Gardner-Webb University

Digital Commons @ Gardner-Webb University

Doctor of Education Dissertations College of Education

Spring 2022

An Exploratory Quantitative Study of the Impact of STEM-Focused

Middle Schools on Student Persistence and Performance in STEM
Shandua Ellis
Gardner-Webb University, sellis6@gardner-webb.edu

Follow this and additional works at: https://digitalcommons.gardner-webb.edu/education-dissertations

Part of the Curriculum and Instruction Commons, Educational Assessment, Evaluation, and Research
Commons, and the Educational Methods Commons

Recommended Citation
Ellis, Shandua, "An Exploratory Quantitative Study of the Impact of STEM-Focused Middle Schools on
Student Persistence and Performance in STEM" (2022). Doctor of Education Dissertations. 109.
https://digitalcommons.gardner-webb.edu/education-dissertations/109

This Dissertation is brought to you for free and open access by the College of Education at Digital Commons @
Gardner-Webb University. It has been accepted for inclusion in Doctor of Education Dissertations by an authorized
administrator of Digital Commons @ Gardner-Webb University. For more information, please see Copyright and
Publishing Info.
AN EXPLORATORY QUANTITATIVE STUDY OF THE IMPACT OF STEM-
FOCUSED MIDDLE SCHOOLS ON STUDENT PERSISTENCE AND
PERFORMANCE IN STEM

By
Shandua Brown Ellis

A Dissertation Submitted to the

Gardner-Webb University College of Education
in Partial Fulfilment of the Requirements
for the Degree of Doctor of Education

Gardner Webb University

2022
 Approval Page

This dissertation was submitted by Shandua Brown Ellis under the direction of the
persons listed below. It was submitted to the Gardner-Webb University College of
Education and approved in partial fulfillment of the requirements for the degree of
Doctor of Education at Gardner-Webb University.

__________________________________ ________________________
Benjamin Williams, EdD Date
Committee Chair

_________________________________ ________________________
Joey Lord, EdD Date
Committee Member

_________________________________ ________________________
Jenni Corn, PhD Date
Committee Member

_______________________________ ________________________
Rodney Trice, EdD Date
Committee Member

_________________________________ ________________________
Prince Bull, PhD Date
Dean of the College of Education

ii
 Acknowledgements

In Deuteronomy 31:8, it says that the Lord himself goes before you and

will be with you; he will never or forsake you. Do not be afraid, do not be

discouraged. I would like to first and foremost give praise and thanks to the

Lord, the Almighty, for his blessings and for giving me the courage, strength,

and perseverance to endure this challenging yet rewarding experience of

completing my dissertation research successfully.

I would like to express my deep and sincere gratitude to my dissertation

coach and chair, Dr. Benjamin Williams, for always sharing words of inspiration

to help keep me motivated, for ensuring I stayed abreast of all deadlines and

requirements, and for always responding to my many questions throughout the

research process with patience and kindness. I would also like to thank my

dissertation committee members, Dr. Joey Lord, Dr. Jenni Corn, and Dr. Rodney

Trice, for their time, invaluable guidance, support, and intellectual contributions

to my research. In addition, I’d like to thank Dr. Colleen Paeplow for providing

the data needed for my research and for sharing her knowledge and skills to

assist me in getting through the most challenging component of my research—

using descriptive and multiple regression statistics to describe and interpret my

data. Your time, patience, and kindness will never be forgotten.

I am extremely grateful to my wonderful husband, Reggie, and my

beautiful daughters, Khaylan and Tearra, for their love, encouragement,

understanding, prayers, and continuing support throughout my entire research

journey. Reggie, you are always there for me, every step of the way, and I love

iii
and appreciate you more than words can ever express. Thank you for your

undying patience and understanding. I am greatly appreciative as well for my

parents, Barbara and George Brown, for always loving me, praying for me, and

supporting each and every one of my endeavors in life. Also, I express my thanks

to all other family members for their support and valuable prayers. My special

thanks go to my friend, Dr. Renita Griffin-Jordan, for sharing this whole doctoral

experience with me. You were always there to give an encouraging word, and

your success along this journey too helped keep me motivated as well! Thank

you for always being a good friend and role model.

Moreover, I’d like to thank Mr. Paul Domenico. Thank you so much Mr.

Domenico for all your support on all the big projects for many of my classes and all

your advice and assistance throughout my research work as well! Thank you as well

to Ms. Jill Herbst and Mrs. Susan Martin for participating in all my action research

projects! I learned so much from our experiences and will be forever grateful for

how you have contributed to my success.

Finally, my thanks go to all the people who have supported me in some

way, directly or indirectly, throughout my entire doctoral journey.

iv
 Abstract

AN EXPLORATORY QUANTITATIVE STUDY OF THE IMPACT OF STEM-

FOCUSED MIDDLE SCHOOLS ON STUDENT PERSISTENCE AND

PERFORMANCE IN STEM. Ellis, Shandua Brown, 2022: Dissertation, Gardner-Webb

University.

Science, technology, engineering, and math (STEM) continue to be a major challenge for

the United States (U.S.), as the U.S. continues to lag behind other countries in this area

due to ongoing lack of individuals who are entering STEM fields and many who are

entering STEM fields but lack the skills necessary to perform adequately in these roles.

The problem related to this study involved the ongoing need to identify how well STEM

education programs are addressing this need and increasing the number of students,

especially underrepresented minorities, in following STEM career pathways and

developing the knowledge and skills needed to persist in the field. The purpose of this

quantitative study was to explore the persistence and performance of students who

attended a STEM-focused middle school in North Carolina. The Social Cognitive Career

Theory is the theoretical framework for this study and provides a foundation for how

career decisions developed over time. This study was a quantitative, nonexperimental

investigation of student data on high school student STEM persistence and academic

performance in STEM-related courses throughout high school, after attending a STEM-

focused middle school for their sixth-, seventh-, and eighth-grade years. The study results

found that male students have twice the amount of STEM persistence on average than

female students and that African-American students had the least amount of STEM

persistence as all other subgroups of students, while the White subgroup had the greatest

v
STEM persistence than all other subgroups. In addition, study results also found that

female students STEM academic performance was comparable to that of male students,

with females having slightly better performance in mathematics courses. Also, the White

subgroup outperformed all other subgroups, while the African-American and Hispanic

subgroups’ academic performance was the lowest. The standard multiple regression

resulted in very low significance, however, between the gender and race/ethnicity of

students and their STEM persistence and STEM academic performance. Implications of

the study for school districts include ensuring STEM-focused middle schools have

effective practices that significantly impact students’ interest, especially underrepresented

subgroups of students, in STEM, explicitly identifying and acknowledging and

identifying solutions for barriers that impact STEM persistence and academic

performance, and ensuring the instructional practices employed by STEM teachers are

equitable through the use of culturally and gender-responsive pedagogy. Overall, the

study has the potential to assist districts with STEM-focused middle schools to improve

data trends on students’ academic performance and persistence in STEM.

Keywords: STEM, STEM education, STEM skills, career aspirations, STEM

dispositions, STEM self-efficacy, STEM persistence, 21st century skills, underrepresented

minorities

vi
 Table of Contents
Page
Chapter 1: Introduction ........................................................................................................1
Background ..............................................................................................................2
Problem Statement ...................................................................................................6
Theoretical Framework ............................................................................................8
Purpose.....................................................................................................................9
Research Questions and Hypotheses .......................................................................9
Methodology ..........................................................................................................10
Rationale ................................................................................................................11
Significance of the Study .......................................................................................12
Assumptions...........................................................................................................12
Scope ......................................................................................................................13
Limitations .............................................................................................................13
Delimitations ..........................................................................................................13
Definition of Terms................................................................................................14
Summary ................................................................................................................17
Organization of Study ............................................................................................17
Chapter 2: Literature Review .............................................................................................19
Introduction ............................................................................................................19
Literature Search Strategy......................................................................................22
Theoretical Foundation .........................................................................................23
Defining STEM from a Historical Context ............................................................32
STEM Education in the U.S...................................................................................34
Importance of STEM Education During Middle School Years .............................45
STEM-Focused Schools.........................................................................................47
Controversies with STEM and STEM Education ..................................................48
Student Dispositions in STEM...............................................................................48
Factors Impacting STEM Self-Efficacy in Students ..............................................53
Student Persistence in STEM.................................................................................55
Summary ................................................................................................................56
Chapter 3: Methodology ....................................................................................................58
Purpose...................................................................................................................58
Research Questions and Hypotheses .....................................................................59
Setting ....................................................................................................................59
Research Design and Rationale ............................................................................ 60
Role of the Researcher ...........................................................................................62
Methodology ..........................................................................................................63
Population ..............................................................................................................63
Sampling Strategy Identification and Justification ................................................63
Participant Selection Criteria .................................................................................64
Number of Participants and Rationale ...................................................................64
Participant Identification, Contact, and Recruitment .............................................65
Instrumentation ......................................................................................................66
Data Collection Analysis and Rationale ................................................................67
Ethical Considerations ...........................................................................................67

vii
 Summary ................................................................................................................68
Chapter 4: Results ..............................................................................................................69
Introduction ............................................................................................................69
Data Collection ......................................................................................................69
Variables ................................................................................................................74
Description of Study Participants ..........................................................................75
STEM Persistence Data Results.............................................................................76
STEM Academic Performance Data Results .........................................................91
Results and Analysis ............................................................................................135
Summary ..............................................................................................................139
Chapter 5: Discussion ......................................................................................................141
Overview ..............................................................................................................141
Interpretation of Findings ....................................................................................143
Connection to Theoretical Framework ................................................................147
Implications..........................................................................................................148
Limitations of the Study.......................................................................................151
Recommendations for Action and Further Studies ..............................................153
Conclusions ..........................................................................................................155
References ........................................................................................................................158
Appendices .......................................................................................................................173
A High School Planning Guide Student Course Requirements ...............................173
B Middle/High School Student Attitudes Towards STEM Survey .........................176
Tables
1 High School STEM-Oriented Courses...................................................................71
2 Frequencies and Percentages for Gender and Race/Ethnicity ...............................76
3 High School Course Requirements (College and Career Ready) ..........................77
4 Number of High School STEM Course Options by Category...............................78
5 Descriptive Stats for STEM Course Count Data ...................................................79
6 STEM Course Enrollment Means Comparison by Gender....................................81
7 STEM Count Means Comparison by Race/Ethnicity ............................................84
8 STEM Count Assumption Check for Multicollinearity: Correlation
Coefficients ............................................................................................................88
9 STEM Count Assumption Check for Multicollinearity: Tolerance/VIF ...............88
10 STEM Count Assumption Check for Residuals for STEM Count ........................89
11 Cook’s Distance and Leverage for STEM Count ..................................................89
12 STEM Count Multiple Regression Results ............................................................90
13 STEM Course Performance Means by Subject Area .............................................92
14 NC Required High School EOC Test Scales and Performance Levels .................93
15 NC Math I EOC Performance Level Counts .........................................................94
16 NC Math III EOC Performance Level Counts .......................................................96
17 Biology EOC Performance Frequency, Mean and Median ...................................98
18 STEM Course Mean Comparison by Gender ......................................................100
19 STEM Mean Course Performance by Race/Ethnicity .........................................113
20 STEM Academic Performance Assumption Check for Correlation ....................126
21 STEM Academic Performance Assumption Check for Tolerance/VIF ..............127
22 STEM Mean Overall Residuals Check for Possible Outliers ..............................128

viii
23 STEM Mean Cook’s Distance and Leverage Value ............................................129
24 Multiple Regression Results for STEM Academic Performance ........................134
Figures
1 Career-Related Interests and Choice Development Over Time .............................26
2 SCCT Performance Model .....................................................................................28
3 Profiles of the 21st Century Learner and Learning Environment ...........................40
4 A Priori Power Analysis Results ...........................................................................65
5 Frequency of Number of STEM Courses Taken ...................................................80
6 STEM Course Count by Gender Box Plot .............................................................82
7 Number of STEM Courses Taken by Gender ........................................................83
8 STEM Course Enrollment by Race/Ethnicity Box Plot.........................................85
9 STEM Course Count by Race/Ethnicity ................................................................86
10 STEM Count P-P Plot Assumption for Normality ................................................90
11 NC Math I EOC Performance Levels ....................................................................95
12 NC Math III EOC Performance Levels .................................................................97
13 Biology EOC Performance Levels.........................................................................99
14 Math Mean Performance Comparison by Gender ...............................................102
15 NC Math I EOC Performance by Gender ............................................................103
16 NC Math III EOC Performance by Gender .........................................................104
17 Science Mean Performance by Gender ................................................................106
18 Biology EOC Performance by Gender ................................................................107
19 STEM Electives Mean Performance by Gender ..................................................109
20 Overall STEM Mean Course Performance by Gender ........................................110
21 Math Mean Performance by Race/Ethnicity ........................................................114
22 NC Math I EOC Performance by Race/Ethnicity ................................................116
23 NC Math III EOC Performance by Race/Ethnicity .............................................118
24 Science Mean Performance by Race/Ethnicity Box Plot .....................................119
25 Biology EOC Performance by Race/Ethnicity ....................................................121
26 STEM Electives Mean by Race/Ethnicity ...........................................................122
27 STEM Course Overall Mean Performance by Race/Ethnicity ............................123
28 Overall STEM Mean P-Plot Assumption for Normality .....................................129
29 NC Math I EOC P-Plot Assumption for Normality .............................................129
30 NC Math III EOC P-Plot Assumption for Normality ..........................................130
31 Biology EOC P-Plot Assumption for Normality .................................................130

ix
 1

Chapter 1: Introduction

Background

For many years now, school districts worldwide have celebrated the advancement

and integration of science, technology, engineering, and math (STEM) education in their

schools and how they are educating students to prepare for future STEM challenges they

will face, both in the workplace and in society. Significant STEM challenges our students

will face in their future include securing cyberspace, sustaining lands and oceans,

ensuring economic clean energy, accessing clean water, developing and delivering better

medicines, sustaining cities and communities, quantifying uncertainty with climate

change, and much more (Department of Economic & Social Affairs, 2018). While I

believe every accomplishment, no matter how great or small, should be celebrated in

regard to STEM advancement, the United States Department of Education (2016b) shares

the importance of continuing to assist our youth in preparing for challenges they will face

in this increasingly complex world.

Although the U.S. has been a world leader in the economy, it continues to fall

behind in preparing students with the knowledge and skills needed to solve the current

and future problems through some of the most important areas of education.

Over the last decade, the U.S. has seen nearly 2 million new STEM jobs, but

students’ math and science scores continue to lag behind other nations. China has

been the leader in producing STEM graduates in the world—with about a 1 to 29

ratio between the U.S. and China, with the U.S. lagging far behind. (Jones, 2020,

p. 1)

Taking this into consideration, it is necessary to conduct studies to determine if STEM

 2

programs are preparing students to be the more highly skilled students and employees

sought by colleges, businesses, and organizations. To recruit these types of students, it is

important to have students who are first interested in STEM and have attained STEM

career pathways and have the confidence, along with the skills needed, to be successful in

STEM-related courses, which leads to the purpose of my research study.

The purpose of this study was to explore student persistence and academic

performance in STEM after attending a STEM-focused middle school during their sixth-,

seventh-, and eighth-grade years. The study also sought to discover if any gaps in

performance and persistence exist between traditionally represented and underrepresented

subgroups of students. This study has implications for students, teachers, and

administrators, especially those in secondary STEM-focused schools and district leaders

who support curriculum enhancement programs within the district. Identifying how

STEM-focused schools are impacting students can help school districts determine

whether students who attend STEM-focused schools are persistent in STEM and have

high levels of academic performance in STEM courses in which they enroll in high

school. It can also provide data that school districts need to determine how to further

improve and/or enhance their STEM programs. Doing so can help encourage more

students to pursue STEM career pathways and perform well in STEM subjects. Such

improvement can play a role in the larger goal of assisting the U.S. in developing more

highly skilled STEM professionals.

Background

In 2011, President Barack Obama, in his State of the Union Address, made the

statement, “In a single generation, revolutions in technology have transformed the way
 3

we live, work and do business” (para. 14). He went on to describe how China recognized

this change and had strengthened math and science education and how their investments

have led to their leadership in developing the largest solar facility and fastest computer.

He further stated, “the future is not a gift, but it’s an achievement” (State of the Union

Address, 2011, para. 18); and “we know what it takes to compete for the jobs and

industries of our time. We need to out-innovate, out-educate, and out-build the rest of the

world” (State of the Union Address, 2011, para. 19). He tells the nation that “this is our

Sputnik” (State of the Union Address, 2011, para. 23) moment and challenged us to get

behind the science and engineering innovations of our time and that the way to do so is

through improving how we are educating our children in the U.S.

Since Obama’s address to the nation, there have been some scholarly research

studies published that have focused on some of the nation’s efforts in improving the

impact STEM programs are having on student interest and performance in specific areas

of STEM. There are research studies that were based on Lent et al.’s (2002) social

cognitive career theory (SCCT). Blotnicky et al. (2018) conducted a study to explore

student math self-efficacy, future career interests, preferences for particular career

activities, and their likelihood to pursue STEM. Mueller et al. (2015) conducted a study

that tests the validity of utilizing SCCT to examine the career goals and choices of middle

school students who were already expressing an interest in math- and science-related

subjects and careers. Fouad and Santana (2016) also conducted a research study that

focused on factors with early choices and consisted of studying existing literature related

to SCCT and underrepresented minorities and identified barriers impacting their career

choices. These students are a very small part of the greater issue our nation has in
 4

continuing the need to improve our STEM workforce through STEM education. The

literature review shows SCCT, along with research studies that have utilized it as the

foundation for their research, provides an appropriate foundation. My study is needed as

it adds to the research literature in that it examines dispositions, self-efficacy, persistence,

and performance in students who attend STEM-focused middle schools, utilizing SCCT

as a foundation as well and provides an investigation into how those schools are

influencing students through their STEM programs despite other factors that can

negatively impact students to develop their STEM knowledge, skills, and dispositions.

In 2014, the U.S. Bureau of Labor Statistics announced that the number of people

employed in careers related to STEM was expected to increase to more than 9

million between 2012 and 2022. Also, more recently, the United States’ PISA

rankings placed the U.S. 38th out of 71 countries in math and 24th in science.

(Gunn, 2020, para. 4)

The research literature in this study provides a further examination into the

current ongoing need to ensure we are preparing our students to meet the demands of our

nation in the future workforce in which they will have an impact. Despite U.S. efforts to

advance STEM, Jones (2020) reported that “more than half of U.S. patents still go to

foreign nationals and the U.S. continues to be the net importer of high-tech products” (p.

5). Also, U.S. employers continue to express the growing need to recruit and attain

employees with technical skills, and colleges and universities, as well as many businesses

and organizations, continue to report the need for more highly skilled students and

employees with STEM skills (Jones, 2020).

Data from various research studies in the literature have led to conclusions on the
 5

dispositions and self-efficacy of students that can impact the persistence and performance

of students after they participated in a variety of STEM-related education activities.

STEM education has proven, through research studies over the past 5 years, that STEM

continues to have a significant impact on the academic achievement of students in

STEM-related courses. Yaki et al. (2019) and Acara et al. (2018) have implemented and

reported data and findings from studies that show that STEM approaches to science

improve science achievement. Yang and Baldwin (2020) provided data supporting

student achievement in technology and the positive impacts that technology has had on

other subject areas as well.

In addition, the research literature emphasizes the importance of assisting students

in developing interests in STEM and STEM careers during their middle school years.

(Almeda & Baker, 2020). Student attitudes and their experiences during middle school

impact their dispositions and self-efficacy and can impact their belief in their probability

of success in STEM courses and the likelihood of pursuing STEM fields of study and

careers (Van Tuijl & van der Molen, 2016). The research literature also focuses on the

impact of STEM-focused middle schools on students and provides data supporting the

importance of developing student dispositions and self-efficacy towards certain fields

during the middle school years as it impacts student interests and career pathways

(Blotnicky et al., 2018).

Furthermore, the literature review includes factors that impact the dispositions

and self-efficacy of students, especially underrepresented minorities in STEM. Racial and

gender gaps continue to exist in regard to the lack of equity in traditional classroom

environments which impact dispositions and self-efficacy (Christensen & Knezek, 2017;
 6

National Research Council, 2011). Some studies have even shown low self-confidence in

Hispanics in STEM (Student Research Foundation, 2019). Other studies have been done

to address the disproportionality of underrepresented minorities in STEM (Christensen &

Knezek, 2017; U.S. Department of Education, 2016a; Verriden, 2017). Factors such as

expectancy; stereotypes; and family, school, and community values can greatly impact

student dispositions and self-efficacy as well (Bronfenbrenner, 2005; Lambert, 2018;

Lent et al., 2002).

Moreover, the literature review focuses on the persistence of students in STEM

and factors that can influence that as well. Exposure to STEM, such as what students who

attend STEM-focused schools would receive, can impact student persistence in STEM

due to the ongoing opportunities to engage in STEM activities that can impact their

knowledge and success in STEM over time. The more knowledgeable a student becomes

in STEM, the more likely they are to develop an interest in it and engage in more STEM

activities that can lead to an interest in pursuing a STEM-related career (Blotnicky et al.,

2018).

Problem Statement

The problem related to the study is the lack of students graduating from high

school who are prepared with the STEM knowledge and skills to handle the growing

need for STEM occupations. With the ongoing increase in global challenges that students

will inevitably encounter, especially with significant advances continuing to surface in

STEM areas, there is much need to ensure our students are developing the knowledge,

desire, and skills to be successful in addressing issues relating to sustainability, climate

change, accessing clean water, the balance of population growth and resources, global
 7

foresight during unprecedented change (much like what our students are facing with our

current global pandemic), global convergence of information and communications

technology, reducing new and emerging diseases, growing change of women status,

science and technological breakthroughs, and the list continues (Millennium Project

Group, 2017).

Current STEM education research continues to express the need for the U.S. to

better prepare students for college and careers in STEM. “If America’s STEM

proficiency continues to decline, not only will the skills gap be detrimental to the

workforce, but it will also erode its potential future for economic and scientific

leadership” (Jones, 2020, p. 8). This continues to be a growing concern as STEM

continues to grow rapidly in our society and proposes a wide variety of challenges for our

students and their future. In preparing students for a future in STEM, many school

systems have developed and/or adopted STEM education programs in the elementary,

middle, and high school levels, which focus on developing and enhancing student STEM

knowledge and skills. Experts also suggest that the introduction of STEM at an earlier

age and educating students on the diversity of STEM careers are crucial elements in

preparing a more capable workforce (Jones, 2020).

School districts must be aware of the extent of the contribution they are making to

address the issues of the need for students to be more interested in STEM and be better

prepared for the STEM workforce. There is very little available research that evaluates

the effectiveness of STEM-focused middle schools on student dispositions towards

STEM and student career pathways and performance in STEM. This leaves school

districts unaware of the level of success in which STEM education in the district is
 8

helping to develop student persistence and performance in STEM.

Theoretical Framework

SCCT is the theoretical framework for this study. Developed by Lent et al.

(2002), the study originated from Bandura’s (1989) social cognitive theory. Overall,

SCCT supports the idea that people’s career paths are influenced by beliefs about their

self-efficacy and career path options based upon their experiences, culture, and

environment (Borgen, 1991). Self-efficacy beliefs, outcome expectations, and goals serve

as the foundation upon which SCCT was developed. “SCCT seeks to explain three

interrelated aspects of career development which include how basic academic and career

interests develop, how educational and career choices are made, and how academic and

career success is obtained” (Social Cognitive Career Theory, 2021, para. 1).

The theory explains the three interrelated aspects of careers using an interest

model, choice model, and performance model (Lent et al., 2002). SCCT uses the interest

model to describe how interests can be developed through people’s feelings of self-

efficacy based on childhood exposures and experiences and whether they anticipate

positive outcomes from their performance in certain skills (Lent et al., 2002; Social

Cognitive Career Theory, 2021). The theory goes on to explain how educational and

career choices are made through the choice model, where it describes how interest can

foster career choices. It describes how a person’s environment can influence what career

choices are made based on what’s seen as acceptable or unacceptable. The theory then

proposes how academic and career success is obtained through the interrelationship of

interest and choice utilizing the performance model to describe a person’s ability as

compared to the performance goals they set for themselves. All these models are
 9

impacted by self-efficacy and outcome expectations (Lent et al., 2002).

In the literature review, I provide a more detailed explanation of each of the

components of the theory. Also, STEM education-based research studies are described in

terms of how factors within the theory such as self-efficacy and performance expectations

impact student persistence and performance.

Purpose

As aforementioned, research studies express the significance of better preparing

our students to successfully develop STEM knowledge and skills, thus playing a role in

addressing the nation’s problem. The purpose of this quantitative study aims to explore

the extent that STEM-focused middle schools are preparing middle school students who

have attended a STEM-focused middle school throughout their sixth-, seventh-, and

eighth-grade years, to develop STEM persistence and high academic achievement in

STEM education as they enter and transition through their high school years.

In accomplishing this purpose, this study reviewed the dependent variables of

STEM persistence and STEM academic performance in STEM-related courses from their

freshman to their junior year in high school as based on data provided through

deidentified LEA student information rosters. Understanding how students, regardless of

their backgrounds, who attend STEM-focused middle school persist and perform in

STEM could provide information on how well the STEM schools are having an impact

on students and what improvements may be needed, if any, based on the results.

Research Questions and Hypotheses

I developed the following research questions, along with my hypothesis for each

question, to address the problem and align with the purpose of this research study:
 10

1. What are high school seniors’ STEM persistence in high school after a 3-year

enrollment in a STEM middle school?

H1: High school seniors, after a 3-year enrollment in a STEM-focused

middle school, will enroll in advanced STEM-related core courses and

STEM-oriented electives throughout high school.

2. What are high school seniors’ academic performance in STEM courses in

high school after a 3-year enrollment in a STEM middle school?

H2: High school seniors, after a 3-year enrollment in a STEM-focused

middle school, will enroll in advanced STEM-related core courses and

STEM-oriented electives and experience high academic achievement in

those courses throughout high school.

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

H3: Due to the ongoing nature of educational inequality, minority subgroups

will have less positive results in persistence and performance than other

subgroups.

Methodology

For this study, a quantitative paradigm was used. A nonexperimental quantitative

design was used to analyze data to determine the extent of high school student STEM

persistence and academic performance. In this study, I explored the dependent variables

of student STEM persistence and STEM academic performance in high school after

attending their STEM-focused middle school for their sixth-, seventh-, and eighth-grade

years. Independent variables in the student included gender and race/ethnicity. These
 11

variables were explored to compare similarities and differences in STEM persistence and

STEM academic performance.

In addition, descriptive statistics including means of central tendency such as

median, mode, and percentages, along with multiple regressions models were utilized to

report and analyze the study’s data. A nonexperimental quantitative design was utilized

to identify student persistence in STEM, based on advanced STEM core and STEM

elective courses in which students have enrolled in high school and their achievement in

those courses which was identified through deidentified data provided on student grades

and end-of-course (EOC) performance.

Rationale

Despite an abundance of research supporting the middle school years as the

optimal time for assisting students in developing dispositions and self-efficacy in STEM

so students may persist and perform well in STEM, there is a gap in the literature that

focuses on STEM-focused middle schools in which students attend through their entire

middle school years and the impact or influence these schools are having on student

persistence and academic performance in STEM beyond middle school. There are school

districts, such as the one included in this research study, that provide opportunities for

students to develop their STEM skills and knowledge, not just through STEM programs,

but through STEM-focused schools in which students can develop and enhance their

STEM knowledge and skills throughout their education. This study contributes to and

extends the research on how STEM education in our nation is helping to address the need

to better educate and motivate students to develop STEM career pathways.

 12

Significance of the Study

This research study has potential contributions that can advance knowledge,

practices, and policies in STEM education in the middle school setting. It can assist in

determining how STEM programs at STEM-focused middle schools are impacting

underrepresented minorities in STEM, despite the impact of outside factors that can

influence their persistence and academic performance in STEM.

This research study also has potential implications for positive social change that

is consistent with and bounded by the scope of study in that it can help determine if

middle schools in the district are having a positive impact on student STEM persistence

and academic performance in STEM. It can also assist in identifying if we are

contributing to society by preparing more students to be interested in STEM college and

career pathways which could ultimately contribute to the ongoing shortage of highly

qualified STEM candidates in STEM fields of study and STEM careers.

Assumptions

I have made several assumptions on which I have based components of this

research study. The first assumption made is that students who attended 3 years of a

STEM-focused middle school in the same district have similar experiences. Another

assumption is that all the high schools the students in the study attend offer equal

opportunities for students to participate in advanced STEM-related courses and that the

teachers of the courses used best instructional practices to teach those courses allowing

for equitable opportunities for success. Without this, it could impact the overall STEM

persistence and academic performance of students.

 13

Scope

The scope of this study includes the boundary of the research where the purpose

was to specifically examine the persistence and academic performance in STEM courses

for students based on a roster of deidentified data. Another boundary of the study is that it

included only students in a school district in North Carolina who were enrolled in STEM-

focused middle schools in the district for their sixth-, seventh-, and eighth-grade years

and were enrolled as high school seniors in the district as well.

Limitations

One limitation of the study is that student data that were used for the study were

for those students who had been enrolled in a STEM-focused middle school. The study

does not include any preexisting data that identified if there was already a desire to

persist in STEM and attain high academic achievement in STEM prior to attending a

STEM-focused middle school. Another limitation for the study was the possibility that

seniors were in high school long enough to possibly have been more influenced by their

high school experiences that have impacted their persistence and performance in STEM

in addition to their middle school experience. Collecting enrollment and performance

data from the time the students first entered high school as freshmen assisted with this

limitation.

Delimitations

The delimitations of this study involved the selection of student subjects who

attended a STEM-focused middle school for 3 years. I only included students who were

currently high school seniors in the study. Based on SCCT (Lent et al., 2002), these

students may have already had a variety of influences and experiences outside the walls
 14

of the 3-year STEM-focused middle schools they attended that could have had some type

of impact on their STEM persistence and academic performance. Also, although it would

have been very beneficial to examine individual perceptions through interviews, focus

groups, and an in-depth review of each of the components of the STEM programs at each

of the STEM-focused middle schools during the time frames in which the students in the

study attended those schools, limited time and limited resources prevented this type of

data from being collected for this research study in particular, therefore leaving the

results of this study to be based only on quantitative data from the review of student

STEM courses taken and grades and assessment performances of students in STEM

courses.

Definitions of Terms

21st Century Skills

Abilities that today’s students need to succeed in their careers during the

Information Age. The 21st century skills are critical thinking, creativity,

collaboration, communication, problem-solving, information literacy, media

literacy, technology literacy, flexibility, leadership, initiative, productivity, and

social skills. These skills are intended to help students keep up with the lightning

pace of today’s modern markets. Each skill is unique in how it helps students, but

they all have one quality in common. They’re essential in the age of the Internet.

(Stauffer, 2020, para. 1)

Career Aspirations

Career aspirations are the desire and intention to pursue an occupation or a

particular position within an occupation. Aspirations play an important role in

 15

career decisions because they reflect the goals and intentions that influence

individuals toward a particular course of action. (Aspirations in Career Decisions,

2021, para. 1)

College and Career Readiness

“Student who is ready for college and career can qualify for and succeed in entry-

level, credit-bearing college courses leading to a baccalaureate or certificate, or career

pathway-oriented training programs without the need for remedial or developmental

coursework” (Conley, 2012, p. 1).

STEM

An interdisciplinary approach to learning where rigorous academic concepts are

coupled with real-world lessons as students apply science, technology,

engineering, and mathematics in contexts that make connections between school,

community, work, and the global enterprise enabling the development of STEM

literacy and with it the ability to compete in the new economy. (Hallinen, 2020,

para. 7)

STEM Education

According to the National Science Teachers Association (2020),

A common definition of STEM education…is an interdisciplinary approach to

learning where rigorous academic concepts are coupled with real-world lessons as

students apply science, technology, engineering, and mathematics in contexts that

make connections between school, community, work, and the global enterprise

enabling the development of STEM literacy and with it the ability to compete in

the new economy. (para. 1)

 16

STEM Skills

They are a combination of the ability to produce scientific knowledge, supported

by mathematical skills, to design and build (engineer) technological and scientific

products or services. Although STEM skills overlap with basic and higher-order

cognitive skills, they merit separate treatment in a policy-oriented context to

target specific requirements in the education and labor market. (Siekmann &

Korbel, 2016, p. 45)

Underrepresented Minorities

(URM), which refers to the low participation rates of racial and ethnic groups in

fields such as computing relative to their representation in the U.S. population.

African Americans/Blacks, Hispanics/Latino(a), and Native Americans/Alaskan

Natives are most commonly defined as URMs, which aligns with the National

Science Foundation's definition. (Williams, 2020, para. 4)

Dispositions

A leaning toward a way of thinking or a state of mind regarding something

(Merriam-Webster, n.d.a).

STEM Self-Efficacy

A person’s perception or belief that they have the ability to complete STEM-

related tasks or can be successful in STEM (Falco & Summers, 2017).

Persistence

The act of keeping at something despite difficulties, opposition, or

discouragement (Merriam-Webster, n.d.b).

 17

Academic Performance

For the purpose of this study, academic performance pertained to students’ overall

grades in courses in which they have enrolled and their scores on EOC assessments.

Summary

The future of U.S. students will be full of STEM challenges as time goes on, and

there continues to be a need to prepare students to address those challenges successfully.

The increase of STEM challenges in our society has led to ongoing growth in STEM jobs

in the U.S. that need graduates who are competent with the knowledge and skills needed

to do those jobs well. Unfortunately, there are not ample graduates in the U.S. who are

interested or prepared for these roles. In addition, the U.S. continues to be outperformed

by other countries in the STEM subject areas of math and science as well as in STEM

advancements. SCCT provides a foundation for this study of factors that impact student

dispositions, self-efficacy, perseverance, and performance in STEM. Many STEM

research studies support this theory and the importance of ensuring a strong foundation of

STEM exposure, experience, and education, most especially in the middle school years

can assist in improving the motivation and interest in STEM. This can be beneficial in

assisting school districts in contributing to the growing need for students to develop

STEM career pathways and developing the skills needed to be successful in STEM fields

and careers.

Organization of Study

This dissertation includes five chapters. Chapter 1 provided an overview of the

study including background on STEM challenges that the U.S. is facing with STEM and

the need to better motivate and prepare our students to develop the knowledge and skills
 18

needed for success in STEM fields and careers. Chapter 2 provides a review of the

literature focusing on SCCT and its impact on dispositions, self-efficacy, perseverance,

and performance and ensuring that these factors as it relates to STEM are addressed in

the middle school years. Chapter 3 discusses the quantitative methodology used in this

research and the data collection methods. Chapter 4 explains how the quantitative data

are coded and aggregated and what statistical tests are used to analyze and explain the

data. Chapter 5 concludes with a discussion of the findings, implications, and suggestions

for future research.

 19

Chapter 2: Literature Review

Introduction

The purpose of this quantitative research study was to explore the persistence and

academic performance of high school students who attended a STEM-focused middle

school throughout the entire 3 years. Specifically, this study focused on the interactions

of variables from SCCT (Lent et al., 2002) and student participation in 3 years of a

STEM program and the impact it has had on their STEM persistence and STEM

academic performance in high school.

The problem is there continues to be a need to ensure our students are developing

the knowledge, skills, and motivation to attain STEM career pathways in high schools to

encourage more students to be interested in STEM careers. Research studies continue to

show that the U.S. struggles with this compared to other countries. The United States

Department of Education (2019) provided highlights from the 2019 Trends in

International Mathematics Study (TIMMS), which provided data on math and science

achievement of U.S. students in fourth and eighth grades compared to those of other

countries, showed that the U.S. had significantly large score gaps between the top- and

bottom-performing students, the highest of almost every other country except Turkey.

The score gaps have increased since the 2015 administration of the test. There were also

drops in performance in mathematics and science since the past two prior

administrations. In addition, average scores overall have not changed significantly since

the 2015 administration in both math and science, with a decrease for fourth graders

(U.S. Department of Education, 2019).

In addition to the TIMMS studies that take place every few years, the Programme
 20

for International Student Assessment (PISA), where over 80 countries participate and 15-

year-olds take a test developed by educators and researchers all over the world to test

their reading, math, and science knowledge and 21st century skills to meet real-world

challenges, also compared U.S. student performance to the performance of students in

other countries. In the 2018 PISA results, the latest administration of the PISA

assessment, the U.S. average mathematics literacy score was lower than the average (30

of 77) of other educational systems including countries such as China, Switzerland,

Germany, Italy, and France; and the U.S. score was 11 points lower than the overall

average of scores of participating countries. For the science component of the assessment,

the U.S. scored lower than 11 of the 77 participating countries, and the average score was

13 points higher than the average score of all participating countries. Results show a

continuing need to advance science and math knowledge and skills in education in the

U.S. Due to the COVID-19 pandemic, the PISA was not administered in 2021 and may

not be administered until 2022 (Organisation for Economic Cooperation and

Development, 2021).

David (2019) shared that companies are faced with the challenge of moving

forward due to the lack of meeting the demand of the changing workforce. This is a result

of limited candidates who are qualified and capable of handling STEM-related issues.

David also emphasized the fear of automation of jobs through the invasion of

technologies into almost every business in the world. The world is becoming more

technologically advanced and along with that will come an ever-increasing need for

workers who can handle those changes, among other global issues. Furthermore, as

mentioned in Chapter 1, the U.S. continues to lag other countries in advancing and
 21

developing competent STEM professionals. With this in mind, the more students we can

encourage to develop strong interests, skills, and abilities in STEM, the more likelihood

there will be of increasing the percentages of high school graduates who are going into

the STEM field of study, which ultimately could have a positive impact on our nation’s

global competitiveness. To maintain its status as a global leader in STEM, be competitive

in the STEM space, and address global challenges, the U.S. must make it a priority to

diversify STEM education to a STEM career pathway. The U.S. must take intentional and

strategic action to not be left behind in the innovative, global STEM space (Coleman,

2020). The National Science Foundation (2014) stated, “The U.S. STEM workforce must

be considered in the context of an expanding and vibrant global scientific and

technological enterprise” (p. 19). Coleman (2020) suggested that “without the

participation of individuals of all races and genders, the increasing demand for workers in

STEM fields will not be met, potentially compromising the position of the United States

as a global leader” (p. 1).

Research has demonstrated that there is a strong correlation between student

attitudes, dispositions, and self-efficacy and student career pathway choices and

performance in STEM-oriented courses. There are also correlations between developing

student career interests during the middle school years and their career pathway choices

in high school; however, there are gaps in research that do not show the impact student

participation in a STEM-focused middle school throughout their entire 3 years of middle

school has on their dispositions, self-efficacy, persistence, and performance once they

enter high school.

In this study, I examined the STEM persistence and academic performance of

 22

students who attended a STEM-focused middle school from Grades 6-8. Chapter 2 begins

with the review of relevant literature strategy and theoretical foundation for the study. In

the literature review, I outline relevant concepts from current research related to the

problem and purpose of this study. First, I describe Lent et al.’s (2002) SCCT that serves

as a framework for this research study and explain how the theory applies to the study as

well. Next, I define STEM and STEM education in their historical to contemporary

context. Then, I present STEM instructional best practices. Additionally, I discuss the

importance of STEM during middle school and the impact of participating in STEM-

focused middle schools. Following that discussion, I share controversies with STEM and

STEM education. Finally, I discuss student dispositions, factors affecting STEM self-

efficacy, and persistence in STEM.

Literature Search Strategy

A variety of databases, search engines, and other resources were utilized in this

review of the literature. The types of literature and sources included empirical research

articles from peer-reviewed journals, books, electronic newspapers, electronic

encyclopedias, dissertations based on similar research, STEM organizations’ websites,

and published reports. Database and search engines used were Ebsco Academic Search;

Google Scholar; Education Resource Information Center (ERIC); Research Gate;

ProQuest Research Library; ProQuest Central; ProQuest Dissertations & Theses Global:

The Humanities and Social Sciences Collection; and Gardner Webb University’s Bulldog

One-Search. Relevant documents from 1977 through 2022 uncovered several key themes

for this research study: STEM crisis, why STEM matters, student participation in STEM,

advancing STEM education, middle school student STEM interests, underrepresented

 23

minorities and the gender gap in STEM, and influences that impact student participation

in STEM. Key search terms and combinations of search terms for these topics included

SCCT, underrepresented minorities in STEM, self-efficacy, STEM education, middle

school and STEM, gender gap in STEM, STEM policy, STEM skills, STEM career

pathways, STEM pipeline, STEM integration, STEM dispositions, STEM persistence,

STEM academic performance, and STEM in the United States. To further expand my

research, the key terms, along with references from individual documents, were used to

help further expand the research to ensure all relevant concepts, to the extent possible,

were included in the literature review. In my initial search, I noticed there was very little

research that focused specifically on middle school student self-efficacy, disposition,

persistence, and performance in STEM. Of those research areas, specific STEM programs

were studied, and only one of the four areas was researched in an individual study on the

impact of an individual STEM program.

Theoretical Foundation

The theoretical framework for this research study is SCCT, developed by Lent et

al. (2002). The theory complements and builds on to incorporate a variety of concepts

and ideas from earlier career development theories. This theory also builds upon

Bandura’s (1989) social cognitive theory. The goal of the theory was to “adapt, elaborate,

and extend the aspects of Bandura’s theory that seemed most relevant to the processes of

interest formation, career selection and performance” (Lent et al., 2002, p. 257). SCCT is

supported by the notion that people’s beliefs about themselves, their environments, and

possible career paths help construct their career outcomes (Borgen, 1991). “An array of

factors such as culture, gender, genetic endowment, socio-structural considerations, or

 24

disability and health status operate in tandem with people’s cognitions, affecting the

nature and range of their career possibilities” (Lent et al., 2002, p. 256).

Major Theoretical Propositions or Major Hypotheses

Self-efficacy beliefs, outcome expectations, and goals serve as the foundation

upon which SCCT was developed. Self-efficacy refers to an individual’s personal beliefs

about their capabilities to perform particular behaviors or courses of action (Bandura,

1986; Social Cognitive Career Theory, 2021). “The strength of people’s convictions in

their effectiveness is likely to affect whether they will even try to cope with given

situations” (Bandura, 1977, p. 193). Outcome expectations refer to a person’s beliefs or

estimates about the consequences or outcomes of performing a given behavior (Bandura,

1977; Social Cognitive Career Theory, 2021). “An efficacy expectation is the conviction

that one can successfully execute the behavior required to produce the outcomes”

(Bandura, 1977, p. 194). Personal goals may be defined as one’s intentions to engage in a

particular activity or to attain a certain level of activities (Social Cognitive Career

Theory, 2021).

Self-efficacy can play a role in not only how you feel about yourself, but whether

or not you successfully achieve your goals in life (Cherry, 2020). “Self-efficacy also

determines what goals we choose to pursue, how we go about accomplishing those goals,

and how we reflect upon our performance” (Cherry, 2020, para. 9). SCCT seeks to

explain “three interrelated aspects of career development: (1) how basic academic and

career interests develop, (2) how educational and career choices are made, and (3) how

academic and career success is obtained” (Social Cognitive Career Theory, 2021, para.

1).
 25

How Basic Academic and Career Interests Develop: Interest Model

“SCCT’s interest model emphasizes both the experiential and cognitive factors

that give rise to career-related interests while tracing the role of interests in helping to

motivate choice behavior and skill acquisition” (Lent et al., 2002, p. 256). “Interests in

career-relevant activities are seen as the outgrowth of self-efficacy and outcome

expectations” (Social Cognitive Career Theory, 2021, para. 5). Throughout childhood and

adolescence, people are exposed, directly and vicariously, to an array of activities such as

crafts, music, sports, mathematics, and mechanical tasks that have potential relevance to

occupational behavior in school, at home, and in communities (Lent et al., 2002; Social

Cognitive Career Theory, 2021). They are also differentially reinforced for pursuing

certain activities, continuing their engagement, developing their skills, and achieving

particular levels of performance in different activity domains. The types and varieties of

activities to which children and adolescents are exposed are partly a function of the

context and culture in which they grow up (Social Cognitive Career Theory, 2021).

SCCT emphasizes that people form a lasting interest in an activity when they

view themselves as competent at it and when they anticipate that performing it will

produce valued outcomes (Bandura, 1986; Lent et al., 2002). “Through continued activity

exposure, practice, and feedback, people refine their skills, develop personal performance

standards, form a sense of their efficacy in particular tasks, and acquire certain

expectations about the outcomes of activity engagement” (Social Cognitive Career

Theory, 2021, para. 6). Figure 1 provides an overview of how interests develop over

time, according to SCCT.

 26

Figure 1

Career-Related Interests and Choice Development Over Time

Adapted from R. W. Lent, S. D. Brown, and G. Hackett (2002).

How Education and Career Choices Are Made: Choice Model

Mostly resulting from self-efficacy and outcome expectations, career-related

interests foster academic and career choice goals, especially when those goals are

supported in a person’s environment.

Choice goals are sometimes influenced more directly and potently by self-efficacy

beliefs, outcome expectations, or environmental variables than they are by

interests. Interests are expected to exert their greatest impact on academic and

occupational choice under supportive environmental conditions, which enable

people to pursue their interests. (Social Cognitive Career Theory, 2021, para. 10)

People’s free agency to choose can also be restricted by environmental conditions such as

cultural values, economic need, family pressures, or educational limitations, resulting in

them making choices that do not necessarily align with their interests as the type of work

that is available to them (Lent et al., 2002; Social Cognitive Career Theory, 2001). The
 27

choice model is built into the interest model as you see in Figure 1.

How Academic and Career Success Are Obtained: Performance Model

SCCT’s performance model is concerned with predicting and explaining two

primary aspects of performance: the level of success that people attain in

educational and occupational pursuits and the degree to which they persist in the

face of obstacles. SCCT focuses on the influences of ability, self-efficacy,

outcome expectations, and performance goals on success and persistence. (Social

Cognitive Career Theory, 2021, p. 1; Lent et al., 2002)

Also, “ability” is assumed to affect performance either directly or indirectly, considering

that performance involves both ability and motivation. It can directly influence if a

person does something at a higher level and more persistently than those who do not. On

the other hand, it can have an indirect influence on performance and persistence

depending upon a person’s self-efficacy and outcome expectations, especially

considering self-efficacy is a co-determinant of performance. These factors influence

what performance goals people set for themselves (Social Cognitive Career Theory,

2021). “Stronger self-efficacy beliefs and more favorable outcome expectations promote

more ambitious goals, which help people mobilize and sustain their performance

behavior” (Lent et al., 2002, p. 277). Figure 2 outlines how a person’s ability or past

performance, along with their self-efficacy and outcome expectations, impacts their

performance choices and goals.

 28

Figure 2

SCCT Performance Model

Adapted from R. W. Lent, S. D. Brown, and G. Hackett (2002).

Research and Practical Applications of SCCT Theory

In reviewing available research relevant to the application of SCCT, a

considerable amount of research has been accumulated suggesting that SCCT is a useful

framework for explaining various aspects of educational and vocational interest

development, choice-making, and performance; however, there is not substantial research

focusing specifically on middle school students, which seems to be a disadvantage,

especially considering research studies that will be discussed later in this literature review

that support the middle school years as a vital time to assist youth in developing their

career interests and goals. The theory has also recently been extended to the

understanding of academic and work satisfaction. SCCT has motivated and encouraged

researchers to design and test interventions aimed at various facets of career

development.

In a research study of the correlation between STEM career knowledge,

mathematics self-efficacy, career interests, and career on the likelihood of middle school
 29

students pursuing a STEM career, Blotnicky et al. (2018) conducted a study with a

sample of 1,448 public school students in Atlantic Canada in Grades 7 and 9. The

purpose of this study was to explore student knowledge of science and mathematics

requirements needed for STEM. Using SCCT as a theoretical framework, the study also

explored student math self-efficacy, future career interests, preferences for particular

career activities, and their likelihood to pursue STEM. The study resulted in students

lacking knowledge of requirements for math and science skills needed to pursue STEM

careers. It was also noticed that students with higher mathematics self-efficacy and

STEM career knowledge were more likely to pursue STEM careers. It also resulted in

students with greater interest in technology and science being more likely to pursue

STEM careers. The research’s conclusion was that there was a need to improve the

STEM knowledge and skills as well as awareness of STEM careers to middle school-age

students. The more exposure middle school students have to STEM, the more their

interest overall in STEM and their likelihood of pursuing a STEM career increase.

In a study conducted by Mueller et al. (2015), an adapted model of SCCT was

tested with a self-selected, diverse group of 186 middle school students in Grades 6-8

attending a STEM Saturday Academy located in a mid-southern city. The goal of the

study was to test the validity of utilizing SCCT to examine the career goals and choices

of middle school students who were already expressing an interest in math- and science-

related subjects and careers. It utilized the ideals of SCCT that the interaction of personal

and contextual factors, especially during the middle school years, may function

differently for self-selected students. A pre- and post-survey was utilized, and the

findings of the study revealed that math and science motivation, family support for
 30

engineering, outcome expectations, and interest were significant predictors of goal

intentions. Self-efficacy, on the other hand, was interesting and non-significant as other

research has supported that it was. An assumption from this research is that there may be

some measurement issues from SCCT that may need to be further researched. It does not

discredit SCCT, as it concluded that SCCT is a robust way to examine career goal

intentions among self-selected students but provides results that show that it is not always

as straightforward as it seems.

In examining underrepresented minorities in STEM, Fouad and Santana (2016)

utilized the SCCT model which they believed explained STEM choices and career

decisions for women and racial-ethnic minorities as well as barriers that may exist to

prevent entry into the STEM workforce. This research focused on factors with early

choices and consisted of studying existing literature related to SCCT and

underrepresented minorities. The research on SCCT with middle and high school students

provides rather consistent evidence that successful learning experiences help to promote

the development of self-efficacy and outcome expectations and that self-efficacy in math

and science is important in career development, specifically around supporting vocational

choices, interests, goals, and actions starting in adolescence. The evidence also suggests

that interventions to promote math and science career interests with underrepresented

racial-ethnic minorities should attempt to build parental support.

Collectively, these studies support Lent et al.’s (2002) SCCT that person inputs

(e.g., gender and/or race/ethnicity) play a significant role in both self-efficacy and

outcome expectations, especially for those in STEM fields. They further suggest that

efforts to build self-efficacy and outcome expectations via performance

 31

accomplishments, vicarious influences, and parental supports can, to some degree,

promote increased math and science interests and intentions among middle and high

school girls and racial and ethnic minority students. More research is needed to

understand career choices and the intersectionality of contextual factors with

developmentally and racially diverse adolescents. Future research can build on these

findings through examining intersections of race-ethnicity and gender to create

interventions to increase science/math-related self-efficacy, outcome expectations,

interests, goals, and actions.

Rationale for Theory Use

SCCT is appropriate for this study because it aligns with the purpose of the study

to discover the impact that STEM participation in a 3-year middle school program may

have on student dispositions, self-efficacy, persistence, and performance. The variables

being studied are all embedded within SCCT. Furthermore, this research adds to the

justification for studying these variables. As previously mentioned, dispositions, self-

efficacy, persistence, and performance are variables of SCCT. The results of the presence

of each of these variables within middle school students included in this study can

support the extent to which STEM-focused middle schools may be intervening to have a

positive impact on those variables.

My research study can assist in further building upon research on a variety of

STEM programs and other interventions that can motivate students to become engaged

and remain engaged in STEM-oriented courses and to develop the skills necessary to

build their self-efficacy and outcomes in those courses, which can have a positive

influence on their decisions to follow STEM career pathways once they enter high
 32

school. Utilizing SCCT in the STEM context can provide more support for interventions

that can take place to address barriers or other issues that may negatively impact student

interests, self-efficacy, and persistence in STEM. Furthermore, SCCT provided a

validated, well-established theory on which data analysis could be examined. According

to Fouad and Santana (2016), SCCT has been proven to be consistent across all

subgroups of people, including minorities, for predicting STEM career interest and

choice.

Defining STEM From a Historical Context

The origins of STEM go back as far as the Morrill Act of 1862 which promoted

agricultural science, and eventually engineering, as more land grant universities were

developed. In 1958, during the Cold War and space race, the U.S. began recognizing

science education on a national level following Russia’s launch of Sputnik that year.

Eisenhower initiated the National Aeronautics and Space Administration and Kennedy

later promoted scientific advancement leading to the 1969 landing on the moon.

Technology advances became more prevalent in the 1970s and 1980s where computers,

cell phones, the development of the first artificial heart, and the first space shuttle landing

helped the U.S. realize the importance and need for improving science education (STEM

School, 2021).

According to Hallinen (2020), STEM, which currently stands for science,

technology, engineering, and mathematics, was introduced in 2001 by the U.S. National

Foundation originally as SMET and later rearranged to STEM. In the early 2000s, it

became increasingly integrated into topics of education in the U.S. due to research at the

time that emphasized the links between STEM, prosperity, and knowledge-intensive jobs.
 33

In addition, many definitions of STEM were spreading around and there were concerns

that no one could agree on one solid definition. With this in mind, the Claude

Worthington Benedum Foundation conducted a study, administered jointly by Carnegie

Mellon University and the Intermediate Unit 1 Center for STEM Education, where they

administered and collected data from surveys to examine educator knowledge of STEM

and identify K-12 system needs in southwestern Pennsylvania (Tsupros et al., 2008). The

study consisted of a survey and focus groups of 350 educators in the region. The study’s

results provided conclusions that educators on all levels of education needed further

professional learning in STEM including increasing their awareness of STEM, assisting

them in further understanding how their work as educators contributes to STEM (Tsupros

et al., 2008).

Results also found that educators were very interested in collaborating with

postsecondary institutions to help them and their students learn more about what it means

to be a STEM professional and how the content they taught their students aligned with

the work of STEM professionals. There was expressed interest to gain support in

developing interdisciplinary units of study with the existing curriculum to provide

students opportunities to develop their innovation and team-based problem-solving skills

(Tsupros et al., 2008). As a result of this study, a definition for STEM was developed and

has been widely used by researchers due to its links to education goals and workforce

needs. STEM is defined as,

an interdisciplinary approach to learning where rigorous academic concepts are

coupled with real-world lessons as students apply science, technology,

engineering, and mathematics in contexts that make connections between school,

 34

community, work, and the global enterprise enabling the development of STEM

literacy and with it the ability to compete in the new economy. (Hallinen, 2020,

para. 7)

STEM Education in the U.S.

STEM Education Acts

In 2002, President George W. Bush signed into law the No Child Left Behind Act

of 2001 (a reauthorization of the Elementary and Secondary Education Act) requiring

state-level standardized reading and math testing in Grades 3-8 and once in high school.

The goal of the act was to maintain high academic standards in those subject areas. All

students were expected to meet or exceed state standards in reading and math by 2014

(No Child Left Behind Act, n.d.; STEM School, 2021). Although there were some

positives to this controversial act, including the “long-term Nation's Report Card (NAEP)

results, released in July 2005, showing elementary school student achievement in reading

and math at all-time highs and the achievement gap closing” (U.S. Department of

Education, 2006, para. 2), U.S. students were still falling behind in science (STEM

School, 2021).

In 2009, President Obama established the Educate to Innovate campaign. Its goal

involved “moving American students from the middle to the top of the pack in science

and math achievement over the next decade” (STEM School, 2021, para. 4). The

initiative “included preparing 100,000 STEM teachers by 2021 and called for increasing

federal funding toward STEM education” (Office of Secretary of State, 2009, para. 1).

Also, through the initiative, President Obama’s goals were to increase STEM literacy so

all students can think critically in science, math, engineering, and technology, “thus
 35

improving the quality of math and science teaching so American students are no longer

outperformed by those in other nations; and to expand STEM education and career

opportunities for underrepresented groups, including women and minorities” (Office of

Secretary of State, 2009, para. 6).

To further support the initiative, Obama established a $4.35 billion Race to the

Top fund. This fund incentivized states to commit to a comprehensive strategy to

improve STEM education.

It challenged states to dramatically improve their schools and student achievement

by raising standards, using data to improve decisions and inform instruction,

improving teacher effectiveness, using innovative and effective approaches to turn

around struggling schools and making it possible for STEM professionals to bring

their experience and enthusiasm into the classroom. (Office of the Secretary of

State, 2009, para. 13)

In December 2015, President Obama signed into law the Every Student Succeeds

Act as a replacement and update of the No Child Left Behind Act to ensure fair,

equitable, and high-quality education for all children. The purpose of the act was to

address and close achievement gaps as well. This act funds and enforces the Title I-Title

IX requirements including improving basic programs operated by state and local

agencies; preparing training and recruiting high-quality teachers, principals, and other

school leaders; ensuring language instruction for English learners and immigrant

students; establishing U.S. schools as 21st century learning facilities; providing

opportunities for state innovation and local flexibility; ensuring native Hawaiian, Indian,

and Alaskan education; laws for the homeless; and many other provisions (U.S.
 36

Department of Education, 2016b).

In addition, the STEM Education Act of 2015, which added computer science to

the STEM curriculum and provided more teacher training, was signed into law (STEM

School, 2021, para. 9). This act provided a scholarship program funded by the National

Science Foundation to math and science teachers to expand and boost research and

training opportunities on formal and informal STEM education through the National

Science Foundation scholarship program. Also, for federal purposes, it explicitly

integrated computer science into the definition of STEM education (Henry, 2015).

In 2017, the Inspiring the Next Space Pioneers, Innovators, Researchers and

Explorers (INSPIRE) Women Act was signed into law by President Trump. This law

authorizes the head of NASA “to support initiatives that will encourage women and girls

to study science, technology, engineering and mathematics and to pursue careers that will

further advance America's space science and exploration efforts” (Bryner, 2017, para. 1;

Inspire Women Act, 2017). The act supported initiatives such as NASA Boys; NASA

Girls; and Summer Institute in Science, Technology, Engineering, and Research. In the

same year, Trump also signed into law the Promoting Women in Entrepreneurship Act

which encouraged the National Science Foundation to help women succeed beyond

working in laboratories and succeed in the commercial world with jobs in STEM (Bryner,

2017; Promoting Women in Entrepreneurship Act, 2017). This bill made “education and

skills-training programs more accessible for women and other underrepresented groups

and makes it clear that we can and should do more to support women when it comes to

commercializing great ideas, starting small businesses, and creating jobs” (Promoting

Women in Entrepreneurship Act, 2017, p. 1).

 37

STEM Curriculum

STEM is an education curriculum that focuses mostly on helping students develop

and/or enhance their 21st century skills in the subjects of science, technology,

engineering, and mathematics. It is interdisciplinary and assists students in applying 21st

century skills based on real-world concepts, issues, and problems. Around the world,

STEM is a growing movement in education (Bybee, 2010; Hom, 2014). STEM

emphasizes problem-solving, innovation, and design, which are significant to every

country (Bybee, 2010). STEM learning environments are those that are globally

competitive and promote deep understanding and transfer of knowledge among

disciplines through providing education that is personalized, inclusive, flexible,

collaborative, student-centered, engaging, and exciting (Bybee, 2010).

Within the STEM curriculum, there are certain skills students are developing and/

or enhancing. When the literature is examined, it is seen that some skills are accepted as

STEM skills and there is a common understanding. These skills are emerging in the form

of engineering-based problem-solving, association skills, engineering-based design,

innovation, digital competence, creativity, and communication and collaboration (Sen et

al., 2018). STEM skills also include analyzing, asking questions, and drawing

conclusions on research; developing project plans and timelines; breaking down complex

systems into smaller pieces; identifying cause and effect; defending opinions with facts;

using math skills for measurements and calculations; paying attention to details;

accurately recording data; writing instruction; troubleshooting technical issues; repairing

machines; debugging operation systems; and staying abreast of current software and

equipment (Understanding STEM Skills, 2021).

 38

In addition to these very technical skills, there are also “soft skills” students

would need to learn in STEM fields which include communication, cooperation,

listening, collaboration, creativity, problem-solving, innovation, leadership, and

organizational skills (Understanding STEM Skills, 2021). Other important STEM skills

students are taught in STEM education include statistics, argumentation, logical

reasoning, intellectual curiosity, data-driven decision-making, and flexibility aimed at

training individuals to meet the needs of the 21st century workforce (Adams, 2017;

Moore, 2009).

Multiple Perceptions of STEM Education

The STEM curriculum and concept continues to flourish and change. Schools

increasingly provide application and problem-solving experiences to create more

awareness of STEM among students of diverse backgrounds. “Some educators advocate

for the inclusion of arts and humanities, suggesting that the acronym be changed to

STEAM. Other educators argue that a STEM curriculum should include the history of

science, particularly the contributions of women scientists” (STEM School, 2021, para.

12).

Some possibilities, all related to one another, include increased emphasis on

technology and engineering, the opportunity to stress 21st century skills, and the

development of an integrated curricular approach to studying grand challenges of

our era, such as energy efficiency, resource use, and other socio-environmental

topics. These areas can all be useful in developing and supporting STEM literacy.

(Bybee, 2010, p. 31)

 39

STEM Education Best Practices and Impact on Academic Achievement

21st Century Skills Development. STEM education supports the development of

21st century skills in all subject areas, including STEM literacy, problem-solving,

creativity, critical thinking, adaptability, overcoming adversity, innovation, and design.

Due to our world consistently evolving globally and becoming increasingly complex, our

students must be able to know how to use the knowledge skills they attain to make sense

of information and solve complex problems (Ball et al., 2016; Dede, 2010; Jr. Tech,

2020; Pearson, 2014). A 21st century education provides an “interdisciplinary,

personalized, inclusive, flexible, collaborative, student-centered, engaging and exciting

teaching environment” (Cookson, 2009, as cited in Sen et al., 2018, p. 81). Twenty-first

century skills are vital for students to master because they will also make them more

employable and lifelong learners (DeAngelis, 2015).

According to Jr. Tech (2020), “students become competitive and marketable with

the critical 21st century STEM skills--broad-based critical thinking, problem solving,

innovating and inventing, analytical and logical thinking, technological literacy,

collaboration, and leadership” (para. 3). Figure 3 outlines the unifying themes evidenced

in both 21st century instruction and the 21st century classroom.

 40

Figure 3

Profiles of the 21st Century Learner and Learning Environment

Note: Unifying themes of 21st century education and the 21st century learning

environment. Reprinted from Jr. Tech, 2020. Retrieved July 10, 2021, from

https://juniortech.org/jr-tech-mission/stem-education/

Ball et al. (2016) conducted an exploratory study that investigated 21st century life

and career skills and 21st century skill classroom environments. It also explored middle

school student perspectives of their self-efficacy in these skills. Through surveys

administered to 262 middle schoolers who participated in the study, they discovered that

the instruments reliably measured leadership; responsibility; collaboration; adaptability;

goal setting; self-regulation; social, cross-cultural, productivity, and accountability skills;

and their perspectives of the 21st century learning environments in which they were

engaged and how it impacted their skill development. This study supports the importance

of assisting students in developing 21st century skills within their classroom

 41

environments. These types of environments can assist in building student self-efficacy

and persistence in their subject areas and can have a positive impact on their academic

achievement overall.

Integrated STEM Approach to Science. What separates STEM from the

traditional science and math education is the blended learning environment and showing

students how the scientific method can be applied to everyday life. It teaches students

computational thinking and focuses on the real-world applications of problem-solving

(Hom, 2014, para. 9). Several studies have reported that STEM-based approaches

improve student achievement in science (Yaki et al., 2019) because it is a nontraditional

instructional strategy that provides opportunities for active engagement (Krajcik, 2015).

Acara et al. (2018) adopted a quasi-experimental design to determine the effects of

STEM-based instruction on mathematics and science achievements. Their findings

indicated that students improved their science scores.

Integrate STEM Approach to Technology. The technology component in

STEM refers to the tools and delivery modalities that help students become

technologically proficient learners, users, and consumers, as well as being the technical,

disciplinary training related to the needs of industry and workforce preparation. When

students create, design, and program products with the support of technology, it deepens

their learning. (Yang & Baldwin, 2020, p. 1)

Advancements in educational technology have provided various opportunities for

supporting student learning, and they offer unique affordances for complex, integrated

STEM learning environments. Technology can bring remote subject content experts into

the classroom to make up for the potential lack of content knowledge on the instructor’s
 42

part in an integrated STEM learning environment (Smith & Mader, 2017; Yang &

Baldwin, 2020). Technology offers various opportunities for students to become engaged

in the subject content. Technology can facilitate the exploration of STEM subjects and

provide support for students to connect different disciplinary ideas, for example, when

using simulations. Technology also encourages students to reorganize scientific and

mathematical ideas in a new way (e.g., building robots or creating content).

Integrated STEM Approach to Engineering. The most effective and engaging

way to help students develop their problem-solving skills is through engineering projects.

These types of projects require students to design solutions for authentic, real-world

problems. Problem-solving is the main goal of engineering and also aids in developing

creativity; organization and logic; clear and precise problem formulation; and knowledge

of math, science, and technology (Truesdale, 2014).

An instructional design strategy that many STEM programs utilize, the

Engineering Design Process, assists students in learning how to be creative and

innovative in their thinking. The process is iterative as it involves students asking

questions to identify the problem they are trying to resolve or address, imagining or

brainstorming ideas, planning designs, creating models and prototypes based on their

designs, testing out their models with data collection, and then making improvements on

their results by going through the process over and over again until they have solved the

problem. The goal of this process is to encourage students to work together in teams and

utilize their creativity and reasoning skills to develop solutions to problems (Hafiz &

Ayop, 2019).

Syukri et al. (2018) conducted a quasi-experimental study on the impact of the

 43

integration of the engineering design process on improving secondary school students’

understanding of solving physics problems. There were two groups of students in the

experiment who were taught a physics unit, one using the engineering design process and

one using another teaching method. The study resulted in the experimental group–the one

that utilized the engineering design process–yielding higher mean scores when assessing

problem-solving skills, showing it as the most effective of the two methods. In another

study, Goktepe Yildiz and Ozdemir (2020) investigated the effects of engineering design-

based instruction on spatial abilities of eighth-grade students through an experimental

study where 75 students were divided into groups with the experimental group utilizing

the engineering design process to complete mathematical activities to demonstrate their

spatial skills. This study resulted in the engineering design process having a positive

impact on student development of spatial skills. Many other research studies on the

impact of the engineering design process integration into learning activities yielded

positive results as well.

Integrate STEM Approach to Mathematics. Mathematics, in K-12, assists

students in making sense of the world around them through developing problem-solving,

reasoning, communicating, and mathematical modeling skills. For STEM learning,

mathematics is seen as more than simply a set of tools for these disciplines. To better

connect mathematics and other disciplines in STEM, we should focus on ideas and

thinking development in mathematics, unifying instruction from the student perspective

(Li & Schoenfeld, 2019).

Project-Based Learning. Another STEM-based instructional practice is project-

based learning (PBL). PBL combines standards-based content with a real-world

 44

challenge so students can use what they have learned to investigate concepts further and

create real-world connections (STEM & Project-Based Learning, 2021, para. 2). Kolk

(2018) described PBL as follows:

In a project-based approach to learning, students are first presented with a real-

world problem or issue and then learn the content necessary to answer questions

they have derived in response to the problem. During the process of questioning,

research, ideation, and developing solutions, students build the problem-solving,

project management, collaboration, and leadership skills necessary for success in

the world beyond the classroom. PBL helps students bridge thinking across

disciplines, promotes deeper connections to content, fosters the inquiry skills

necessary for success in STEM, and fosters reflection and metacognition. Taking

a project-based approach to STEM learning can help students form deeper

connections to content, connect ideas across disciplines, and build the

questioning, thinking, and metacognitive skills necessary for success in today’s

rapidly-changing world. (paras. 2-3)

There are many advantages to using PBL for instruction. One of the main

advantages of PBL is that it assists students in developing their technical, personal, and

contextual competencies. Another advantage is that it engages students in solving real-

world problems from professional contexts. One final advantage is that it promotes

collaboration between students. These advantages are important for student development

of their 21st century skills which include problem-solving, communication, collaboration,

and critical thinking (Li & Schoenfeld, 2019).

 45

Importance of STEM Education During Middle School Years

Research on the relationship between student interest in and the pursuit of STEM

careers has increased in recent decades. This may be due to the ongoing increasing need

for skilled workers in STEM fields. With this comes a strong need as well to inspire and

encourage young students to pursue a career in STEM fields (Almeda & Baker, 2020).

One reason students may not pursue STEM careers is a lack of early awareness and

exposure to STEM fields and careers, therefore lacking the knowledge they need to

consider a career in a STEM field (Christensen & Knezek, 2017). Middle school is an

opportune time to assist students in developing an awareness of STEM, provide them

opportunities to explore STEM fields and occupations that can be pursued, and more

importantly guide student interests towards STEM disciplines, as they begin to think

about their future and careers they may want to attain (Almeda & Baker, 2020; Hom,

2014). “Student exploration of STEM-related careers begins at this level, particularly for

underrepresented populations” (Hom, 2014, para. 10).

Attitudes students develop during their middle school years largely influence

student academic performance. Research suggests that students can be motivated if their

beliefs about their probability of success are improved (outcomes expectancy) and if

students are interested in a task or see the value or worth of the task for themselves. This

can affect student career aspirations (Lent et al., 2002; Van Tuijl & van der Molen,

2016). In addition, feedback from parents and teachers on effective stereotypical values

about work in STEM fields are important for building career interests and career

development in children ages 8 to 16 and can assist in building negative or positive

dispositions towards STEM (Van Tuijl & van der Molen, 2016); therefore,
 46

“understanding middle school students’ perceptions regarding STEM dispositions is vital

to preparing our future STEM workforce as well as future citizens” (Choi & Chang,

2011, p. 2).

Sadler et al. (2012) conducted a study on a cohort of 34 beginning college

students from 2- and 4-year colleges and universities who were enrolled in mandatory

English courses at their universities; 6,860 PRiSE surveys were completed which asked

students what different items BEST described what they want(ed) to be at different points

in their lives and provided them a detailed list of 19 career fields from which to choose.

Two of the points in their lives included the beginning and end of high school. The

survey results showed 75% of the male respondents and only 25% of the female

respondents favored engineering or science careers at the end of high school. Of the

males interested in those careers at the end of their senior year, nearly three-quarters had

already been interested at the start of their freshman year. Of the females interested in

those careers at the end of their senior year, approximately half had been interested at the

start of their freshman year.

Sadler et al. (2012) suggested that initial interest in particular careers may be

predictive of a greater or lesser likelihood of pursuing a STEM career. Using a variety of

statistical methods to further examine the variables in the study, Sadler et al. also

reported,

The odds of reporting a STEM career interest (rather than a career interest outside

of STEM) at the end of high school are about nine times as high for students who

reported an interest in engineering or science careers at the start of high school as

for students who did not report such an interest at the start of high school. (p. 419)
 47

They concluded the study stating that whatever student career aspirations are when they

begin high school strongly predicts their career aspirations at the end of their senior year,

thus finding evidence for the significance of establishing early career interest in science

for students.

In a study by Almeda and Baker (2020), an examination was conducted of 467

participants’ career pathways from their early learning, affect, and behaviors while using

ASSISTments online mathematics formative assessment and tutoring program in middle

school to whether or not they end up in a STEM field beyond college. The study resulted

in students who had a significantly higher mathematics proficiency pursued a STEM-

related career than those with lower proficiencies and suggested that “developing aptitude

in middle school math is positively associated with the decision to enroll in college,

pursue a STEM major, and participate in a STEM career after college” (p. 43).

Understanding how students’ early learning, affect, and disengaged behavior influence

their eventual choices of occupation will help provide a more comprehensive picture of

student pathways towards STEM fields. This study further supports the importance of

STEM experiences during middle school and their importance in impacting student

career pathways towards STEM fields.

STEM-Focused Schools

Many school districts have STEM-focused schools at the elementary, middle, and

high school levels. At STEM-focused schools, students learn through collaboration and

PBL, and they work in teams to engage in hands-on learning to come up with real-world

solutions to real-world problems. In addition, learning that involves assisting students in

developing problem-solving, creativity, and critical-thinking skills is at the core of any

 48

true STEM school. STEM-focused or STEM specialty schools can help students gain the

skills necessary for success in STEM fields. In STEM-focused schools, the entire

school’s focus is on STEM. Every student is given opportunities to participate in STEM

curriculum. STEM-focused schools encourage students to think and behave like

scientists. In addition, STEM-focused schools connect STEM learning to STEM-related

careers, integrate STEM with other subjects, and make use of technology as well

(Yednak, 2012).

Controversies With STEM and STEM Education

The literature has revealed mixed findings regarding the effects of STEM-based

approaches on student achievement (Berland et al., 2014; Guzey et al., 2017). Guzey et

al. (2017), in their study on the effects of design-based STEM instruction on student

achievement in middle school, found significant learning gains in physical science

content but no significant learning gains in life science and mathematics. They

highlighted that adding engineering casually into science instruction did not promote

meaningful learning but addressed the absence of meaningful integration.

In contrast, James (2014) found no significant differences between groups

instructed with STEM and non-STEM. Lachapelle et al. (2011) found no

significant differences between the post-test scores of students who participated in

STEM-based engineering design integration and the control group in a science

unit on organisms. (Yaki et al., 2019, p. 184)

Student Dispositions in STEM

Experts on the President’s Committee of Advisors on Science and Technology

stress the importance of improving the overall interest and attitude toward STEM among
 49

young students. This is just as important as increasing the overall level of academic

proficiency in STEM (National Research Council, 2011). Gaining “an understanding

middle school students’ perceptions regarding STEM dispositions, and the role attitudes

play in establishing STEM career aspirations, is imperative to preparing the STEM

workforce of the future” (Christensen & Knezek, 2017, p. 2).

Gender Gaps in STEM Careers

Although there has been and continues to be substantial growth in STEM jobs,

research experts have identified gender and racial gaps in STEM careers. This discovery

leads to the ongoing need to increase the number of women and ethnic minorities in

STEM (STEM School, 2021). In 2011, the National Research Council called on

educators to increase the number of students pursuing STEM career pathways after high

school, including students from groups traditionally underrepresented in STEM—

students of color, women, and students from low socioeconomic backgrounds.

Although beneficial for all students, the STEM curriculum is aimed toward

attracting underrepresented minorities in STEM, such as female students who are

significantly less likely to pursue a college major or career. Though this is nothing new,

the gap between male and female students pursuing STEM careers is increasing at a

significant rate.

Male students are also more likely to pursue engineering and technology fields,

while female students prefer science fields, like biology, chemistry, and marine

biology. Overall, male students are three times more likely to be interested in

pursuing a STEM career. (Hom, 2014, para. 11)

In a Going Green! Middle Schoolers Out to Save the World (MSOSW) project,
 50

Christensen and Knezek (2017) shared that the results of this project concur with the

ACT findings that a gap exists among young people across the U.S. regarding positive

interest in STEM as a career versus stated intent to pursue a STEM career. The findings

from the current study also provide evidence that progress can be made toward

eliminating the existing gender gap in STEM career interest and intent and that hands-on

science activities, such as those embedded in the MSOSW project, are particularly

effective in enhancing STEM career interests among middle school girls. This is true for

girls whether or not they begin project activities planning to pursue a career in STEM or

not.

Racial Gaps in STEM Careers

Historically, Asian and American Indian students have displayed the highest level

of interest in STEM fields. Before 2001, students of an African American background

also showed high levels of interest in STEM fields, second only to the Asian

demographic, but has since dropped dramatically to lower than any other ethnicity (Hom,

2014). “The lack of equity or the presence of achievement gaps between high-, medium-,

and low-ability students, especially in traditional classroom environments, could

negatively affect the interest of low- and medium-ability science students and their

subsequent choice of STEM careers” (Lin & Lin, 2016, p. 1375).

Coleman (2020) conducted a study through the Illinois Math and Science

Academy that was purposed to assist in gaining a better understanding of how to

diversify STEM education to career pathways to address the disproportionality of Black

and Latino access and exposure to STEM, thus impacting their STEM literacy. They

gathered the perspectives of 415 students, parents, teachers, and Black and Latino
 51

professionals. Through two STEM Think Tanks, participants shared their stories related

to the intersection of race and STEM. Data collected from the study resulted in themes

including

obligation to Black/Latinx communities to break the negative stigma and be

different; future success because STEM is a prominent, progressive field;

learning/discovery of STEM knowledge and real-life applicability; STEM passion

and enjoyment; and solve problems to advance humanity. (Coleman, 2020, p.

283)

Verriden (2017) conducted a qualitative study of African American girls ranging

from ages 13-15 on their sense of belongingness as related to teachers, family, and peers.

The purpose of the study was to identify factors relevant to their self-concept and sense

of belongingness in math and science and how to help them overcome barriers including,

but not limited, to sexism and racism. In addition to examining teacher, family, and peer

influences, prep programs, school atmosphere, community neighborhood atmosphere,

STEM interests, identity and beliefs, academic self-concept, and future ambition were

examined as well. Results of the study were that African American students have low

self-efficacy and STEM career knowledge that significantly impact adolescent pursuits of

STEM-related careers; limited knowledge of STEM preparation needs and STEM

careers; and lower self-efficacy and interest in STEM as they progress through high

school. Students with higher STEM career knowledge and math self-efficacy were

slightly more likely to pursue a STEM career. The study provided implications for

developing and/or increasing knowledge and experience in employing STEM skills

important for the STEM workforce. Finally, the study provided overall conclusions that
 52

middle school students have limited career knowledge, low math-self efficacy, and a

declining interest in STEM.

The U.S. Department of Education (2016b, as cited in Coleman, 2020) provided

the following statement in STEM 2026: A Vision for Innovation:

How STEM is messaged to youth and their families is transformed. Research

shows that repeated exposure to images, themes, and ideas affects people’s

beliefs, behaviors, and attitudes. In STEM 2026, popular media, toy developers,

and retailers consider issues of racial, cultural, and gender diversity and identity in

portrayals of STEM professionals and STEM-themed toys and games. These

images counter-historical biases that have prevented the full participation of

certain groups of individuals in STEM education and career pathways. These

portrayals include diverse pictures, descriptions, or images of what STEM work

entails, including the array of jobs and activities that use STEM; and who is seen

doing and leading STEM-related work. Communities and youth in all

neighborhoods and geographic locations around the country are equally exposed

to social and popular media outlets that focus on STEM, and a wide diversity of

STEM-themed toys and games that are accessible and inclusive and effectively

promote a belief among all students that they are empowered to understand and

shape the world through the STEM disciplines. (p. 278)

As school districts continue to motivate more students to become interested in

STEM, it will be important that they take into consideration the discoveries made through

these studies, along with the Vision 2026 statement, to ensure STEM programs and

STEM-focused schools put practices into place that positively impact underrepresented
 53

minorities and their dispositions towards STEM.

Factors Impacting STEM Self-Efficacy in Students

As aforementioned in a previous section of this study, the middle school years,

when students are 12 to 15 years old, are important for assisting students in developing

interest and efficacy in STEM careers because this is when student beliefs about

competency and interests begin to solidify (Blotnicky et al., 2018). In revisiting the

theoretical framework for this study, SCCT (Lent et al., 2002) provides support for this in

its proposal that career interests, choice, and personal goals have a major impact on

performance, self-efficacy, and outcome expectations. Self-efficacy is positively related

to academic performance and impacts activities in which one decides to pursue,

ultimately impacting their success and continued interest in specific activities or careers.

In a 2017-2018 STEM study conducted by the Student Research Foundation

(2019), statistics were shared regarding barriers that exist between high school Hispanic

students and STEM careers. Although Hispanic Americans make up 21% of the 18- to

24-year-old population in the U.S., this study revealed that 12% of all college graduates

who earn STEM degrees are Hispanic Americans (Student Research Foundation, 2019).

This may be due to Hispanic students taking fewer STEM courses than other subgroups

of students. It also revealed that Hispanic high school students are less confident in their

STEM self-efficacy than other subgroups of students. Female Hispanic students

expressed less confidence in their STEM abilities than male Hispanic students. The study

revealed that only 22% of female Hispanic high school students express confidence about

their abilities in STEM subjects, compared to 30% of male Hispanic students. In addition,

female Hispanic high school students lag behind non-Hispanic female students, with 28%
 54

saying they are confident about their abilities in STEM (Student Research Foundation,

2019).

Societal Influences on STEM Self-Efficacy

Both STEM career knowledge and career interests are also influenced by society

at large. These society influencers include role models to whom students are exposed

either in person or through the media; the individuals students interact with daily such as

teachers, family members, and peers; and student extracurricular experiences. SCCT

proposes that these influencing factors predict the self-efficacy youth hold about their

career options as well as their outcome expectancies (Lent et al., 2002, as cited in

Blotnicky et al., 2018). Self-efficacy is considered a major predictor guiding the selection

of majors during high school and postsecondary education (Kelly et al., 2013, as cited in

Blotnicky et al., 2018).

Another theory, expectancy value theory, expanded from Vroom’s (1964)

expectancy theory, is much like the SCCT framework of this study, in that it proposes

that significant life choices people make and the decision-making processes they used in

making those choices are influenced by how much success a person believes they will

have and how much value they have in an activity (Eccles & Wigfield, 2022; Gottlieb,

2018; Lent et al., 2002). The theory goes on to propose that choice, such as what type of

career to pursue, can be impacted by factors such as cognitive processes, affective

memories and reactions, cultural stereotypes, and socialization. In regard to the pursuit of

STEM careers, girls tend to have lower self-efficacy beliefs due to low “task value,”

beliefs that science was not involved with helping other people, culturally shaped notions

of women in science, and misconceptions about women’s abilities in science (Eccles &
 55

Wigfield, 2002; Kijanka, 2009).

In addition to the expectancy theory is Bronfenbrenner’s (2005) bioecological

model. Established by psychologist and scholar Urie Bronfenbrenner, this theory

proposes how family, school, and community environmental factors can influence and

shape middle school Hispanic girls’ STEM interests (Bronfenbrenner, 2005; Lambert,

2018). This model provides specifics on microsystems (family and friends),

macrosystems (relationships within the microsystems in a person’s life), mesosystems

(relationships among teachers, parents, and peers), and macrosystems (overarching

beliefs and values passed from one generation to the next) and their influences on human

dispositions and self-efficacy.

Student Persistence in STEM

STEM career knowledge can be defined as familiarity with a STEM career and

can vary significantly based on the STEM career guidance a school provides. Without

adequate knowledge, there is a risk that students will dismiss a STEM-based career path

as a potential option for their future (Compeau, 2016, as cited in Blotnicky et al., 2018).

In addition to their knowledge, student career interests and their preferred future career

activities will also affect their intention of pursuing a STEM career.

In an Atlantic Canada public school study by Blotnicky et al. (2018), 1,448

middle school students in Grades 7-9 were surveyed, and the results revealed that while

older students had more knowledge about mathematics and science requirements for

STEM careers, this knowledge was lacking overall. Also, students with higher math self-

efficacy were more knowledgeable about STEM career requirements and more likely to

choose a STEM career. Students with greater interest in technical and scientific skills
 56

were also more likely to consider a STEM career than those who preferred career

activities that involved practical, productive, and concrete activities. The results of this

study show that students in middle school have limited STEM career knowledge

concerning subject requirements and what sort of activities these careers involve. In

addition, students with low mathematics self-efficacy have a declining interest in STEM

careers.

Overall, the results of the study support the need to improve access to knowledge

to facilitate student understanding of STEM careers and the nature of STEM work,

especially between the ages of 12 and 15. Exposure of students to STEM careers can

enhance their interest in and the likelihood of pursuing careers involving science,

technology, engineering, and mathematics. There seems to be a correlation between a

student’s interest in STEM entering high school and their interest upon graduation. In

addition, self-efficacy in STEM further determines student career choices (Blotnicky et

al., 2018).

Summary

An overview of Lent et al.’s (2002) SCCT provided a theoretical framework for

exploring the dispositions, self-efficacy, and persistence of the study population of

middle school students after enrollment in a 3-year STEM-focused middle school. This

included providing research background on the practical applications of the theory as it

relates to STEM careers. In my review of related literature, major themes aligning with

the purpose of the study included defining STEM and its historical to contemporary

context, STEM education in the U.S. including STEM acts that have influenced the

advancement of STEM education, STEM instructional best practices, importance of

 57

STEM during the middle school years, STEM-focused middle schools, controversies with

STEM and STEM education, student dispositions in STEM, factors impacting STEM

self-efficacy in students, and student persistence in STEM.

Research continues to show the growing need to improve STEM education in the

U.S. and to motivate more students to become interested in this steadily growing field. If

the U.S. is to improve its global standing in STEM and improve our economy overall, it

must continue to invest in STEM and ensure that STEM practices across the nation are

yielding positive results. It is known that different STEM practices can yield positive

outcomes on student STEM knowledge and skills and can enhance their academic

performance overall. STEM-focused schools have a great advantage of integrating STEM

into every facet of education and the culture of the school. From the social cognitive

theory and studies utilizing the theory, much is known about how student dispositions,

self-efficacy, and persistence can be impacted by many factors, but the middle school

years are the prime time for schools to have a significant impact on student dispositions,

self-efficacy, and persistence towards STEM regardless of external factors that may

influence student attitudes and interests. There is very little, if any, research on the impact

of the influence of STEM-focused middle schools on students who attend those schools

for their entire middle school experience. This study seeks to add to this gap in the

literature by providing data and results on STEM-focused middle schools in the largest

school district in North Carolina and the STEM dispositions, self-efficacy, persistence,

and performance of students who attended those schools.

 58

Chapter 3: Methodology

Purpose

The purpose of this quantitative study was to explore the impact STEM-focused

middle schools have had on student STEM persistence and performance in STEM as they

transition through their high school years. This includes students who have attended a

STEM-focused middle school for all three of their middle school years. Data collection

was in the form of requesting from the Data, Research, and Accountability (DRA) office

of the district for this study information rosters of current high school seniors that include

courses they took since they enrolled in high school and their performance in those

courses, more specifically the STEM-oriented courses (i.e., advanced math, advanced

science, technology courses, and other STEM-related elective courses). The data

collected address the persistence and performance of students in STEM beyond their

middle school years. Other variables included student gender and ethnicity. Analysis of

the data collected data was also compared to determine relationships between the

different variables.

In Chapter 3, I describe my research methodology for this study. First, I describe

the research design for the study and the rationale for implementing that design. Then, I

review how I collected a sample from my research population. Next, I discuss the

instrumentation that was used to collect data. Furthermore, I provide details on how I

analyzed the data collected for the study. Lastly, I conclude Chapter 3 by providing an

overview of threats to validity for the data and ethical considerations that are relevant to

this study.
 59

Research Questions and Hypotheses

The following is a reiteration of the research study questions and hypotheses:

1. What are high school seniors’ STEM persistence in high school after a 3-year

enrollment in a STEM middle school?

H1: High school seniors, after a 3-year enrollment in a STEM-focused

middle school, will enroll in advanced STEM-related core courses and

STEM-oriented electives throughout high school.

2. What are high school seniors’ academic performance in STEM courses in

high school after a 3-year enrollment in a STEM middle school?

H2: High school seniors, after a 3-year enrollment in a STEM-focused

middle school, will enroll in advanced STEM-related core courses and

STEM-oriented electives and experience high academic achievement in

those courses throughout high school.

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

H3: Due to the ongoing nature of educational inequality, minority subgroups

will have less positive results in persistence and performance than other

subgroups.

Setting

The research study took place in a school district in North Carolina. This school

district has been known and continues to be touted as one of the largest and leading

school districts in North Carolina, especially when it comes to innovation and

technology. The school district has several STEM-focused elementary, middle, and high
 60

schools and has invested, over the past 10 years, much time, money, and effort into

supporting STEM schools in the district. Despite its investments, there are very few, if

any, studies showing the level of impact or influence its STEM schools are having on

students. As one of the leading school districts in the state, I believed this setting was

appropriate for the study because it could provide useful feedback to all school districts in

North Carolina, especially in magnet and curriculum enhancement program offices and to

school administrators, STEM coordinators, parents, staff, and teachers at STEM-focused

middle schools.

Research Design and Rationale

In this nonexperimental quantitative research study, I examined the impact

STEM-focused middle schools have had on the dependent variables of STEM persistence

and academic performance of the independent variable of high school seniors who

attended their STEM-focused middle school for their sixth-, seventh-, and eighth-grade

years. While student enrollment was the primary independent variable, the data were also

aggregated by other independent variables including gender and ethnicity. The research

design I selected for this study was appropriate in that it not only aligned with the

research questions for my study but also purposefully included subjects who were

students who participated in a STEM-focused middle school throughout all their middle

school years. In addition, because this study is nonexperimental, it did not include the

random assignment of students to a control or treatment group.

Also, I did not influence the study as the researcher because the students included

had already completed their 3 years in the STEM-focused middle school and were

already in high school; therefore, the data I sought was historical since they had already
 61

participated. In addition, the data collected on the STEM-oriented courses in which they

had enrolled and/or completed, in addition to their performance in STEM-oriented

courses they already completed in high school, were collected beyond their attendance in

the STEM-focused middle school. This study also provided correlational data in that the

gender and ethnicity of students and their STEM persistence and academic performance

were reviewed and compared among subgroups.

Participants in this research study had already completed 3 years of middle school

at a STEM-focused school and had completed 3 or more years of high school. A posttest-

only design was appropriate for this study, considering the participants were completing

this survey after they had participated for 3 years already in a STEM-focused middle

school. Also, the data collected on these students, which again included their STEM-

oriented course selections and academic performance in STEM-oriented courses provided

data to support student persistence in STEM and their academic achievement as well.

Using this type of design aligned with the intended purpose of the research study and

assisted with examining the impact of STEM-focused middle schools on student STEM

persistence and performance. This contributes to further advancing research on STEM

education and how STEM-focused middle schools are assisting the nation in moving

towards its ongoing goals to attract more students towards STEM fields of study and

careers.

The student information roster, which includes data on student course enrollments

since entering high school after attending a STEM-focused middle school for 3 years,

their academic performance in those courses, and other variables including their gender

and ethnicity, assisted in answering the research questions for this study:
 62

1. What are high school seniors’ STEM persistence in high school after a 3-year

enrollment in a STEM middle school?

2. What are high school seniors’ academic performance in STEM courses in

high school after a 3-year enrollment in a STEM middle school?

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

Role of the Researcher

For this research study, it is important to note that during the years the students

attended their STEM-focused middle school, I was also a STEM coordinator for a

STEM-focused middle school. At the end of the 2018-2019 academic year, the school

was no longer a STEM-focused school and has since then transitioned into a Global

Studies and World Languages magnet school as well. Bias was reduced, however, since I

did not utilize this school in the study because it had not been a STEM school for the past

2 years, thus providing invalidating results if I had included it. Also, bias would have

existed if I interpreted data for my school versus others if it were included in the study.

In addition, the high school students whose academic data were collected had no

direct contact with me as the researcher and remained anonymous to me as well. Also, I

was never employed at the STEM middle school where those students attended;

therefore, I do not know who the students are whose data were used for the study. In

addition, I connected with the director of curriculum enhancement and magnet programs

for the district, who had a genuine interest in the research and the results of the research

as well. I believed that his involvement with assisting in obtaining the necessary data

needed to complete the study was important.

 63

Methodology

In the methodology section of Chapter 3, I provide information about the

population for this study. I also share my sampling procedures for obtaining a reasonable

sample including inclusion and exclusion criteria and recruitment procedures. I provide

an overview of the instruments that were used in collecting the data, how the variables in

this study were operationalized, and how the data were analyzed.

Population

The target population for this study were current high school seniors who attended

a STEM-focused middle school in the district for all 3 of their middle school years. Data

for this research were collected from high school seniors, a subset of the target

population, who were enrolled in the district for the 2021-2022 academic year and who

attended one of the STEM-focused middle schools in the district for their sixth-, seventh-,

and eighth-grade years. High school seniors in this population attended their STEM-

focused middle school from the 2015-2016 to the 2017-2018 academic years.

Sampling Strategy Identification and Justification

Although it would have been ideal to gather information from the entire target

population of high school seniors who attended STEM-focused middle schools, doing so

would have been unlikely and would not have been feasible for me if it were likely. For

this study, a probability sampling method was utilized where data from a specific group

of high school seniors from the sampling frame was used. A cluster sampling strategy

was used where the entire target population of high school seniors was divided into

clusters based on which STEM-focused middle school they attended and the primary base

school to which the majority of students from each middle school transferred for high
 64

school. Because the individuals in each sample must have similar characteristics, this

strategy seemed most appropriate considering that all the students were high school

seniors, included multiple ethnicities and genders, and transferred to their primary base

high school upon completing their sixth-, seventh-, and eighth-grade years at a STEM-

focused middle school (McCombs, 2019). Then, instead of sampling high school seniors

from each of the clusters, one of the clusters was selected. Because it was practically

possible to do so, data from each individual student in the selected cluster were able to be

utilized.

Participant Selection Criterion

The sampling procedures included both inclusion and exclusion criteria. For their

data to be included in the study, participants must have been high school seniors, must

have attended all 3 of their middle school years in a STEM-focused middle school, and

must be high school seniors at the primary base high school to which the majority of

students attending that STEM-focused middle school transfer upon completion of middle

school. The data used were from all the students included in the cluster sample described

in the sampling strategies and justification section of this research.

Number of Participants and Rationale

Using G*Power software, I conducted an a priori power analysis to determine

how much data I needed for my study for the results to be generalizable to the target

population (Faul et al., 2009). F tests were selected as the test family with a linear

multiple regression with a fixed model and an R2 deviation from zero as the statistical

test. Based on norms for educational research, the effect size selected was .25 (Lipsey et

al., 2012) with an alpha error of 0.05 and a power of .8. This resulted in a sample size of
 65

42 needed for the study (see Figure 4). The cluster of students totaled a sample size of 90

students of the total population of high school students in the district who attended a

STEM-focused middle school during the 2015-2016, 2016-2017, and 2017-2018

academic years. Doing so allowed me to generalize results to the target population,

ensuring that every time frame was covered.

Figure 4

A Priori Power Analysis Results

Participant Identification, Contact, and Recruitment

Several important steps were taken in the participation identification process.

Before collecting data, I ensured that I had completed and submitted all the appropriate

documents to Gardner Webb’s IRB for approval. The IRB approval process required that

the research already be approved by the district being used in the study. The district’s

DRA office requires researchers have university IRB approval first; therefore, I was able

to receive approval from the IRB to conduct the research and also met all the

requirements to obtain permission from the district’s DRA office to conduct the study.

Following approval, no parental permission was necessary to utilize the student

data requested from the district’s DRA office because the data provided were deidentified
 66

and my research required no active student participation to collect the necessary data;

therefore, no student recruitment efforts were necessary. The cluster sample of students

whose data were used for the research were all students who attended the same STEM

middle school for their sixth-, seventh-, and eighth-grade years from 2015-2016 to 2017-

2018. These students were all current high school seniors at the primary base high school

where most of the students from that middle school attend.

Procedures for closing this study were minimal considering the students from

whom the data were collected did not participate actively in the study (they did not

complete a survey, interview, focus study, etc.). The data collection process only

included reviewing, organizing, interpreting, and analyzing the student data to determine

trends in student STEM persistence and performance.

Instrumentation

Only one quantitative instrument was used to collect discreet data for this research

study. Because historical or existing data were already available, document review was

the best type of quantitative instrument to utilize for this study. I attempted earlier in my

research efforts to utilize a student survey in addition to document review, but due to

several limitations, I was unable to move forward with collecting data; how I could have

used this instrument will be further described in the recommendations for future research

section of this study.

To obtain the data needed for the research, I reached out to the district’s DRA

office to request and receive written permission to use rosters of archived data for

students who attended one of the STEM-focused middle schools during their sixth-,

seventh-, and eighth-grade years and were high school seniors at the primary high school
 67

into which the STEM-focused middle school feeds. The information requested for the

data reports utilized deidentifiers to protect student identity, gender, and ethnicity; all

courses in which they enrolled each year since leaving middle school; and grades and

EOC scores in each of those courses.

Data Collection Analysis and Rationale

In analyzing the data from the deidentified rosters of student information, which

included student ethnicity and gender, courses in which they enrolled since high school,

and their performance in those courses, descriptive statistics including measures of

central tendency were utilized to describe the data. The course enrollments were

compared using mean, median, and mode values. In addition, comparisons between

gender and race subgroups’ persistence and performance were compared to identify

similarities and differences between subgroups based on ethnicity and gender. This was

done using standard multiple regression tests to identify correlations between the

dependent and independent variables of the study.

Ethical Considerations

Researchers are required to develop competency in conducting ethical research

for students. This includes ensuring their research provides insight into academic issues

as well as ensuring the scientific credibility of research through there appropriate

utilization of research methods (Kaiser, 2019). Only deidentified rosters of student data

were utilized in this study; therefore, no parent or guardian permission was necessary due

to the anonymity of the students to whom the data belonged. In addition, the STEM-

focused middle school in which the students, whose data were provided to me from the

DRA office, were enrolled and the high school in which they were currently enrolled are
 68

not identified in the study. All data provided to me for the research will be kept in a file

cabinet locked and discarded after 2 years. Beneficence will be accomplished where

findings from the research study can be utilized by the districts to inform decision-

making around making improvements, if any, to STEM-focused middle schools and their

practices in motivating and supporting students to become interested in STEM and

develop STEM career pathways, especially once they are promoted to high school.

Summary

In Chapter 3, I provided specifics on quantitative research design and my

rationale for using that design. It also provided an overview of my methodology,

including the target research population, sampling procedures, and recruitment. Details

were provided on my role as the researcher in the study and detailed what participants in

the study did. Furthermore, data collection procedures along with a description of the

instrumentation used to collect the data were described and how each of the variables in

my study were operationalized. My plan for analyzing the data was provided along with a

discussion of the anticipated threats to the validity of my study. The chapter concluded

with the procedures I followed to alleviate any ethical concerns or issues raised by this

study.

In Chapter 4, I discuss my data collection procedures in more depth. I provide

descriptive statistics for my data set and an overview of participant demographics for my

study. My overall statistical model is presented along with the appropriate statistical data

for each independent variable. Chapter 4 ends with an overall summary of the data I

collected during my study.

 69

Chapter 4: Results

Introduction

The purpose of this quantitative, nonexperimental study was to explore student

persistence and academic performance in STEM after attending a STEM-focused middle

school from sixth through eighth grades. In this chapter, I provide details on the data

collection procedures for this study including data management and a description of the

study sample. A summary of the data results and analysis are presented followed by a

detailed discussion of the results. In revisiting the research questions for this study, I

sought to discover the following:

1. What are high school seniors’ STEM persistence in high school after a 3-year

enrollment in a STEM middle school?

2. What are high school seniors’ academic performance in STEM courses in

high school after a 3-year enrollment in a STEM middle school?

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

Data Collection

The data collected from the district’s DRA office consisted of a deidentified roster

of archived student data. The data set included student gender and ethnicity; STEM

course enrollment since entrance into high school; grades in STEM courses; and NC

Math I, NC Math III, and biology EOC performance. I used the district’s high school

planning guide to identify STEM courses to include in the analyses. All 90 high school

seniors included within the study had attended the same STEM-focused middle school in

sixth, seventh, and eighth grades and all had transferred to the base high school utilized
 70

for this study.

The study was focused on identifying the STEM persistence and STEM academic

performance of students included in the study as they transferred from a STEM-focused

middle to high school and throughout their high school years; thus, it was important to

ensure that the appropriate data were identified that would align with the meaning of

“persistence” and “academic performance.” For this study, STEM persistence pertained

to student course selections throughout high school and whether students engaged in what

would be considered as STEM-related or STEM-oriented courses. STEM-related courses

included advanced mathematics and science courses, enrollment in math and science

courses beyond what is required for high school graduation, and elective courses that are

related to science, technology, engineering, and/or mathematics (Shetay et al., 2016).

STEM academic performance pertains to how well students performed in STEM-oriented

courses which include the same types of courses aforementioned for STEM persistence.

It also pertains to student performance on EOC exams associated with STEM courses.

Preliminary Data Management

The STEM-oriented courses identified from the district’s high school planning

guide and included in this study to measure STEM persistence are outlined in Table 1.

Courses that were in the data set that were not STEM-oriented were removed. Many of

the courses identified as core courses could be considered STEM elective courses as well

if students enrolled in these courses as their STEM elective choice. Students who were

considered persistent in STEM were those who took honors or AP level mathematics and

science courses and have participated in a STEM-oriented elective from the time they

entered high school until the end of their junior year.

 71

Table 1

High School STEM-Oriented Courses

Math Science STEM electives

Honors NC Math II Intro to Meteorology Computer Art and
Honors NC Math III Honors Biology Animation
Honors NC Math IV Advanced Placement Sports Medicine
Honors Pre-Calculus Biology Animal Science
Advanced Placement Anatomy and Physiology Horticulture
Calculus Honors Chemistry Agriscience Applications
Advanced Placement Advanced Placement Comptia IT Certification
Statistics Chemistry Python Program
Advanced Placement Honors Physics Apparel and Textile
Physics I-Algebra Based Honors Earth/ Production
Environmental Science Food and Nutrition
Advanced Placement Biomedical Technology
Earth/Environmental Health Sciences
Science Construction and Masonry
Marine Ecology Carpentry and Drafting
Astronomy Cisco Network
Engineering Technology
Computer Engineering
Technology
Adobe Visual, Digital, and
Video Design
Advanced Manufacturing
Technology Engineering
and Design
Digital Design and
Animation
Game Art Design

In operationalizing STEM academic achievement for this study, students who

scored at least a B and above on their report card grades for their STEM courses and a

Level 4 and Level 5 (on a scale ranging from 2-5) on their math and science EOCs were

considered as having high academic performance in their STEM-related courses in high

school. Numerically, an A is given for final grades of 90-100 and a B is given for final

grades of 80-90 in a course. Students having a C would be considered as having average

 72

academic performance. Numerically, a C is given for final grades of 70-80 in a course.

Students scoring anything below a C are not performing well enough to meet the

standards for a course. In addition to report card grades, EOC levels of performance were

based on scale scores students received on their Math I, Math III, and biology courses.

These are the only STEM courses that require students to take an EOC assessment.

The data set was screened for inaccuracies, outliers, and missing values. The

quantitative data collected initially included data that were not utilized in this study.

Descriptive statistics were conducted to confirm that all values were within the range of

feasible values. All data points were within range, therefore no values were removed

because of inaccuracy. Throughout the regression analysis, one outlier was examined but

not removed due to its lack of significant impact on the data overall. I evaluated the data

set to confirm that no missing values that exhibited nonrandom patterns in the data set

were present. None existed. The data collected were imported and analyzed within the

Statistical Package for the Social Sciences (SPSS).

Data Collection Analysis Procedure: Standard Multiple Regression

Standard multiple regression tests were used for prediction and to determine if

each dependent or outcome variable in the study--STEM course counts, average mean

performance in STEM courses, and EOC performance in NC Math I, NC Math III, and

biology--could be predicted by the independent variables gender and ethnicity. In other

words, how much of the variation in each of the dependent variables can be explained by

the independent variables “as a whole,” but also the relative contribution of each of these

independent variables in explaining the variance. The multiple regression analysis

assisted in providing information on the accuracy of my predictions, testing how well the
 73

regression model fit my data, determining the variation in the dependent variables

explained by my independent variables, and testing my hypotheses (Laerd Statistics,

2015).

Before performing the multiple regression analyses on each of my dependent

variables, eight multiple regression assumptions were checked for each to ensure the data

met all the requirements for using multiple regression to statistically analyze the data.

The first two assumptions related to whether my study design had a continuous

dependent variable and two or more independent variables which were either continuous

or categorical. The third assumption checked included whether there was independence

of errors (residuals) or observations; in other words, the observations in my study could

not be related. To check the independence of observations, I used the linear regression

procedure in SPSS to check the Durban-Watson statistic. The Durbin-Watson statistic can

range from 0 to 4; I looked for a value of approximately 2 to indicate whether there was

correlation between residuals.

Assumptions 4 and 5, pertaining to whether there is linearity and

homoscedasticity, were met due to the fact that I created dummy variables to represent

gender and the different ethnicities using the numbers 0 and 1. The inclusion of dummy

independent variables automatically creates linearity, thus by default indicating

homoscedasticity as well (Hardy, 1993; Morgan, 2017). The sixth assumption checked

was whether the data showed multicollinearity (that the independent variables were not

highly correlated to each other) to avoid problems in determining which independent

variable contributes to any variance explained in the dependent variable. This was

checked through an inspection and interpretation of correlation coefficients using

 74

Pearson’s correlation and tolerance/VIF values.

The seventh assumption I checked was that there were no significant outliers,

high leverage points, or highly influential points that can negatively impact the regression

line. I examined case-wise diagnostics, studentized deleted residuals, and Cook’s to

check for leverage. For the final assumption, I checked to ensure all the residuals were

approximately normally distributed. This was checked using a histogram with a

superimposed normal curve and a P-P plot. After running the procedures to test that my

data met all the multiple regression assumptions for each dependent variable, I then

reported and interpreted the results of my multiple regression. The results of my

assumption tests were included in the regression findings.

Variables

This research study includes five dependent and two independent variables. The

first dependent variable in the study is student STEM persistence which relates to the

number of STEM-oriented courses in which they were enrolled throughout high school.

The second dependent variable is high school student overall mean academic

performance for STEM-oriented courses. The third dependent variable is high school

student performance on the NC Math I EOC assessment. The fourth dependent variable is

high school student performance on the NC Math III EOC assessment. The fifth

dependent variable is high school student performance on the biology EOC assessment.

My study sought to examine these variables to see what significance the high school

student attendance at a 3-year STEM-focused middle school has had on their STEM

persistence and academic performance as they entered and proceeded through high

school.
 75

Within this study, I sought to examine whether there are differences among high

school student persistence and performance based on two independent variables—student

gender and ethnicity. For this research, only male and female were used to identify

gender, and the races/ethnicities only included Hispanic, Black or African American,

White, and Other. The other subgroup consisted of a mix of American Indian, Asian, and

Multiple ethnicities. The data were limited to these student subgroups based on the small

number of students in the remaining subgroup. Only the genders and races that were

included in the data set provided were used. To examine the contribution of ethnicity in

explaining the variance within the dependent variables, the Hispanic, Black, and Other

student subgroups were compared to the White student subgroup.

Description of the Study Participants

The participant sample included 89 students who had participated in a STEM-

focused middle school during their sixth-, seventh- and eighth-grade years and transferred

to the same high school. All the participants were current high school seniors at the high

school used for this study. Of the 89 students, 38 (42%) identified as female, and 51

(58%) identified as male. In reviewing the ethnicity/race of student participants, 30 (33%)

were Hispanic, 41 (46%) were White, 13 (15%) were Black or African American, and

five (3%) were Other. Table 2 outlines the frequencies and percentages for gender and

race/ethnicity.
 76

Table 2

Frequencies and Percentages for Gender and Race/Ethnicity

Variable N %
Gender
Male 51 58
Female 38 42

Ethnicity/Race
Hispanic 30 33
White 41 46
Black 13 15
Other 5 3

STEM Persistence Data Results

This study examined the STEM persistence (measured by course enrollment) of

study participants overall and by gender and race/ethnicity to address Research Questions

1 and 3:

1. What are high school seniors’ STEM persistence in high school after a 3-year

enrollment in a STEM middle school?

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

In examining student persistence in STEM, I first had to look at the high school

course guidelines for the high school in which the students were enrolled. The high

school runs on a common block schedule, where students are required to enroll in four

courses per semester (eight courses per academic year). The data provided from the

district’s DRA office included all the courses students took during their freshman,

sophomore, and junior years; therefore, students had already completed 24 courses. At

the time the data were requested, senior year data were not available. Generally, by the
 77

time students reach the end of their junior year, they have already established a career

pathway which is evident through the courses they have taken, especially those beyond

what is required for all students; therefore, although it would have been interesting to

include senior year data, it was not necessary for this study.

Opportunities for STEM Course Enrollment

It was important to take into consideration the opportunities students have to take

STEM courses, which would impact the number of opportunities they had to develop a

STEM career pathway throughout high school. There are specific courses in which all

students are required to enroll regardless of their career pathways to fulfill high school

graduation requirements (see Appendix A), especially to meet general college entrance

requirements. Table 3 shows course requirements students must fulfill to be promoted

from one grade level to the next for the high school in which the study participants

attended. With existing course requirements, and considering students are required to

enroll in health and PE and world language courses at some point during high school,

there remain four to six opportunities for students to enroll in STEM courses each year.
 78

Table 3

High School Course Requirements (College and Career Ready)

Grade Promotion criteria Typical annual Number of STEM

course course enrollment
enrollment options
Freshman English I; two credits in the areas of 8 6
mathematics, social studies, or science;
and three additional credits
Sophomore English II; one credit in mathematics; 8 6
one credit in social studies; one credit in
science; and two additional credits
Junior English III; enrollment in a program 8 6
which, if successfully accomplished,
will result in the completion of
graduation requirements

Note. Table is adapted from the district’s high school planning guide.

In considering the number of opportunities the students had to enroll in STEM-

oriented courses, it was also important to consider the number of STEM course options

that were provided to students. Based on my review of the courses in the high school

planning guide, I identified 60 STEM-oriented courses offered at the study participants’

school for students to select throughout their high school career. This provides ample

opportunities for students to persist in taking STEM courses had they decided to follow a

STEM career pathway. Table 4 shows a breakdown of those options.

 79

Table 4

Number of High School STEM Course Options by Category

Variable N
Math options 6
Science options 12
STEM elective options 42
Total 60

Note. These are the number of STEM options available at the high school in which the

study participants attended. Other high schools in the district may offer more or fewer

options.

Study Participant STEM Course Enrollment

Once student opportunities to engage in STEM courses throughout high school

were discovered, I ran descriptive statistics from the data set to identify student trends in

enrolling in STEM courses. Using the count variables by cases command in SPSS to

create a dummy variable that included a count of all the STEM courses each student in

the data set had completed, I ran an analysis of the data to describe the data set as seen in

Table 5. The mode enrollment was five STEM courses; thus, this was the most common

number of STEM courses enrolled in by study participants. The average number of

STEM courses in which the students had enrolled was approximately four courses

(mean=3.8, SD=2.3). The skewness and kurtosis of the data are very close to 0, showing

that the data are relatively normally distributed. The number of STEM courses in which

students enrolled ranged from 0-13; thus, some students had not enrolled in any STEM

courses, while other students enrolled in as many as 13 over the course of 3 years.
 80

Table 5

Descriptive Stats for STEM Course Count Data

N
Valid 90
Missing 0
Mean 3.86
Median 4.00
Mode 5.00
Standard deviation 2.33
Skewness .62
Standard error of skewness .25
Kurtosis 1.33
Standard error of kurtosis .503
Range 13.00
Minimum .00
Maximum 13.00

After reviewing the descriptive statistics for the data, it was evident that my data

were normally distributed, and I could draw valid conclusions from my analyses of the

data. I next created a histogram of the number of STEM courses taken which shows an

approximately normal curve with only a slight positive skew (see Figure 5). There is one

outlier that does not seem to impact the data overall. The histogram shows that the

students had mostly enrolled in four or five STEM courses over the course of 3 years.
 81

Figure 5

Frequency of Number of STEM Courses Taken

STEM Course Enrollment by Gender

In examining STEM course enrollment by gender, I ran a means comparison

between students identified as males and females. Table 6 shows that with a mean of 4.7

for students identified as male and 2.7 for students identified as female, the male

students, on average, had enrolled in almost twice the number of STEM courses than

female students. Students identifying as male enrolled in two additional STEM courses

on average than students identifying as female. In addition, the STEM count data for

male students had a standard deviation of 2.42, whereas the STEM count data for female

students had a standard deviation of 1.62; thus, the larger standard deviation among male

students shows greater variance in the number of STEM courses taken for students
 82

identifying as male than for students identifying as female.

Table 6

STEM Course Enrollment Means Comparison by Gender

Gender Mean N Standard deviation

Male 4.71 52 2.43
Female 2.71 38 1.63
Total 3.87 90 2.34

Figure 6 shows the distribution of the STEM course count data for the male and

female study participants. The number 74 with the circle icon next to it is interpreted in

SPSS as a mild outlier. There were no extreme outliers in this data set. The female

median on the box plot has a value of 2.5 which is half of the median value of 5.0 shown

for males, showing that the upper half of course counts were higher for males than

females. The interquartile range of the STEM course counts was the same for both males

and females with a value of 3. The upper quartile range of STEM course counts for the

males, with six counts in that range, were greater than those of females as well with four

counts. Visually, the box plot for the males is higher on the chart than for females. In

addition, the whisker lengths appear to be half as short for females as they are for males,

showing less variation in STEM course selections for females than for males which seem

to show a greater variation in the upper quartile for females than males. There is a greater

variation as well in the upper quartile for females.

 83

Figure 6

STEM Course Count by Gender Box Plot

In further breaking down the STEM course data to compare male and female

STEM course counts, a bar chart (see Figure 7) was created to show the distributions of

STEM course counts between male and female students. The total number of STEM

courses is provided on the x-axis with the percentage of male or female students on the y-

axis. From this data set, the number of STEM courses most females are taking is smaller

than the number of courses for male students. The larger the number of STEM courses

taken over the 3-year period the lower the percentage of female students. While this

pattern is also seen among students identifying as male, it is more pronounced among

students identifying as female. The highest number of courses taken by female students in

this sample was three, whereas for male students the highest number was 13.
 84

Figure 7

Number of STEM Courses Taken by Gender

STEM Course Enrollment by Race/Ethnicity

Next, I completed a means comparison for race/ethnicity. Table 7 shows that with

a mean of 4.40, students in the Other subgroup had a higher average mean than the Black

or African American subgroup with a mean of 3.07 the Hispanic subgroup with a mean of

3.83 and the White subgroup with a mean of 4.07. The Other subgroup took an average

of one more STEM course counts than the Black or African American subgroup, a little

less than one more STEM course than the Hispanic subgroup, and approximately the

same number of STEM courses as the White STEM subgroup.

In comparing the standard deviations, the Black or African American subgroup

had a standard deviation of 2.02, the Hispanic subgroup had 2.04, the Other subgroup

had 3.58, and the White subgroup had 2.49. The Other subgroup data, having a higher
 85

standard deviation, show that the Other subgroup STEM course count data are more

spread out or dispersed than all other subgroups. This could be because of the small

number of students included in the subgroup.

Table 7

STEM Count Means Comparison by Race/Ethnicity

Race/ethnicity Mean N Standard deviation

Black or African American 3.08 13 2.02
Hispanic 3.83 30 2.04
Other 4.40 5 3.58
White 4.07 42 2.49
Total 3.87 90 2.34

Figure 8 shows the distribution of the STEM course count data for each subgroup

of students. The number 74 above the White subgroup, again, is interpreted in SPSS as a

mild outlier. There were no extreme outliers in this data set that would significantly skew

the data. The Black or African American subgroup median on the box plot has a value of

3.0, the Hispanic subgroup a median of 4.0, the Other subgroup a median of 4.0, and the

White subgroup a median of 4.0. The Black or African American student subgroup had a

median one-quarter lower than those of other student subgroups; thus, the upper half of

course counts were lower for this subgroup than for all other student subgroups.

The interquartile range of the STEM course counts was 4 for Black or African

American, 3.25 for Hispanic, 7 for Other, and 3.25 for White. The Other subgroup could

be more significant in comparison due to the much smaller number of student participants

in the group. Visually, the box plots for the Other and White subgroups are higher on the

chart than for the Black or African American and Hispanic subgroups. In addition, the

whisker lengths appear to be much larger for the White subgroup showing a greater
 86

variation in the upper quartile than lower, and the whisker length much shorter for the

Black or African American student subgroup showing a much smaller variation of STEM

course count.

Figure 8

STEM Course Enrollment by Race/Ethnicity Box Plot

In further breaking down the STEM course data to compare STEM course counts

by race and ethnicity, a clustered bar chart (see Figure 9) was created to show the

distributions of STEM course counts between the different subgroups. The total number

of STEM courses is provided on the x-axis with the percentage of students from each

subgroup on the y-axis. It is important to keep in mind that the Other subgroup has a

much smaller number of students than the other ethnic subgroups in the sample. In

comparing student ethnicity, the White subgroup included the highest count of STEM

course enrollments, with one student taking nine courses and another student enrolled in
 87

13 courses. All subgroups included students who did not take a single STEM course

across the span of their 3 years in high school. As the number of STEM courses in which

the students could have enrolled increased, the percentage of students from each

subgroup became lower (the percentage of students from the Other subgroup is much

smaller than what it appears on the chart due to the small size of the subgroup).

Figure 9

STEM Course Count by Race/Ethnicity

Multiple Regression Analysis for STEM Persistence

Assumptions Checks. Using SPSS, I ran a multiple regression to predict

VO2max (persistence) from gender and ethnicity. Dependent variables for student STEM

persistence included the overall STEM course count over a 3-year period. This part of the

study also contained two categorical independent variables—gender and ethnicity.

Ethnicity was entered into the model as three dichotomous variables (Hispanic, Black or
 88

African American, and Other) and these dummy-coded variables were compared to the

White student subgroup. There was independence of residuals, as assessed by the Durbin-

Watson statistic of 1.96. In checking for linearity, because dummy variables had to be

created from the data in order to make race/ethnicity comparisons throughout the study,

dummy variables do not need checks for linearity (Hardy, 1993). They automatically

meet the assumption of linearity by definition, because they create two data points and

two points define a straight line (Hardy, 1993; Morgan, 2017); therefore, this would meet

the assumption for homoscedasticity as well.

There was no evidence of multicollinearity, as assessed by tolerance values,

studentized deleted residuals, leverage values, and Cook's distance. In reviewing the

independent variables in Table 8, none of the independent variables are above 0.7,

showing no correlations between the independent values. In Table 9, all the tolerance

values are greater than .1 and VIFs less than 10; therefore, I was fairly confident that I

did not have a problem with collinearity.

 89

Table 8

STEM Assumption Check for Multicollinearity: Correlation Coefficients

STEM Gender Hispanic Black or Other

course African
count American
Pearson STEM course count 1.00 -.43 -.01 -.14 .06
correlation
Gender -.43 1.00 -.03 .16 .09

Hispanic -.01 -.03 1.00 -.29 -.17

Black or African -.14 .16 -.29 1.00 -.10

American

Other .06 .08 -.17 -.10 1.00

Table 9

STEM Count Assumption Check for Multicollinearity: Tolerance/VIF

Tolerance VIF
(Constant)
Gender2 .96 1.04
Hispanic .87 1.15
Black or African American .87 1.16
Other .94 1.07

There was one case where the standard residual was slightly above the 3.0

threshold for residuals with a value of 3.8 (see Table 10). After a review of this case, I

determined that the student’s scores used to create the overall STEM course count

variable were valid data and thus should remain in the study. Additionally, I reran the

assumption without the participant’s data and there was no significant change in the

results showing that although that student’s data was an outlier, it did not significantly

impact the data results. In reviewing Cook’s distance and leverage, there were no values

for Cook’s above 1. There was also a returned maximum value of .206 for leverage
 90

which was within the leverage threshold (see Table 11).

Table 10

Assumptions Check for Residuals for STEMCount

Case number Standard residual STEMCount Predicted value Residual

79 3.83 13.00 4.78 8.22

Note. Dependent Variable: STEMCount.

Table 11

Cook’s Distance and Leverage for STEM Count

Minimum Maximum Mean Standard N

deviation
Cook's distance .00 .19 .01 .03 90
Centered leverage value .02 .21 .04 .04 90

Note. Dependent Variable: STEMCount

The P-P Plot depicted in Figure 10 shows that the assumption of normality was

met. The data within the P-P Plot is aligned (although not perfectly) along the diagonal

line, indicating that the residuals are close enough to normal for the analysis to proceed;

therefore, no transformations were needed.

 91

Figure 10

STEM Count P-P Plot Assumption for Normality

Multiple Regression Results. The multiple regression model statistically significantly

predicted STEMCountALL, F(4, 89) = 5.117, p < .001, adj. R2 = .156. R2 for the overall

model was 19.4% with an adjusted R2 of 15.6%, a small size effect according to Cohen

(1988). All four variables added statistically significantly to the prediction, p < .05.

Regression coefficients and standard errors can be found in Table 12.

 92

Table 12

STEM Count Multiple Regression Results

STEMCountALL B 95% CI for B SE B  R2 R2

LL UL
Model .20 .16***
Constant 4.78*** 4.40 5.517 0.371
Gender -1.98* -2.91 -1.057 .47 -.42**
Hispanic -.15*** -1.18 .87 .51 -.03***
Black or African American -.48*** -1.86 .89 .69 -.07***
Other .81*** -1.22 2.843 1.022 .08***

Note. Model=“Enter” method in SPSS Statistics; B=unstandardized regression

coefficient; CI=confidence interval; LL=lower limit; UL=upper limit; SE B=standard

error of the coefficient; =standardized coefficient; R2=coefficient of determination;

R2=adjusted R2. *p<.05, **p<.01, ***p<.001.

STEM Academic Performance Data Results

The next set of results of this research study focus on providing data to answer the

second and third research questions:

2. What are high school seniors’ academic performance in STEM courses in

high school after a 3-year enrollment in a STEM middle school?

3. How do high school seniors’ STEM persistence and academic performance in

STEM courses compare by gender and ethnicity?

STEM Course Performance Data

In reviewing the course performance data, it was neither feasible nor helpful to

analyze each student’s individual scores for all STEM courses. Instead, the average mean

performance for the clusters of math, science, and STEM elective courses students

completed was computed by using the mean function available under the calculate

variable action in SPSS. After the means for each cluster of courses was computed, the
 93

overall mean of student grades on all courses was computed. I then ran a means analysis

on each which provided descriptives on the mean course grades as seen in Table 13. All

the means for each category of STEM courses fell into the category of high academic

performance, as they were all a mean score of 80 or above.

As aforementioned in the study, any scores between 80-90 are considered a B

score and are considered to be a high-performance rating. Although high performing,

they are still closer to a C average, which is considered average performance. In addition,

not all the students in the data set took what is considered a STEM-oriented math course.

Only about half of the study participants have enrolled in advanced math courses. In

reviewing the standard deviations, the average standard deviation for grades was high for

math, science, and STEM elective courses, as well as all course grades together.

Although the mean performance for all areas falls in the high-performance range, it is

evident through the minimum and maximum values that not all grades have been high

performing, with some classes having grades ranging between 50 to 75 points.

Table 13

STEM Course Performance Means by Subject Area

Math Science STEM electives All courses

Mean 81.57 81.19 81.88 80.68
N 42 90 85 90
Standard deviation 10.97 10.44 16.77 12.73
Minimum 43.00 42.00 24.50 32.60
Maximum 93.50 96.00 100.00 95.00
Range 50.50 54.00 75.50 62.40
Kurtosis 2.58 1.22 2.73 2.50
Skewness -1.52 -1.05 -1.700 -1.56
 94

NC EOC Test Performance

In addition to examining student performance in STEM-oriented math, science,

and elective classes, student STEM academic achievement was examined through their

performance in any STEM classes with a state-required EOC assessment. To test their

mastery of subject-related concepts, high school students in North Carolina are only

required to take EOC assessments for NC Math I, NC Math III, and biology for the

STEM courses included in this study. These assessments count towards the students’

final grades in the course in which the assessment corresponds. It represents a weight of

20% of the final grade for the course (North Carolina Department of Public Instruction,

2022).

In reviewing study participant performance, I first had to convert the scale scores

that were provided for each assessment into the corresponding levels of performance.

Table 14 shows the breakdown of scale scores which have been the same since the

students have enrolled in high school; therefore, all the EOC scores were accurate

regardless of whether they completed the assessment in their freshman, sophomore, or

junior years. Students were required to score a Level 3 to show they have mastered

course content, and anything below that was considered as non-mastery or not proficient.

Level 1 and Level 2 scores are no longer provided, and students scoring in that range

were considered low performing on that EOC. Students who scored a Level 3, although

considered proficient, were considered as having average performance on the assessment,

and students scoring a Level 4 and 5 were considered as high performing on the

assessment and were also considered as college and career ready (North Carolina

Department of Public Instruction, 2022).

 95

Table 14

NC Required High School EOC Test Scales and Performance Levels

Description Not proficient Level 3 Level 4 Level 5

Math I Less than or equal to 548-554 555-562 Greater than or
547 equal to 563
Math III Less than or equal to 550-555 556-562 Greater than or
549 equal to 563
Biology Less than or equal to 250-251 252-260 Greater than or
249 equal to 261

NC Math I EOC Performance. All NC high school students, regardless of what

level of the NC Math I course they take (college prep, honors, or advanced placement),

are required to take the NC Math I EOC assessment. For the NC Math I performance,

Table 15 shows the number of students from the sample who did not score proficiently

and scored either a Level 3, Level 4, or Level 5. It is evident that none of the students

scored a Level 5 as that category of performance was not a part of the table output from

SPSS. Only two of the students scored a Level 4, which is considered high performing;

19 scored a Level 3, which is average performing; and the remaining 69, the majority of

the sample, did not score proficient on the assessment.

In addition, the median and mode were reported as well. The median score for

student performance on the NC Math I EOC was 0.00 and the mode was 0 which is

equivalent to not proficient. These results were not positive as these numbers fall within

the range of what would be considered as low performing. Overall, the majority of

students who completed the assessment were in the low-performing range. The mean was

not included as there are technically no Level 1 or Level 2 score values provided for EOC

scores as there were historically. All students who did not score proficient were assigned
 96

a number 0 in SPSS to allow a valid mean to be calculated.

Table 15

NC Math I EOC Performance Level Counts

EOC performance level N

Not proficient 69
3 19
4 2
5 0
Valid 90
Missing 0
Median 0
Mode 0

In addition to the frequency, median, and mode for EOC level, Figure 11 displays

the percentage of students scoring at each performance level. The majority of the students

did not perform proficiently on the NC Math I EOC assessment. Approximately three

quarters (76.6%) of all student participants were considered low performing on the

assessment. Additionally, given the EOC scores account for 20% of student course

grades, the low performance on the assessment also had an impact on their overall course

grade for NC Math I. Only 21.1% of the students overall had an average performance

with a Level 3 and an even smaller 2.2% of students were considered high performing

with a Level 4.
 97

Figure 11

NC Math I EOC Performance Levels

NC Math III EOC Performance. All NC high school students, regardless of

what level of the NC Math III course they take (college prep, honors, or advanced

placement), are required to take the NC Math III EOC assessment in addition to the NC

Math I EOC. For the NC Math III, Table 16 shows the number of students for each EOC

level (not proficient, Level 3, Level 4, and Level 5). Thirty-three of the 67 students from

the sample who took the NC Math III EOC did not score proficient and were considered

low performing. Seven students scored a Level 3 and were considered as average

performing. Seventeen students scored a Level 4, and 10 students scored a Level 5 and

thus were considered as high performing on this assessment. Twenty-three students did

not complete the assessment and therefore had no scores included in the results.

In addition, the median and mode were reported. The median score for student
 98

performance on the NC Math I EOC was 3.0 and the mode was 0, which is equivalent to

not proficient. The median results fall within the range of what would be considered as

average performance and the mode of what would be considered as low performing.

Overall, the majority of students who completed the assessment were in the low-

performing range. The mean was not included as there are technically no Level 1 or

Level 2 score values provided for EOC scores as there were historically.

Table 16

NC Math III EOC Performance Level Counts

EOC performance level N

Not proficient 33
3 7
4 17
5 10
Missing from system 23
Valid 67
Median 3.00
Mode 0

In addition to the frequency, median, and mode for EOC level, Figure 12 displays

the percentage of students scoring at each performance level. The majority of the students

did not perform proficiently on the NC Math III EOC assessment. Approximately half

(49.3%) of all student participants were considered as low performing on the assessment.

Additionally, given the EOC scores account for 20% of student course grades, the low

performance had an impact on their overall course grade for NC Math I: 10.4% of the

students overall had an average performance with a Level 3; 25.4% of the students scored

a Level 4; and 14.9% of the students scored a Level 5. More students were high

performing on the NC Math III assessment than the NC Math I.

 99

Figure 12

NC Math III EOC Performance Levels

Biology EOC Performance. As is the case for NC Math I and NC Math III, all

NC high school students, regardless of what level of the NC Math III course they take

(college prep, honors, or advanced placement), are required to take the biology EOC

assessment. For the biology EOC assessment, Table 17 shows the number of students by

proficiency level (not Proficient, Level 3, Level 4, and Level 5). Six of the 46 students

from the sample who took the NC Math III EOC did not score proficient and were

considered low performing. Two students scored a Level 3 and were considered as

average performing. Twenty-one students scored a Level 4, and 17 students scored a

Level 5 and were considered as high performing on this assessment. Forty-four students

did not complete the assessment and therefore had no scores included in the results.

In addition, the median and mode were reported. The median score for student
 100

performance on the biology EOC was 4.00 and the mode was a Level 4. This was

positive, as these numbers fall within the range of what would be considered as high

performing. Overall, the majority of students who completed the assessment were in the

high-performing range. The mean was not included, as there are technically no Level 1 or

Level 2 score values provided for EOC scores as there were historically.

Table 17

Biology EOC Performance Frequency, Mean, and Median

EOC performance level N

Not proficient 6
3 2
4 21
5 17
Valid 46
Missing 44
Median 4.00
Mode 4

In addition to the frequency count, median, and mode for the biology EOC level,

Figure 13 displays the percentage of students scoring at each performance level. The

majority (82.7%) of the students performed proficiently on the biology EOC: 13% of all

student participants were considered as low performing on the assessment, and such a

low performance had an impact on their overall course grade as well for biology; 4.3% of

the students overall had an average performance with a Level 3; 45.7% of the students

scored a Level 4; and 37% of the students scored a Level 5. It is evident through a

comparison of the EOC scores, that the students performed significantly higher on the

biology assessment than their math assessment.

 101

Figure 13

Biology EOC Performance Levels

STEM Course Performance by Gender

In addition to examining student STEM academic performance overall, the data

were further broken down to compare student performance in courses and on the EOCs

by gender and ethnicity. As was reported in the problem and literature review of this

research study, there are underrepresented subgroups including female, Black or African

American, and Hispanic students in STEM fields of study, and the STEM workforce and

STEM efforts across the U.S. seek to address this issue through the education system.

The data review in the upcoming sections examines student performance of these

subgroups compared to male and White subgroups that are not underrepresented in

STEM. First, I go back and review student performance in math, science, STEM

electives, and all courses combined but include comparisons between the subgroups.
 102

Table 18 provides descriptive and comparative data on overall mean performance

between male and female student subgroups. For students identifying as male, the overall

mean performance in math was a 79, science an 80, STEM electives an 81, and for all

STEM courses combined a 79.8 which can be rounded to an 80. In comparing the female

performance to the males, the mean performance for math is approximately 6 points

higher than that of males on average, 3 points higher for science performance on average,

2 points lower on STEM electives performance on average, and 2 points higher overall

for all STEM courses on average. The standard deviation of the scores on average was 3

points higher for males than females for math, 1 point higher for science for males than

females, 2 points higher for STEM electives for males than females, and approximately

the same for males and females for STEM courses overall.

Table 18

STEM Course Mean Comparison by Gender

Gender Math Science STEM All

elective courses
Male Mean 79.02 80.09 81.06 79.89
N 23 52 50 52
Standard deviation 12.17 10.86 16.08 12.52

Female Mean 84.68 82.69 83.06 81.78

N 19 38 35 38
Standard deviation 8.63 9.78 17.87 13.11

Math Mean Performance by Gender. In further examining male versus female

student performance, a mean performance gender comparison was completed as well

using a box and whisker plot. Figure 14 shows the distribution of the STEM math mean

course performance data for each subgroup of students. On the chart, three outliers are

identified, one with a score of 59 for males and 1 for females which is considered a mild
 103

outlier as indicated by the O icon next to the number, and the final outlier an average

score of 34 which is considered possibly significant as indicated by the star icon next to

the number. There were no extreme outliers in this data set. A multiple regression was

performed to identify the impact any outliers had on my data set and is discussed in a

later section of this paper.

The gender median on the box plot has a value of 87 for female students and a

value of 82 for male students, showing the median value to be 5 points higher for the

female subgroup overall. Visually, the box plot for the female subgroup is higher for the

females with a small box size and an almost equal amount of variation between the

interquartile ranges with most of the scores falling equally above and below the median

value. The male subgroup appears to have more variation in their scores with the majority

of the scores falling below the median value with a greater amount of their scores falling

in the average to low performing range. In addition, the whisker lengths appear to be

much longer for lower scores for the male subgroup than the female subgroup whose

whiskers are much shorter and along a much higher range of scores.
 104

Figure 14

Math Mean Performance Comparison by Gender

NC Math I EOC Performance by Gender. In addition to comparing male and

female performance in math courses, I also compared their NC Math I EOC performance.

Figure 15 displays the percentage of students by gender for each proficiency level on the

NC Math I EOC. The data displayed within Figure 15 show that 75% of males compared

to 79% of females did not score proficiently on the NC Math I EOC. This includes the

scores of 52 male and 38 female students who took the NC Math I EOC, thus 39 male

students were not proficient compared to 29 female students. In reviewing Level 3

performance scores, 21% of males compared to 21% of females scored a Level 3 on this

assessment. While the percentage is approximately equal, it represents, eight females and

11 males, which are close in number meaning slightly more males performed on a Level

3 than females. For Level 4, 3.8% (n=2) of males compared to 0% (n=0) of females
 105

performed on this level. Overall, 25% of males were high performing on the NC Math I

EOC (13 males) as compared to 21% of females (eight females). The data between the

males and females on the NC Math I EOC do not seem to be too far apart, although more

males scored high performing than females.

Figure 15

NC Math I EOC Performance by Gender

NC Math III EOC Performance by Gender. Next, a comparison of male and

female student performance on the NC Math III EOC was performed. There were 39

male students who took the assessment compared to 29 female students. Figure 16

displays the percentage of students at each proficiency level on the NC Math III EOC by

gender. The data displayed by the histogram first show that 51% of male students

compared to 46.4% of female students did not score proficiently, meaning 20 males were

not proficient compared to 13 females. In reviewing Level 3 performance scores, 12.8%

 106

of males compared to 7.1% of females scored a Level 3 on this assessment representing

five males and two females. For Level 4, 25.6% (n=10) of male students compared to

25% (n=7) of females performed on this level. Overall, 35.9% of male students were high

performing on the NC Math III EOC (n=14) as compared to 46.4% of female students

(n=13). The data between the males and females on the NC Math III EOC do not seem to

be far apart.

Figure 16

NC Math III EOC Performance by Gender

Science Mean Performance by Gender. In further examining male versus

female performance, the science mean performance comparison was completed as well

using a box and whisker plot. Figure 17 shows the distribution of the STEM science

performance data for the male and female subgroups. On the plot, there is one outlier, a

mean score of 39, which is considered a mild outlier as indicated by the O icon next to
 107

the number (an SPSS output signal). There were no other outliers in this data set. Later in

this analysis, a multiple regression was performed to identify the impact any outliers had

on my data set.

The gender median on the box plot has a value of 86 for females and the males a

value of 82, showing the median value to be 4 points higher for the female subgroup

overall. Visually, the box plot for the female subgroup is almost an equal height with the

whisker, median, interquartile ranges, and lower whiskers all being slightly higher for the

females than males. The male subgroup appears to have more variation in their scores

with the majority of the scores falling below the median value with a greater amount of

their scores falling in the average to low-performing range. The upper whiskers for both

the male and female subgroups appear to be very similar in regard to the higher value of

scores.
 108

Figure 17

Science Mean Performance by Gender

Biology EOC Performance by Gender. Next, a comparison of male and female

performance on the biology EOC was performed. There were 21 males who took the

assessment compared to 25 females. Figure 18 displays the percentage of students

between the two subgroups who either did not score proficiently or scored a Level 3, 4, or

5 on the NC Math III EOC. The data displayed by the histogram first show that 19% of

males compared to 8% of females did not score proficiently, meaning four males were

not proficient compared to two females. In reviewing Level 3 performance scores, 0% of

males compared to 8% of females scored a Level 3 on this assessment, representing zero

males and two females. For Level 4, 53% of males compared to 40% of females which

equals 11 males compared to 10 females performed on this level. For Level 5, 29% of

males compared to 44% of females, which is six males and 11 females, almost double the
 109

number of males, received this highest performance rating. Overall, 81% of males were

high performing on the NC Math III EOC (17 males) as compared to 84% of females (21

females). The data between the males and females on the NC Math III EOC do not seem

to be far apart either compared to NC Math I and NC Math III performance.

Figure 18

Biology EOC Performance by Gender

STEM Electives Mean Performance by Gender. Next, a STEM elective mean

course performance gender comparison was completed as well using a box and whisker

plot. Figure 19 shows the distribution of the STEM elective mean course performance

data for each subgroup of students. On the chart, three outliers are identified for the male

and female subgroups. For the male subgroup, the mean outlier scores were 43, 59, and

86, with 86 being identified as a mild outlier and 43 and 59 as more significant. For the

female subgroups, the mean outlier scores were 3, 16, and 33, the 3 indicated as a
 110

significant outlier and the 16 and 33 as mild outliers. There were no extreme outliers in

this data set. Later in this analysis, a multiple regression was performed to identify the

impact any outliers had on my data set.

The gender median on the box plot has a value of 90 for females and the males a

value of 87, showing the median value to be 3 points higher on average for the female

subgroup overall. Visually, the box plot for the female subgroup is slightly higher than

the male subgroup, with slightly more variation in the scores below the median than

above and a large number of students falling within the fourth quartile. The male

subgroup appears to have more variation in their scores, with the majority of the scores

falling below the median value with a greater amount of their scores falling in the

average to low performing range.

 111

Figure 19

STEM Electives Mean Performance by Gender

Overall STEM Course Performance by Gender. In further examining male

versus female performance, a mean performance gender comparison was completed as

well using a box and whisker plot. Figure 20 shows the distribution of the STEM math

mean course performance data for each subgroup of students. On the chart, there are four

possible outliers shown. One possible outlier has a score of 43 for males and two others

have scores of 9 and 16 for females. These are considered as potential mild outliers as

indicated by the O icon next to the number. The final possible outlier has an average

score of 3, which is considered possibly significant as indicated by the star icon next to

the number. There were no extreme outliers in this data set. Later in this analysis, a

multiple regression was performed to identify the impact any outliers had on my data set.

The gender median on the box plot has a value of 86 for females and the males a
 112

value of 84, showing the median value to be 2 points higher for the female subgroup

overall. Visually, the box plot for the female subgroup is slightly higher for the females

with a small box size and a greater variation of scores falling below the median. The male

subgroup appears to have more variation in their scores with the majority of the scores

falling below the median value with a greater amount of their scores falling in the

average to low performing range. In addition, the whisker lengths appear to be much

longer for lower scores for the male subgroup than the female subgroup whose whiskers

are much shorter and along a much higher range of scores.

Figure 20

Overall STEM Mean Course Performance by Gender

STEM Course Performance by Race/Ethnicity

In further examining student STEM academic performance overall and by gender,

the data were further broken down to compare student performance in courses and on the
 113

EOCs by ethnicity. As was reported in the problem and literature review of this research

study, there are underrepresented subgroups including females, Black or African

American, and Hispanic students in STEM fields of study, and the STEM workforce and

STEM efforts across the U.S. seek to address this issue through the educational system.

The data review in the upcoming sections examines student performance in the Black or

African American, Hispanic, and Other subgroups (i.e., underrepresented subgroups in

STEM) as compared to the White subgroup. First, I provide a review of student mean

performance in math, science, STEM electives, and all courses combined and make

comparisons between the underrepresented and White subgroups. Table 19 provides

descriptive and comparative data on overall mean performance for Black or African

American, Hispanic, White, and Other subgroups. The Black or African American

subgroup had a math mean performance of 88, the Hispanic subgroup a 75, the other

subgroup an 87, and the White subgroup an 82. The Black or African American subgroup

scored 6 points higher, on average, than the White subgroup, but it has to be taken into

consideration that the Black or African American subgroup only has two students

compared to 27 students in the White subgroup. The Hispanic subgroup performed

approximately 7 points lower on average than the White subgroup but also with a much

smaller number of students—only nine—compared to 27 in the White subgroup. The

Other subgroup, with five or less students, scored 5 points more, on average, than the

White subgroup.

In comparing the science mean performance, the Black or African American

student subgroup, which consisted of 13 students who participated in advanced science

courses in high school, had a mean score of 77; the Hispanic subgroup with 30 students
 114

had a mean score of 78; the Other subgroup with five students had a mean score of 86;

and the White subgroup with 42 students had a mean score of 84. The Black or African

American subgroup scored 7 points lower; the Hispanic subgroup 6 points lower, on

average; and the Other subgroup 2 points higher, on average than the White subgroup.

In comparing the STEM elective courses mean performances, the Black or

African American subgroup with 12 students who participated in one or more STEM

courses had a mean performance grade of 78; the Hispanic subgroup, with 29 students,

had a mean performance grade of 74; the Other subgroup with four students, had a mean

performance grade of 90; and the White subgroup, with 40 students, had a mean

performance grade of 87. The Black or African American subgroup’s mean performance

was 9 points lower, the Hispanic subgroup 12 points lower, and the Other subgroup 3

points higher than the White subgroup.

Finally, in comparing the STEM mean performance for all STEM courses, the

Black or African American subgroup, with 13 students, had an overall mean of 79. The

Hispanic group, with 30 students, had an overall mean of 74. The Other subgroup, with

five students, had a mean of 88; and the White subgroup, with 42 students, had a mean of

85. The Black or African American subgroup’s mean score was 6 points lower, the

Hispanic subgroup’s 9 points lower, and the Other subgroup’s 3 points higher than the

White subgroup.
 115

Table 19

STEM Mean Course Performance by Race/Ethnicity

Ethnicity Math Science STEM electives STEM mean

mean mean mean overall
Black or Mean 88.00 77.10 78.28 78.55
African N 2 13 12 13
American Standard deviation 1.41 9.42 15.69 8.69

Hispanic Mean 75.01 77.81 74.66 74.34

N 9 30 29 30
Standard Deviation 15.03 10.30 21.54 16.42

Other Mean 87.08 86.46 90.15 87.98

N 4 5 4 5
Standard Deviation 4.45 7.92 7.39 5.36

White Mean 82.47 84.24 87.37 85.01

N 27 42 40 42
Standard Deviation 9.70 10.13 10.85 8.94

Math Mean Performance by Race/Ethnicity. In further breaking down the math

mean performance by race and ethnicity, a box and whisker plot was created (see Figure

21). From this data set, it is important to keep in mind that the Other subgroup has a

much smaller number of students than the other subgroups in the data set. First, the

median scores were observed for each subgroup. On the plots, the medians appeared to be

very close for all subgroups. The Black or African American subgroup had a median of

88, the Hispanic subgroup had a median of 82, the Other subgroup an 87, and the White

Subgroup an 85.

In comparing variance, it appears on the plot that the African American subgroup

has a much smaller variance of mean scores around the median than all the other

subgroups. The Hispanic and White subgroups appear to have a greater variance of scores

below the median, although the variance seems greater for the White subgroup than the
 116

Hispanic subgroup due to the long whiskers on the lower side of the plot. All the top

scores for each plot appear to have the same value for each subgroup and appear to be

much lower for the Hispanic and White subgroups than the Black or African American

and Other subgroups. Only one group had a potential outlier, the Hispanic subgroup, that

was labeled a mild outlier with a score of 59. I ran multiple regressions, which are

presented later in the study, to determine if outliers existed that impacted the data.

Figure 21

Math Mean Performance by Race/Ethnicity

NC Math EOC I Performance by Race/Ethnicity. In addition to comparing

race/ethnicity mean performances in math courses, I also compared their NC Math I EOC

performance. Figure 22 displays the percentage of students by gender for each

proficiency level. Thirteen Black or African American, 30 Hispanic, five Other, and 42

White students took the NC Math I EOC. The data displayed within Figure 22 show that

85% of Black or African American, 83% of Hispanic, 100% of Other, and 67% of White
 117

students did not score proficiently on the NC Math I EOC. This represents 11 Black or

African American, 25 Hispanic, five Other, and 28 White students. In reviewing Level 3

performance scores, 15% of Black or African American, 17% of Hispanic, 0% of Other,

and 29% of White students scored a Level 3 on this assessment. This represents two

Black or African American, five Hispanic, zero Other, and 12 White students. For Level

4, 0% of Black or African American, 0% of Hispanic, 0% of Other, and 5% of Whites

scored a Level 4. Only the White subgroup, with two students, had students who scored a

Level 4 on the NC Math I EOC. None of the students in all subgroups performed at a

Level 5. Overall, 0% of the Black or African American, Hispanic, and Other subgroups

and 5% of the White subgroup were high performing on the NC Math I EOC, meaning

only the White subgroup, with only two students, had students in this category. The

underrepresented subgroups were slightly lower performing than the White subgroup.
 118

Figure 22

NC Math I EOC Performance by Race/Ethnicity

NC Math III EOC Performance by Race/Ethnicity. In addition to comparing

race/ethnicity mean performances on the NC Math I EOC, the NC Math III EOC

performances were compared as well. Figure 23 displays the percentage of students at

each proficiency level on the NC Math III EOC by gender. Ten Black or African

American, 21 Hispanic, four Other, and 32 White students took the NC Math III EOC.

The data displayed by the histogram first show that 70% of Black or African American,

76% of Hispanic, 25% of Other, and 28% of White students did not score proficiently on

the NC Math I EOC. This represents seven Black or African American, 16 Hispanic, one

Other, and nine White students. In reviewing Level 3 performance scores, 20% of Black

or African American, 0% of Hispanic, 0% of Other, and 16% of White students scored a

Level 3 on this assessment. This represents two Black or African American, zero
 119

Hispanic, zero Other, and give White students. For Level 4, 0% of Black or African

American, 24% of Hispanic, 50% of Other, and 31% of White students scored a Level 4.

This represents zero Black or African American, five Hispanic, three Other, and 10 White

students. For Level 5, 10% of Black or African American, 0% of Hispanic, 25% of Other,

and 25% of White students performed at this level. This represents one Black, zero

Hispanics, one Other, and eight White. Overall, 10% of the Black or African American,

24% of Hispanic, 75% of Other, and 56% of the White subgroups were high performing

on the NC Math I EOC. This represents one Black or African American, five Hispanic,

four Other, and 18 White students. The majority of the students in the underrepresented

subgroups were lower performing on this assessment than White students. There were 17

more White students than Black or African American students and four times as many

White students who were high performing than Hispanic and Other students.
 120

Figure 23

NC Math III EOC Performance by Race/Ethnicity

Science Mean Performance by Race/Ethnicity. In further breaking down the

Science mean performance by race and ethnicity, a box and whisker plot was created (see

Figure 24). From this data set, it is important to keep in mind that the Other subgroup has

a much smaller number of students than the other subgroups in the data set. First, the

median scores were observed for each subgroup. The Black or African American

subgroup had, on average, a median of 77; the Hispanic subgroup a 76; the Other

subgroup a 91; and the White Subgroup an 87. The median was 10 points higher for the

African American subgroup, 9 points higher for the Hispanic subgroup, and 4 points

higher for the Other subgroup than the White subgroup.

In comparing variance, it appears on the plot that the Other subgroup has a much

smaller variance of mean scores around the median than all the other subgroups. The
 121

Black or African American and Hispanic subgroups appear to have a greater variance of

scores below the median. The variance along the bottom of the plots appears to be greater

for the lower scores for Blacks than Hispanics. Only one group had potential outliers, the

White subgroup, one with a value of 85 and labeled with the O icon as a potential mild

outlier, and another with a value of 39 labeled with a star icon as a potential extreme

outlier. Multiple regressions were run later in the study that assisted in determining

whether or not outliers existed that impacted the data.

Figure 24

Science Mean Performance by Race/Ethnicity Box Plot

Biology Performance by Race/Ethnicity. Next, a comparison of race/ethnicity

mean performances on the biology EOC was done. Figure 25 displays the percentage of

students between the subgroups who either did not score proficiently or scored a Level 3,

4, or 5 on the NC Math I EOC. Seven Black or African American, 11 Hispanic, four

Other, and 24 White students took the biology EOC. The data displayed by the histogram
 122

first show that 29% of Black or African American, 76% of Hispanic, 25% of Other, and

28% of White students did not score proficiently on the NC Math I EOC. This represents

seven Black or African American, 16 Hispanic, one Other, and nine White students. In

reviewing Level 3 performance scores, 20% of Black or African American, 0% of

Hispanic, 0% of Other, and 16% of White students scored a Level 3 on this assessment.

This represents two Black or African American, zero Hispanic, zero Other, and five

White students. For Level 4, 0% of Black or African American, 24% of Hispanic, 50% of

Other, and 31% of White students scored a Level 4. This represents zero Black or African

American, five Hispanic, three Other, and 10 White students. For Level 5, 10% of Black

or African American, 0% of Hispanic, 25% of Other, and 25% of White students

performed at this level. This represents one Black, zero Hispanics, one Other, and eight

White students. Overall, 10% of the Black or African American, 24% of Hispanic, 75%

of Other, and 56% of the White subgroups were high performing on the NC Math I EOC.

This represents one Black or African American, five Hispanic, four Other, and 18 White

students. The majority of the students in the underrepresented subgroups were lower

performing on this assessment than White students. There were 17 more White students

than Black or African American students and four times as many students who were high

performing than the Hispanic and Other subgroups.

 123

Figure 25

Biology EOC Performance by Race/Ethnicity

STEM Electives Mean Performance by Race/Ethnicity.

Next, a STEM electives mean course performance comparison by race/ethnicity

was completed as well using a box and whisker plot. Figure 26 shows the distribution of

the STEM electives mean course performance data for each subgroup of students. On the

chart, four potential outliers were identified for the White subgroup only. Those mean

outlier scores were 39, 84, and 85 being identified as potentially mild outliers, and 86 as a

potentially significant outlier. This is discussed in more detail in the regression analysis

of the student performance data for race/ethnicity.

Figure 26 displays a median of 78 for Black or African American, 81 for

Hispanic, 90 for Other, and 89 for White students. Visually, the box plots fall within the

average to above performance for the Black or African American and the Hispanic
 124

subgroups and above performance overall for the Other and White subgroups. The

Hispanic subgroup appears to have more variation in their scores with the majority of the

scores falling below the median value with a greater amount of their scores falling in the

average to low-performing range. The same is true for the African American group, with

a slightly lower variation below the median.

Figure 26

STEM Electives Mean by Race/Ethnicity

Overall STEM Course Mean Performance by Race/Ethnicity. In further

breaking down the STEM course data to compare STEM course counts by race and

ethnicity, a clustered bar chart (see Figure 27) was created to show the distributions of

STEM course counts between the different subgroups. The total number of STEM

courses is provided on the x-axis with the percentage of students from each subgroup

who took those number of STEM courses on the y-axis. From this data set, it is important

to keep in mind that the Other subgroup has a much smaller number of students than the
 125

other subgroups in the data set. In comparing the subgroup data, the White subgroup

included a student with the highest count of STEM course enrollments of nine courses

and another with 13 courses. All subgroups included students who did not take a single

STEM course across the span of their 3 years in high school. As the number of STEM

courses in which the students could have enrolled increased, the smaller the percentage or

number of students from each subgroup became (the % of students from the other

subgroup is much smaller than what it appears on the chart due to the small size of the

subgroup).

Figure 27

STEM Course Overall Mean Performance by Race/Ethnicity

Multiple Regression for Overall Student Mean Academic Performance by Gender and

Ethnicity

Assumptions Checks. A multiple regression was run to predict academic

 126

performance from gender and ethnicity for Hispanic, Black or African American, and

Other student subgroups. I measured student academic performance with four student

assessment scores: overall mean performance in STEM classrooms, NC Math I, NC Math

III, and biology EOC performance. Each measure of academic performance was included

as a dependent variable within a separate regression analysis. I ran each regression model

using two categorical independent variables—gender and ethnicity—to explain the

amount of variation within the performance measure (i.e., overall mean performance in

STEM classrooms, NC Math I, NC Math III, and biology EOC performance). To

measure ethnicity, I created three independent dichotomous variables (Hispanic, Black or

African American, and Other) and used the White student subgroup as the reference

group. The linearity assumption was met by including only dichotomous independent

variables. Dichotomous variables do not require checks for linearity (Hardy, 1993); thus,

they automatically meet the assumption of linearity by creating two data points, and two

points define a straight line (Hardy, 1993; Morgan, 2017). Therefore, the assumption for

homoscedasticity was also met.

Next, I checked to ensure that there was no correlation between residuals and no

evidence of multicollinearity. There was independence of residuals, as assessed by a

Durbin-Watson statistic of 2.225 for STEM overall mean, 1.857 for NC Math I EOC

performance, 2.051 for NC Math III performance, and 1.735 for biology EOC

performance. There was no evidence of multicollinearity, as assessed by examining

tolerance values, studentized deleted residuals, leverage values, and Cook's distance for

each dependent variable—overall STEM mean, NC Math I, NC Math III, and biology

EOC performance. In reviewing the independent variables in Table 20, none of the
 127

independent variable correlation coefficients were above 0.7, showing no problem with

the correlations between the independent variables. In Table 21, all the tolerance values

are greater than .1 and VIFs less than 10; therefore, I was fairly confident that I did not

have a problem with collinearity in this particular sample.

 128

Table 20

STEM Academic Performance Assumption Check for Correlation

STEM Gender Hispanic Black or Other

mean African
overall American

Pearson STEM Mean Overall 1.00 0.07 -0.35 -0.07 0.14

correlation Gender 0.07 1.00 -0.03 0.16 0.09
Hispanic -0.35 -0.032 1.00 -0.29 -0.17
Black or African American -0.07 0.16 -0.291 1.00 -0.10
Other 0.14 0.09 -0.17 -0.10 1.00
NC Math I EOC 1.00 -0.11 0.28 0.20 -0.18
Gender -0.11 1.00 -0.03 0.16 0.09
Hispanic 0.28 -0.03 1.00 -0.29 -0.17
Black or African American 0.10 0.16 -0.29 1.00 -0.10
Other -0.18 0.08 -0.17 -0.10 1.00
NC Math III EOC 1.00 0.140 -0.40 -0.14 0.15
Gender 0.14 1.00 0.07 0.07 0.04
Hispanic -0.40 0.07 1.00 -0.29 -0.17
Black or African American -0.13 0.07 -0.29 1.00 -0.11
Other 0.15 0.04 -0.17 -0.12 1.00
Biology EOC 1.00 0.19 -0.38 -0.14 0.15
Gender 0.19 1.00 -0.20 0.15 0.13
Hispanic -0.38 -0.20 1.00 -0.24 -0.17
Black or African American -0.14 0.15 -0.24 1.00 -0.13
Other 0.15 0.13 -0.17 -0.13 1.00
 129

Table 21

STEM Academic Performance Assumption Check for Tolerance/VIF

Dependent variable Collinearity statistics

Tolerance VIF
STEM mean (Constant)
overall Gender 0.96 1.04
Hispanic 0.87 1.14
Black or African American 0.87 1.16
Other 0.93 1.07
NC Math I EOC (Constant)
score Gender 0.92 1.04
Hispanic 0.87 1.15
Black or African American 0.87 1.16
Other 0.94 1.07
NC Math III (Constant)
EOC score Gender 0.98 1.02
Hispanic 0.86 1.16
Black or African American 0.88 1.13
Other 0.94 1.07
Biology EOC (Constant)
score Gender 0.94 1.07
Hispanic 0.88 1.13
Black or African American 0.90 1.11
Other 0.93 1.08

Next, I reviewed standard residuals to check for potential outliers. There was one

case—for the dependent variable STEM mean overall—where the standard residual was

slightly above the 3.0 threshold for residuals. I reviewed this case to determine whether

or not to remove the student’s data from this sample, and after a careful review of the

student’s grades, I determined that the data for the student were valid. Additionally, I

tested the assumptions after removing the outlier, and there was no significant change

showing that although that student’s data was an outlier, it did not have a significant

impact on the results. No output for case-wise diagnostics was generated by the
 130

regression analyses conducted on NC Math I, NC Math III, and biology EOC, indicating

no possible outliers in the sample that would skew my results.

Table 22

STEM Mean Overall Residuals Check for Possible Outliers

Case number Standard residual STEM mean overall Predicted value Residual
15 -3.63 32.60 75.71 -43.11

Note: Dependent Variable: STEMMeanOverall.

Next, Cook’s distance and centered leverage values were checked (see Table 23).

In reviewing Cook’s distance and leverage, there were no values for Cook’s above 1 for

each category of academic performance. There was also a returned maximum value of

.206 for overall STEM academic mean and NC Math I performance, .251 for NC Math

III EOC performance, and .281 for biology EOC performance for leverage which was

within the leverage threshold.

 131

Table 23

STEM Mean Cook’s Distance and Leverage Value

Dependent Minimum Maximum Mean Standard N

variable deviation
STEM mean Cook's 0.00 0.15 0.01 0.02 90
overall distance
Centered 0.02 0.21 0.04 0.04 90
leverage
value
NC Math I Cook's 0.00 0.16 0.01 0.02 90
EOC score distance
Centered 0.02 0.21 0.04 0.04 90
leverage
value
NC Math III Cook's 0.00 0.18 0.01 0.02 66
EOC score distance
Centered 0.03 0.25 0.06 0.06 66
leverage
value
Biology EOC Cook's 0.00 0.23 0.03 0.043 46
distance
Centered 0.04 0.28 0.09 0.06 46
leverage
value

Finally, the assumption of normality was met, as assessed by the P-P Plots

produced in the multiple regression output conducted within SPSS. The P-P Plots shown

in Figure 28 for each of the dependent variables showed although the points are not

aligned perfectly along the diagonal line, they resemble a hand-drawn line thus indicating

that the residuals are close enough to normal for the analysis to proceed for overall

STEM mean academic performance. Therefore, no transformations or other corrections

on the data needed to take place, as the assumption of normality was not violated.

Figure 28
 132

Overall STEM Mean P-P Plot Assumption for Normality

Figure 29

NC Math I EOC P-Plot Assumption for Normality

 133

Figure 30

NC Math III EOC P-Plot Assumption for Normality

Figure 31

Biology EOC P-Plot for Assumption for Normality

Multiple Regression Results. Four multiple regression models were conducted,

 134

one for each of the dependent variables used as a measure of academic performance

within this study—STEM mean overall, NC Math I, NC Math III, and biology EOC. The

first multiple regression model statistically significantly predicted STEM mean overall, F

(4, 45) = 3.022, p < .001, adj. R2 = .129. R2 for the overall model was 16.8% with an

adjusted R2 of 12.9% for STEM mean overall performance. The second multiple

regression model statistically significantly predicted NC Math I EOC performance, F(4,

89) = 3.397, p < .001, adj. R2 = .097. R2 for the overall model was 13.8% with an

adjusted R2 of 9.7% for NC Math I EOC performance. The third multiple regression

model statistically significantly predicted NC Math III EOC performance, F(4, 65) =

5.540, p < .001, adj. R2 = .218. R2 for the overall model was 26.6% with an adjusted R2

of 21.8% for NC Math III EOC. The forth multiple regression model statistically

significantly predicted biology EOC performance, F(4, 45) = 3.022, p < .001, adj. R2 =

.152. R2 for the overall model was 22.8% with an adjusted R2 of 15.2% for biology EOC

performance.

Based on widely accepted guidance on effect size (i.e., 0.10 – < 0.30 = small, 0.30

– < 0.50 = medium, and ≥ 0.50 = large), the regression results for each dependent

variable resulted in a small size effect as they all fell within the .10 to .30 or 10% to 30%

range (Cohen, 1988), thus indicating the independent variables had low statistical

significance to the prediction for each dependent variable, although it was somewhat

close to a medium effect for the NC Math I, NC Math III, and biology EOC scores

(Wilson & Lipsey, 2001). Regression coefficients and standard errors for each regression

model output are found in Table 24.

 135

Table 24

Multiple Regression Results for STEM Academic Performance

Regression B 95% CI for B SE B  R2 R2

model
LL UL
Overall 0.17 0.13
STEM mean
Constant 84.20 80.12 88.28 2.05
Gender 2.27 -2.87 7.41 2.59 0.08
Hispanic -10.76 -16.42 -5.12 2.84 -0.40
Black or -7.03 -14.66 0.58 3.83 -0.20
African
American
Other 2.42 -8.82 13.68 5.66 0.04

NC Math I 0.138 0.097

Constant 396.90 350.83 442.96 23.17
Gender -35.91 -93.92 22.11 29.18 -0.12
Hispanic 93.57 29.78 157.35 32.08 0.31
Black or 79.36 -6.59 165.30 43.22 0.20
African
American
Other -58.75 -185.68 68.17 63.84 -0.10

NC Math III 0.266 0.218

Constant 553.99 550.71 557.27 1.64
Gender 3.58 -0.53 7.69 2.05 0.193
Hispanic -9.69 -14.34 -5.05 2.32 -0.49
Black or -7.37 -13.35 -1.40 2.99 -0.29
African
American
Other 0.97* -7.73 9.68 4.35 0.02

Biology 0.23 0.15

Constant 258.20* 254.16 262.25 2.001
Gender 2.39*** -2.31 7.09 2.33** 0.15
Hispanic -7.89*** -13.55 -2.23 2.80 -0.41
Black or -5.91**** -12.57 0.74 3.29 -0.26
African
American
Other 0.75 -7.61 9.11 4.14 0.02

Note. Model=”Enter” method in SPSS Statistics; B=unstandardized regression

 136

coefficient; CI=confidence interval; LL=lower limit; UL=upper limit; SE B=standard

error of the coefficient; =standardized coefficent; R2=coefficent of determination;

R2=adjusted R2. *p<.05, **p<.01, ***p<.001.

Results and Analysis

This section is organized to present the study’s findings by research question;

thus, I present the results in response to each of the study’s research questions and

whether the results supported or rejected the null hypothesis associated with each

question.

Research Question 1: What Are High School Seniors’ STEM Persistence in High

School After a 3-Year Enrollment in a STEM Middle School?

Before examining student course enrollment, student opportunities to enroll in

STEM were explored to ensure students had opportunities to develop STEM career

pathways. Without opportunities to persist, this would have limited my ability to measure

student persistence. High school students in the district have approximately 60 STEM

course options in math, science, and a wide variety of STEM-oriented electives to select

from throughout their high school career. They also have 16 opportunities within their

first 3 years to establish a STEM career pathway. After determining that the total number

STEM courses taken by students was normally distributed, it was observed that a large

number of study participants enrolled in four STEM courses on average, with the largest

number of participants enrolling in four to six courses. There was only one student of the

90 participants who took advantage of 13 of the 16 opportunities they had to take STEM

courses, but a slightly greater number of students on the other end of the spectrum did not

take a single STEM-oriented course.

 137

The district’s CTE concentrators guide, a guide that shows how many courses in a

particular pathway a student would have to take to be considered a “concentrator” in that

pathway, indicates that students have to take at least four courses in a particular career

pathway including one completer course (i.e., a course that has a prerequisite course).

Similar to expectations for concentrators, many study participants participated in at least

four or more courses, four being the mean amount of STEM courses taken by study

participants. The majority of participants, 63 of 90 (70%), completed at least four or more

STEM courses. The results support my research hypothesis that high school seniors, after

a 3-year enrollment in a STEM-focused middle school, will enroll in advanced STEM-

related core courses and STEM-oriented electives throughout high school.

Research Question 2: What Are High School Seniors’ Academic Performance in

STEM Courses in High School After a 3-Year Enrollment in a STEM Middle School?

Study participant performances in the STEM-oriented courses in which they

enrolled and completed were observed along with their EOC performances for three of

the required STEM courses—Math I, Math III, and biology. These courses were the only

STEM courses in which students were required to take an EOC as 20% of their final

grade for that course. In regard to classroom performance, student participants were high

performing based on their mean score for all STEM courses in which they enrolled;

however, they were close to average performing with an overall low B of 81. This

average was consistent as well when reviewing their performances on their math cluster,

science cluster, and STEM electives cluster of courses.

All the study participants took the NC Math I assessment and the majority of them

did not score proficient (70% were low performing). Only 2% of the students were high
 138

performing on the assessment, with the remaining 28% demonstrating average

performance. The nonproficient scores could have impacted the overall mean for math,

which could have further impacted the overall mean for all STEM courses together

considering the assessment again does carry a 20% weight on final course grades. The

NC Math III performance was much improved as a greater percentage of students were

high performing on this assessment than the NC Math I. The percentage of students who

were high performing was higher for the biology assessment, although only half of the

students completed the assessment. Eighty-three percent of students who completed the

biology assessment were proficient.

In revisiting my hypothesis for this research question—High school seniors, after

a 3-year enrollment in a STEM-focused middle school will experience high academic

achievement in those courses throughout high school—the results support that the

students, overall and by average, are high performing. A closer look at the data, however,

does support that there is still room for improvement, as the students are very close to the

average performance range.

Research Question 3: How Do High School Seniors’ STEM Persistence and Academic

Performance in STEM Courses Compare by Gender and Ethnicity?

The results for STEM persistence by gender showed that based on the number of

courses taken over a 3-year period by female students as compared to male students, the

male subgroup was twice as persistent in taking STEM courses than the female subgroup,

meaning they enrolled in and completed, on average, twice as many STEM courses. In

addition, there were no females that took more than six STEM-oriented courses

compared to a large number of males enrolled in STEM courses. In viewing persistence

 139

in math, science, and STEM electives, females were slightly less persistent on average in

enrolling in advanced mathematics courses than males. The males enrolled, on average,

in significantly more science (73% more) and STEM elective courses (70% more) than

females. In comparing performance in STEM courses, females slightly outperformed

males across the board in math, science, and STEM electives as well as overall course

grades. Females were more high performing than males on the NC Math I (4% more),

NC Math III (10% more), and biology EOC (3% more) assessments.

A review of STEM persistence by race/ethnicity showed that the White subgroup

of student participants enrolled and completed one more course on average than the

African American students; however, White students enrolled and completed a

comparable number of courses compared to the Hispanic and Other student subgroups.

The African American group had a significantly lower rate of students enrolled in STEM-

oriented math classes than any other student subgroups. The African American subgroup

did not have any students enroll in more than six STEM courses compared to the other

student subgroups. The White subgroup had the largest number of students enroll in a

higher number of courses above the average STEM course enrollment count (between

seven and 13).

In reviewing the data results for STEM academic performance, the Black or

African American subgroup performed lower than the White subgroup in science, STEM

electives, and STEM courses overall, but slightly higher in mathematics. There were only

two African Americans who took advanced mathematics courses compared to 27 White

students. The Hispanic subgroup performed the lowest across the board compared to all

other subgroups although they had twice as many students enrolled in math, science, and
 140

STEM electives than the African American subgroup. In comparing EOC assessment

performance, the White subgroup had a significantly lower percentage of students who

were proficient on the NC Math I EOC.

My research hypothesis for this research question was that due to the ongoing

nature of educational inequality, underrepresented subgroups will have less positive

results in persistence and performance than other subgroups. In the case of the African

American subgroup, this hypothesis was correct, as the African American and Hispanic

subgroups had lower performance rates than the Other and White subgroups.

Summary

The purpose of this study was to explore student persistence and academic

performance in STEM after attending a STEM-focused middle school from sixth through

eighth grades. Throughout Chapter 4, data were presented that shared student STEM

persistence based on the number of STEM courses in which they enrolled over a 3-year

period in high school. The course enrollment was examined not only for all STEM

courses but broken down into the categories of math, science, and STEM electives. In

addition, results of student STEM academic performance were presented that provided

student mean performance on all the STEM courses in which they have enrolled. The

data were broken down to further examine differences between STEM persistence and

academic performance between male and female subgroups and then between Black or

African American, Hispanic, White, and Other subgroups.

The results of average persistence for student participants overall show that male

students have a higher STEM persistence than female students and White and Other

subgroups have a higher STEM persistence than Black or African American students. In
 141

addition, the results showed slightly higher STEM academic performance for females

than males. Findings also indicated higher STEM performance for White and Other

subgroups than Black or African American and Hispanic subgroups. A multiple

regression analysis was run as well to test the correlation between gender, ethnicity, and

STEM persistence and academic performance. All regression tests resulted in a low effect

size for all models indicating gender and ethnicity had a small but significant contribution

to explaining student academic performance. The overall study findings supported my

hypothesis for each research question.

In Chapter 5, I provide an overview of the study and an interpretation of the

findings. In addition, I discuss the connections to the theoretical framework and the

implications of the study results for STEM education. Lastly, I provide recommendations

for action and further study.

 142

Chapter 5: Discussion

Overview

The purpose of this quantitative research study was to explore the persistence and

academic performance of high school students who attended a STEM-focused middle

school throughout their sixth-, seventh-, and eighth-grade years. The research questions

for the study focused on examining student STEM persistence (dependent variable) by

reviewing data pertaining to the STEM courses in which they were enrolled and student

STEM academic performance (dependent variable) by reviewing data pertaining to

student grades and EOC performance in STEM-oriented courses. In addition, I sought to

explore similarities and/or differences between student STEM persistence and STEM

academic performance by reviewing these data by gender (independent variable) and

race/ethnicity (independent variable).

This study employed a quantitative and nonexperimental design. The data utilized

for the research included a deidentified roster of student course enrollments, grades, and

EOC performance for students who all attended the same STEM-focused middle school

for their sixth-, seventh-, and eighth-grade years and who all transferred from that STEM

middle school to the same high school. I analyzed the data using descriptive statistics and

a standard multiple regression model. Through the multiple regression, I sought to

discover differences that may exist between two independent variables over five

dependent variables. I conducted this study to discover the impact of attending a STEM-

focused middle school on student STEM persistence and academic focus and how

traditionally underrepresented subgroups of students in STEM compare to students who

are not traditionally underrepresented. The overall goal was to examine whether student
 143

attendance for 3 years at a STEM-focused middle school has the intended impact of

encouraging students to follow a STEM-career pathway in high school and to be prepared

to perform well in STEM-oriented courses in which they enroll in high school.

For STEM persistence, key findings revealed that overall, the majority of

participants took, on average, an ample amount of STEM courses to be considered as

being persistent. In comparing gender differences in STEM persistence, there was a

significant number of males showing persistence in the number of STEM courses overall

they completed compared to females and triple the number of advanced science or

science beyond what is required and STEM-elective courses. In comparing race/ethnicity,

African American students enrolled in significantly lower STEM courses in high school

than White students—more so in advanced mathematics courses.

In regard to STEM academic performance, on average, the students were high

performing in all of their STEM classes combined and in each category of STEM courses

including advanced mathematics, advanced science courses, and STEM electives.

Although these students have high performance on their STEM course averages, the

majority of students were low performing on the math EOCs, but the majority were

average to high performing on the science EOC. In comparing gender differences in

STEM academic performance, females slightly outperformed males in math, science, and

STEM elective courses on average, as well as on their performance on the math and

science EOCs.

In comparing race/ethnicity subgroups, the traditionally underrepresented

groups—African American and Hispanic subgroups—performed lower academically in

STEM courses on average than the White subgroup.

 144

Interpretation of Findings

In reviewing the outcomes of the data, it was important not to overgeneralize the

results but to explain the STEM persistence and STEM academic performance for this

sample of students to the total population of students who have attended STEM-focused

middle schools. Some of the findings support the literature review research, while others

negate it. In this section, I provide an interpretation of the results of the data based on the

research questions for this study.

The first literature review questions related to what STEM persistence of students

would look like in high school after attending a 3-year STEM middle school during their

sixth-, seventh-, and eighth-grade years. A big challenge STEM educators encounter is

trying to engage older students with STEM for the first time (Earth Networks, 2017).

According to Morgan (2015), and also as previously mentioned in the literature review,

students’ middle school years are pivotal to the development of student career pathways

as students begin to seriously consider what they want to do or be in life. Their STEM

education in middle school can impact their opinions about STEM subjects and subjects

they will take in years beyond middle school. In observing the outcomes, the measure of

student STEM persistence was operationalized by comparing the number of courses

students in which enrolled during their first 3 years in high school to the number of

courses in which students generally enroll that would allow the student to be considered

as a “concentrator” of a specific career pathway. There was a wide variety of STEM

courses students could take to show their persistent interest in STEM by enrolling in four

or more STEM courses. The majority of the participants had enrolled in four to six

courses.
 145

The second research question related to what the STEM academic performance of

students looks like in high school after attending a 3-year STEM-focused middle school

during their sixth-, seventh-, and eighth-grade years. Research findings in a study

conducted by Hacioglu and Gulhan (2021), where middle schoolers were exposed to

engineering-designed-based activities, resulted in skills such as open-mindedness, truth-

seeking, listening to others’ opinions, and other critical-thinking skills improving as a

result of the experience. Their experiences were based on improving their critical-

thinking and problem-solving skills. “When educators teach children how to find

solutions, they gain confidence and go on to achieve academic success in high school and

college” (Howard, 2021, para. 11). With this in mind and other research from the

literature review that supports STEM education’s impact on improving student academic

abilities, I anticipated that their 3 years of STEM education in middle school would assist

in developing the skills necessary to be successful in STEM courses in which they would

engage in high school, much like those participating in the Hacioglu and Gulhan (2021)

study.

On average, students were high performing in the STEM core and elective

courses they completed. This was based on the final grades they scored for their STEM

courses. Their classroom performance, however, did not align overall with their math and

science EOC assessment performance. The majority of students were low performing on

the NC Math I assessment, but they were high performing on the biology assessment

(keeping in mind that only half of the students completed this assessment). Although high

performing, on average, their overall performance was only slightly above the average

performance threshold.
 146

The third research study question related to how the high school senior’s STEM

persistence and academic performance in STEM courses compare by gender and

ethnicity 3 years after attending a STEM-focused middle school during their sixth-,

seventh-, and eighth-grade years. In comparing student results by gender, male students

had completed, on average, double the number of courses than females over a 3-year

period in high school. The enrollment statistics were far more significant for STEM

science and elective courses for males over females than math. The persistence results for

females compared to males in this study align with this idea. Increasing the number of

females who choose to study STEM subjects at school is the first step in reducing the

STEM gender gap in the workplace. This is one of the goals of STEM education in

STEM schools.

In comparing academic performance by gender, females outperformed males

overall, on average, and in math, science, and STEM electives when reviewing the data

by those clusters of courses. Recent studies are showing that female students perform

well in STEM subjects at school (Liberatore & Wagner, 2020). While women perform at

the same or higher level in math and science as men, their performance in the humanities

is markedly better. This may be the reason they are choosing not to pursue STEM

careers.

In comparing race/ethnicity differences in STEM persistence, the African

Americans were less persistent in completing STEM courses than the other subgroups,

with the African American subgroup being the least persistent of all, especially in

mathematics. These results align with current research that reports that “students from

marginalized groups sometimes don’t see themselves going into science, engineering, or
 147

technology. STEM Education can empower them by giving them the knowledge and

skills necessary to succeed in those types of positions” (Howard, 2021, p.15). Although

these students were exposed to STEM for 3 years in middle school, those experiences

seemed not to be enough to encourage a significant number of African Americans to

follow a STEM career pathway in high school by enrolling in just as many STEM-

oriented courses throughout high school as other subgroups of students. Temming (2021)

recently shared research that aligns with this research study result where from 2017-2019,

there were only 9% of Blacks or African Americans in the U.S. who worked in STEM

jobs or careers, with Hispanics at a lower rate of 8%, while Asians and Whites continue

to overrepresent in the STEM field.

In comparing race/gender ethnicities and academic performance, the Hispanic

subgroup was the lowest performing of all other subgroups in math, science, and STEM

electives. The African American subgroup was outperformed in science and STEM

electives but slightly outperformed the White subgroup in math, although the African

American subgroup size was significantly smaller. The Hispanic subgroup was the lowest

performing of all subgroups. Although the goal of STEM education has been targeted to

increase interest and performance in STEM, African American students continue to lag

significantly behind White students. STEM education still has some improvements to

make overall in addressing the disparities that continue to have a negative impact on

African American student performance compared to other ethnicities to assist African

Americans in being resilient to other societal factors that can impact their performance.

Hispanic students continue to experience lower academic achievement in STEM subjects

as compared to White subgroups too.

 148

Connection to Theoretical Framework

SCCT is applicable to the results of the study as well as the study’s results focus

on student STEM persistence and STEM academic performance, especially with its

overall focus on how students make choices about their interests, especially when

developing a career path in high school (Lent et al., 2002). How students have persisted

and performed in STEM throughout high school could have been impacted by how they

have seen themselves in STEM, especially as they experienced STEM through their

experience in attending a STEM-focused middle school for 3 years. SCCT focuses on

tracing the role of interest and its impact on choice and skills development. After 3 years

in a STEM middle school, my desire was to discover whether that experience may have

impacted student interests in STEM and even their academic performance in STEM,

despite other factors that may be a hindrance, especially with STEM efforts focusing on

especially attracting underrepresented subgroups of students. I believe that if those

experiences throughout their experience in their STEM middle school were able to

strongly influence their career choices, we would see a greater influx of underrepresented

groups of students taking an increasing number of STEM courses throughout high school.

Many student choices are based on what they are exposed to that have potential relevance

to occupational behavior in school (Lent et al., 2002; Social Cognitive Career Theory,

2021).

In addition, my study’s results of lower persistence and lower academic

performance for the underrepresented subgroups, compared to the overall results for all

student data that were used in the study, could have been further impacted by student

self-efficacy and competency in STEM throughout their middle school years (Bandura,
 149

1986; Lent et al., 2002). SCCT describes the impact that “ability” has on student

performance, where if a person does not believe they are capable of performing well in

something, they tend to either avoid their participation in it or they do not believe they

can do well at something and do not put forth their best effort, even when given the

opportunity to do so (Social Cognitive Career Theory, 2021). While unknown for the

particular students whose persistence and performance data were reviewed in this study,

this component of the theory could further explain the study’s results. As reported in the

literature review, there is still not substantial research on middle school STEM

experience and its impact on high school choices and performance in STEM; therefore,

allowing my study to add to this body of knowledge and extend the research in this field.

Implications

The problem and results of this research study related to the problem provide

several implications for STEM education, especially in middle schools with a STEM

focus. These implications relate to how middle school STEM programs are developing

student interest in STEM (especially for underrepresented students), how other factors

that can impact STEM interest despite a school’s STEM education focus are being

addressed, and what STEM education looks like in the STEM-focused middle school’s

courses in regard to developing student skills and performance in STEM. The

implications would not only improve student individual persistence and performance but

would increase underrepresented student persistence and performance as a whole,

therefore having the potential to impact student high school choices and performance and

possibly increase the number of underrepresented students in the STEM college to career

pipeline, thus contributing to the nation’s goal of increasing the presence of

 150

underrepresented gender and race/ethnic groups in STEM fields with the skills and ability

to perform successfully in their roles.

The first implication is the importance of ensuring that STEM-focused middle

schools have ample and effective practices in place to assist in developing student interest

in STEM throughout their middle school experience. While this includes all students,

extra focus is needed for underrepresented subgroups of students, so they, like their peers

in traditionally represented subgroups, have ample exposure to STEM. Students need

explicit connections to STEM through their core and elective courses throughout middle

school and consistent and ample exposure to STEM professionals. Teaching students

about Black or African American and Hispanic excellence in STEM can further motivate

and inspire students to gain more interest in STEM versus just focusing on excellence in

STEM in general. Females, Black or African American, and Hispanic students need

affirmation in a way that is relevant to their lives. There are many female, Black or

African American, and Hispanic people in history and today who have influenced STEM

in our world and continue to do so (Holly, 2021). This does not mean to leave out the

contributions of other subgroups that already have overrepresentation in STEM, but it

means to showcase a balance of diversity in STEM excellence. We are already doing a

great job of STEM representation in White subgroups and will not decrease our efforts

for that group but will ensure equity of representation across all subgroups.

To do so does not necessarily automatically increase student interest as one may

believe that it could or should. That leads to the second implication. It is vital for STEM

districts to explicitly identify and acknowledge the specific barriers that exist within

society as a whole and in the communities in which the students in the district live. These
 151

are all the same types of barriers described in SCCT. We cannot ignore the biases,

prejudices, and racism—whether explicit or systematic—that exist that continue to have

more of an impact on underrepresented students than the current efforts that may be in

place in STEM middle schools and beyond. STEM middle schools cannot ignore this

factor and at the same time say that they desire through their efforts to improve student

interest, knowledge, and skills in STEM. The goals in our STEM middle schools must

align with the goals for STEM education in our nation in order to address our greater

cause for advancing STEM education in our public schools. STEM leaders and educators

must ensure that they are well-educated on the barriers and work collaboratively to not

just acknowledge the barriers that exist that are impacting student STEM persistence and

performance, but actively address the issues head on to determine how our schools can

assist in helping students overcome many, if not all, of those barriers. Failure to do so

could be considered a disservice to female, Black or African American, and Hispanic

students in STEM schools.

The third implication is the importance of ensuring equity in instructional

practices that are being utilized within classrooms to assist students in realizing their

abilities and skills to be successful in math, science, and other STEM-related subjects.

The study’s results extend STEM education impact research, whose findings continue to

reveal the need to improve the gender and race/ethnicity gaps through culturally and

gender-responsive pedagogy. Traditional approaches to teaching math and science have

for many years proven ineffective for female, Black or African American, and Hispanic

students, due to the characteristics of scientific language, competition, and rigor without

social relevance to student lives and experiences, therefore devaluing the learning
 152

experiences for these students causing them to lose interest and to become disengaged in

their learning which generally leads to underperformance (Ananga, 2021; Holly, 2021).

Limitations of the Study

The first limitation of the study was the location. This study only included a small

cluster of students who attended a STEM-focused middle school in the district during

their sixth-, seventh-, and eighth-grade years. These students also went on to attend the

same high school. Including a more widespread sample of students would be more

beneficial for the study.

The second limitation included the inability for me to include one of the research

methodologies I initially proposed for this research study. In addition to examining

student persistence and academic performance in STEM, I originally wanted to collect

data from students using a STEM interest and attitude survey. I was able to be successful

in getting permission from the author to use an already-validated STEM survey that

perfectly aligned with research questions that pertained to student dispositions and self-

efficacy. I even had the added benefit of having one of the survey creators sit on my

dissertation committee. I ran into a roadblock because, as the researcher, I was not

permitted to know the identities of the students completing the survey and I had to

anonymously attain parental permission for the students to participate in the survey as

well. I was unable to attain cooperation from the high school the students attended to

assist with attaining parental permission and administering the survey electronically to

the students. While administration and staff valued the research and were willing to help,

doing so was not feasible; therefore, I attained permission from the university IRB, the

district’s DRA office, and my dissertation committee to remove the STEM dispositions
 153

and self-efficacy component from the research and to consider it as part of the

recommendation for further research.

The third limitation of this research study was time. The time available for me to

complete the research was not ample for me to include other factors that were of interest

to me to explore which would have provided more background on STEM education

programs and practices at STEM middle schools in the district, most specifically the

STEM middle school in which the students whose data was utilized in this study,

attended for their sixth-, seventh-, and eighth-grade years. Instead of just a quantitative

study, a mixed methods research design would have provided much more insight as I

could have had the opportunity to possibly interview school STEM coordinators,

administrators, and/or teachers and could have possibly included focus groups with

students. In addition, the timeline in which I was able to receive data needed to review

STEM persistence and performance placed me on a very tight schedule to review,

interpret, and analyze the results of just those two components alone.

The fourth limitation of the study was the sample size. A larger sample size would

have allowed me to be able to have a greater effect size on the research results, especially

when running the multiple regressions on the independent and dependent variables of the

research study. The additional data would have provided more information allowing my

results to be more precise, thus decreasing any uncertainty about the research (Little,

2014). Due to staffing limitations in the DRA office and time, it was not feasible during

this study to obtain additional student data to increase the sample size and effect. With

the district being one of the largest school districts in North Carolina, the DRA office is

inundated with data requests among other services such a small department has to
 154

provide to such a large district.

The fifth limitation was the potential impact the COVID-19 pandemic may have

had on our middle school STEM education programs, thus possibly impacting school

abilities to provide highly influential and effective STEM learning experiences. In 2019,

students had to receive instruction across the world in a way the majority has never had to

experience. School district leaders, administrators, teachers, parents, and students were

faced with challenges that greatly impacted access to instruction and social and emotional

well-being. Many courses that were generally available to students may not have been

fully available to students throughout the pandemic due to the impact on staffing needed

to provide both remote and in-person instruction. The negative impacts of the pandemic

were felt across the globe and are still having an effect on education although districts

have mostly returned to in-person instruction.

Recommendations for Action and Further Studies

The implications for each of the study’s research questions and limitations

provide the foundation for the recommendations for actions. Although STEM education

is proven through many research studies to have a positive effect on student STEM

persistence, knowledge, and performance in STEM overall, there continues to be gender

and race/ethnicity disparities. The district supports the advancement of STEM education

in the district and one recommendation is for the district to more closely examine

culturally and gender-responsive teaching practices, especially within its STEM-focused

schools, considering its goal of addressing underrepresented students in STEM. This can

be done collaboratively with the district’s office of equity affairs, its curriculum

enhancement and magnet programs office, and professional learning. Instructional

 155

walkthroughs throughout the district’s STEM middle schools could provide valuable

insight into the prevalence of culturally and gender-responsive strategies and practices

taking place within the programs and practices being implemented. This is recommended

to all school districts that have STEM-focused middle schools. Walkthroughs can provide

beneficial data and trends that could provide some guidance as to areas of improvement

that could be addressed through district-provided professional development, coaching,

and support.

In addition, I recommend that further research on this topic is carried out to

examine student STEM dispositions and self-efficacy to provide additional information

that has the potential to explain student STEM persistence and performance in high

school after attending a STEM-focused middle school for their sixth-, seventh-, and

eighth-grade years from the lens of SCCT. The theory proposes that dispositions and self-

efficacy can have an impact on persistence and performance and therefore gathering data

on these variables would be beneficial to extend this research. This purpose can be

accomplished through the administration of a STEM survey where students would

provide input relating to their attitudes towards STEM which will provide input regarding

their dispositions, persistence, and self-efficacy. I highly recommend using the

Middle/High School Student Attitudes Towards STEM survey.

This instrument was developed by researchers from North Carolina State

University’s (NCSU) Friday Institute for Educational Innovation--Malinda Faber, Dr.

Alana Unfried, Dr. Eric N. Weibe, and Dr. Jeni Corn. Malinda Faber is a research

associate at the Friday Institute for Educational Innovation at NCSU (see Appendix B).

Dr. Alana Unfried was a graduate research student at the Friday Institute as well during
 156

their development of this instrument and is currently a professor in the mathematics

department at California State University Monterey Bay. Dr. Eric Weibe is a professor in

the Department of STEM Education and a Senior Research Fellow at the Friday Institute

at NCSU. Lastly, Dr. Jeni Corn, during the development of this instrument, was the

Director of Evaluation Programs at the Friday Institute for Educational Innovation at

NCSU and is currently the Director of Strategic Initiatives at myFutureNC.

The survey was created, with permission, from other validated instruments to

ensure alignment with the National Science Foundation outreach program evaluation

goals. This survey has also been piloted with 9,108 sixth through 12th graders to further

assess and establish validity. Given the specifics provided on this instrument leads to why

I would highly recommend using it for a further research study to further extend the

results of my study (Faber et al., 2013).

Lastly, this study was limited to one North Carolina school district and only

included data from one of several clusters of students who would have qualified to have

their data used in this study. It would be beneficial for this study to be replicated in other

school districts to see if results were similar which has the potential to further validate my

research findings and implications for STEM education.

Conclusions

The major finding for this quantitative, nonexperimental research study is that

students who attend a STEM-focused middle school for their sixth-, seventh-, and eighth-

grade years show positive STEM persistence and performance in high school. Although

this is the case, further examination of the results continues to show gender and race/

ethnicity disparities where traditionally underrepresented subgroups of students in STEM

 157

are showing the same underrepresentation in STEM throughout their high school years.

There were no major differences in the relationships between gender and race/ethnicity

according to regression models presented in the research findings for either persistence or

academic performance; therefore, all our underrepresented groups should be able to

persist and perform at a similar rate to our traditionally represented groups in STEM. The

results showed this to be the case for females for academic performance but an ongoing

issue for the African American and Hispanic subgroups. In addition, although females are

outperforming males in STEM subjects, they are still not persisting in STEM in high

school. Further examination into SCCT, as it relates to student participation in STEM

education throughout middle school, can significantly impact the role educational

stakeholders can play in addressing other factors that may be diminishing the impact

middle school STEM education can have on underrepresented students.

In a recent report, the U.S. Department of Education (2022) shared the continued

importance of preparing students to solve some of the most complex challenges of our

nation through quality STEM education regardless of a student’s gender, race/ethnicity,

or location where they live. An abundance of STEM education efforts is increasingly

being provided to school districts across the nation to ensure education stakeholders have

the tools, resources, and support needed to succeed in increasing the number of students

moving into the STEM education to career pipeline. Ensuring that all students from all

backgrounds have equitable opportunities to advance in STEM is equally important.

Doing so can have a significant impact on our nation’s ability to become more globally

completive and rise to the top of the STEM workforce. My research study results provide

that there is still some work to do to ensure this happens. The abilities of STEM schools
 158

to break down barriers that impact equity in STEM can increase opportunities for

students and improve data trends in future research studies on student STEM persistence

and academic performance overall.

 159

References

Acara, D., Tertemizb, N., & Taşdemirc, A. (2018). The effects of STEM training on the

academic achievement of 4th graders in science and mathematics and their views

on STEM training teachers. International Electronic Journal of Elementary

Education, 10(4), 505–513. https://doi.org/10.26822/iejee.2018438141

Adams, C. (2017, February 23). The 7 most important STEM skills we should be

teaching our skills. We are Teachers. https://www.weareteachers.com/important-

stem-skills-teaching-kids/

Almeda, M. V., & Baker, R. S. (2020). Predicting student participation in STEM careers:

The role of affect and engagement during middle school. Journal of Educational

Data Mining, 12(2). https://files.eric.ed.gov/fulltext/EJ1267599.pdf

Ananga, E. D. (2021, April). Gender responsive pedagogy for teaching and learning: The

practice in Ghana’s initial teacher education programme. Creative Education, 12,

848-864. https://doi.org/10.4236/ce.2021.124061

Aspirations in Career Decisions. (2021). IResearchNet: Career research.

http://career.iresearchnet.com/career-development/aspirations-in-career-

decisions/#:~:text=Thus%2C%20the%20intention%20to%20pursue%20the%20g

oal%20and,occupation%20or%20a%20particular%20position%20within%20an%

20occupation.

Ball, A., Joyce, H. D., & Anderson-Butcher, D. (2016). Exploring 21st century skills and

learning environments for middle school youth. International Journal of School

Social Work, 1(1). https://doi.org/10.4148/2161-4148.1012

 160

Bandura, A (1977). Self-efficacy: Toward a unifying theory of behavioral change.

Psychological Review, 84(2), 191–215.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.

Prentice Hall.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist,

44(9), 1175–1184. https://doi.org/10.1037/0003-066X.44.9.1175

Berland, L., Steingut, R., & Ko, P. (2014). High school student perceptions of the utility

of the engineering design process: Creating opportunities to engage in engineering

practices and apply math and science content. Journal of Science Education and

Technology, 23(6), 705–720. https://doi.org/10.1007/s10956-014- 9498-4

Blotnicky, K. A., Franz-Odendaal, T., French, F, & Joy, P. (2018). A study of the

correlation between STEM career knowledge, mathematics self-efficacy, career

interests, and career activities on the likelihood of pursuing a STEM career among

middle school students. International Journal of STEM Education, 5(1), 22.

https://doi.org/10.1186/s40594-018-0118-3

Borgen, F. H. (1991). Megatrends and milestones in vocational behavior: A 20-year

counseling psychology retrospective. Journal of Vocational Behavior, 39, 263–

290.

Bronfenbrenner, U. (2005). Making human beings human: Bioecological perspectives on

human development. SAGE Publications, Inc.

Bryner, J. (2017). Women in STEM fields get boost from President Trump. LiveScience.

https://www.livescience.com/58066-president-trump-supports-women-in-

science.html
 161

Bybee, R. W. (2010, September). Advancing STEM education: A 2020 vision.

Technology and Engineering Teacher, 70(1), 30-35.

Cherry, K. (2020, July 22). Self-efficacy and why believing in yourself matters. Very

Well Mind. https://www.verywellmind.com/what-is-self-efficacy-

2795954#:~:text=Self%2D%20Efficacy%20Overview,succeed%20in%20a%20pa

rticular%20situation.

Choi, N., & Chang, M. (2011). Interplay among school climate, gender, mathematics

attitude, and mathematics performance of middle school students. Middle-Grade

Research Journal, 6, 15-28.

https://www.researchgate.net/publication/265784315_Interplay_among_school_cl

imate_gender_mathematics_attitude_and_mathematics_performance_of_middle_

school_students

Christensen, R., & Knezek, G. (2017). Relationship of middle school student STEM

interest to career intent. Journal of Education in Science, Environment and Health

(JESEH), 3(1), 1- 13.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).

Lawrence Erlbaum Associates.

Coleman, A. (2020, July). D-STEM equity model: Diversifying the STEM education to

career pathway. Athens Journal of Education, 7(3), 273-296.

https://files.eric.ed.gov/fulltext/EJ1256642.pdf
 162

Conley, D. T. (2012, May 2). A complete definition of college and career readiness.

Educational Policy Improvement Center.

http://schools.shorelineschools.org/swstaff/files/2014/08/College-and-Career-

Ready-Conley.pdf

David, A. (2019, April 16). The push for STEM education: Why it matters. Teach

Thought. https://www.teachthought.com/education/the-push-for-stem- education-

why-it-matters/

DeAngelis, S. (2015, September 4). STEM education helps teach skills necessary for 21st

century success. Supply Chain Brief. https://enterrasolutions.com/blog/stem-

education-helps-teach-skills-necessary-for-21st-century-success/

Dede, C. (2010). Comparing frameworks for 21st-century skills. In J. Bellanca & R.

Brandt (Eds.), 21st-century skills: Rethinking how students learn (pp. 51–76).

Solution Tree Press.

Department of Economic & Social Affairs. (2018). Sustainable development: The 17

goals. United Nations. https://sdgs.un.org/goals

Earth Networks. (2017, January 9). STEM education: 4 ways to inspire STEM learners.

https://www.earthnetworks.com/blog/inspire-stem-learners/

Eccles, J. S., & Wigfield, A. (2002). Motivational beliefs, values, and goals. Annual

Review of Psychology, 53(1), 109–132.

Faber, M., Unfried, A., Weibe, E, Corn, J., Townsend, L., & Collins, T. L. (2013, June

23-26). Student attitudes toward STEM: The development of upper elementary

school and middle/high school student surveys [Paper Presentation]. 120th ASEE

Conference & Exposition, Atlanta, GA.

 163

Falco, L. D., & Summers, J. J. (2017). Improving career-decision self-efficacy in high

school girls: Evaluation of an intervention. Journal of Career Development,

46(1), 62-76. https://doi.org/10.1177%2F0894845317721651

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses

using G*Power 3.1: Tests for correlation and regression analyses. Behavior

Research Methods, 41, 1149-1160.

Fouad, N. A., & Santana, M. C. (2016, August 5). SCCT and underrepresented

populations in STEM fields: Moving the needle. Journal of Career Assessment,

25(1), 24-39. https://doi.org/10.1177%2F1069072716658324

Goktepe Yildiz, S., & Ozdemir, A. S. (2020). The effects of engineering design processes

on spatial abilities of middle school students. International Journal of Technology

& Design Education, 30(1), 127-148. https://doi.org/10.1007/s10798-018-9491-y

Gottlieb, J. J. (2018, March 30). STEM career aspirations in Black, Hispanic, & White

ninth-grade students. Journal of Research in Science Teaching, 55(10), 1365-

1392. https://doi-org.ezproxy.gardner-webb.edu/10.1002/tea.21456

Gunn, J. (2020). The evolution of STEM and STEAM in the United States. Resilient

Educator. https://resilienteducator.com/classroom-resources/evolution-of-stem-

and-steam-in-the-united-

states/#:~:text=In%202014%2C%20the%20U.S.%20Bureau,million%20between

%202012%20and%202022
 164

Guzey, S. S., Harwell, M., Moreno, M., Peralta, Y., & Moore, T. J. (2017). The impact of

design-based STEM integration curricula on student achievement in engineering,

science, and mathematics. Journal of Science Education and Technology, 26(2),

207–222. https://doi.org/10.1007/s10956-016-9673-x

Hacioglu, Y., & Gulhan, F. (2021). The effects of STEM education on the students’

critical thinking skills and STEM perceptions. Journal of Education in Science,

Environment and Health (JESEH), 7(2), 139-155.

https://doi.org/10.21891/jeseh.771331

Hafiz, N. R. M., & Ayop, S. K. (2019). Engineering design process in STEM education:

A systematic review. International Journal of Academic Research in Business

and Social Sciences, 9(5), 676–697. https://doi.org/10.6007/IJARBSS/v9-i5/5998

Hallinen, J. (2020, July 15). STEM. Encyclopedia Britannica.

https://www.britannica.com /topic/STEM-education

Hardy, M. A. (1993). Regression with dummy variables. SAGE Publications, Inc.

https://dx.doi.org/10.4135/9781412985628

Henry, M. S. (2015, October 8). President signs STEM education bill into law. American

Institute of Physics. https://www.aip.org/fyi/2015/president-signs-stem-education-

bill-law

Holly, J. (2021, June 28). 3 ways schools can improve STEM learning for Black students.

The Conversation. https://theconversation.com/3-ways-schools-can-improve-

stem-learning-for-black-students-156094

Hom, E. (2014, February 11). What is stem education? LiveScience.

https://www.livescience. com /43296-what-is-stem-education.html

 165

Howard, C. (2021, September 27). Studies weekly. https://www.studiesweekly.com/

stem-education/

Inspiring the Next Space Pioneers, Innovators, Researchers, and Explorers (INSPIRE)

Women Act. (2017). Public Law 115–7. https://www.congress.gov/115/

plaws/publ7/PLAW-115publ7.pdf

Jones, K. (2020, January 6). How the STEM crisis is threatening the future of work.

Visual Capitalist. https://www.visualcapitalist.com/stem-education-crisis-future-

work/

Jr. Tech. (2020). What is STEM education? https://juniortech.org/jr-tech-mission/stem-

education/

Kaiser, M. (2019, March 13). Quantitative methods. The Norwegian National Research

Ethics Committees. https://www.forskningsetikk.no/en/resources/the-research-

ethics-library/methods/quantitative-methods/

Kijanka, L. (2009). Exploring the influence of middle school leaders on middle school

girls' interest in high school science course enrollment (Publication No. 3353762)

[Doctoral dissertation, Florida Atlantic University].ProQuest Dissertations &

Theses Global.

Kolk, M. (2018, April 26). Taking a STEM-Based approach to learning. Creative

Educator. https://creativeeducator.tech4learning.com/2018/articles/taking-a-

project-based-approach-to-STEM-learning

Krajcik, J. (2015, January). Project based science. Science Teacher, 82(1), 25-27.

https://my.nsta.org/resource/98362/project-based-science
 166

Laerd Statistics. (2015). Multiple regression using SPSS Statistics. Statistical tutorials

and software guides. https://statistics.laerd.com/premium/spss/mr/multiple-

regression-in-spss.php

Lambert, G. (2018). What are the issues affecting Hispanic middle school girls when

considering a future career in the STEM fields (Publication No. 10812709)

[Doctoral dissertation, Cardinal Stritch University]. ProQuest Dissertations &

Theses Global

Lent, R. W., Brown, S. D., & Hackett. G. (2002). Social cognitive career theory. In

Duane Brown & Associates (Eds.), Career choice & development (4th ed., pp.

255-311). Jossey-Bass.

Li, Y., & Schoenfeld, A. (2019). Problematizing teaching and learning mathematics as

“given” in STEM education. International Journal of STEM Education, 6.

https://doi.org/10.1186/s40594-019-0197-9

Liberatore, M. J., & Wagner, W. P. (2020). Gender, performance, and self-efficacy: A

quasi-experimental field study. Journal of Computer Information Systems,

[online] 2380–2057. https://doi.org/10.1080/08874417.2020.1717397

Lin, S., & Lin, H. (2016). Learning nanotechnology with texts and comics: The impacts

on students of different achievement levels. International Journal of Science

Education, 38(8), 1373–1391. https://doi.org/10.1080/09500693.2016.1191089

 167

Lipsey, M., Puzio, K., Yun, C., Herbert, M. A., Steinka-Fry, K., Cole, M., Roberts, M.,

Anthony, K. S., & Busick, M. (2012, November). Translating the statistical

representation of the effects of education interventions into more readily

interpretable forms. U. S. Department of Education.

https://ies.ed.gov/ncser/pubs/20133000/pdf/20133000.pdf

Little, D. (Academic). (2014). Effect size [Video]. SAGE Research Methods Video

https://dx.doi.org/10.4135/9781473997448

McCombs, S. (2019, September 19). Sampling methods: Types and techniques explained.

Scribbr. https://www.scribbr.com/methodology/sampling-methods/

Merriam-Webster. (n.d.a). Disposition. In Merriam Webb Word Central Student

Dictionary. Retrieved January 15, 2020, from http://www.wordcentral.com/cgi-

bin/student?disposition

Merriam-Webster. (n.d.b). Persistence. In Merriam Webb Word Central Student

Dictionary. Retrieved January 15, 2020, from http://www.wordcentral.com/cgi-

bin/student?peristence

Millennium Project Group. (2017). 15 Global challenges. The Millennium Project.

http://www.millennium-project.org/projects/challenges/

Morgan, D. (2017, January 30). Check linearity between the dependent and dummy coded

variables? ResearchGate. https://www.researchgate.net/post/Check-linearity-

between-the-dependent-and-dummy-coded-variables

Morgan, T. (2015, January 15). The importance of STEM education for middle school

students. (Blog). Teen Life. https://www.teenlife.com/blog/encouraging-an-

interest-in-stem-in-your-middle-school-child/
 168

Mueller, C., Hall, A., & Miro, D. (2015). Testing an adapted model of social cognitive

career theory: Findings and implications for a self-selected, diverse middle-school

sample. Journal of Research in STEM Education, 1(2). https://doi.org/10.51355/

jstem.2015.17

National Research Council. (2011). Successful k-12 STEM education: Identifying

effective approaches in science, technology, engineering, and mathematics.

Committee on Highly Successful Science Programs for K-12 Science Education.

Board on Science Education and Board on Testing and Assessment, Division of

Behavioral and Social Sciences and Education. The National Academies Press.

https://www.ltrr.arizona.edu/~sheppard/TUSD/NRC2011.pdf

National Science Foundation. (2014). Revisiting the STEM workforce: A companion to

science and engineering indicators. https://bit.ly/2ziqQOL

National Science Teachers Association. (2020, February). STEM education teaching and

learning. National Science Teacher Association. https://www.nsta.org/nstas-

official-positions/stem-education-teaching-and-learning

No Child Left Behind Act. (n.d.). Washington Office of Superintendent of Public

Instruction. https://www.k12.wa.us/policy-funding/grants-grant-

management/every-student-succeeds-act-essa-implementation/elementary-and-

secondary-education-act-esea/no-child-left-behind-act-2001#:~:text=The%

20No%20Child%20Left%20Behind,Elementary%20and%20Secondary%20Educ

ation%20Act.&text=Flexibility%3A%20Allows%20school%20districts%20flexib

ility,funds%20to%20improve%20student%20achievement
 169

North Carolina Department of Public Instruction. (2022). End-of-course.

https://www.dpi.nc.gov/districts-schools/testing-and-school-accountability/state-

tests/end-course-eoc.

Office of the Secretary of State. (2009, November 23). President Obama launches

"Educate to Innovate" campaign for excellence in science, technology,

engineering & math (STEM) education. The White House.

https://obamawhitehouse. archives.gov/the-press-office/president-obama-

launches-educate-innovate-campaign-excellence-science-technology-en

Organisation for Economic Cooperation and Development. (2021). 21st-Century readers:

Developing literacy skills in a digital world. PISA. OECD Publishing.

https://doi.org/10.1787/a83d84cb-en

Pearson, S. (2014). The process secondary administrators use to implement twenty-first

century learning skills in secondary schools (Publication No. 3628276) [Doctoral

dissertation, University of Southern California]. Proquest Dissertations & Thesis

Global.

Promoting Women in Entrepreneurship Act. H.R. 4742, Section 3. (2017). House GOP.

https://www.gop.gov/bill/h-r-4742-promoting-women-in-entrepreneurship-act/

Sadler, P. M., Sonnert, G., Hazari, Z., & Tai, R. (2012). Stability and volatility of STEM

career interest in high school: A gender study. Science Education, 96(3), 411-427.

https://doi.org/10.1002S/sce.21007
 170

Sen, C., & Ay, Z., & Kıray, S. (2018, December). Stem skills in 21st-century education.

Research Gate, p. 81-101.

https://www.researchgate.net/publication/332574347_STEM_SKILLS_in_the_21

_ST_CENTURY_EDUCATION

Shetay, N., Lanehart, R. E., Kersaint, G. K., Lee, R. S., Kromrey, J. D. (2016, November

11). STEM pathways: Examining persistence in rigorous math and science course

taking. Journal of Science and Education Technology, 25, 961-975.

http://stelar.edc.org/sites/stelar.edc.org/files/Ashford%20et%20al%202016.pdf

Siekmann, G., & Korbel, P. (2016). Defining “STEM” skills: Review and synthesis of the

literature. National Center for Vocational Education Research.

https://files.eric.ed.gov/fulltext/ED570655.pdf

Smith, B., & Mader, J. (2017). Align your curriculum with the ISTE standards. The

Science Teacher, 84(5), 8. https://my.nsta.org/resource/109836/science-2-0-align-

your-curriculum-with-the-iste-standards

Social Cognitive Career Theory. (2021). Iresearch. http://psychology.iresearchnet.com/

counseling-psychology/career-counseling/social-cognitive-career-theory/

State of the Union Address. (2011). Huffington Post. Retrieved May 15, 2021, from

https://www.huffpost.com/entry/obama-state-of-the-union-_1_n_813478

Stauffer, B. (2020, March 19). What are 21st-century skills? Applied Educational

Systems. https://www.aeseducation.com/blog/what-are-21st-century-skills

STEM & Project-based learning: 6 tips for students. (2021). STEM Village.

https://www.stemvillage.com/blog/stem-and-project-based-learning-6-tips-for-

student-success
 171

STEM School. (2021, February 16). The rich history of stem education in the United

States. https://www.stemschool.com/articles/rich-history-of-stem-education-in-

the-united-states

Student Research Foundation. (2019, January 29). New research on Hispanic students

considering STEM careers. https://www.studentresearchfoundation.org

/blog/hispanic-students-considering-stem-

careers/#:~:text=Hispanic%20Students%20Generally%20Take%20Fewer%20ST

EM%20Courses.%20The,Students%20Are%20Less%20Confident%20in%20their

%20STEM%20Abilities

Syukri, M., Halim, L., Mohtar, L. E., & Soewarno, S. (2018). The impact of the

engineering design process in teaching and learning to enhance students’ science

problem-solving skills. Jurnal Pendidikan IPA Indonesia, 7(1), 66-75.

https://doi.org/10.15294/jpii.v7i1.12297

Temming, M. (2021, April). STEM’s racial, ethnic, and gender gaps still strikingly large.

Science News. https://www.sciencenews.org/article/science-technology-math-

race-ethnicity-gender-diversity-gap

Truesdale, P. (2014). Engineering essentials for STEM instruction. Association for

Curriculum Development & Supervision.

http://www.ascd.org/publications/books/sf114048/chapters/Engineering-and-Its-

Role-in-STEM.aspx
 172

Tsupros, N., Kohler, R., & Hallinen, J. (2008). STEM education in southwestern

Pennsylvania report of a project to identify the missing components. The

Intermediate Unit 1 STEM Education Center & Carnegie Mellon University.

https://www.cmu.edu/gelfand/documents/stem-survey-report-cmu-iu1.pdf

Understanding STEM Skills. (2021). Minnesota State University.

https://careerwise.minnstate.edu/careers/stemskills.html

United States Department of Education. (2006). No child left behind act is working.

https://www2.ed.gov/nclb/overview/importance/nclbworking.html

United States Department of Education. (2016a). Every student succeeds act.

https://www.ed.gov/essa?src=ft

United States Department of Education (2016b, September). STEM 2026: A Vision for

Innovation in STEM Education. https://www.air.org/system/files/downloads/

report/STEM-2026-Vision-for-Innovation-September-2016.pdf

United States Department of Education. (2019). TIMSS 2019 U.S. highlights web report

(NCES 2021-021). Institute of Education Sciences, National Center for Education

Statistics. https://nces.ed.gov/timss/results19/index.asp

United States Department of Education. (2022). Science, technology, engineering, and

math, including computer science. https://www.ed.gov/stem

Van Tuijl, C., & van der Molen, J. H. W. (2016). Study choice and career development in

STEM fields: an overview and integration of the research. International Journal

of Technology & Design Education, (26), 159–183.

https://doi.org/10.1007/s10798-015-9308-1
 173

Verriden, A. W. (2017). Perceptions of belongingness in STEM subjects and career

development for middle school African American girls (Publication No.

10250528) [Doctoral dissertation, University of Wisconsin-Milwaukee]. ProQuest

Dissertations & Theses Global

Williams, T. L. (2020, June 19). Underrepresented minority considered harmful, racist

language. Communications of the ACM. https://cacm.acm.org/ blogs/blog-

cacm/245710-underrepresented-minority-considered-harmful-racist-

language/fulltext

Wilson, D. B., & Lipsey, M. W. (2001). The role of method in treatment effectiveness

research: Evidence from meta-analysis. Psychological Methods, 6, 413–429.

Yaki, A. A., Saat, R. M., Sathasivam, R. V., & Zulnaidi, H. (2019, March 26). Enhancing

science achievement utilizing an integrated STEM approach. Malaysian Journal

of Learning and Instruction, 16(1), 181-205.

https://files.eric.ed.gov/fulltext/EJ1219799.pdf

Yang, D., & Baldwin, S. J. (2020). Using technology to support student learning in an

integrated STEM learning environment. International Journal of Technology in

Education and Science (IJTES), 4(1), 1-11. https://doi.org/10.46328/ijtes.v4i1.22

Yednak, C. (2012, December 6). The lowdown on STEM schools. GreatSchools.

https://www.greatschools.org/gk/articles/what-is-stem-

school/#:~:text=A%20STEM%20specialty%20school%3A%20The,study%20ST

EM%20in%20more%20depth
 174

Appendix A

High School Planning Guide Student Course Requirements

 175

Content For Ninth Graders For Ninth Graders Entering 2017 or later
Area Entering in 2013 or Later
FUTURE-READY OCCUPATIONAL COURSE OF STUDY
CORE
English 4 Credits 4 Credits
English I, II, III, IV English I, II, III, IV
Mathematics 4 Credits 3 Credits
NC Math 1, NC Math 2, Introduction to Mathematics, Algebra 1 (NC
NC Math 3, and a 4th Math Math I), Financial Management
Course to be aligned with
the students' post high
school plans.

In the rare instance a

principal exempts a
student from the FRC math
sequence, the student
would be required to pass
NC Math 1 and NC Math 2
and two other application-
based math courses.

Science 3 Credits 2 Credits

A physical science course, Applied Science, Biology
Biology,
Earth/Environmental
Science
Social 4 Credits 2 Credits
Studies Entering high school Fall Students Entering 9th grade prior to 2017-
2019 and before: World 2018 - American History I AND American
History (or AP World History II
History), American Students Entering 9th grade for the first time
History: Founding in 2017-2018 – American History I or
Principles, Civics & American History II AND American History:
Economics (or Civic Founding Principles, Civics & Economics
Literacy), AND Students Entering 9th grade for the first time
American History I: in 2020-2021 - Founding Principles of the
Founding Principles, United States of America and North Carolina:
American History II (or AP Civic Literacy or American History: Founding
U.S. History & 1 additional Principles, Civics & Economics AND
social studies elective). Economics and Personal Finance
Entering high school Fall
2021: World History (or
 176

AP World
History), Founding
Principles of the US/NC:
Civic Literacy, American
History, (or AP U.S.
History), and Economics
& Personal Finance (EPF).
World 2 Credits are required to Not Required
Language meet Minimum
Application Requirements
for the UNC System.

Health & 1 Credit 1 Credit

Physical Healthful Living I Healthful Living I
Education Successful Completion of Successful Completion of CPR requirement
CPR requirement outlined in NCGS 115C-81.
outlined in NCGS 115C-
81.
Specific 6 Credits Required 6 Credits
Electives 2 elective credits of any Occupational Prep I, II, III, IV
combination from either:
- Career & Technical Completion of Work-Based Hours as follows:
Education (CTE) Students Entering 9th Grade 2014 or later:
- Arts Education 600 Hours
- World Languages School-Based Vocational 150 Hours
4 elective credits strongly Training =
recommended (four Community-Based Vocational 225 Hours
course concentration) Training =
from one of the following:
Competitive Paid Employment = 225 Hours
- Career & Technical
Education (CTE)
- JROTC Students Entering 9th Grade 2013 or earlier:
- Arts Education (e.g., 900 Hours
dance, music, theatre, School-Based Vocational 300 Hours
visual arts) Training =
- Any other subject area Community-Based Vocational 240 Hours
(e.g., mathematics, Training =
science, social studies, Competitive Paid Employment = 360 Hours
English, or cross-
disciplinary) Completion and presentation of a Career
Portfolio containing all the required
components.
Career & 4 Credits
Technical CTE Electives
 177

Education
Additional 4 Credits
Electives
Total 26 Credits 22 Credits
 178

Appendix B

Middle/High School Student Attitudes Toward STEM Survey

179
180
181
182