Content uploaded by Angela Duckworth
Author content
All content in this area was uploaded by Angela Duckworth on Jun 22, 2016
Content may be subject to copyright.
Content uploaded by Angela Duckworth
Author content
All content in this area was uploaded by Angela Duckworth on Jun 22, 2016
Content may be subject to copyright.
Teaching a lay theory before college narrows
achievement gaps at scale
David S. Yeager
a,1,2
, Gregory M. Walton
b,1,2
, Shannon T. Brady
c
, Ezgi N. Akcinar
b
, David Paunesku
d
, Laura Keane
e
,
Donald Kamentz
f
, Gretchen Ritter
g
, Angela Lee Duckworth
f,h
, Robert Urstein
i
, Eric M. Gomez
j
, Hazel Rose Markus
b
,
Geoffrey L. Cohen
b,c
, and Carol S. Dweck
b
a
Department of Psychology, University of Texas at Austin, Austin,TX 78712;
b
Department of Psychology, Stanford University, Stanford, CA 94305;
c
Graduate
School of Education, Stanford University, Stanford, CA 94305;
d
Project for Education Research That Scales, Department of Psychology, Stanford University,
Stanford, CA 94305;
e
uAspire, Boston, MA 02109;
f
The Character Lab, Philadelphia, PA 19104;
g
Department of Government, Cornell University, Ithaca, NY
14853;
h
Department of Psychology, University of Pennsylvania, Philadelphia, PA 19104;
i
Graduate School of Business, Stanford University, Stanford, CA
94305; and
j
Department of Psychology, University of Washington, Seattle, WA 98195
Edited by Michael Hout, New York University, New York, NY, and approved April 26, 2016 (received for review December 10, 2015)
Previous experiments have shown that college students benefit
when they understand that challenges in the transition to college
are common and improvable and, thus, that early struggles need
not portend a permanent lack of belonging or potential. Could
such an approach—called a lay theory intervention—be effective
before college matriculation? Could this strategy reduce a portion
of racial, ethnic, and socioeconomic achievement gaps for entire
institutions? Three double-blind experiments tested this possibil-
ity. Ninety percent of first-year college students from three insti-
tutions were randomly assigned to complete single-session, online
lay theory or control materials before matriculation (n>9,500). The
lay theory interventions raised first-year full-time college enrollment
among students from socially and economically disadvantaged back-
grounds exiting a high-performing charter high school network or
entering a public flagship university (experiments 1 and 2) and, at a
selective private university, raised disadvantaged students’cumula-
tive first-year grade point average (experiment 3). These gains corre-
spond to 31–40% reductions of the raw (unadjusted) institutional
achievement gaps between students from disadvantaged and non-
disadvantaged backgrounds at those institutions. Further, follow-up
surveys suggest that the interventions improved disadvantaged stu-
dents’overall college experiences, promoting use of student support
services and the development of friendship networks and mentor
relationships. This research therefore provides a basis for further tests
of the generalizability of preparatory lay theories interventions and
of their potential to reduce social inequality and improve other major
life transitions.
inequality
|
behavioral science
|
field experiment
|
social psychology
|
lay theories
Many students face significant challenges during the transi-
tion to college, increasing their risk of dropping out (1–4)
and undermining their future financial security, health, and con-
tributions to society (5, 6). Students who contend with social and
economic disadvantages can experience worse outcomes even
when they enter with identical academic qualifications (2, 3, 7,
8). Of course, institutions provide supports and resources to help
students navigate the transition to college, but this often occurs
after they have entered the institution. Is it possible to improve
the transition in advance? Could this mitigate inequality?
Students may not benefit from additional support before a
transition, for instance, more college advising, because they have
not yet faced the challenges associated with a new role or setting.
They might also forget material taught months in advance, such
as step-by-step directions for studying or for choosing classes.
However, taking a psychological approach, it may be helpful to
provide a lay theory (9) of the transition—a starting hypothesis
that many challenges in the transition are common and not cause
to doubt one’s prospects of belonging and success (10, 11). This
kind of lay theory may help people make sense of challenges they
later face and take steps to overcome them, for instance, by helping
students feel comfortable accessing support services and reaching
out to peers and professors.
If preparatory lay theory interventions were effective, it would sug-
gest a route for improving the college transition and, perhaps, other
life transitions. Moreover, because many colleges have structural
opportunities to reach students en masse before matriculation—
such as first-year student registration or orientation—this strategy
might remedy a portion of group disparities at institutional scale.
Not all psychological interventions are effective as preparation.
Writing expressively about emotions elicited by a past trauma helps
people cope and improves health, but expressive writing before a
trauma (e.g., a life-altering surgery) does not (12). Likewise, people
undergoing exposure therapy for anxiety disorders do not seem to
benefit from time anticipating an upcoming aversive stimulus, even
though processing emotions after exposure to the aversive stimulus
does improve outcomes (13). This may be because people have
difficulty forecasting the nature and emotional intensity of future
experiences (14).
However, lay theory interventions do not require people to
preprocess emotions. Instead, they provide a basis for assigning
Significance
In the United States, large, persistent gaps exist in the rates at
which racial, ethnic, and socioeconomic groups complete post-
secondary education, even when groups are equated on prior
preparation. We test a method for preventing some of those gaps
by providing individuals with a lay theory about the meaning of
commonplace difficulties before college matriculation. Across
three experiments, lay theory interventions delivered to over
90% of students increased full-time enrollment rates, improved
grade point averages, and reduced the overrepresentation of
socially disadvantaged students among the bottom 20% of class
rank. The interventions helped disadvantaged students become
more socially and academically integrated in college. Broader
tests can now be conducted to understand in which settings lay
theories can help remedy postsecondary inequality at scale.
Author contributions: D.S.Y., G.M.W., S.T.B., E.N.A., L.K., D.K., R.U., H.R.M., G.L.C., and C.S.D.
designed research; D.S.Y., G.M.W., S.T.B., E.N.A., D.P., L.K., D.K., G.R., A.L.D., R.U., and E.M.G.
performed research; D.S.Y. and D.P. contributed new reagents/analytic tools; E.N.A., L.K., D.K.,
G.R., and R.U. contributed to development of materials; A.L.D. facilitated collection of primary
outcome data; D.S.Y. and S.T.B. analyzed data; and D.S.Y., G.M.W., S.T.B., H.R.M., G.L.C., and
C.S.D. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Freely available online through the PNAS open access option.
1
D.S.Y. and G.M.W. contributed equally to this work.
2
To whom correspondence may be addressed. Email: dyeager@utexas.edu or gwalton@
stanford.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1524360113/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1524360113 PNAS Early Edition
|
1of8
PSYCHOLOGICAL AND
COGNITIVE SCIENCES
PNAS PLUS
meaning to experiences. Just as a scientific theory allows a
researcher to understand the causes and effects they observe in
their data (15), lay theories help people interpret adversities they
encounter—what caused the adversities and what they mean for
their future (9). Even if people cannot remember step-by-step
directions long in advance, it may be memorable and helpful to
learn that struggling in class or having difficulty making friends
does not mean that they, or people like them, are unintelligent or
do not belong.
Lay theory interventions should benefit students most who
have the greatest reason to draw negative inferences from adver-
sities in college and, therefore, may help reduce group-based in-
equalities (7). In the United States, students from racial/ethnic
minority groups and those who would be the first in their families to
earn a college degree (i.e., first-generation students) can face neg-
ative stereotypes about their intellectual ability, numeric un-
derrepresentation, and other group-based threats on campus (2, 3,
7, 16). This circumstance can lead students to worry whether they
and people like them can fully belong (11, 17), will be seen as
lacking intelligence (7, 18, 19), or will be a poor cultural fit in col-
lege (20, 21). Although students from majority groups can experi-
ence personal worries about belonging and potential, they are less
likely to experience these at the group level. Such concerns seed
harmful inferences for even commonplace challenges in college such
as feelings of loneliness, academic struggles, or critical feedback:
“Maybe this means people like me do not belong or cannot suc-
ceed.”These inferences sap motivation and undermine achievement
through a cycle that gains strength through its repetition (22, 23).
This process is depicted in Fig. 1, which highlights how prob-
lematic lay theories—arising from group-based experiences that
give rise to persistent worries about belonging, potential, and cul-
tural fit—can impede the social and academic integration critical to
success in college (4, 24), contributing to group-based inequality.
As we have suggested, a corollary of the process in Fig. 1 is that
providing people with an alternative lay theory—one that charac-
terizes adversities as common and improvable—should reduce those
inequalities. In small-scale trials, this strategy has been effective
when carried out as treatment. Specifically, interventions teaching
an alternative lay theory have raised achievement for negatively
stereotyped racial minority and first-generation college students
(11, 18, 20), in one case reducing the racial achievement gap by half
over 3 years. Furthermore, effects appeared to be mediated by the
psychological and behavioral processes depicted in Fig. 1—for in-
stance, asking professors or teaching assistants for help, attending
office hours, or taking chances on making friends (11, 17, 25).
Although promising, the scale and timing of past lay theory
interventions in college have been limited by the constraints of
face-to-face delivery. Each past trial was implemented at a se-
lective 4-y university and involved fewer than 50 treated disad-
vantaged students. The present studies, by contrast, randomized
more than 90% of students at three diverse institutions to online
lay theory or control conditions (n>9,500).
Thus, in addition to our primary contribution of testing the
potential of preparatory rather than reactive lay theory interven-
tions, these studies provide an unparalleled test of the scalability,
replicability, and generalizability of lay theory interventions to
lessen postsecondary inequality. Further, we administer a much
broader range of measures of social and academic integration than
available in past lay theory research, directly testing the processes
predicted in Fig. 1.
Overview of Present Research
Three double-blind, randomized experiments tested the effects
of prematriculation (i.e., before arriving on campus), Internet-
administered lay theory interventions on postsecondary outcomes
for students who face greater social and economic disadvantages
in college. Randomization was at the student level.
The primary objective was to test the general efficacy of pre-
paratory lay theory interventions, so the studies tested several
variants, including (i) two used in past research with college
students (addressing social belonging and growth mindset of
intelligence) (11, 18, 26), (ii) one used previously with younger
students (addressing lay theories about critical feedback) (27),
and (iii) a novel intervention based on prior research and theory
with college students (addressing lay theories about cultural fit)
(21). All interventions had a common structure, format, and
length (Methods and SI Appendix, Appendix 1).
Each experiment had four conditions—one control and three
interventions. The primary analysis involved the contrast of re-
ceiving any intervention versus the control. Post hoc comparisons
tested for differences among them. In the one case where a dif-
ference was found, interventions were interpreted separately.
Evaluations were conducted in diverse settings—high school
seniors exiting high-performing charter school networks and at-
tending 70+mostly low-selectivity 2- and 4-y colleges, first-year
students at a public flagship with low 4-y graduation rates, and
first-year students at a highly selective university. Analyses fo-
cused on core first-year college outcomes appropriate to these
contexts. In each setting, students were known to be college-
ready. This was crucial for testing our theory; lay theories will not
remedy inadequate preparation but may allow prepared students
to achieve in college.
As in past experiments, negatively stereotyped racial/ethnic
minority students and first-generation students were expected to
benefit. Although these groups differ in important respects, they
share a common psychological predicament in college: group-based
fears about belonging, potential, or cultural fit (7, 17, 21). However,
social and economic disadvantages are not inherent to groups but
defined by the social context (28). Therefore, in experiments 2 and
3, the groups expected to benefit were defined using (i) theory and
(ii) institutional data revealing which groups have, in the past,
performed less well at each university (see analyses in SI Appendix,
Appendix 3, pp. 30–31, and Appendix 4, pp. 49–50). In experiment
1, all students were either racial minority or first-generation stu-
dents and, based on low graduation rates in previous cohorts, were
known to face disadvantages in college.
Throughout the paper, all estimates are raw percentages or
means, unadjusted for covariates. All analyses are intent-to-treat.
Statistical tests for treatment effects are from regression models
with preintervention covariates (SAT score, high-school class rank,
and gender). Significance levels do not differ without covariates.
Results
Experiment 1: Urban Charter Graduates. Participants were two co-
horts of outgoing seniors at four high-performing urban charter
Membership in a group facing
social disadvantage in higher
education
Experiences of challenge or
setback (e.g. critical feedback,
low score, feelings of loneliness)
With a lay theory of the
transition to college
"I/people like me don't
succeed at my
college."
"It's common to go
through challenges like
this and overcome
them."
Withdrawal from the social/
academic environment
Sustained engagement in the
social/academic environment
Worse achievement, lower
persistence, college completion.
Higher achievement, greater
persistence, college completion.
Psychological
Interpretation
Behavioral
Response
Academic
Outcome
Recursive Process
Recursive Process
Fig. 1. Theoretical model: the process through which lay theories affect
disadvantaged students’behavior and academic outcomes across the tran-
sition to college.
2of8
|
www.pnas.org/cgi/doi/10.1073/pnas.1524360113 Yeager et al.
schools (n=584). They were admitted to more than seventy
4- and 2-y public and private colleges. This study involved initial
correlational analysis of the effects of the social belonging lay
theory, followed by an experiment testing the effects of a social
belonging lay theory intervention, customized for urban charter
students (SI Appendix, Appendix 2a). Fully crossed with this was
a growth mindset of intelligence intervention, a week later.
Correlational analysis. First, correlational data among untreated
students from the first cohort of data collection (n=185) showed
that those who had a lay theory that they might not belong in
college (i.e., belonging uncertainty, e.g., “Sometimes I worry that
I will not belong in college”) (17) were less likely to persist
through the first year of college [logistic regression odds ratio
(OR) =0.66, Z=−2.60, P=0.009]. This was true controlling for
a number of covariates (SI Appendix, Appendix 2b, Table S7).
No other personality trait or lay theory, including growth mindset
(see below), predicted college persistence in this sample, nor did
intelligence quotient, net of SAT, and grade point average
(GPA). Beyond prior preparation, it was not students’traits or
attitudes that predicted success in the transition to college. It was
the presence of an initial doubt about whether they would fit in
(for related analyses, see ref. 3).
Primary intervention outcome: Full-time enrollment. Among students
from both cohorts, only 32% of students in the randomly assigned
active control condition were enrolled full-time continuously for the
first year of college, even though 100% were admitted to college.
For students who received any preparatory lay theory intervention,
that number was 41%, a significant increase (logistic regression
OR =1.59, Z=2.17, P=0.030).
Post hoc tests revealed that this treatment contrast masked
significant heterogeneity (growth mindset only =36%; social
belonging only =45%; social belonging plus growth mindset =
41%) (Fig. 2). The growth mindset-only condition showed sig-
nificantly poorer outcomes compared with the two social be-
longing conditions (Z=2.00, P=0.046) and did not differ from
active controls (P>0.50) (see below for possible explanation).
Furthermore, the two social belonging conditions did not differ on
the basis of whether students also received a growth mindset (Z =
1.13, P=0.26), and both social belonging conditions combined
differed significantly from the active control, logistic regression
(OR =1.87, Z=2.70, P=0.007). Hence, only the social be-
longing intervention was effective in the present experiment.
Social belonging intervention. Why was the social belonging inter-
vention effective? First, survey questions answered moments
after completing the social belonging materials showed that it
effectively taught the lay theory that many students feel that they
do not belong at first in college but come to do so over time (SI
Appendix, Appendix 2a, Table S5; for analogous self-reports in
experiments 2 and 3, see SI Appendix, Appendix 3, Table S9, and
Appendix 4, Table S17).
Next, an exploratory analysis of survey data at 6-mo follow-up
suggests that the social belonging intervention increased stu-
dents’social and academic integration on campus, supporting
the model in Fig. 1. Students who received a social belonging
intervention were more likely than students who did not to re-
port that they had used academic support services, had joined an
extracurricular group, and had chosen to live on campus [com-
posite index of social and academic integration, t(50) =2.76, P=
0.008, d=0.78] (Fig. 3). This index statistically mediated effects
of the social belonging intervention on year-end full-time con-
tinuous enrollment [indirect effect b=0.15 (95% confidence
interval (CI) =0.03, 0.29), P<0.01] (28). There was no effect of
the mindset intervention on this metric (P>0.50). This media-
tion analysis is limited by sample size, effects in only two of three
intervention conditions, and narrow scope of measures. These
are improved upon in experiment 2.
Growth mindset intervention. The growth mindset of intelligence
intervention has previously been effective when delivered via the
Internet (26, 29) but, as noted, was not here. This was not expected.
Two possibilities are germane, but we cannot definitively confirm
one or the other. One is that the intervention was redundant with
growth mindset messages already taught by the high-performing
charter network: the school already administered similar materials
and messages, and students in this population already endorsed a
growth mindset (SI Appendix, Appendix 2b, p. 28). A second is that
the growth mindset message was represented as a private belief,
not a reflection of their college’s values. Even apart from private
beliefs, the perception that an organization endorses a fixed
mindset can lead people who face negative stereotypes to worry
that their intelligence will be questioned (30).
Experiment 2: Public University. Experiment 2 extended experiment
1by(i) testing lay theory interventions with incoming students at
a high-quality 4-y public university, instead of outgoing students
at a high school network; (ii) having interventions come from the
university instead of the high school, which may increase efficacy
(31); (iii) testing whether students from socially and economi-
cally disadvantaged backgrounds benefit disproportionately and,
thus, whether lay theory interventions remedy a portion of in-
equality; (iv) increasing the sample size by a factor of 12 (n=
7,335); (v) investigating the model in Fig. 1 with a broad measure
of social and academic integration; and (vi) comparing students
in the randomized cohort to students in previous and later co-
horts not randomized to condition, to examine reductions in
inequality for the entire institution. For details on intervention
materials and the sample, see SI Appendix, Appendix 3. The focal
lay theory interventions were social belonging and growth
mindset of intelligence interventions. The latter was theoretically
distinct from that tested in experiment 1 in that it represented
growth mindset as the ethos of the university.
Primary outcome: Full-time enrollment. A leading predictor of on-time
graduation is first-year full-time enrollment (i.e., completing
Fig. 2. A prematriculation social belonging intervention improves full-time
college enrollment among urban charter high school graduates in experi-
ment 1. The growth mindset intervention was not effective in this experi-
ment. Bars represent raw, unadjusted means or percentages.
Yeager et al. PNAS Early Edition
|
3of8
PSYCHOLOGICAL AND
COGNITIVE SCIENCES
PNAS PLUS
12+credits both semesters of the first year) (32). Therefore, as in
experiment 1, this was the primary outcome.
First, inequality at this institution in the randomized control
group was large. Disadvantaged students in the control condition
were 10 percentage points less likely to complete the first year
full-time enrolled in both terms compared with advantaged stu-
dents (69% versus 79%; raw values; χ
2
(1) =27.32, P<0.001)
(Table 1).
Next, the one-time prematriculation lay theory interven-
tions cut this inequality by 40%, increasing the percentage
of full-time enrolled disadvantaged students over the first year
to 73%, a significant increase compared with the randomized
control group (OR =1.23, Z=2.26, P=0.024). Unlike ex-
periment 1, there was no heterogeneity across intervention
conditions, meaning that they appeared to be equally effective
(mindset only =74%; belonging only =72%; social belonging
AB
C
Fig. 3. Prematriculation lay theory interventions improve disadvantaged students’social and academic integration in college in the first year. (A) Experiment
1: social belonging, participants from two conditions that received a social belonging intervention; control, participants from two conditions that did not
receive a social belonging condition (because the growth mindset intervention did not affect full-time enrollment or survey responses, as noted in the text).
(Band C) Experiments 2 and 3: any intervention, participants receiving any of the three interventions; control, participants receiving no intervention. Bars
represent raw means or percentages, unadjusted for covariates. The yaxes for unstandardized scores represent full scale ranges.
Table 1. Percent completing fall semester full-time enrolled in experiment 2, by year or condition
Study year or condition Advantaged students Disadvantaged students
2011: no intervention (n=6,896) 90% 81%
2012: randomized control (n=2,062) 90% 82%
2012: randomized intervention (n=5,356) 90% 86%***
2013: no intervention (n=6,719) 88% 81%
2014: nonrandomized intervention (n=6,244) 90% 84%***
***P<0.001, significantly different from disadvantaged students who did not receive the intervention.
4of8
|
www.pnas.org/cgi/doi/10.1073/pnas.1524360113 Yeager et al.
plus growth mindset =73%; all pairwise contrasts among in-
tervention conditions Ps>0.50).
There was no effect of the lay theory interventions among
advantaged students (OR =1.03, Z=0.22, P=0.79), nor was
there expected to be, replicating past research (11). However,
because the trend for advantaged students was positive, the
group ×condition interaction did not reach significance (Z=
1.59, P=0.11), as in some past research (18) (Fig. 4).
Why was the growth mindset lay theory intervention beneficial
here (SI Appendix, Appendix 3, Table S10) but not in experiment
1? One possibility is that this sample had not been systematically
exposed to prior growth mindset messaging (although we do not
have a direct test of this). A second and related possibility is that
a growth mindset was more consequential for this sample: here,
unlike experiment 1, growth mindset predicted full-time enroll-
ment in the control group (SI Appendix, Appendix 3, p. 36) (be-
longing uncertainty was also a significant predictor, P<0.01,
replicating experiment 1). Third, here the growth mindset message
signaled the beliefs of key people in the institution (e.g., profes-
sors), as shown by analyses of the manipulation check items (30).
Furthermore, students’beliefs about the university’sgrowth
mindset ethos predicted outcomes in the control group net of SAT
scores (SI Appendix, Appendix 3, p. 36).
The condition that combined growth mindset and belonging
treatments was no more (or less) effective than either interven-
tion on its own, replicating findings from analogous research
(26). One explanation for this comes from manipulation check
analyses, which suggest that by teaching two messages in a single
session, neither came through as clearly as when they were
taught alone (SI Appendix, Appendix 3, p. 40).
Full-scale institutional change. Year-over-year comparisons (n=
14,216 advantaged and disadvantaged students) examined
whether the lay theory interventions could contribute to reduc-
tions in institutional-level inequality when delivered to the entire
incoming student body without randomization (analyses focused
on full-time fall term enrollment because spring term data were
unavailable for the nonexperimental years; Table 1). There were
no over-time changes among disadvantaged students who did not
receive an intervention (81–82%), but disadvantaged students
who received the randomized treatment (86%) and who received
the nonrandomized treatment (84%) both differed from all un-
treated disadvantaged students (Ps<0.001), and this produced
significant disadvantaged group ×condition interactions (Ps<
0.001) because there was no corresponding increase for advan-
taged students (intervention year, 90%; nonintervention years,
88–90%). Hence, the lay theory interventions appear to have led
to full-scale reduction in institutional inequality.
Social and academic integration. Among the quarter of students
randomized to the control group, disadvantaged students were
more likely to be identified as at risk for dropping out due to
their reports of social and academic integration (disadvantaged
students: 13% at risk of dropping out; advantaged students: 8%),
a marginally significant inequality (Z=1.92, P=0.055). How-
ever, among the three-quarters of students randomized to a lay
theory intervention, only 7% of disadvantaged students were
designated at risk, a significant decrease (Z=2.46, P=0.014),
eliminating the group difference (Fig. 3). This improvement in
social and academic integration mediated the intervention effect
on continuous full-time full-year enrollment among disadvan-
taged students [indirect b=0.01 (95% CI =0.001, 0.03), P=
0.040] (28).
Experiment 3: Selective University. Experiment 3 extended these
results by (i) testing preparatory lay theory interventions at a
selective private university (n=1,592); (ii) examining an out-
come appropriate for selective colleges (cumulative first-year
GPA instead of full-time enrollment because fewer than 1% of
participating students at this institution fail to persist through
their first year); (iii) testing social belonging and two other lay
theory interventions, focused on cultural fit and the experience
of receiving critical feedback, but not growth mindset; and (iv)
providing a more detailed assessment of social and academic
integration than available in experiment 2 and assessing this at
the end of the academic year instead of during the fall term. This
precludes mediational analyses but assesses the durability of
benefits.
Primary outcome: First-year GPA. In the randomized control condi-
tion, disadvantaged students earned lower GPAs than advantaged
3.62 3.61 3.33 3.42
3.00
3.10
3.20
3.30
3.40
3.50
3.60
3.70
3.80
3.90
4.00
Control Intervention Control Intervention
(N = 303) (N = 881) (N = 103) (N = 309)
Advantaged students Disadvantaged students
Experimental Condition, by Demographic Group
Experiment 3: Selective Private University
First Year Cumulative GPA, Fall through Spring
79% 80% 69% 73%
60%
65%
70%
75%
80%
85%
90%
95%
100%
Control Intervention Control Intervention
(N = 1,310) (N = 3,372) (N = 752) (N = 1984)
Advantaged students Disadvantaged students
Experimental Condition, by Demographic Group
Experiment 2: Large Public University
First-Year Continuous, Full-Time Enrollment,
Fall and Spring
Fig. 4. Prematriculation lay theory interventions narrow first-year achievement gaps when delivered to entire incoming classes (>90%). Each experiment
involved three intervention conditions that are presented in composite form because they did not differ significantly. Experiment 2 intervention condition
separate effects: mindset only, 74%; belonging only, 72%; social belonging plus growth mindset, 73%. Experiment 3 intervention condition separate effects:
social belonging, 3.39; cultural fit, 3.47; critical feedback, 3.39. Bars represent raw, unadjusted means or percentages.
Yeager et al. PNAS Early Edition
|
5of8
PSYCHOLOGICAL AND
COGNITIVE SCIENCES
PNAS PLUS
students (disadvantaged students: M=3.33 on a 4.0 scale; advan-
taged students: M=3.62; raw means), planned contrast t(1,591) =
6.99, P<0.001, d=0.80—a significant achievement gap.
However, receiving a lay theory intervention raised disadvan-
taged students’first-year GPAs by 0.09 grade points to 3.42, a
significant improvement [t(1,591) =2.16, P=0.031, d=0.25].
Like experiment 2, post hoc tests found no heterogeneity across
treatment conditions (all contrast ps>0.30; social- belonging =
3.39; cultural fit =3.47; critical feedback =3.39). The im-
provement from the combined preparatory lay theory interven-
tions corresponds to a 31% reduction in the raw achievement
gap (47%, covariate-adjusted; SI Appendix, Appendix 4, Table
S17). There was no intervention effect for advantaged students,
t<1, and the subgroup ×condition interaction was significant
[F(1,1588) =4.04, P=0.045] (Fig. 4).
Previous national surveys found that disadvantaged students at
elite colleges disproportionally end up in the lowest quintile of
class rank (the bottom 20%) (2). Replicating this, in the control
condition, 46% of disadvantaged students fell in the bottom
quintile for first-year GPA. This percentage was reduced to 29%
by the lay theory interventions [χ
2
(1) =8.76, P=0.003] (for
results for all class rank increments, see SI Appendix, Appendix 4,
Figs. S4 and S5). There was no harm or benefit on class rank for
advantaged students (χ
2
<1).
Comparison with preintervention years. Historical data were obtained
for two preintervention years (n=3,029). The cumulative first-
year GPAs of disadvantaged students randomized to the control
condition in the study year did not differ from those in the prior
years (3.33 vs. 3.32, respectively; P>0.80). However, disadvan-
taged students who received a lay theory intervention earned
significantly higher GPAs relative to peers from nonintervention
years [planned contrast t(4517) =4.02, P<0.001, d=0.26].
Social and academic integration. Analyses of a spring survey, com-
pleted roughly 11 mo postintervention, showed that treated
disadvantaged students reported greater social and academic
integration on campus at the end of the first year—having made
more close friends, being more likely to have developed a close
mentor relationship, being more involved in extracurricular groups,
and making greater use of academic support services—compared
with randomized controls [composite index planned contrast
t(473) =3.20, P=0.001, d=0.72] (SI Appendix, Appendix 4,
Table 21)(Fig.3).
Although consistent with our process model (Fig. 1), these
responses were collected after most grades had been recorded
and so cannot be used in mediation analysis of first-year GPA.
However, friends, mentors, and extracurricular involvement are
important outcomes in their own right and facilitate later college
achievement, as well as postcollege opportunities above and be-
yond classroom performance (4, 24).
Discussion
The present research has two major contributions. First, it shows
that preparatory lay theory interventions can be effective in the
college transition. Second, it directly informs theory about soci-
etal inequality (6).
Preparatory Intervention. Three institutional-scale, double-blind,
randomized, active-controlled experiments showed improvement
in disadvantaged students’achievement in the first year of
college. In experiments 2 and 3, the treatments reduced raw
achievement gaps by 31–40%. These effects were obtained with
samples of 90%+of incoming students and reproduced when the
intervention was administered to all students on a nonrandomized
basis (experiment 2). Extending past research, they represent the
most rigorous replication of lay theory interventions to date. These
characteristics make the quality of these data unusually high and
highly relevant for policy as well as for cumulative science.
On a practical level, these effect sizes are important. When a
single college student does not graduate, he or she can forfeit
between $500,000 to $1 million dollars in lifetime wages (33).
Likewise, colleges and universities invest significant resources to
recruit and retain students from underrepresented groups, and
society bears financial responsibility for adults who cannot find
work due to a lack of training. The present effects thus signify the
recovery of meaningful financial losses for individuals, colleges,
and the public. Moreover, once evaluated in a given context,
the present interventions can be administered at low marginal
cost on an ongoing basis. Indeed, this has occurred in the post-
intervention years with the schools that participated in these
experiments. This said, the present results derive from a specific
set of postsecondary settings. A critical next step, for both theory
and application, is to systematically interrogate settings to de-
termine where effects are greater or weaker.
Can preparatory lay theory interventions improve other major
life transitions? This is an exciting possibility. If a new mother
struggles to cope with a difficult baby, will she benefit from
knowing in advance that this does not make her a bad mother but
is common and can improve with time? If a veteran struggles to
find work, will he or she benefit from hearing in advance stories
about struggles other veterans faced returning to civilian society
and how they overcame these? What kinds of people are most at
risk for drawing harmful inferences in these and other major life
transitions (e.g., to retirement and in divorce), who may benefit
most from lay theory interventions? Is it possible to deliver such
interventions at a socially meaningful scale? Research to answer
these questions could uncover new ways to reduce long-standing
social problems.
Societal Inequality. Educational attainment is the best predictor of
upward mobility in the developed world (6); thus, inequality in
the rates at which qualified students from different racial, ethnic,
and social class backgrounds complete college threatens the ideal
of meritocracy and undermines both individuals’life prospects
and national economic growth (2, 3, 8). Our findings directly
speak to ongoing debates about the bases of inequality in soci-
ology, economics, and law.
For instance, sociologists have found that prior social and
economic disadvantages faced by African-American, Latino, and
lower-social class college students predict their belief that people
in college will view them as not having the intellectual potential
to succeed. These beliefs, in turn, predict poor college achieve-
ment and persistence (2, 3). Relatedly, using correlational data,
economists have estimated that nearly half of dropouts after the
first year of college among socioeconomically disadvantaged
students can be attributed to a loss of confidence in the first term
(1). However, the lack of causal evidence for these relationships
led one leading scholar to lament, “How strong is the evidence
that beliefs matter? Unfortunately, not very strong”(ref. 8,
p. 69). Using random assignment experiments, the present re-
search directly confirms the role of student beliefs (aka lay theo-
ries) arising from social disadvantage in causing postsecondary
inequality, at least within the settings studied here.
Next, one influential economic model suggests that struggling
first-year students learn that they truly lack intellectual ability,
which allows them to make the informed choice to withdraw (1).
Our experiments show that this choice need not reflect a lack of
ability. When students learned that early difficulties are common
and not necessarily diagnostic of a lack of ability or belonging,
they showed they could succeed.
Our research also complements so-called behavioral economic
approaches, which show that remedying procedural barriers (like
simplifying the completion of financial aid forms and text messages
to remind students to pay registration fees or select courses) can
increase college persistence (34, 35). These are not thought to
operate through the mechanisms depicted in Fig. 1 and therefore
6of8
|
www.pnas.org/cgi/doi/10.1073/pnas.1524360113 Yeager et al.
represent a complementary (not competing) cause of and remedy
for postsecondary inequality.
Finally, in law, some scholars and Supreme Court justices have
argued that it does racial, ethnic, and socioeconomic minority
students a disservice to admit them to elite institutions where
they will subsequently earn low class rank or drop out (36, 37).
However, experiment 3 showed that low class rank is not auto-
matic for students from disadvantaged backgrounds. When se-
lective colleges create environments in which students’belonging
and potential are not constantly in question, they can admit a
diverse class, and members of this class can succeed (38).
Implications, Limitations, and Future Directions. A critical mis-
interpretation of our research would be to conclude that psy-
chological disadvantage is simply in the head. To the contrary,
worries about belonging and potential are pernicious precisely
because they arise from awareness of real social disadvantage
before and during college, including biased treatment, university
policies and practices that inadvertently advantage some groups
of students over others, and awareness of negative stereotypes
and numeric underrepresentation (2–4, 7). Some lay theory
interventions highlight how students’experiences can differ
along group identity lines and how students can overcome group-
specific challenges to belong and succeed (20, 25). Further, in
implementing lay theory interventions, institutions recognize
how awareness of social disadvantage can facilitate threatening
interpretations of adversities disproportionately for students
from disadvantaged backgrounds. These interventions represent
institutional efforts to help all students navigate challenges in
college effectively.
Another misinterpretation would be to think of lay theories
interventions as “magic bullets”that will work universally for all
racial, ethnic, and socioeconomic minority students in all settings
without adaptation or that they work alone, absent other student
supports (23). Lay theory interventions should be effective only
when (i) materials successfully redirect problematic lay theories
in a given context (hence the customization carried out in each
study) (29); (ii) student motivation is hindered by the targeted
lay theory in a given context; and (iii) colleges afford instruc-
tional opportunities, relationship opportunities, and financial
supports necessary for better outcomes (23).
A critical question then is, in what types of institutions will
different groups of students benefit most and least from which lay
theory interventions? The present research did not fully answer this,
but it did find variability that may be informative for future theory
development and replication attempts. A growth mindset inter-
vention was ineffective in experiment 1, where students had already
been taught a growth mindset and where the mindset intervention
was not represented as reflective of their colleges’values. However,
it was effective in experiment 2, where students had not been sys-
tematically exposed to growth mindset ideas and where a growth
mindset was represented as the ethos of the university. First-
generation Asian-American students benefitted in experiment 2
but performed well and were not classified as disadvantaged in
experiment 3.
We did not test whether students who were unprepared for
college or who did not desire a college credential would benefit
from the intervention because they were not expected to. It will
be exciting in future research to expand to more heterogeneous
samples of students and institutions, and, ideally probability-
based samples, to allow for generalizable inferences.
Finally, it is essential to recognize that there are many psy-
chologically meaningful events in the transition to college, which
can either support or thwart students’growth and belonging.
Preparatory interventions have a special importance because
they provide students a first lens for making sense of events they
later experience. However, how a professor introduces difficult
course material, provides critical feedback, or responds to a
struggling student; how a dean welcomes the entering class; and
how a university frames an academic probation letter, among
many other events, can have independent effects (27, 28). These
all represent opportunities to improve the psychological envi-
ronment of college and thus student success.
Methods
General Intervention Procedures for All Experiments. Each intervention used
the same basic grammar to teach a lay theory (SI Appendix, Appendix 1). Each
(i) clearly described specific difficulties in college and persuasively repre-
sented these as common and as changeable; (ii) provided vivid stories from
upper-year students who experienced and overcame common struggles;
(iii) conveyed data—either results of surveys or summaries of the neuroscience of
learning—in support of these messages; and (iv) allowed participating students
to take ownership of the lay theory after having reflected on the stories and
data by writing about challenges they anticipated and how these challenges are
common and likely to change over time. These essays, students were told, might
be shared with future students to improve their transition (exemplary ones
were). This final element, called the “saying-is-believing”technique, allows
students to personalize generic materials, promoting internalization (10, 17,
18). The institutional review boards at the University of Texas at Austin and
at Stanford University approved all procedures. Students provided active
consent on the first page of the surveys. The data reported in this paper are
tabulated in the SI Appendix and are available upon request from the
authors to investigators, provided that requestors secure the necessary
institutional review board approvals.
Experiment 1.
Participants and procedure. Experiment 1 included two consecutive cohorts of
seniors (n=584, or 97% of students) graduating from high-performing “no-
excuses”public urban charter high schools. This charter network has con-
sistently produced large gains in state test scores compared with local district
schools, it graduates almost all of its students, and college admission is a
requirement for graduation. Hence, students were, by some standards,
college-ready. However, only approximately one-fourth earn any postsec-
ondary credential within 6 y.
Participants were predominately African-American (88%) and first-gen-
eration college students (67%) (all but one student were one, the other, or
both). Continuous (fall and spring) full-time enrollment was tracked through
an objective, third-party database: the National Student Clearinghouse (NSC).
The NSC does not report GPA (39).
Intervention. From students’perspectives, the intervention was a survey about
the transition to college in which they would hear from upper-year students.
To control for placebo effects, all students were told the survey might be
interesting and might help them succeed in college. Students completed it in
their classrooms in May of their senior year of high school. After this session,
students were not contacted again by researchers. Students in cohort 1 were
later recruited by school staff (blind to condition) for the fall survey.
One-half of students were randomized to a social belonging lay theory
intervention within gender ×high school GPA (high vs. low) strata. The exact
stories were based on prior research but revised extensively through a de-
sign thinking process (29) described in SI Appendix, Appendix 2a. The other
half of participants were randomized to a control condition, which included
similar tasks but focused on adjusting to the physical rather than social envi-
ronment of college (e.g., weather and campus buildings). It also involved
reading stories from upper-year students and writing an essay to future stu-
dents transitioning to college. As noted, fully crossed with the social belonging
intervention was a growth mindset of intelligence intervention. This was
delivered 1 wk later (SI Appendix, Appendix 2a).
Fall survey. For the first cohort of students (n=172), staff at the partner
charter network contacted students in late fall of their first year of college to
invite them to complete a survey assessing social and academic integration
on campus. The survey link was sent to students via email, text message, and
social media. Thirty percent of students in this cohort completed the survey
(n=52 students). See SI Appendix, Appendix 2a, for questions.
Experiment 2.
Participants and procedure. Participants were 7,335 (91%) of the first-year
students entering a high-quality public university. Over 85% were in the top
10% of their high school class, but only 50% typically graduated in 4 ys (32).
Students completed online orientation materials in private between May and
August in the summer before entering college. The materials were embedded
among other preorientation tasks, including reporting immunization records,
Yeager et al. PNAS Early Edition
|
7of8
PSYCHOLOGICAL AND
COGNITIVE SCIENCES
PNAS PLUS
signing the honor code, registering for classes, etc. The presentation of the in-
tervention mirrored experiment 1.
Interventions. Randomization to one of four conditions occurred with equal
probability within gender ×SAT (high vs. low) ×college major ×race/ethnicity
strata. The social-belonging intervention (11) was based on prior research and
adapted to fit the prematriculation context. The growth-mindset intervention
(18, 26) used neuroscientific information to convey that intelligence is not a fixed
quantity but can be developed with effort on challenging tasks. In the combined
condition, participants read growth mindset intervention content and upper-
year student stories drawn from the belonging condition. The control condition
mirrored experiment 1 and past research (11).
Survey. Independent of the research team, the university hired a private firm
to administer a web-based survey to all first-year students 2 mo into the fall
semester. The survey was never associated with the present intervention and
differed significantly in format and presentation. All first-year students were
told in an email that they were required to complete it; 1,722 students (23%)
didso.Student-leveldataonanaggregated“at-riskness”index were
obtained by the research team (SI Appendix, Appendix 3).
Experiment 3.
Participants and procedure. Participants were 1,592 matriculating students
(90% of the incoming class) at a highly selective private university. Analogo us
to experiment 2, the intervention materials were embedded as a link on the
matriculation website hosted by the university’s undergraduate advising office.
Between mid-May and early June, students were required to complete a number
of forms on this website. Randomization to one of four conditions occurred with
equal probability within gender ×race/ethnicity strata.
Interventions. Students were randomized to a control condition, to a social-
belonging intervention nearly identical to that tested in experiment 2, or to
one of two other lay theory interventions, which addressed contextually
specific beliefs relevant to social and academic belonging. One was a cultural
fit intervention. This emphasized that although college has an indepen-
dent cultural focus (e.g., “choose your path”), students can maintain interde-
pendent social relationships with home communities and join interdependent
communities in college. This message was designed to forestall the inference
for students coming from interdependent cultural communities (many eth-
nic minority and first-generation college students) that early social difficul-
ties mean that they have a fundamental cultural mismatch with college (21).
The second was a critical feedback intervention. This emphasized that critical
academic feedback from professors and other instructors reflects instructors’
high standards and confidence students can meet those standards, not a
negative judgment of the student or his or her potential (27).
Survey. In the spring of students’first year, a follow-up survey was conducted
with participating students. Students were invited to participate in exchange
for a small token (a class sticker). A total of 31% of the sample (n=491 out
of 1,592) elected to do so. The survey also included additional items assessing
attitudinal/affective measures such as feelings of belonging and perceived
stress; for consistency with experiments 1 and 2, and to understand behavioral
outcomes, analyses reported here focus on behavioral measures of social and
academic integration.
ACKNOWLEDGMENTS. We thank A spire, KIPP, Mastery , and Yes Pre p charter
schools and the Stanford University d.school for assistance piloting and proto-
typing interventions. We thank PERTS for logistical and IT support. We also
thank students and colleagues who contributed to design processes, including
K. Belden, R. Crandall, C. Gabrieli, J. Gabrieli, A. Ericcsson, A. House, J. Powers,
A. Royalty, C. Steele, and U. Treisman, and those who assisted with data
collection and analysis,including J. Beaubien, K. Breiterman-Loader, C. Macrander,
S. Keller, M. Niemasz-Cavanagh, P. McGowan, and O. Zahrt-Omar. Critical
feedback was provided by B. Bigler, R. Crosnoe, C. Muller, J. Pennebaker, and
the social and personality area at the University of Texas at Austin. This
research received support from the William and Melinda Gates Foundation,
Stanford University, and the University of Texas at Austin.
1. Stinebrickner R, Stinebrickner T (2014) Academic performance and college dropout:
Using longitudinal expectations data to estimate a learning model. J Labor Econ
32(3):601–644.
2. Espenshade TJ, Radford AW (2009) No Longer Separate, Not Yet Equal: Race and Class
in Elite College Admission and Campus Life (Princeton Univ Press, Princeton).
3. Massey DS, Charles CZ, Lundy G, Fischer MJ (2011) The Source of the River: The Social
Origins of Freshmen at America’s Selective Colleges and Universities (Princeton Univ
Press, Princeton).
4. Tinto V (1994) Leaving College: Rethinking the Causes and Cures of Student Attrition
(Univ of Chicago Press, Chicago, IL).
5. Hayward MD, Miles TP, Crimmins EM, Yang Y (2000) The significance of socioeco-
nomic status in explaining the racial gap in chronic health conditions. Am Sociol Rev
65(6):910–930.
6. D’Addio AC (2007) Intergenerational Transmission of Disadvantage: Mobility or
Immobility Across Generations? A Review of the Evidence for OECD Countries
(Organ for Eco n Co-op and Dev , Paris).
7. Steele CM (1997) A threat in the air. How stereotypes shape intellectual identity and
performance. Am Psychol 52(6):613–629.
8. Morgan SL (2005) On the Edge of Commitment: Educational Attainment and Race in
the United States (Stanford Univ Press, Stanford, CA).
9. Ross L, Nisbett RE (1991) The Person and the Situation: Perspectives of Social
Psychology (Pinter & Martin Ltd, London), 2nd Ed.
10. Wilson TD, Linville PW (1985) Improving the performance of college freshmen with
attributional techniques. J Pers Soc Psychol 49(1):287–293.
11. Walton GM, Cohen GL (2011) A brief social-belonging intervention improves aca-
demic and health outcomes of minority students. Science 331(6023):1447–1451.
12. Pennebaker JW, Chung CK (2011) Expressive writing and its links to mental and
physical health. The Oxford Handbook of Health Psychology, ed Friedman HS (Oxford
Univ Press, London), pp 417–437.
13. Tiwari S, Kendall PC, Hoff AL, Harrison JP, Fizur P (2013) Characteristics of exposure
sessions as predictors of treatment response in anxious youth. J Clin Child Adolesc
Psychol 42(1):34–43.
14. Wilson TD, Gilbert DT (2005) Affective forecasting: Knowing what to want. Curr Dir
Psychol Sci 14(3):131–134.
15. Kuhn TS (1970) The Structure of Scientific Revolutions (Univ of Chicago Press, Chicago, IL).
16. Murphy MC, Steele CM, Gross JJ (2007) Signaling threat: How situational cues affect
women in math, science, and engineering settings. Psychol Sci 18(10):879–885.
17. Walton GM, Cohen GL (2007) A question of belonging: Race, social fit, and achieve-
ment. J Pers Soc Psychol 92(1):82–96.
18. Aronson JM, Fried CB, Good C (2002) Reducing the effects of stereotype threat on
African American college students by shaping theories of intelligence. J Exp Soc
Psychol 38(2):113–125.
19. Dweck CS, Leggett EL (1988) A social-cognitive approach to motivation and person-
ality. Psychol Rev 95(2):256–273.
20. Stephens NM, Hamedani MG, Destin M (2014) Closing the social-class achievement
gap: A difference-education intervention improves first-generation students’aca-
demic performance and all students’college transition. Psychol Sci 25(4):943–953.
21. Stephens NM, Fryberg SA, Markus HR, Johnson CS, Covarrubias R (2012) Unseen disad-
vantage: How American universities’focus on independence undermines the academic
performance of first-generation college students. J Pers Soc Psychol 102(6):1178–1197.
22. Cohen GL, Garcia J, Purdie-Vaughns V, Apfel N, Brzustoski P (2009) Recursive pro-
cesses in self-affirmation: Intervening to close the minority achievement gap. Science
324(5925):400–403.
23. Yeager DS, Walton GM (2011) Social-psychological interventions in education: They’re
not magic. Rev Educ Res 81(2):267–301.
24. Coleman JS (1988) Social capital in the creation of human capital. Am J Sociol 94:95–120.
25. Walton GM, Logel C, Peach JM, Spencer SJ, Zanna MP (2015) Two brief interventions
to mitigate a “chilly climate”transform women’s experience, relationships, and
achievement in engineering. J Educ Psychol 107(2):468–485.
26. Paunesku D, et al. (2015) Mind-set interventions are a scalable treatment for aca-
demic underachievement. Psychol Sci 26(6):784–793.
27. Yeager DS, et al. (2014) Breaking the cycle of mistrust: Wise interventions to provide
critical feedback across the racial divide. J Exp Psychol Gen 143(2):804–824.
28. Murphy MC, Walton GM, From prejudiced people to prejudiced places: A social-
contextual approach to prejudice. Frontiers in Social Psychology Series: Stereotyping
and Prejudice, eds Stangor C, Crandall C (Psychology Press, New York).
29. Yeager DS, et al. (2016) Using design thinking to improve psychological interventions:
The case of the growth mindset during the transition to high school. JEducPsychol
108(3):374–391.
30. Murphy MC, Dweck CS (2010) A culture of genius: How an organization’s lay theory
shapes people’s cognition, affect, and behavior. Pers Soc Psychol Bull 36(3):283–296.
31. Castleman BL, Owen L, Page LC (2015) Stay late or start early? Experimental evidence
on the benefits of college matriculation support from high schools versus colleges.
Econ Educ Rev 47:168–179.
32. Diehl R (2012) Final Report of the Task Force on Undergraduate Graduation Rates
(Univ of Texas at Austin, Austin, TX).
33. Julian TA (2012) Work-Life Earnings by Field of Degree and Occupation for People
with a Bachelor’s Degree: 2011 (US Census Bureau, Washington, DC).
34. Castleman BL, Page LC (2015) Summer nudging: Can personalized text messages and
peer mentor outreach increase college going among low-income high school grad-
uates? J Econ Behav Organ 115:144–160.
35. Bettinger EP, Long BT, Oreopoulos P, Sanbonmatsu L (2012) The role of application
assistance and information in college decisions: Results from the H&R Block FAFSA
experiment. Q J Econ 127(3):1205–1242.
36. Sander R, Jr, ST (2012) Mismatch: How Affirmative Action Hurts Students It’s Intended
to Help, and Why Universities Won’t Admit It (Basic Books, New York).
37. Liptak A (2015) Supreme court justices’comments don’t bode well for affirmative acti on.
NY Times. Available at www.nytimes.com/2015/12/10/us/politics/supreme-court-to-
revisit-case-that-may-alter-affirmative-action.html. Acc es se d De cem be r 10 , 20 15.
38. Walton GM, Spencer SJ, Erman S (2013) Affirmative meritocracy. Soc Issues Policy Rev
7(1):1–35.
39. Dynarski SM, Hemelt SW, Hyman JM (2013) The Missing Manual: Using National
Student Clearinghouse Data to Track Postsecondary Outcomes (Natl Bur of Econ Res,
Cambridge, MA), NBER Work Pap 19552.
8of8
|
www.pnas.org/cgi/doi/10.1073/pnas.1524360113 Yeager et al.
