Stereotypes Contribute to Gender Imbalances in STEM Fields

Full text

1
Stereotypes Contribute to Gender Imbalances in STEM Fields
Objectives
The gender gap in participation in science, technology, engineering, and mathematics
(STEM) remains a large and persistent educational problem. Women are less likely to major in
STEM fields in college, and less likely to enter STEM careers (U.S. Department of Education,
2017). This gender gap can be traced back to children’s motivation. Girls report lower interest
and confidence in STEM, as early as elementary school (Authors, 2017; Mullis et al., 2008;
Nagy et al., 2010).
Yet STEM fields are not homogenous, and differ in their representation of women. Some
STEM fields have achieved near-equality in gender representation at the college level. For
example, in 2016, women in the United States earned 60% of bachelor’s degrees in biological
sciences and 43% of degrees in math and statistics (U.S. Department of Education, 2017). In
contrast, women’s representation was significantly lower in fields like computer science (19%)
and engineering (20%). Any full explanations for women’s underrepresentation in STEM must
account for these large variations among STEM fields (Ceci et al., 2014). Previous explanations
for women’s underrepresentation have pointed to the math-based nature of STEM fields (e.g.,
Wang et al., 2013), but this does not explain why women are proportionally more likely to major
in math than computer science. Recent evidence suggests that the most likely explanations
involve gender differences in preferences and choices (i.e., motivation) rather than abilities and
performance (Riegle-Crumb et al., 2012).
One potential source of variation in motivation is differences in the cultural stereotypes
about particular STEM fields (Cheryan et al., 2017). There are pervasive cultural stereotypes in
Western countries associating boys and men with STEM (Authors, 2015). Recent research

2
findings have demonstrated that these stereotypes are linked to gender differences in
participation and interest in different academic fields. For example, the more that people within a
field endorse the belief that “natural” talent or genius is required for that field (with “genius”
stereotypically linked to males), the lower the proportion of women in the field (Leslie et al.,
2015). These beliefs are more frequent within fields such as computer science and engineering.
Science-gender stereotypes are also correlated with women’s lower participation in science at the
college and professional levels cross-nationally (Miller et al., 2015). Stereotypes may also be
important for children’s budding interests about academic domains. One study found that 6-year-
old children were more likely to endorse stereotypes that boys were better at programming and
robotics, compared to math and science, a pattern that paralleled adult women’s participation in
those fields (Authors, 2017).
We measured adolescents’ gender stereotypes about who is better and their interest across
four STEM fields, including (as a novel contribution of this project) computer science and
engineering. We predicted that girls would report the strongest stereotypes and the lowest
interest for computer science and engineering compared to math and science. We predicted that
adolescents’ stereotypes would correlate with their interest in those specific fields. We also
expected that this relation would be moderated by gender, with stereotypes more negatively
related to interest among girls compared to boys.
Theoretical Framework
Our model (see Figure 1) is rooted in expectancy-value theory, which shows how
individuals’ perceptions of stereotypes and gender influence their expectations of success in a
domain, which in turn influence their interest in engaging with that domain (Eccles, 2011).
Stereotypes about gender (e.g., that girls have less ability in computer science than boys) may

3
make girls less confident that they have the ability to succeed in that field. With lower
confidence and expectations of future success, girls may be less interested in pursuing
opportunities in that field. This model is also consistent with balanced identity theory, in which
children’s gender identity, gender-STEM stereotypes, and STEM self-concepts are organized in
a consistent way (Cvencek et al., 2011; Cvencek et al., 2014). Previous research has found that
girls’ gender stereotypes about ability are correlated with ability self-concepts in STEM fields
including math, physics, computer science, and engineering (Authors, 2017; Passolunghi et al.,
2014; Plante et al., 2013; Steffens et al., 2010). In turn, ability self-concepts predict motivation to
pursue STEM fields (Beyer, 2014; Eccles & Wang, 2016; Simpkins et al., 2006).
Middle-school students often report counter-stereotypes that girls are better in math and
science than boys (Evans et al., 2011; Martinot & Désert, 2007; Passolunghi et al., 2014; Rowley
et al., 2007). Few studies have examined computer science and engineering stereotypes, yet we
have reason to believe that they may be particularly strong and influential in current society
(Authors, 2015). Middle and high school is also an important time frame for career trajectories.
The choice of whether students intend to pursue STEM careers in particular often occurs during
middle (Tai et al., 2006) or high school (Hall et al., 2011). Thus, we examined students’
stereotypes and interest across both middle and high school.
Partnership and Collaboration
This research project is well-suited for this year’s AERA Theme, “When Researchers and
Organizational Stakeholders Collaborate.” This project is a result of a partnership between the
Authors (university-based researchers) and the Computer Science For Rhode Island (CS4RI)
team. In 2016, the Rhode Island Department of Education’s (RIDE) Office of Innovation reached
out to the Authors to solicit their expert advice on how to improve equity in computer science

4
education while enacting a governor’s initiative to put computer science into all K-12 schools.
We began a Researcher-Practitioner Partnership, in which the Authors conducted research within
Rhode Island schools in close consultation with the CS4RI organization within RIDE. The
CS4RI team is composed of organizational leaders from education, government, and local
industry who have contributed research questions. The Authors have attended CS4RI team
meetings to learn more about the challenges of broadening access to computer science education
in authentic educational settings on a statewide level. Research findings and data have
continuously been shared.
Methods
Participants
Participants were 1,233 students in grades 6 through 12 in a suburban school district in
Rhode Island. A minimum target sample size (N = 882) was chosen based on effect sizes from a
pilot study. Families were sent an opt-out consent form about the research project. Students took
the survey using the online program Qualtrics. Participants were excluded from analyses if they
failed an attention check (“Please mark ‘slightly disagree’ to show that you read this item”),
leaving a final sample of N = 1,041 students. For this sample, 509 students self-identified as
boys, 503 students as girls, and 26 students as using a different word, with 3 missing. Only
children who identified as boys or girls were included in these analyses. In this district, 43% of
students receive free/reduced lunch. Students self-reported their race/ethnicity as 36% White,
26% Latinx, 17% Multiple ethnicities, 10% Asian American, 8% Black/African American; 1%
Other, and 1% Native American, with 1% missing. In terms of maternal education level, 9%
reported no high school degree; 18% high school degree; 38% some college/college degree; 17%
some grad school/graduate degree; and 18% missing.

5
Measures
Stereotypes. Stereotypes were measured with two questions: “How good do you think
most girls are at the following subjects?” and “How good do you think most boys are at the
following subjects?” on a Likert scale from 1 (Really not good) to 6 (Really good). Students
separately rated four different fields for each question: math, science, computer coding, and
engineering. We defined computer coding and engineering for the students. The order of the four
fields was random and balanced across participants; each participant saw the same order for all
questions. Stereotype score was calculated by taking the difference between ratings about boys
minus ratings about girls, so that positive scores indicated traditional stereotypes that boys were
better, 0 represented neutral, and negative scores represented counter-stereotypes that girls were
better. Test-retest correlations for middle-schoolers were stable over four months, ps ≤ .001.
Interest. Interest was measured with two items, “How much do you agree with the
following statements: I like to do [field] activities,” “I am interested in [field] activities,” on a
Likert scale from 1 (Strongly disagree) to 6 (Strongly agree). Correlations between the two items
for each field ranged from r = .83 to .88, ps < .001, indicating high reliability, so items were
averaged for each field.
Results
See Tables 1 and 2 for descriptive statistics. One strength of this study is that the
hypotheses and analyses were pre-registered on the Open Science Framework. All results below
support the pre-registered hypotheses.
Correlations between stereotypes and interest. Boys showed positive correlations
between stereotypes and interest in each field (significant for all fields except computer coding),
while girls showed negative correlations for each field (significant for all fields). The gender

6
difference in correlations was significant for each field, all zs > 2.95, ps < .001. See Table 3 for
full correlations, and Table 4 for same-field correlations by school level.
Differentiating stereotypes in different STEM fields. Because we were interested in
comparing computer coding and engineering, r(1,036) = .37, to math and science, r(1,035) = .48,
we grouped the fields into pairs for analyses and calculated the average stereotype for both pairs.
To see stereotype means for each field separately, see Figure 2A. Girls reported significantly
stronger stereotypes (that boys are better than girls) for coding/engineering, M = .43, SD = 1.04,
compared to math/science, M = -.40, SD = 0.96, t(502) = -15.58, p < .001, dz = .70. (Boys also
reported stronger stereotypes for coding/engineering than math/science, p < .001.)
Effects of gender and field on interest. Again, we grouped coding/engineering, r(1,038) =
.37, and math/science, r(1,038) = .39, and calculated the average interest for both pairs. To see
means for each field separately, see Figure 2B. A 2 × 2 (Participant gender [girl, boy] × Field
[coding/engineering, math/science]) mixed-model ANOVA on interest revealed a significant
interaction, F(1,1009) = 60.47, p < .001, ηp2 = .06. Simple effects for girls revealed that interest
in coding/engineering, M = 3.50, SD = 1.40, was significantly lower than interest in
math/science, M = 4.01, SD = 1.26, p < .001, dz = .38. In contrast, boys’ interest was higher for
coding/engineering, M = 4.30, SD = 1.26, compared to math/science, M = 4.17, SD = 1.18, p =
.022, dz = .10. Looking at the interaction the other way, the gender difference in math/science
interest, p = .041, d = .13, was smaller than the gender difference in coding/engineering interest,
p < .001, d = .61.
Scholarly Significance
We found that girls had stronger stereotypes linking ability with males and reported lower
interest for computer science/engineering compared to math/science. We also replicated previous

7
research showing that middle-school students believe girls are better than boys at math and
science. Students’ stereotypes and their interest were also directly correlated for each field. Boys
who believed that boys were better than girls at math, science, and engineering reported higher
interest in these fields, similar to a “stereotype lift” effect (Walton & Cohen, 2003). In contrast,
girls who believed that boys were better than girls at each field reported lower interest. Because
these results are correlational, the next step involves experimental follow-ups to assess more
closely the causal effects of stereotypes (planned longitudinal intervention studies).
Improving diversity in STEM is an important educational issue. Women are most
strongly underrepresented in fields like computer science and engineering, which leads to deep
inequities because these are high-paying and high-status jobs. These findings provide insight into
one possible mechanism underlying women’s underrepresentation across STEM fields. These
findings have implications for the design of educational interventions to improve diversity in
computer science education. To have the greatest impact in reducing gender disparities in STEM,
we might want to redirect our efforts into computer science and engineering, rather than solely to
math or science. Successful approaches may include changing girls’ perceptions of computer
science and engineering by broadening the image of who succeeds in these fields, and by giving
girls more opportunities to experience these fields and build self-efficacy (Cheryan et al., 2017).
Changing (and diversifying) messages about who can succeed in STEM can help increase equity
in STEM.
Word Count: 1,999

8
References
Authors. (2015).
Authors. (2017).
Beyer, S. (2014). Why are women underrepresented in Computer Science? Gender differences in
stereotypes, self-efficacy, values, and interests and predictors of future CS course-taking
and grades. Computer Science Education, 24, 153-192.
Ceci, S. J., Ginther, D. K., Kahn, S., & Williams, W. M. (2014). Women in academic science: A
changing landscape. Psychological Science in the Public Interest, 15, 75-141.
Cheryan, S., Ziegler, S. A., Montoya, A. K., & Jiang, L. (2017). Why are some STEM fields
more gender balanced than others? Psychological Bulletin, 143, 1-35.
Cvencek, D., Meltzoff, A. N., & Greenwald, A. G. (2011). Math–gender stereotypes in
elementary school children. Child Development, 82, 766-779.
Cvencek, D., Meltzoff, A. N., & Kapur, M. (2014). Cognitive consistency and math–gender
stereotypes in Singaporean children. Journal of Experimental Child Psychology, 117, 73-
91.
Eccles, J. (2011). Gendered educational and occupational choices: Applying the Eccles et al.
model of achievement-related choices. International Journal of Behavioral Development,
35, 195-201.
Eccles, J. S., & Wang, M. T. (2016). What motivates females and males to pursue careers in
mathematics and science? International Journal of Behavioral Development, 40, 100-
106.
Evans, A. B., Copping, K. E., Rowley, S. J., & Kurtz-Costes, B. (2011). Academic self-concept
in Black adolescents: Do race and gender stereotypes matter? Self and Identity, 10, 263-

9
277.
Hall, C., Dickerson, J., Batts, D., Kauffmann, P., & Bosse, M. (2011). Are we missing
opportunities to encourage interest in STEM fields? Journal of Technology
Education, 23, 32-46.
Leslie, S.-J., Cimpian, A., Meyer, M., & Freeland, E. (2015). Expectations of brilliance underlie
gender distributions across academic disciplines. Science, 347, 262-265.
Martinot, D., & Désert, M. (2007). Awareness of a gender stereotype, personal beliefs and self-
perceptions regarding math ability: When boys do not surpass girls. Social Psychology of
Education, 10, 455-471.
Miller, D. I., Eagly, A. H., & Linn, M. C. (2015). Women’s representation in science predicts
national gender-science stereotypes: Evidence from 66 nations. Journal of Educational
Psychology, 107, 631-644.
Mullis, I. V. S., Martin, M. O., & Foy, P. (with Olson, J. F., Preuschoff, C., Erberber, E., Arora,
A., & Galia, J.) (2008). TIMSS 2007 international mathematics report: Findings from
IEA’s trends in international mathematics and science study at the fourth and eighth
grades. Chestnut Hill, MA: TIMSS & PIRLS International Study Center, Boston College.
Nagy, G., Watt, H. M. G., Eccles, J. S., Trautwein, U., Lüdtke, O., & Baumert, J. (2010). The
development of students’ mathematics self-concept in relation to gender: Different
countries, different trajectories? Journal of Research on Adolescence, 20, 482-506.
Passolunghi, M. C., Rueda Ferreira, T. I., & Tomasetto, C. (2014). Math-gender stereotypes and
math-related beliefs in childhood and early adolescence. Learning and Individual
Differences, 34, 70-76.
Plante, I., De la Sablonnière, R., Aronson, J. M., & Théorêt, M. (2013). Gender stereotype

10
endorsement and achievement-related outcomes: The role of competence beliefs and task
values. Contemporary Educational Psychology, 38, 225-235.
Riegle-Crumb, C., King, B., Grodsky, E., & Muller, C. (2012). The more things change, the more
they stay the same? Prior achievement fails to explain gender inequality in entry into
STEM college majors over time. American Educational Research Journal, 49, 1048-
1073.
Rowley, S. J., Kurtz
-
Costes, B., Mistry, R., & Feagans, L. (2007). Social status as a predictor of
race and gender stereotypes in late childhood and early adolescence. Social
Development, 16, 150-168.
Simpkins, S. D., Davis-Kean, P. E., & Eccles, J. S. (2006). Math and science motivation: A
longitudinal examination of the links between choices and beliefs. Developmental
Psychology, 42, 70-83.
Steffens, M. C., Jelenec, P., & Noack, P. (2010). On the leaky math pipeline: Comparing implicit
math-gender stereotypes and math withdrawal in female and male children and
adolescents. Journal of Educational Psychology, 102, 947-963.
Tai, R. H., Liu, C. Q., Maltese, A. V., & Fan, X. (2006). Planning early for careers in science.
Science, 312, 1143-1144.
U.S. Department of Education (2017). Table 318.30. Bachelor’s, master’s, and doctor’s degrees
conferred by postsecondary institutions, by sex of student and discipline division: 2015-
16. National Center for Education Statistics. Retrieved from https://nces.ed.gov/programs
/digest/d17/tables/dt17_318.30.asp
Walton, G. M., & Cohen, G. L. (2003). Stereotype lift. Journal of Experimental Social
Psychology, 39, 456-467.

11
Wang, M.-T., Eccles, J. S., & Kenny, S. (2013). Not lack of ability but more choice: Individual
and gender differences in choice of careers in science, technology, engineering, and
mathematics. Psychological Science, 24, 770-775.

12
Table 1
Descriptive Statistics for STEM–Gender Stereotypes
Overall
Girls
Boys
Gender
Field
M
SD
M
SD
M
SD
difference
Math
-.39
1.24
-.42
1.19
-.36
1.29
p = .40
Science
-.27
1.09
-.37
1.02
-.20
1.15
p = .011
Computer coding
.45
1.29
.34
1.21
.55
1.35
p = .008
Engineering
.60
1.27
.52
1.25
.68
1.29
p = .041
Note. Range: -5 (girls a lot better) to 5 (boys a lot better). All measures were significantly
different from neutral (0), ps < .001.

13
Table 2
Descriptive Statistics for Interest
Overall
Girls
Boys
Gender
Field
M
SD
M
SD
M
SD
difference
Math
3.81*
1.55
3.78*
1.57
3.84*
1.53
p = .54
Science
4.38*
1.36
4.24*
1.44
4.50*
1.28
p = .003
Computer coding
3.84*
1.56
3.45
1.54
4.21*
1.48
p < .001
Engineering
3.97*
1.49
3.54
1.50
4.40*
1.36
p < .001
Note. Range: 1 (strongly disagree) to 6 (strongly agree) for items including “I like to do [field]
activities.”
Overall significantly different from scale midpoint: *p < .001

14
Table 3
Correlations Between Stereotypes and Interest
1
2
3
4
5
6
7
8
1. Math stereotypes
.50***
.21***
.18***
-.13**
-.07
-.03
-.02
2. Science stereotypes
.26***
.27***
.20***
-.01
-.09*
-.01
-.01
3. Computer coding
stereotypes
.12**
.30***
.43***
-.02
-.09*
-.14**
-.11*
4. Engineering
stereotypes
.12**
.14**
.53***
-.05
-.06
-.08
-.18***
5. Math interest
.22***
.06
.07
.04
.40***
.32***
.39***
6. Science interest
.02
.13**
.07
.00
.40***
.36***
.42***
7. Computer coding
interest
.01
.00
.05
.00
.23***
.34***
.70***
8. Engineering interest
.00
.09*
.14**
.11*
.32***
.43***
.56***
Note. Girls (N = 503) are above the diagonal, and boys (Ns = 506-508) are below the diagonal.
Stereotypes are difference scores for ratings about boys minus ratings about girls. Results shown are from
students (84%) who passed an attention check question in the survey. Same-field correlations are in black,
and different-field correlations are in gray.
*p ≤ .05, **p ≤ .01, ***p ≤ .001

15
Table 4
Same-Field Stereotype-Interest Correlations by Gender and School Level
Math
Science
Computer
Coding
Engineering
Middle
Girls
-.12
-.14*
-.08
-.20***
Boys
.30***
.11
.06
.06
High
Girls
-.14*
-.06
-.16*
-.16*
Boys
.14*
.14*
.02
.14*
Note. *p ≤ .05, **p ≤ .01, ***p ≤ .001

16
Figure 1. Theoretical model showing how stereotypes influence interest in STEM.
INTEREST
ABILITY
STEREOTYPES
ABILITY SELF-CONCEPT
(Will I be good at this?)

17
Figure 2. (A) Stereotypes and (B) Interest by gender and field. Error bars are +/− SE.
-1
-0.5
0
0.5
1
Math Science Computer
Coding
Engineering
Stereotypes
A
Boys
Girls
Boys better
Girls better
1
2
3
4
5
6
Math Science Computer
Coding
Engineering
Interest
Boys
Girls
Strongly
Strongly not
Slightly
Slightly not
B