefore, it is imperative that persistence
efforts address motivation and conﬁ dence
(see the ﬁ gure).
The framework identifies learning and
professional identiﬁ cation as determinants
of persistence. Research demonstrates their
importance in predicting student behavior ( 2,
4, 10), and both can be modulated by myriad
interventions ( 7). Some of the most success-
ful STEM retention initiatives pay careful
attention to both elements ( 3). Moreover,
both learning and professional identifica-
tion increase conﬁ dence and, consequently,
motivation, which in turn spur academic suc-
cess and feeling like a scientist, thus creating
mutually reinforcing experiences (see the
ﬁ gure). This contrasts with student reports
of many current introductory STEM courses
that obscure the subject, diminish students’
confidence, and discourage them from
becoming scientists ( 7). Although the con-
ceptual elements are well established, unify-
ing them into a single persistence framework
for guiding STEM education is new.
Answering PCAST’s call to increase
STEM student retention requires wide-
spread attention. Departments and institu-
tions also need ﬂ exibility in the approaches
they take and a look at working examples
they can model. Because the framework uni-
ﬁ es principles that may be implicit in even
the most successful programs, we highlight
a few intensively studied ones with quantiﬁ -
able success (although many other success-
ful models exist).
For African-American students and other
underrepresented groups, the University of
Maryland–Baltimore County Meyerhoff
Scholars Program has dramatically increased
student achievement, retention, and graduate
study in STEM ﬁ elds. Of their 508 STEM
majors between 1993 and 2006, Meyerhoff
boasts 86% retention in STEM ( 3), twice
the nationwide average for all students and
more than four times the average retention
for African-American students. Other pro-
grams such as the Biology Scholars Program
at University of California, Berkeley ( 11),
more broadly target gender, racial, and eth-
nic groups. Another approach is the peer-led
Gateway Science Workshops at Northwest-
ern University ( 12), which are open to all
beginning STEM students. The Posse pro-
grams that focus on urban-schooled science
students ( 1), and the LA-STEM and Howard
Hughes Medical Institute (HHMI) Research
Scholars Programs at Louisiana State Uni-
versity that focuses on promoting student
diversity in STEM are other outstanding
Such successful programs commonly use
three interventions widely recognized for
inspiring STEM students: (i) early research
experiences, (ii) active learning in introduc-
tory courses, and (iii) membership in STEM
learning communities (see the ﬁ gure).
Early research experiences. Despite well-
known beneﬁ ts of research experience, most
undergraduates are not offered research
opportunities until late in college, after
the critical period of attrition from STEM
Increasing Persistence of College
Students in STEM
Mark J. Graham,
1, 2 Jennifer Frederick,
1 Angela Byars-Winston,
3 Anne-Barrie Hunter,
An evidence-based framework offers a guide
for efforts to increase student persistence
in STEM majors.
1Center for Scientific Teaching, Department of Molec-
ular, Cellular and Developmental Biology, Yale Uni-
versity, New Haven, CT 06520, USA. 2Department of
Psychiatry, School of Medicine, Yale University, New
Haven, CT 06511, USA. 3Department of Medicine, School
of Medicine and Public Health, University of Wisconsin,
Madison, WI 53705, USA. 4Ethnography and Evaluation
Research, University of Colorado, Boulder, CO 80309, USA.
*Corresponding author. E-mail: email@example.com
www.sciencemag.org SCIENCE VOL 341 27 SEPTEMBER 2013 1455
The Persistence Framework. Conﬁ dence is belief
in one’s own ability; motivation is intention to take
action in pursuit of goals; learning is acquiring
knowledge and skills; and professional identiﬁ cation
is feeling like a scientist.
Published by AAAS
majors ( 13). Students who engage
in research in the first 2 years of
college are more likely to persist in
STEM majors ( 14). Research expe-
rience is a powerful learning tool,
engaging students and stimulating
curiosity; and it naturally encour-
ages professional identification
because students are being scien-
tists, not just studying products of
other people’s science.
The PCAST report recommends
implementing research courses for
all beginning undergraduates ( 1).
Research courses provide students
with the project ownership and intel-
lectual challenges of empirical pur-
suit. At the same time, these courses
use teaching time and materials efﬁ ciently by
having student teams work on parallel prob-
lems that require similar techniques.
In research courses, students engage in
authentic research—they design experi-
ments, collect and analyze data, and some-
times make significant discoveries ( 15);
thus, undergraduates in research courses
experience the same dramatic gains in learn-
ing and positive attitudes toward science as
those who conduct research in faculty labo-
ratories ( 10, 16).
A variety of research courses have been
implemented successfully at large and small
institutions. Faculty who are understandably
hesitant to accept inexperienced undergrad-
uates into their research laboratories find
research courses feasible and rewarding to
teach. One is the multi-institutional HHMI–
Science Education Alliance (SEA) PHAGES
in which freshman discover new bacterio-
phages from soil ( 15). The University of
Texas at Austin’s Freshman Research Initia-
tive demonstrates that research courses can
also be cost-effective on a large scale when
they replace traditional introductory lab
courses. In the Austin model, faculty provide
projects, derived from their own research, as
the basis for student research projects in lab
sections of 20 to 30 students.
Active learning in introductory courses.
Many talented college students ﬂ ee STEM
majors because they find introductory
courses uninspiring ( 7). This can be cor-
rected by incorporating classroom teaching
practices that engage students in the learning
process, known as “active learning,” which
has been shown to reduce STEM attrition
( 17). Active learning includes any activity
in which every student must think, create, or
solve a problem. For example, brief lectures
interspersed with opportunities for students
to reﬂ ect on or apply their own knowledge
induces active engagement in large lec-
ture courses ( 18). Active learning improves
understanding and retention of concepts and
information ( 1), and it helps students identify
as scientists because they participate in sci-
entiﬁ c thinking with peers who create a sci-
entiﬁ c community.
Faculty are often reluctant to try active
learning because of lack of experience, so it
is essential to provide training. Opportunities
exist at many universities, professional soci-
eties, and in the National Academies Sum-
mer Institutes on Undergraduate Science
Education, a national program that trains
instructors in evidence-based instructional
methods ( 19).
Membership in STEM learning commu-
nities. Learning communities are typically
virtual or physical structures that provide
gathering places or events that enable stu-
dents to work with and learn from each
other ( 12).
Forming learning communities often
requires a small financial investment that
produces substantial impacts on student
achievement and persistence. Establishing
learning communities can be as simple as
ensuring that all students have access to a
study group outside of class or providing an
online environment where students can dis-
cuss course content. Learning communities
can be constructed in tutoring centers where
students congregate by course or discipline,
science clubs, or science-based residential
communities ( 12). All of these activities
stimulate intellectual growth. Involvement
with other students who are aspiring scien-
tists also strengthens professional identity.
To ensure inclusion of all students in
learning communities, attention must be paid
to being impartial. Students from groups typ-
ically underrepresented in science are less
likely to form study groups, may be unaware
of the academic beneﬁ ts of group
work outside of class, and confront
unintentional biases that may make
it challenging to break into estab-
lished cliques ( 20). Instructors can
reduce this tendency by confronting
the issue in class and encouraging
students to form and join inclusive
Proven interventions exist, so
now it is time for all stakeholders
to contribute to increasing student
persistence in STEM majors (see
table). The elements of the persis-
tence framework are universal and
can be tailored for any classroom. In
the United States, a concerted effort
to implement evidence-based strat-
egies will pay off by advancing the goal of
having sufﬁ cient STEM college graduates to
meet projected workforce needs.
References and Notes
1. PCAST, Engage to Excel: Producing One Million
Additional College Graduates with Degrees in Science,
Technology, Engineering, and Mathematics (PCA ST,
Washington, DC, 2012).
2. M. Estrada, A. Woodcock, P. R. Hernandez, P. W. Schultz,
J. Educ. Psychol. 103, 206–222 (2011).
3. M. F. Summers, F. A. Hrabowski III, Science 311,
4. H. Thiry, S. L. Laursen, A.-B. Hunter, J. Higher Educ. 82,
5. A. Byars-Winston, B. Gutierrez, S. Topp, M. Carnes, CBE
Life Sci. Educ. 10, 357–367 (2011).
6. A. Bandura, Am. Psychol. 44, 1175–1184 (1989).
7. E. Seymour, N. M. Hewitt, Talking About Leaving (Wes t-
view Press, Boulder, CO, 1997).
8. National Academy of Sciences, National Academy of
Engineering, Institute of Medicine, Expanding Underrep-
resented Minority Participation: America’s Science and
Technology Talent at the Crossroads (National Academies
Press, Washington, DC, 2011).
9. C. S. Dweck, Am. Psychol. 41, 1040–1048 (1986).
10. D. Lopatto et al., Science 322, 684–685 (2008).
11. J. Matsui, R. Liu, C. M. Kane, Cell Biol. Educ. 2, 117–121
12. G. Light, M. Micari, Making Scientists (Harvard Univ.
Press, Cambridge, MA, 2013).
13. S. H. Russell, M. P. Hancock, J. McCullough, Science 316,
14. B. A. Nagda, S. R. Gregerman, J. Jonides, W. Von Hippel,
J. S. Lerner, Rev. Higher Educ. 22, 55 (1998).
15. G. F. Hatfull et al., PLoS Genet. 2, e92 (2006).
16. A.-B. Hunter, S. L. Laursen, E. Seymour, Sci. Educ. 91,
17. D. C. Haak, J. HilleRisLambers, E. Pitre, S. Freeman,
Science 332, 1213–1216 (2011).
18. J. Handelsman, S. Miller, C. Pfund, Scientiﬁ c Teaching
(W. H. Freeman, New York, 2007).
19. C. Pfund et al., Science 324, 470–471 (2009).
20. C. A. Moss-Racusin, J. F. Dovidio, V. L. Brescoll, M. J.
Graham, J. Handelsman, Proc. Natl. Acad. Sci. U.S.A.
109, 16474–16479 (2012).
Acknowledgments: This work was supported by a grant
to J.H. from the HHMI Professors Program and NIH Grant
1R13GM090574-01. We thank E. Baraban and J. Young for
comments on an earlier version of this manuscript.
27 SEPTEMBER 2013 VOL 341 SCIENCE www.sciencemag.org
(i) Faculty and instructional staff should teach undergraduate research
courses, use active learning in introductory STEM courses, and encourage
students to form learning communities;
(ii) Students should be educated about the benefits of learning communities
and supported to create their own;
(iii) Departments should examine curricula and reward structures to
incentivize effective teaching, and then align them to enable early research
and active learning in introductory courses;
(iv) Provosts, deans, and chairs should advocate for and dedicate resources
to changing classroom practice by creating opportunities for instructors to
learn new teaching techniques;
(v) Public and private funding entities should apply the persistence
framework to evaluation of new initiatives in STEM undergraduate
(vi) Accreditation agencies should incorporate measurements of STEM
persistence into their periodic institutional reviews.
Published by AAAS