Encouraging Girls in Math and Science: IES Practice Guide

Encouraging Girls in
Math and Science

IES Practice Guide
NCER 2007-2003
U.S. DEPARTMENT OF EDUCATION
Math and Science
IES Practice Guide
SEPTEMBER 2007
Diane F. Halpern (Chair)

Claremont McKenna College
Joshua Aronson
New York University
Nona Reimer
John Malcom Elementary School, California
California State University, Fullerton
Sandra Simpkins
Arizona State University
Jon R. Star
Harvard University
Kathryn Wentzel
University of Maryland, College Park
NCER 2007-2003
U.S. DEPARTMENT OF EDUCATION
This report was prepared for the National Center for Education Research, Institute of Education
Sciences, under contract no. ED-05-CO-0026 to Optimal Solutions Group, LLC.
Disclaimer
The opinions and positions expressed in this practice guide are the authors’ and do not necessarily
represent the opinions and positions of the Institute of Education Sciences or the U.S. Department
of Education. This practice guide should be reviewed and applied according to the specific needs of
the educators and education agencies using it and with full realization that it represents only one
approach that might be taken, based on the research that was available at the time of publication.
This practice guide should be used as a tool to assist in decision-making rather than as a “cookbook.”
Any references within the document to specific education products are illustrative and do not imply
endorsement of these products to the exclusion of other products that are not referenced.
U.S. Department of Education

Margaret Spellings
Secretary
Institute of Education Sciences

Grover J. Whitehurst
Director
National Center for Education Research

Lynn Okagaki
Commissioner
September 2007
This report is in the public domain. While permission to reprint this publication is not necessary,
the citation should be:
Halpern, D., Aronson, J., Reimer, N., Simpkins, S., Star, J., and Wentzel, K. (2007). Encouraging Girls
in Math and Science (NCER 2007-2003). Washington, DC: National Center for Education Research,
Institute of Education Sciences, U.S. Department of Education. Retrieved from http://ncer.ed.gov.
This report is available for download on the IES website at http://ncer.ed.gov.
Alternative Formats
On request, this publication can be made available in alternative formats, such as Braille, large print,
audiotape, or computer diskette. For more information, call the Alternative Format Center at
(202) 205-8113.
Math and Science
Contents
Preamble from the Institute of Education Sciences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
About the authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Disclosures of potential conflicts of interest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Encouraging girls in math and science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Scope of the practice guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Checklist for carrying out the recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Recommendation 1: Teach students that academic abilities are expandable and improvable. . . . . . . . . 11
Recommendation 2: Provide prescriptive, informational feedback. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Recommendation 3: Expose girls to female role models who have succeeded in math and science. . . . 19
Recommendation 4: Create a classroom environment that sparks initial curiosity

and fosters long-term interest in math and science. . . . . . . . . . . . . . . . . . . . . . . . . 23
Recommendation 5: Provide spatial skills training.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Appendix: Technical information on the studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
( iii )
List of Tables and Figures
Tables
Table 1. Institute of Education Sciences Levels of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Table 2. Recommendations and corresponding Levels of Evidence to support each . . . . . . . . . . . . . . . 7
Table 3. Percent of degrees awarded to women in engineering subfields in 1966, 1985, and 2004 . . . . 20
Figures
Figure 1. Percent of degrees awarded to women by major field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Figure 2. Percent of public high school graduates who completed various mathematics
and science courses in high school, by gender: 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Figure 3. NAEP mathematics scores by highest course completed and gender: 2005 . . . . . . . . . . . . . . . 5
Figure 4. NAEP science scores by highest course completed and gender: 2005 . . . . . . . . . . . . . . . . . . . 5
( iv )
Encouraging Girls in Math and Science
Preamble from the Practice guides can also be distinguished from

systematic reviews or meta-analyses, which employ
Institute of Education statistical methods to summarize the results of studies
obtained from a rule-based search of the literature.
Sciences Authors of practice guides seldom conduct the types of
systematic literature searches that are the backbone of a
meta-analysis, though they take advantage of such work
What is a practice guide? when it is already published. Instead they use their
The health care professions have embraced a mecha- expertise to identify the most important research with
nism for assembling and communicating evidence- respect to their recommendations, augmented by a
based advice to practitioners about care for specific search of recent publications to assure that the research
clinical conditions. Variously called practice guidelines, citations are up-to-date. Further, the characterization
treatment protocols, critical pathways, best-practice of the quality and direction of evidence underlying
guides, or simply practice guides, these documents are a recommendation in a practice guide relies less on a
systematically developed recommendations about the tight set of rules and statistical algorithms and more on
course of care for frequently encountered problems, the judgment of the authors than would be the case in
ranging from physical conditions such as foot ulcers to a high-quality meta-analysis. Another distinction is that
socioemotional issues such as navigating the demands a practice guide, because it aims for a comprehensive
of adolescence.1 and coherent approach, operates with more numerous
and more contextualized statements of what works than
Practice guides are similar to the products of typical does a typical meta-analysis.
expert consensus panels in that they reflect the views
of those serving on the panel, as well as the social Thus practice guides sit somewhere between consensus
decision processes that come into play as the positions reports and meta-analyses in the degree to which
of individual panel members are forged into statements systematic processes are used for locating relevant
that all are willing to endorse. However, practice guides research and characterizing its meaning. Practice guides
are generated under three constraints that do not are more like consensus panel reports than meta-
typically apply to consensus panels. The first is that a analyses in the breadth and complexity of the topic
practice guide consists of a list of discrete recommenda- that is addressed. Practice guides are different from
tions that are intended to be actionable. The second is both consensus reports and meta-analyses in providing
that those recommendations taken together are intend- advice at the level of specific action steps along a
ed to comprise a coherent approach to a multifaceted pathway that represents a more-or-less coherent and
problem. The third, which is most important, is that comprehensive approach to a multifaceted problem.
each recommendation is explicitly connected to the
level of evidence supporting it (e.g., strong, moderate, Practice guides in education at the Institute
and low). The levels of evidence are usually constructed of Education Sciences
around the value of particular types of studies for
drawing causal conclusions about what works. Thus The Institute of Education Sciences (IES) publishes
one typically finds that the top level of evidence is practice guides in education to bring the best avail-
drawn from a body of randomized controlled trials, able evidence and expertise to bear on the types of
the middle level from well-designed studies that do not systemic challenges that cannot be addressed by single
involve randomization, and the bottom level from the interventions or approaches. Although IES has taken
opinions of respected authorities. Levels of evidence advantage of the history of practice guides in health
can also be constructed around the value of particular care to provide models of how to proceed in education,
types of studies for other goals, such as the reliability education is different from health care in ways that
and validity of assessments. may require that practice guides in education have
somewhat different designs. Even within health care,
where practice guides now number in the thousands,
1
Field and Lohr (1990).
(v)
PRACTICE GUIDE
there is no single template in use. Rather, one finds carried out independently of the office and staff within
descriptions of general design features that permit IES that initiated the practice guide.
substantial variation in the realization of practice
guides across subspecialties and panels of experts.2 Because practice guides depend on the expertise of their
Accordingly, the templates for IES practice guides authors and their group decision-making, the content
may vary across practice guides and change over of a practice guide is not and should not be viewed as
time and with experience. a set of recommendations that in every case depends
on and flows inevitably from scientific research. It
The steps involved in producing an IES-sponsored is not only possible but also likely that two teams of
practice guide are first to select a topic, which is recognized experts working independently to produce a
informed by formal surveys of practitioners and practice guide on the same topic would generate
spontaneous requests from the field. Next, a panel chair products that differ in important respects. Thus con-
is recruited who has a national reputation and up-to- sumers of practice guides need to understand that they
date expertise in the topic. Third, the chair, working in are, in effect, getting the advice of consultants. These
collaboration with IES, selects a small number of panel- consultants should, on average, provide substantially
ists to co-author the practice guide. These are people better advice than educators might obtain on their
the chair believes can work well together and have the own because the authors are national authorities who
requisite expertise to be a convincing source of recom- have to achieve consensus among themselves, justify
mendations. IES recommends that at least one of the their recommendations in terms of supporting
panelists be a practitioner with considerable experience evidence, and undergo rigorous independent peer
relevant to the topic being addressed. The chair and the review of their product.
panelists are provided a general template for a practice
guide along the lines of the information provided in
this preamble. They are also provided with examples of
practice guides. The practice guide panel works under
a short deadline of 6 to 9 months to produce a draft
document. The expert panel interacts with and receives
feedback from staff at IES during the development of
the practice guide, but the panel members understand
that they are the authors and thus responsible for the
final product.
One unique feature of IES-sponsored practice guides is

that they are subjected to rigorous external peer review
through the same office that is responsible for indepen-
dent review of other IES publications. A critical task of
the peer reviewers of a practice guide is to determine
whether the evidence cited in support of particular
recommendations is up-to-date, and that studies
of similar or better quality that point in a different
direction have not been ignored. Peer reviewers also are
asked to evaluate whether the evidence grades assigned
to particular recommendations by the practice guide
authors are appropriate. A practice guide is revised as
necessary to meet the concerns of external peer reviews
and to gain the approval of the standards and review
staff at IES. The process of external peer review is
2
American Psychological Association (2002).
( vi )
About the authors

Diane F. Halpern is the Director of the Berger
Institute for Work, Family, and Children, and Professor
of Psychology at Claremont McKenna College. She
earned her Ph.D. in Psychology from the University
of Cincinnati. Her research focuses on the development
of critical thinking skills and on gender differences.
Joshua Aronson is an associate professor of Develop-

mental, Social, and Educational Psychology at New
York University. He earned his Ph.D. in Psychology
from Stanford University. His research focuses on
“stereotype threat” and minority student achievement.
Nona Reimer has been an elementary school teacher

in Capistrano Unified School District, California, for
20 years. She is also a part-time faculty member in Disclosures of potential
education at the California State University at Fuller-
ton. She earned her B.A. in History and Environmental
conflicts of interest
Studies from the University of California, Santa Practice guide panels are composed of individuals
Barbara and her Multiple Subjects Credential from who are nationally recognized experts on the topics
California State University, Hayward. She currently about which they are rendering recommendations.
teaches fifth grade at John S. Malcom Elementary IES expects that such experts will be involved profes-
School, a National Blue Ribbon School. sionally in a variety of matters that relate to their work
as a panel. Panel members are asked to disclose their
Sandra Simpkins is an assistant professor in the
professional involvements and to institute deliberative
School of Social and Family Dynamics at Arizona State
processes that encourage critical examination of the
University. She earned her Ph.D. in Psychology from
views of panel members as they relate to the content
the University of California, Riverside. Her research
of the practice guide. The potential influence of panel
focuses on children’s pursuits, such as their participa-
members’ professional engagements reported by each
tion in after-school activities or enrollment in an
panel member is further muted by the requirement that
elective science course.
they ground their recommendations in evidence that
Jon R. Star is an assistant professor in the Graduate is documented in the practice guide. In addition, the
School of Education at Harvard University. He practice guide is subjected to independent external peer
earned a Ph.D. in Education and Psychology from the review prior to publication, with particular focus on
University of Michigan. Star is a former middle and whether the evidence related to the recommendations
high school mathematics teacher. His research focuses in the practice guide has been appropriately presented.
on students’ learning of middle and secondary school
The professional engagements reported by each panel
mathematics.
member that appear most closely associated with the
Kathryn Wentzel is a professor of Human Develop- panel recommendations are noted below.
ment at the University of Maryland, College Park. She
Dr. Aronson has developed interventions aimed at
earned a Ph.D. in Psychological Studies in Education
changing students’ conceptions of intelligence in
from Stanford University. Her research focuses on
order to improve their academic achievement. He
teachers, peers, and parents as social motivators of
has also developed interventions to reduce stereotype
young adolescents’ classroom behavior and academic
threat. These interventions are not referenced in the
accomplishments.
practice guide.
( vii )
Introduction assurance to the target population because, for example,

the findings have not been sufficiently replicated, or
The goal of this practice guide is to formulate specific (b) studies that are generalizable but have more causal
and coherent evidence-based recommendations that ambiguity than offered by experimental designs, for
educators can use to encourage girls in the fields of example, statistical models of correlational data or
math and science. The target audience is teachers group comparison designs where equivalence of the
and other school personnel with direct contact with groups at pretest is uncertain. For assessments, moder-
students, such as coaches, counselors, and principals. ate refers to high quality studies from a small number
The practice guide includes specific recommendations of samples that are not representative of the whole
for educators and the quality of evidence that supports population. Low refers to expert opinion based on
these recommendations. reasonable extrapolations from research and theory on
other topics and/or evidence from studies that do not
We, the authors, are a small group with expertise meet the standards for moderate or strong evidence.
on this topic. The range of evidence we considered
in developing this document is vast, ranging from For each recommendation, we include an appendix
experiments, to trends in the National Assessment of that provides more technical information about the
Educational Progress (NAEP) data, to correlational and studies and our decisions regarding level of evidence for
longitudinal studies. For questions about what works the recommendation. To illustrate the types of studies
best, high-quality experimental and quasi-experimental reviewed, we describe one study in considerable detail
studies, such as those meeting the criteria of the What for each recommendation. Our goal in doing this is to
Works Clearinghouse, have a privileged position. In all provide interested readers with more detail about the
cases, we pay particular attention to findings that are research designs, the intervention components, and
replicated across studies. how impact was measured. By including a particular
study, we do not mean to suggest that it is the best
Although we draw on evidence about the effectiveness study reviewed for the recommendation or necessarily
of specific practices, we use this information to make an exemplary study in any way.
broader points about improving practice. In this
document, we have tried to take findings from research
or practices recommended by experts and describe
how the use of this recommendation might actually
unfold in school settings. In other words, we aim to
provide sufficient detail so that educators will have
a clear sense of the steps necessary to make use of the
recommendation.
A unique feature of practice guides is the explicit and

clear delineation of the quality and quantity of evidence
that supports each claim. To this end, we adapted a
semi-structured hierarchy suggested by the Institute
of Education Sciences. This classification system helps
determine whether the quality and quantity of available
evidence in support of a practice is of strong, moderate,
or low quality. This system appears in table 1 below.
Strong refers to consistent and generalizable evidence

that an approach or practice causes improved
performance in math or science among girls or that an
assessment is reliable and valid. Moderate refers either
to evidence from (a) studies that allow strong causal
conclusions but which cannot be generalized with
(1)
PRACTICE GUIDE
Table 1. Institute of Education Sciences Levels of Evidence

In general, characterization of the evidence for a recommendation as strong requires both studies
with high internal validity (i.e., studies whose designs can support causal conclusions), as well as
studies with high external validity (i.e., studies that in total include enough of the range of participants
and settings on which the recommendation is focused to support the conclusion that the results
can be generalized to those participants and settings). Strong evidence for this practice guide is
operationalized as:
• A systematic review of research that generally meets the standards of the What Works
Clearinghouse (see http://ies.ed.gov/ncee/wwc/) and supports the effectiveness of a program,
practice, or approach with no contradictory evidence of similar quality; OR
Strong
• Several well-designed, randomized, controlled trials or well-designed quasi-experiments that
generally meet the standards of the What Works Clearinghouse and support the effectiveness of a
program, practice, or approach, with no contradictory evidence of similar quality; OR
• One large, well-designed, randomized, controlled, multisite trial that meets the standards of the
What Works Clearinghouse and supports the effectiveness of a program, practice, or approach,
with no contradictory evidence of similar quality; OR
• For assessments, evidence of reliability and validity that meets the Standards for Educational and
Psychological Testing.3
In general, characterization of the evidence for a recommendation as moderate requires studies

with high internal validity but moderate external validity, or studies with high external validity but
moderate internal validity. In other words, moderate evidence is derived from studies that support
strong causal conclusions but where generalization is uncertain, or studies that support the generality
of a relationship but where the causality is uncertain. Moderate evidence for this practice guide is
operationalized as:
• Experiments or quasi-experiments generally meeting the standards of the What Works
Clearinghouse and supporting the effectiveness of a program, practice, or approach with small
sample sizes and/or other conditions of implementation or analysis that limit generalizability, and no
contrary evidence; OR
Moderate • Comparison group studies that do not demonstrate equivalence of groups at pretest and therefore
do not meet the standards of the What Works Clearinghouse but that (a) consistently show
enhanced outcomes for participants experiencing a particular program, practice, or approach and
(b) have no major flaws related to internal validity other than lack of demonstrated equivalence at
pretest (e.g., only one teacher or one class per condition, unequal amounts of instructional time,
highly biased outcome measures); OR
• Correlational research with strong statistical controls for selection bias and for discerning influence
of endogenous factors and no contrary evidence; OR
• For assessments, evidence of reliability that meets the Standards for Educational and
Psychological Testing 4 but with evidence of validity from samples not adequately representative of
the population on which the recommendation is focused.
In general, characterization of the evidence for a recommendation as low means that the
recommendation is based on expert opinion derived from strong findings or theories in related
Low areas and/or expert opinion buttressed by direct evidence that does not rise to the moderate or
strong levels. Low evidence is operationalized as evidence not meeting the standards for the
moderate or high levels.
3
American Educational Research Association, American Psychological Association, and National Council on Measurement in Education
(1999).
4
Ibid.
(2)
Encouraging girls in chemistry, 46 percent in mathematics, 25 percent

in computer sciences, 22 percent in physics, and
math and science 21 percent in engineering.7 In general, women earn
substantial proportions of the bachelor’s degrees in
math and the sciences, except in computer sciences,
Overview physics, and engineering. At the master’s level, women
Although there is a general perception that men do earned 59 percent of all master’s degrees. The pattern
better than women in math and science, researchers at the master’s degree level is similar (see figure 1). At
have found that the differences between women’s and the doctoral level, however, gender imbalances become
men’s math- and science-related abilities and choices more prevalent, including in math and chemistry (see
are much more subtle and complex than a simple figure 1). Women earned 45 percent of all doctoral
“men are better than women in math and science.” 5 degrees, but they earn less than one-third of all doctoral
In fact, experts disagree among themselves on the degrees in chemistry, computer sciences, math, physics,
degree to which women and men differ in their and engineering.8 In contrast, women earn 67 percent
math- and science-related abilities.6 A quick review of of the doctoral degrees in psychology and 44 percent
the postsecondary paths pursued by women and men in other social sciences.9 This disproportionate repre-
highlight the areas in math and science where women sentation in math and science graduate degrees is also
are not attaining degrees at the same rate as men. reflected in math and science career pathways. While
women make up nearly half of the U.S. workforce,
In 2004, women earned 58 percent of all bachelor’s they make up only 26 percent of the science and
degrees, 78 percent of bachelor’s degrees in psychol- engineering workforce.10 The question many are asking
ogy, 62 percent in biological sciences, 51 percent in is why women are choosing not to pursue degrees
Figure 1. Percent of degrees awarded to women by major field
90 Bachelor’s
Master’s
80 77.8 78.1
Doctorate
70 67.3
62.5
59.1 58.6
60 57.6 54.5 55.9
51.1
50 46.3 45.9 45.4
45.4 46.2
44
40
31.7 31.2 25.2
30 28.4
25.1 21.8 20.5 21.1 17.6
20.5
20 15.5
10
0
All Fields Biological Chemistry Computer Mathematics Physics Engineering Psychology Social Sciences
Sciences Science
SOURCE: National Science Foundation, Science and Engineering Degrees: 1966-2004.
5
See Hyde (2005), Spelke (2005), and Halpern (2000) for recent discussions of the literature.
6
See Gallagher and Kaufman (2005) for a collection of chapters representing different researchers’ views.
7
National Science Foundation (2006a, 2006b).
8
Ibid.
9
Ibid.
10
National Science Foundation (2006d).
(3)
PRACTICE GUIDE
and careers in the physical sciences, engineering, or likely to complete biology (girls, 93 percent; boys, 89
computer science. Several potential reasons for the percent), advanced placement (AP) or honors biology
gender disparity include previous coursework, ability, (girls, 19 percent; boys, 14 percent), and chemistry
interests, and beliefs. (girls, 66 percent; boys, 58 percent) than boys were.
Although some gender differences are present in high
An examination of course-taking patterns shows that school math and science course enrollments, similari-
girls are taking math and science courses in high ties between the genders is also common. This gender
school. On the 2005 National Assessment of Educa- parity in course-taking patterns may be less surprising
tional Progress (NAEP) High School Transcript Study, than it appears, given that high school graduation
girls who graduated from high school, on average, requirements typically include multiple science courses,
earned slightly more credits in mathematics and science as well as mathematics.
(7.3) than boys earned (7.1). Boys, however, earned
slightly more credits in computer-related courses (1.1) A second reason for the observed differences in college
than girls earned (0.8).11 Figure 2 shows the percent- and occupational choices may be that males and
ages of female and male high school graduates in 2000 females have variant math and science abilities, as
that completed math and science courses. Although measured by standardized tests. Although girls gener-
a greater percentage of boys completed physics (34 ally do as well as, or better than, boys on homework
percent) and calculus (12 percent) than girls (physics, assignments and course grades in math and science
29 percent; calculus, 11 percent), girls were more classes,12 boys tend to outscore girls when tested on
Figure 2. Percent of public high school graduates who completed various mathematics and science courses in
high school, by gender: 2000
100
90
80
70
60
50
40
30
20
10
11
Shettle, Roey, Mordica, et al. (2007).
12
College Board (2006, August 29); Shettle, Roey, Mordica, et al. (2007).
(4)
Figure 3. NAEP mathematics scores by highest course completed and gender: 2005
300
250
200
150
100
50
Figure 4. NAEP science scores by highest course completed and gender: 2005
300
250
200
150
100
50
(5)
PRACTICE GUIDE
the same content in high-pressure situations, such as in math or science careers.16 This gender difference is
standardized tests with time limits. These tests are typi- interesting, and somewhat puzzling, given that males
cally not linked to instructed curriculum, and so can be and females generally enroll in similar courses and
understood to be measures of more general abilities in display similar abilities (at least as measured by course
math and science.13 For example, on the 2005 NAEP grades). In other words, girls, particularly as they move
math and science assessments, girls scored lower than out of elementary school and into middle and high
boys when controlling for highest course completed at school and beyond, often underestimate their abilities
all levels, except the lowest level (see figures 3 and 4).14 in mathematics and science. However, it is important
Performance differences on timed standardized tests to note that not all girls have less confidence and inter-
do not necessarily mean that girls are not as capable as est in mathematics and science, and that girls, as well
boys in math or science. Researchers have found, for as boys, who have a strong self-concept regarding their
instance, that SAT math scores underpredict young abilities in math or science are more likely to choose
women’s performance in college math courses.15 This and perform well in elective math and science courses
suggests that it is not ability, per se, that hinders girls and to select math- and science-related college majors
and women from pursuing careers in math and science. and careers.17 This is noteworthy because it suggests
If not ability, then what? that improving girls’ beliefs about their abilities could
alter their choices and performance. Theory and
Areas where consistent gender differences have emerged empirical research suggest that children’s beliefs about
are children’s and adolescents’ beliefs about their their abilities are central to determining their interest
abilities in math and science, their interest in math and performance in different subjects, the classes they
and science, and their perceptions of the importance choose to take, the after-school activities they pursue,
of math and science for their futures. In general, and, ultimately, the career choices they make.18
researchers have found that girls and women have less
confidence in their math abilities than males do and What can teachers do to encourage girls to choose
that from early adolescence, girls show less interest career paths in math- and science-related fields? One
major way to encourage girls to choose careers in
math and science is to foster the development of
strong beliefs about their abilities in these subjects–
beliefs that more accurately reflect their abilities –
and more accurate beliefs about the participation of
women in math- and science-related careers (see table
2). Our first two recommendations, therefore, focus
on strategies that teachers can use to strengthen girls’
beliefs regarding their abilities in math and science: (1)
Teach students that academic abilities are expandable
and improvable (Level of Evidence: Moderate); and
(2) Provide prescriptive, informational feedback (Level
of Evidence: Moderate). Our third recommendation
addresses girls’ beliefs about both their own abilities
and the participation of women in math- and science-
related careers: (3) Expose girls to female role models
who have succeeded in math and science (Level of
Evidence: Low).
13
See Halpern, Benbow, Geary, et al. (2007) for a more thorough discussion of this point.
14
Shettle, Roey, Mordica, et al. (2007).
15
Wainer and Steinberg (1992).
16
Andre, Whigham, Hendrickson, et al. (1999); Herbert and Stipek, (2005); Jacobs, Lanza, Osgood, et al. (2002); Simpkins and Davis-Kean
(2005); Wigfield, Eccles, Mac Iver, et al. (1991).
17
Simpkins and Davis-Kean (2005); Updegraff and Eccles (1996).
18
Pajares (2006).
(6)
Table 2. Recommendations and corresponding Levels of Evidence to support each

Recommendation Level of Evidence
1. Teachers should explicitly teach students that academic abilities are expandable
and improvable in order to enhance girls’ beliefs about their abilities. Students
who view their cognitive abilities as fixed from birth or unchangeable are more
likely to experience decreased confidence and performance when faced with
difficulties or setbacks. Students who are more confident about their abilities in Moderate
math and science are more likely to choose elective math and science courses
in high school and more likely to select math and science-related college
majors and careers.
2. Teachers should provide students with prescriptive, informational feedback

regarding their performance. Prescriptive, informational feedback focuses
on strategies, effort, and the process of learning (e.g., identifying gains in
children’s use of particular strategies or specific errors in problem solving). Moderate
Such feedback enhances students’ beliefs about their abilities, typically
improves persistence, and improves performance on tasks.
3. Teachers should expose girls to female role models who have achieved in math
or science in order to promote positive beliefs regarding women’s abilities in
math and science. Even in elementary school, girls are aware of the stereotype
that men are better in math and science than women are. Exposing girls to Low
female role models (e.g., through biographies, guest speakers, or tutoring by
older female students) can invalidate these stereotypes.
4. Teachers can foster girls’ long-term interest in math and science by choosing
activities connecting math and science activities to careers in ways that do
not reinforce existing gender stereotypes and choosing activities that spark
initial curiosity about math and science content. Teachers can provide ongoing Moderate
access to resources for students who continue to express interest in a topic
after the class has moved on to other areas.
5. Teachers should provide opportunities for students to engage in spatial skills

training. Spatial skills training is associated with performance in mathematics Low
and science.
(7)
PRACTICE GUIDE
Girls are more likely to choose courses and careers

in math and science if their interest in these fields
Scope of the
is sparked and cultivated throughout the school practice guide
years.19 Our fourth recommendation focuses on the
importance of fostering long-term interest (Level of This practice guide provides five recommendations for
Evidence: Moderate) and provides concrete strategies encouraging girls in math and science. These recom-
that teachers might use to promote greater interest in mendations together form a coherent statement: To
math and science. encourage girls in math and science, we need to begin
first with their beliefs about their abilities in these areas,
A final way to encourage girls in math and science second with sparking and maintaining greater interest
is to help them build the spatial skills that are crucial in these topics, and finally with building associated
to success in many math- and science-related fields, skills. Our specific recommendations cover these three
such as physics, engineering, architecture, geometry, domains in a representative but not exhaustive way. In
topology, chemistry, and biology. Research suggests particular, we have chosen to focus on specific recom-
that spatial skills, on which boys have typically outper- mendations that have the strongest research backing
formed girls, can be improved through specific types available. In addition, we limit our focus to recom-
of training. Thus, our final recommendation is that mendations that teachers can carry out in the classroom
teachers provide students, especially girls, with specific and that do not require systemic change within a
training in spatial skills (Level of Evidence: Low). school district. We remind the reader that students’
choices to pursue careers in math and science reflect
multiple influences that accumulate over time. We
have identified practices that elementary, middle, and
high school teachers can implement during instruction
that we believe would increase the likelihood that girls
and women will not prematurely decide that careers in
math and science are not for them.
19
Wigfield, Eccles, Schiefele, et al. (2006).
(8)
Checklist for carrying out Provide girls and young women with information
about mentoring programs designed to support
the recommendations students who are interested in mathematics and
science.
Recommendation 1: Teach students that Encourage parents to take an active role in provid-
academic abilities are expandable and ing opportunities for girls to be exposed to women
improvable. working in the fields of math and science.
Teach students that working hard to learn new

Recommendation 4: Create a classroom
knowledge leads to improved performance.
environment that sparks initial curiosity
Remind students that the mind grows stronger and fosters long-term interest in math and
with use and that over time and with continued science.
effort, understanding the material will get easier.
Embed mathematics word problems and science
activities in contexts that are interesting to both
Recommendation 2: Provide prescriptive,
boys and girls.
informational feedback.
Provide students with access to rich, engaging
Provide students with feedback that focuses on
relevant informational and narrative texts as they
strategies used during learning, as opposed to
participate in classroom science investigations.
simply telling them whether they got an answer
correct. This strategy encourages students to correct Capitalize on novelty to spark initial interest. That
misunderstandings and learn from their mistakes. is, use project-based learning, group work, innova-
tive tasks, and technology to stir interest in a topic.
Provide students with positive feedback about
the effort they expended on solving a difficult Encourage middle and high school students to
problem or completing other work related to their examine their beliefs about which careers are
performance. typically female-oriented and which are typically
male-oriented. Encourage these students to learn
Avoid using general praise, such as “good job,”
more about careers that are interesting to them
when providing feedback to individual students or
but that they believe employ more members of the
an entire class.
opposite gender.
Make sure that there are multiple opportunities for
Connect mathematics and science activities to
students to receive feedback on their performance.
careers in ways that do not reinforce existing
gender stereotypes of these careers.
Recommendation 3: Expose girls and young
women to female role models who have
Recommendation 5: Provide spatial skills
succeeded in math and science.
training.
Invite older girls and women who have succeeded
Recognize that children may not automatically
in math- or science-related courses and professions
recognize when spatial strategies can be used to
to be guest speakers or tutors in your class.
solve problems and that girls are less likely to
Assign biographical readings about women use spatial strategies than boys. Teach students
scientists, mathematicians, and engineers, as part of to mentally image and draw spatial displays in
students’ assignments. response to mathematics and science problems.
Call attention to current events highlighting the Require students to answer mathematics and
achievements of women in math or science. science problems using both verbal responses and
spatial displays.
When talking about potential careers, make
students aware of the numbers of women who Provide opportunities for specific training in spatial
receive advanced degrees in math- and science- skills such as mental rotation of images, spatial
related disciplines. perspective, and embedded figures.
(9)
( 10 )
Recommendation 1: Teach students that academic abilities

are expandable and improvable.
To enhance girls’ beliefs about their abilities, we recommend that
teachers understand and communicate this understanding to students:
Math and science abilities—like all abilities—can be improved through
consistent effort and learning. Research shows that even students
with considerable ability who view their cognitive abilities as fixed or
unchangeable are more likely to experience greater discouragement,
lower performance, and, ultimately, reduce their effort when they
encounter difficulties or setbacks. Such responses may be more likely in
the context of math, given stereotypes about girls’ innate mathematics
abilities.20 In contrast, students who tend to view their abilities as
expandable tend to keep trying in the face of frustration in order to
increase their performance. To help girls and young women resist negative reactions to the difficulty of
math and science work, it can be very helpful for them to learn that their math and science abilities can
improve over time with continuous effort and engagement.
Level of evidence: Moderate or ability is that it is not fixed but can be improved
through hard work and effort. Research shows that
The panel judges the level of evidence supporting this children who view intelligence as a fixed trait or believe
recommendation to be moderate, based on the two that their own abilities cannot be changed tend to
small experimental studies that examined the effect of pursue “performance goals.” That is, they tend to be
this practice for improving K–12 students’ performance more concerned with demonstrating their intelligence
on math,21 one experimental study that examined the and prefer to complete tasks that will show that they
effect of this strategy for improving college students’ are “smart.” In contrast, students who believe that intel-
general academic performance,22 and supporting cor- ligence or ability can be improved with effort are more
relational research demonstrating the relation between likely to pursue learning goals.25 That is, they tend to
students’ beliefs about the stability and malleability of be more concerned with learning new material and are
intellectual abilities and their performance.23 more likely to seek to master difficult material, even if
doing so does not make them look “smart” (e.g., they
Brief summary of evidence to support might not be able to solve problems initially). When
the recommendation tasks become more challenging, students who believe
that abilities or intelligence cannot be changed are more
What students believe about the nature of intelligence likely to become anxious, downgrade their assessment
and ability affects their achievement.24 Some students of their ability, and give up. However, students who
believe that people, in general, are born with a fixed believe that abilities can be improved through effort
amount of intelligence, such that some people are born and hard work are more likely to respond to challenge
smart and others less smart and that little can be done with increased effort. In the long run, the students who
to change this. Similarly, some students believe that are able to persist in their attempts to master difficult
their abilities were determined at birth and cannot be material perform better than the students who doubt
changed. An alternative way to think about intelligence their ability and give up.26
20
See Dweck (2006) for a recent discussion focused on girls’ beliefs about intelligence.
21
Good, Aronson, and Inzlicht (2003); Blackwell, Trzesniewski, and Dweck (2007).
22
Aronson, Fried, and Good (2002).
23
Weiner (1986); Graham (1991); see Dweck (1999) for an overview of research in this area.
24
Grant and Dweck (2003).
25
Dweck (1999); Dweck and Leggett (1988).
26
Utman (1997).
( 11 )
PRACTICE GUIDE
These different orientations toward learning may mind grows stronger as a result of doing
have long-term implications. For example, if a child’s hard work, and that over time and with
goal is to look smart, she may shy away from challeng- continued effort, understanding the material
ing tasks with potential for failure in favor of easier and solving the problems will get easier.29
tasks with higher potential for success. In addition,
if a child believes that intelligence or abilities are fixed, • Teachers can also remind students about the
then she is likely to attribute failure to lack of ability, malleability of intelligence when they make
and her belief in her own abilities may eventually progress, pointing out that their brains are
decline. Thus, failures or challenges can have a negative actively building new connections as they study.
impact on children who view intelligence or abilities
as a fixed trait. In contrast, research shows that when Potential roadblocks and solutions
students are taught that intelligence and abilities can
be increased with hard work, their test scores and their Roadblock 1.1. Some adults may believe that
grades improve.27 intelligence and abilities are innate or fixed and that
people who are “naturally” good at something will
Finally, why is it important to foster girls’ belief in excel in that domain. It can be difficult to convince
the malleability of intellectual abilities? As discussed students that effort will make a significant difference
in the overview, girls tend to lack confidence in their when some adults seem to favor “natural ability”
math and science abilities even when they do well in explanations for success over “effort and hard work”
their math and science courses. Teaching girls that explanations.
knowledge and intellectual skills increase, for example,
Solution. Some teachers may view abilities and
when students learn how to solve problems that they
intelligence as static characteristics that are fixed at
previously could not do explicitly provides girls with a
birth. Neurologists used to believe this as well, but
way to interpret failure that does not discourage them
they no longer do. Thus, teachers as well as students
from persevering to master new material in class.
need to take the neuroscience seriously and examine
and modify their preconceptions about the nature of
How to carry out the recommendation human intelligence. Only then can they genuinely help
both students and parents understand that our brains
To help modify students’ beliefs about their intelligence
are constantly creating and refining new synaptic
or abilities, teachers can:
connections, based on our experiences and the activi-
• Expose students to and discuss the neuroscience ties we regularly practice. Teachers should consider
research that shows that brains grow new synaptic the following two studies. In one, researchers found
connections when new material is learned and that cab drivers had enlarged portions of the part of
practiced, thus making the brain more complex the brain that is important in performing spatial tasks
and “smarter”—that working hard to learn new (right posterior hippocampus) relative to a control
knowledge leads to improved intelligence. Sports group of adults whose employment required less use of
analogies can buttress this learning: practicing spatial navigational skills.30 Among the cab drivers in
academic skills, like solving math problems, this study, the number of years spent driving taxis was
improves performance much like practicing positively correlated with the size of the right posterior
free throws improves basketball performance hippocampus. In a related study, researchers found
or practicing serves helps one’s tennis game.28 that in a sample of adults who were not taxi drivers,
there was no correlation between size of the posterior
• When students are struggling, teachers can hippocampus and navigational expertise.31 These two
explicitly remind their students that the sets of findings suggest that experience with complex
27
Aronson, Fried, and Good (2002); Good, Aronson, and Inzlicht (2003); Blackwell, Trzesniewski, and Dweck (2007).
28
Good, Aronson, and Inzlicht (2003).
29
Ibid.
30
Maguire, Gadian, Johnsrude, et al. (2000).
31
Maguire, Spiers, Good, et al. (2003).
( 12 )
route finding caused the increased size of the relevant

brain structure. This is just one example of research
that underlines the fundamental flexibility of the
human brain in creating new synaptic connections for
tasks that are repeated and practiced over time.
Roadblock 1.2. By the time students enter high

school, some girls perceive themselves to be less capable
than boys in math and science. They may believe
that their abilities in this domain are not significantly
expandable or that they are innately less likely than
boys to do well in these domains.
Solution. Teachers will need to keep in mind that girls

perform as well as or even better than boys in school,
including on exams and course grades in math and
science classes; boys outperform girls only when we
look at scores on advanced standardized tests. Experts
disagree as to why girls’ equal or superior classroom
performance in math and science does not carry over
to their performance on high-stakes standardized
tests.32 Thus, teachers can emphasize to students that
high scores on advanced standardized tests do not
in the long run determine success in science- and
math-related fields. Many different skills are needed
for success in these domains, including the content
knowledge gained in coursework and through experi-
ence, as well as excellent writing and communication
skills. There is no reason to believe that girls are
biologically programmed to perform less well than
boys in math- and science-related careers, and there are
many reasons to believe that women, once rare in math
and science careers, will continue to close the gap, as
they have for the past several decades. Again, the best
way for teachers to respond with genuine encourage-
ment to girls’ doubts about their innate aptitudes in
math and science is to be clear that the male advantage
in math has little grounding in science and is limited
to standardized tests. Appreciating the fact that girls
perform better on standardized math tests when they
believe that their math abilities are not fixed is one
powerful way to start.
32
Gallagher and Kaufman (2005).
( 13 )
( 14 )
Recommendation 2: Provide prescriptive, informational feedback.

We recommend that teachers provide students with prescriptive,
informational feedback regarding their performance in math and
science courses. Prescriptive, informational feedback focuses on
strategies, effort, and the process of learning. Examples include
identifying gains in children’s strategy use, praising effort, or identify-
ing gaps or errors in problem-solving. Although this type of feedback
overlaps with the type of feedback that teachers provide during forma-
tive assessment, this recommendation specifically targets feedback
that focuses students’ attention on their beliefs about why they did or
did not perform well on a particular task. Prescriptive, informational
feedback enhances students’ beliefs about their abilities, typically
improves persistence, and improves performance on tasks. In addition, students’ beliefs about their
abilities are related to their math- and science-related choices.33
Level of evidence: Moderate a response. However, all forms of feedback are not
equal in their impact on students’ beliefs about their
The panel judges the quality of the evidence on the abilities in a given domain, such as math or science, nor
relation between prescriptive, informational feedback in their impact on performance. In particular, when
and students’ beliefs about their math and science teachers provide specific, informational feedback in
abilities and their performance on math- and science- terms of strategies, effort, and the process of learning
related tasks to be moderate, based on a set of small (e.g., “you worked really hard at that subtraction
experimental studies using random assignment that problem”), rather than general praise (e.g., “good job”)
focus specifically on children performing math or or feedback regarding global intelligence (e.g., “you’re 15
math-related tasks 34 and supporting research on the smart”), students’ beliefs about their abilities and their
effects of different types of feedback on a variety of performance are positively influenced.36
tasks.35 The supporting research on feedback includes
many studies that vary in terms of design, including Many teachers know that providing informational
small experimental studies, longitudinal and cross- feedback helps create a positive learning environment.
sectional correlational studies, and qualitative studies. Indeed, the use of classroom formative assessment is
Many of the experimental studies on the effects of linked to substantial learning gains.37 When teachers
different types of feedback have been conducted give informational feedback (e.g., pointing out to a
with children. student a specific problem in her logic rather than
simply noting that the answer is incorrect) students’
achievement and attitudes improve.38 During whole-
Brief summary of evidence to support
class instruction, when teachers combine positive
the recommendation
comments with specific information about how to
Students often receive feedback regarding their solve a problem, students are less likely to report that
performance in the form of grades, test scores, or they engage in self-defeating behaviors (e.g., putting
statements from teachers regarding the accuracy of off doing their homework until the last minute) or
33
See Hackett (1985) for a classic study supporting this conclusion in the context of mathematics; Fouad and Smith (1996) discuss this
relationship in middle school students.
34
Mueller and Dweck (1998); Elawar and Corno (1985); Miller, Brickman, and Bolen (1975).
35
See Bangert-Drowns, Kulik, Kulik, et al. (1991) for a synthesis of studies on feedback and Henderlong and Lepper (2002) for a recent
review on the effects of praise on children’s intrinsic motivation.
36
Mueller and Dweck (1998); Elawar and Corno (1985); Miller, Brickman, and Bolen (1975).
37
See Black and Wiliam (1998) for a recent discussion of the literature on feedback and formative assessment.
38
Elawar and Corno (1985).
( 15 )
PRACTICE GUIDE
avoid asking for help when they don’t understand abilities? As discussed in the overview, girls tend to
assignments.39 In addition, research suggests that lack confidence in their math and science abilities even
positive substantive feedback that provides information when they do well in math and science courses. Provid-
about students’ progress toward goals and progress in ing informative feedback focuses students’ attention on
learning is related to children’s motivational beliefs, what to do when they do not solve a problem correctly
such as their self-concept of ability and self-efficacy. An rather than letting girls attribute wrong answers to
observational study of math classrooms illustrates how a lack of ability. When students experience success,
including such feedback during instruction can support providing informative feedback directs their attention
students’ self-efficacy in mathematics.40 Even though to what they did to achieve that success (e.g., worked
the research demonstrates the critical and potentially hard, tried multiple strategies, used the procedures in
powerful role that appropriate feedback can play, it the correct order) rather than allowing girls to attribute
does not appear that teachers typically use prescriptive, that success to having a certain amount of ability.
informational feedback. In fact, a recent descriptive
study of teacher feedback used in 58 third-grade How to carry out the recommendation
mathematics classrooms suggests that the primary form
of feedback teachers use during instruction is general What can teachers do to make sure that the feedback
praise, such as “that’s very good,” which does not they give students will help improve both their motiva-
provide any useful information to students.41 tion to learn new material and their performance, even
in the face of failure?
Experimental work suggests that feedback given in
the form of praise focused on global intelligence (e.g., • Provide positive, substantive feedback to students
“you’re smart”) may have a negative impact on future as they solve problems to encourage students to
learning behavior in comparison to praise about effort correct misunderstandings and learn from their
(e.g., “you must have worked hard”).42 Elementary mistakes.43 Teachers should create a classroom
school students who were given praise about their environment in which learning, improving,
intelligence after correctly solving a problem were likely and understanding are emphasized. In such an
to attribute future failures to lack of ability, have lower environment, when children give an incorrect
interest, show less persistence on future tasks, and have answer, it becomes an opportunity for learning.
a goal for future tasks of looking smart. In contrast,
children who were given praise about their effort were • Highlight the importance of effort for succeeding
more likely to believe that subsequent failure was due at difficult tasks. By attributing success to effort
to lack of effort, show higher persistence on difficult rather than to global intelligence, expectations
tasks, and have a goal of mastering challenging tasks for future success are supported. Praising general
or concepts rather than just “looking smart.” Thus, intelligence implies that natural intellectual gifts
teacher feedback that attributes student success to effort determine success (and failure) rather than effort;
(e.g., “you’ve been working hard”) and task-specific this can be a debilitating mindset for students
ability (e.g., “you did very well at solving this division when confronted with failure on a difficult task.44
problem”) strengthens self-efficacy beliefs about
• Keep a balance between learning on the one hand
mathematics. These beliefs, in turn, influence a child’s
and performance on the other. Grades matter,
future persistence on difficult tasks and, ultimately,
but students who focus single-mindedly on their
overall performance.
grades may come to care so much about perfor-
Finally, why is prescriptive, informational feedback mance that they sacrifice learning opportunities.45
important to enhancing girls’ beliefs about their
39
Turner, Midgley, Meyer, et al. (2002).
40
Schweinle, Turner, and Meyer (2006).
41
Foote (1999).
42
Mueller and Dweck (1998).
43
Schweinle, Turner, and Meyer (2006); Turner, Midgley, Meyer, et al. (2002).
44
Mueller and Dweck (1998); see Foersterling (1985) for a review of research on attributional feedback.
45
Dweck (2002); Mueller and Dweck (1998).
( 16 )
Potential roadblocks and solutions so that appropriate adjustments can be made through-
out the year. Lessons should include formative and
Roadblock 2.1. Some teachers may find it difficult to summative assessments of student progress, providing
focus on effort and strategy use rather than on perfor- feedback to students long before the annual standard-
mance. Too often, the attention of many students (and ized assessments are taken. Teachers can also use peer
sometimes their parents) is on report card grades and feedback and critique as a classroom activity—provid-
exam scores. In addition, many teachers are required ing clear criteria for feedback and critique to students
to assess and report performance in terms of grades or at the outset.
exam scores.
Solution. Teachers can draw attention to students’

efforts when possible. When explaining exam scores or
grades on an assignment, teachers can provide com-
ments on effort and strategy. Teachers can routinely
comment on the combined efforts of a class as students
are working on assignments or projects.46 Feedback
specific to individual students is best delivered in a
one-on-one context.47 Teachers can also design assign-
ments that reward effort. For example, students can be
encouraged to submit drafts on which feedback can be
given and then revised versions submitted for a grade.
Roadblock 2.2. Teachers whose schedules are already

stretched may find it difficult in the course of the aver-
age school day to give each student detailed feedback
on problem-solving and strategy use.
Solution. Feedback or praise does not need to be

given all the time.48 In fact, informative feedback,
and particularly praise focused on effort, should be
given only when it is genuine. Giving students praise
on simple tasks may undermine motivation. When
praise is warranted, teachers can focus on effort,
using phrases such as, “you worked really hard.”
Teachers can be strategic in when and how they
provide detailed informative feedback. For example,
it often is appropriate to give such feedback to an entire
class after a test or exam, especially when most students
make a specific error. A class review after an assignment
or test also is a good way to provide all students with
informative feedback.
Roadblock 2.3. Many teachers rely heavily on stan-

dardized assessment techniques, which provide little
feedback and can foster a performance rather than a
learning orientation regarding scores.
Solution. Effective math and science programs provide

continual, multiple assessments of student knowledge
46
Turner, Midgley, Meyer, et al. (2002).
47
Ward (1976).
48
Henderlong and Lepper (2002) provide a recent discussion of the research on praise.
( 17 )
( 18 )
Recommendation 3: Expose girls to female role models who have

succeeded in math and science.
We recommend that teachers expose girls to female role models
who have succeeded in math and science. Research demonstrates
that triggering negative gender stereotypes can create problems for
girls and women on tests of mathematics and spatial reasoning.49
Exposure to female role models who have succeeded in math has
been shown to improve performance on math tests and to invalidate
these stereotypes.50
Level of evidence: Low Brief summary of evidence to support

the recommendation
We rated the level of evidence that supports this
recommendation as low. This recommendation is Researchers have found that negative stereotypes
based on our extrapolation of relevant research, can affect performance in test-taking situations 54
including four small experimental studies with college and have labeled this phenomenon “stereotype threat.”
students.51 Although the experiments that support the Stereotype threat arises from a psychologically threaten-
recommendation have strong research designs (internal ing concern about confirming a negative stereotype,
validity) for supporting causal claims, these studies both in one’s own eyes and in the eyes of others. For
were conducted with college students rather than girls both self and others, the existence of the stereotype
from kindergarten through high school and were short fosters negative beliefs about the meaning of difficulty
laboratory experiments, rather than real-world class- or low performance—namely, that one lacks ability.
room studies conducted with students over extended Thus, when a woman is told that her math abilities
periods of time. Thus, the applicability of these studies are being evaluated, she is likely to perform worse
to the effects of exposing girls to female role models in on a standardized math test than a man with similar
natural contexts (e.g., classrooms) is limited. course grades and performance on assignments, because
of the anxiety, deficits in short-term memory, and
In addition to these studies that explicitly address negative thoughts that have been shown to accompany
the effect of exposure to female role models on stereotype threat. Studies also show that stereotype
young women’s math performance and beliefs about threat can lead young adolescent girls and women to
their math abilities, there is related research that choose unchallenging problems to solve,55 lower their
supports this recommendation. This research includes performance expectations,56 and devalue mathematics
experimental evidence that negative stereotypes can as a career choice.57 Thus, negative stereotypes can
impede performance 52 and one small, cross-sectional impair engagement and confident performance of
observational study showing that children are aware of girls and women in science, technology, engineering,
math-related gender stereotypes.53 and mathematics.
49
Aronson (2002); Aronson and Steele (2005); Steele, Spencer, and Aronson (2002).
50
Marx and Roman (2002); McIntyre, Paulson, and Lord (2003).
51
Ibid.
52
For reviews of the research, see Aronson and Steele (2005); Steele, Spencer, and Aronson (2002).
53
Steele (2003).
54
Several experimental studies have been conducted with college students that demonstrate the stereotype threat phenomenon. For example,
see Gonzales, Blanton, and Williams (2002); Spencer, Steele, and Quinn (1999); Steele and Aronson (1995).
55
Aronson and Good (2002).
56
Stangor, Carr, and Kiang (1998).
57
Davies, Spencer, Quinn, et al. (2002).
( 19 )
PRACTICE GUIDE
Research also indicates that when some women take the math or science performance of girls, experimental
tests, a certain amount of stereotype threat is generally studies demonstrate that calling attention to gender
operative—that is, stereotype threat is the default decreased performance on math tests in 12-year-old
unless measures are taken to counter it in the testing girls but not in younger girls (10- and 11-year-olds).61
situation.58 It is also the case that circumstances can be In addition, researchers have found that by age 10 or
more or less threatening depending on the number of so, simply being evaluated is enough to evoke stereo-
stereotype cues in the environment. For example, when type threat with respect to academically stigmatized
men outnumber women in the room when taking a ethnic minority students.62 These studies demonstrate
test, women perform worse than when they outnumber that stereotype threat can be a problem by the time
men, and women seem to perform better still when no girls reach middle school.
men are present.59
Because the research on the effects of stereotype
Evidence from four random-assignment experiments threat and exposure to positive female role models
indicate that exposing women to female role models has primarily been documented with adult women,
who are high-achieving or who are perceived as math we considered related research on awareness of gender
experts can mitigate the effects of stereotype threat stereotypes by girls in general. Evidence from one
on math test performance. 60 These studies show that small, cross-sectional observational study suggests that
even brief exposure to women who are perceived to be elementary-school-aged girls are aware of the stereotype
experts in math can improve female students’ perfor- that men are considered to be better at math than
mance on math tests. women, but that they view girls and boys to be equally
good at math. 63
Although we did not identify experimental studies test-
ing the effect of exposure to successful role models on
Table 3. Percent of degrees awarded to women in engineering subfields in 1966, 1985, and 2004
Bachelor’s Master’s Doctorate
1966 1985 2004 1966 1985 2004 1966 1985 2004
Aeronautical Engineering .3 8.4 17.8 .8 5.1 17.1 0 4.0 11.9
Chemical Engineering .8 23.4 35.4 .7 15.7 27.7 .5 8.1 23.9
Civil Engineering .4 13.8 24.2 .4 11.7 27.2 0 5.1 19.6
Electrical Engineering .3 11.5 14.2 .6 8.8 19.6 .4 4.9 13.5
Industrial Engineering .4 29.1 33.2 .5 15.5 21.3 0 6.5 19.4
Mechanical Engineering .2 10.5 13.6 .3 7.0 12.4 .2 5.1 11.1
Materials & Metallurgical

.9 22.4 31.2 .8 15.8 24.9 .9 10.6 17.7
Engineering
Other Engineering Subfields .8 16.0 29.9 .8 11.5 24.7 .4 6.3 24.9
SOURCE: National Science Foundation. Division of Science Resources Statistics. 2006. Science and Engineering
Degrees: 1966-2004.
58
E.g., Inzlicht and Ben-Zeev (2000); Inzlicht, Aronson, Good, et al. (2006).
59
Ibid.
60
61
Good and Aronson (2007).
62
McKown and Weinstein (2003).
63
Steele (2003).
( 20 )
Although the negative stereotype that women cannot In addition, we suggest that teachers invite women or
perform as well as men in math and science exists and older students who can serve as role models in math or
can affect students’ performance, the stereotype is not science to be guest speakers or tutors. Learning about
necessarily accurate. As we indicated in the overview, role models who have achieved in math or science,
women earn a substantial proportion of the bachelor’s whether through biographies or personal conversations:
and master’s degrees in math and in many science
disciplines (see figure 2). Even in engineering, the pro- • Teaches students that struggle and eventual
portions of degrees that women now earn have substan- success are normal. This knowledge seems to
tially changed over the last 40 years.64 Table 3 shows the reduce anxiety and boost motivation when
percent of degrees awarded to women by engineering the student encounters challenges in work
subfields in 1966, 1985, and 2004. In 1966, women related to science, technology, engineering, or
earned less than 1 percent of the bachelor’s degrees mathematics. A role model who communicates
in any engineering subfield. By 2004, women earned this may serve as a greater inspiration to
about one-third of the bachelor’s degrees in chemical persist through difficulty than someone for
engineering (35 percent), industrial engineering (33 whom achievement appears effortless.66
percent), and materials and metallurgical engineering
• Conveys to students that becoming good at
(31 percent). Exposing girls to female role models may
math or science takes hard work and that
help negate the stereotype and encourage more girls to
self-doubts are a normal part of the process of
pursue math- and science-related careers.
becoming expert at anything worthwhile.67
How to carry out the recommendation At least some female role models should be “attainable.”
Research supports the idea that older students who
To counteract the negative stereotypes regarding overcame initial difficulty with hard work eventually to
women’s math and science abilities, teachers should become high performers can be effective role models.68
provide exposure to female role models who are experts A famously gifted female engineer for whom math and
in math and science fields. Experimental studies with science always came easily and naturally is not always
college women indicate that learning about women the best role model because she can be written off as
who have achieved success in math or science can help an the exception—a rare case who triumphed through
attenuate the effects of negative stereotypes.65 Teachers talent rather than hard work.
can expose students to female role models in a number
of ways: In addition, many mentoring programs have been
created for young women in an attempt to provide
• Assign biographical readings about women them with role models and foster their interest in
scientists, mathematicians, and engineers. mathematics and science.69 Mentoring is a broad
term used to describe formal and informal programs
• Call attention to current events highlighting the in which mentors, who are people with expertise in a
achievements of women in math or science. field, help people develop and accomplish their educa-
• When talking about potential careers, make tional and career goals. Little rigorous research exists
students aware of the numbers of women who assessing mentors’ effectiveness in math and science
receive advanced degrees in math- and science- per se, but mentoring programs may provide many
related disciplines. The National Science Founda- high-school-age girls with exposure to and connections
tion publishes statistics on women in the sciences, with a woman who has succeeded in math and science.
math, and engineering each year. This information However, some promising research illustrates the effects
is available on its website (www.nsf.gov/statistics). of mentoring programs in increasing the numbers
64
National Science Foundation (2006a).
65
66
Ibid.
67
Ibid.; Wilson and Linville (1985).
68
69
Building Engineering and Science Talent (2004).
( 21 )
PRACTICE GUIDE
of minority students pursuing advanced degrees in

science, technology, mathematics, and engineering.70
In addition, rigorous experimental research has shown
that mentors can positively affect young adolescents’
behaviors (e.g., school attendance, drug and alcohol
use).71 Teachers may choose to support a young girl’s
interest in math or science by helping her to find a
suitable mentoring program.
Potential roadblocks and solutions
Roadblock 3.1. School hours alone may not provide

enough opportunities for girls to be exposed to female
role models, particularly if many of the science and
math teachers in a given school are male.
Solution. Teachers need to encourage parents to take

an active role in providing opportunities for girls to be
exposed to women working in the fields of math and
science. Teachers might encourage parents to sign up
girls for activities where women work in math- and
science-related careers, such as at an aquarium, a
hospital, or a scientific laboratory. In addition, several
national organizations provide conferences designed
to encourage girls’ participation in math and science
careers.
Roadblock 3.2. Many girls see math- and science-

related careers as male oriented. If girls believe that
careers in mathematics and science are nontraditional
choices for women, they may be less likely to take
advanced courses in math and science in high school
or to choose a college path that leads to careers in math
and science.
Solution. It is certainly true that the stereotype of the

high achieving woman in math is a nontraditional one,
and media images do not help this.72 To counteract
this stereotype, teachers can call attention to the fact
that many women today are becoming mathematicians
and scientists. The National Science Foundation
publishes current statistics on the numbers of women
and men receiving bachelor’s, master’s, and doctoral
degrees in math and science.73
70
Maton and Hrabowski (2004); Summers and Hrabowski (2006).
71
Tierney and Grossman (2000).
72
E.g., Davies, Spencer, Quinn, et al. (2002).
73
See National Science Foundation at http://www.nsf.gov/statistics/wmpd/ for current statistics on the numbers of women and men receiving
degrees in math and science.
( 22 )
Recommendation 4: Create a classroom environment that sparks

initial curiosity and fosters long-term interest in math and science.
To encourage more girls to choose careers in the fields of science,
technology, engineering, and math, we recommend that teachers use
strategies designed to generate initial interest in specific math and
science activities and build on this initial interest to foster sustained
interest in math and science content. Few teachers would be surprised
by research indicating that students’ interest is linked to academic
performance and choices for both girls and boys. When students are
interested in mathematics and science, they tend to get better grades
in mathematics and science, take more advanced mathematics and
science courses, and are more likely to pursue mathematics- and
science-related college majors.74
Level of evidence: Moderate activities and careers that involve social and artistic
pursuits.78 These differences in interests are apparent
A number of studies have shown that students’ interest by middle school and are reflected in undergraduate
predicts both concurrent and long-term performance degree and career choices. Even highly mathematically
in and choice of math and science courses, majors, and able youth who were identified in adolescence and
careers.75 In addition, several small-scale experimental followed longitudinally showed gender effects in their
studies have examined the effect of specific strategies choice of degrees. The ratio of male to female under-
(e.g., using practical problems as a context for learn- graduate degrees in math and science is 2.4:1.79
ing math or science skills) on students’ interest and
learning. The panel judges the quality of the evidence To make math and science content more interesting
supporting this recommendation to be moderate, based to students, it is helpful to think about what “interest”
on five small-scale experiments that focus specifically means. In research on learning, the word interest is used
on children learning in math and science contexts 76 in two different ways. The first, which might be called
as well as supporting experimental studies on interest long-term or individual interest, is a relatively stable and
in other domains or with adult populations and personal preference toward certain types of activities. A
correlational studies of students’ interest and course- second kind of interest, which might be called curiosity
taking or career choices.77 or situational interest, is a more immediate response
to particular aspects of situations or problems.80 It is
true, as noted above, that gender differences have been
found in students’ long-term interests. But many expert
the recommendation
teachers realize that an important way to cultivate
Researchers have found gender differences in students’ students’ long-term interests in math and science is to
interests, with boys typically more interested in build upon their initial curiosity. Research (as well as
activities and careers involving scientific, technical, intuition) suggests that curiosity can serve as a hook to
and mechanical pursuits and girls more interested in engage students in math and science content.81 Once
74
Simpkins, Davis-Kean, and Eccles (2006); Updegraff and Eccles (1996).
75
Ibid.
76
Cordova and Lepper (1996); Parker and Lepper (1992); Ginsburg-Block and Fantuzzo (1998); Turner and Lapan (2005); Phelps and
Damon (1989).
77
For a recent review of this literature, see Hidi and Renninger (2006).
78
Lapan, Adams, Turner, et al. (2000).
79
Webb, Lubinski, and Benbow (2002).
80
Mitchell (1993); Hidi and Renninger (2006).
81
Hidi and Renninger (2006).
( 23 )
PRACTICE GUIDE
students’ interest in a topic or content area is sparked, • Use project-based learning, group work, in-
teachers can then build on that curiosity, providing novative tasks, and technology to stir interest
students with opportunities to engage with interesting in a topic. Using web-based presentations of
material and potentially transforming that initial content and animated presentations of changes
curiosity into long-term interest. that occur during chemical bonding supports
student exploration of key chemical concepts
and is associated with improved learning. Other
How to carry out the recommendation research shows that when students are asked
Teachers can take several actions to create a classroom to solve the series of problems presented in a
video-based adventure as members of teams,
environment that sparks initial interest and may foster
they outperform their counterparts who are
long-term interest.82 To spark initial interest, teachers
asked to solve the problems individually. 87
can provide students the opportunity to see how the
skills and knowledge they are learning in mathematics • Examine the large variety of tools designed for
and science classes can be used to solve interesting and teachers to use with girls and young women to
meaningful problems. For example, teachers can: spark initial curiosity about science and mathemat-
ics content. Many of these tools incorporate the
• Embed mathematics word problems and science principles that have been discussed in this practice
activities in interesting contexts.83 For elementary guide. A listing of these tools is published in a
school children, embedding mathematical practice volume entitled New tools for America’s workforce:
in fantasy contexts (e.g., saving a planet from an girls in science and engineering and can be down-
alien invasion, searching for buried treasure or loaded at http://www.nsf.gov/publications/. 88
criminals) has been found to be highly motivating
and interesting and has led to improved perfor- Teachers can capitalize on that initial situational
mance.84 For middle school children, embedding curiosity by providing students with opportunities to
mathematical practice in the contexts of real- deepen their knowledge and understanding of par-
ticular math and science content and to broaden their
world problems (e.g., figuring out how to build an
understanding of how what they are learning today
effective skateboard ramp given a limited budget
connects to their future goals. Teachers should also
or a hovercraft that “flies” multiple students)
consider providing students with access to female role
can be both motivating and support learning.85 models who have been successful in mathematics and
• Provide students with access to rich, engaging, science careers to nurture long-term individual interest
(see Recommendation 3).
relevant informational and narrative texts as they
participate in classroom science investigations. • Encourage middle and high school students to
For example, during weeklong series of science examine their beliefs about which careers are
investigation activities where elementary school typically female-oriented and which are typically
children learn about owls and birds and have the male-oriented. Encourage these students to learn
opportunity to dissect owl pellets and discuss more about careers that are interesting to them
what they find, the children are also asked to read but that they believe employ more members of the
from both informational and narrative texts that opposite gender. Connect mathematics and sci-
focus on owls. Embedding the reading activities ence activities to careers in ways that do not rein-
into the context of science investigation serves to force existing gender stereotypes of these careers.89
generate situational interest and further curiosity.86
82
Mitchell (1993); Hidi and Renninger (2006).
83
Renninger, Ewen, and Lasher (2002).
84
Cordova and Lepper (1996); Parker and Lepper (1992).
85
Bottge, Rueda, Serlin, et al. (2007).
86
Guthrie, Wigfield, Humenick, et al. (2006).
87
Barron (2000); Linn, Lee, Tinker, et al. (2006); Phelps and Damon (1989); Kaelin, Huebner, Nicolich, et al. (2007).
88
National Science Foundation (2006c).
89
Turner and Lapan (2005); Ji, Lapan, and Tate (2004); Lapan, Adams, Turner, et al. (2000).
( 24 )
Potential roadblocks and solutions and covering the required curriculum. But, despite
increased time demands, using interesting tasks and
Roadblock 4.1. Teachers may be reluctant to incorpo- instructional formats may be worth the extra effort:
rate aspects of youth culture into a lesson because they Research indicates a strong link between interest and
lack familiarity with students’ interests. Teachers may academic performance, for both girls and boys. In
be concerned that they might look silly by connecting addition, as a teacher and her students become more
to something long out-of-date in students’ minds. accustomed to doing things a bit differently (e.g., using
group work or using project-based learning), the extra
Solution. Certainly it is a challenge to follow cultural
effort required to implement these kinds of activities
trends and fads in students’ interests; it is likely that
can decrease.
students do not expect their teachers to be able to
keep up and may even be surprised when one is able
to do so. However, connecting to trends in popular
culture is only one potential strategy for arousing
students’ curiosity. Teachers can also seek to connect
math and science content with other contexts that
are of interest to students of all ages and generations,
including history and current events. In addition,
incorporating hands-on activities, group work, and
technology into lessons can make the content more
interesting for students.
Roadblock 4.2. Teachers may be reluctant to incor-

porate interesting activities, because of the belief that
some activities might distract students and not serve
the goals of the lesson.
Solution. Teachers should not interpret this recom-

mendation as suggesting that all activities in math and
science classes should take the form of games or fun
activities. While games can be fun and exciting for
students, it is also true that some games and activities
that are not tightly linked to learning objectives may
indeed distract students and be counterproductive to
learning. Teachers should strive to achieve a balance
between enhancing activities so that they incorporate
some material that sparks interest while ensuring that
content learning remains the primary goal of the lesson.
Roadblock 4.3. Teachers may be hesitant to use

time-intensive methods such as hands-on activities and
group work, given the pressure that many feel to keep
up with curriculum pacing guides.
Solution. It is true that creating interesting and

innovative tasks is time-consuming on many levels—
increased preparation time for teachers, more time
needed to cover the material in class, and frequently
more time for teachers to grade or evaluate students’
work. Teachers need to achieve a balance between
incorporating interesting tasks and lesson formats
( 25 )
( 26 )
Recommendation 5: Provide spatial skills training.

We recommend that teachers provide spatial skills training for girls.
Researchers have found that spatial skills are associated with
performance on math tests and that spatial skills can be improved
with practice on certain types of tasks.90
Level of evidence: Low We find the largest between-gender differences with

standardized examinations that do not match any
We consider the level of evidence to support this particular curriculum. We are referring to the standard-
recommendation to be low because of the lack of ex- ized tests that are used for admission to undergraduate
perimental or rigorous quasi-experimental studies that and graduate programs and for state and international
directly examine the effects of spatial skills training on comparisons. As already discussed, one possible reason
girls’ math or science performance. Our recommenda- that girls and women perform less well on these tests
tion to improve math and science performance through in mathematics and science is stereotype threat, but
spatial skills training represents an extrapolation based the differences are found on specific items on these
on the broader relevant research literature.91 tests, which suggests that there are gender differences
in how certain types of questions are solved or not
solved. For example, numerous researchers have found
that when math items are highly spatial in nature, boys
the recommendation solve more of these questions correctly.93 Consider
When we look at girls’ math and science performance the conclusion from a study of 24,000 ninth-graders,
which showed that males perform better on items
in school, we see that they are doing well on tests
that require significant spatial processing and females
that are closely related to the curriculum they are
outperform males on items requiring memorization.94
taught in school. As noted earlier, girls, on average,
Additional support comes from an analysis of the
graduate from high school with slightly more math
mathematical test questions that showed the largest
and science classes than boys and with higher grades. gender differences favoring males on an international
Girls now obtain almost half of all undergraduate math assessment.95 The items included calculating the
degrees in mathematics and the majority of degrees in height of a mountain, calculating the distance between
health-related fields. But, as noted at the beginning of two intercepts on a plane, calculating the length of
this guide, girls and women are underrepresented in a string, calculating the perimeter of a polygon, and
physics, computer science, engineering, and chemistry other similar problems that are spatial in nature. There
at the undergraduate level, and their numbers fall off as is evidence that gender differences in math problem-
they move beyond the undergraduate degree.92 solving strategies begin as early as first grade, with girls
90
Doolittle (1989); Newcombe (2002); McGraw, Lubienski, and Struchens (2006).
91
Ben-Chaim, Lappan, and Houang (1988); Piburn, Reynolds, McAuliffe, et al. (2005).
92
A summary of research can be found in Halpern, Benbow, Geary, et al. (2007).
93
E.g., Bielinksi and Davison (2001); Doolittle (1989); Gallagaher, De Lisi, Holst, et al. (2000); Geary, Saults, Liu, et al. (2000); Gierl,
Bisanz, Bisanz, et al. (2003).
94
Gierl, Bisanz, Bisanz, et al. (2003).
95
Casey, Nuttall, and Pezaris (2001).
( 27 )
PRACTICE GUIDE
using more overt strategies, such as counting, and boys Several studies have documented that women who
using more conceptual spatial strategies.96 undergo specific spatial skills training programs show
improved performance in the specific domain of their
A large research literature shows that boys outperform training.105 In one study, for example, researchers
girls on many tests of spatial skills, especially ones that designed and implemented a course to improve the
require visualizing what an object will look like when it spatial visualization skills of first-year engineering
is rotated in space.97 Researchers have established that students.106 In this course, students learned effective
spatial skill performance is correlated with performance strategies for mentally representing objects, and for
in mathematics and science.98 For example, researchers using graphs, diagrams, charts, and maps as tools for
have found that kindergarteners’ ability to perceive thinking about topics in science and mathematics.
and discriminate among various shapes and geometric Notably, retention in the engineering program for
forms predicts their later performance in fourth-grade female engineering students who took the spatial
math.99 Scores on a spatial visualization test, for visualization course was 77 percent, whereas retention
example, have correlated with subsequent test scores was only 47 percent among those who did not take the
in geology.100 Other evidence supporting the idea that course. Although both females and males participated
spatial abilities are important in math and science was in the spatial training course, more females scored
provided by researchers who found that scores on a test low on tests of spatial visualization at the start of the
of mental rotation, which measures how well students program, and females showed greater gains in retention
can visualize an irregular shape when it is rotated in in the engineering program as well as in grades. Other
space, accounted for a large proportion of the male- researchers examined the effects of spatial training
female difference on standardized examinations.101 when using information about maps and found that
Additionally, geometry items comprise close to one- when students were taught how to project lines of
longitude and latitude and to visualize the curve of the
third of the questions on the math portion of the SAT,
Earth’s surface, their map skills improved.107
and the largest differences between males and females
are found on geometry items.102 Spatial skills, therefore, Several studies, including a few experiments using
figure importantly in the male-female test score gap.103 random assignment, have directly examined spatial
skills training in school-age children. In these studies,
In mathematics, researchers have studied the way
children who received training in a specific spatial skill
that representations affect the learning of complex (e.g., mental rotation, spatial perspective, embedded
mathematical ideas. For example, in a study of the figures) improved their performance on related visual
strategies used to solve mathematical problems, and spatial measures more than children who did not
researchers found that overall, the male students were receive this training.108 In a recent synthesis of the
more likely than female students to use a flexible set spatial skills training literature, such training was found
of general strategies and more likely to solve problems to improve the visual and spatial skills of both children
correctly when the solution required a spatial represen- and adults.109 Thus, targeted training can serve to
tation, a short cut, or the maintenance of information improve spatial skills performance beginning in early
in spatial working memory.104 childhood and continuing into adulthood.
96
Carr, Jessup, and Fuller (1999).
97
E.g., Battista (1990); Bielinksi and Davison (2001); Doolittle (1989); Harris and Carlton (1993); McGraw, Lubienski, and Struchens
(2006).
98
Casey, Nuttall, and Pezaris (2001); Kurdek and Sinclair (2001); Piburn, Reynolds, McAuliffe, et al. (2005).
99
Kurdek and Sinclair (2001).
100
Piburn, Reynolds, McAuliffe, et al. (2005).
101
Casey, Nuttall, and Pezaris (2001).
102
Harris and Carlton (1993).
103
E.g., Bielinksi and Davison (2001); McGraw, Lubienski, and Struchens (2006).
104
Gallagher, De Lisi, Holst, et al. (2000).
105
Halpern (2000).
106
Sorby (2001).
107
Muehrcke and Muehrcke (1992).
108
Connor, Schackman, and Serbin (1978); DeLisi and Wolford (2002).
109
Marulis, Liu, Warren, et al. (2007).
( 28 )
To reiterate, the mathematical test items that show the offering ideas to teachers for incorporating spatial skills
greatest difference favoring boys are spatial in nature. training in their everyday curriculum.115
Spatial skills can be improved with training. When
spatial skills training was given to college students in Roadblock 5.2. Learners in any classroom will vary in
engineering, grades in courses that use spatial skills their spatial abilities, making it difficult for teachers to
improved and retention in engineering programs, know how to target their teaching in this domain.
especially for women students, improved. In addition,
Solution. Tools and lesson plans available on the Web
spatial ability predicted which courses high school
can be used by learners at different levels of ability.
students liked best (with math and science courses
In addition to published materials, special workshops
positively related to spatial ability) and the careers in
for teachers that vary by grade level could also help
which students were employed when they were in their
teachers begin to plan lessons, as well as ready-to-use
30s (with spatial ability positively related to careers in
sample exercises and online training programs.
math, science, and engineering).110
How to carry out the recommendation
Teachers can provide spatial skills training with a

variety of age-appropriate activities. In particular,
teachers can:
• Encourage young girls to play with toys that

require the application of spatial knowledge,
such as building toys.111
• Teach older girls to mentally image and draw

mathematics or other assignments so that they
become as comfortable with spatially displayed
information as they are with verbal information.112
• Require answers that use both words and a

spatial display.113
• Provide opportunities for specific training in

spatial skills, such as mental rotation of images,
spatial perspective, and embedded figures.114
Potential roadblocks and solutions
Roadblock 5.1. Some teachers may not feel prepared

to provide spatial skills training.
Solution. Many materials are available to help teachers

include spatial skills in their everyday curriculum. Free
software and lesson plans for grades K to 12 are avail-
able on websites dedicated to the topic of spatial skills
training. In addition, published materials are available
110
Shea, Lubinski, and Benbow (2001).
111
Deno (1995).
112
Gerson, Sorby, Wysocki, et al. (2001).
113
Casey, Nutall, and Pezaris (2001).
114
Sorby and Baartmans (2000).
115
Casey (2003).
( 29 )
( 30 )
Conclusion In addition to beliefs about abilities, girls are more

likely to choose courses and careers in math and science
To conclude, let us revisit our recommendations. if their interest in these fields is sparked and cultivated
What do we recommend that teachers do to encour- throughout the school years.116 Our fourth recom-
age girls and young women to choose career paths mendation focuses on the importance of fostering both
in math- and science-related fields? One major way situational and long-term interest in math and science,
is to foster girls’ development of strong beliefs about and provides concrete strategies that teachers can use
their abilities in these subjects—beliefs that more to do so.
accurately reflect their abilities and more accurate In addition to beliefs and interests, a final way to
beliefs about the participation of women in math- and encourage girls in math and science is to help them
science-related careers. Our first two recommendations, build the spatial skills that are crucial to success in
therefore, focus on strategies that teachers can use to many math- and science-related fields, such as physics,
strengthen girls’ beliefs regarding their abilities in math engineering, architecture, geometry, topology, chemis-
and science: (1) Teach students that academic abili- try, and biology. Research suggests that spatial skills, on
ties are expandable and improvable; and (2) Provide which boys have typically outperformed girls, can be
prescriptive, informational feedback. Our third recom- improved through specific types of training. Thus, our
mendation addresses girls’ beliefs about both their final recommendation is that teachers provide students,
abilities and the participation of women in math- and especially girls, with specific training in spatial skills.
science-related careers: (3) Expose girls to female role
models who have succeeded in math and science.
116
Wigfield, Eccles, Schiefele, et al. (2006).
( 31 )
( 32 )
Appendix: Example of an intervention that teaches students

that academic abilities are expandable and
Technical information improvable.
on the studies In our example study,121 shortly after the school year
began, 138 seventh-graders (both boys and girls) were
randomly assigned a mentor, with whom they commu-
nicated throughout the school year. The mentors were
Recommendation 1: 25 college students who were completing a required
Teach students that academic abilities are mentor-training course designed by the school district.
The mentors’ role was to offer advice to their assigned
expandable and improvable.
students regarding study skills and the transition to
junior high, explicitly teach one of four messages, and
Level of evidence: Moderate help the students design and create a web page for their
computer skills class in which the students advocated
The panel rated the level of evidence as Moderate. the experimental message conveyed to them by the
We were able to locate two small experiments demon- mentor throughout the year.
strating support for the practice of improving students’
performance by teaching them that academic abilities The participating seventh-graders were randomly
are expandable and improvable.117 In the first study, assigned to one of four conditions, representing a
138 girls and boys in seventh grade participated in the message taught by a mentor: (a) the incremental
study that examined the effect of teaching students condition, in which students learned that intelligence
is an expandable capacity that increases with mental
about the expandability of their abilities on end-of-year
effort; (b) the attribution condition, in which partici-
math and reading achievement test scores.118 The
pants learned that many students experience difficulty
second study was conducted with 95 girls and boys when they move to a new educational level (such as
in seventh grade and examined the effect of such in- junior high), but then improve their performance
struction on students’ end-of-semester math grades.119 once they are familiar with their new environment;
In addition, we identified a small experiment with (c) a combined condition, in which students learned
79 female and male college students that demonstrated both the incremental and attribution messages; and
support for this practice improving students’ end-of- (d) an antidrug control condition, in which students
term grade point averages.120 All three studies included learned about the dangers of drug use. Each mentor
students from ethnically and socioeconomically diverse had about six randomly assigned students across three
backgrounds. of the four messages.
These studies share the following features: Students or Students’ math achievement was measured at the
groups of students were randomly assigned to treat- end of the school year using the Texas Assessment of
Academic Skills, a statewide standardized achievement
ment and control conditions, where some were taught
test administered to all students in the district.
that intelligence is malleable and abilities are expand-
able, while others were not taught this but spent an Study results indicate that the boys in the antidrug
equal amount of time engaged in learning something condition performed significantly better on the math
new. In all three studies, students in the group that was test than the girls; that is, the math score mean for the
taught that abilities are expandable performed better boys was 81.55, but the mean math score for girls was
than students in the control condition. 74.00. However, in the three other conditions, the
gender gap in math performance disappeared. The
117
Blackwell, Trzesniewski, and Dweck (2007); Good, Aronson, and Inzlicht (2003).
118
119
Blackwell, Trzesniewski, and Dweck (2007).
120
Aronson, Fried, and Good (2002).
121
( 33 )
PRACTICE GUIDE
math test scores of the girls in the experimental variation in math course-taking.126 Students examined
conditions were statistically equivalent to the scores of in this study (n = 1,039) were followed from 9th to
the boys across all conditions (i.e., in the incremental 12th grade, and were primarily from middle and
condition, the average math score for males was 85.25 lower middle class European American backgrounds.
and for females was 82.11; in the attribution condition, Self-concept of ability in math, utility of math, and
the average math scores of both boys and girls was interest in math were measured in the ninth grade
84.53; in the combined condition, the average math using questionnaires. Grades in math, standardized
scores for females was 84.06 and for males was 82.30). test scores in math, specific course enrollment choices,
Importantly, the large difference in test scores between and the number of math classes taken throughout high
boys and girls that was found in the control condition school were gathered from school record data. Findings
(i.e., antidrug message) was eliminated in the three indicated a strong and significant association between
experimental conditions. The experimental manipula- self-concept of math ability and performance in
tions were particularly beneficial for the female students mathematics. In addition, the perceived utility of math
and demonstrated significant effects closing the gender had the strongest and most consistent association with
gap in performance, indicating that the intervention the number of high school math courses taken, with
procedures meaningfully increased girls’ math scores boys perceiving math as having significantly greater
compared to the control condition. utility than girls.
In the second longitudinal study, researchers analyzed

Recommendation 2: data from a sample of 227 youths from 5th through
12th grade to address longitudinal associations among
Provide prescriptive, informational feedback. students’ math- and science-related activities, their
math and science related beliefs, and math and science
Level of evidence: Moderate courses taken throughout high school.127 The research-
ers focused on the physical sciences. Results indicated
The panel rated the level of evidence as Moderate. that youths’ participation in out-of-school math
We were able to locate one high-quality study, and science activities during fifth grade significantly
which presented a series of six random assignment predicted their math and science beliefs at sixth grade
experiments demonstrating support for the practice (e.g., math and science self-concepts, perceptions of
of providing K–12 students with prescriptive, informa- math and science importance, and interest in math
tional feedback as a way to improve student motivation and science). These beliefs measured at 6th grade
and performance.122 Two additional classroom-based predicted beliefs at 10th grade, and the 10th-grade
experimental studies provide support for providing beliefs predicted the number of high school courses
prescriptive, informational feedback during mathemat- students took, even after taking into account the
ics instruction.123 We also drew on a recent substantive predictive power of their math and science grades in
review of the literature that discusses the effects of 10th grade.
praise on children’s intrinsic motivation.124 In addition,
two high-quality longitudinal studies demonstrated Together, these two studies suggest a strong relation
the link between students’ self-efficacy beliefs and their between the math and science courses that students
math- and science-related choices.125 choose to take in high school and their beliefs regarding
their abilities and interest in these subjects, as well as
In one of the longitudinal studies, researchers used their perception of the importance of these subjects.
a large random sample of students in 10 different Although these studies have high external validity, they
school districts in southeastern Michigan to test the do not answer questions regarding the direction of
utility of students’ math-related self-beliefs to explain causality between students’ self-concepts in math and
122
123
Elawar and Corno (1985); Miller, Brickman, and Bolen (1975).
124
Henderlong and Lepper (2002).
125
Simpkins, Davis-Keans, and Eccles (2006); Updegraff and Eccles (1996).
126
Updegraff and Eccles (1996).
127
Simpkins, Davis-Kean, and Eccles (2006).
( 34 )
science and their performance in these areas. The third given a third and easier set of matrices to complete,
study 128 presents a series of six randomized controlled thus yielding a measure of post-failure performance.
experiments that demonstrate that performance on a
math-related skill can be improved by manipulating Findings from this series of studies indicate that
fifth-graders’ attributions regarding success and failure, fifth-graders who were praised for ability performed
thus increasing their beliefs regarding the likelihood less well following subsequent failure, whereas students
that they will succeed on future related tasks. who were praised for effort did not. Following failure,
students praised for effort rather than ability also
Example of an intervention that uses prescriptive, showed greater task persistence and enjoyment and
informational feedback. greater orientation toward mastery learning. Students
in the control condition typically had scores between
In a series of six experiments, the impact of ability those of students in the ability-praise and effort-praise
versus effort feedback on fifth-graders’ subsequent conditions.
performance on the Standard Progressive Matrices,129
designed to measure ability to form perceptual rela- In this series of studies, several alternative explanations
tions and to reason by analogy, was examined. The were ruled out, including the possibility that: (a) praise
Standard Progressive Matrices is a nonverbal measure for ability might have led children to have higher
that correlates strongly with the Stanford Binet and expectations for future performance, which might then
Wechsler Scales of intelligence. In addition, studies of have led to greater disappointment following failure;
the Wechsler Preschool and Primary Test of Intelligence (b) having the same teacher administer both positive
(Revised) show that math performance is moderately and negative feedback following task completion might
related to perceptual organization for every age group have increased the desire of ability-praised children
sampled. Thus, the Standard Progressive Matrices to perform well on the second set of matrices; and (c)
measures perceptual relations skills that are related to ability-praised children might have interpreted the task
mathematics performance. as an intelligence test, thus leading to greater disap-
pointment following failure. With these competing
The number of participating fifth-graders for each of explanations ruled out through experimental manipula-
the six studies was: 128, 51, 88, 51, 46, and 48. The tion, the authors conclude that ability-focused praise
ethnically and socioeconomically diverse students (both led the fifth-graders to view performance as an index of
boys and girls) were drawn from public schools in the ability, resulting in decreased motivation and perfor-
midwest and the northeast. Students were randomly mance following failure. Students who were praised for
assigned to one of three conditions: (a) ability or effort rather than ability did not show these post-failure
intelligence praise; (b) effort praise; and (c) a control deficits in motivation and performance.
condition in which students received a general state-
ment of praise with no attributions regarding either
ability or effort.
Recommendation 3:
In this series of studies, students were asked to do Promote positive beliefs regarding women’s
three sets of Standard Progressive Matrices. Following abilities by exposing girls to female role
the first set, students received either positive praise models who have succeeded in math
regarding their intelligence or ability; positive praise and science.
regarding their effort; or nonspecific praise that did
not offer an attribution for success. Students then
Level of evidence: Low
completed a second set of matrices, following which
they were told that they had performed “a lot worse” The panel rated the level of evidence that supports
on this set than they had on the first set. After complet- this recommendation as Low, based on four small
ing items measuring dimensions such as desire to experimental studies with college students. Although
persist on the problems, enjoyment of the problems, the experiments that support the recommendation have
and attributions for their failure, students were then strong internal validity for supporting causal claims,
128
129
Ibid.
( 35 )
PRACTICE GUIDE
these studies were conducted with college students strategy (i.e., exposing girls to positive role models) in
rather than girls in kindergarten through high school general has been shown to be effective for elementary to
and were short laboratory experiments, rather than secondary school students in some outcome domains,
real-world classroom studies conducted with students although not specifically for improving outcomes in
over extended periods of time. Thus, the generalizabil- math or science.
ity of these studies to the effects of exposing girls
to female role models over, for example, the course Example of intervention using competent female
of a school year is limited. role models with college students.
In addition to these studies that explicitly address In our example study, a research team conducted
the effect of female role models on young women’s a series of three random-assignment experiments
math performance and beliefs about their math examining the impact of competent female role models
abilities, there is evidence that exposure to positive on the performance of college women in a setting that
role models or mentors can have a positive effect was represented to participants as a diagnostic testing
on students’ behaviors.130 situation for math.134 The participants in each of these
experiments were students who considered themselves
In a set of three small experimental studies with to be good in math and to be motivated to do well
college students, researchers demonstrated that the on math exams. In experiment 1, 43 undergraduate
presence of female role models who are competent women and men were randomly assigned to undergo
in math mitigated the negative impact of gender the math testing situation with either a female or a
stereotypes on the math performance of female male experimenter, both of whom were presented to
college students.131 These studies provide evidence the participants as being competent in math. Women
that exposure to competent female role models can performed as well as their male counterparts when the
close the female-male math gap in testing and improve experimenter was a woman; women did not do as well
young women’s beliefs about their math abilities. on the math test when the experimenter was a man.
However, these studies do not examine the efficacy
of female role models with girls from kindergarten In experiments 2 and 3, participants were provided
through high school. with the biographical sketch of a fictitious female
experimenter who was not actually present; in the
In another small, random-assignment experiment biographical sketches, the female experimenter’s level of
with college women, a team of researchers found math competence was manipulated to be either high or
significantly improved performance on a test of items low. The results of both experiments demonstrated that
similar to those used for the GRE math test when the the improved performance on the math exam obtained
students first read about women’s accomplishments in in experiment 1 was dependent on the undergraduate
architecture, law, medicine, and invention.132 women perceiving the female role model to be very
compentent in math, as opposed to simply taking the
These four studies provide evidence of the positive test with a female experimenter with average math
effect of female role models on the test performance skills. In addition, in experiment 3, the researchers
of young college women and on their beliefs about found that young women’s perceptions of their math
their math abilities. Some evidence suggests that ability were higher when they were exposed to a female
exposure to positive role models or mentors can have role model who was competent in math.
a positive effect on students’ behaviors. For example, a
large-scale evaluation of Big Brothers/Big Sisters found These studies provide evidence that a brief exposure
positive effects of mentoring on young adolescents’ to a compentent female role model can affect college
school attendance and drug and alcohol use.133 women’s performance on tests and their perceptions of
We cite these studies as providing evidence that the their math abilities. An important limitation of these
130
E.g., Tierney and Grossman (2000).
131
Marx and Roman (2002).
132
McIntyre, Paulson, and Lord (2003).
133
Tierney and Grossman (2000).
134
Marx and Roman (2002).
( 36 )
studies for our purposes is that the participants were of science texts.141 Finally, another set of studies
students who self-identified as being good in math and examined 133 third-grade and 106 fifth-grade students
being motivated to do well on math tests. from two low-income schools with predominantly
African-American populations.142
Across these studies, students in the high-interest

Recommendation 4: conditions outperformed students in control condi-
Create a classroom environment that sparks tions on study-specific indices of interest, motivation,
initial curiosity and fosters long-term interest and learning. The learning activities in each of these
in math and science. interventions share some or all of the following
features: They were designed around contexts that
Level of evidence: Moderate students found engaging, made available a range of
texts, used real-life settings, used technology, provided
The panel rated the level of evidence as Moderate. There students’ with choice, and used group work.
is a long and rich tradition of exploring ways
to increase student interest in mathematics and science Example of intervention that improves math
content in education research. In determining the level performance through increasing interest.
of evidence supporting this recommendation, we drew In this study, mathematics instruction was embedded
on several experimental and quasi-experimental studies. in fantasy or nonfantasy computer games in order to
For example, we found two small-scale, randomized study the impact of intrinsic interest on math perfor-
controlled trials that demonstrate support for the mance.143 Seventy students (both boys and girls) were
practice of providing elementary school students with randomly assigned to play computer games designed
an environment that improves learning in the areas to teach specific math skills (the hierarchy of the order
of mathematics and computer science by fostering of operations and the proper use of parentheses in
greater interest.135 Others have experimentally evaluated arithmetic expressions) in one of five conditions: (a) in
the facilitative effect of peer collaboration on solving a basic, unembellished version that served as a control
mathematics problems;136 while yet another study 137 condition, students played a math game on the com-
examined how participating in a career exploration puter in which they were required to learn and correctly
module fostered middle school girls’ interest in use the target skills in order to win the game; (b) in two
nontraditional careers (e.g., those in the areas of generic fantasy conditions, the activity was presented
mathematics and science). Two additional quasi-experi- within fantasy contexts (spacetrip and treasure island)
mental studies provide support for the benefits of this designed to increase children’s intrinsic interest in
practice on students’ motivation and comprehension the game; in the first generic fantasy condition, the
of science texts.138 students had some choice regarding at least six details
One study examined the impact of three interest- that were incidental to the game itself (e.g., type and
enhancing strategies on the motivation and perfor- color of spaceship to use), whereas (c) in the second
mance of 70 fourth- and fifth-grade students in a generic fantasy condition, these details were provided
related math task.139 Another examined the influence randomly by the computer; (d) in two personalized
of embedding a computer programming task in a fantasy conditions, several generic referents in the
fantasy context with 27 children.140 In another program program were replaced with details that were personally
of research, two studies involving a total of 885 third- relevant to the child (e.g., child’s name, favorite foods
grade students investigated reading comprehension to bring along on the spacetrip), obtained from a
135
Cordova and Lepper (1996); Parker and Lepper (1992).
136
Ginsburg-Block and Fantuzzo (1998); Phelps and Damon (1989).
137
Turner and Lapan (2005).
138
Guthrie, Wigfield, Barbosa, et al. (2004); Guthrie, Anderson, Alao, et al. (1999).
139
Cordova and Lepper (1996).
140
Parker and Lepper (1992).
141
Guthrie, Wigfield, Barbosa, et al. (2004).
142
Guthrie, Anderson, Alao, et al. (1999).
143
Cordova and Lepper (1996).
( 37 )
PRACTICE GUIDE
pretest questionnaire; in the first personalized fantasy considered important in math and science achieve-
condition, the students had some choice regarding at ment. One study was a small-scale, random-assignment
least six details incidental to the game itself, whereas experiment with college students144; the other was a
(e) in the second personalized fantasy condition, quasi-experiment with elementary or middle school
the students did not have a choice. Thus, the five students.145 Both studies focused on the improvement
conditions were: no-fantasy control condition; generic of specific spatial skills. Although there is agreement
fantasy, no choice; generic fantasy, choice; personalized among experts that the relevant skills are important to
fantasy, no choice; personalized fantasy, choice. specific aspects of math and science performance, the
transfer of these skills more generally to achievement in
The intervention consisted of five sessions. In the first math and science courses has not been directly demon-
session, participating students were pretested on their strated. The evidence supporting the practice of spatial
knowledge of the target math skills, and also completed skills training as a way to improve performance in math
several measures of their global motivational orienta- and science courses in K–12 is thus indirect.
tion. In the next three sessions, each child played with
one randomly assigned version of the computer games. In the first study, 103 college students taking an
These sessions were scheduled about five days apart introductory geology course were in one of four
and lasted about 30 minutes each. In the fifth session, sections randomly assigned to either an experimental
students completed posttests on the target math skills, or a control condition.146 Both experimental and
as well as several additional attitudinal measures. control sections used the same laboratory manual;
Dependent measures included self-reports of interest however, the two experimental sections used additional
and enjoyment, behavioral commitments to continued computer-based modules that allowed extensive student
task engagement, preferences for increasingly chal- involvement with images that could be manipulated.
lenging tasks, and direct measures of students’ online Students were administered a content assessment and
involvement in the activities. two spatial-visual measures as pretests and posttests.
Study results indicate that students in the experimental
The results of the intervention indicate that students sections improved their spatial visualization skills
exposed to each of the three specific strategies designed significantly more than did students in the control
to heighten student interest (contextualization, choice, sections and that pre-existing gender differences with
and personalization) showed significantly greater mo- regard to spatial visualization were eliminated in the
tivation, involvement, and learning of the target math experimental sections. That is, final posttest scores of
skills than those students in the control condition. males and females on the spatial visualization tasks were
not different from one another.
Recommendation 5: In the second study, researchers examined the impact

of spatial visualization skills training on about 1,000
Provide spatial skills training. fifth- through eighth-grade students.147 Before instruc-
tion, there were significant differences by sex in spatial
Level of evidence: Low visualization performance (favoring boys). After the
intervention, middle school students, regardless of sex,
The panel rated the level of evidence as Low. That is, gained significantly from the training program. Reten-
the evidence for the recommendation is based on the tion of the impact persisted after a 4-week period and
expert opinion of panel members, justified by high- after 1 year. The persistent gender differences found
quality research in related domains. in performance after training in this younger sample
are important to note. This research suggests that the
The panel located two high-quality studies of spatial
spatial skills of both boys and girls are susceptible to
skills training focused on skills that are generally
144
145
Ben-Chaim, Lappan, and Houang (1988).
146
147
Ben-Chaim, Lappan, and Houang (1988).
( 38 )
training; however, it leaves open the question of how

to close the gender gap in spatial skills performance.
Example of intervention that improves

spatial skills.
The quasi-experiment supports the practice of using

spatial skills training to improve specific skills that
experts consider to bear a relationship to performance
in mathematics, science, and engineering.148 Spatial
visualization is a subset of spatial skills that involves the
ability to mentally manipulate and rotate objects. In
this study, about 1,000 fifth- to eighth-grade students
at three sites in and around a large midwestern city
participated in a 3-week spatial visualization unit that
engaged them in concrete activities with small cubes,
such as building and drawing representations. The
spatial visualization test was used as a pretest and post-
test measure. After the instruction intervention, middle
school students, regardless of sex, gained significantly
from the training program in spatial visualization
tasks. Boys and girls responded similarly to the training
program, indicating that spatial visualization skills can
be improved when appropriate training is provided.
The retention of the training effects 4 weeks and 1 year
later underscores the potential long-term benefits of
spatial skills training.
148
Ibid.
( 39 )
( 40 )
References
American Educational Research Association, American Psychological Association, & National Council on Measure-
ment in Education. (1999). Standards for educational and psychological testing. Washington, DC:
AERA Publications.
American Psychological Association (2002). Criteria for practice guideline development and evaluation.
American Psychologist, 57, 1048-1051.
Andre, T., Whigham, M., Hendrickson, A., and Chambers, S. (1999). Competency beliefs, positive affect, and
gender stereotypes of elementary students and their parents about science versus other school subjects.
Journal of Research in Science Teaching, 36, 719-747.
Aronson, J. (2002). Stereotype threat: Contending and coping with unnerving expectations. In J. Aronson (Ed.),
Improving academic achievement: impact of psychological factors on education. San Diego, CA: Academic Press.
Aronson, J., Fried, C. and Good, C. (2002). Reducing the effects of stereotype threat on African American college
students by shaping theories of intelligence. Journal of Experimental Social Psychology, 38, 113-125.
Aronson, J., and Good, C. (2002). The development and consequences of stereotype vulnerability in adolescents.
In F. Pajares and T. Urdan (Eds.) Academic motivation of adolescents (pp.299-330). Greenwich, CT: Information
Age Publishing.
Aronson, J., and Steele, C.M. (2005). Stereotypes and the fragility of human competence, motivation, and
self-concept. In C. Dweck & E. Elliot (Eds.), Handbook of competence & motivation. New York: Guilford.
Bangert-Drowns, R.L., Kulik, C-L. C., Kulik, J.A., and Morgan, M. (1991). The instructional effect of feedback
in test-like events. Review of Educational Research, 61(2), 213-238.
Barron, B. (2000). Problem solving in video-based microworlds: Collaborative and individual outcomes of
high-achieving sixth-grade students. Journal of Educational Psychology, 92(2), 391-398.
Battista, M. (1990). Spatial visualization and gender differences in high school geometry. Journal for Research in
Mathematics Education, 21, 47–60.
Ben-Chaim, D., Lappan, G., and Houang, R. T. (1988). Effect of instruction on spatial visualization skills of
middle school boys and girls. American Educational Research Journal, 25, 51 – 71.
Bielinksi, J., and Davison, M.L. (2001). A sex difference by item interaction in multiple choice mathematics items
administered to national probability samples. Journal of Educational Measurement, 38, 51–77.
Black, P. and Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education: Principles,
Policy & Practice, 5(1), 7-74.
Blackwell, L., Trzesniewski, K., and Dweck, C.S. (2007). Implicit theories of intelligence predict achievement
across an adolescent transition: A longitudinal study and an intervention. Child Development, 78, 246-263.
Bottge, B.A., Rueda, E., Serlin, R.C., Hung, Y., and Kwon, J.M. (2007). Shrinking achievement differences with
anchored math problems: Challenges and possibilities. The Journal of Special Education, 41(1), 31-49.
Building Engineering and Science Talent. (2004). What it takes: PreK-12 design principles to broaden participation
in science, technology, engineering, and mathematics. Retrieved September 12, 2007 from
http://www.bestworkforce.org/.
( 41 )
PRACTICE GUIDE
Carr, M., Jessup, D.L., and Fuller, D. (1999). Gender differences in first grade strategy use: Parent and teacher
contributions. Journal for Research in Mathematics Education, 30, 20–46.
Casey, B. (2003). Mathematics problem-solving adventures: A language-arts based supplementary series for early
childhood that focuses on spatial sense. In D. Clements, J. Sarama, and A.-M. DiBaise (Eds.), Engaging young
children in mathematics: Standards for early childhood mathematics education (pp. 7-75). Mahwah, NJ: Erlbaum.
Casey, M.B., Nutall, R.L., and Pezaris, E. (2001). Spatial-mechanical reasoning skills versus mathematics self-
confidence as mediators of gender differences on mathematics subsets using cross-national gender-based items.
Journal for Research in Mathematics Education, 32, 28–57.
College Board. (2006, August 29). College Board announces scores for new SAT with writing section. Retrieved
November 15, 2006, from http://www.collegeboard.com/press/releases/150054.html.
Connor, J.M., Schackman, M., and Serbin, L.A. (1978). Sex-related differences in response to practice on a
visual-spatial test and generalization to a related test. Child Development, 49, 24–29.
Cordova, D., and Lepper, M.R. (1996). Intrinsic motivation and the process of learning: Beneficial effects of
contextualization, personalization, and choice. Journal of Educational Psychology, 88(4), 715–730.
Davies, P.G., Spencer, S.J., Quinn, D.M., and Gerhardstein, R. (2002). Consuming images: How television
commercials that elicit stereotype threat can restrain women academically and professionally. Personality
and Social Psychology Bulletin, 28, 1615–1628.
DeLisi, R., and Wolford, J.L. (2002). Improving children’s mental rotation accuracy with computer game playing.
Journal of Genetic Psychology, 163(3), 272–282.
Deno, J.A. (1995). The relationship of previous experiences to spatial visualization ability. Engineering Design
Graphics Journal, 59, 5–17.
Doolittle, A.E. (1989). Gender differences in performance on mathematics achievement items. Applied Measure-
ment in Education, 2, 161–178.
Dweck, C.S. (1999). Self-theories: Their role in motivation, personality, and development. Philadelphia:
Taylor & Francis.
Dweck, C.S. (2002). Messages that motivate: How praise molds students’ beliefs, motivation, and performance (in
surprising ways). In J. Aronson (Ed.), Improving academic achievement: Impact of psychological factors on educa-
tion (pp. 37–60). San Diego: Academic Press.
Dweck, C.S. (2006). Is math a gift? Beliefs that put females at risk. In S.J. Ceci and W. Williams (Eds.),
Why aren’t more women in science? Top researchers debate the evidence. Washington, DC: American
Psychological Association.
Dweck, C.S., and Leggett, E.L. (1988). A social cognitive approach to motivation and personality. Psychological
Review, 95, 256–273.
Elawar, M.C., and Corno, L. (1985). A factorial experiment in teachers’ written feedback on student homework:
Changing teacher behavior a little rather than a lot. Journal of Educational Psychology, 77(2), 162–173.
Field, M.J., and Lohr, K.N. (Eds.). (1990). Clinical practice guidelines: Directions for a new program. Washington,
DC: National Academy Press.
Foersterling, F. (1985). Attributional retraining: A review. Psychological Bulletin, 98, 495–512.
( 42 )
Foote, C.J. (1999). Attribution feedback in the elementary classroom. Journal of Research in Childhood Education,
13(2), 155–166.
Fouad, N.A., and Smith, P.L. (1996). A test of a social cognitive model for middle school students math and
science. Journal of Counseling Psychology, 43(3), 338–346.
Gallagher, A.M., De Lisi, R., Holst, P.C., McGillicuddy-De Lisi, A.V., Morely, M., and Cahalan, C. (2000).
Gender differences in advanced mathematical problem solving. Journal of Experimental Child Psychology,
75, 165–190.
Gallagher, A.M., and Kaufman, J.C. (Eds.) (2005). Gender differences in mathematics: An integrative psychological
approach. New York: Cambridge University Press.
Geary, D.C., Saults, S.J., Liu, F., and Hoard, M.K. (2000). Sex differences in spatial cognition, computational
fluency, and arithmetical reasoning. Journal of Experimental Child Psychology, 77, 337–353.
Gerson, H.B.P., Sorby, S.A., Wysocki, A.F., and Baartmans, B.J. (2001). The development and assessment of
multimedia software for improving 3-D spatial visualization skills. Computer Applications in Engineering
Education, 9, 105–113.
Gierl, M.J., Bisanz, J., Bisanz, G.L., and Boughton, K.A. (2003). Identifying content and cognitive skills that
produce gender differences in mathematics: A demonstration of the multidimensionality-based DIF analysis.
Journal of Educational Measurement, 40, 281–306.
Ginsburg-Block, M., and Fantuzzo, J. (1998). An evaluation of the relative effectiveness of NCTM standards-based
interventions for low-achieving urban elementary students. Journal of Educational Psychology, 90(3), 560–569.
Gonzales, P.M., Blanton, H., and Williams, K.J. (2002). The effects of stereotype threat and double-minority status
on the test performance of Latino women. Personality and Social Psychology Bulletin, 28(5), 659–670.
43
Good, C., and Aronson, J. (2007). The development of stereotype threat: Consequences for educational and social
equality. In E. Turiel, J. Smetana, and C. Wainryb (Eds.), Social development, social inequalities, and social
justice (pp. 155-185). Mahwah, NJ: Erlbaum.
Good, C., Aronson, J., and Inzlicht, M. (2003). Improving adolescents’ standardized test performance:
An intervention to reduce the effects of stereotype threat. Applied Developmental Psychology, 24, 645–662.
Graham, S. (1991). A review of attribution theory in achievement contexts. Educational Psychology Review,
3, 5–39.
Grant, H., and Dweck, C.S. (2003). Clarifying achievement goals and their impact. Journal of Personality and
Social Psychology, 85(3), 541–553.
Guthrie, J.T., Anderson, E.A., Alao, S., and Rhinehart, J. (1999). Influences of concept-oriented reading instruc-
tion on strategy use and conceptual learning from text. Elementary School Journal, 99(4), 343–366.
Guthrie, J.T., Wigfield, A., Barbosa, P., Perencevich, K., Taboada, A., Davis, M.H., Scafiddi, N.T., and
Tonks, S. (2004). Increasing reading comprehension and engagement through concept-oriented reading
instruction. Journal of Educational Psychology, 96(3), 403–423.
Guthrie, J.T., Wigfield, A., Humenick, N.M., Perencevich, K.C., Taboada, A., and Barbosa, P. (2006).
Influences of stimulating tasks on reading motivation and comprehension. The Journal of Educational Research,
99(4), 232–245.
( 43 )
PRACTICE GUIDE
Hackett, G. (1985). The role of mathematics self-efficacy in the choice of math-related majors of college women
and men: A path analysis. Journal of Counseling Psychology, 32, 47–56.
Harris, A.M., and Carlton, S.T. (1993). Patterns of gender differences on mathematics items on the Scholastic
Aptitude Test. Applied Measurement in Education, 6, 137–151.
Halpern, D.F. (2000). Sex differences in cognitive abilities (3rd ed.). Mahwah, NJ: Erlbaum.
Halpern, D.F., Benbow, C.P., Geary, D.C., Gur, R.C., Hyde, J.S., and Gernsbacher, M.A. (2007). The science of
sex differences in science and mathematics. Psychological Science in the Public Interest, 8(1), 1–51.
Henderlong, J., and Lepper, M.R. (2002). The effects of praise on children’s intrinsic motivation: A review and
synthesis. Psychological Bulletin, 128(5), 774–795.
Herbert, J., and Stipek, D. (2005). The emergence of gender differences in children’s perceptions of their academic
competence. Applied Developmental Psychology, 26, 276–295.
Hidi, S., and Renninger, K.A. (2006). The four-phase model of interest development. Educational Psychologist,
41(2), 111–127.
Hyde, J.S. (2005). The gender similarities hypothesis. American Psychologist, 60(6), 581–592.
Inzlicht, M., Aronson, J., Good, C., and McKay, L. (2006). A particular resiliency to threatening environments.
Journal of Experimental Social Psychology, 42, 323–336.
Inzlicht, M., and Ben-Zeev, T. (2000). A threatening intellectual environment: Why females are susceptible to
experiencing problem-solving deficits in the presence of males. Psychological Science, 11, 365–371.
Jacobs, J.E., Lanza, S., Osgood, D.W., Eccles, J.S., and Wigfield, A. (2002). Changes in children’s self-competence
and values: Gender and domain differences across grades one through twelve. Child Development, 73, 509–527.
Ji, P.Y., Lapan, R.T., and Tate, K. (2004). Vocational interests and career efficacy expectations in relation to occupa-
tional sex-typing beliefs for eighth grade students. Journal of Career Development, 31(2), 143–154.
Kaelin, M.A., Huebner, W.W., Nicolich, M.J., and Kimbrough, M.L. (2007). Field test of an epidemiology cur-
riculum for middle school students. American Journal of Health Education, 38(1), 16–31.
Kurdek, L.A., and Sinclair, R.J. (2001). Predicting reading and mathematics achievement in fourth-grade children
from kindergarten readiness scores. Journal of Educational Psychology, 93, 451–455.
Lapan, R. T., Adams, A., Turner, S.L., and Hinkleman, J.M. (2000). Seventh graders’ vocational interest and
efficacy expectation patterns. Journal of Career Development, 26(3), 215–229.
Linn, M.C., Lee, H.S., Tinker, R., Husic, F., and Chiu, J.L. (2006). Teaching and assessing knowledge integration
in science. Science, 313(5790), 1049–1050.
Maguire, E.A., Gadian, D.G., Johnsrude, I.S., Good, C.D., Ashburner, J., Frackowiak, R.S., and Frith, C.D.
(2000). Navigation-related structural change in the hippocampi of taxi drivers. Proceedings of the National
Academy of Sciences of the United States of America, 97, 4398–4403.
Maguire, E.A., Spiers, H.J., Good, C.D., Hartley, T., Frackowiak, R.S.J., and Burgess, N. (2003). Navigation
expertise and the human hippocampus: A structural brain imaging analysis. Hippocampus, 13(2), 250–259.
( 44 )
Marulis, L., Liu, L., Warren, C., Uttal, D., and Newcombe, N. (2007, March). Effects of training or experience on
spatial cognition in children and adults: A meta-analysis. Poster presented at the biennial meetings of the Society
for Research in Child Development, Boston.
Marx, D.M., and Roman, J.S. (2002). Female role models: Protecting women’s math performance. Personality and
Social Psychology Bulletin, 28(9), 1183–1193.
Maton, K., and Hrabowski III, F.A. (2004). Increasing the number of African American PhDs in the sciences and
engineering: A strengths-based approach. American Psychologist, 59(6), 547–556.
McGraw, R., Lubienski, S.T., and Strutchens, M.E. (2006). A closer look at gender in NAEP mathematics achieve-
ment and affect data: Intersections with achievement, race/ethnicity, and socioeconomic status. Journal for
Research in Mathematics Education, 37, 129–150.
McIntyre, R.B., Paulson, R.M., and Lord, C.G. (2003). Alleviating women’s mathematics stereotype threat through
salience of group achievements. Journal of Experimental Social Psychology, 39(1), 83–90.
McKown, C., and Weinstein, R.S. (2003). The development and consequences of stereotype-consciousness in
middle childhood. Child Development, 74, 498–515.
Miller, R.L., Brickman, P., and Bolen, D. (1975). Attribution versus persuasion as a means for modifying behavior.
Journal of Personality and Social Psychology, 31(3), 430–441.
Mitchell, M. (1993). Situational interest: Its multifaceted structure in the secondary school mathematics classroom.
Journal of Educational Psychology, 85(3), 424–436.
Muehrcke, P.C., and Muehrcke, J.O. (1992). Map use: Reading, analysis, and interpretation (3rd ed). Madison, WI:
JP Publications.
Mueller, C.M., and Dweck, C.S. (1998). Praise for intelligence can undermine children’s motivation and perfor-
mance. Journal of Personality and Social Psychology, 75(1), 33–52.
National Science Foundation. (2006a). Science and engineering degrees: 1966–2004. Retrieved September 12,
2007, from http://www.nsf.gov/statistics/nsf07307/pdf/nsf07307.pdf.
National Science Foundation. (2006b). Women, minorities, and persons with disabilities in science and engineering.
Retrieved June 25, 2006, from http://www.nsf.gov/statistics/wmpd/.
National Science Foundation. (2006c). New tools for America’s workforce: Girls in science and engineering. Retrieved
August 15, 2007, from http://www.nsf.gov/publications/.
National Science Foundation. (2006d). Science and engineering indicators 2006. Retrieved August 17, 2007,
from http://www.nsf.gov/statistics/seind06/c3/c3s1.htm.
Newcombe, N. (2002). Maximization of spatial competence: More important than finding the cause of sex differ-
ences. In A. McGillicuddy-De Lisi and R. De Lisi (Eds.), Biology, society, and behavior: The development
of sex differences in cognition. Advances in applied developmental psychology, Vol. 21. (pp. 183–206). New York:
Ablex Publishing.
Pajares, F. (2006). Self-efficacy during childhood and adolescence: Implications for teachers and parents.
In F. Pajares and T. Urdan (Eds.), Self-efficacy beliefs of adolescents (pp. 339–367). Greenwich, CT: Information
Age Publishing.
( 45 )
PRACTICE GUIDE
Parker, L., and Lepper, M.R. (1992). Effects of fantasy contexts on children’s learning and motivation: Making
learning more fun. Journal of Personality and Social Psychology, 62, 625–633.
Phelps, E., and Damon, W. (1989). Problem solving with equals: Peer collaboration as a context for learning
mathematics and spatial concepts. Journal of Educational Psychology, 81(4), 639–646.
Piburn, M.D., Reynolds, S.J., McAuliffe, C., Leedy, D.E., Birk, J.P., and Johnson, J.K. (2005). The role of visualiza-
tion in learning from computer-based images. International Journal of Science Education, 27, 513–527.
Renninger, K.A., Ewen, L., and Lasher, A. (2002). Individual interest as context in expository text and mathemati-
cal word problems. Learning and Instruction, 12, 467–491.
Schweinle, A., Turner, J.C., and Meyer, D.K. (2006). Striking the right balance: Students’ motivation and affect in
elementary mathematics. The Journal of Educational Research, 99(5), 271–290.
Shea, D.L., Lubinski, D., and Benbow, C.P. (2001). Importance of assessing spatial ability in intellectually talented
young adolescents: A 20-year longitudinal study. Journal of Educational Psychology, 93, 604–614.
Shettle, C., Roey, S., Mordica, J., Perkins, R., Nord, C., Teodorovic, J., Brown, J., Lyons, M., Averett, C.,
and Kastberg, D. (2007). The nation’s report card: America’s high school graduates (NCES 2007-467).
National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.
Washington, DC.
Simpkins, S.D., and Davis-Kean, P.E. (2005). The intersection between self-concept and values: Links between
beliefs and choices in high school. New Directions for Child and Adolescent Development, 110, 31–47.
Simpkins, S.D., Davis-Kean, P.E., and Eccles, J.S. (2006). Math and science motivation: A longitudinal examina-
tion of the links between choices and beliefs. Developmental Psychology, 42, 70–83.
Sorby, S.A. (2001). A course in spatial visualization and its impact on the retention of women engineering students.
Journal of Women and Minorities in Science and Engineering, 7, 153–172.
Sorby, S.A., and Baartmans, B.J. (2000). The development and assessment of a course for enhancing the 3-D spatial
skills of first year engineering students. Journal of Engineering Education, 89, 301–307.
Spelke, E.S. (2005). Sex differences in intrinsic aptitude for mathematics and science? A critical review.
American Psychologist, 60(9), 950–958.
Spencer, S.J., Steele, C.M., and Quinn, D.M. (1999). Stereotype threat and women’s math performance. Journal of
Experimental Social Psychology, 35(1), 4–28.
Stangor, C., Carr, C., and Kiang, L. (1998). Activating stereotypes undermines task performance expectations.
Journal of Personality and Social Psychology, 75, 1191–1197.
Steele, C.M., and Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans.
Journal of Personality and Social Psychology, 69(5), 797–811.
Steele, C.M., Spencer, S., and Aronson, J. (2002). Contending with group image: The psychology of stereotype
and social identity threat. In M. Zanna (Ed.), Advances in experimental social psychology, Vol. 37. San Diego:
Academic Press.
Steele, J. (2003). Children’s gender stereotypes about math: The role of stereotype stratification. Journal of Applied
Social Psychology, 33(12), 2587–2606.
( 46 )
Summers, M.F., and Hrabowski III, F.A. (2006). Preparing minority scientists and engineers. Science, 311(5769),
1870–1871.
Tierney, J.P., and Grossman, J.B. (2000). Making a difference: An impact study of Big Brothers/Big Sisters.
Retrieved May 18, 2007, from http://www.ppv.org/ppv/publications/publications_description.asp?search_
id=7&publication_id=111.
Turner, J.C., Midgley, C., Meyer, D.K., Gheen, M., Anderman, E.M., Kang, Y., and Patrick, H. (2002). The
classroom environment and students’ reports of avoidance strategies in mathematics: A multimethod study.
Journal of Educational Psychology, 94(1), 88–106.
Turner, S.L., and Lapan, R.T. (2005). Evaluation of an intervention to increase non-traditional career interests and
career-related self-efficacy among middle-school adolescents. Journal of Vocational Behavior, 66, 516–531.
Updegraff, K.A., and Eccles, J.S. (1996). Course enrollment as self-regulatory behavior: Who takes optional high
school math courses? Learning and Individual Differences, 8, 239–259.
Utman, C.H. (1997). Performance effects of motivational state: A meta-analysis. Personality and Social Psychology
Review, 1, 170–182.
Wainer, H., and Steinberg, L.S. (1992). Sex differences in performance on mathematics section of the Scholastic
Aptitude Test: A bidirectional validity study. Harvard Educational Review, 62(3), 323–336.
Ward, J. (1976). Behaviour modification in education: An overview and a model for programme implementation.
Bulletin of the British Psychological Society, 29, 257–268.
Webb, R.M., Lubinski, D., and Benbow, C.P. (2002). Mathematically facile adolescents with math-science aspira-
tions: New perspectives on their educational and vocational development. Journal of Educational Psychology,
94(4), 785–794.
Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer-Verlag.
Wigfield, A., Eccles, J.S., Mac Iver, D., Reuman, D.A., and Midgley, C. (1991). Transitions during early adoles-
cence: Changes in children’s domain-specific self-perceptions and general self-esteem across the transition to
junior high school. Developmental Psychology, 27, 552–565.
Wigfield, A., Eccles, J.S., Schiefele, U., Roser, R.W., and Davis-Kean, P. (2006). Development of achievement
motivation. In W. Damon and R. M. Lerner (Series Eds.) and N. Eisenberg (Vol. Ed.), Handbook of child
psychology, Vol. 3: Social, emotional, and personality development (6th ed). New York: Wiley.
Wilson, T.D., and Linville, P.W. (1985). Improving the performance of college freshmen with attributional tech-
niques. Journal of Personality and Social Psychology, 49, 287–293.
( 47 )

Encouraging Girls in Math and Science: IES Practice Guide

Uploaded by

Copyright:

Available Formats

Encouraging Girls in Math and Science: IES Practice Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Encouraging Girls in Math and Science: IES Practice Guide

Uploaded by

Copyright:

Available Formats

Encouraging Girls in

Math and Science

Diane F. Halpern (Chair)

U.S. Department of Education

Institute of Education Sciences

National Center for Education Research

This report is available for download on the IES website at http://ncer.ed.gov.

About the authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Disclosures of potential conﬂicts of interest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Encouraging girls in math and science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Recommendation 2: Provide prescriptive, informational feedback. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Recommendation 4: Create a classroom environment that sparks initial curiosity

Recommendation 5: Provide spatial skills training.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Appendix: Technical information on the studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Table 1. Institute of Education Sciences Levels of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Table 2. Recommendations and corresponding Levels of Evidence to support each . . . . . . . . . . . . . . . 7

Figure 1. Percent of degrees awarded to women by major ﬁeld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Preamble from the Practice guides can also be distinguished from

One unique feature of IES-sponsored practice guides is

About the authors

Joshua Aronson is an associate professor of Develop-

Nona Reimer has been an elementary school teacher

Introduction assurance to the target population because, for example,

A unique feature of practice guides is the explicit and

Strong refers to consistent and generalizable evidence

Table 1. Institute of Education Sciences Levels of Evidence

In general, characterization of the evidence for a recommendation as moderate requires studies

Encouraging girls in chemistry, 46 percent in mathematics, 25 percent

Figure 1. Percent of degrees awarded to women by major ﬁeld

SOURCE: National Science Foundation, Science and Engineering Degrees: 1966-2004.

Table 2. Recommendations and corresponding Levels of Evidence to support each

2. Teachers should provide students with prescriptive, informational feedback

5. Teachers should provide opportunities for students to engage in spatial skills

Girls are more likely to choose courses and careers

Teach students that working hard to learn new

Recommendation 1: Teach students that academic abilities

route finding caused the increased size of the relevant

Roadblock 1.2. By the time students enter high

Solution. Teachers will need to keep in mind that girls

Recommendation 2: Provide prescriptive, informational feedback.

Solution. Teachers can draw attention to students’

Roadblock 2.2. Teachers whose schedules are already

Solution. Feedback or praise does not need to be

Roadblock 2.3. Many teachers rely heavily on stan-

Solution. Effective math and science programs provide

Recommendation 3: Expose girls to female role models who have

Level of evidence: Low Brief summary of evidence to support

Aeronautical Engineering .3 8.4 17.8 .8 5.1 17.1 0 4.0 11.9

Chemical Engineering .8 23.4 35.4 .7 15.7 27.7 .5 8.1 23.9

Civil Engineering .4 13.8 24.2 .4 11.7 27.2 0 5.1 19.6

Electrical Engineering .3 11.5 14.2 .6 8.8 19.6 .4 4.9 13.5

Industrial Engineering .4 29.1 33.2 .5 15.5 21.3 0 6.5 19.4

Mechanical Engineering .2 10.5 13.6 .3 7.0 12.4 .2 5.1 11.1

Materials & Metallurgical

Other Engineering Subﬁelds .8 16.0 29.9 .8 11.5 24.7 .4 6.3 24.9

of minority students pursuing advanced degrees in

Potential roadblocks and solutions

Roadblock 3.1. School hours alone may not provide

Solution. Teachers need to encourage parents to take