Paper ID #40243

Why Students Choose STEM: A Study of High School Factors That Influence
College STEM Major Choice
Dr. Joyce B. Main, Purdue University
Joyce B. Main is Associate Professor of Engineering Education at Purdue University. She received an
Ed.M. in Administration, Planning, and Social Policy from the Harvard Graduate School of Education,
and a Ph.D. degree in Learning, Teaching, and Social Policy.
Tram Dang, Purdue University
Tram Dang is a PhD student of Engineering Education at Purdue University as well as a tenured professor
of physics and engineering at Santa Monica College (SMC), a two-year transfer-focused Hispanic serving
institution in California. Her research interests lie in using mixed methods techniques to understand
the factors leading to STEM and engineering student retention and university transfer at the community
college level. Tram Dang currently holds her Master’s in Materials Science and Engineering as well as
her Bachelor’s in Physics from UCSD.
Beata Johnson, Campbell University

Beata Johnson is an Engineering Education Ph.D. student at Purdue University and recipient of an NSF
Graduate Research Fellowship. She received her BS in Chemical Engineering from Purdue University.
Her research interests include extracurricular and exper
Qian Shi, Purdue University
-

Dr. Cesare Guariniello, Purdue University

Cesare Guariniello is a Research Scientist in the School of Aeronautics and Astronautics in Purdue Uni-
versity. He holds two Master’s degrees, in Automation and Robotics Engineering and in Astronautical
Engineering, from the University of Rome ”La Sapienza”, and a PhD in Aeronautics and Astronautics
from Purdue University. His research ranges from System-of-Systems design and architecting to space
applications, cybersecurity, and defense and includes projects with NASA, the US DoD, the US Navy.
Cesare recently expanded his research in the field of Earth Sciences, where he is pursuing a Master’s
degree in Planetary Geology. He is a senior member of IEEE and AIAA, and member of INCOSE.
Daniel Delaurentis, Purdue University

Dr. Daniel DeLaurentis is a Professor at the School of Aeronautics and Astronautics, Purdue University,
where he also serves as Vice President for Discovery Park District (DPD) Institutes. His research centers
on design and system engineering methods for aerospace systems and systems-of-systems. Dr. DeLau-
rentis is Chief Scientist in the DoD Systems Engineering Research Center (SERC) and a Fellow of both
INCOSE and AIAA.

©American Society for Engineering Education, 2023

 High School Student Academic Factors Associated with College-Going
and STEM Major Choice
Abstract

As our national workforce needs continue to grow, attracting and retaining postsecondary
students in the science, technology, engineering, and mathematics (STEM) fields remains a
priority. A student’s interest in a STEM major often begins at the precollege level, and their
precollege experience can determine their later academic trajectory. While this interest often
develops in middle school or earlier, a student’s high school experience can affect whether a
student maintains or loses their interest. In order to understand a student’s high school
experience, this study focuses on the high school factors, student demographic characteristics,
and academic achievement factors that inform college-going and STEM major choice. For this
study, data come from the High School Longitudinal Study of 2009 (HSLS:09), which is a
nationally representative longitudinal study following over 23,000 students from 2009 to 2016.
The data were analyzed using multiple regression analyses to correlate high school,
demographic, academic achievement factors from the 2009 and 2012 data collection waves to a
student’s likelihood of attending college and majoring in a STEM field. The high school level
factors that were found to be significant predictors for college STEM major declaration include
the student’s family background, high school STEM GPA, and measures for math/science
identity. The findings are mixed and suggest further research is needed, particularly in
disaggregating the math/science self-efficacy, identity, and utility measures, as well as in
investigating potential differences in major choice by field separately, rather than STEM in the
aggregate. Research findings can be used to inform policies and programs aimed at increasing
diversity and inclusivity in STEM fields, as well as to identify areas where additional support
and resources may be needed to help students succeed.

Introduction

Despite recent dips in the economy due to COVID-19, the U.S. expects to see
considerable occupational growth over the next decade. More specifically, the U.S. Bureau of
Labor Statistics predicts that STEM occupations will grow at over twice the rate (10.8%) of non-
STEM occupations (4.9%) between 2021-31 [1]. As our national workforce needs continue to
grow, attracting and retaining postsecondary students in science, technology, engineering, and
mathematics (STEM) fields remain a top priority.

Methods for attracting and retaining students in STEM fields has been debated for
decades and remains a focus of the U.S’s current educational policies; the recent passage of the
CHIPS and Science Act of 2022, for example, reaffirms the White House’s commitment to
growing a pool of STEM workforce graduates by investing in STEM education and training at
all educational levels [2]. Despite the dedication of resources towards STEM, many students still
fall behind in academic achievement, particularly in math. While these educational policies have
led to increasing secondary school science literacy in the U.S., the average U.S. math
achievement scores still falls below the international averages and has remained largely
unchanged since 2003 [3]. Findings regarding math and science interest and confidence seem to
convey a similar story; in one study, while science interest remained unchanged throughout a
student’s high school years, significant losses occurred in science confidence and math interest
and confidence during these years [4]. These are alarming results, as these high school level
math and science courses act as gatekeepers to STEM degrees and careers.

One step towards overcoming this issue is to develop an understanding of what impacts a
student’s choice to attend college and to major in a STEM field. It is well understood that a
student’s interest in STEM develops through exposure at an early age [5]. While this interest is
maintained through middle school, loss of interest often occurs during a student’s high school
years [6]. It becomes imperative that educators, researchers, and policymakers understand the
high school factors that influence college-going and STEM major declaration. Many studies have
looked at the high school experience and its implications on STEM major declaration in college
(e.g. [7]), but few studies follow students longitudinally to examine their pathways from high
school through college. We help address this gap by examining the high-school level factors,
student demographic characteristics, and academic achievement factors that are associated with
college-going and STEM major declaration to contribute to the conversations regarding how we
can collectively help promote STEM workforce talent development. Specifically, we address the
following research questions:

1. Which high school-level factors, student demographic characteristics, and student academic
achievement factors are associated with college-going?
2. Among students who attend a four-year college institution, which high school-level factors,
student demographic characteristics, and student academic achievement factors are
associated with STEM major choice?

Our research findings contribute to the literature that investigates the factors that promote
STEM major choice by leveraging comprehensive, nationally representative data that provide
insights into students’ longitudinal trajectories from high school through postsecondary
education. Identifying the factors that are associated with college-going has the potential to
increase participation in higher education and to expand the pool of students who might be
interested in STEM majors. Examining STEM major choice factors provides stakeholders with
critical information regarding which potential interventions may be more likely to promote
STEM interest, major declaration, and persistence. Our research findings can inform the
development of programs and contribute to our national commitment to growing and
diversifying the STEM workforce.

Literature Review

The focus of this investigation is on student high school experiences, student

demographics, and high school characteristics as factors for predicting a student’s declaration of
a STEM major in postsecondary education. Previous studies using nationally-represented data
that have investigated the relationship between student and high school characteristics on STEM
major choice have drawn data from the Educational Longitudinal Survey of 2002 (ELS:2002)
(e.g., [8]–[10]). Wang [10], for example, studied student demographics and student high school
experiences in 10th and 12th grades as factors for entrance into a STEM field at the postsecondary
level. Her findings showed that math and science exposure, as well as 12th grade math efficacy
and achievement were predictors for STEM major intention. Similar to our study, other studies
draw data from HSLS:09 to study similar high school factors (e.g., [11]–[15]).. Much like studies
drawn from ELS:2002, these studies found similar results; for example, high school students
with an interest in math and science and math ability were more likely to have an intention to
major in STEM at the postsecondary level [12]. Many of the studies using HSLS:09 data,
however, are limited to using data from the first, second, and third waves of data collection,
which do not include information on postsecondary STEM major enrollment and instead
provides information on STEM major aspirations.

Our study extends the literature by incorporating longitudinal data from the HSLS from
the first data collection wave in 2009 through the fourth and final data collection wave in 2016.
Using this data, we examine student demographics, student experiences, and high-school level
factors that lead to a declaration of a STEM major in their postsecondary years.

A. Student high school experience and STEM major choice

High school STEM course taking

Studies on the relationship between high school STEM courses and STEM major choice
highlight the importance of STEM course participation over STEM course availability. Having a
wide variety of STEM courses available at the high school level does not necessarily ensure
student participation or course taking. As a result, having more STEM courses available at the
high school alone does not increase the proportion of students selecting a STEM major later,
which is confirmed by numerous studies [16], [17].

Student participation in STEM courses, however, is an important factor for STEM major
selection later [18]–[22]. Crisp et al. (2009), for example, shows that exposure to math and
science at the high school level is a strong predictor of STEM major choice even when compared
to other strong predictors, such as math achievement. It follows, then, that greater participation in
STEM courses would lead greater odds of STEM major selection. Previous research supports
this trend. Wang [10], for example, found that the number of STEM courses in high school was a
strong indicator of STEM major choice, while Zhang et al. [15], found that earning more credits
in STEM-related courses was significantly related to STEM major choice. Specific to
engineering, Jewitt and Chen [24] showed that students taking more AP STEM courses increased
the odds of becoming an engineering major later. Existing literature suggests, then, that high
school student STEM course taking is a greater indicator of STEM major selection than high
school STEM course offerings alone.

Math and science identity and efficacy

Research on identity and efficacy shows a relationship between math and science identity
and efficacy with STEM major choice. It is possible that a student’s math and science identity is
related though STEM exposure, which in turn can predict a student’s likelihood of declaring a
STEM major later. One study, for example, identified the role of science identity in STEM
participation, which found that science identity significantly predicts student participation in
optional science learning experiences in secondary students [25]. Science identity, in turn,
predicts a commitment to science careers [26]. In addition, students reporting higher math- and
science-related interest and higher self-assessment were more likely to declare a STEM major
later in life [27].

STEM achievement scores

 Interestingly, research shows that a student’s identity plays an important role in academic
achievement [28]–[30]. It follows, then, that a student’s science and math identity can affect a
student’s STEM achievement scores. Current research supports this idea as well; math identity
are strongly and positively associated with math high school GPA in Black [31] and Latine [32]
secondary school students. Since a student’s math and science identity is linked to STEM major
choice, a case can also be made for linking math and science scores to STEM major choice. This
relationship has been confirmed in many studies [10], [14], [15], [20].

B. Student demographic characteristics and STEM major choice

Gender and Race/Ethnicity

The relationship between STEM major choice and gender is well-studied. Studies agree
that gender is a significant predictor of STEM major selection. Mau and Li (2018) showed that
being White and male are strong predictors for selecting STEM majors. Other studies support
this finding, such as from Zhang et al. [15], which showed that half as many females choose
STEMS compared to their male counterparts.

Studies, however, are inconsistent with their findings regarding race as a predictor for
STEM major selection. One study, for example, showed that female, Black, and Hispanic
students are less likely to develop and maintain an interest in STEM careers in high school [33].
In contrast, another study found that Asians are more likely to enter in STEM, but that there are
otherwise no measurable differences between White, Black, and Hispanic students [34], [35].
The inconsistency may be because demographic factors can have a complex relationship with
outcomes such as STEM major choice, and instead mediate other factors such as identity. In one
study, for example, Black female students see lower science identity in secondary school, and in
turn, lower participation in science activities [31, p. 202]. The link between identity,
participation, and STEM major choice is discussed in a previous section.

Parents’ Education

Most studies agree that there is a correlation between a parent’s education level and a
student’s STEM major choice. Many studies report that a parent’s education level is a strong
predictor for STEM degree selection [7], [20]. Other researchers, such as Moakler Jr. and Kim
[35], reported that parental education level is not a strong predictor of STEM major selection, as
the correlation was not significant. Interestingly, this complex relationship may be explained
through other student experience factors such as science and math identity and efficacy. The
same study by Moakler Jr. and Kim [35] noted that students with parents in occupations
requiring STEM degrees are more confident in their math abilities, which in turn, is a statistically
significant indicator for STEM major selection.

Socioeconomic Status

Most researchers agree that there is a positive correlation between a student’s

socioeconomic status (SES) and STEM major selection. G. Saw et al. [33] showed that students
with lower SES are less likely to show, maintain, and develop an interest in STEM during high
school. This agrees with other studies showing that students with low SES are less likely to
select a STEM degree [14], [21]. In contrast, Sahin et al. [36] saw no correlation between
parents’ education and houshold income on STEM degree selection. As seen with parent
education and race and ethnicity, SES may have a complex relationship with STEM major
selection. This is support by Niu [37], who showed that family SES does not predict STEM
enrollment by itself, but that it interacts with several other predicting factors, such as gender,
race, and achievement scores.

C. High school characteristics

Studies on the relationship between a high school’s urbanicity and a student’s interest in
STEM highlight the varying access and availability of resources in rural areas. For example, our
previous discussion shows that student STEM participation is a strong factor for later STEM
major selection; as such, students attending STEM-oriented high schools are more likely to
declare a STEM major later [38]. The availability of STEM-oriented high schools, however, vary
by an area’s urbanicity and are less common in rural areas [39, p. 201], [40]. In agreement with
this, Saw and Agger [41] found that high schools in rural and small-town settings have less AP
math and science class; have less math and science fairs; are less likely to sponsor after school
programs; and are less likely to inform students about extracurricular math and science
programs. Students attending rural high school, then, tend to have less opportunities to engage in
STEM-related activities, and therefore, lower STEM participation.

Conceptual Framework

We draw on Main et al.’s [42] conceptual framework on the factors associated with
engineering major choice across life stages to inform this study. Whereas Main et al.’s
framework focuses on engineering major choice, this study focuses on college-going and STEM
major choice, and thus, we adapt the framework accordingly. The adapted framework is
summarized in Figure 1. Consistent with the life course perspective, we also emphasize the
importance of the different stages of the student pathway from high school through college in
contributing to STEM major choice [43], [44]. Together with findings from previous studies
highlighted in the literature review, Main et al.’s [42] conceptual framework informs the factors
to investigate for this study. Thus, our regression models include factors related to the student’s
demographic characteristics, family background, high school characteristics, academic
achievement, and attitudes and perspectives toward math and science.

Our conceptual framework (Figure 1) highlights the importance of demographic

characteristics and family background on STEM major choice, and we therefore include the
following variables in our regression models: gender, race/ethnicity, parental education, and
family income. We also investigate high school-level factors, including high school level of
urbanicity, percentage of racially minoritized students enrolled at the high school, whether the
high school is public or private, and the percentage of students eligible for free or reduced-priced
lunch (e.g.,[10], [16], [41]). The conceptual framework also indicates that individual student
academic achievement and math and science interest and perceived proficiency are important to
STEM major choice. To model this, we include several variables drawn from the second wave of
HSLS data (11th grade): math identity, math utility, math self-efficacy, science identity, science
utility, and science self-efficacy. We also include academic achievement variables, such as
STEM grade point average, 9th grade math scores, SAT math score, SAT reading score, whether
the student took an Advanced Placement math course during high school, and whether the
student took and Advanced Placement science course during high school.

Figure 1
Factors Associated with STEM Major Choice Across Life Stages
Adapted from Main et al. (2022)

Methods

Our study addresses two research questions surrounding student STEM participation:
First, which high school-level factors, student demographic characteristics, and student academic
achievement factors are associated with college-going? And among students who attend a four-
year college institution, which high school-level factors, student demographic characteristics,
and student academic achievement factors are associated with STEM major choice? Our data
come from the High School Longitudinal Survey, which includes a nationally-representative
sample of students followed longitudinally from their 9th grade year (2009) through
postsecondary education (2016) with transcript information (2017-2018).

We use linear probability models to identify factors that are associated with (1) college-
going and (2) STEM major choice. College-going is defined as whether the student attended any
type of postsecondary institution, 4-year college-goin is whether the student attended a four-year
academic institution, and STEM major choice is whether the student enrolled in a four-year
college declared a major in science, technology, engineering, or mathematics during the 2016
survey wave. The resulting sample size for the linear regression model on college-going is
19,770, which includes all students in the HSLS sample who enrolled in any type of
postsecondary institution (e.g., community college, four-year institution). The outcome variable
for the 4-year college-going model is enrollment at a four-year institution, and the resulting
sample size is 13,050. For the model on STEM declaration, the resulting sample size is 8,890,
which includes all students attending a four-year college and declared any major. The models
were weighted using the “student longitudinal analytic weight” provided by the HSLS.

Following our conceptual model (Figure 1), our explanatory variables are related to the
student’s demographic characteristics, family background, high school characteristics, academic
achievement, and attitudes and perspectives toward math and science. Explanatory variables
include gender, race/ethnicity, parents’ education level (at least one parent who completed high
school, associate’s degree, bachelor’s degree, or master’s degree) with no high school degree as
the comparison group, whether parents have a STEM degree, and categories for family income in
2012 (< $35k, $35-$75k, $75-$155K, and >$155K) with <$35K as the base category. The HSLS
indicates male and female for gender, and the following groups for race/ethnicity: White,
Black/African American, Asian American/Asian, Hispanic/Latine, and Other Race/Ethnicity. In
terms of high school level factors, our models include whether the school is public or private, the
percentage of URM students, percentage of students eligible for free or reduced-price lunch, and
level of urbanicity (city, suburban, town, and rural) with city as the base category.

We also included indices that measure student math identity, math utility, and math self-
efficacy, as well as science identity, science utility, and science self-efficacy [45]. Identity
measures the level of agreement with statements associated with being a “math person” or
“science person,” whereas utility indicates the students’ perceptions of the usefulness of math, as
well as the usefulness of science. Finally, higher self-efficacy represents a higher sense of
competence in the given subject—math or science. These indices are normed with a standardized
mean of 0. For the individual student academic achievement variables, we include STEM high
school grade point average (GPA) on a scale from 0 to 4.0, 9th grade math score (“mathematics
standardized theta score,” which is a norm-referenced measurement of math achievement in the
9th grade), and SAT verbal and math scores, each measured from a scale of 200 to 800 and
normalized for analyses. We also considered whether the student took at least one Advanced
Placement course in math, as well as at least one Advanced Placement course in science, during
high school. We used the same explanatory variables for both models, except that for the model
focusing on college-going, we excluded SAT math and SAT reading because taking these tests
are strongly correlated with college-going. That is, most students who take the SAT have
intentions to pursue postsecondary education. We also did not include the variables associated
with taking advance placement courses because they are also strongly correlated with college-
going.

Data

The HSLS is comprised of a nationally representative sample of 23,000 high school

students from 944 high schools in the U.S. The HSLS provides comprehensive longitudinal data
collected during multiple time points (1) Base Year: 9th grade (2009); (2) First Follow-up: 11th
grade (2012 spring); (3) 2013 Update: high school graduation (2013); (4) Second Follow-up: 3rd
year of college (2016), and (5) Postsecondary Transcripts (2017-2018). For this study, we
primarily draw on the second follow-up wave (2016) to investigate factors associated with
college-going and STEM major choice. In particular, we focus on students who attended college
(n = 13,050) at four-year institutions (n = 8,890). Among students who attended four-year
academic institutions, 2,590 declared a STEM major.

Table 1 presents the summary statistics of our analytic sample. Column 2 includes
summary statistics for all students who enrolled in college. The values indicated are proportions,
except for the following normalized variables, which are shown as means: math/science self-
efficacy, math/science utility, math/science identity, grade point average, math scores, and SAT
scores for math and science. Column 3 includes summary statistics for the sample of students
who attended four-year academic institutions, and column 4 focuses on those who attended four-
year academic institutions and declared a STEM major. Overall, the summary statistics are
consistent with those reported from previous studies and from the National Science Foundation.
Women comprise 49% of students majoring in STEM at four-year colleges, while 9% of STEM
students identified as African American/Black, 7% as Asian American/Asian, and 13%
Hispanic/Latine (Column 4). Students who declared STEM majors at a four-year institution are
more likely to have parents with a bachelor’s degree (33%) and to have parents with a Master’s
degree (28%) compared to students who declared other majors or who attended other types of
postsecondary institutions. Notably, 38% of parents of STEM students have a STEM degree.
STEM students are also more likely to have higher SAT math and reading scores and to have
taken AP math and science courses compared to students attending four-year colleges in other
majors.

Table 1
Summary Statistics

Column 1 Column 2 Column 3 Column 4

Attended 4-
year college
Attended Attended 4- & declared
All students college year college STEM major
Women 49.9 52.9 54.1 48.5
African American/Black 12.9 11.8 11.2 8.6
Asian American/Asian 3.5 4.5 5.2 6.6
Hispanic/Latine 22.0 19.9 15.1 13.4
Other Race/Ethnicity 8.6 8.3 7.7 6.7
Parents' Education: Less than
High School 6.5 4.8 2.3 1.4
Parents' Education: High
School 36.0 30.4 24.8 21.4
Parents' Education:
Associate's Degree 15.1 16.1 14.9 14.2
Parents' Education:
Bachelor's Degree 20.9 25.0 29.1 32.5
Parents' Education: Master's
Degree or Higher 14.7 19.0 24.6 27.4
Parents have STEM Degree 24.3 28.3 32.0 37.7
 Family Income in 2012:
<$35K 31.2 28.6 22.2 17.6
Family Income in 2012: $35-
75K 31.3 28.7 27.6 29.1
Family Income in 2012: $75-
155K 28.4 31.4 36.1 37.8
Family Income in 2012:
$155K+ 9.1 11.2 14.1 15.5
Public School 92.3 89.8 86.0 86.9
Urbanicity: City 31.6 32.3 30.8 29.5
Urbanicity: Suburban 33.5 24.7 37.3 37.1
Urbanicity: Town 11.5 10.5 9.4 9.8
Urbanicity: Rural 23.3 22.4 22.5 23.6
Percentage URM Students 33.0 28.8 25.5 22.9
Percentage Students Eligible
for Free/Reduced Lunch 39.3 32.7 28.8 26.8
11th Grade Math Identity
Index 0.0 0.1 0.2 0.6
Math Utility Index 0.02 0.06 0.11 0.30
Math Self-Efficacy Index 0.02 0.10 0.19 0.45
Science Identity Index 0.04 0.10 0.21 0.61
Science Utility Index 0.00 0.00 0.01 0.03
Science Self-Efficacy Index 0.03 0.08 0.16 0.37
STEM GPA 2.44 2.62 2.88 3.08
9th Grade Math Score 50.8 53.0 55.7 58.3
SAT Math 534 568
SAT Reading 530 550
Taken AP Math Course(s) 13.2 20.4
Taken AP Science Course(s) 18.5 26.3
N 19,770 13,050 8,890 2,590
Notes. The sample sizes have been rounded to the nearest 10. The values indicated are
proportions, except for the following normalized variables, which are shown as means:
math/science self-efficacy, math/science utility, math/science identity, grade point average, math
scores, and SAT scores for math and science.
Data source: U.S. Department of Education, National Center for Education Statistics, High
School Longitudinal Study of 2009 (HSLS:09), Second Follow-up, 2016.

Results

Table 2 shows the marginal effects regression coefficients for our models with column 2
showing the results for whether the student attended a four-year college and column 3 showing
the results for whether students enrolled at a four-year college declared a STEM major versus
another major. We present our findings by research question below.
Research Question 1: Which high school-level factors, student demographic characteristics, and
student academic achievement factors are associated with college-going?

We found that African American/Black students are 13.5 percentage points more likely
than White students to attend a four-year college, all else held constant. Compared to students
whose parents did not receive a high school degree, students whose parents completed high
school or a postsecondary degree are also more likely to attend college. Family income is also
positively associated with college-going. Students who had higher grade point averages from
their high school STEM courses, as well as have higher 9th grade math scores, are also more
likely to pursue STEM majors. The results related to the measures on math/science self-efficacy,
utility, and identity indices are mixed, with the findings suggesting that students who score
higher on math identity, math self-efficacy and science identity are more likely to attend a four-
year academic institution.

Research Question 2: Among students who attend a four-year college institution, which high
school-level factors, student demographic characteristics, and student academic achievement
factors are associated with STEM major choice?

Among students who attended a four-year college, women are less likely than men to
choose a STEM major by 3.1 percentage points, all else held constant. Consistent with previous
studies, students whose parents have a STEM degree are more likely than students whose parents
have a degree from another field to pursue STEM. Again, the results for the math/science self-
efficacy, utility, and identity indices are mixed, where higher math identity and math self-
efficacy, as well as higher science identity and science utility, are associated with STEM major
choice. Higher STEM GPA, higher SAT math scores, and taking at least one advance placement
course in math and science are positively associated with STEM major choice.
Table 2
Regression coefficients on college-going, 4-year college going and STEM major declaration

Column 1 Column 2 Column 3

STEM major
College-going 4-year college declaration at a 4-
model going model year college model
Coefficient SE Coefficient SE Coefficient SE
Women 0.039*** 0.007 0.013 0.009 -0.031** 0.011
African American/Black 0.080*** 0.012 0.135*** 0.015 0.011 0.019
Asian American/Asian 0.079*** 0.019 0.015 0.021 0.026 0.024
Hispanic/Latine 0.051*** 0.010 0.015 0.013 0.032 0.017
Other Race/Ethnicity 0.031* 0.013 0.013 0.016 -0.009 0.020
Parents' Education: High School -0.002 0.015 0.057** 0.022 0.057 0.036
Parents' Education: Associate's Degree 0.103*** 0.017 0.119*** 0.023 0.086* 0.038
Parents' Education: Bachelor's Degree 0.089*** 0.017 0.147*** 0.023 0.086* 0.037
Parents' Education: Master's Degree or Higher 0.088*** 0.018 0.169*** 0.024 0.072 0.038
Parents have STEM Degree 0.013 0.008 0.002 0.010 0.034** 0.012
Family Income in 2012: $35-75K 0.033*** 0.009 0.038** 0.012 0.050** 0.016
Family Income in 2012: $75-155K 0.080*** 0.010 0.068*** 0.013 0.014 0.016
Family Income in 2012: $155K+ 0.094*** 0.015 0.054** 0.017 -0.001 0.021
Public School -0.050*** 0.014 0.139*** 0.016 0.041* 0.017
Urbanicity: Suburban -0.010 0.009 0.040*** 0.011 0.014 0.013
Urbanicity: Town -0.037** 0.012 0.058*** 0.016 0.031 0.020
Urbanicity: Rural -0.044*** 0.010 0.001 0.012 0.028 0.015
Percentage URM Students 0.140*** 0.019 0.010 0.024 -0.027 0.029
Percentage Students Eligible for Free/Reduced Lunch -0.187*** 0.022 0.014 0.027 -0.071* 0.034
11th Grade Math Identity Index -0.016*** 0.005 0.026*** 0.006 0.042*** 0.007
Math Utility Index 0.004 0.004 0.005 0.005 -0.007 0.007
Math Self-Efficacy Index 0.011* 0.005 0.012* 0.006 0.026*** 0.007
Science Identity Index 0.006 0.005 0.017** 0.006 0.079*** 0.007
Science Utility Index -0.010 0.067 0.137 0.084 0.627*** 0.105
Science Self-Efficacy Index 0.000 0.004 0.002 0.005 -0.009 0.007
 STEM GPA 0.192*** 0.005 0.181*** 0.007 0.060*** 0.010
9th Grade Math Score 0.031*** 0.004 0.088*** 0.006 -0.003 0.008
SAT Math 0.056*** 0.011
SAT Reading -0.029** 0.010
Taken AP Math Course(s) 0.065*** 0.017
Taken AP Science Course(s) 0.030* 0.015
N 19,770 13,050 8,890
*p<0.1; **p<0.05; ***p<0.01
Notes. The sample sizes have been rounded to the nearest 10.
Data source: U.S. Department of Education, National Center for Education Statistics, High School Longitudinal Study of 2009
(HSLS:09), Second Follow-up, 2016.
Discussion and Conclusion

Consistent with our conceptual model, we found that demographic characteristics, math
and science academic achievement, and family characteristics, are correlated with STEM major
choice. Overall, our results related to math/science self-efficacy, utility, and identity are mixed.
We found that 11th grade math identity, math self-efficacy, and science identity are each
positively associated with attending a four-year institution and pursuing a STEM major at a four-
year institution. Previous studies that have investigated the impact of these constructs on STEM
major choice tend to focus on only one or two measures, and it may be the combination of the six
different measures that have led to the mixed results (e.g., [25], [26]). It may also be due to the
nature of the sample representing over 944 high schools and a variety of college contexts, which
differs from many previous studies, which often focus on single or multiple institutions. Future
work should unpack the interaction effects of these measures.

Women’s lower likelihood of pursuing STEM majors is consistent with previous studies
and statistics typically reported by the National Science Foundation (e.g., [45]). We also found
that family background—whether parents have a STEM degree—and family income matters in
STEM major choice. This is consistent with literature on intergenerational transfer and the
importance of social and economic capital on students’ academic trajectories, especially parents’
educational levels (e.g., [7], [20], [35]). The findings suggest further research is needed,
particularly in disaggregating the math/science self-efficacy, identity, and utility measures, as
well as in investigating potential differences in major choice by field separately, rather than
STEM in the aggregate. Our research findings can be used to inform policies and programs
aimed at increasing diversity and inclusivity in STEM fields, as well as to identify areas where
additional support and resources may be needed to help students succeed.

Acknowledgments

The authors acknowledge that this work is supported, in whole or in part, by the U.S. Department
of Defense through SERC/AIRC on research task WRT 1068, contract no. HQ0034-19-D-0003.
This material is based upon work supported by the National Science Foundation under Grant
Number 2142697. Any opinions, findings, and conclusions or recommendations expressed in this
material are those of the authors and do not necessarily reflect the views of the National Science
Foundation.
References

[1] U.S. Bureau of Labor Statistics, “Employment in STEM occupations,” Sep. 08, 2022.
https://www.bls.gov/emp/tables/stem-employment.htm (accessed Feb. 23, 2023).
[2] The White House, “FACT SHEET: CHIPS and Science Act will lower costs, create jobs,
strengthen supply chains, and counter China,” Aug. 09, 2022.
https://www.whitehouse.gov/briefing-room/statements-releases/2022/08/09/fact-sheet-
chips-and-science-act-will-lower-costs-create-jobs-strengthen-supply-chains-and-counter-
china/ (accessed Feb. 23, 2023).
[3] National Science Board, “Science and engineering indicators 2022: The state of U.S.
Science and Engineering.,” National Science Foundation, Alexandria, VA, NSB-2022-1,
2022. Accessed: Feb. 23, 2023. [Online]. Available: https://ncses.nsf.gov/pubs/nsb20221
[4] V. Genareo, J. Mitchell, B. Geisinger, and M. Kemis, “University science partnerships:
What happens to STEM interest and confidence in middle school and beyond,” K-12 STEM
Education, vol. 2, no. 4, pp. 117–127, Oct. 2016.
[5] A. VanMeter-Adams, C. L. Frankenfeld, J. Bases, V. Espina, and L. A. Liotta, “Students
who demonstrate strong talent and interest in STEM are initially attracted to STEM through
extracurricular experiences,” LSE, vol. 13, no. 4, pp. 687–697, Dec. 2014, doi:
10.1187/cbe.13-11-0213.
[6] P. M. Sadler, G. Sonnert, Z. Hazari, and R. Tai, “Stability and volatility of STEM career
interest in high school: A gender study,” Science Education, vol. 96, no. 3, pp. 411–427,
2012, doi: 10.1002/sce.21007.
[7] H. Woo, N. Heo, H. Jang, and Y. Jang, “Parental and school factors on American high
school students’ academic and career intentions in STEM fields,” Int J Educ Vocat
Guidance, Sep. 2021, doi: 10.1007/s10775-021-09498-9.
[8] R. Lowinger and H. Song, “Factors Associated with Asian American Students’ Choice of
STEM Major,” Journal of Student Affairs Research and Practice, vol. 54, no. 4, pp. 415–
428, Oct. 2017, doi: 10.1080/19496591.2017.1345754.
[9] C. Sublett and M. A. Gottfried, “Individual and institutional factors of applied STEM
coursetaking in high school,” Teachers College Record, vol. 119, no. 10, pp. 1–38, 2017.
[10] X. Wang, “Why students choose STEM majors: Motivation, high school learning, and
postsecondary context of support,” American Educational Research Journal, vol. 50, no. 5,
pp. 1081–1121, Oct. 2013, doi: 10.3102/0002831213488622.
[11] M. Edwin, D. J. Prescod, and J. Bryan, “Profiles of high school students’ STEM career
aspirations,” The Career Development Quarterly, vol. 67, no. 3, pp. 255–263, 2019, doi:
10.1002/cdq.12194.
[12] E. R. Kurban and A. F. Cabrera, “Building readiness and intention towards STEM fields of
study: Using HSLS:09 and SEM to examine this complex process among high school
students,” The Journal of Higher Education, vol. 91, no. 4, pp. 620–650, Jun. 2020, doi:
10.1080/00221546.2019.1681348.
[13] D. M. Mangu, A. R. Lee, J. A. Middleton, and J. K. Nelson, “Motivational factors
predicting STEM and engineering career intentions for high school students,” in 2015 IEEE
Frontiers in Education Conference (FIE), Oct. 2015, pp. 1–8. doi:
10.1109/FIE.2015.7344065.
[14] W.-C. J. Mau and J. Li, “Factors influencing STEM career aspirations of underrepresented
high school students,” The Career Development Quarterly, vol. 66, no. 3, pp. 246–258,
2018, doi: 10.1002/cdq.12146.
[15] J. Zhang, G. Bohrnstedt, X. Zheng, Y. Bai, D. Yee, and M. Broer, “Choosing a college
STEM major: The roles of motivation, high school STEM coursetaking, NAEP
mathematics achievement, and social networks.,” American Institutes for Research,
Washington, DC, AIR-NAEP Working Paper #2021-02, Jun. 2021. Accessed: Aug. 26,
2022. [Online]. Available: https://www.air.org/resource/report/choosing-college-stem-
major-roles-motivation-high-school-stem-coursetaking-naep
[16] R. Darolia, C. Koedel, J. B. Main, J. F. Ndashimye, and J. Yan, “High school course access
and postsecondary STEM enrollment and attainment,” Educational Evaluation and Policy
Analysis, vol. 42, no. 1, pp. 22–45, Mar. 2020, doi: 10.3102/0162373719876923.
[17] J. B. Main, R. Darolia, C. Koedel, J. Yan, and J. F. Ndashimye, “The role of high school
math and science course access in student college engineering major choice and degree
attainment,” presented at the 2017 ASEE Annual Conference and Exposition, Columbus,
Ohio, Jun. 2017. Accessed: Sep. 18, 2022. [Online]. Available: https://peer.asee.org/board-
93-the-role-of-high-school-math-and-science-course-access-in-student-college-engineering-
major-choice-and-degree-attainment
[18] C. A. Ethington and L. M. Woffle, “Women’s selection of quantitative undergraduate fields
of study: Direct and indirect influences,” American Educational Research Journal, vol. 25,
no. 2, pp. 157–175, Jun. 1988, doi: 10.3102/00028312025002157.
[19] M. A. Gottfried, “The influence of applied STEM coursetaking on advanced mathematics
and science coursetaking,” The Journal of Educational Research, vol. 108, no. 5, pp. 382–
399, Sep. 2015, doi: 10.1080/00220671.2014.899959.
[20] S. A. Maple and F. K. Stage, “Influences on the choice of math/science major by gender
and ethnicity,” American Educational Research Journal, vol. 28, no. 1, pp. 37–60, Jan.
1991, doi: 10.3102/00028312028001037.
[21] J. Trusty, “Effects of high school course-taking and other variables on choice of science and
mathematics college majors,” Journal of Counseling & Development, vol. 80, no. 4, pp.
464–474, 2002, doi: 10.1002/j.1556-6678.2002.tb00213.x.
[22] W. Tyson, R. Lee, K. M. Borman, and M. A. Hanson, “Science, technology, engineering,
and mathematics (STEM) pathways: High school science and math coursework and
postsecondary degree attainment,” Journal of Education for Students Placed at Risk, vol.
12, no. 3, pp. 243–270, Oct. 2007, doi: 10.1080/10824660701601266.
[23] G. Crisp, A. Nora, and A. Taggart, “Student characteristics, pre-college, college, and
environmental factors as predictors of majoring in and earning a STEM degree: An analysis
of students attending a Hispanic serving institution,” American Educational Research
Journal, vol. 46, no. 4, pp. 924–942, Dec. 2009, doi: 10.3102/0002831209349460.
[24] E. C. Jewett and R. Chen, “Examining the relationship between AP STEM course‐taking
and college major selection: Gender and racial differences,” J of Engineering Edu, pp. 1–
19, Apr. 2022, doi: 10.1002/jee.20464.
[25] P. Vincent-Ruz and C. D. Schunn, “The nature of science identity and its role as the driver
of student choices,” International Journal of STEM Education, vol. 5, no. 1, p. 48, Nov.
2018, doi: 10.1186/s40594-018-0140-5.
[26] M. M. Chemers, E. L. Zurbriggen, M. Syed, B. K. Goza, and S. Bearman, “The role of
efficacy and identity in science career commitment among underrepresented minority
students,” Journal of Social Issues, vol. 67, no. 3, pp. 469–491, 2011, doi: 10.1111/j.1540-
4560.2011.01710.x.
[27] E. Seymour and N. M. Hewitt, Talking About Leaving: Why Undergraduates Leave the
Sciences. Boulder, Colorado: Westview Press, 2000.
[28] H. W. Marsh, “A multidimensional, hierarchical model of self-concept: Theoretical and
empirical justification,” Educ Psychol Rev, vol. 2, no. 2, pp. 77–172, Jun. 1990, doi:
10.1007/BF01322177.
[29] H. W. Marsh, “The multidimensional structure of academic self-concept: Invariance over
gender and age,” American Educational Research Journal, vol. 30, no. 4, pp. 841–860,
Dec. 1993, doi: 10.3102/00028312030004841.
[30] H. W. Marsh, B. M. Byrne, and R. J. Shavelson, “A multifaceted academic self-concept: Its
hierarchical structure and its relation to academic achievement,” Journal of Educational
Psychology, vol. 80, no. 3, pp. 366–380, 1988, doi: 10.1037/0022-0663.80.3.366.
[31] L. Gonzalez, S. Chapman, and J. Battle, “Mathematics identity and achievement among
Black students,” School Science and Mathematics, vol. 120, no. 8, pp. 456–466, 2020, doi:
10.1111/ssm.12436.
[32] L. Gonzalez, N. Lucas, and J. Battle, “A quantitative study of mathematics identity and
achievement among LatinX secondary school students,” Journal of Latinos and Education,
pp. 1–16, May 2022, doi: 10.1080/15348431.2022.2073231.
[33] G. Saw, C.-N. Chang, and H.-Y. Chan, “Cross-sectional and longitudinal disparities in
STEM career aspirations at the intersection of gender, race/ethnicity, and socioeconomic
status,” Educational Researcher, vol. 47, no. 8, pp. 525–531, Nov. 2018, doi:
10.3102/0013189X18787818.
[34] X. Chen, Students who study science, technology, engineering, and mathematics (STEM) in
postsecondary education. National Center for Education Statistics, 2009. Accessed: Apr.
28, 2022. [Online]. Available: https://eric.ed.gov/?id=ED506035
[35] M. W. Moakler Jr. and M. M. Kim, “College major choice in STEM: Revisiting confidence
and demographic factors,” The Career Development Quarterly, vol. 62, no. 2, pp. 128–142,
2014, doi: 10.1002/j.2161-0045.2014.00075.x.
[36] A. Sahin, A. Ekmekci, and H. C. Waxman, “The relationships among high school STEM
learning experiences, expectations, and mathematics and science efficacy and the likelihood
of majoring in STEM in college,” International Journal of Science Education, vol. 39, no.
11, pp. 1549–1572, Jul. 2017, doi: 10.1080/09500693.2017.1341067.
[37] L. Niu, “Family socioeconomic status and choice of STEM major in college: An analysis of
a national sample,” College Student Journal, vol. 51, no. 2, pp. 298–312, Jun. 2017.
[38] L. Vaval, A. J. Bowers, and V. Snodgrass Rangel, “Identifying a typology of high schools
based on their orientation toward STEM: A latent class analysis of HSLS:09,” Science
Education, vol. 103, no. 5, pp. 1151–1175, 2019, doi: 10.1002/sce.21534.
[39] M. C. Bottia, E. Stearns, R. A. Mickelson, and S. Moller, “Boosting the numbers of STEM
majors? The role of high schools with a STEM program,” Science Education, vol. 102, no.
1, pp. 85–107, 2018, doi: 10.1002/sce.21318.
[40] M. F. Rogers-Chapman, “Accessing STEM-focused education: Factors that contribute to
the opportunity to attend STEM high schools across the United States,” Education and
Urban Society, vol. 46, no. 6, pp. 716–737, Sep. 2014, doi: 10.1177/0013124512469815.
[41] G. K. Saw and C. A. Agger, “STEM pathways of rural and small-town students:
Opportunities to learn, aspirations, preparation, and college enrollment,” Educational
Researcher, vol. 50, no. 9, pp. 595–606, Dec. 2021, doi: 10.3102/0013189X211027528.
[42] J. B. Main, A. L. Griffith, X. Xu, and A. M. Dukes, “Choosing an engineering major: A
conceptual model of student pathways into engineering,” Journal of Engineering
Education, vol. 111, no. 1, pp. 40–64, 2022, doi: 10.1002/jee.20429.
[43] G. H. Elder and J. Z. Giele, Eds., The Craft of Life Course Research. New York ; London:
Guilford Press, 2009.
[44] J. T. Mortimer and M. J. Shanahan, Eds., Handbook of the Life Course. in Handbooks of
sociology and social research. New York: Kluwer Academic/Plenum Publishers, 2003.
[45] L. Tan, J. B. Main, and R. Darolia, “Using random forest analysis to identify student
demographic and high school‐level factors that predict college engineering major choice,” J
Eng Educ, vol. 110, no. 3, pp. 572–593, Jul. 2021, doi: 10.1002/jee.20393.