International Journal of Education in Mathematics, Science and Technology

International Journal of Education in Mathematics, Science and Technology
Volume 6, Number 3, July 2018 ISSN: 2147-611X
6
3
2018
2147-611X
EDITORIAL BOARD
Editor
Mack SHELLEY - Iowa State University, U.S.A.
Guest Editor (Special Issue on Engineering Education)
Şenay PURZER- Purdue University, U.S.A.
Section Editors
Arthur POWELL - Rutgers University, U.S.A. Lina TANKELEVICIENE - Siauliai University, Lithuania
Chun-Yen CHANG - National Taiwan Normal University, Taiwan Sandra ABEGGLEN - London Metropolitan University, U.K.
Jacqueline MCDONNOUGH - Virginia Commonwealth Univ., U.S.A. Robert M. CAPRARO - Texas A&M University, U.S.A.
Editorial Board
Ann D. THOMPSON - Iowa State University, U.S.A Mary B. NAKHLEH - Purdue University, U.S.A.
Bill COBERN - Western Michigan University, U.S.A. Muteb ALQAHTANI - Rutgers University, U.S.A.
Douglas B. CLARK - Vanderbilt University, U.S.A. Ok-Kyeong KIM - Western Michigan University, U.S.A.
Gokhan OZDEMIR - Nigde University, Turkey Pasha ANTONENKO - Oklahoma State University, U.S.A.
Hakan AKCAY - Yildiz Technical University, Turkey Paul ERNEST - University of Exeter, UK
Huseh-Hua CHUANG - National Sun Yat-sen University, Taiwan Pornrat WATTANAKASIWICH - Chiang Mai University, Thailand
Igor M. VERNER - Technion - Israel Institute of Technology, Israel Robert E. YAGER - University of Iowa, U.S.A.
Ilhan VARANK - Yildiz Technical University, Turkey Sanjay SHARMA - Roorkee E&M Technology Institute, India
James M. LAFFEY - University of Missouri, U.S.A. Sinan ERTEN - Hacettepe University, Turkey
Kamisah OSMAN - National University of Malaysia, Malaysia Tsung-Hau JEN - National Taiwan Normal University, Taiwan
Lynne SCHRUM - George Mason University, U.S.A. William F. MCCOMAS - University of Arkansas, U.S.A.
International Journal of Education in Mathematics, Science and Technology (IJEMST)
The International Journal of Education in Mathematics, Science and Technology (IJEMST) is a peer-reviewed scholarly online journal. The IJEMST is
published quarterly in January, April, July and October. The IJEMST welcomes any papers on math education, science education and educational technology
using techniques from and applications in any technical knowledge domain: original theoretical works, literature reviews, research reports, social issues,
psychological issues, curricula, learning environments, research in an educational context, book reviews, and review articles. The articles should be original,
unpublished, and not in consideration for publication elsewhere at the time of submission to the IJEMST. Access to the Journal articles is free to individuals,
libraries and institutions through IJEMST’s website.
Abstracting/ Indexing
The IJEMST is indexed by the following abstracting and indexing services: Ulrich Index, ASOS Index, Journal Seek, JournalRate, Directory of Research
Journals Indexing (DRJI), Infobase Index, ResearchBib, Index Copernicus, TUBITAK ULAKBIM Dergipark, ERIH, Scientific Indexing Service (SIS),
Education Resources Information Center (ERIC), and Emerging Sources Citation Index (ESCI).
Submissions
All submissions should be in electronic (.Doc or .Docx) format. Submissions in PDF and other non-editable formats are not acceptable. Manuscripts can be
submitted through the journal website. All manuscripts should use the latest APA style. The manuscript template for formatting is available on the journal
website.
Contact Info
International Journal of Education in Mathematics, Science and Technology (IJEMST)

Email: ijemst@gmail.com
Web: http://www.ijemst.com
TABLE OF CONTENTS
Culture, Identity, and Motivation in Engineering Education [Editorial Essay] i

Mack Shelley, Senay Purzer
Relationship of Mathematics Self-efficacy and Competence with Behaviors and Attitudes of 200
Engineering Students with Poor Mathematics
Gustavo Morán-Soto, Lisa Benson
An Instrument for Examining Elementary Engineering Student Interests and Attitudes 221
Cathy Pauline Lachapelle, Robert T Brennan
Exploring Academic Performance Paths and Student Learning Strategies in a Large Foundational 241
Engineering Course
Jacob R Grohs, David B Knight, Glenda D Young, Michelle M Soledad
Engineering Identity Development: A Review of the Higher Education Literature 254

Sarah L Rodriguez, Charles Lu, Morgan Bartlett
From Deficit Thinking to Counter Storying: A Narrative Inquiry of Nontraditional Student 266
Experience within Undergraduate Engineering Education
Angela Minichiello
“Just put it together to make no commotion:” Re-imagining Urban Elementary Students’ 285
Participation in Engineering Design Practices
Christopher Wright, Kristen B. Wendell, Patricia P. Paugh
Exploring Moments of Agency for Girls during an Engineering Activity 302

Gina Navoa Svarovsky, Catherine Wagner, Monica E. Cardella
Volume 6, Number 3, 2018 DOI:10.18404/ijemst.428160
Culture, Identity, and Motivation in Engineering Education

Mack Shelley, Şenay Purzer
Article Info Abstract

Article History The collection of papers in this special issue examine student learning and
experiences in the engineering component of STEM (science, technology,
Received: engineering, and mathematics). This editors’ essay provides an overview of each
07 July 2017 of the articles included in this special issue and their contributions to the expanding
research on engineering education. Two main themes emerge: (1) how self-
Accepted:
efficacy beliefs, interest, study strategies, and academic performance are
12 January 2018
manifested when learning engineering, and (2) the experiences of unique groups
in engineering and contexts within which students acquire knowledge about key
Keywords
elements of engineering. Implications are discussed for faculty professional
STEM development; how best to create, evaluate, and use measurement instruments; the
Engineering education use of contemporary methods in both qualitative and quantitative analyses of
Student achievement student learning about and experiences with engineering; and the need to “speak
truth to power” to influence policy decisions about workforce development and to
encourage better- informed student career choices.
Introduction
Engineering impacts every sector of a modern society, ranging from transportation, construction, and
manufacturing to healthcare, energy, and communication. Hence, improving engineering education through
research has benefits beyond those who are influenced directly. The timing of this special issue is also noteworthy
as we approach the year 2020—a critical era for engineering as 2020 is the year that the National Academy of
Engineering in the United States predicted would see a critical turn in the technological and innovation landscape
fueled by globalization (NAE, 2004, 2005). The NAE reports articulate that the engineers of 2020 would need
knowledge, skills, and abilities beyond technical knowledge while highlighting that engineering education is
highly linked to a country’s economic progress and well-being. Yet, many authors report the need for caution in
addressing engineering education given the limited public understanding of the engineering profession and of its
impact on society (Cunningham, Lachapelle, & Lindgren-Streicher, 2005; Davis & Gibbon, 2002; Wulf, 1999),
the declining interest of youth in pursuing careers in engineering (Becker, 2010; NAE, 2008), and lack of diversity
in the engineering workforce (Sadler, Sonnert, Hazari, & Tai, 2012; Sherriff & Binkley, 1997).
One of the motivators of research in engineering education has been the need to diversify the engineering
workforce and attract students into engineering. Another motivator of research is advancing student learning and
enriching their experiences when they finally decide to pursue engineering. However, attracting students into
engineering and diversifying the demographic make-up of the engineering workforce has been an uphill challenge
especially in Western countries such as the United States and Europe. Scholars in more recent studies argue that
achieving equity necessitates the examination of culture, identity, and motivation—not just at the university level
but also at early stages of education as children start to develop agency, efficacy, and disciplinary identities.
Themes of This Special Issue
This special issue brings together a variety of studies and scholars in an effort to exemplify how the disciplinary
cultures shape students’ experiences, how they form their disciplinary identities, and how their self-efficacy and
agency develop as they learn engineering. The seven papers selected for this special issue present diverse vantage
points on the circumstances that influence students. While painting a collective picture of the manifestation of
culture, identity, and motivation in relationship to learning engineering, the papers also center around two distinct
themes: (1) self-efficacy, academic performance, and student study strategies and (2) unique groups and contexts.
ii Shelley & Purzer
Four articles examine the manifestation of student academic performance, self-efficacy, and study strategies, and
explore the development of identity and changes in student interest and attitudes. Three articles in this special
issue examine the educational experiences of non-traditional, often marginalized groups in engineering in a variety
of contexts such as higher education institutions, K-12 classrooms, and museums.
Self-efficacy, Academic Performance, and Student Study Strategies
In “Relationship of Mathematics Self-efficacy and Competence with Behaviors and Attitudes of Engineering
Students with Poor Mathematics Preparation,” Morán-Soto and Benson examine factors that influence
engineering students’ decisions to pursue and complete an engineering major. Their research specifically focuses
on mathematics, which is a gateway subject to high-level engineering coursework. They examined the relationship
between student mathematics self-efficacy and performance in mathematics courses through a mixed-methods
study. They found mismatches between students’ mathematics self-efficacy beliefs and mathematics competence
levels. Interestingly, the discrepancies between the self-efficacy beliefs and competence levels explained students’
study strategies. Students who spent extra time working and seeking extra help reported that their self-efficacy
beliefs are in alignment with their competence levels. In contrast, students who procrastinated and put little effort
into improving their competence reported self-efficacy beliefs that were not in congruence with their competence
levels and externalized their poor performance. Despite varying mathematics competence levels, the interviewed
participants reported relatively high mathematics self-efficacy and reported their persistence to continue taking
mathematics courses required for their major even after failing their first college mathematics course.
Similar to self-efficacy, interest and attitudes are strong predictors of engagement and persistence. In “An
Instrument for Examining Elementary Engineering Student Interests and Attitudes,” Lachapelle and Brennan
argued for the need to develop an instrument designed to assess children’s interest and attitudes in engineering.
Their article presents validity evidence for the Engineering Interest and Attitudes (EIA) survey as a pretest and
posttest to measure changes in student interests and attitudes after participating in engineering activities, programs,
and curricula. Such instruments designed specifically for engineering are necessary to understand and evaluate
interventions and programs regarding their impact on student interest and attitudes.
In “Exploring Academic Performance Paths and Student Learning Strategies in a Large Foundational
Engineering Course,” Grohs, Knight, Young, and Soledad examined learning strategies of second-year
engineering students in a statics course. Cluster analysis identified groups exhibiting distinct performance paths,
and the most important differences across those clusters were found in how students spent time rather than the
total time they spent studying. The strategy of solving problems independently was used significantly more often
by the highest-performing students. Unsuccessful students spent less time in independent problem-solving and
more time solving problems with peers. This study is another example showing that how students study—as
opposed to how much they study or how high their self-efficacy beliefs are—is important in shaping their
academic success.
In “Engineering Identity Development: A Review of the Higher Education Literature,” Rodriguez, Lu, and Barlette
analyzed the body of research on identity development in engineering and STEM. They identified 88 articles
focused on engineering identity development in higher education. While research in engineering identity has
increased over the last decade, it has been conducted primarily with qualitative methods and has focused on the
learning contexts and experiences of women and underrepresented racial/ethnic minorities with less focus on men,
international, or graduate students. The authors concluded that research on undergraduate engineering identity
development should expand beyond qualitative methods and study a variety of academic communities such as
mathematics, science, and engineering education. This extensive review of the literature reminds us that studying
the identity development of a variety of groups within an institution including majority groups is as important as
studying the experiences of minority groups based on gender and ethnicity to form a comprehensive view of
minority students’ experiences.
Unique Groups and Contexts
When it comes to identity and agency development, it is also important to examine different agents at work in
different educational contexts. The next three articles in this special issue examine such unique groups and
contexts such as students who transition from two-year colleges to four-year universities, elementary students
learning engineering in urban schools, and young girls playing with their parents when learning engineering, with
critical insights suggesting the existence of bias even when all participants in these situations are well-intended.
Int J Educ Math Sci Technol iii
In “From Deficit Thinking to Counter Storying: A Narrative Inquiry of Nontraditional Student Experience within
Undergraduate Engineering Education,” Minichiello focused on the experiences of nontraditional undergraduate
engineering students through narrative inquiry research. The research participants were enrolled in a two-year
engineering transfer program offered in the evenings, via synchronous broadcast distance instruction. These
students first attended two-year programs at regional campuses near their local communities, and later transitioned
to the four-year university campus located in another city away from these communities. Their study presents
personal, social, and institutional tensions arising from instances of thinking of themselves as deficient and
underprepared through a narrative inquiry of the participants’ lived and told stories. Despite deep, personal
reactions to the bias they experienced, participants overcame perceptions of personal deficiency to pursue and
achieve their goal of becoming engineers. Moreover, participant counter stories further revealed ways in which
their unique life experiences alternately served to enhance and deepen their engineering education.
In “Just Put It Together to Make No Commotion: Re-imagining Urban Elementary Students’ Participation in
Engineering Design Practices,” Wright, Wendell, and Paugh examined cultural norms and teacher expectations
in an elementary classroom. Their findings suggest that students conceptualized urban engineering learning
environments as spaces for risk management. This notion of managing risks informed the students’ participation
in collaborative decision-making and in the ways in which they viewed themselves as doers of engineering.
However, this form of risk-averse participation created a tension between disciplinary practices expected in
engineering design and the teacher’s need to manage a classroom in an urban school. The authors argue for the
need to develop methodologies and frameworks that provide opportunities to uncover these conflicts and the
potential risks that students need to take as they participate in engineering design practices.
In "Exploring Moments of Agency for Girls During an Engineering Activity,” Svarovsky, Wagner, and Cardella
examine parent-child interactions, specifically focused on moments when girls express agency during an
engineering design process. They followed these interactions during a museum exhibit that engages visitors in
engineering design activity. In their study, children expressed agency and led interactions by directing, proposing
design ideas, and asking questions. However, young women tended to direct their mothers more than they directed
their fathers, suggesting differences in parent-child interactions should be considered to support girls’ agency in
engineering activities.
Summary
In summary, this special issue brings together seven articles that provide a wide range of conceptual,
methodological, and situational perspectives on how culture, identity, and motivation intersect with learning and
engagement. Together, these papers argue for two critical points. First, these papers collectively suggest that
beyond students’ cognitive abilities, many other considerations—such as culture, identity, and motivation—play
important roles in the education of students. In addition, these papers show that learning occurs in many places:
higher education institutions, K-12 classrooms, and child-parent interactions in informal settings such as
museums. Second, the papers in this special issue illustrate mismatches between student abilities and perceptions,
between student abilities and educator expectations, and between individual identities and institutional identities.
A limited understanding of such discrepancies had perhaps been one of the reasons for slow progress in increasing
student interest and success in engineering.
Conclusions
A number of lessons learned can be gleaned from this research. These articles suggest that it is imprudent to lump
students into an undifferentiated mass summarized by just a set of mean scores. Instead, it is essential to examine
sources of variation in student performance that are based on a mix of demographics, motivational differences,
and other key variables that distinguish one student’s foundation for learning engineering principles from that of
another student. That perspective provides part of the agenda for future research in the area of engineering
education specifically and STEM education at all levels more generally.
A related conclusion regarding future research is that we will need to develop and apply measurement instruments
and research methods that are sufficiently sensitive to handle the magnitude and full range of individual variation,
to provide more valid and useful inferences about students’ performance and experiences. Both qualitative and
quantitative research methods as well as their mixing are appropriate for teasing out the specifics of individuation
within the broader context of systemic and structural differences in how students learn and differences in the
iv Shelley & Purzer
resources available to their families, instructors, and institutions. On the quantitative side, multilevel (mixed)
models provide a well-developed approach that nests individual student traits within higher-level metrics.
Qualitative methods are essential, particularly in the absence of the extensive and detailed data needed about
student individual characteristics and institutional characteristics for multilevel statistical modeling, but more
importantly to provide the rich context for understanding how students understand the process of learning and
how best to improve outcomes of student success, equity, and workforce development.
The implications of this work for teaching will need to be embedded in teacher and faculty professional
development to promote increased awareness of educators in their interactions and expectations based on cultural
norms and gender roles. In addition, the education and training of education administrators, recruitment staff,
academic counselors, and parents must recognize that efforts would have limited success if data are lumped
without a recognition of variation among students. Professional development efforts can be informed by enhanced
knowledge of what motivates students and how individual students react differently to the same instructional
activities. Administrators could benefit from knowing what mix of formal and informal education may be optimal
for students’ ability to process and comprehend the complexities of engineering design and analysis. The process
of “speaking truth to power” is essential to inform those who establish education policies and funding levels that
they need to be mindful of the compelling lessons learned from studies such as these about how to encourage
educators, students, and parents to ensure the best possible circumstances for developing a deep understanding of
what engineering education contributes to societal growth and a more equitable future. Students must be provided
with the resources and support mechanisms that will make it possible for them to succeed at as high a level as
possible and thereby become more productive contributing members of their chosen profession.
References
Becker, F. S. (2010). Why don’t young people want to become engineers? Rational reasons for disappointing
decisions. European Journal of Engineering Education, 35(4), 349–366.
Cunningham, C. M., Lachapelle, C., & Lindgren-Streicher, A. (2005). Assessing elementary school students’
conceptions of engineering and technology. Proceedings of the American Society of Engineering
Education. Portland, OR.
Davis, L. A., & Gibbon, R. D. (2002). Raising public awareness of engineering. Washington, DC: National
Academies Press.
National Academy of Engineering (2004). The engineer of 2020: Visions of engineering in the new century.
Washington, DC: National Academies Press.
National Academy of Engineering (2005). Educating the engineer of 2020: Adopting engineering education to
the new century. Washington, DC: National Academies Press.
National Academy of Engineering (2008). Changing the conversation: Messages for improving public
understanding of engineering. Washington, DC: National Academies Press.
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. Retrieved from 10.1002/sce.21007
Sherriff, B. L., & Binkley, L. (1997). The irreconcilable images of women, science, and engineering: A Manitoban
program that is shattering the stereotypes. Journal of Women and Minorities in Science and Engineering,
3, 21-36.
Wulf, W. (1999). The image of engineering. Issues in Science and Technology online. Issues in Science and
Technology, 15(2). http://issues.org/15-2/wulf-2/
Author Information
Mack Shelley Şenay Purzer
Iowa State University Purdue University
Ames, Iowa West Lafayette, Indiana
U.S.A. U.S.A.
Contact e-mail: purzer@purdue.edu
Relationship of Mathematics Self-efficacy and Competence with Behaviors

and Attitudes of Engineering Students with Poor Mathematics Preparation
Gustavo Morán-Soto, Lisa Benson

Article History This mixed methods research study focuses on two relevant factors in students’
decisions to pursue and complete an engineering major: mathematics
Received: preparation and mathematics self-efficacy. This study describes the relationship
07 July 2017 of mathematics self-efficacy on engineering students’ performance, behavior,
and attitudes in their first college mathematics courses. Participants completed a
Accepted:
mathematics self-efficacy survey (n=408); 11 were selected to be interviewed.
12 January 2018
A grounded theory approach was used to examine participants’ behaviors and
attitudes in college mathematics courses, their mathematics self-efficacy
Keywords
beliefs, and how these beliefs aligned with their mathematics competence.
Mathematics self-efficacy Interview participants reported relatively high mathematics self-efficacy, but
Engineering many revealed a mismatch between their mathematics self-efficacy beliefs and
Mixed methods mathematics competence levels. Participants who had a balance between these
two factors reported being more likely to spend extra time working to overcome
their mathematics deficiencies and seeking extra help. However, participants
with a mismatch reported being more likely to procrastinate and put little effort
into improving their mathematics competence, blaming external factors for their
struggles. Despite showing different behaviors and attitudes, all participants
reported being likely to continue taking mathematics courses required for their
major even after failing their first college mathematics course.
Introduction
Mathematics Self-Efficacy
Self-efficacy beliefs have become central to research into motivational factors that could affect students’ interest
in pursuing science, technology, engineering, or mathematics (STEM) majors. Findings of these studies show
that students who felt that they had good mathematics preparation in their precollege academic experiences,
which helped them to develop a high mathematics self-efficacy, were more likely to show interest in pursuing a
mathematics-related major like engineering (Hackett, 1985; Lent et al., 1991; Lent, Brown, & Hackett, 1994).
The importance of understanding engineering students’ mathematics self-efficacy levels can be seen in the
different reactions that people normally have when they have low or high self-efficacy for performing specific
mathematics tasks. Low mathematics self-efficacy levels have been shown to have an influence on students’
decisions to avoid mathematics-related activities that may lead them to have feelings of stress and anxiety
(Cooper & Robinson, 1991). Conversely, the literature suggests that after experiencing success in mathematics
activities and seeking assistance when they are uncertain, students’ mathematics self-efficacy may increase
(Bandura, 1986).
This relationship between mathematics self-efficacy and performance has been shown to work both ways, with
performance affecting mathematics self-efficacy and with mathematics self-efficacy modifying choices,
performance, and persistence in specific mathematics tasks (Williams & Williams, 2010). The influence of
mathematics self-efficacy on students’ performance in mathematics courses and activities is well-documented,
and has been confirmed in different contexts, cultures, and populations (Cooper & Robinson, 1991). The
consistency in the relevance of mathematics self-efficacy on not only students’ performance but also on their
choices and persistence when they are struggling to understand mathematics topics, is a good indicator of the
importance of this motivational factor in students’ mathematics preparation. Prior research has shown that
students’ mathematics self-efficacy was lower for students leaving STEM majors, and this factor was more
significant for students leaving college during their first semesters (Eris et al., 2010).
Int J Educ Math Sci Technol 201
The relationship of self-efficacy beliefs with performance could be interactive, with performance affecting self-
efficacy and vice versa (Bandura, 1980), and there may be some discrepancies between these two factors during
the students’ mathematics learning process (Moran & Benson, 2016). These discrepancies may change and
evolve according to students’ experiences; ideally, students would have a more equilibrated and realistic
perception of their mathematics abilities after having enough experience taking challenging mathematics
courses (Redmond et al., 2007).
Motivation for this Study
Engineers are important to the economic development and growth of a country (Committee on STEM
Edcucation, 2013). Without enough engineers having the necessary skills to design and create new technological
solutions to current society needs and problems, a country will face setbacks in its development (National
Academy of Engineering, 2015). In spite of the clear importance of training engineers, the projected demands
for professionals in STEM fields for the next decade surpasses the projected supply of trained STEM
professionals in some countries like the U.S. (Committee on STEM Edcucation, 2013; Snyder & Dillow, 2010).
The need for professionals in engineering fields is constantly growing in the biggest economies of the European
Union, which currently employ 640,000 to 1.2 million engineering workers in countries like the United
Kingdom and Germany, respectively (Erdmann & Schumann, 2010). If countries such as these decrease their
current engineering students’ graduation rates, they could face a shortage of qualified professionals to fill new
emerging engineering jobs.
Understanding how to keep students in their engineering degree programs is of importance to engineering
educators around the globe (Chen & Soldner, 2013). The lack of adequate mathematics background for some
students who want to become engineers is a great challenge (Kokkelenberg & Sinha, 2010). Once students
decide to pursue an engineering major, mathematics competence plays a key role in their academic success.
However, there is much variability in the level of high school mathematics preparation for students entering
college (Casey, Nuttall, & Pezaris, 1997; Porter, 2011). Once enrolled, mathematics courses like calculus I and
II have been identified as barrier courses for engineering students, and struggling to complete these courses
could discourage students trying to get an engineering degree (Seymour & Hewitt, 1997; Suresh, 2006). The
quality of the mathematics education that some students receive before college could affect their mathematics
competence development, and some populations such as under-served groups, minorities, and first generation
students are more likely to have some deficiencies in their mathematics preparation due to the lack of
opportunities and resources available to them (Flores, 2007; Lee, 2012; May & Chubin, 2003).
Purpose
The purpose of this research is to deepen our understanding of mathematics self-efficacy beliefs for engineering
students with poor mathematics preparation, and how these beliefs relate to academic performance, behaviors,
and attitudes in their first mathematics course. This research is designed to help inform efforts to decrease
engineering attrition, especially for students who experience difficulties with mathematics courses due to a poor
mathematical high school background that is reflected in low mathematical competence. Here we are using
academic performance as an indicator of competence. Although there is much that is known in general about the
relationship between students’ mathematics self-efficacy and their performance in college mathematics courses
(Hackett & Betz, 1989; Siegel, Galassi, & Ware, 1985; Williams & Williams, 2010), there is little research
analyzing the specific case where engineering students start their college mathematics courses with an
overestimation of their mathematics self-efficacy. According to mathematics self-efficacy literature, having high
mathematics self-efficacy can help engineering students build confidence in their ability to successfully
complete a mathematics-related major such as engineering (Hackett, 1985; Lent et al., 1991). However, there is
a lack of research analyzing engineering students’ reactions to poor performance and struggles understanding
the material in mathematics courses due to a poor mathematics preparation despite reporting high mathematics
self-efficacy levels. The mismatch between mathematics self-efficacy and competence could negatively affect
students choice of behavior and attitudes in mathematics courses (Vancouver & Kendall, 2006), creating a sense
of overconfidence in students with higher mathematics self-efficacy beliefs than competence, or generating
feelings of anxiety and the need to avoid mathematics related activities in students with lower mathematics self-
efficacy levels (Bandura, 1977, 1980).
202 Morán-Soto & Benson
Engineering students’ first experience taking a mathematics course in college could influence their perceptions
of their mathematics abilities, and their performance and the mathematics course level have shown to be
significant predictors of student retention in engineering majors (Gardner, Pyke, Belcheir, & Schrader, 2007;
Middleton et al., 2015; Van Dyken, Benson, & Gerard, 2015). If mathematics educators and professors are more
familiar with engineering students’ possible reactions to mismatches between their mathematics self-efficacy
and competence, then these professors could better prepare to advise students to avoid negative behaviors that
may lead them to failure experiences in their first college mathematics courses.
This research was guided by the following research questions:
R.Q. 1 What relationship exists between mathematics self-efficacy of first-year engineering students with
poor mathematics preparation and their performance, behavior, and attitudes in their first college
mathematics course?
R.Q. 2 What are student behaviors and attitudes in mathematics courses when there is a mismatch between
mathematics self-efficacy and competence (low mathematics self-efficacy and high competence; high
mathematics self-efficacy and low competence)?
R.Q. 3 What are student behaviors and attitudes in mathematics courses when their mathematics self-
efficacy and competence knowledge levels are well matched?
R.Q. 4 What are the effects of a mismatch or a match between mathematics self-efficacy and competence on
students’ confidence in completing the mathematics courses required for an engineering degree?
For the purpose of this study, student behaviors are defined as the decisions that students make regarding time
expended trying to understand their class material, time expended working on their homework assignments,
their reaction if they struggle with specific topics, and their decisions to take advantage of or ignore the extra
resources available for the class (software, teaching assistants, supplemental instructors, office hours, tutors,
etc.). Student attitudes refer to their feelings about mathematics in general and about their current mathematics
course.
Despite the importance of mathematics self-efficacy and its influence on engineering students’ choices of
behavior and performance, the interactions between poor mathematics preparation and mathematics self-
efficacy for engineering students have not been studied before. This study will expand current literature by
analyzing the different behaviors and attitudes of engineering students with poor mathematics preparation based
on their level of mathematics self-efficacy and how well these mathematics self-efficacy beliefs match their
actual competence.
Differences in students’ mathematics self-efficacy could explain why students with similar cognitive abilities
perform at different levels; this difference in performance could be due to differences in the effort that students
exerted when they struggled or faced difficulties performing mathematics (Lent et al., 1994). Finding a way to
keep engineering students motivated to persist and continue taking mathematics courses despite facing struggles
could help engineering educators to address the high attrition rates in engineering majors (Geisinger & Raman,
2013), especially for under-served groups and minorities that are more likely to quit if they face obstacles due to
their poor mathematics preparation (Flores, 2007; Lent, Brown, & Hackett, 2000).
Theoretical Frameworks
The concept of self-efficacy emerged in the late 1970’s when Bandura (1977) suggested that beliefs in one’s
abilities to successfully perform a specific task could be a major determinant in one’s decisions to attempt that
task, and could also affect the effort that one is willing to expend to complete that task when faced with
struggles. According to Bandura (1986), self-efficacy refers to “people’s judgments of their capabilities to
organize and execute courses of action required to attain designated types of performances” (p. 391). Findings
of early research in self-efficacy positioned it as an important motivational factor for achievement that may
influence students’ interest in pursuing specific majors and persisting in their desire to complete that major
despite struggling and facing obstacles in the process (Bandura, 1986; Hackett & Betz, 1989; Lent, Lopez, &
Bieschke, 1991; Multon, Brown, & Lent, 1991).
Developed by Bandura (1986), Social Cognitive Theory highlights cognitive, vicarious, self-regulatory, and
self-reflective processes that modify people’s behavior and choices. Within Bandura´s social cognitive
constructs, self-efficacy has a key role predicting one´s choice of activities, effort expenditure, persistence, and
emotional reactions when facing difficulties performing any particular task. These beliefs are task-specific and
one’s skills in a particular task may influence only the individual performance on that task (Bandura, 1986). It is
important to understand that self-efficacy beliefs could be related to objectively assessed skills, but these two
factors could also be very different depending on one’s experiences and circumstances.
Bandura’s theory hypothesizes that self-efficacy beliefs that slightly exceed one’s current skill level could
encourage people to try challenging activities that may promote better skill development. On the other hand,
self-efficacy beliefs that grossly exceed or underestimate current skill level could lead to maladaptive behavior;
a discrepancy with high self-efficacy without good skills that support these beliefs could lead to failure or poor
performance in challenging activities, while pessimistic beliefs or lack of confidence in well-developed skills
may prompt avoidance of activities that are within one’s competence range.
Research Design
A sequential explanatory mixed methods design was selected to conduct this research (see Figure 1) (Creswell
& Clark, 2011). Although the quantitative phase was performed first, there was an emphasis on the qualitative
phase of this mixed methods design. To get a thorough description of students’ experiences taking college
mathematics courses, a constructive grounded theory methodology was followed in the qualitative phase of this
research to develop a theory based on the data (Charmaz, 2006). The sequential explanatory design allowed the
researchers to elaborate and expand the conclusions about the participants’ mathematics self-efficacy levels
reported during the quantitative results. By comparing the quantitative results with the qualitative findings, the
researchers were able to find and further explain possible relationships between students’ mathematics self-
efficacy levels and their chose of behavior and attitudes in mathematics courses.
Inputs Outcomes
MSE surveys n = 408 Classification of mathematics self-
Quantitative data collection
Course performance efficacy and mathematics competence
(clusters)
Descriptive statistics Quantitative analysis Next participant

Cluster analysis according to theory
Theoretical selection
Qualitative results Mixing Maximal variation within mathematics
self-efficacy and competence
Interviews n = 11 Qualitative data collection
Coding Memos, codes, and categories

Qualitative analysis
Comparative method Theory grounded in the data
Deep understanding of students’

Qualitative and Mixing behaviors and attitudes according to
quantitative findings their mathematics self-efficacy and
mathematics competence.
Figure 1. Inputs and outcomes of the sequential explanatory mixed methods design
A grounded theory approach was selected for the qualitative phase of this study due to the lack of literature and
prior studies about mathematics self-efficacy of engineering students with poor mathematics preparation
(Corbin & Strauss, 2008). Developing a theory based on students’ experiences is an important step for
engineering and mathematics educators seeking to decrease engineering attrition caused by poor mathematics
preparation. Designing mathematics courses and possible interventions to help engineering students understand
how their mathematics self-efficacy beliefs might influence their behavior and performance in mathematics
courses if these beliefs are not aligned with their mathematics abilities could help students identify and avoid
struggles and failing experiences in mathematics courses.
Participants
Participants for this study were selected based on the mathematics course in which they were enrolled during
their first year at a southeastern United States university. Students who have the lowest scores on the university
mathematics placement test are placed in a precalculus class (“Precalc”). Those scoring slightly higher on the
placement test but who are not ready to take a full semester of calculus as their first college mathematics course
are placed in a two-semester course combining precalculus and calculus. The first semester course in this series
is referred to as “Long Calc,” and the following course is “Long Calc II.” Both Precalc and Long Calc are
considered non-college-level mathematics courses and are designed to help prepare students with poor
mathematics backgrounds to succeed in calculus courses.
Students from all majors enrolled in Precalc and Long Calc for the fall and spring semesters over the course of
one academic year were stratified (Onwuegbuzie & Collins, 2007), aiming to collect information about students
with poor mathematics preparation in their first college mathematics course. A total of 408 students participated
in the quantitative phase of this study: 101 students in three sections of Precalc and 207 students in 5 sections of
Long Calc in the fall semester; and 20 students in one section of Precalc and 80 students in 3 sections of Long
Calc in the following spring semester. Six participants were randomly selected for interviews from the Precalc
and Long Calc courses, and these interviews were analyzed, before a theoretical sampling could be conducted,
in keeping with a grounded theory approach (Bryant & Charmaz, 2007). Preliminary findings from these six
participants formed the baseline of the emerging theory that helped refine the research questions and modify the
study’s research design. This method of participant selection is in line with the first steps of a grounded theory
approach, where there is not enough data to perform theoretical selection (Bryant & Charmaz, 2007).
Preliminary findings of this first stage of the grounded theory approach indicated the need for a more accurate
measurement of participants’ mathematics self-efficacy beliefs for comparison with their mathematics
competence and to enable theoretical selection based on the mathematics self-efficacy level.
Following a sequential data collection design (Onwuegbuzie & Collins, 2007), five additional participants were
purposely selected from the previously stratified sample following the guidelines determined by the theoretical
sampling after the analysis of quantitative survey data from prior participants. These five participants’ data was
added to the previous six participants’ (n = 11) data to be analyzed following a grounded theory approach. The
last five participants of the qualitative phase of this study were selected with the goal of reaching maximum
variation within students’ mathematics self-efficacy beliefs and competence levels to develop a thorough theory
(Teddlie & Yu, 2007).
Quantitative Phase Methods
Data Collection
This research used the Mathematic Self-Efficacy Survey (MSES), developed by Betz and Hackett (1983), to
measure participants’ mathematics self-efficacy for the theoretical selection. The MSES consists of 52 items
measuring three different mathematics self-efficacy subscales: mathematics problem-solving (18 items),
everyday mathematics tasks (18 items), and mathematics courses (16 items). Participants rated their level of
confidence performing different mathematics-related activities for each question on a scale ranging from no
confidence at all (0) to complete confidence (10). The survey was selected for this study due to its high
reliability in measuring mathematics self-efficacy in prior studies (e.g. Brown & Burnham, 2012). Both the full-
scale original survey and its subscales independently have been examined previously with coefficient alpha
values ranging from 0.90 to 0.96 (Brown & Burnham, 2012; Kranzler & Pajares, 1997; Lent, Lopez, &
Bieschke, 1991; Pajares & Miller, 1995). These Cronbach’s alpha values demonstrated the internal consistency
reliability of the MSES for different populations and contexts, providing confidence that the survey could
accurately measure engineering students’ mathematics self-efficacy for this research population. Some MSES
items were adapted to the context and population of this study aiming to get a self-efficacy measurement that
could be related to participants’ performance in these specific courses (Bandura, 1986). For example, an item
asking students about their level of confidence working with a slide rule was modified to include a scientific
calculator (CASIO fx-350MS) instead of the outdated slide rule. These changes helped participants to interpret
the survey questions and relate these questions to the mathematics activities that they perform in their daily
lives.
Students from all majors at the university in the sections of mathematics courses in which instructors agreed to
distribute the MSES completed the survey during class time. Instructors distributed and collected the paper-
based surveys at the end of their class, and they returned them immediately to the researcher for data analysis.
The six participants who were interviewed before the mathematics self-efficacy survey was used for theoretical
sampling purposes did not complete the survey; these six participants were missing this quantitative
mathematics self-efficacy measurement.
Participants’ mathematics self-efficacy level was determined using both survey and interview responses. An
average of participants’ interview responses about their mathematics self-efficacy for different mathematics
related activities from 1 to 10 (e.g., How confident do you feel that you can solve your homework after being
taught a difficult topic? Could you select your level of confidence in a scale between 1 (not at all) and 10 (very
confident)) were used as their mathematics self-efficacy level for the mixed analysis to be compared with
participants’ mathematics competence. All participants provided a mathematics self-efficacy estimate as part of
their interview; the number of interview questions that participants answered to get that mathematics self-
efficacy estimate changed from interview to interview because of the evolving nature of the grounded theory
methodology. The seventh and all subsequent interview participants completed the MSES to classify their
mathematics self-efficacy beliefs before the interview. Results of the MSES were used to purposely select the
most appropriate participants to complement and further develop the emerging theory. Only the last five
participants answered the MSES, and these survey averages were used for the purpose of theoretical selection
only.
Participants’ mathematics competence was estimated based on their performance in their college mathematics
course at the time of the interview. Participants’ performance information was obtained from the online course
software used to teach the course for the Precalc students, and from the course test scores for the Long Calc
students. Some grades were not final due to the time of the semester in which the interview was conducted, so
interim grades were used. The competence scores were on a scale from 1 to 100, so the mathematics
competence estimation was calculated by dividing the interim grades by 10 to be compared with participants’
mathematics self-efficacy beliefs for the mixed analysis of the data.
Data Analysis
The data collected with the MSES (n=408) were analyzed using descriptive statistics and cluster analysis to
classify participants’ mathematics self-efficacy level using the statistical software R (Team, 2012). A k-means
cluster analysis (Maechler, Rousseeuw, Struyf, Hubert, & Hornik, 2015) was conducted to classify participants
in three homogenous subgroups based on their mathematics self-efficacy levels for each semester that the MSES
was used to collect data. K-means is a partitioning type of clustering method and is appropriate for cases in
which the number of clusters can be assumed based on knowledge of underlying theory and prior results (Ehlert,
Faber, Benson, & Kennedy, 2017). The three clusters identified through the k-means cluster analysis classified
participants’ mathematics self-efficacy beliefs as medium-low, medium-high, and high according to their
average MSES scores. These clusters were used to purposefully select individuals with the appropriate
characteristics for the qualitative phase interviews, aiming to get a maximum variation within mathematics self-
efficacy beliefs. Participants’ mathematics competence was classified as high or low according to course
performance data.
An additional mathematics self-efficacy average was calculated using the data from the participants’ interview
responses in which they rated their confidence in performing mathematics related activities. This mathematics
self-efficacy average was used for the mixed analysis of the reasons behind participants’ behaviors and
performance within different mathematics activities. All eleven interviewed participants reported a mathematics
self-efficacy average from their interviews, and the last five participants were selected based on the MSES
results and emerging theory. The difference between these two mathematics self-efficacy measurements is that
interview data was used as the participants’ mathematics self-efficacy average during the mixing phase, and the
MSES data were used for theoretical selection purposes only.
Quality Considerations
The adapted version of the MSES was validated using Cronbach’s alpha values to evaluate the internal
consistency reliability of the survey items for this specific population (Thorndike & Thorndike-Christ, 2010).
All MSES constructs were expected to have a Cronbach’s alpha value above 0.7 (Nunnally & Bernstein, 1994),
showing that they are measuring similar theoretical concepts. Content and face validity were also evaluated by
asking qualitative experts and mathematics instructors at the university about the clarity and possible students’
interpretation of the survey items. This feedback guided modifications and rewording of some survey items to
ensure the items were understandable for participants and were all measuring the same construct in a thorough
way (Creswell, 2009).
Qualitative Phase Methods
Data Collection
The data for the qualitative phase of this research were collected through semi-structured interviews, including
questions about participants’ reasons for choosing engineering, persistence in their desire to become engineers,
mathematics self-efficacy based on Bandura’s guide for constructing self-efficacy scales (Bandura, 1986), and
social supports (Lent et al., 1994). During the interview, participants were asked to rate their self-efficacy on a
scale of 1-10 for specific mathematics skills, such as solving word problems and finding errors in their
calculations. The interview questions related to mathematics self-efficacy were designed to let the participants
express their feelings and rate their level of confidence performing mathematics-related activities. The questions
used the phrase “level of confidence” instead of mathematics self-efficacy to avoid misunderstandings with
respect to the term “self-efficacy.” Self-efficacy interview questions focused on challenging mathematics
activities and common assignments that students normally face in their college mathematics course, with
questions such as: “How confident do you feel that you can do the most challenging mathematics exercises of
your class if you have all the time and resources you need to work to your satisfaction? How is your level of
confidence in your mathematics abilities when you are solving mathematics problems in class compared to
solving mathematics problems on a test?” At the end of every mathematics self-efficacy question, the
participants were asked to rate their mathematics self-efficacy using a scale suggested by Bandura (1986)
ranging from 1 to 10, with the following question: “What is your level of confidence for this question on a scale
from 1(not at all) and 10 (very confident)?”
The interview protocol was modified as needed at certain points during the qualitative data collection to expand
participants’ responses relative to previous participants. The last five interview participants were purposefully
selected according to the quantitative phase results, with the goal of maximum variation in mathematics self-
efficacy. Each participant was selected after analyzing and coding the data from the previous participant;
theoretical selection was performed based on the emerging codes and categories from previous participants.
Eleven participants were interviewed between the Fall 2015 and Spring 2016 semesters.
Data Analysis
Following a constructivist grounded theory approach, the qualitative data was initially coded line-by-line using
open coding (Charmaz, 2006). Codes emerged directly from the data, and they were named using gerunds and
participants’ words (in vivo coding) during the interview, in order to develop a theory that could accurately
represent participants’ experiences and actions in their college mathematics courses (Bryant & Charmaz, 2007).
The first interview was coded and analyzed before conducting the next interview (Bryant & Charmaz, 2007),
and this process continued for all subsequent interviews. The codes, memos, and responses from the interviews
analysis were used to modify the interview protocol according to emerging themes and missing information
before conducting the next interview.
This coding philosophy facilitated the theoretical interpretation of the data based on what was reported by the
participants. The initial codes were constantly analyzed and compared between participants. During this analysis
process, memos were written as notes that described the researcher’s interpretation of the emerging codes
(Charmaz, 2007). These memos helped the researcher develop categories based on the constant comparison of
codes, memos, and data, allowing the categories to emerge as the more representative actions of engineering
students with poor mathematics preparation in college mathematics courses. After a constant comparison of
codes, memos, categories, and data from different participants, the final categories became the core of a theory
grounded in the data that described mathematics self-efficacy perceptions of engineering students with poor
mathematics preparation, and how their self-efficacy beliefs are related to their behavior and attitudes in college
mathematics classes. Preliminary findings of this qualitative analysis were used together with the quantitative
results to inform the theoretical sampling. This theoretical sampling and continual refining of the interview
protocol were part of a cycle for selecting the next, most appropriate participant.
According to Johnson (1997), there are five types of validity that need to be considered to validate qualitative
studies: descriptive validity, interpretive validity, theoretical validity, internal validity, and external validity.
Three of these five types of validity were applied in this study. Following Johnson’s validation framework an
additional person was present in the interviews to take notes and help triangulate and discuss the data after the
interviews, which provided descriptive validity. Additionally, participants’ feedback was used to demonstrate
interpretive validity and accuracy of the conclusions. Theoretical sampling to purposefully select specific types
of participants, which enriched and complemented the emerging theory, provided theoretical validity.
Mixed Phase Methods
Data Analysis
The mixing between the quantitative and the qualitative phases occurred in two different places in this research.
The first mixing phase took place during data collection. During this mixing phase the quantitative results
informed the participant selection for the qualitative phase to perform theoretical sampling accurately, using the
MSES results to purposefully select the most adequate characteristics of the subsequent participants for the
qualitative phase (Sandelowski, Voils, & Knafl, 2009). Additionally, after finalizing both quantitative and
qualitative phases, the findings and data of these two phases were analyzed together, aiming to expand our
understanding of the possible influences of mathematics self-efficacy on engineering students and answer the
research questions. This final analysis thoroughly integrated the two methods, helping the researcher
consolidate, compare, and integrate into a coherent whole both qualitative and quantitative findings
(Onwuegbuzie & Combs, 2010). To accomplish this, survey data (clusters) and interview data (categories,
codes, diagrams) were compared side by side for each participant, aiming to triangulate the data to compare and
expand the findings of each phase (Moran-Ellis, 2006). This second mixing phase helped the researcher to
classify participants in different groups depending on their mathematics self-efficacy beliefs and mathematics
competence relationship. If participants showed a gap larger than 1 point between their mathematics self-
efficacy beliefs assessment (between 1 and 10) and their mathematics competence assessment (between 1 and
100 divided by 10), then these participants were considered having a mismatch between these two factors; if the
gap between these two factors was lower than 1 point, then these participants were considered having a match
between their mathematics self-efficacy and competence (see Table 2).
This study used the Onwuebuzie and Johnson legitimation framework (2006) for mixed methods quality
considerations. For this framework, nine legitimation types are considered to improve the quality of mixed
methods studies, including: (1) sample integration, (2) inside-outside, (3) weakness minimization, (4)
sequential, (5) conversion, (6) paradigmatic mixing, (7) commensurability, (8) multiple validities, and (9)
political. Five of these nine legitimation types were applied in this study. The sample for the qualitative phase
was selected from a subset of the quantitative sample by using the quantitative results to inform the theoretical
sampling for the qualitative phase, which provided sample integration and sequential legitimization. The
relationship between participants’ behaviors and attitudes and their mathematics self-efficacy level were
described using the richness of the data collected in the qualitative phase. These findings were complemented
with quantitative results that helped to develop a thorough theory and minimized demonstrated weaknesses.
Data conversion (quantitizing nor qualitizing (Sandelowski et al., 2009)) was not used to analyze the whole
data, but a constant side-by-side data comparison was conducted before making inferences. Finally, individual
quantitative and qualitative quality assessments were performed prior to the final integration of the final results
to provide multiple sources of validity.
Results
Quantitative Results
Almost all the MSES’s constructs showed good internal consistency reliability in the fall survey, with
Cronbach’s α values above 0.8 (see Table 1). The only construct that was slightly below the acceptable cutoff of
0.7 (Trochim & Donnelly, 2007) was the mathematics activities construct for the spring semester, with α = 0.69.
Although the mathematics activities construct had a low Cronbach’s α value for the spring semester, this
construct showed good internal consistency reliability in the fall with α = 0.91. This mathematics activities
construct was included in the final data analysis due to the smaller sample in the spring semester (n = 100),
which is less reliable than the fall Cronbach’s α value for the same construct using a bigger sample (n = 308)
(Onwuegbuzie & Leech, 2007).
Table 1. Cronbach’s α values of the items of each construct on the MSES

Semester MSES Construct Cronbach’s α
Mathematics Activities (18 items) 0.91
Fall 2015 Mathematics Courses (14 items) 0.83
Mathematics Problems (18 items) 0.90
Mathematics Activities (18 items) 0.69
Spring 2016 Mathematics Courses (14 items) 0.80
Mathematics Problems (18 items) 0.90
Participants’ Mathematics Competence and Mathematics Self-Efficacy
Table 2 summarizes participants’ mathematics competence knowledge based on their performance in the
mathematics course that they were taking at the time of the interview and their mathematics self-efficacy levels
for both the survey and the interview.
Table 2. Comparison between participants’ mathematics self-efficacy and their mathematics competence
Mathematics self-efficacy Mathematics Comparison
Participants
Interview MSES Competence Self-efficacy Competence
P Spring 2014 6.6 NA 8.1 Medium Higher
L Summer 2014 7 NA 9.1 Medium Higher
P Fall 2014 7.6 NA 1.5 High Lower
P Fall 2014 8 NA 2.9 High Lower
L Spring 2015 6.2 NA 7.3 Medium Higher
P Spring 2015 6.8 NA 6.7 Medium Matched
L Fall 2015 7 5.8 8.5 Medium Higher
L Spring 2016 6.5 7.9 6.6 Medium Matched
P Spring 2016 9 9 8.3 High Matched
L Spring 2016 6.9 8.8 6.5 Medium Matched
L Spring 2016 7.8 8.6 8.2 High Matched
P = Precalc; L = Long Calc; NA is for the six participants that did not take the MSES.
The k-means cluster analysis separated all participants enrolled in an engineering major in three different
clusters based on their mathematics self-efficacy level. The average MSES score (ranging from 1 – 10) was used
to cluster participants into three groups for the semesters that the MSES was distributed and analyzed. These
groups were labeled medium-low (n = 24 for the fall and n = 6 for the spring), medium-high (n = 37 for the fall
and n = 13 for the spring), and high (n = 31 for the fall and n = 5 for the spring) mathematics self-efficacy
groups (see Table 3). None of the three groups was labeled as low due to the relatively high confidence that
participants reported about their mathematics abilities on the survey. The lowest mathematics self-efficacy mean
found by the k-means cluster analysis was 6.3, and this value was not low enough to be considered as poor
mathematics self-efficacy by the researchers. The low mathematics self-efficacy label was purposely avoided to
emphasize the high mathematics self-efficacy levels in general and facilitate interpretation of the final findings
and conclusions. The overall high mathematics self-efficacy levels and the lack of participants reporting a low
mathematics self-efficacy were considered enough evidence to classify these participants as having relatively
high mathematics self-efficacy levels. This relatively high mathematics self-efficacy included participants with
medium and high mathematics self-efficacy levels, emphasizing the confidence that most students reported
about their mathematics abilities.
Table 3. Mathematics self-efficacy survey mean scores separated in 3 clusters by the k-means cluster analysis
Semester Clusters Mean (MSES) Students Participants Selected
1  Medium Low 6.6 24 1
2  Medium High 8 37 0
Fall 2015
3  High 9.2 31 0
1  Medium Low 6.3 6 0
2  Medium High 8 13 2
Spring 2016
3  High 9.3 5 2
Following theoretical selection guidelines (Corbin & Strauss, 2008), participants were selected from each of the
three mathematics self-efficacy groups to complement and test the emerging theory. One participant was
selected from the medium low mathematics self-efficacy group for the fall semester, and two participants were
selected from both the medium high and high self-efficacy groups for the spring semester (see Table 3).
Theoretical sampling was performed to reach maximum variation within these three groups (Teddlie & Yu,
2007). However, it was difficult to reach theoretical saturation due to the lack of participants reporting low
mathematics self-efficacy levels during the survey and some difficulties recruiting participants from the required
groups to keep developing the theory. These difficulties and theoretical saturation issues are described as
limitations for this study later in this paper.
Qualitative Findings
Summaries of each of the 11 participants’ interviews were developed after analyzing all of the qualitative data.
These summaries describe participants’ behavior during their experiences with college mathematics courses,
classifying their experiences in terms of how they reacted to different situations, and contrasting these for
different mathematics self-efficacy and competence levels. After a constant comparison between participant
summaries and a deep analysis of all qualitative data, codes, and memos, the following categories were
established to describe participants’ behaviors and attitudes: persisting in college mathematics courses, adjusting
mathematics self-efficacy, failing a college mathematics course, refreshing basic mathematics competence,
spending extra time working on mathematics, taking struggles as opportunities, seeking help, lacking college-
level mathematics experience, feeling overconfident, lacking effort, blaming other factors, and experiencing
stress working on mathematics.
These categories were the baseline of the theoretical ideas (Charmaz, 2008b) used to develop a theory that
described students’ behavior and attitudes in their college mathematics courses, and identified differences
between students with matched mathematics self-efficacy and competence and students showing a gap between
these two factors based on the results of the mixed methods analysis.
Mixed Methods Results
All participants reported mathematics self-efficacy ranging from medium to high, and they revealed high
confidence about their general mathematics abilities during both their interview and their qualitative
measurements (see Table 2). High mathematics self-efficacy beliefs could influence students’ decision to pursue
an engineering major (Lent et al., 1991; Siegel et al., 1985); thus it was expected that these participants
indicated high levels of confidence in their mathematics skills. Despite participants’ strong confidence in their
mathematics abilities, some of them revealed having some deficiencies in their mathematics competence after
the mixed analysis comparing their interview responses with their mathematics grades (see Table 2). Other
participants revealed a better mathematics competence that they thought they had. Thus, they were rating their
mathematics self-efficacy lower than the actual mathematics competence they possessed (see Table 2). This
mismatch between participants’ mathematics self-efficacy beliefs and their real competence affected these
participants’ behavior and attitude in mathematics courses, making them more likely to behave in certain ways
depending on the direction of the mismatch.
Some participants grouped together after analyzing and comparing their mathematics self-efficacy beliefs with
their mathematics competence (see Figure 2). Participants that clustered together showed similar behaviors and
attitudes while taking college mathematics courses, and these behaviors and attitudes were used to refine and
expand the categories described in the qualitative results. These refined categories were the key to developing a
theory that explains how mathematics self-efficacy of first-year engineering students with poor mathematics
preparation influences their performance and behavior in college mathematics courses.
Figure 2. Participants’ mathematics self-efficacy compared with their mathematics competence and
how they grouped according these two factors
Participants’ behaviors based on their mathematics self-efficacy and mathematics competence relationship
Specific behaviors and attitudes described by participants are categorized for the four groups of participants with
similar mathematics self-efficacy and competence levels, as shown in Figure 2.
Mismatch between High Mathematics self-efficacy and Lower Mathematics Competence
Participants in this group demonstrated overconfidence about their abilities to perform well in their mathematics
courses. They normally struggle to complete their mathematics assignments on time because they believe they
will not need much time to understand their mathematics course material than they really need:
I wasn't putting in all the time [on my math course], and I wasn't practicing enough. So like I said
when I went into the test, even if I could do the problems, I went into all four tests being able to do
the problems, but not being able to do them fast enough to finish the test on time.
It is also typical for this type of engineering student to struggle during their first experience taking a college
mathematics course, and they normally blame external factors for their poor understanding of the class material,
such as the course format or the lack of an experienced professor to set the class pace and explain challenging
topics:
It's online [the math course] so that's my biggest problem. It's all taught online, so that's not really my
style. So I'm a little bit behind in like the pace they want me to be at, but as for understanding the
material I'm doing pretty well.
These participants started their mathematics college classes believing the mathematics material would be easy to
understand because they had developed the belief that mathematics courses are not very challenging based on
their high school experiences. Students who avoid taking calculus in high school may have an inflated sense of
their mathematics abilities because of their experience taking basic mathematics topics that did not challenge
their mathematics understanding. Although current literature suggests that a high mathematics self-efficacy may
have a positive impact on students’ behaviors in mathematics classes (Gore, 2006; Schunk, 1991), engineering
students with high mathematics self-efficacy need to be cautious about thinking that their mathematics abilities
are better than their mathematics competence, especially if their beliefs are based on their performance in basic
mathematics courses. If these students’ mathematics competence is considerably lower than their mathematics
self-efficacy beliefs, they could experience setbacks in their mathematics education once they try to perform
well in college-level mathematics courses such as calculus with the wrong idea that they do not need to spend
extra time working to improve their mathematics abilities.
Engineering students with a mismatch between high mathematics self-efficacy and lower mathematics
competence are likely to experience feelings of stress and anxiety when struggling on their mathematics material
because they believe they should be able to learn the course material and perform well without much effort. One
participant noted,
I'm relatively OK until I get a problem wrong, and then, when I rework it and it's wrong in a different
way, then I get like really stressed because I tried it in two different ways and it didn't work.
The stress and feelings of anxiety usually led these participants to stop working on their mathematics material,
which in turn made them fall behind and prioritize other academic activities instead of trying to improve their
mathematics abilities (Jameson, 2013; Meece, Wigfield, & Eccles, 1990). Although engineering students of this
type are likely to fall behind schedule in their mathematics course and experience failure for the first time due to
their overconfidence about their mathematics abilities, our participants in this group took this failing experience
as an opportunity to refresh their basic mathematics knowledge. This positive attitude about failing their first
college mathematics course helped these participants maintain high confidence in their mathematics abilities so
they could complete the same course the following semester:
Since I'm taking the class again, I'm putting a lot more effort into it, a lot more practice, um, and like
really being involved with the course.
Realizing that their high school mathematics preparation was not as strong as they believed may influence their
mathematics self-efficacy beliefs in future college mathematics courses:
I knew I've always been good at mathematics from like elementary school up to high school, but now
it is getting in more depth and in the hardest stuff I kind of... kind of don't feel like I’m that good
anymore.
Students will more likely approach the course with a more realistic view of their mathematics abilities and a
better understanding of the time that they would need to spend on their mathematics material to complete the
course.
Mismatch between Medium Mathematics self-efficacy and Higher Mathematics Competence
Participants with medium mathematics self-efficacy and higher mathematics competence were more reluctant to
get involved in challenging mathematics activities or to take advanced college mathematics courses for which
they did not feel prepared (Manley & Rosemier, 1972), even when some of them have taken courses like
calculus in high school. These students’ decision to start in a non-college-level mathematics course, even though
they may have been prepared for higher levels, may be because they are trying to refresh their prior mathematics
knowledge before taking a college-level mathematics course:
I'm now taking Precalc, just because I want. Also because I just want to make sure of my basic skills
before I try advance math courses. You don't want to try and push your way through this math, and
maybe pass it, but not really get it.
They are usually confident enough about their mathematics abilities to sufficiently complete the mathematics
courses required by their engineering major, but they consciously decided to be cautious about the possibility of
struggling with more advanced mathematics topics in advanced mathematics courses. These participants viewed
struggles in learning mathematics as opportunities to test their real understanding and improve their
mathematics abilities, considering struggling as a normal part of learning advanced mathematics topics.
Although these participants believe that college mathematics courses were more challenging than their high
school courses, they were likely to have positive behaviors and attitudes in their college mathematics classes,
spend extra time working on their mathematics assignments, seek help in case they struggle to understand any
mathematics topic, and remain motivated to improve their mathematics abilities even when they were facing
struggles in understanding new mathematics topics:
Usually I would go back through the homework assignments that we have done. And also there are
extra assignments that we don’t get to in class, so I’ll just go back and practice those, and if I have a
problem with one, then I'll go to my teacher and say can you help me with this and she will guide me
through it.
These engineering students normally put in more effort to learn new and challenging mathematics material, but
they are likely to wait until they get help from someone to continue working on a mathematics problem they fail
to solve initially on their own:
When I get helped, it just helps me to remember more, 'cause when you go to ask someone you kind
of recall the conversation. I do office hours definitely with the teacher, but if it is a minor homework
problem. I go with my friends, or my roommate, or anyone that I know they are good at math, or they
are in a higher level of math than I am. So you know they will know the problem or the topic I'm
working on and they can help me.
Although seeking help before giving up helps these students believe that they can complete advanced
mathematics courses if they find the right support systems, these beliefs may also generate poor confidence in
their own mathematics abilities when they are working on their own and there is no one to help them, such as on
an exam, in case they struggle:
In class I'll say about an 8 [mathematics self-efficacy] and in a test I'll say about a 5 [mathematics
self-efficacy]. In class I just ask for some help and then I understand it, and during the test, you know,
they [mathematics problems] don't look the same to me, or I don't see them as the same and I just get
confused. I kind of like… I just answer what I really know how to answer.
This type of student is more likely to look for a mathematics course that could let them adapt to the different
teaching style and difficulty level of college mathematics courses without feeling rushed or getting behind when
transitioning into college.
Matched High Mathematics self-efficacy and High Mathematics Competence
Participants with matched high mathematics self-efficacy and mathematics competence demonstrated positive
behaviors and attitudes in college mathematics courses. Such behaviors included spending extra time working
on their mathematics material, seeking extra help from tutors, and attending office hours (Zimmerman, 2000):
I´m normally done with my classes by twelve, so I'm at the library doing work or in my room doing
work with a friend. I do work at least an hour a day, at least one hour per day for math class if not
more.
These engineering students usually view their struggles learning mathematics as part of the mathematics
learning process, and they consider these struggles as opportunities to improve their mathematics competence
(Lent et al., 1991). They enjoy the challenge in learning new and/or advanced mathematics topics:
I feel like things [mathematics courses] are getting harder sometimes, but as long as you want to
learn, and you want to get better then you can do it. And I've always loved math so I always try to be
better at it.
Students of this type are very confident about completing the mathematics courses required by their engineering
majors, and they usually show the will to take advanced mathematics courses more than once if they struggle to
complete the course the first time (Multon et al., 1991; Suresh, 2006):
I will [finish all math courses required by my major] even if that takes an extra semester, or a year of
college. I think I would because this [engineering] is for sure what I want to do.
All participants in this category were taking their mathematics course for the second time after failing their first
college mathematics course. They reported changing their approach to taking a college mathematics course after
experiencing setbacks and failure the first time they took the course. They used this failing experience as a
wake-up call that helped them to realize they would need to spend extra time working on their mathematics
material; they sought help to improve their mathematics abilities to stay on pace with the class at the college
level:
I didn't pass the first part of the Long Calc. It was kind of... kind of a reality check, 'cause I guess I
could have studied more, went to tutoring more, and stuff to make my grade better in that class.
These participants did not lose confidence in their mathematics abilities after failing their first college
mathematics course, but rather they used their failing experience to match their real mathematics competence
with more realistic mathematics self-efficacy beliefs. Keeping a high mathematics self-efficacy that was better
aligned with their mathematics competence helped these participants feel motivated to work harder the second
time that they took the same mathematics course (Bandura & Schunk, 1981; Walker, Greene, & Mansell, 2006).
This result was confirmed by their improved performance when taking the same mathematics course that they
failed the semester before.
Matched Medium High Mathematics self-efficacy and Medium High Mathematics Competence
Participants who had a matched medium high mathematics self-efficacy and mathematics competence lacked
college-level mathematics skills coming from high school, and that poor understanding of advanced
mathematics topics was a problem when they took their first college mathematics course. These engineering
students usually think their mathematics competence is acceptable to pass the course without working hard to
improve their mathematics abilities causing them to remain in a comfort zone where they overlook the
importance of hard work and practice in mathematics courses (Lent et al., 1984):
Well the first time [taking a college math course], I was under the assumption that ‘Oh I know this
stuff, I don't have to study as much as thought I did,’ and I didn’t put as much practice as you need to
do it.
Participants in this group typically blamed external factors such as the professor’s teaching style and the fast
class pace for their failing experiences in college mathematics courses, but they usually showed limited interest
in spending extra time working on their own to improve their mathematics abilities.
All participants in this category failed their first college mathematics course. After this failing experience, these
engineering students realized they would need to work harder and spend more time working on their coursework
to be able to complete all the mathematics courses required by their engineering majors.
Failing the course made me feel pretty bad, and after that I told myself I will work better the next
time, so I can pass it and not have that feeling again.
Their failing experience showed them that their mathematics courses were more difficult than expected, and
they would need to lower their mathematics self-efficacy beliefs to align with their ability. This mathematics
self-efficacy adjustment brought their mathematics self-efficacy beliefs closer to their real mathematics
competence, making these students more likely to seek help when struggled with mathematics topics (Lopez &
Lent, 1992). Although these participants showed more interest on improving their mathematics abilities the
second time they took a college mathematics course, they remained more likely to stop trying to solve
mathematics problems if they were struggling and there was no one there to help them understand what was
wrong with their approach. They were likely to experience stress if they struggled working on mathematics
problems:
I don't even think that I hate math, 'cause I enjoy talking about mathematical concepts with
professors. I just hate the struggle. I hate when you can't do it, and you sit there over ten online
homework problems for hours and is like... if I could do this in 20 minutes, math would be my
favorite subject.
However, they reported to be likely to persist in repeating mathematics courses they had failed, viewing their
failing experiences as opportunities to refresh their mathematics competence and perform better next time.
Discussion
The population of engineering students in this study reported relatively high mathematics self-efficacy,
especially when they were referring to their abilities to perform mathematics in general and their confidence in
completing all the mathematics courses required by their majors. These results support current literature about
students who choose engineering as their major feeling confidence that their mathematics competence will be
good enough to perform well in college mathematics-related activities (Hackett & Betz, 1989; Lent, Lopez, &
Bieschke, 1991.) Most of the participants in this study were very confident that they would successfully
complete their first college mathematics course, even when they showed poor mathematics preparation coming
from high school. The gap between high mathematics self-efficacy and lower mathematics competence was
related to participants’ negative behaviors and attitudes in mathematics courses such as feeling overly confident
about their mathematics competence, blaming other factors for their struggles and failing experiences, and not
putting forth effort to improve their mathematics abilities (see Figure 3). These negative behaviors and attitudes
created a stressful environment in the participants’ mathematics courses, making them more likely to perform
poorly and fail their first college mathematics course.
Having a failing experience in their first college mathematics course put participants in a position where they
had to adjust their mathematics self-efficacy to a lower level, closer to their real mathematics competence. Their
mathematics self-efficacy remained high enough to make them think that they still could complete all the
mathematics courses required by their majors, but after failing their first college mathematics course, they were
more likely to adjust their mathematics self-efficacy to match their actual mathematics competence. This new
and more realistic perception of their mathematics competence was related to a noticeable change in
participants’ behavior and attitudes when taking a college mathematics course the following semester.
Participants with mathematics self-efficacy beliefs aligned with their mathematics competence were more likely
to have positive behaviors and attitudes in mathematics courses such as spending extra time working on their
mathematics material and seeking help to address their doubts, and viewing their struggles in learning new
mathematics topics as opportunities to develop better their mathematics abilities and gain a deeper
understanding of the course material (see Figure 3). By adjusting their approach in the same mathematics
course, participants were better able to develop their mathematics competence and they usually enjoyed the
challenge involved in learning new difficult mathematics topics.
Although some participants reported struggling with the pace of the course and understanding advanced
mathematics topics in their first college mathematics course, their mathematics self-efficacy remained always
relatively high. This medium to high mathematics self-efficacy level was related to participants’ confidence
about satisfactorily completing all the mathematics courses required by their majors if they worked hard and put
in extra effort to improve their mathematics abilities. This result is supported in the self-efficacy literature where
students feel more motivated to put extra effort to overcome difficulties when they are facing challenging
activities if they believe they can succeed performing such activity (Bandura, 1980; Marra, Rodgers, Shen, &
Bogue, 2009; Walker et al., 2006). These engineering students were confident about their abilities to complete
their college mathematics courses even if they would need to repeat some of the most advanced courses due to
failing to pass them the first time.
Figure 3. Usual behaviors and attitudes in college mathematics courses for first-year engineering students with
high mathematics self-efficacy and poor mathematic preparation as demonstrated by lower mathematics
competence.
Note that in Figure 3, “Positive Behaviors and Attitudes” refers to spending extra time, seeking help, and taking
struggles as opportunities; “Negative Behaviors and Attitudes” refers to lacking effort, experiencing stress,
blaming other factors, and overconfidence.
The relationship between high mathematics self-efficacy and lower mathematics competence suggested by our
findings is not accounted for in the current mathematics self-efficacy literature (Bandura, 1986; Hackett & Betz,
1989). These research findings are inconsistent with the most common findings in self-efficacy studies, where
students are more likely to be engaged and try to spend more time on activities where they think that they can
perform well (Bouffard-Bouchard, 1990; Liem, Lau, & Nie, 2008). Most prior research about self-efficacy is
based on this factor alone, and how this factor relates to or affects other behavioral variables such as
performance (Vancouver & Kendall, 2006), anxiety (Pajares & Miller, 1994), or academic choices (Schunk,
1991). The mixing phase in our study, which analyzes matches and mismatches of mathematics self-efficacy
and competence, helped the researchers to expand current knowledge about self-efficacy. The comparison of
mathematics self-efficacy levels with engineering students’ mathematics competence showed that there is
usually a mismatch between these two factors for first year engineering students, suggesting that the analysis of
students’ mathematics self-efficacy levels alone is not enough to determine the possible students’ behaviors and
attitudes in mathematics courses. It may be beneficial for future mathematics self-efficacy research to analyze
their participants’ mathematics self-efficacy together with their mathematics competence to determine effects of
a possible mismatch between these two factors.
Conclusion and Recommendations

Engineering and mathematics educators should be aware that engineering students are likely to have high
mathematics self-efficacy beliefs coming from high school, believing they can perform well in college
mathematics courses even when they have some deficiencies in their mathematics competence. Mathematics
instructors should be prepared to advise and effectively motivate engineering students with a poor mathematics
preparation who are not ready to take a calculus course when they start their college education. First year
engineering students tend to be overconfident about their mathematics abilities the first time they take a college
mathematics course; they tend to procrastinate in class, put little effort trying to improve their mathematics
abilities, and blame other factors for their struggles in learning new mathematics topics. The combination of
these beliefs, attitudes, and actions may result in them failing their first college mathematics course.
Mathematics instructors should be prepared to inform students about the potential negative behaviors and
attitudes related to engineering students with overestimated mathematics self-efficacy beliefs. With this
knowledge, these students may reflect on their competence more honestly and could adjust their mathematics
self-efficacy beliefs before experiencing a failing grade in their first college mathematics course. If they are
advised to align their mathematics self-efficacy beliefs closer to their mathematics competence, they may be
more likely to avoid the negative behaviors described above and be more engaged in activities that could help
them to improve their mathematics competence and be less likely to fail their first mathematics course in
college. Although engineering students will normally use a failing experiences in college mathematics courses
to adjust their mathematics self-efficacy beliefs closer to their real mathematics competence level, positively
changing their behaviors and attitudes in future mathematics courses, mathematics professors should warn
engineering students with poor mathematics preparation about the risk of having a mismatch between high
mathematics self-efficacy beliefs with lower mathematics competence during the first weeks of the semester.
This knowledge may allow engineering students to avoid failing experiences in their first college year and avoid
second thoughts and negative feelings about their abilities to successfully complete all their mathematics
courses required by their major. By decreasing the risk of first-year engineering students becoming
overconfident about their mathematics competence, engineering educators may also seek ways to reduce the
high failing rates in college mathematics courses, having a positive impact in the engineering graduation rates
(Seymour & Hewitt, 1997; Suresh, 2006). It may be helpful to have students assess their own mathematics self-
efficacy using an instrument like the MSES. We demonstrated validity and reliability for results from the MSES
for a population of first-year engineering students.
This research contributes to the mathematics self-efficacy literature by suggesting that high mathematics self-
efficacy beliefs that are not supported by a correspondingly high mathematics competence could lead to
negative behaviors and attitudes such as overconfidence, lack of effort in mathematics classes, laying blame on
other factors for their struggles learning mathematics, and feelings of stress and anxiety when performing
mathematics activities, which ultimately lead students to perform poorly and fail their first college mathematics
courses.
Despite the negative behaviors and attitudes related to engineering students who show a gap between high
mathematics self-efficacy and lower mathematics competence, these students are very likely to keep trying to
complete the mathematics courses required by their majors regardless of facing failing experiences because of
their high levels of mathematics self-efficacy. Engineering and mathematics educators should try to take
advantage of this persistence shown by engineering students by trying to engage them in mathematics related
activities. Students with high levels of mathematics self-efficacy are likely to respond positively to a professor’s
constructive feedback about their mathematics deficiencies, as their mathematics self-efficacy enables them to
enjoy working on mathematics related activities but also feel capable of performing well on them.
Limitations
This research should be viewed considering its limited sample, interpretive nature, and the complexity of the
theoretical sampling to select the most appropriate participants for the qualitative phase of this research. Firstly,
a bigger and more diverse sample would be required to complement the emerging theory with a deeper
understanding of the causality of these behaviors and attitudes. Collecting additional data from more students
would be especially beneficial for expanding groups that showed some contradictions in student’s behaviors.
For example, the small sample size (n = 2) of the high mathematics self-efficacy with lower mathematics
competence group made it challenging to determine the most likely behaviors and attitudes of participants in
this cluster.
Additional data would also support the use of k-means cluster analysis for purposeful sampling of participants,
as k-means is appropriate for large sample sizes (Ehlert et al., 2017). The use of k-means cluster analysis with
small sample size is a limitation of this study. Secondly, recruiting the right participant that could fit the
theoretical sampling specifications was a major challenge due to the lack of response from students to the e-mail
invitation. Failing to recruit the right participants to continue developing the emerging theory was the biggest
limitation of this research; this issue affected the sample size for groups that needed more participants,
particularly those with both low mathematics self-efficacy and competence. Finally, there are more Spring
semester (seven) than Fall semester (four) participants in this research. The last qualitative data collection was
performed during the Spring 2016 semester, and the last four participants were repeating their mathematics
course after failing it the first time. This inability to recruit a participant who had not had a failing experience
affected the theoretical sampling. This issue could be addressed in future research by aiming to develop a better
understanding of engineering students’ behaviors and attitudes prior to having a failure experience.
Future Work
Future research should be conducted to determine the reasons why students in different groups are likely to
choose certain behaviors in mathematics classes depending on their mathematics self-efficacy level and how
well this mathematics self-efficacy is representing their mathematics competence. Additional research with
participants from diverse backgrounds should be added to this grounded theory study aiming to expand the
emerging theory, and trying to reach theoretical saturation. More data from engineering students in different
backgrounds will help to describe students experiences in a richer way (Bryant & Charmaz, 2007), which was
not possible to reach during this research due to its time and participant recruitment limitations. Future research
aiming to shed light on the possible relationships between mathematics self-efficacy and engineering students’
behaviors and attitudes should consider the importance of comparing students’ mathematics self-efficacy with
their mathematics competence. This research may guide researchers focusing on only self-efficacy
measurements rather than analyzing supports for students’ beliefs about their abilities. Some students
commented on factors outside their control, such as teaching styles and classroom environments, on their
academic performance. This indicates a need for future research into the relationships between a mathematics
self-efficacy and competence mismatch and student perceptions of locus of control.
Acknowledgements
We wish to thank Dr. Courtney Faber and Catherine McGough, who helped collect interview data and review
coded transcripts, and Aubrie Pfirman and Dr. Rachel Anderson for their constant reviewing of the manuscript.
We are indebted to Dr. Sandra Linder for generously sharing her expertise on research methods as we designed
and conducted this mixed methods study. This research was supported through funding by the Mexican Consejo
Nacional de Ciencia y Tecnología.
References
Bandura, A. (1977). Self-efficacy: Toward a Unifying Theory of Behavioral Change. Psychological Review, 84,
191–215.
Bandura, A. (1980). Gauging the Relationship Between Self-Efficacy Judgment and Action. Cognitive Therapy
and Research.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ:
Prentice Hall.
Bandura, A., & Schunk, D. H. (1981). Cultivating Competence, Self-Efficacy, and Intrinsic Interest Through
Proximal Self-Motivation. Journal of Personality and Social Psychology, 41(3), 586–598.
Betz, N. E., & Hackett, G. (1983). The Relationship of Mathematics Self-Efficacy Expectations to the Selection
of Science-Based College Majors. Journal of Vocational Behavior.
Bouffard-Bouchard, T. (1990). Influence of Self-Efficacy on Performance in a Cognitive Task. The Journal of
Social Psychology.
Brown, S., & Burnham, J. (2012). Engineering Student’s Mathematics Self-Efficacy Development in a
Freshmen Engineering Mathematics Course. International Journal of Engineering Education, 28(1),
113–129.
Bryant, A., & Charmaz, K. (2007). The SAGE Handbook of Grounded Theory. Thousand Oaks, CA: Sage
Publications, Inc.
Casey, M. B., Nuttall, R. L., & Pezaris, E. (1997). Mediators of Gender Differences in Mathematics College
Entrance Test Scores: A Comparison of Spatial Skills with Internalized Beliefs and Anxieties.
Developmental Psychology, Vol 33(4), 669–680.
Charmaz, K. (2006). Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis. London:
Sage Publications.
Charmaz, K. (2007). Constructionism and the Grounded Theory Method. In Handbook of Constructionism
Research. (pp. 397–412). New York: Guilford Press.
Charmaz, K. (2008a). Grounded Theory as an Emergent Method. In Handbook of Emergent Methods (pp. 155–
170). New York: Guilford Press.
Charmaz, K. (2008b). The Legacy of Anselm Strauss for Constructivist Grounded Theory. Bingley, UK:
Emerald.
Chen, X., & Soldner, M. (2013). STEM Attrition: College Students’ Paths Into and Out of STEM Fields
Statistical Analysis Report. National Center for Education.
Committee on STEM Edcucation. (2013). Federal Science, Technology, Engineering, and Mathematics (STEM)
Education. Washington, D.C.: National Science and Technology Council.
Cooper, S. E., & Robinson, D. A. (1991). The Relationship of Mathematics Self-Efficacy Beliefs to
Mathematics Anxiety and Performance. Measurement & Evaluation in Counseling & Development,
24(1), 4–8.
Corbin, J., & Strauss, A. (2008). Basics of Qualitative Research: Grounded Theory Procedures and Techniques
(3rd ed.). Thousand Oaks, CA: Sage.

Creswell, J. W. (2009). Research Design: Qualitative, Quantitative and Mixed Methods Research. Thousand
Oaks, CA: Sage Publications, Inc.
Creswell, J. W., & Clark, V. (2011). Designing and Conducting Mixed Methods Research (2nd ed.). Thousand
Oaks, CA: Sage Publications, Inc.
Ehlert, K. M., Faber, C. J., Benson, L. C., & Kennedy, M. S. (2017). Utilizing Cluster Analysis of Closed-Ended
Survey Response to Select Participants for Qualitative Data Collection. In Proceedings of the 124th
American Society for Engineering Education Annual (ASEE) Conference & Exposition, Columbus, OH.
Erdmann, V., & Schumann, T. (2010). European Engineering Report. Retrieved from
https://www.vdi.de/uploads/media/2010-04_IW_European_Engineering_Report_02.pdf
Eris, O., Chachra, D., Chen, H. L., Sheppard, S., Ludlow, L., & Rosca, C. (2010). Outcomes of a Longitudinal
Administration of the Persistence in Engineering Survey. Journal of Engineering Education, 99(4), 371.
Flores, A. (2007). Examining Disparities in Mathematics Education: Achievement Gap or Opportunity Gap?
High School Journal, 91(1), 29–42.
Gardner, J., Pyke, P., Belcheir, M., & Schrader, C. (2007). Testing Our Assumptions: Mathematics Preparation
and its Role in Engineering Student Success. In Proceedings of the 2007 American Society for
Engineering Education Annual Conference and Exposition, Honolulu, HI.
Geisinger, B. N., & Raman, D. R. (2013). Why They Leave : Understanding Student Attrition from Engineering
Majors. International Journal of Engineering Education, 29(4), 914–925.
Gore, P. a. (2006). Academic Self-Efficacy as a Predictor of College Outcomes: Two Incremental Validity
Studies. Journal of Career Assessment, 14(1), 92–115.
Hackett, G. (1985). Role of Mathematics Self-Efficacy in the Choice of Math-Related Majors of College
Women and Men: A Path Analysis. Journal of Counseling Psychology, 32(1), 47–56.
Hackett, G., & Betz, N. E. (1989). An Exploration of the Mathematics Self-Efficacy/Mathematics Performance
Correspondence. Journal for Research in Mathematics Education, 20(3), 261–273.
Jameson, M. M. (2013). Contextual Factors Related to Math Anxiety in Second-Grade Children. The Journal of
Experimental Education, (March 2014), 1–19.
Johnson, R. B. (1997). Examining the Validity Structure of Qualitative Research. Education, 118, 282–292.
Kokkelenberg, E. C., & Sinha, E. (2010). Economics of Education Review Who succeeds in STEM studies ? An
analysis of Binghamton University undergraduate students. Economics of Education Review, 29(6), 935–
946.
Kranzler, J. H., & Pajares, F. (1997). An Exploratory Factor Analysis of the Mathematics Self-Efficacy Scale-
Revised (MSES-R). Measurement & Evaluation in Counseling & Development, 29(4), 215–229.
Lee, J. (2012). Educational Equity and Adequacy for Disadvantaged Minority Students: School and Teacher
Resource Gaps Toward National Mathematics Proficiency Standard. Journal Education Research,
105(1), 64–75.
Lent, R., Lopez, F., & Bieschke, K. (1991). Mathematics Self-Efficacy Sources and Relation to Science-Based
Career Choice. Journal of Counseling Psychology, 38(4), 424–430.
Lent, R. W., Brown, S. D., & Hackett, G. (1994). Toward a Unifying Social Cognitive Theory of Career and
Academic Interest, Choice, and Performance. Journal of Vocational Behavior, 45, 79–122.
Lent, R. W., Brown, S. D., & Hackett, G. (2000). Contextual Supports and Barriers to Career Choice: A Social
Cognitive Analysis. Journal of Counseling Psychology, 47(1), 36–49.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1984). Relation of Self-Efficacy Expectations to Academic
Achievement and Persistence. Journal of Counseling Psychology, 31, 356–362.
Lent, R. W., Lopez, F. G., & Bieschke, K. J. (1991). Mathematics Self-Efficacy: Sources and Relation to
Science-Based Career Choice. Journal of Counseling Psychology, 38(4), 424–430.
Liem, A. D., Lau, S., & Nie, Y. (2008). The Role of Self-Efficacy, Task Value, and Achievement Goals in
Predicting Learning Strategies, Task Disengagement, Peer Relationship, and Achievement Outcome.
Contemporary Educational Psychology, 33, 486–512.
Lopez, F. G., & Lent, R. W. (1992). Sources of Mathematics Self-Efficacy in High School Students. Career
Development Quarterly, 41(1), 3.
Maechler, M., Rousseeuw, P., Struyf, A., Hubert, M., & Hornik, K. (2015). Cluster: Cluster Analysis Basics
and Extensions. R Package Version 2.0.1.
Manley, M. J., & Rosemier, R. A. (1972). Developmental Trends in General and Test Anxiety Among Junior
and Senior High School Students. Journal of Genetic Psychology, 120, 219–226.
Marra, R. M., Rodgers, K. a, Shen, D., & Bogue, B. (2009). Women Engineering Students and Self-Efficacy: A
Multi-Year, Multi-Institution Study of Women Engineering Student Self-Efficacy. Journal of
Engineering Education, 98(January), 27–38.
May, G. S., & Chubin, D. E. (2003). Undergraduate Engineering Success for Underrepresented Minority
Students. Journal of Engineering Education, 92(January), 27–39.
Meece, J. L., Wigfield, A., & Eccles, J. S. (1990). Predictors of Math Anxiety and its Influence on Young
Adolescents’ Course Enrollment Intentions and Performance in Mathematics. Journal of Educational
Psychology, 82(1), 60–70.
Middleton, J. A., Krause, S., Maass, S., Beeley, K., Collofello, J., & Culbertson, R. (2015). Early Course and
Grade Predictors of Persistence in Undergraduate Engineering Majors. Proceedings - Frontiers in
Education Conference, FIE, 2015–Febru(February).
Moran-Ellis, J. (2006). Triangulation and Integration: Processes, Claims and Implications. Qualitative Research,
6(1), 45–59.
Moran, G., & Benson, L. (2016). Effects of Math Self-Efficacy on Behaviors of Engineering Students with Poor
Math Preparation. In International Group for the Psychology of Mathematical Education. Szeged,
Hungary.
Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of Self-Efficacy Beliefs to Academic Outcomes: A
Meta-Analytic Investigation. Journal of Counseling Psychology, 38(1), 30–38.
National Academy of Engineering. (2015). NAE Grand Challenges for Engineering. Retrieved from
http://www.engineeringchallenges.org/challenges.aspx
Nunnally, J., & Bernstein, I. (1994). Psychometric Theory (3rd ed.). New York: McGraw-Hill.
Onwuebuzie, A. J., & Johnson, R. B. (2006). The Validity Issue in Mixed Research. Research in the Schools,
13(1), 48–63.
Onwuegbuzie, A. J., & Collins, K. M. (2007). A Typology of Mixed Methods Sampling Designs in Social
Science Research . The. Qualitative Report, 12(2), 281–316.
Onwuegbuzie, A. J., & Combs, J. P. (2010). Emergent Data Analysis Techniques in Mixed Methods Research.
In I. Los Angeles: Sage Publications (Ed.), SAGE Handbook of Mixed Methods in Social & Behavioral
Research (pp. 397–430).
Onwuegbuzie, A., & Leech, N. (2007). Sampling Designs in Qualitative Research : Making the Sampling
Process More Public. The Qualitative Report, 12(2), 19–20.
Pajares, F., & Miller, M. D. (1994). Role of Self-Efficacy and Self-Concept Beliefs in Mathematical Problem
Solving: A Path Analysis. Journal of Educational Psychology, 86(2), 193–203.
Pajares, F., & Miller, M. D. (1995). Mathematics Self-Efficacy and Mathematics Performances: The Need for
Specificity of Assessment. Journal of Counseling Psychology, 42(2), 190–198.
Porter, C. H. (2011). An Examination of Variables Which Influence High School Students to Enroll in an
Undergraduate Engineering or Physical Science Major. Dissertation, Clemson University Eugene Moore
School of Education, Education Leadership Program.
Redmond, A., Thomas, J., High, K., Scott, M., Jordan, P., & Dockers, J. (2007). Enriching Science and Math
Through Engineering. School Science and Mathematics., 111(8), 399–408.
Sandelowski, M., Voils, C. I., & Knafl, G. (2009). On Quantitizing. Journal of Mixed Methods Research, 3(3),
208–222.
Schunk, D. (1991). Self-Efficacy and Academic Motivation. Educational Psychologist.
Seymour, E., & Hewitt, N. M. (1997). Talking About Leaving: Why Undergraduates Leave the Sciences.
Westview Press, Boulder, CO.
Siegel, R. G., Galassi, J. P., & Ware, W. B. (1985). A Comparison of Two Models for Predicting Mathematics
Performance: Social Learning Versus Math Aptitude-Anxiety. Journal of Counseling Psychology, 32(4),
531–538.
Snyder, T. D., & Dillow, S. A. (2010). Digest of Education Statistics 2009 (NCES 2010-013). National Center
for Education Statistics, Institute of Education Sciences, Washington, DC: U.S. Department of
Education.
Suresh, R. (2006). the Relationship Between Barrier Courses and Persistence in Engineering. Journal of College
Student Retention: Research, Theory and Practice, 8(2), 215–239.
Team, R. C. (2012). R: A Language and Environment for Statistical Computing. Vienna: R Foundation for
Statistical Computing. Retrieved from http://www.r-project.org/.
Teddlie, C., & Yu, F. (2007). Mixed Methods Sampling: A Typology With Examples. Journal of Mixed
Methods Research, 1(1), 77–100.
Thorndike, R. M., & Thorndike-Christ, T. (2010). Measurement and Evaluation in Psychology and Education
(8th ed.). Boston: Pearson.
Trochim, W., & Donnelly, J. P. (2007). The Research Methods Knowledge Base. Atomic Dog Publishing.
Van Dyken, J., Benson, L., & Gerard, P. (2015). Persistence in Engineering: Does Initial Mathematics Course
Matter? American Society Engineering Education, 2015 Annual Conference and Exposition.
Vancouver, J. B., & Kendall, L. N. (2006). When Self-Efficacy Negatively Relates to Motivation and
Performance in a Learning Context. Journal of Applied Psychology, 91(5), 1146–1153.
Walker, C. O., Greene, B. a., & Mansell, R. a. (2006). Identification with Academics, Intrinsic/Extrinsic
Motivation, and Self-Efficacy as Predictors of Cognitive Engagement. Learning and Individual
Differences, 16, 1–12.

Williams, T., & Williams, K. (2010). Self-Efficacy and Performance in Mathematics: Reciprocal Determinism
in 33 Nations. Journal of Educational Psychology, 102(2), 453–466.
Zimmerman, B. (2000). Self-Efficacy: An Essential Motive to Learn. Contemporary Educational Psychology,
25, 82–91.
Author Information
Gustavo Morán-Soto Lisa Benson
Instituto Tecnológico de Durango Clemson University
Mexico U.S.A.
Contact e-mail: gmorans@clemson.edu
An Instrument for Examining Elementary Students’ Engineering Interests

and Attitudes
Cathy P. Lachapelle, Robert T. Brennan

Article History Engineers and policymakers have expressed concern that too few students enter
the engineering pipeline. This has led to many efforts to engage students in
Received: engineering in after-school programs, summer programs, and more recently, in
22 July 2017 school curricula. The expectation is that, through these efforts, greater numbers
of more demographically diverse children will become aware of engineering as a
Accepted:
possible career option, and some will decide to pursue it, thereby increasing and
01 February 2018
diversifying the population pursuing engineering careers. This expectation makes
the assumption that students will become more interested in and form more
Keywords
positive attitudes towards engineering as they encounter it in formal and informal
Elementary school settings. To measure this assumption, we have developed an Engineering Interest
Engineering education and Attitudes (EIA) survey, drawing from earlier surveys used to measure
Student interests student interest in and attitudes toward science. We show that the subscales
Student attitudes developed from EFA and CFA are reliable, and considerable evidence is present
Gender differences for the validity of use of EIA for measuring young students’ engineering interests
and attitudes. We also present evidence that EIA can be used by researchers and
curriculum developers with students ages 8-11 to measure change in student
interests and attitudes towards the goal of evaluating engineering activities,
programs, and curricula.
Introduction
Context
Engineers and policymakers have expressed concern that too few students enter the engineering pipeline. The
Organization for Economic Co-operation and Development notes that the proportion of students in OECD
countries choosing to enter all STEM fields has been dropping since the mid-1990s (OECD, 2008). A recent
report by the U.S. National Academies summarizes reports and surveys of employers, industry groups, and
government agencies that have expressed concern about an insufficient supply of engineers and other skilled
technology workers (National Academies of Sciences, Engineering, and Medicine, 2017).
An additional concern is the lack of diversity among engineers (Buccheri, Gurber, & Bruhwiler, 2011; National
Academy of Sciences, National Academy of Engineering (NAE), & Institute of Medicine, 2010; National
Research Council & NAE, 2014). Women are underrepresented in most nations, and in the United States there is
particular concern about the paucity of African American and Latino/a engineers compared to the general
population.
It can be argued that an important time to introduce children to career options is during childhood. Research
shows that many engineers and scientists form their career choices before adolescence (Maltese & Tai, 2010;
Royal Society, 2006; Venville, Wallace, Rennie, & Malone, 2002). Using longitudinal datasets and methods,
several researchers have found that before children reach middle school they already have settled on a career
path, whether or not that will be a STEM field (Lindahl, 2007; Lyons, 2006; Tai, Liu, Maltese, & Fan, 2006).
Children’s interest in and aptitude for science has generally been found to be high for both girls and boys
younger than 10; however, interest drops over time as children progress through school (Murphy & Beggs,
2003). The drop in interest is particularly pronounced for female students—among adolescent and older
students, the attitudes of males toward the physical sciences and engineering are consistently more positive than
those of females (Tytler, 2014). This may be due to the content of the curriculum, which often does not connect
well to the concerns of people and societies—concerns that girls consistently rate as more compelling than
222 Lachapelle & Brennan
content detached from such concerns, particularly as compared to boys (Burke, 2007; Häussler & Hoffmann,
2002). College-bound girls have been shown to prefer biological sciences and engineering majors, particularly
those relating to health careers, the environment, and other “helping” professions (Buccheri, Gurber, &
Bruhwiler, 2011; Drechsel, Carstensen, & Prenzel, 2011; Miller, Blessing, & Schwartz, 2006).
Children’s Interests in and Attitudes toward Science and Engineering
Vaughan and Hogg (2013, p. 169) explain that “Theories of attitude structure generally agree that attitudes are
lasting general evaluations of socially significant objects (including people and issues).” In engineering
education, socially significant objects include engineers and the work of engineering, which can have social
significance both for students themselves—their life experiences and future expectations—and in their effects
on aspects of the world that matter to students, such as transportation, the environment, or medicine. In this
paper, we focus on student attitudes toward three socially significant objects: engineers, engineering as a
profession, and learning experiences in engineering. Research on student attitudes conducted in science
education shows that it is important to attend separately to students’ attitudes toward school science versus
science and scientists more generally, because these can be quite different, and students’ attitudes toward each
can vary accordingly (Lindahl, 2007; Tytler, 2014), and we have taken this finding into consideration as we
investigate children’s attitudes toward engineering.
Most research has focused on interest in and attitudes toward science, though some findings have extended to
STEM careers more generally. Despite the relative lack of work specifically on young students’ engineering
interests, engineering advocates interested in increasing the flow of students through the engineering pipeline
have chosen to see science findings as applicable to engineering, leading to many efforts to engage pre-
adolescent students in engineering in after-school programs, summer programs, and, more recently, in-school
curricula. The expectation is that by engaging students in engineering greater numbers of more demographically
diverse children will become aware of it as a career option, and some students will find a special affinity to
engineering and ultimately pursue it. Given the goal of increasing interest in engineering through interventions,
it is important to develop instruments capable of measuring change in student attitudes toward and interest in
engineering for a given intervention.
Most available STEM attitude measures, like most STEM research studies, have focused on student attitudes
toward science, as well as their interest in (or aspirations toward) future study of science or careers in STEM
fields. According to two recent literature reviews, the quality and validity evidence for these surveys of attitudes
varies greatly (Blalock et al., 2008; Tytler, 2014). Among the topics surveyed by science attitudes instruments
as reported in these reviews are: (1) desire to learn science, (2) interest in science careers, (3) positive emotions
(e.g., enjoyment or “liking”) toward science generally, (4) positive emotions toward doing science in school, (5)
valuing of the work of scientists and the outcomes of science, and (6) valuing scientific perspectives.
The most common instrument found in the literature (Tytler, 2014) is the Scientific Attitude Inventory (SAI),
which was designed for use with middle and high school students (Moore & Sutman, 1970). This instrument
was later revised and improved as the SAI-II (Moore & Foy, 1997). However, the revision did not result in a
factor structure that matched the author’s original theorized structure of 12 factors (six factors with a positive
and negative version of each). Lichtenstein and colleagues revisited the survey (2008) with a new sample
collected from more than 500 middle and high school students; using exploratory factor analysis (EFA) and
confirmatory factor analysis (CFA), they found and confirmed a structure of three factors with only one of the
three having acceptable psychometric properties.
The instrument with the strongest characteristics (Blalock et al., 2008) was found to be the Attitude toward
Science in School Assessment (ATSSA), which was designed for use with high school students (Germann,
1988). The purpose of the ATSSA is to assess the attitudes of adolescent students toward school science. As
with all the instruments we surveyed, the ATSSA employs a five-point Likert scale. Germann chose 14 items
from an initial set of 34 based on expert review. EFA on those 14 items resulted in a single factor with high
internal consistency (Cronbach’s α>.95 in four experimental samples).
The Simpson-Troost Attitude Questionnaire (STAQ), which was developed originally for use with high school
students, was designed to measure changes in students’ commitment to learning science over time, and to
identify influences on students’ commitment to and interest in science among the student’s teachers, peers, and
family (Simpson & Troost, 1982). Recently, the instrument was reevaluated and shortened from 58 items in 14
subscales to 22 items in 5 subscales using methods of EFA and CFA (Owen et al., 2008).
Because of our interest in younger students, we paid particular attention to the modified Attitudes Toward
Science Inventory (mATSI), which assesses changes in the attitudes of urban fifth-grade students (ages 10–11)
due to an intervention (Weinburgh & Steele, 2000). This instrument is a simplified version of the ATSI,
developed for use with college students not majoring in science (Gogolin & Swartz, 1992); Weinburgh and
Steele cut questions and simplified phrasing of items to be appropriate for pre-adolescent children. The mATSI
designates 5 subscales, including one addressing students’ attitudes toward school science and another asking
about the value of science to the world.
The Middle School Students’ Attitude to Mathematics, Science, and Engineering (MS-AMSE) Survey was
developed to investigate students’ interest in and knowledge about potential careers in engineering (Gibbons,
Hirsch, Kimmel, Rockland, & Bloom, 2004). The survey was adapted from a longer version developed for use
with high school students (Hirsch, Gibbons, Kimmel, Rockland, & Bloom, 2003). In addition to asking students
about engineering careers, the survey included items probing students’ attitudes and feelings of efficacy toward
mathematics and science.
We adapted the MS-AMSE (Cunningham & Lachapelle, 2010) for use with elementary school students to
measure attitudes toward and interest in engineering careers before and after participation in the Engineering is
Elementary curriculum (EiE). We used this instrument, the Elementary Engineering Attitudes (EEA) survey, a
precursor to the EIA, throughout the development of EiE as part of formative evaluation, and found that girls
showed interest in more socially or environmentally responsive engineering fields (e.g., biomedical engineering)
while boys were more likely to express interest in engineering of vehicles or structures. We also found that
interests and attitudes of EiE participants became more positive, with the attitudes of girls lower on the pretest
than those of boys; however, the gap in interest and attitudes closed after participation. Pretest scores for
subscales had much lower reliability, however, than posttest scores.
Our research presumes that student interest and attitudes toward engineering will vary with the context. A
personally relevant, engaging context is likely to affect students’ attitudes positively (Ainley & Ainley, 2011);
but some of the impact may be of short duration. An intervention that focuses primarily on “fun,” in particular,
may have only short-term effects (Appelbaum & Clark, 2001). Therefore, an instrument (and an intervention)
should focus on more aspects of attitude and interest than simply the emotional impact of an intervention.
Purpose
Many proponents of increasing student exposure to engineering claim that introducing engineering to greater
numbers of young students will increase and diversify the population pursuing engineering careers. An
important assumption of this claim is that students will become more interested in engineering and more
positive in their attitudes as they engage in engineering experiences in and out of school. To measure this
assumption, we have developed an Engineering Interest and Attitudes (EIA) questionnaire, intended to be used
to measure the impact of an engineering intervention on the interests, attitudes, and gender biases of elementary
school students.
In this paper, we lay out the evidence for the quality of the EIA instrument. This includes evidence for internal
consistency reliability and validity of the subscales. We detail evidence of content validity, including the use of
prior instruments and research to form the questionnaire, and the interpretation and judgment of survey
questions by content experts. We also include evidence for response processes gathered from individual
interviews with students, and we describe evidence based on the internal structure of the instrument by
comparing our original theoretical constructs with the results of EFA and CFA.
Method
Research
The EIA questionnaire was designed in the context of an efficacy study of the EiE curriculum, Exploring the
Efficacy of Engineering is Elementary (E4). EiE had been under intensive development and formative
evaluation from 2004 to 2010. The E4 study was designed as a cluster randomized trial (CRT) between EiE and
a comparison curriculum. The study collected data on student achievement outcomes and fidelity of
implementation, as well as student interests and attitudes in engineering.
Students participated in either the treatment or comparison engineering curriculum. Teacher volunteers were
recruited for this study through their principals and superintendents. Teachers applied to participate as teams of
2–4 teachers from the same school. Only teachers from schools that had not implemented engineering curricula
were accepted. Once the recruitment and acceptance process was completed, cohorts of teachers at the school
level were randomized into either the treatment or comparison group.
The treatment curriculum is designed from a social constructivist theoretical framework, based on the belief that
students learn deeply the key practices and content of a discipline through meaningful engagement in its
epistemic practices at a developmentally appropriate level (Duit & Treagust, 1998; Sawyer, 2006). The
treatment curriculum meets the criteria for project-based learning, where students focus on a design challenge
that engages them with key ideas in science and engineering. The central project is open-ended, where students
are engaged in the problem with a realistic context, and heavy scaffolding is provided to students, to support
them as they use engineering practices and reasoning. Although the comparison curriculum also includes hands-
on challenges, the challenges are not motivated with a context, no scaffolding is provided, many challenges are
not open-ended, and information is given through direct instruction.
Instrument Development
The E4 project was to collect data from upper elementary students aged 8–11, so the E4 project team searched
the literature for instruments addressing interest in and attitudes toward engineering and science that were
designed for younger populations of students. Some of the best instruments we found, however, were designed
for older students, those in middle and high school. From such instruments, we chose scales with simpler
phrasing, and avoided scales with complex language that we deemed would exceed the reading abilities of our
younger subjects.
Instruments addressing science were more abundant and better tested than those addressing engineering, so we
decided to duplicate our chosen science scales and items, replacing the word “science” with “engineering.” We
initially worked to assemble an instrument that included both science and engineering items, because our units
were testing both science and engineering content knowledge—teachers participating in E4 were required to
teach science content that was relevant to their engineering unit. In looking for scales and items, we chose to
address five of the most common topics surveyed in science attitude instruments as reported by recent literature
reviews (Blalock et al., 2008; Tytler, 2014), as noted in the first column of Table 1. The second and third
columns of Table 1 show the evolution of subscales over time, and will be explained further in subsequent
sections.
From the SAI-II (Lichtenstein et al., 2008; Moore & Foy, 1997), we chose to use items from the “I want to be a
scientist” scale, as identified by Lichtenstein et al. (2008), to measure students’ interest in pursuing engineering
and science careers. This scale consists of eight items (Cronbach’s α=.810), some of which are expressed as
negatives, such as, “Scientific work would be too hard for me.” It also includes an item more aligned with
attitude than aspirations, “I enjoy studying science.”
Table 1. Subscales identified and named at each stage of analysis

After Literature Review After Qualitative Analysis After Item Reduction
Desire to Learn Science Value of Engineering to Me Value of Engineering to Me
Self-Efficacy in Science Self-Efficacy in Engineering (not retained)
Enjoyment of Science Enjoyment of Engineering Enjoyment of Engineering
Interest in Science Careers Aspirations for Engineering Aspirations for Engineering
Attitudes toward School Science Attitudes toward School Attitudes toward School
Engineering Engineering
Value of Science in Society Value of Engineering to Society Value of Engineering to Society
(added) Gender Bias Gender Bias

We chose to test the 14 items addressing attitudes toward school science from the ATSSA (Germann, 1988).
Items include “I would like to learn more about science,” and “Science is fascinating and fun.” From the
mATSI, we pulled 17 items from three of five subscales as candidates for testing (Weinburgh & Steele, 2000);
scales included “Value of Science in Society,” “Self-Concept of Science,” and “Desire to Do Science.” From the
STAQ (Owen et al., 2008), we chose to examine 14 items from the three subscales “Motivating Science Class,”
“Self-Directed Effort,” and “Science is Fun for Me.” Some items were redundant across scales, either exactly or
similarly, but we used such items only once in the questionnaire, blending or choosing between similar
questions. We also incorporated eleven of the items from the EEA (2010) addressing the value of science and
engineering. Finally, we chose to develop five new questions to assess gender biases in engineering attitudes.
We chose to implement the survey as a post-only Likert-scale questionnaire, with a range of prompts from
“Strongly Disagree” to “Strongly Agree” (see Figure 1). Students were asked to answer each question twice:
once to the prompt “Last summer, I would have said:” and also to the prompt “Now I would say.” We chose to
implement the survey in this way knowing that the students in our study were likely to know little to nothing
about engineering before engaging in the curriculum, and we had learned from prior experience with the EEA
that children’s responses regarding engineering before engineering instruction tended to be unreliable given
their lack of a clear sense of what engineering is. We considered that this may be due to a “response shift,”
whereby students have a better sense of how to self-evaluate at the end of an intervention than they do prior to
the intervention (the “retrospective pre”), as other researchers have found to be the case (e.g., Bhanji,
Gottesman, de Grave, Steinert, & Winer, 2012; Sibthorp, Paisley, Gookin, & Ward, 2007). By asking about
“before” and “now” after engineering instruction, we hoped to get more reliable data about students’ change in
attitudes by having them compare their current attitudes and interests to what they remembered of their prior
interests and attitudes. Though we expect this retrospective will introduce some bias to “before” responses, we
expect this will be outweighed by students’ ability to give more informed responses.
Figure 1. Image from the EIA assessment
Qualitative Evidence of Content Validity
Our first goal was to gather expert opinions on the content validity of the 64 items we had collected, plus ten
more we developed to investigate gender stereotypes. We solicited opinions from four experts on science and
engineering assessment and education within our institution, as well as from a former engineer, now an
educator. We asked the experts to read the items, think about how they and their students might answer them,
and comment on possible problems with content, readability, or wording of the items. We also asked engineers
to comment on the original subscale naming. With the assistance of these experts, we flagged items that were
possibly inappropriate or likely to be misinterpreted by our target age group, made adjustments to the scale
names, and confirmed that assignments of items to subscales was considered appropriate. Thirty-three items
from this list were dropped, generally because they were not appropriate for the age group, for example, items
that referenced a “science course” or “science teacher,” because American elementary school children are often
taught all subjects by one or two teachers in their primary classroom. Seven items were modified to simplify
vocabulary or sentence structure; for example, “Science is of great importance to a country’s development” was
modified to read “Science is of great importance to my country.” Twenty-one items were added to the list,
duplicating some items but referencing “engineering” instead of “science”; for example, “I enjoy studying
engineering” was added to the list in parallel to the item “I enjoy studying science.”
After expert review, we had a list of 62 items to test with students. To test the items for validity of response
processes, we conducted cognitive interviews with 15 students in the target grade range (grades 3–5, ages 8–11),
some from classrooms that had implemented in-school engineering curricula, and others from out-of-school
time (OST) programs engaged in engineering units of exploration. During the interviews, we asked students to
read each question aloud and talk about it. We asked them to explain any confusing aspects of the question, and
to talk aloud about what they were thinking as they chose answers from the Likert scales. Based on the results of
these interviews, we dropped one question that students had difficulty reading, “Engineering solutions to
problems would be boring work.” We revised eight questions to simplify the phrasing and re-tested them: for
example, “No matter how hard I try, I cannot understand engineering” was changed to “Engineering is really
hard to understand.” We also had several classes of students in the target age range complete the questionnaire
without interviews, and found that it was taking them much too long—more than 45 minutes. To shorten the
questionnaire we decided to drop all 31 questions that referenced science rather than engineering, as data about
science attitudes were less important to us than data about engineering interests and attitudes for our engineering
curriculum study.
Once again, after our revisions based on testing of suitability with the target population, we asked our experts to
review the resulting 30 candidate items and subscales for validity of content. After combining and dropping
items, we finalized six candidate subscales (see Table 1). The second review resulted in few additional
suggestions for revision—all minor edits of wording.
Data Collection
As part of the Exploring the Efficacy of Engineering is Elementary (E4) study, we collected post-surveys of
students’ interests in and attitudes toward engineering. Over two years, we collected surveys from almost 11,000
students in grades 3, 4, and 5. Students spanned a wide range of racial and SES demographic groups, from rural,
urban, and suburban areas of several geographically non-contiguous American states. See Table 2 for the
demographic breakdown of the sample.
Table 2. Student demographic breakdown of sample, reported as percentages

Male Minority1 FRL2 English Grade 3 Grade 4 Grade 5 Total N
Learners
Initial Sample
Comparison 50.3 37.1 46.3 6.5 23.8 31.4 44.7 5,994
Treatment 51.4 31.7 43.8 5.8 32.3 36.7 31.0 4,912
Total 50.8 34.6 45.1 6.2 27.6 33.8 38.5 10,906
Final Sample after drops (due to incomplete surveys):
Comparison 49.7 35.4 45.4 6.3 22.6 31.9 45.6 5,385
Treatment 51.4 30.2 42.6 5.5 31.9 36.8 31.3 4,417
Total 50.5 33.1 44.1 5.9 26.8 34.1 39.1 9,802
1
Percentage of students from underrepresented minority groups (African American, Latino/a, Mixed-race,
Other).
2
Percentage of students receiving Free or Reduced-Price Lunch.
Item Reduction
Initially, we tested the instrument with a portion of the first year of data collected for the E4 study, returned in
the first three months of the study. The purpose of the initial testing was to drop items not adding to the value of
the questionnaire, to shorten it and reduce the burden on class time. The initial sample totaled 1,563 students
from grade 3–5 classrooms that had implemented one or two engineering curricular units. Students completed
the questionnaire independently as a written assessment. To gather evidence for the validity of the internal
structure of the questionnaire, we examined the internal consistency reliability of items (Cronbach’s α) and
conducted a Principal Components Analysis (PCA) in SPSS version 22 (IBM Corporation, 2013) to determine
which items contributed least to the total variance within the set of items (Dillon & Goldstein, 1984). Toward
our goals, we dropped six items from the 30 tested that performed particularly poorly, failing to load onto a
component, or detracting from internal consistency reliability. For example, the item “I do not want to be an
engineer” was dropped because it had particularly low initial and extraction communalities (<.1); its removal
increased Cronbach’s α, and it failed to load onto any component. E4 subjects completing the EIA after item
reduction analysis received the 24-question version.
Exploratory Factor Analysis
Before final analysis of the instrument, we randomly split our sample in half to conduct an EFA and CFA on
separate samples. The purpose of EFA is to describe or explore the relationships between items that are
interrelated, to describe common factors (groupings of items) that are expected to correspond to theorized latent
(unobserved) variables. EFA was conducted for this study because the items used in the instrument had not been
previously analyzed together for the purpose of ensuring that an interpretable factor structure was possible, and
CFA should not be run until the structure has been studied using EFA with a separate, independent sample
(Bandalos & Finney, 2010).
With the first subsample, we used Parallel Analysis (PA), a method of comparing the eigenvalues of a specific
sample with estimated population eigenvalues, to determine what number of factors was likely to be significant.
To conduct PA, we used a script from https://people.ok.ubc.ca/brioconn/nfactors/nfactors.html (O’Connor,
2000) in SPSS 24 to assist in estimating the number of factors, backed up by examination of the scree plot and
eigenvalues, and comparison to the intended subscales; Bandalos and Finney (2010) recommend the use of
multiple methods and the criterion of theoretical plausibility to determine the number of factors, with preference
given to a choice for which multiple methods and theoretical plausibility converge.
EFA was conducted in Mplus 7.4 (Muthén & Muthén, 2015). We used the MLR estimator, an extension of
Maximum Likelihood (ML) estimation that is robust to multivariate non-normality, and adjusts for missing data
using Full Information Maximum Likelihood (FIML), to handle the non-normality of our ordinal 5-point Likert-
scale data. The ratio of sample size (5,390) to expected factors (<10) is quite high (539:1) so we expect that the
sample size is sufficient for this procedure, even when extracted communalities are low (MacCallum, Widaman,
Zhang, & Hong, 1999). We used an oblique rotation (Geomin) with the EFA because we expected the resulting
factors to be correlated. We examined the structure matrices for correlations between items and factors, and the
pattern matrices for item loadings and cross-loadings, using base thresholds of structure coefficient >.450 and
pattern coefficient >.300 for considering an item as loading onto a factor; because our factor correlations are
strong and sample size is large, structure coefficients generally are expected to be larger than pattern coefficients
(Bandalos & Finney, 2010; Brown, 2006).
We also considered goodness-of-fit information that is available with an ML-based EFA. Three kinds of fitness
measures are available for testing models: measures of absolute fit, comparative fit, and parsimonious fit
(Kelloway, 2015). Measures of absolute fit test how closely the covariance matrix for the model matches the
covariance matrix for the input (baseline) data. Measures of comparative fit give information about which of
two competing models better matches the covariance matrix for the input data. Measures of parsimonious fit are
a type of comparative measure that adjust negatively for the loss of degrees of freedom due to specifying more
parameters for a model—because, all else being equal, the specification of more parameters will always lead to
a better fit to the covariance matrix. We used the root mean square error of approximation (RMSEA) and the
standardized root mean square residual (SRMR) as measures of absolute fit, with cutoffs of <.05 for the former
and <.80 for the latter; we used the comparative fit index (CFI) as a measure of comparative fit, with a cutoff of
>.95 indicating good fit; and we used the Akaike information criterion (AIC) as a measure of parsimonious fit to
compare models, with a smaller value indicating a superior model (Kelloway, 2015). We also report the χ 2
statistic, which can be used as a measure of absolute fit, with a difference between the fitted model and baseline
model of p<.05 traditionally indicating good fit. Our purpose was to explore possible factor structures and
compare the fit of a variety of candidate factor solutions, so we could ensure that a structure could be specified
where the correlations of items with factors was interpretable and reasonably congruent to theorized latent
dimensions before embarking on CFA.
Confirmatory Factor Analysis
Our purpose in conducting CFA was to cross-validate the factor structure developed by theory and refined with
EFA (Brown, 2006). For CFA, in contrast to EFA, all indicators (survey items) and their relationships to latent
variables (factors) must be specified in advance, to “confirm” the validity of the theorized model. Kelloway
(2015) recommends that, because CFA is strongest for comparing models, the best approach is to identify
ambiguous aspects of the model to be tested, and to specify nested models that remove ambiguous aspects of the
full model to test their contribution to the model. The theorized relationships between latent variables can be
tested by specifying nested models that contain a subset of the parent model’s parameters. The parent and nested
models can then be compared to determine which model specification is the best fit for the data (Brown, 2006).
Therefore, before beginning CFA analysis, we generated a nested, competing model from a single, fully
specified parent model containing the relationships between all latent variables freely specified, as well as the
full set of relationships between observed indicators and latent variables that resulted from EFA. The nested
model set all cross-loading parameters to zero, effectively removing them from the model. We used CFA to
compare the fit of the parent and nested models.
Using the second subsample that we generated before EFA, we conducted CFA using Mplus 7.4. Sample data
was input to Mplus, which generated variance-covariance matrices for analysis. All models used MLR as
estimator. We used RMSEA and SRMR as measures of absolute fit, CFI as a measure of comparative fit, and
AIC as a measure of parsimonious fit to compare models. We also report the χ2 statistic; however, we compared
our nested models using an adjusted scaled difference χ2 test statistic (Satorra & Bentler, 2010), which is
necessary because the simple difference between two scaled χ 2 statistics from MLR does not have a χ2
distribution.
To determine the quality of the model, we examined the parameter estimates for significance and interpretability
(Brown, 2006). Mplus provides, in addition to the unstandardized and standardized parameter estimates, the
standard error, z-statistic, and z-test p-value for each parameter for each parameter estimate, which we inspected
and reported. Non-significant parameters should be considered for removal from the model. Standard errors
were inspected for excessively large values, which would indicate an unreliable parameter estimate. For items
that are not cross-loading, the completely standardized factor loading represents the correlation between the item
and factor, and the R2 statistic represents the communality, that is the proportion of variance of the item that is
explained by the factor. We examined the size of factor loadings and R2 values as further evidence for whether
item-factor relationships are strong enough to be meaningful. Finally, we examined the factor determinacy of
factor scores as a measure of factor score quality, with a threshold of >.8 for a good-quality factor score and >.9
as preferred (Grice, 2001). Mplus provides factor score determinacies, which are a measure of the correlation
between generated factor scores and the latent factor estimate (available on request by email with the first
author).
Results and Discussion

Exploratory Factor Analysis and Subscale Development
EFA was conducted concurrently on each set of items (PRE and NOW), to ensure a factor structure that fit both
the PRE, “Last summer, I would have said,” and NOW, “Now I would say,” responses. We expected to see
differences between the PRE and NOW sets of items, because we knew that students were likely not to have had
prior engineering experiences so were likely to report weaker opinions (or possibly stronger in the case of
Gender bias) on the PRE questions. However, to facilitate PRE-NOW comparisons, we would need one
consistent model. Therefore, throughout the EFA process, we worked to find the best-fitting factor structures for
both the PRE and NOW sets of items that also made the most sense theoretically. Where there were differences
in the pattern of coefficients for each set, a compromise was made, and we chose the thematically most sensible
placement. This led to the choice of factor structure that may not have been the statistically best fit for either set
of items. However, the final structure chosen was a good fit for each set of items, and made sense given the
theoretical framework.
To begin, we ran PA with a 99% probability cutoff on the random half-sample 1 prior to EFA. To determine the
number of factors indicated for analysis, we compared the sample data eigenvalues to the randomly generated
data percentile eigenvalues (Table 3). A factor where the sample data eigenvalue exceeds the random data
percentile eigenvalue is retained. Analysis indicates that seven factors can be extracted for the PRE variables,
and eight for the NOW items. With the traditional cutoff of eigenvalues >1, however, only 2 factors are
indicated. Examination of the scree plots with the PA simulated data superimposed as a gently sloped line
(Figure 2) shows a sharp drop in sample data eigenvalues after one factor, with a softer elbow curving down to
near horizontal including another five factors. To examine a range of possible factor structures as indicated by
PA (7-8 factors), our theorized factor structure (6 factors), the scree plots (6 factors), and the much smaller
number of eigenvalues>1 (2 factors), we decided to conduct EFA to fit 4 to 8 factors. We chose not to examine
2 or 3 factors because this was so much less than our theorized 6 factors and the results of other methods for
estimating the number of factors.
7
PRE Eigenvalues 1-24
6
Raw Data Eigenvalue
5 Generated Data 95 percentile

Eigenvalue
-1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
7
NOW Eigenvalues 1-24
6
Raw Data Eigenvalue
5 Generated Data 95 percentile

Eigenvalue
-1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 2. Scree Plots of “Pre” (top) and “Now” factors extracted, Random Half 1
Examination of goodness-of-fit estimates showed that structures with more factors tended to fit the data better,
especially as compared to 4 or 5 factors, although the difference between the best-fitting 7- and 8-factor
structures was very small (see Table 4). Table 5 shows the structure and pattern matrices for PRE/NOW, with
coefficients for the chosen scales in bold, and cross-loading items shown in italics for the secondary loading.
We were unable to calculate χ2 or other fit coefficients for 5 factors for the PRE data, but as other factor
structures fit the data better and had better fit statistics than the 5-factor NOW structure, we decided not to
explore this further. The factor structure that best mirrored our theorized subscales, which we had tested
qualitatively and examined using PCA (Table 1), included 7 factors. All 6 factors we had theorized were
identifiable in the analysis for NOW data, and for the PRE data 5 of 6 were present (the seventh factor had no
significant loadings). We expect that students were expressing weaker opinions about items loading on the two
factors “Enjoyment” and “Aspirations” before they participated in engineering in school. Another difference
between the factor structure and our theorized structure is that the Gender bias factors were, in each case, split
across two factors: Male bias (items asking whether boys were better/girls had a harder time with engineering)
and Female bias (with girls and boys swapped in the items, but using the same phrasing). The split of the
Gender bias factor accounts for the remaining factor in each case.
Table 3. Parallel Analysis for PRE and NOW items

“PRE” Eigenvalues “NOW” Eigenvalues
Random Data Random Data Random Data Random Data
Factor Sample Data Sample Data
Mean Percentile Mean Percentile
1 7.067699 .142976 .175269 6.384192 .142844 .174591
2 1.196063 .122523 .145149 1.433322 .122370 .143114
3 .986179 .106976 .126354 .886926 .106737 .125686
4 .571639 .093268 .110253 .778632 .092936 .110052
5 .369565 .081129 .097095 .341110 .080612 .096031
6 .204882 .069291 .084135 .203052 .069121 .083231
7 .084881 .058362 .072131 .117400 .058076 .073497
8 .053478 .047825 .061380 .100461 .047626 .062739
9 .005873 .037199 .050885 .026821 .037760 .051062
Table 4. Fit indices for EFA models, PRE and NOW

# Factors # Parameters χ2 df AIC CFI RMSEA SRMR
4 PRE 138 1383.8 186 403257 0.96 0.034 0.02
4 NOW 138 1294.2 186 364866 0.96 0.034 0.02
5 PRE Could not be computed.
5 NOW 158 821.6 166 364113 0.98 0.027 0.02
6 PRE 177 634.5 147 402353 0.99 0.025 0.01
6 NOW 177 580.6 147 363774 0.99 0.023 0.01
7 PRE 195 395.6 129 402112 0.99 0.019 0.01
7 NOW 195 366.5 129 363512 0.99 0.018 0.01
8 PRE 212 364.8 112 402026 0.99 0.020 0.01
8 NOW 212 281.1 112 363402 0.99 0.017 0.01
All χ2 tests of model fit were significant (p<.0001).
Both the structure and pattern matrices were interpreted in making decisions for subscale loadings. Items that
loaded on either the PRE or NOW structure matrix with a coefficient >.450 were considered, as were items that
loaded on either the PRE or NOW pattern matrix with a coefficient >.300. In comparing the content of items
with the strongest coefficients, it became clear that the PRE factor corresponding to the “Enjoyment of
engineering” subscale also included the “Aspirations” items, while the NOW coefficients for those items were
loaded across two factors, one “Enjoyment” and the other “Aspirations.” For this reason, PRE coefficients in the
Table 5 “Enjoyment” column are bolded or italicized as half pairs when they are intended to be paired with the
NOW coefficients in the “Aspirations” column in the final, compromise model.
All item-factor loadings with at least three out of four (PRE, NOW, pattern, and structure) coefficients larger
than the thresholds were chosen for subscales. Only two item-factor loadings with fewer than three above-
threshold coefficients were chosen as secondary, cross-loading items on factors: Item 5 “I would like to work
with other engineers to solve engineering problems” (which loaded on the “Aspirations” and “Value to me”
subscales) and Item 10 “It is important to understand engineering in order to get a good job” (which loaded
primarily on the “Value to society” subscale and secondarily on the “Aspirations” subscale). Each of these two
item-factor loadings was chosen because of the content of the item, which sensibly cross-loaded, and because
some of the below-threshold coefficients were similar in value to the above-threshold coefficients for the same
item’s primary factor loading.
Table 5. Structure and pattern matrices for PRE/NOW

Item # Value to Aspirations
Value to me Enjoyment Male bias Female bias School
-Type society (NOW only)
1-S .506/.552 .470/.345 .454/.418
1-P .378/.404 .105/.018 .261/.156
2-S .493/.521 .465/.461
2-P .408/.400 .385/.298
3-S .563/.576 .658/.675 .603 .495/.353
3-P .332/.359 .502/.497 .077 .016/-.055
4-S .758/.808
4-P .759/.826
5-S .434/.501 .503/.465 .559
5-P .257/.293 .280/.079 .351
6-S .534/.557 .527/.528 .482 .626/.618
6-P .334/.311 .012/.228 -.006 .476/.394
8-S .484/.496 .710/.794 .481/.358 .661 .550/.439
8-P .200/.183 .509/.666 .010/.000 .050 .025/.042
9-S .498/.443 .478/.360 .570/.489
9-P .023/.252 .095/.077 .352/.298
10-S .463/.261 .521/.372 .374 .474/.283
10-P .091/-.074 .335/.239 .238 .112/.073
13-S .673/.734 .522/.394 .650 .634/.491
13-P .383/.510 .009/.042 .173 .287/.154
14-S .692/.650 .478/.310
14-P .682/.669 -.024/-.065
15-S -.450/-.399
15-P -.443/-.359
17-S .496/.340 .567/.486 .479/.343
17-P .160/.223 .397/.403 .044/.054
18-S .659/.486 .490/.287 .665
18-P .645/.007 .179/.068 .688
19-S .640/.768
19-P .641/.753
20-S .929/.827
20-P .929/.843
21-S .461/.283 .660/.559 .555/.436
21-P .029/-.032 .557/.430 .180/.173
22-S .557/.431 .594/.496 .456 .723/.687
22-P .029/.008 .178/.135 .148 .585/.538
23-S .478/.373
23-P .362/.281
24-S .617/.569 .487/.342
24-P .530/.521 .111/.042
25-S .530/.430 .527/.448 .598/.469
25-P .198/.167 .145/.220 .381/.225
26-S .765/.688 .516/.371 .806 .607/.469
26-P .747/.152 -.005/-.014 .683 .134/.088
28-S .734/.757
28-P .736/.750
30-S .734/.657 .461/.346 .805 .530/.378
30-P .779/.102 .005/-.031 .700 .006/.024
Bold/italics marks items in subscales. Coefficients were omitted when none of the set met minimum thresholds.
Because the intended “Gender bias” subscale split across two factors, we decided to drop the “Female bias”
factor, as only two items loaded on it, while the “Male bias” factor also captured Item 15, and the wording of its
items more strongly corresponded to traditional gender stereotypes for engineering. Table 6 lists the final
choices and full text of items with their primary loadings onto subscales, and, where applicable, a secondary
loading.
Table 6. Items in final subscales

Item Subscale Cross-Loading Text of Item from the EIA Questionnaire
#
8 Enjoyment Engineering is fun
13 Enjoyment I am interested when we do engineering in school
23 Enjoyment Engineering is easy for me
3 Enjoyment Value to me I enjoy studying engineering
1 Value to me It is important for me to understand engineering
2 Value to me Value to society Engineering helps me understand today’s world
6 School Value to me We learn about interesting things when we do engineering in school
9 School When we do engineering, we use a lot of interesting materials & tools
22 School We learn about important things when we do engineering in school
25 School Value to society I try hard to do well in engineering
14 Value to society Engineers help make people’s lives better
17 Value to society I know what engineers do for their jobs
21 Value to society Engineering is useful in helping to solve the problems of everyday life
24 Value to society Engineering is really important to my country
10 Value to society Aspirations It is important to understand engineering in order to get a good job
30 Aspirations I really want to learn engineering
18 Aspirations I would enjoy being an engineer when I grow up
26 Aspirations I would like to learn more about engineering
5 Aspirations Value to me I would like to work with other engineers to solve engineering
problems
4 Gender bias Boys are better at engineering than girls
28 Gender bias Girls have a harder time understanding engineering than boys
15 Gender bias Girls and boys are equally good at engineering
19 Dropped Boys have a harder time understanding engineering than girls
20 Dropped Girls are better at engineering than boys
To assess the relations between factors, we examined the correlation matrices for the PRE and NOW 7-factor
extractions (Table 7). None of the factors were excessively correlated, which would be an indication that factors
should be combined. Only “Male bias” and “Female bias” had low enough correlations with other factors to be
statistically insignificant. “Value to me” was the only factor correlated with the two gender bias subscales on the
PRE, a relationship that may require analysis by gender to understand. On the other hand, most of the other
variables (except “Aspirations”) were significantly and negatively (though only mildly) correlated with the
NOW gender bias subscales, indicating that less bias was associated with more positive attitudes of enjoyment
of engineering in general and at school, and more positive assessment of the value of engineering to society.
Table 7. Correlations between factors extracted by EFA with MLR estimators PRE/NOW
Value to me Enjoyment Value to society School Aspirations Male bias
Value to me 1.000
Enjoyment .451/.411 1.000
Value to society .259/.384 .643/.373 1.000
School .382/.411 .724/.461 .706/.527 1.000
Aspirations X/.497 X/.757 X/.431 X/.425 1.000
Male bias .049/-.047 .000/-.109 -.029/-.138 -.051/-.200 X/-.021 1.000
Female bias .061/-.034 .030/-.051 -.021/-.117 -.005/.199 X/.022 .004/-.109
X: No factor corresponding to “Aspirations” was found in the PRE data. Bold correlations are significant
(p<.05).
Confirmatory Factor Analysis
As with EFA, CFA was conducted on the PRE and NOW datasets in parallel, to ensure that the final model
would fit well for both datasets. Using Mplus 7.4, we tested the factor structure specified in Table 6 with the
second random half of the full dataset, to cross-validate the factor structure with new data. All CFA models
were estimated with MLR, which allows for the estimation of missing values using FIML.
Table 8. Model Information

Degrees of # Free #Observations # Missing data patterns
Freedom Parameters PRE NOW PRE NOW
Model 1 188 87 5495 5507 310 294
Model 2 194 81 5495 5507 310 294
We chose to examine and compare two nested models. The models are illustrated in Figure 3, model
information is given in Table 8, and model specifications are presented in Table 9. The primary difference
between Model 1 and Model 2 is the specification of cross-loading terms. Model 1 includes all of the cross-
loading item-factor relationships specified in Table 6, and has 188 degrees of freedom. Model 2 contains no
cross-loading items, and therefore has more degrees of freedom: 194. For each latent variable, we chose a
marker indicator, an item that had a high pattern coefficient on the corresponding EFA factor and low cross-
loading pattern coefficients (see Table 5). Our initial choice of marker indicators was successful for all but one
latent variable: Value to me (initial choice EIA_1). In this case, examination of the Modification Indices for the
model showed it would be substantially improved by freeing EIA_1 and substituting EIA_2 as the marker
indicator.
Table 9. Model specifications

Model 1 Indicators Model 2 Indicators
Marker (Fixed) Freely estimated Marker (Fixed) Freely estimated
Enjoyment EIA_8 EIA_3, 13, 23 EIA_8 EIA_3, 13, 23
Value to me EIA_2 EIA_1, 3, 5, 6 EIA_2 EIA_1
School EIA_22 EIA_6, 9, 25 EIA_22 EIA_6, 9, 25
Value to society EIA_14 EIA_2, 10, 17, 21, 24, 25 EIA_14 EIA_10, 17, 21, 24
Aspirations EIA_30 EIA_5, 10, 18, 26 EIA_30 EIA_5, 18, 26
Gender bias EIA_4 EIA_4, 15, 28 Variance@1 EIA_4, 15, 28
Overall goodness-of-fit was very good for both models (Table 10). All fit indices met threshold requirements,
except the CFI measure of parsimonious fit: values for the PRE models slightly missed the threshold (0.94 <
0.95). CFI values for Model 1 NOW were better than those for Model 2. RMSEA 95% confidence intervals
were under the 0.05 threshold.
The AIC measure of comparative fit was smaller for the Model 1 PRE and NOW than for the corresponding
values for Model 2, indicating that Model 1 is the better fit to the data. To further compare the models for
goodness-of-fit, we calculated the Satorra-Bentler scaled χ2 difference test (TRd), to compensate for the MLR χ2
having a different distribution than the standard χ2 distribution. The value for the PRE models is TRd=243, and
the value for the NOW models is TRd=291; positive values indicate that the model with fewer degrees of
freedom (Model 1) is the better model. As all indications are that Model 1, with cross-loading indicators, is
superior to Model 2 without cross-loading, we proceed to specify Model 1 in the remainder of this paper.
Table 10. Fit indices for nested models, PRE and NOW
Model χ2 df AIC CFI RMSEA RMSEA 95% CI SRMR
1 PRE 1953.6 188 367792 0.94 0.041 0.040 - 0.043 0.060
1 NOW 1294.2 194 328411 0.96 0.033 0.031 - 0.034 0.046
2 PRE 2169.2 188 368076 0.94 0.043 0.041 - 0.045 0.061
2 NOW 1539.5 194 328777 0.95 0.035 0.034 - 0.037 0.049
All χ2 tests of model fit were significant (p<.0001).
Figure 3. Two nested models to compare

Table 11. Standardized parameter estimates for Model 1: PRE / NOW.

Parameter Residual Variance
Name Estimate S.E. P-Value Estimate S.E. P-Value R-Square
Enjoyment BY
EIA_8 0.740/0.804 0.009/0.009 0.000/0.000 0.452/0.354 0.013/0.014 0.000/0.000 0.548/0.646
EIA_3 0.504/0.612 0.043/0.027 0.000/0.000 0.528/0.484 0.012/0.015 0.000/0.000 0.472/0.516
EIA_13 0.716/0.767 0.009/0.010 0.000/0.000 0.488/0.412 0.013/0.015 0.000/0.000 0.512/0.588
EIA_23 0.492/0.375 0.013/0.016 0.000/0.000 0.758/0.859 0.013/0.012 0.000/0.000 0.242/0.141
Value to me BY
EIA_2 0.447/0.361 0.030/0.035 0.000/0.000 0.676/0.684 0.016/0.017 0.000/0.000 0.324/0.316
EIA_1 0.654/0.715 0.018/0.023 0.000/0.000 0.572/0.489 0.023/0.032 0.000/0.000 0.428/0.511
EIA_3 0.215/0.155 0.046/0.033 0.000/0.000 0.528/0.484 0.012/0.015 0.000/0.000 0.472/0.516
EIA_5 0.341/0.211 0.040/0.029 0.000/0.000 0.682/0.643 0.014/0.014 0.000/0.000 0.318/0.357
EIA_6 0.367/0.150 0.056/0.055 0.000/0.006 0.559/0.504 0.015/0.018 0.000/0.000 0.441/0.496
School BY
EIA_22 0.701/0.666 0.010/0.014 0.000/0.000 0.509/0.557 0.015/0.018 0.000/0.000 0.491/0.443
EIA_6 0.341/0.588 0.055/0.047 0.000/0.000 0.559/0.504 0.015/0.018 0.000/0.000 0.441/0.496
EIA_9 0.620/0.569 0.011/0.015 0.000/0.000 0.616/0.676 0.014/0.017 0.000/0.000 0.384/0.324
EIA_25 0.501/0.423 0.071/0.042 0.000/0.000 0.628/0.688 0.014/0.017 0.000/0.000 0.372/0.312
Value to society BY
EIA_14 0.655/0.604 0.011/0.015 0.000/0.000 0.571/0.636 0.015/0.018 0.000/0.000 0.429/0.364
EIA_10 0.423/0.262 0.029/0.022 0.000/0.000 0.683/0.814 0.013/0.012 0.000/0.000 0.317/0.186
EIA_17 0.608/0.511 0.011/0.017 0.000/0.000 0.631/0.739 0.014/0.017 0.000/0.000 0.369/0.261
EIA_21 0.671/0.600 0.011/0.015 0.000/0.000 0.550/0.640 0.015/0.018 0.000/0.000 0.450/0.360
EIA_24 0.606/0.571 0.012/0.015 0.000/0.000 0.633/0.674 0.015/0.017 0.000/0.000 0.367/0.326
EIA_2 0.164/0.263 0.027/0.033 0.000/0.000 0.676/0.684 0.016/0.017 0.000/0.000 0.324/0.316
EIA_25 0.120/0.165 0.074/0.046 0.106/0.000 0.628/0.688 0.014/0.017 0.000/0.000 0.318/0.357
Aspirations BY
EIA_30 0.728/0.813 0.009/0.009 0.000/0.000 0.471/0.339 0.014/0.014 0.000/0.000 0.529/0.661
EIA_5 0.269/0.452 0.040/0.026 0.000/0.000 0.682/0.643 0.014/0.014 0.000/0.000 0.318/0.357
EIA_10 0.175/0.230 0.029/0.021 0.000/0.000 0.683/0.814 0.013/0.012 0.000/0.000 0.317/0.186
EIA_18 0.665/0.632 0.011/0.011 0.000/0.000 0.557/0.601 0.014/0.014 0.000/0.000 0.443/0.399
EIA_26 0.768/0.818 0.009/0.008 0.000/0.000 0.411/0.331 0.013/0.014 0.000/0.000 0.589/0.669
Gender bias BY
EIA_4 0.769/0.780 0.017/0.014 0.000/0.000 0.409/0.391 0.026/0.023 0.000/0.000 0.591/0.609
EIA_15 -0.424/-0.408 0.018/0.017 0.000/0.000 0.820/0.834 0.015/0.014 0.000/0.000 0.180/0.166
EIA_28 0.726/0.775 0.017/0.015 0.000/0.000 0.473/0.399 0.025/0.023 0.000/0.000 0.527/0.601
Standardized parameter estimates are listed in Table 11. All parameters are statistically significant (p<.001)
except for Value to me by EIA_6 - NOW (p<.01) and Value to society by EIA_25 - PRE (p=.106). The lack of
significance of the Value to society by EIA_25 parameter for the PRE data also corresponds with an unusually
high standard error for Model 1 (S.E.=0.074). EIA_25 also has a high standard error for its primary loading on
the PRE, with the latent variable School (S.E.=.074). The content of EIA_25 is “I try hard to do well in
engineering” and it may make sense that children who have just completed a questionnaire about their first
experience with engineering might provide unreliable answers to this question. However, we decided not to
drop this indicator because EIA_25 still fits well with the NOW data, with a highly significant estimate (p<.001)
and much smaller standard error (S.E.=.046), much more in line with the range of standard errors for the rest of
the parameter estimates.
For the most part, Model 1 parameters explained substantial item variance: R2 for the non-cross-loading terms
ranged between 0.141 (EIA_23 NOW) and 0.669 (EIA_26 NOW). Most R2 values ranged between .3 and .5,
which corresponds to approximately 30% to 50% of each observed indicator’s variance explained by Model 1.
Disattenuated correlations between the factors are presented in Table 12. All factors except Gender bias had
statistically significant positive correlations with each other (p<.001), indicating that positive interest and
attitudes tend to go hand-in-hand. The Gender bias correlations are theoretically interesting in the pattern of
changes from PRE to NOW, which indicates that before participating in engineering student gender bias was
unrelated to their attitudes and interest, but after participating in engineering students who expressed more
positive attitudes also tended to indicate that they feel less biased about gender. Gender bias is measured in the
negative, as indicated by the signs of the associated indicators: Items 4 and 28, for which larger values indicate
more gender bias (see Table 6), are positively associated with the Gender bias latent factor, while Item 15,
“Girls and boys are equally good at engineering,” is negatively associated with it. Therefore a negative
correlation of Gender bias with the attitude and interest latent factors indicates that less gender bias is associated
with more positive attitudes and interest.
Table 12. Disattenuated correlations between factors Model 1 PRE/NOW

Factor Enjoyment Value to me School Value to society Aspirations
Enjoyment 1.00
Value to me .791**/.621** 1.00
School .899**/.840** .754**/.720** 1.00
Value to society .768**/.584** .668**/.612** .887**/.762** 1.00
Aspirations .932**/.891** .702**/.566** .819**/.734** .732**/.539** 1.00
Gender bias .009 /-.105** -.013 /-.148** -.060* /-.208** -.041 /-.180** -.021 /-.046t
t
p<.05; * p<.01; **p<.001
Table 13 displays the factor determinacies for the refined factor scores derived by Mplus from Model 1 with the
random half 2 sample data. All factors exceed the threshold of 0.8, indicating good quality and replicability.
More than half of the factor determinacies exceed the preferred threshold of 0.9.
Table 13. Model 1 factor determinacies

Enjoyment Value to me School Value to society Aspirations Gender bias
PRE 0.934 0.865 0.918 0.908 0.908 0.858
NOW 0.939 0.839 0.917 0.873 0.936 0.876
Conclusion
Intentions about careers can be shaped as early as elementary school, when few children have a clear sense of,
never mind enthusiasm toward, engineering. Strategies for addressing shortages in the STEM pipeline that target
high school or even middle school students may come too late, particularly to tap into the interests of presently
underrepresented groups, including women, African Americans, and Latinos/as. Therefore many researchers,
policymakers, funders, and educators are working to bring engineering curricula, environments, clubs, and
activities to elementary school children, both in and out of school, to address the pressing need for more young
people interested in and prepared to pursue further education for STEM careers.
Given high interest in addressing the existing STEM pipeline shortages through interventions with younger
children in and out of school intended to positively affect interests and attitudes, an instrument is needed to
measure the impact of such programs. This study shows that the EIA questionnaire has strong evidence of
content and structural validity. The instrument can be used with students ages 8–11 to measure changes in
student enjoyment of engineering, desire to learn engineering, interest in school engineering, aspirations to
become an engineer, and attitudes toward the value of engineering to society. It can also be used to measure
self-reported changes in the level of student gender bias regarding participation in engineering, and the
relationship of gender bias to other engineering attitudes and interests. We expect that researchers and
curriculum developers will want to use this instrument to measure changes in student interests and attitudes after
participation in engineering activities, programs, and curricula.
Recommendations
The sample for this study was diverse with regard to race, ethnicity, and socioeconomic status as well as
geography within the United States, and the EIA instrument likely will perform similarly with like populations.
The sample of English learners was about 6%, so use of the measure in samples with a higher proportion of
English learners may differ. Translations for bilingual classrooms may be necessary to preserve the
characteristics of the questionnaire. No translations of the questionnaire have been tested and the properties of
the instrument for use outside the United States have not been investigated.
These results are specific to school-aged children enrolled in grades 3–5, although many items were adapted
from questionnaires for older children. Use with younger children in a written format might be challenging. Use
of the instrument with older students might be possible, but the factor structure, reliability, and validity
exercises should be revisited.
The largest concern in the use of this assessment is the one experienced in the process of development and
reported by others, which is that among naive elementary-aged children, there is insufficient knowledge of the
concept of engineering to respond meaningfully to questions that use the term. In our circumstances, this was
resolved by exposing the students to an engineering curriculum and administering the questionnaire afterward.
To assess a “pre” time point, students were asked to reflect back to a specific time before their exposure to the
engineering curriculum. An effective method to assess constructs such as desire to learn engineering, perceived
value of engineering, and gendered attitudes around engineering in elementary-/middle-school-aged children
who have not been exposed to an engineering curriculum is yet to be demonstrated.
Our intention is to use the refined factor scores as outcome variables in future work where we explore the
impact of engineering curricula on student interest in, aspirations for, and attitudes toward engineering. Others
may similarly use the instrument with children ages 8-11 to gather data on changes in children’s interests and
attitudes following engineering interventions in formal and informal settings. We recommend, when evaluating
the attitudes and interest of students who are new to engineering, that the retrospective post version of the EIA
be administered after completion of an engineering intervention. However, in cases where students already have
enough engineering experience that they can reasonably interpret the questions, the instrument could be adapted
and used as a pre-post survey. In either case, “NOW” subscales should be used as outcome variables, with the
corresponding “PRE” subscales as covariates.
Notes
This material is based upon work supported by the National Science Foundation under Grant No. [1220305].
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors
and do not necessarily reflect the views of the National Science Foundation.
References
Ainley, M., & Ainley, J. (2011). Student engagement with science in early adolescence: The contribution of
enjoyment to students’ continuing interest in learning about science. Contemporary Educational
Psychology, 36(1), 4–12. https://doi.org/10.1016/j.cedpsych.2010.08.001
Appelbaum, P., & Clark, S. (2001). Science! Fun? A critical analysis of design/content/evaluation. Journal of
Curriculum Studies, 33(5), 583–600.
Cunningham, C. M., & Lachapelle, C. P. (2010, June). The impact of Engineering is Elementary (EiE) on
students’ attitudes toward engineering and science. Paper presented at the American Society for
Engineering Education Annual Conference & Exposition, Louisville, KY. Retrieved from
https://peer.asee.org/15989
Bandalos, D. L., & Finney, S. J. (2010). Factor analysis: exploratory and confirmatory. In G. R. Hancock & R.
O. Mueller (Eds.), The reviewer’s guide to quantitative methods in the social sciences. New York, NY:
Routledge.
Bhanji, F., Gottesman, R., de Grave, W., Steinert, Y., & Winer, L. R. (2012). The retrospective pre–post: A
practical method to evaluate learning from an educational program. Academic Emergency Medicine,
19(2), 189–194.
Blalock, C. L., Lichtenstein, M. J., Owen, S., Pruski, L., Marshall, C., & Toepperwein, M. (2008). In pursuit of
validity: A comprehensive review of science attitude instruments 1935–2005. International Journal of
Science Education, 30(7), 961–977.
Brown, T. A. (2014). Confirmatory factor analysis for applied research. Guilford Publications.
Buccheri, G., Gurber, N. A., & Bruhwiler, C. (2011). The impact of gender on interest in science topics and the
choice of scientific and technical vocations. International Journal of Science Education, 33(1), 159–
178. https://doi.org/10.1080/09500693.2010.518643
Burke, R. J. (2007). Women and minorities in STEM: A primer. In R. J. Burke & M. C. Mattis (Eds.), Women
and minorities in science, technology, engineering and mathematics: Upping the numbers (pp. 3–27).
Northampton, MA: Edward Elgar Publishing, Inc. https://doi.org/10.4337/9781847206879.00008
Dillon, W. R. & Goldstein, M. (1984). Multivariate analysis: methods and applications. New York, NY: Wiley.
Drechsel, B., Carstensen, C., & Prenzel, M. (2011). The role of content and context in PISA interest scales: A
study of the embedded interest items in the PISA 2006 science assessment. International Journal of
Science Education, 33(1), 73–95. https://doi.org/10.1080/09500693.2010.518646
Duit, R., & Treagust, D. F. (1998). Learning in science: From behaviourism towards social constructivism and
beyond. International Handbook of Science Education, 1(Part 1), 3–25.
Germann, P. J. (1988). Development of the attitude toward science in school assessment and its use to
investigate the relationship between science achievement and attitude toward science in school. Journal
of Research in Science Teaching, 25(8), 689–703.
Gibbons, S. J., Hirsch, L. S., Kimmel, H., Rockland, R., & Bloom, J. (2004). Middle school students’ attitudes
to and knowledge about engineering. Presented at the International Conference on Engineering
Education, Gainesville, FL.
Gogolin, L., & Swartz, F. (1992). A quantitative and qualitative inquiry into the attitudes toward science of
nonscience college students. Journal of Research in Science Teaching, 29(5), 487–504.
Grice, J. W. (2001). Computing and evaluating factor scores. Psychological Methods, 6(4), 430.
Häussler, P., & Hoffmann, L. (2002). An intervention study to enhance girls’ interest, self‐concept, and
achievement in physics classes. Journal of Research in Science Teaching, 39(9), 870–888.
Hirsch, L. S., Gibbons, S. J., Kimmel, H., Rockland, R., & Bloom, J. (2003). High school students’ attitudes to
and knowledge about engineering (p. F2A7-12). Presented at the 33rd ASEE/IEEE Frontiers in
Education Conference, Boulder, CO: IEEE.
IBM Corporation. (2013). IBM SPSS Statistics for Windows (Version 22.0). Armonk, NY: IBM Corporation.
Kelloway, E. K. (2015). Using Mplus for structural equation modeling: A researcher’s guide (2nd ed.). SAGE
Publications, Inc.
Lichtenstein, M. J., Owen, S. V., Blalock, C. L., Liu, Y., Ramirez, K. A., Pruski, L. A., … Toepperwein, M. A.
(2008). Psychometric reevaluation of the scientific attitude inventory-revised (SAI-II). Journal of
Research in Science Teaching, 45(5), 600–616. https://doi.org/10.1002/tea.20244
Lindahl, B. (2007). A longitudinal study of students’ attitudes towards science and choice of career. Presented at
the NARST Annual Conference, April 15-18, 2007, New Orleans.
Lyons, T. (2006). Different countries, same science classes: Students’ experiences of school science in their own
words. International Journal of Science Education, 28(6), 591–613.
MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological
Methods, 4(1), 84–99. https://doi.org/10.1037/1082-989X.4.1.84
Maltese, A. V., & Tai, R. H. (2010). Eyeballs in the fridge: Sources of early interest in science. International
Journal of Science Education, 32(5), 669–685. https://doi.org/10.1080/09500690902792385
Miller, P. H., Blessing, J. S., & Schwartz, S. (2006). Gender differences in high-school students’ views about
science. International Journal of Science Education, 28(4), 363–381.
Moore, R. W., & Foy, R. L. H. (1997). The scientific attitude inventory: A revision (SAI II). Journal of
Research in Science Teaching, 34(4), 327–336.
Moore, R. W., & Sutman, F. X. (1970). The development, field test and validation of an inventory of scientific
attitudes. Journal of Research in Science Teaching, 7(2), 85–94.
Murphy, C., & Beggs, J. (2003). Children’s perceptions of school science. School Science Review, 84(308),
109–116.
Muthén, L. K., & Muthén, B. O. (2015). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén.
National Academies of Sciences, Engineering, and Medicine. (2017). Building America’s skilled technical
workforce. The National Academies Press. https://doi.org/10.17226/23472
National Academy of Sciences, National Academy of Engineering, & Institute of Medicine. (2010). Expanding
underrepresented minority participation: America’s science and technology talent at the crossroads.
https://doi.org/10.17226/12984
National Research Council, & National Academy of Engineering. (2014). Career choices of female engineers: A
summary of a workshop. https://doi.org/10.17226/18810
OECD. (2008). Encouraging students’ interest in science and technology studies. OECD. Retrieved from
http://dx.doi.org/10.1787/9789264040892-en
Owen, S. V., Toepperwein, M. A., Marshall, C. E., Lichtenstein, M. J., Blalock, C. L., Liu, Y., … Grimes, K.
(2008). Finding pearls: Psychometric reevaluation of the Simpson–Troost Attitude Questionnaire
(STAQ). Science Education, 92(6), 1076–1095.
Royal Society, The. (2006). Taking a leading role. The Royal Society.
Satorra, A., & Bentler, P. M. (2010). Ensuring positiveness of the scaled difference chi-square test statistic.
Psychometrika, 75(2), 243–248. https://doi.org/10.1007/s11336-009-9135-y
Sawyer, R. K. (2006). Introduction: The new science of learning. In R. K. Sawyer (Ed.), The Cambridge
handbook of the learning sciences (pp. 1–16). New York, NY: Cambridge University Press.
Sibthorp, J., Paisley, K., Gookin, J., & Ward, P. (2007). Addressing response-shift bias: Retrospective pretests
in recreation research and evaluation. Journal of Leisure Research, 39(2), 295.
Simpson, R. D., & Troost, K. M. (1982). Influences on commitment to and learning of science among
adolescent students. Science Education, 66(5), 763–781.
Tai, R. H., Liu, C. Q., Maltese, A. V., & Fan, X. (2006). Planning early for careers in science. Science,
312(5777), 1143–1144. https://doi.org/10.1126/science.1128690
Tytler, R. (2014). Attitudes, identity, and aspirations toward science. In Handbook of research on science
education (pp. 82–103).
Vaughan, G. M., & Hogg, M. A. (2013). Social Psychology. Pearson Higher Education AU.
Venville, G. J., Wallace, J., Rennie, L. J., & Malone, J. A. (2002). Curriculum integration: Eroding the high
ground of science as a school subject? Studies in Science Education, 37, 43–83.
Weinburgh, M. H., & Steele, D. (2000). The Modified Attitudes toward Science Inventory: Developing an
instrument to be used with fifth grade urban students. Journal of Women and Minorities in Science and
Engineering, 6(1), 87–94.
Author Information
Cathy P. Lachapelle Robert T. Brennan
Museum of Science, Boston Harvard T. H. Chan School of Public Health
Engineering is Elementary Department of Global Health and Population
1 Science Park, Boston, MA 02114 Harvard Humanitarian Initiative
U.S.A. 14 Story Street
Contact e-mail: clachapelle@mos.org Cambridge, MA 02138
U.S.A.
Appendix
Engineering Interest and Attitudes Assessment
Somewhat
Somewhat
We are interested in learning about your opinions of engineering. Please answer
Not Sure
Disagree
Strongly
Strongly
Agree
Agree
Agree
each question honestly. Mark how strongly you agree or disagree after each
statement. Thank you very much!
1. It is important for me to understand Last summer, I would have said: 0 1 2 3 4

engineering. Now I would say: 0 1 2 3 4
2. Engineering helps me to understand Last summer, I would have said: 0 1 2 3 4
today’s world. Now I would say: 0 1 2 3 4
Last summer, I would have said: 0 1 2 3 4
3. I enjoy studying engineering.
Now I would say: 0 1 2 3 4
4. Boys are better at engineering than girls.
5. I would like to work with other Last summer, I would have said: 0 1 2 3 4
engineers to solve engineering problems. Now I would say: 0 1 2 3 4
6. We learn about interesting things when Last summer, I would have said: 0 1 2 3 4
we do engineering in school. Now I would say: 0 1 2 3 4
7. I really want to learn engineering.
8. Girls are better at engineering than boys.
9. Engineering is useful in helping to solve Last summer, I would have said: 0 1 2 3 4
the problems of everyday life. Now I would say: 0 1 2 3 4
10. We learn about important things when Last summer, I would have said: 0 1 2 3 4
we do engineering in school. Now I would say: 0 1 2 3 4
11. Engineering is easy for me.
12. Engineering is fun.
13. When we do engineering, we use a lot of Last summer, I would have said: 0 1 2 3 4
interesting materials and tools. Now I would say: 0 1 2 3 4
14. It is important to understand engineering Last summer, I would have said: 0 1 2 3 4
in order to get a good job. Now I would say: 0 1 2 3 4
15. Girls have a harder time understanding Last summer, I would have said: 0 1 2 3 4
engineering than boys. Now I would say: 0 1 2 3 4
16. I would like to learn more about Last summer, I would have said: 0 1 2 3 4
17. I am interested when we do engineering Last summer, I would have said: 0 1 2 3 4
in school. Now I would say: 0 1 2 3 4
18. Engineers help make people’s lives Last summer, I would have said: 0 1 2 3 4
better. Now I would say: 0 1 2 3 4
19. Girls and boys are equally good at Last summer, I would have said: 0 1 2 3 4
20. I try hard to do well in engineering.
21. I know what engineers do for their jobs.
22. I would enjoy being an engineer when I Last summer, I would have said: 0 1 2 3 4
grow up. Now I would say: 0 1 2 3 4
23. Boys have a harder time understanding Last summer, I would have said: 0 1 2 3 4
engineering than girls. Now I would say: 0 1 2 3 4
24. Engineering is really important for my Last summer, I would have said: 0 1 2 3 4
country. Now I would say: 0 1 2 3 4
Exploring Academic Performance Paths and Student Learning Strategies

in a Large Foundational Engineering Course
Jacob R. Grohs, David B. Knight, Glenda D. Young, Michelle M. Soledad

Article History Situated in the second year of an engineering curriculum, undergraduate
engineering mechanics courses represent a significant barrier to persistence in
Received: engineering. This study seeks to inform and improve these educational
01 July 2017 environments by examining academic performance paths over time in a course
and explore how students in each path compare in the learning strategies they
Accepted:
employ to engage with course content. Through online surveys, we gathered data
08 January 2018
on self-reported time spent engaging with course content before high-stakes
testing in four large sections of a Statics course that were all taught by the same
Keywords
instructor. Cluster analysis identified groups exhibiting distinct performance
Middle years paths, and one-way Welch’s F-tests with post-hoc comparisons explored
Statics differences between these clusters based on time spent engaging with course
Metalearning content through specific learning strategies. Differences across performance
Large classes clusters were found primarily in the ways in which students spent time rather
Student strategies than total time spent. Solving problems independently was a strategy employed
significantly more often by the highest-performing cluster of students. In
contrast, a group of unsuccessful students in the course spent comparably less
time solving problems independently but comparably more time solving
problems with peers. From these results, we suggest how leveraging these
findings might impact educational practice and guide future research.
Introduction
Engineering mechanics courses frequently represent a significant challenge to students along the pathway to an
undergraduate engineering degree and thus can be major barriers for student persistence and success in
engineering (Lord & Chen, 2014). After completing general engineering coursework and university-level
studies (e.g., mathematics, chemistry, physics), students are tasked with mastering more specific and technical
engineering content, typically in the second year of study. Coupled with the challenging content, these classes
tend to be characterized by large enrollments and are structured primarily in a lecture-based format because they
are required across several engineering disciplines. Such educational environments are not always conducive for
students to digest conceptually difficult course content and receive timely individualized feedback. For example,
Halpern and Hakel (2003) claim that lecture-style approaches can be “one of the worst arrangements for in-
depth understanding” since “understanding is an interpretive process in which students must be active
participants” (p. 40). Although some programs have been successful with adjusting pedagogical approaches
(e.g., introducing flipped classroom models) (Bishop & Verleger, 2013), the large lecture model still
characterizes the structure of many programs and likely will continue to do so because of resource availability,
organizational inertia, and swelling enrollments.
Within this early stage of their curriculum, students often find themselves at a crossroads of mastering technical
content while simultaneously navigating how to be an effective learner in college (i.e., the development of self-
regulation and metacognition skills). Educational psychologist Paul Pintrich (2000) describes this development
as “an active, constructive process whereby learners set goals for their learning and then attempt to monitor,
regulate, and control their cognition, motivation, and behavior, guided and constrained by their goals and the
contextual features in the environment” (p. 453). Students begin to examine and adjust their thoughts, beliefs,
and behaviors by considering their prior technical knowledge and coursework (e.g., mathematics or physics),
planning and executing various studying approaches, and managing their time across their entire course load.
Although student success in any course is a function of these skills, because engineering mechanics courses
242 Grohs, Knight, Young, & Soledad
represent such an early pivotal point in an engineering curriculum, students’ ability to self-regulate learning—
and how instructors can help students develop this skill—represents a critical area of further study.
The large class scenario, however, places many constraints on how teachers can provide targeted feedback and
monitor individual students’ progress; this educational environment generally forces teachers to provide
minimal individual feedback and instead offer general feedback to the class in the aggregate. When providing
feedback on class performance, many instructors will conduct analyses on individual assessments by
interrogating class averages and show those results to their students following exams. This approach considers
assessments as discrete events (i.e., performance on Test 1, performance on Test 2, etc.). Additionally, many
instructors will pay attention to students’ cumulative performance, generally placing particular emphasis on this
measure at the end of a full course (i.e., students’ final grades in a class). However, both of these approaches
lose information about students’ “academic performance paths” through the class (i.e., a holistic view of grades
over time). If it is important to help students develop self-regulation skills, such as identifying effective
strategies and planning to improve in a course after a poor grade, then understanding how those paths change
from one assessment to the next represents an important consideration. We explore this idea in our study of
students enrolled in large engineering mechanics courses, specifically Statics, and address the following
research question:
RQ 1: What academic performance paths exist in an undergraduate mechanics course?

After establishing predominant academic performance paths, we investigate how students with different paths
differ on a suite of measures related to students’ self-reported learning strategies for engaging with course
content as well as their final exam performance. The objective of this analysis was to determine whether
recommending that teachers pay attention to students’ academic performance paths—as opposed to discrete
grades or only final averages—might provide insights on how students approach learning within large classes.
We addressed the following:
RQ2: How do overall performance and student learning strategies differ across academic performance
paths in an undergraduate mechanics course?
Review of the Literature

The middle years of undergraduate engineering study are formative years that introduce students to foundational
engineering and discipline-specific coursework. The limited prior work that has investigated the middle years
has focused on redesigning curriculum and instruction formats to address student persistence and instructional
challenges (Lord & Chen, 2014). Still, researchers acknowledge the need to further investigate academic
experiences and outcomes in the middle years, as most research in engineering education has focused on
students’ first and final years as undergraduates. We contribute to the literature focused on engineering students’
middle years by exploring the relationships between student learning strategies and students’ academic
performance paths over the duration of a semester in a foundational engineering course.
Student Learning Strategies
A vast body of literature has demonstrated the key role that learning strategies play in student learning. From a
cognitive perspective on learning, all students must become independent learners, be able to engage in
continuous learning, and take control of their own learning (e.g., Bransford et al., 2000; Greeno, Collins, &
Resnick, 1996; Newstetter & Svinicki, 2014). Conceptualizations of constructs such as “self-regulation” or
“metacognition” within the engineering education literature have varied (Lawanto, 2010), including a range of
phenomena such as awareness of knowledge, thinking, learning, and organizing cognitive resources (e.g.,
Cuasay, 1992; Flavell, 1979; Marzano et al., 1988). Although our study investigates domain-specific student
learning strategies, the theoretical backing for exploring learning strategies is strongly informed by literature on
metacognition, self-regulation, and metalearning—thus these are discussed in this review.
As summarized by Meyer et al. (2015a), previous research on metacognition within engineering education has
shown the following benefits for students: (1) helping them recognize how to associate elements of knowledge,
(2) promoting long-term comprehension of concepts, (3) developing self-confidence, (4) increasing awareness
of their own knowledge gaps, and (5) enabling teachers to provide students with actionable metacognitive
feedback. There has been widespread focus on how metacognitive abilities play an important role in problem-
solving and knowledge transfer, as students aim to apply their expertise within different contexts (Lawanto,
2010; McKenna, Johri, & Olds, 2014; Prince & Felder, 2006; Woods, 2000). Litzinger et al. (2010), for
example, identified this pattern when evaluating weak and strong students’ approaches to solving free body
diagrams in a statics course. The academically stronger students utilized metacognitive skills nearly twice as
much as did the academically weaker students in a Statics course. Other authors across a range of engineering
disciplines similarly have shown a relationship between metacognitive abilities and academic performance in
chemical engineering (Ko & Hayes, 1994), statics (Hanson & Williams, 2008), and civil engineering (Meyer et
al., 2015a). Previous researchers have linked metacognitive skills and abilities to higher grade point average,
expert and novice practices (Larkin, McDermott, Simon, & Simon, 1980; Voss, Greene, Post, & Penner, 1983),
enhanced problem-solving (Cooper, Sandi–Urena, & Stevens, 2008) and higher average exam scores (Zulkiply,
Kabit, & Ghani, 2009). Thus, literature supports the connection between students’ metacognitive activities and
their learning and performance in courses.
Captured under this broad “metacognition” umbrella, metalearning refers to the more specific notion that
individuals have awareness of their own approaches to learning and their abilities to control it. As articulated by
Biggs (1985), metalearning consists of two stages: awareness of learning processes within a specific learning
context (i.e., how am I learning?), and self-regulatory (internal) control over those processes (i.e., how can I
develop a successful learning strategy?). As argued by Meyer et al. (2015b) within the undergraduate
engineering context, for students to develop metalearning capacity they need to become aware of their learning
behaviors via some mechanism, they need help interpreting those existing behaviors, and they need to be
empowered to change their behaviors. All of these efforts should be focused within a specific learning context
because approaches to learning vary contextually.
Researchers within engineering education have taken a few different approaches to developing students’
metalearning capacities. Some researchers have advocated for a student success program specifically focused on
metalearning (e.g., Turner, 2001), but that kind of approach does not sit within the context of students’
engineering coursework. Other approaches have included administering self-report instruments such as the
Study Process Questionnaire to explore differences in students’ surface and deep approaches learning across
courses and class years (Jenkins, Edwards, Nepal, & Bolton, 2011; Turner, 2004). In a more recent
advancement, Meyer et al. (2015b) administered the Reflections on Learning Inventory within a Civil
Engineering course in Australia and focused students’ learning contexts on a specific “threshold concept” in the
course. Using this tool helped make learning within a specific context visible for students via an individually
tailored learning profile that yielded benefits in learning engagement and conceptual understanding, as indicated
by student participants. Literature on the awareness and self-regulatory control of the thinking and learning
process sets the stage for our focus on student learning strategies, as self-regulation is a process that involves
goal-setting, planning, motivation, monitoring, appropriate help-seeking, evaluation, and reflection (Ormrod,
2012). These characteristics of self-regulated learning are reflected in strategies adopted by engineering students
in their learning, including independent problem-solving, working with peers, and attending office hours
(McCord & Matusovich, 2013).
While we are not measuring metalearning capacity specifically, our study similarly investigates a mechanism for
helping students visualize how academic performance paths correspond to student learning strategies. We
investigate how leveraging a commonly used measure (i.e., grades) in a new way might serve as an indicator of
students’ learning processes to open up new dialogue between instructors and about metalearning in a large,
middle years engineering course. In this work, we evaluate students’ performances on tests across the semester,
investigate their performance paths within the course, and compare those paths for differences in time spent on
different learning strategies as well as final course grade. We capture data using surveys administered around
high-stakes exams throughout the semester to help students think about their learning processes. Although
researchers have used similar in-class survey approaches before and after lectures (Mazumder, 2012; Mazumder
& Ainsworth, 2010), we instead investigate how learning strategies relate to longitudinal course performance to
understand how students might adjust within the time window of a single semester. Understanding this
relationship provides a mechanism for instructors to understand and explain to students how behavior can be
adjusted throughout a course to support successful academic performance in undergraduate engineering courses.
Data and Methods

Data Collection Context
This study took place during the 2014–2015 academic year at a comprehensive, research-intensive university in
the United States that is predominantly known for its College of Engineering. The institution enrolled about
7,000 undergraduate engineers during that year across all class years. Data for this study were drawn from four
sections of the Statics course (two in Fall 2014 and two in Spring 2015) that were all taught by the same
instructor for consistency of pedagogical style. The Statics sections were taught in a lecture style that followed a
common course schedule across all sections and instructors within the department, but individual instructors had
autonomy in determining their approaches to teaching. Linear Algebra and Vector Geometry are prerequisites
for the course, and Mutivariable Calculus is a co-requisite. Following institutional review board (IRB)
procedures for involving students in research projects, enrolled students were given the option to participate in
this study and allow their exam data to be incorporated in analyses. Table 1 gives a demographic breakdown of
the consenting study sample, which suggests that it is a reasonable demographic representation of the broader
College of Engineering population.
Table 1. Demographic breakdown of consenting student sample compared to recent college enrollment
Consenting College of
Sample Engineering
(n=191) Enrollment
Sex
Female 28% 22%
Male 72% 78%
Racial/Ethnic Background
African American 2% 3%
Asian/Pacific Islander 12% 12%
Hispanic/Latino 3% 5%
White (Non-Hispanic) 78% 62%
Multi-Racial 3% 4%
Prefer Not to Answer/None of the
Above 3% 2%
Nonresident Aliena --- 11%
Note: Numbers may not total 100% because of rounding to nearest integer.
a
The nonresident alien category is reported in the institutional data but was not used as a category in our
research.
The class’s assessment scheme included 24 homework problem sets (accounted for 15% of the grade in total),
four high-stakes tests (each accounted for 15% of the total grade with each test set by the instructor), as well as a
final, common departmental exam (weighted at 25% of the total grade). The four high-stakes tests, which were
written by the instructor, consisted of four to seven open-ended problems similar to assigned homework
exercises, and were graded by awarding partial credit when incorrect answers were characterized by common
identifiable mistakes within otherwise conceptually sound solutions. In contrast, the departmentally written and
standardized final exam consisted of multiple choice-style problems (i.e., about 20 individual questions drawn
from 10-14 problems) with no partial credit awarded for incorrect answers. For example, if a problem required
solving for forces in a truss system, the truss system would comprise one problem but may include several
individual questions to report values for different variables.
Data Sources and Collection Procedures
Outside of the course performance, this study consisted of five online surveys collected over the course of the
semester via Qualtrics. The first survey was administered at the beginning of the semester, included basic
demographic questions, and solicited overarching consent for researchers to link together participant course
performance with any other research data collected (e.g., future surveys). The remaining four surveys were
identical to one another and administered the class period prior to each of the high-stakes tests in the course.
These surveys, hereafter named according to the high-stakes test which they preceded (e.g., Pre-Test 1 Survey),
explored how often (i.e., hours per week) and through what methods (e.g., classroom attendance, office hours,
independent problem solving, group problem solving) students self-reported engagement with course content
throughout the semester. These items were not from existing instrumentation in the literature but instead
leveraged the instructor’s expertise to understand common ways in which students engage with course content.
An open-ended question prompted respondents to consider indicating methods of engagement not specifically
named in the survey but did not result in any submissions that suggested the surveys warranted substantive
revision. The means of engagement included in the surveys also aligned with existing research that examined
intentional ways that students engage in the learning process in the context of conceptual change (McCord &
Matusovich, 2013). Students were given time to complete the surveys during class meetings, and to spur higher
response via an incentive students who participated were entered into a raffle for a gift card drawing at the end
of the semester.
Responses from the surveys were paired with course achievement data (homework average, high-stakes tests 1-
4, final exam, overall course grade). The total number of student responses decreased throughout the course,
which we attribute to students withdrawing from the course, survey fatigue, and dissatisfaction with overall
course performance. Out of a population of 477 students enrolled in the sections at the start of each semester,
responses on the various surveys were: 202 students (42.3%) on the first survey, 182 (38.2%) on the Pre-Test 1
Survey, and 85 students (17.8%) on the Pre-Test 2 survey. Response rates on surveys corresponding to Tests 3
and 4 were low enough to warrant dropping those from our analyses due to concerns that the sample would no
longer be representative across all academic performance paths. Thus, we focus on linking course-long
performance data with the student learning strategies reported on the Pre-Test 1 survey and the Pre-Test 2
survey.
Analytical Procedures
Our first research question focused on the identification of academic performance paths across multiple tests; to
identify such paths and group together students who exhibited similar performance paths, we utilized cluster
analyses. We first performed agglomerative hierarchical cluster analysis on scores from the four high-stakes
tests using the built-in hclust function in R and identified the number of clusters desired using guidance from the
hclust dendrogram. Having identified the number of clusters, we then used the partitioning around medoids
(pam) function from the cluster package in R to identify cluster membership for each individual student.
Partitioning around medoids is a more robust form of commonly used k-means clustering methods and is
advantageous in part because it does not rely on randomly seeded centroids and thus offers solution stability
across analytical runs (Kaufman & Rousseeuw, 2009).
After developing clusters that represent different academic performance paths through the course, we addressed
the second research question by examining differences across clusters on other measures related to achievement,
engagement, and motivation. With the intent of running analysis of variance (ANOVA) and appropriate post-
hoc testing, we examined the distribution of our various dependent variables across the groupings. In most
cases, Bartlett’s test indicated heterogeneity of variance; consequently, one-way Welch’s F-tests were used
instead of the usual parametric F-tests in traditional ANOVAs (Field, Miles, & Field, 2012). Accordingly, when
conducting pairwise t-tests as part of post-hoc testing to compare group means, we assumed unequal variance
and also adjusted p-values for family-wise error rate using the Holm correction to the traditional Bonferroni
method (Holm, 1979). Though one-way tests are considered robust against the normality assumption, because
some of the clusters have relatively small numbers of members (e.g., 10-15) for some dependent variables, the
appropriateness of parametric tests is called into question. To interrogate this issue, non-parametric Kruskal-
Wallis tests with Dunn post-hoc tests and Bonferroni adjustment for family-wise error rates were completed for
all the same comparisons as the one-way tests. In all cases, a threshold of p<0.05 was used as the criterion for
statistical significance. Across the board, the omnibus tests (i.e., Welch’s F and Kruskal-Wallis) were in
agreement on statistically significant findings for all variables. The post-hoc tests across both were in near-
perfect agreement with the rare inconsequential borderline cases (e.g., values slightly above 0.05 on one test and
slightly below on another).
Results
Academic Performance Paths, Not Discrete Academic Performance Points (RQ1)
Our cluster analysis sought to identify emergent groupings of undergraduate engineering mechanics students
based on their multiple test performances over a semester. We identified a five-cluster solution as best fitting
our data using guidance from a dendrogram from the first-stage hierarchical clustering, and proceeded to run a
five-cluster partitioning around medoids analysis. We labeled these five clusters as: high performer, average
performer, low performer, early adapter, and late adapter. Results of this analysis are shown in Figure 1 as a plot
of mean performance on each of the four high-stakes tests broken out by cluster.
Figure 1. Academic performance paths in Statics based on cluster analysis across four high-stakes tests
We named the clusters to characterize the performance paths over the course of the semester. The High
Performer (n=71), Average Performer (n=104), and Low Performer (n=14) clusters are the most intuitive
clusters as they represent consistent performance of A’s, C’s, or F’s across each of the four tests. Of particular
interest are the Early Adapter (n=103) and the Late Adapter (n=58) clusters because they exhibit significant
variation across the semester. The Early Adapter cluster has a failing Test 1 average (53.81%) followed by a
consistent improvement throughout the semester and ends with an overall average grade right at the C- break
point, which is the minimum grade for passing the course. On the other hand, the Late Adapter cluster has a
slightly better Test 1 average compared to the Early Adapter cluster (59.06%), but instead takes a second poor
performance by also failing Test 2 before starting an improvement on Tests 3 and 4. Visualizing students’
academic paths in such a way presents a very different picture to instructors and students compared to analyses
of single-point exam grades. The Early Adapter and Late Adapter paths would not have been identified when
performance is considered discretely as opposed to holistically.
Characterizing Overall Course Performance and Student Learning Strategies by Performance Path
(RQ2)
Performance
Course performance is a composite of each high-stakes test and the final exam. However, the cluster solution
presented in answering RQ1 was developed on high-stakes test performance for Tests 1, 2, 3 and 4 only. Before
investigating this broader performance across the course by performance path (cluster), we first justify
exclusions from the cluster analysis. Overall grades offer a summative performance point but were not included
in the cluster analysis because they are direct artifacts of the test scores themselves. Similarly, scores on a
comprehensive final exam should be related to scores on the individual tests and thus warrant examination by
performance path (cluster) but would not warrant including them as a singular input beside each individual test
in the data-driven clustering algorithm.
Table 2 includes the numeric values from Figure 1 along with the Final Exam scores and Overall Grades for the
class. Additionally, the table includes data for a sixth grouping of students, named “Drop-Out”—this group
includes any student who did not fully complete the course and thus was not included in the cluster analysis
because of incomplete data. Though not part of the cluster analysis, results about this group’s learning strategies
will be compiled with the other five clusters because the Drop-Out group remains a critical group of students to
understand better.
Table 2. Average performance on each test, the final exam, and final grade by cluster
Clusters N Test1 Test2 Test3 Test4 Final Exam Overall Grade
High Performer 71 81.80 85.54 92.34 92.01 80.92 87.01
Average Performer 104 80.20 77.22 72.16 83.61 60.91 74.78
Early Adapter 103 53.81 66.28 79.33 84.33 58.00 69.83
Late Adapter 58 59.06 49.60 60.33 69.11 39.14 56.64
Low Performer 14 34.68 31.77 9.61 7.39 11.00 17.61
Drop-Out 127 43.71 39.60 35.03 -- -- --
The next set of results relies on variables collected through the battery of surveys conducted throughout the
course. The number of student responses varied over time, as shown in Table 3. Overall, each cluster is well
represented across surveys, which helps mitigate concerns that arise when respondents are overwhelmingly from
the same group (e.g., had only the strongest students participated in the research, it would have been difficult to
draw conclusions). Across the administration of the Pre-Test 1 and Pre-Test 2 surveys, the Pre-Test 2 Survey
received the poorer overall response rate, but there were sufficient numbers of responses in most clusters to
allow statistical comparisons. One exception occurred in the Low Performer cluster, which did not have any
respondents to the Pre-Test2 survey and thus was not used. It is not surprising that students in that group did not
engage in the final survey for the course given their course performance.
Table 3. Response rates to surveys across performance path clusters

Clusters N Initial Response Pre-Test 1 Response Pre-Test 2 Response
High Performer 71 43.7% 39.4% 24.0%
Average Performer 104 33.7% 33.7% 14.4%
Early Adapter 103 54.4% 51.5% 28.2%
Late Adapter 58 53.5% 32.8% 19.0%
Low Performer 14 28.6% 28.6% 0.0%
Drop-Out 127 35.4% 33.9% 10.2%
Student Learning Strategies
We first explore how students self-reported spending their time engaging with course content prior to high-
stakes Test 1 (Table 4) and Test 2 (Table 5) independently; next, we consider their time linked across surveys
and consider how students changed their time from one part of the class to the next (Table 6). Statistically
significant (p<.05) differences across the clusters from the series of one-way F-tests are highlighted in gray.
First, looking across Tables 4 and 5 of total hours per week for engagement in specific learning strategies, it
seems that the vast majority of students in all groups attended class for the three nominal credit hours (officially
150 minutes of contact time) per week and otherwise invested significant time in the course as shown. For
example, students’ time engaging with the course ranged from 7.7-11 total hours per week prior to Test 1. These
empirical data run counter to the knee-jerk reaction by many instructors that attributes poor performance on a
high-stakes exam to lack of student engagement or preparation—the “you all just didn’t work hard enough”
narrative. Surprisingly, the significant differences across clusters are not in the total time spent alone for Test 1
or Test 2 as one might expect, but rather how that time is spent.
In preparation for Test 1 (Table 4), a key difference was apparent in time spent solving problems with peers,
F(4,73.7) = 8.40, p<0.001. Students in the High Performer cluster spent the least amount of time on average
(0.55 hours per week), and students in the Drop-Out cluster spent the most (2.06 hours per week).
Table 4. Average hours per week spent engaging with course content in the weeks leading to high-stakes test 1
Clusters
Response
Survey 1
Pre-Test
Attendance
Class
Reading Text
Online
Independently
Problems
Solving
w/ Peers
Problems
Solving
Office Hours
Total Time
High
28 (39.44%) 2.75 0.64 0.41 3.36 0.55 0.27 7.69
Performer
Average
35 (33.65%) 2.71 1.61 0.89 3.06 1.77 0.66 10.70
Performer
Early
52 (51.46%) 2.72 1.00 0.66 2.52 1.61 0.71 9.25
Adapter
Late
19 (32.76%) 2.71 1.53 0.63 2.50 1.84 0.26 9.47
Adapter
Low
4 (28.57%) 3.00 1.25 0.75 3.00 0.00 0.00 8.00
Performer
Drop-Out 43 (33.86%) 2.72 1.66 0.93 2.57 2.06 0.84 10.92
Statistically significant findings related to time activities prior to Test 2 (Table 5) support the notion that the
more time students spent engaged in solving problems independently (F(4,34)= 8.96, p<0.001) is related to
higher scores overall. Post-hoc analysis of the Welch’s F-test result revealed statistically significant differences
between the High Performer cluster (6.47 hours per week), the Late Adapter (2.64 hours per week), and Drop-
Out Clusters (1.54 hours per week) in solving problems independently.
Table 5. Average hours per week spent engaging with course content in the weeks leading to high-stakes test 2
Clusters
Response
Survey 2
Pre-Test
Attendance
Class
Reading Text
Online
Independently
Problems
Solving
w/ Peers
Problems
Solving
Office Hours
Total Time
High Performer 17 (23.94%) 2.88 0.50 0.29 6.47 0.68 0.69 11.31
Average
15 (14.42%) 2.77 2.23 1.03 4.27 1.47 0.67 12.43
Performer
Early Adapter 29 (28.16%) 2.66 1.30 1.09 4.19 2.83 1.36 13.55
Late Adapter 11 (18.97%) 2.64 2.05 0.68 2.64 3.23 0.77 12.00
Low Performer 0 (0.00%) -- -- -- -- -- -- --
Drop-Out 13 (10.24%) 2.81 1.96 0.58 1.54 2.77 0.31 10.04
The essence of this finding is repeated again in Table 6 when investigating how different students in different
clusters changed the amount of time spent on various activities. Once again, we found statistically significant
differences in the change in time spent solving problems independently, F(4,31) = 4.16, p<0.01; post-hoc tests
identified differences between the High Performer cluster (increase in time spent working independently by 2.81
more hours per week on Test 2 than Test 1) and Drop-Out Clusters (0.82 fewer hours per week). We also found
a significant difference in the Total Time change from Test 1 to Test 2, F(4,30) = 2.80, p<0.05, with the Early
Adapter cluster making the largest increased investment of total time on average (4.39 hours per week).
Table 6. Change in how time is spent engaging with course content reported in hours per week between high-
stakes tests 1 and 21
Clusters
Response
Survey
Combined
Attendance
Class
Reading Text
Online
Independently
Problems
Solving
w/ Peers
Problems
Solving
Office Hours
Total Time
-
High Performer 16 (22.54%) -0.06 0.03 2.81 0.16 0.43 3.12
0.11
Average Performer 13 (12.50%) 0.00 0.15 0.08 1.08 0.15 0.23 1.69
Early Adapter 28 (27.18%) -0.11 0.43 0.50 1.52 1.41 0.70 4.39
Late Adapter 11 (18.97%) -0.09 0.45 0.09 0.23 1.82 0.32 2.82
Low Performer 0 (0.00%) -- -- -- -- -- -- --
Drop-Out 13 (10.24%) 0.00 0.77 -0.45 -0.82 0.73 -0.55 -0.23

1
Note: negative values indicate a decrease in time; positive values indicate an increase.
Discussion
The findings from this study lead us to suggest several implications for improving the educational environment
of this Statics course that may also be considered for other large foundational engineering courses. Similarly,
our work leads us to consider important next steps in the research to further understand the importance of
metalearning, particularly in the pivotal middle years of the undergraduate engineering curriculum.
One key finding in answering RQ1 is the existence of academic performance paths that exhibit significant grade
variation on high-stakes tests throughout the semester. In particular, the grade pattern differences between the
Early Adapter and Late Adapter clusters demonstrate that two groups with failing grades on Test 1 take different
patterns—one resulting in steady improvement and a successful overall grade in the course, and one where any
improvement comes too late and results in a failing overall grade in the course. Statics is a challenging
foundational course often taught in a high-stakes large lecture environment, and studies have shown that it
serves as a gatekeeping course that can be a barrier to student success in engineering (e.g., Grohs, Kinoshita,
Novoselich, & Knight, 2015; Grohs, Soledad, Knight, & Case, 2016). From the perspective of faculty members
and administrators, there is an expectation that students may struggle in challenging foundational courses as
they adapt to the academic demands of the engineering curriculum. However, from the point of view of students
who are under-performing in one of their first engineering courses after excelling in high school, that first
failing test grade can be distressing and demoralizing. Identifying academic performance paths in a foundational
undergraduate engineering course, and specifically the existence of the Early Adapter compared to the Late
Adapter cluster, can provide a hopeful, yet sobering, reflective opportunity for students because it positions how
students respond to their Test 1 performance as a potential fork in the road. Instructors can show these kinds of
performance path data to students so that they can see how large numbers of students have recovered from a
poor first test and finished with a passing grade. Other students who do not begin to improve until after their
second test grade were unable to recover in time and did not pass the course.
If nothing else, such a result should demonstrate to students the importance of help-seeking and adjustment of
strategies—engaging in metalearning—earlier rather than later in a course. While we cannot know exactly what
caused the differences between the academic performance paths, we believe that instructors can play a key role
in prompting student reflection by encouraging students to consider their grade after Test 1, to set goals for what
feasible grades on future tests may be, and to think about the means of engaging with course content that can
enable student learning and success. These suggestions are supported by literature and thus our work offers
reinforcement; for example, findings about the MUSIC Model of Academic Motivation have shown giving
students actionable ways they might drive their own learning, and demonstrating faculty investment in student
success both increase student academic motivation (Jones, 2009). The knowledge of academic performance
paths where students show dramatic improvement over the course of the semester may inspire improvement by
demonstrating that it is not only possible, but that many students routinely make such improvement semester
after semester.
Other important findings from this work stem from questions prompted by findings in RQ1—what behavior is
different across these clusters? Do students across clusters engage in different ways or for different amounts of
time? To address these questions, we used the survey information about specific learning strategies students
used and the amount of time they invested in the course across these strategies. Although we cannot make any
causal claims that may be desired by a student facing the fork in the road between Early Adapter and Late
Adapter, an investigation of variation across clusters lends insight into what students in stronger-performing
clusters may do differently than their less successful peers. In a participatory design study exploring how data
may assist students be more successful, first-year engineering students specifically brainstormed such
comparative analyses as something they desired. In particular, they wanted to know how high-performing
students were different in terms of variables like time spent engaging in different class-related activities (Knight,
Brozina, & Novoselich, 2016).
A key finding was that students across all academic performance paths reported investing significant time in the
course, including classroom attendance. This result is somewhat surprising in a large, challenging gatekeeper
course where it is expected that significant numbers of students will not be successful. Narratives rationalizing
dropout in gatekeeper courses often depict a body of students as too unwilling to work hard enough to be
successful. Olson and Riordan (2012) note that “many STEM faculty members believe that this ‘weeding out’
process is in the best interest of their disciplines and the larger national interest.” Shifting these narratives is
important, especially given that they may contribute to reasons why gatekeeping courses have been shown to
disproportionally hinder women and underrepresented groups (e.g., Gainen, 1995; Seymour, 2008). Our
research contribution, that all groups are investing significant time, may come as a surprise to the “weed-out”
notion that students just aren’t working hard enough to be successful. Instead, as noted in our results, the critical
difference appears to be in how students invest their time.
Time spent solving problems with peers, in the context of this course and this assessment structure, was a
learning strategy employed differently by the different clusters. In advance of both Test 1 and Test 2, the High
Performer cluster of students spent the least amount of time solving problems with peers, while students in less
academically successful clusters such as the Drop-Out or Late Adapter clusters spend significantly more. At first
glance, this finding could seem to go against established positive benefits of social learning and collaboration
with peers (e.g., Doolittle, 1997; Gallagher, Weldon, Haller, & Felder, 1999). However, the testing environment
required students to solve problems in an independent manner on a challenging, timed exam. Although
collaboration may often benefit learning, the testing environment of this course specifically, and of foundational
engineering courses at large institutions generally, is in sharp contrast to solving problems with peers. Further,
individuals struggling to learn content may be lured into a false sense of security about what they know when
working with groups since recognizing rationales behind solution steps is different than generating them on
one’s own. Aligned with best practices for cooperative learning, instructors should encourage students to engage
in metacognitive processes to pay close attention to what they do not fully understand without the assistance of
their peers.
Time spent solving problems independently was another student learning strategy that differed significantly
across academic performance pathway clusters. The High Performer cluster spent the most amount of time
solving problems independently in preparation for Test 2 and also invested a significantly larger increase in time
on this activity from Test 1 to Test 2 when contrasted with learning strategy investment shifts other clusters
made from Test 1 to Test 2. For example, the Drop-Out cluster actually decreased the amount of time invested
in solving problems independently but increased other strategies such as solving problems with peers. Though
time investment in a course need not be a zero-sum game, we characterize this behavior as a doubling down on a
seemingly less effective strategy by the Drop-Out cluster that is noteworthy and warrants further study.
These findings about solving problems independently highlight it as a student learning strategy that appears to
be particularly important for student success in this Statics course. We have already argued that, at face value,
solving problems independently in practice is in strong alignment with the mode of performance—sitting alone
in one’s room with a calculator and a problem on a blank page is a great way to prepare for a test consisting of 5
or 6 problems on a blank page without aids. Additionally, the highlighted importance of solving problems
independently as a mode of engagement is supported by broader literature on the role of practice in learning.
Specifically, research has shown that deliberate goal-directed practice is critical to learning and expertise as
compared to more general types of practice (Ambrose, Bridges, DiPietro, Lovett, & Norman, 2010; Ericsson,
Krampe, & Tesch-Römer, 1993). Arguably, independently solving problems is the most goal-directed means of
engagement that students reported with obvious end results of problems being solved, or not, solely attributable
to the learner’s efforts. Thus, it is more about working smarter rather than just working harder.
Limitations and Future Work
These findings should be interpreted with a few cautions. First, the sample sizes decrease when the overall
number of survey responses is broken down into respondents from each of the different clusters. In part because
steps were taken to choose conservative post-hoc procedures to account for family-wise error issues in multiple
pairwise comparisons, larger sample sizes (i.e., more statistical power) would be needed to interrogate more
helpful differences between the Early Adapter and Late Adapter clusters and thus shed insight on the “fork-in-in
road moment” for those groups of students. As a further limitation, self-reported data about course engagement
could be artificially inflated because of social desirability bias, although IRB consent documents indicated that
research findings would only be reported to the instructor in aggregate during the course and thus any bias
would not be a result of a respondent concern for individual judgement about responses. Additionally, the
magnitude of hours students reported investing in course-related efforts across the different kinds of activities
differed as anticipated, so we have more confidence in these values. For example, the low averages for office
hour attendance resonated with what the instructor experienced.
Both students and instructors may be interested to know how much time on average students spend engaging
with course content. In our study, the instructor was surprised that students from all clusters were reporting that
they invest significant time in the course. Further, how little time some students were spending solving problems
independently was also a surprise and prompted explicit discussion in class about ways for students to construct
and attempt their own practice exam of problems from the textbook as formative assessments to gauge readiness
for the next high-stakes test. Although not consistently implemented in this administration, basic analysis
procedures for the course engagement questions could be automated by the research team so that an instructor
could share the hours per week across engagement strategies data with the overall grade distribution following
each test. Students are accustomed to comparing their own grades to the grade distribution of the whole class, so
pairing grade data with engagement data could assist students in proactively shifting strategies and improving
performance throughout the semester.
A primary goal of this research has been to make student learning strategies and class performance paths more
visible to faculty and students alike. To address limitations associated with large foundational engineering
courses noted in the introduction, one highlight of this work is the ability to capture students’ engagement in
real time so instructors might have actionable recommendations to adjust their own teaching, and to present the
data to students as an opportunity for opening a discussion around metalearning. Though this study was limited
to one instructor of several large Statics sections, future work should focus on similar data collection across
varied instructors and related foundational engineering courses to gain a better understanding of how our
findings might relate to similar courses and contexts.
Acknowledgements
The authors would like to gratefully acknowledge the support of a 4-VA Collaborative Research Grant, which
supported the research in this article.
References
Ambrose, S. A., Bridges, M. W., DiPietro, M., Lovett, M. C., & Norman, M. K. (2010). How Learning Works:
Seven Research-Based Principles for Smart Teaching. John Wiley & Sons.
Biggs, J. B. (1985). The role of metalearning in study processes. British Journal of Educational Psychology,
55(3), 185–212.
Bishop, J. L., & Verleger, M. A. (2013). The Flipped Classroom: A Survey of the Research. Presented at the
ASEE Annual Conferance & Exposition, Atlanta, GA.
Bransford, J. D., Brown, A. L., Cocking, R. R., & others. (2000). How people learn. Washington, DC: National
Academy Press.
Cooper, M. M., Sandi–Urena, S., & Stevens, R. (2008). Reliable multi method assessment of metacognition use
in chemistry problem solving. Chemistry Education Research and Practice, 9(1), 18–24.
Cuasay, P. (1992). Cognitive Factors in Academic Achievement. Higher Education Extension Service Review,
3(3), n3.
Doolittle, P. E. (1997). Vygotsky’s zone of proximal development as a theoretical foundation for cooperative
learning. Journal on Excellence in College Teaching, 8(1), 83–103.
Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of
expert performance. Psychological Review, 100(3), 363–406. https://doi.org/10.1037/0033-
295X.100.3.363
Field, A., Miles, J., & Field, Z. (2012). Discovering Statistics Using R. SAGE.
Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry.
American Psychologist, 34(10), 906.
Gainen, J. (1995). Barriers to success in quantitative gatekeeper courses. New Directions for Teaching and
Learning, 1995(61), 5–14. https://doi.org/10.1002/tl.37219956104
Gallagher, V., Weldon, T., Haller, C. R., & Felder, R. M. (1999). Dynamics Of Peer Interactions In Cooperative
Learning (p. 4.210.1-4.210.15). Presented at the 1999 Annual Conference. Retrieved from
https://peer.asee.org/dynamics-of-peer-interactions-in-cooperative-learning
Greeno, J. G., Collins, A. M., & Resnick, L. B. (1996). Cognition and Learning. In D. Berliner & R. Calfee
(Eds.), Handbook of Educational Psychology (pp. 15–46). New York: Macmillan.
Grohs, J., Kinoshita, T., Novoselich, B., & Knight, D. (2015). Exploring Learner Engagement and Achievement
in Large Undergraduate Engineering Mechanics Courses (p. 26.729.1-26.729.11). ASEE Conferences.
https://doi.org/10.18260/p.24066
Grohs, J., Soledad, M., Knight, D., & Case, S. (2016). Understanding the Effects of Transferring In Statics
Credit on Performance in Future Mechanics Courses. ASEE Conferences.
https://doi.org/10.18260/p.27103
Halpern, D. F., & Hakel, M. D. (2003). Applying the science of learning to the university and beyond: Teaching
for long-term retention and transfer. Change: The Magazine of Higher Learning, 35(4), 36–41.
Hanson, J. H., & Williams, J. M. (2008). Using Writing Assignments to Improve Self-Assessment and
Communication Skills in an Engineering Statics Course. Journal of Engineering Education, 97(4), 515–
529.
Holm, S. (1979). A Simple Sequentially Rejective Multiple Test Procedure. Scandinavian Journal of Statistics,
6(2), 65–70.
Jenkins, G. A., Edwards, D., Nepal, K. P., & Bolton, M. (2011). Mapping student approaches to learning within
a civil engineering program. In AAEE 2011: Bridging sustainable community development and social
justice: the next ethical challenge for humanitarian engineers: proceedings of the 22nd annual AAEE
Conference on Engineering Education 5-7 December 2011, Fremantle, Western Australia (pp. 523–529).
Engineers Australia. Retrieved from http://dro.deakin.edu.au/eserv/DU:30054231/nepal-mappingstudent-
2011.pdf
Jones, B. D. (2009). Motivating Students to Engage in Learning: The MUSIC Model of Academic Motivation.
International Journal of Teaching and Learning in Higher Education, 21(2), 272–285.
Kaufman, L., & Rousseeuw, P. J. (2009). Finding Groups in Data: An Introduction to Cluster Analysis. John
Wiley & Sons.
Knight, D. B., Brozina, C., & Novoselich, B. (2016). An Investigation of First-Year Engineering Student and
Instructor Perspectives of Learning Analytics Approaches. Journal of Learning Analytics, 3(3), 215–238.
https://doi.org/10.18608/jla.2016.33.11
Ko, E. I., & Hayes, J. R. (1994). Teaching Awareness of Problem-Solving Skills to Engineering Freshmen.
Journal of Engineering Education, 83(4), 331–336.
Larkin, J., McDermott, J., Simon, D. P., & Simon, H. A. (1980). Expert and Novice Performance in Solving
Physics Problems. Science, 208(4450), 1335–1342. https://doi.org/10.1126/science.208.4450.1335
Lawanto, O. (2010). Students’ metacognition during an engineering design project. Performance Improvement
Quarterly, 23(2), 117–136. https://doi.org/10.1002/piq.20084
Litzinger, T., Van Meter, P., Firetto, C., Passmore, L., Masters, C., Turns, S., … Zappe, S. (2010). A Cognitive
Study of Problem Solving in Statics. Journal of Engineering Education, 99(4), 337–337.
Lord, S. M., & Chen, J. C. (2014). Curriculum Design in the Middle Years. In Cambridge handbook of
engineering education research (pp. 181–195). New York, NY: Cambridge University Press.
Marzano, R. J., & others. (1988). Dimensions of thinking: A framework for curriculum and instruction. ERIC.
Retrieved from https://eric.ed.gov/?id=ED294222
Mazumder, Q. H. (2012). Improving Confidence Level and Performance of First Generation and Female
Students using Metacognition Strategies (p. 25.743.1-25.743.11). Presented at the 2012 ASEE Annual
Conference & Exposition. Retrieved from https://peer.asee.org/improving-confidence-level-and-
performance-of-first-generation-and-female-students-using-metacognition-strategies
Mazumder, Q. H., & Ainsworth, A. (2010). Use Of Metacognition Strategy To Improve Student Learning (p.
15.1306.1-15.1306.8). Presented at the 2010 Annual Conference & Exposition. Retrieved from
https://peer.asee.org/use-of-metacognition-strategy-to-improve-student-learning
McCord, R., & Matusovich, H. M. (2013). Developing an Instrument to Measure Motivation, Learning
Strategies and Conceptual Change. In 120th ASEE Annual Conference & Exposition, Atlanta, GA.
Retrieved from
http://www.asee.org/file_server/papers/attachment/file/0003/3450/ASEE_CAREER_Instrument_Develo
pment_final.pdf
McKenna, A. F., Johri, A., & Olds, B. M. (2014). Adaptive Expertise and Knowledge Fluency in Design and
Innovation. In Cambridge Handbook of Engineering Education Research (pp. 227–242). New York, NY:
Cambridge University Press.
Meyer, J. H., Knight, D. B., Callaghan, D. P., & Baldock, T. E. (2015a). An empirical exploration of
metacognitive assessment activities in a third-year civil engineering hydraulics course. European Journal
of Engineering Education, 40(3), 309–327.
Meyer, J. H., Knight, D. B., Callaghan, D. P., & Baldock, T. E. (2015b). Threshold concepts as a focus for
metalearning activity: application of a research-developed mechanism in undergraduate engineering.
Innovations in Education and Teaching International, 52(3), 277–289.
Newstetter, W. C., & Svinicki, M. D. (2014). Learning Theories for Engineering Education Practice. In
Cambridge handbook of engineering education research (pp. 29–46). New York, NY: Cambridge
University Press.
Olson, S., & Riordan, D. G. (2012). Engage to Excel: Producing One Million Additional College Graduates
with Degrees in Science, Technology, Engineering, and Mathematics. Report to the President. Executive
Office of the President. Retrieved from http://eric.ed.gov/?id=ED541511
Ormrod, J. E. (2012). Metacognition, Self-regulated learning, and Study strategies. In Human learning (6th ed,
pp. 352–388). Boston: Pearson.
Pintrich, P. R. (2000). The role of goal orientation in self-regulated learning. Handbook of Self-Regulation, 451,
451–502.
Prince, M. J., & Felder, R. M. (2006). Inductive Teaching and Learning Methods: Definitions, Comparisons,
and Research Bases. Journal of Engineering Education, 95(2), 123–138. https://doi.org/10.1002/j.2168-
9830.2006.tb00884.x
Seymour, E. (2008). Talking About Leaving: Why Undergraduates Leave The Sciences. Westview Press.
Turner, M. P. (2001). Metalearning in the engineering curriculum: work in progress. Presented at the
Australasian Association for Engineering Education Annual Meeting, Brisbane, Queensland.
Turner, M. P. (2004). The SPQ in monitoring progress. Presented at the Australasian Association for
Engineering Education Annual Meeting, Toowoomba, Queensland.
Voss, J. F., Greene, T. R., Post, T. A., & Penner, B. C. (1983). Problem-solving skill in the social sciences.
Psychology of Learning and Motivation, 17, 165–213.
Woods, D. R. (2000). An Evidence-Based Strategy for Problem Solving. Journal of Engineering Education,
89(4), 443–459. https://doi.org/10.1002/j.2168-9830.2000.tb00551.x
Zulkiply, N., Kabit, M. R., & Ghani, K. A. (2009). Metacognition: What Roles Does It Play in Students’
Academic Performance? International Journal of Learning, 15(11).
Author Information
Jacob R. Grohs David B. Knight
Virginia Tech Virginia Tech
Engineering Education (0218) Engineering Education (0218)
Goodwin Hall Goodwin Hall
635 Prices Fork Road 635 Prices Fork Road
Blacksburg, VA 24061 Blacksburg, VA 24061
U.S.A. U.S.A.
Contact e-mail: jrgrohs@vt.edu
Glenda D. Young Michelle M. Soledad

Virginia Tech Ateneo de Davao University
Engineering Education (0218) Lower Ground Floor
Goodwin Hall Finster Hall, Jacinto Campus
635 Prices Fork Road Philippines
Blacksburg, VA 24061
U.S.A.
Engineering Identity Development: A Review of the Higher Education

Literature
Sarah L. Rodriguez, Charles Lu, Morgan Bartlett

Article History The purpose of this systematic literature review was to appraise and synthesize
the current scholarship pertaining to engineering identity development within the
Received: higher education context and create recommendations for future scholarship
30 June 2017 within engineering education. A review of the literature concluded that research
on engineering identity development has increased over the past ten years, has
Accepted:
been conducted primarily with qualitative methods, and has been primarily
18 December 2017
limited to academic communities focused on mathematics, science, and
engineering education. In addition, current scholarship reflected that most of the
Keywords
work in this area has focused on the learning contexts and experiences of women
Engineering and underrepresented racial/ethnic minorities with less focus on men,
Identity international, or graduate students. Future scholarship in this area should focus
Systematic literature on expanding forms of engineering identity frameworks and focus to facilitate
review greater understanding of engineering identity development.
Higher education
Introduction
Engineering advancements have been central to human progress since the beginning of time. Major historical
events, from building the first canal to the invention of social media, were made possible because of engineers
who designed, innovated, and solved problems throughout the world. The key to the future is no exception. In
the United States, for example, the job outlook of all engineering professions in the U.S. is expected to grow in
the next decade, with Biomedical Engineers leading the way at a 26% market increase (U.S. Bureau of Labor
Statistics, 2016). Without a commitment to educating and sustaining the global pipeline of engineers, the world
will not continue to progress and reach its full potential.
Despite the increasing demand for professional engineers and increasing college enrollments overall,
stakeholders have expressed a shortage in identifying qualified, career-ready individuals (McCave, Gilmore, &
Burg, 2014; United Nations Educational, Scientific and Cultural Organization [UNESCO], 2010). This shortage
of engineers results from a lack of college engineering graduates available to meet the growing need for these
roles. Furthermore, the engineering context is influenced by a need to manage global competition and address an
aging and non-diverse workforce (UNESCO, 2010; World Economic Forum, 2015).
One aspect of engineering education that is still relatively underdeveloped (Patrick & Borrego, 2016) is the use
and discussion of engineering identity. Gee (1999, 2000) defined identity as “the ‘kind of person’ one is seeking
to be and enact in the here and now” (1999, p. 13). In addition, Gee also acknowledged that identity cannot be
claimed in isolation and requires the participation of others to recognize that identity. Identity development
examines how people progress in stages as they grow and develop throughout time (Chickering & Reisser,
1993; Erikson, 1964; Phinney, 1993). Engineering identity development is a rapidly growing area of
international scholarship, drawing on research from the engineering, educational, psychological, and
sociological fields. Prior research (e.g., McCave, Gilmore, & Burg, 2014; Owen & Rolfes, 2015; Tonso, 2014;
Trytten, Pan, Shehab, & Walden, 2015) has shown that students who do not identify with the field or are not
recognized by engineering faculty are more likely to leave the field altogether. To be successful, students must
see themselves as future engineers and be recognized by others as such. However, some of today’s best
potential engineers, specifically women and underrepresented minorities, are not studying engineering or have
chosen to leave the academic pipeline due to lack of identification with engineering (Nosek, Bnaji, &
Greenwald, 2002; Ong, Wright, Espinosa, & Orfield, 2011; Patrick & Borrego, 2016; Pierrakos, Beam,
Constantz, Johri, & Anderson, 2011; Ross & Godwin, 2016). If engineering departments across colleges and
universities can help cultivate stronger engineering identities, more women and minorities can be retained and
ultimately contribute to the domestic shortage of creative and technical talent. As this area of research grows, it
is imperative that scholars examine the current scholarship concerning engineering identity development. Such
an examination has the capacity to make connections between various areas of interdisciplinary research on the
topic to enhance college student engineering identity development.
The purpose of this systematic literature review is to appraise and synthesize the current scholarship pertaining
to engineering identity development specifically within higher education, and create recommendations for future
scholarship within engineering education. The following section addresses the way in which the authors scanned
and analyzed the available literature within this area to present relevant findings about engineering identity
development. The final section contains recommendations for future scholarship and is intended to challenge
scholars to enhance their understanding of engineering identity development to produce greater, more nuanced
understandings of this development process.
Methods
This study was a systematic review of the literature on engineering identity development in higher education,
which provides a synthesis of the available literature and a holistic approach to understanding this topic
(Baumeister, 2013; Baumeister & Leary, 2007). Due to the comprehensive nature of the topic, systematic
literature review was selected as the appropriate methodology; this type of review utilizes a set of clear methods
to identify, evaluate, and synthesize studies addressing a focused topic (Baumeister & Leary, 2007; Grant &
Booth, 2009; Hutchinson, 1993; Slavin, 1995). This type of review seeks to provide a comprehensive and
replicable search to minimize bias (Grant & Booth, 2009). A systematic review enables researchers to establish
the parameters of existing research, identify relationships within the literature, and then provide directions for
addressing gaps in knowledge and forming future research (Baumeister & Leary, 2007).
Data Sources
The researchers performed an initial search of the literature to refine the topic of engineering identity
development in higher education and to organize an overview of what was to be written (Polgar & Thomas,
1995). Initial searches of the literature enabled the researchers to understand published work within this area in
order to refine the topic and aims of this review (DePoy & Gitlin, 1993; Lang & Heiss, 1998). Abstracts were
assessed according to the agreed-upon inclusion criteria and articles were retrieved and critically assessed by the
researchers. Inclusion and exclusion boundaries were set to ensure that the researchers retrieved all relevant
studies but were still defined in a manner that would produce a nuanced understanding of the topic area (DePoy
& Gitlin, 1993; Oxman & Guyatt, 1988). Selection criteria were outlined to ensure that scholarship was
included due to relevance to the topical area rather than simply relying on the researchers’ familiarity or
agreement with the available scholarship (Gehlbach, 2006; Hutchinson, 1993; Oxman, 1994; Oxman & Guyatt,
1988; Slavin, 1995). Exclusion criteria were identified to outline how researchers eliminated studies from
consideration and defined the purpose of the study (DePoy & Gitlin, 1993). For this study, inclusion and
exclusion criteria included:
1. The study was published between 2000 and 2017.

2. The study was considered a peer-reviewed article or published conference proceeding.
3. The study was available in full-text and written in English.
4. The study made use of some form of engineering identity either
a. Exclusively or
b. In conjunction with other concepts
5. The study addressed the nature of engineering identity development in higher education.
These criteria were selected to keep the focus of the literature review clear, defined, and feasible (DePoy &
Gitlin, 1993). Researchers selected this 17-year time frame to reflect the recent literature that has been published
since the turn of the new millennium. Published peer-review articles and engineering education conference
proceedings were reviewed to include only updated, relevant, and rigorous articles. The decision to include only
full-text English studies was a reflection of the language skills of the researchers.
256 Rodriguez, Lu, & Bartlett
Engineering identity development as a topical area was selected due to its emerging popularity within multiple
disciplines and areas of scholarship. As a result, a review of the literature has the opportunity to influence a
greater number of scholars and educational stakeholders across disciplinary boundaries. The topical area is both
comprehensive as well as feasible due to its focus on engineering, rather than multiple fields of study or an
aggregation of fields as is often seen with scholarship on science, technology, engineering, and mathematics
(STEM). Furthermore, this work is narrow in its scope due to its focus on engineering identity development as a
construct at the higher education level. Rather than attempt to review the literature from all sectors of education,
this synthesis focuses on identity development transitioning into and through higher education.
Multiple library databases/digital libraries (Education Full Text, Education Resource Information Center
[ERIC], IEEE Xplore, JSTOR, Project Muse PsycINFO,) and published searchable conference proceedings
(American Society for Engineering Education [ASEE], Frontiers in Education [FIE]) were searched to provide
the most comprehensive understanding of engineering identity development scholarship. ASEE and FIE
conference proceedings were retrieved via the organizational websites (ASEE Conference Proceedings Search,
2017; IEEE Xplore Digital Library, 2017). These library databases/digital libraries and conference proceedings
were selected due to their influence on the topic (Portney & Watkins, 2000). Particular attention was paid to
selected key journals such as (alphabetically): IEEE Transactions on Education, International Journal of
Education in Mathematics, Science and Technology, Journal of College Student Development, and Journal of
Engineering Education.
Data Analysis
The researchers defined the scope of the review, gathering literature from peer-reviewed journal articles and
published conference proceedings to achieve a comprehensive understanding of engineering identity
development in higher education. The search terms used were: “engineering identity,” “development,” “higher
education,” “college,” and “professional identity.”
Abstracts were assessed according to agreed inclusion criteria. Selected articles and conference proceedings
were critically analyzed by members of the project team independently. The initial literature search identified
398 possible sources. On examination of the abstracts, 154 full-text articles and proceedings were retrieved. Our
final selection included only items that met the aforementioned inclusion criteria. The analysis reported here
included 88 articles focused on engineering identity development in higher education.
The researchers used a qualitative, interpretive approach to data analysis in which they read, reflected, and
created notes for each of the sources (Strauss & Corbin, 1990). Studies were critically appraised based on their
appropriateness to the research question. To determine appropriateness for this review, the authors evaluated
articles based on the content of their sections and used unprompted judgment based on their expert opinions. To
organize the scholarship under review, a spreadsheet was created to log the following information from each
article or conference proceeding: author(s), title, year, journal theoretical framework, methodology, key
findings, and calls for future research. When reviewing content for this study, authors sometimes coded articles
and conference proceedings under multiple categories to address their relevancy. Authors used frequency counts
and thematic analysis to evaluate and categorize studies. Once an entry was made for each article or conference
proceeding, thematic categories, relationships, and emergent themes were compared across the sources (Miles &
Huberman, 1994; Ventres, Kooienga, & Marlin, 2006). A conceptual map was created to connect the most
salient concepts across the literature. Throughout the literature review process, the researchers participated in
peer-debriefing activities in which they exchanged ideas on project design, analysis, and interpretation activities.
Collaborators held discussions to refine interpretations, organize emergent themes, and draw conclusions on the
available scholarship.
Trustworthiness & Researcher Positionalities
To conduct a trustworthy literature review, mechanisms were woven in at each stage (Creswell, 2013; Lincoln
& Guba, 1985). To establish trustworthiness, the three authors, who have different backgrounds, scrutinized the
data independently and explored their positionalities. To understand and reduce the influence of biases, values,
and experiences that the researchers were bringing with them to the project, the researchers explored their
“position” in relationship to project through the act of memoing, or short writing exercises (Creswell, 2013, p.
216). All three researchers hold graduate degrees within higher education and either currently or have worked
within the higher education field. The first author works as a faculty member within a higher education program
conducting research focused on STEM identity development. The second author directs academic support at a
research-intensive university and has a background as secondary science teacher. The third author, whose
interests center upon equity in education, has previous work experience in both engineering admissions and
programming at a research-intensive, highly competitive school of engineering. While all three authors have
done work in STEM education, none of the authors received a higher education degree in a science, math, or
engineering discipline. This should be considered with the analytical process because the authors may
theoretically understand or even be able to empathize with science identity development in the articles they
reviewed, but none of them have first-hand knowledge of the experiences students face in the STEM disciplines.
Limitations and Delimitations
Like any empirical study, delimitations existed for this research. The first delimitation, which was constructed
by the authors, was requiring “full-text” articles, which may infuse bias into the search. The “full-text”
definition may have different meaning for different databases, but the authors felt that having access to full-text
documents was important to read and code the articles accurately for data analysis. Another delimitation the
authors created was to read articles that were written in English. This was primarily done due to the authors’
language abilities and comfort levels, but also to avoid any misinterpretation that may occur as a result of
linguistic translations.
Results and Discussion

The results of this systematic literature review demonstrate that engineering identity development is a topic that
has received increasing attention within the past three years, and is primarily limited to discussion in academic
communities focused on math, science, and engineering education. Due to challenges they encounter that
ultimately led to low retention rates in the engineering field, the majority of work on engineering identity has
explored the dynamic in women and underrepresented students, with a focus on African-American and Latina/o
students. In addition, engineering identity literature has mainly been grounded in the context of learning and the
experience of students in the classroom. However, most of the works have examined engineering identity as a
unilateral framework and only few have explored the intersections of engineering identity with other elements of
identity. This results section follows the following outline: (1) the characteristics of included studies on
engineering identity development; (2) the focus of engineering identity research; and (3) the theoretical
approaches to engineering identity research.
Trends on Reviews Journals and Conference Proceedings
The studies, both peer-reviewed articles and published conference proceedings, included in our analysis were
published between 2000 and 2017. Our analysis included 53 peer-reviewed articles and 35 published conference
proceedings. The number of published peer-review articles and conference proceedings addressing engineering
identity has increased steadily over time. Around 2011, the scholarship on identity development began to
increase, and 2016 marked the greatest number of works (17/88) published on the topic. Of the included studies,
19% (17/88) were published in 2016; most were published in engineering education, general education, and
sociological/psychological journals (40/88). Studies were also found in math or science education journals
(8/88). Fewer works have been published in journals focused on career development (2/88), scholarship of
teaching and research (2/88), or research methods (1/88) (see Figure 1).
Most of the peer-reviewed articles on engineering identity development have been published in the Journal of
Engineering Education (18/88). Studies were also found in Cultural Studies of Science Education, Educational
Studies in Mathematics, Harvard Education Review, International Journal of College Student Development,
Journal of Educational Research, New Directions for Institutional Research, and NWSA Journal, among others.
The majority of engineering identity peer-reviewed articles were qualitative in nature (e.g., Bergerson,
Hotchkins, & Furse, 2014; Foor & Walden, 2009; Foor, Walden, & Trytten, 2007; Godwin, Potvin, & Lock,
2016). Fewer peer-reviewed articles were quantitative (e.g., Bix, 2004; Burack & Franks, 2004) or considered
mixed methods (e.g., Litchfield & Javernick-Will, 2015). A growing area of work on engineering identity is
starting to be seen in published conference proceedings (e.g., Cross & Paretti, 2012; Godwin, Potvin, Hazari, &
Lock 2013), with the majority of those conference proceedings being quantitative in nature. Fewer published
conference proceedings center upon theory (e.g., Cross & Paretti, 2012) or mixed methods (e.g., Chachra,
Kilgore, Loshbaugh, McCain, & Chen, 2008).
20
18
16
Number of Articles
14
12
10
8
6
4
2
0
Type of Journal
Figure 1. Number of articles by type of journal
Areas of Focus for Engineering Identity Development
Of the included studies addressing engineering identity, many have focused on the experiences of
underrepresented populations within the engineering disciplines. Overall, 25% (22) of studies addressed
underrepresented racial/ethnic minority students in engineering. There was a strong concentration of engineering
identity studies focused on the experiences of women students (e.g., Hug, Jurow, & Chi, 2011) or
underrepresented racial/ethnic minority students (e.g., Fleming, Smith, Williams, & Bliss, 2013). In addition,
scholars have addressed the unique experiences of women of color (see Black & Williams, 2013; Ross &
Godwin, 2016; Martin, Simmons, & Yu, 2013) and international students (Dutta, 2015), as well as the
experiences of other students with underrepresented or non-normative identities (see Hughes, 2017; Kirn,
Godwin, Benson, Potvin, & Verdin, 2016). Of the 88 studies reviewed, three directly addressed men of color
(Lu, 2015), gay men (Hughes, 2017), or engineering identity related to individual customs and religion (Black &
Williams, 2013). Studies specifically addressing engineering identity for Native American/American Indian,
Native Hawaiian/Pacific Islander, multi-racial, or White students were not found (See Figure 2).
Graduate School
First-Year
Topical Focus
Group Dynamics
Men of Color
International Students
Women and Gender
Racial and Ethnic Minority
0 5 10 15 20 25
Number of Articles
Figure 2. Number of articles by topical focus

Of the available scholarship, 25% (22) of studies focused on women and gender in engineering identity
development (Du, 2006; Johnson, 2011; Settles, O’Connor, & Yap, 2016). Studies also addressed issues of
persistence or retention (8/88) (Cech, Rubineau, Silbey, & Seron, 2011; Ross & Godwin, 2016) as well as group
or team dynamics component (6/88) (Tonso, 2006; Trytten, Pan, Foor, Shehab, & Walden, 2015). Fewer studies
(5/88) were found on first-year (see Atadero, Paguyo, Rambo-Hernandez, & Henderson, 2016; Green,
Mohammadi-Aragh, & Warnock, 2015; Author, 2015) and graduate engineering identity experiences (4/88) (see
Louis Kajfez & McNair, 2014).
Much of the available scholarship on engineering identity development is focused on the context of learning
(12/88) (e.g., Walther, Kellam, Sochacka, & Radcliffe, 2011; Tatar, Van Beek, & Lillienkamp, 2016), including
a focus on competence, curriculum, and learning strategies applied to both in-classroom and online learning
environments. In addition, several studies address the major choice, career goals, and trajectories of students in
relationship to their engineering identities (9/88) (see Lent, Sheu, Singley, Schmidt, Schmidt, & Gloster, 2008;
Thomas, 2014). Similarly, other studies have been written about behavior, aspirations, confidence, expectations,
balance, personality, social capital, engineering research, and self-concept (9/88) (e.g., Cech, Rubineau, Sibley,
& Seron, 2011; Martin, Simmons, & Yu, 2013; McCave, Gilmore, & Burg, 2014).
Fewer studies were written regarding the measurements and relevant factors associated with engineering identity
development (see Godwin, 2016). In particular, few studies addressed the relationships between agency,
motivation, and engagement and engineering identity development or disciplinary differences that might create
unique identity development experiences for students (e.g., Groen, Simmons, & McNair, 2016; Matusovich,
Streveler, & Miller, 2010). The majority of the reviewed studies represented cross-sectional or single-institution
scholarship rather than longitudinal or multi-institution scholarship (Prybutok, Patrick, Borrego, Seepersad, &
Kirisits, 2016; Matusovich, Streveler, & Miller, 2010; Stoup & Pierrakos, 2016).
While research on engineering identity development has continued to grow, much of the scholarship portrays
engineering identity development to include only one type of engineering identity rather than many types or
manifestations. Of the studies included, only three made reference to differing types of engineering identities,
including “leading” and “creative” identities (Black, Williams, Hernandez-Martinez, Davis, Pampaka, & Wake,
2010; Brookstein & Sadeghipour, 2016; Matusovich, Barry, Meyers, & Louis, 2011). In addition, very few
studies addressed or placed their studies at Minority Serving Institutions (MSIs) or Hispanic Serving Institutions
(HSIs) (see Fleming, Smith, Williams, & Bliss, 2013). As a result, much of the available engineering identity
literature does not address development within these institutional contexts.
Theoretical Approaches to Engineering Identity Research
Of the 88 peer-reviewed articles and published conference proceedings that were analyzed, many discussed
engineering identity in terms of other, more established, theories. Few articles, however, established new
theoretical frameworks focused directly on engineering identity development. Furthermore, identity
development theory was often built through foundations from other theories, such as motivation (e.g. desire,
influences to achieve a goal), choice (e.g. selection of college, decision-making), critical (e.g. critique of society
and culture), socialization (e.g. process of learning/participating in a culture), sociocultural (e.g. importance of
society to individual development), asset-based (e.g. anti-deficit concepts, utilizing strengths of individual or
community), developmental (e.g. way in which people grow or change), stage (e.g. distinct patterns of stages
over time, successive order), and persistence (e.g. retention in college or major) theories. Identity or role
theories (e.g., Matusovich, 2010; McClain, 2014; Meyers, 2012) were most often used to frame understandings
of engineering identity development (25/88). Identity or role theory considers how an individual performs as a
result of their desire to fill a socially defined role or category which comes with its own expectations and norms.
Studies approaching engineering identity through this lens often included discussions around crystallized
identity, normativity and non-normative identities, and navigation of identity (Carlone & Johnson, 2007; Louis
Kajfez & McNair, 2014; Forin, Adams, & Hatten, 2012). Role theory was also discussed in terms of discipline-
specific identification, group affiliations, and collective identities (Llewellyn, Pyke, Paterson, Landrum, Scarritt,
Cullers, & Warner, 2016; Foor, Walden, & Trytten, 2007). In this approach, engineering students are thought of
as navigating their identities or roles, and negotiating the various intersecting identities or dynamic experiences
that emerge.
12
10
Number of Articles 8
Theory Type
Figure 3. Number of articles by theory type
Critical theories were also utilized (10/88) to focus on engagement with gender, race, and the influence of
intersectional identities present within the engineering disciplines (e.g., Godwin & Potvin, 2016; Godwin,
Potvin, & Hazari, 2013; Trytten, Wong Lowe, & Walden, 2012). These critical frameworks focused on culture,
discourse, agency, and engagement as key aspects to the identity development process. In these approaches,
engineering students are considered to possess agency and self-efficacy within their educational journeys, yet
scholars also have recognized that context and myriad identities influence the engineering experience and
subsequent identity development process.
Similarly, socialization theories were also prevalent (9/88), particularly those focused on situated learning,
group theory, and communities of practice (Hernandez-Martinez, 2016; Hug, Jurow, & Chi, 2011; Owen &
Rolfes, 2015). Scholars discussed identity in the context of social practices and stages of socialization (or
alienation) of students. In this way, studies sought to understand the ways in which students are socialized into
or marginalized from engineering as a result of their interactions and identity development processes. Several
studies also utilized sociocultural and asset-based models (8/88) which emphasized the backgrounds and
knowledge that students were bringing with them (Llewellyn et al., 2016; Martin, Simmons, & Yu, 2013).
Cultural community wealth, social capital, sociocultural perspectives, and contextualized understandings on
learning were important ways that scholars approached engineering identity development. Learning theories
approached identity development from different epistemological stances as well as examined it in terms of post-
structural thinking, developmental asynchronous learning, and best practice orientation.
Within the broader literature, there is a debate as to how identity might be differentiated from other interrelated
concepts. As a result, persistence (2/88) and a host of other developmental or stage model theories (6/88) were
associated with engineering identity development, including concepts of sustainability, sexual-orientation
development, language, and psychological well-being. Motivation, choice, and expectations models were also
used as theoretical framings for these studies (9/88), particularly those utilizing or focusing on social cognitive
career theory, self-efficacy, self-determination theory, or talent development theory (e.g., Jones, Tendhar &
Paretti, 2016; Lent, et al, 2008; Perez, Cromley, & Kaplan, 2014).
Conclusion
As the demand for engineers grows in our global economy, engineering education stakeholders have increased
efforts to recruit and retain engineering students. Scholarship on engineering identity will enable these efforts to
be successful, but given the unique nature the engineering disciplines more nuanced research is needed to
understand engineering identity development more accurately. Through its review of the available literature on
engineering identity development, this article articulates the landscape of the available literature and explores
ways in which engineering identity development research might be expanded in the future.
Recommendations
Several gaps exist in our understanding of the engineering identity development process. Although scholars
have amassed a robust understanding of multiple aspects of the engineering experience as well as engineering
identity development, there are still areas in which future research should seek to fill gaps in our knowledge. In
particular, scholars may need to shift the design and focus of engineering identity studies to provide a wider
range of understanding of the engineering student experience. In the future, more research should be conducted
focusesd on the aforementioned areas to enhance our understanding of engineering identity development and to
inform local, regional, national, and global engineering policies.
Expanding Forms of Engineering Identity Studies
Future research should focus on longitudinal models of understanding engineering identity development,
specifically looking across the educational pipeline to understand how the concept of engineering identity might
change over time. Longitudinal work could help improve understanding how students, particularly
underrepresented racial/ethnic minorities and women, navigate their identities and how these students may be
asked to negotiate those identities within the engineering environment.
Scholarship needs to focus on defining the concepts and measures related to engineering identity development
with a wide variety of methods. More work is needed on the development and study of constructs and
instruments designed to interpret engineering identity (e.g. Godwin, 2016; Godwin, Potvin, Hazari, & Lock,
2013). Quantitative work should look into experimental designs and studies that incorporate larger samples
across time. In addition, scholarship using mixed-method approaches are greatly needed in the field. Merging
datasets has the capacity to produce more robust understandings, coupled alongside qualitative work that has the
capacity to focus on the lived experiences and voices of engineering students, particularly those who are
marginalized within this context.
Replication and updating is needed to understand if change has occurred since the original studies were done,
particularly thinking about the shift in politics, rhetoric, and colleges becoming more liberal places of thought
with supported diversity efforts. This is coupled with the reality that women and URMs remain
underrepresented within engineering. Future studies need to include a wider variety of student characteristics
(e.g., generational status, literacy, social support) as well as greater attention to a wider range of institutional
contexts (e.g., type, size, focus, geographic). Students across multiple disciplines who possess differing
characteristics could broaden our understanding of the multitude of experiences that students in engineering
have. In addition, examining disciplinary, institutional, and national contexts could help contextualize findings
and enhance transferability as well as explain why/how identity development occurs within certain contexts.
Shifting the Focus of Engineering Identity Studies
Greater attention is needed to focus on understanding identity from a student development standpoint,
understanding that identity development is not static and can change with environment and context, engagement
and interest, and socialization and alienation experiences. Understanding how students understand their roles
within engineering and beyond as well as how they navigate those roles may help our understanding of how
engineering identity develops. Scholarship should focus on what it means to be an engineer and how identity
development may be value-laden and heavily dependent on the held identities and contexts in which students
find themselves.
Future scholarship should focus on the need for looking within subdisciplines to understand how students
develop within these contexts. For instance, developing a civil engineering identity may be different from
developing a mechanical or electrical engineering identity. Few research studies examine subdisciplines of
engineering identity development and no empirical studies to date have looked at similarities or differences
across various subdisciplines. Furthermore, given the move towards interdisciplinary work, future identity work
should focus on how interdisciplinary engineering identities might form, particularly within group contexts.
Scholarship should focus on the organizational and engagement experiences of engineering students,
particularly of students involved in identity-based engineering groups and the ability of engineering students,
given heavy course loads, to engage in extra-curricular and often engineering identity-growing activities.
Careful look should be given to various engagement experiences, both within and outside of the curriculum, that
seek to promote professional identity development among students.
Future studies should seek to understand the expectation and realities of engineering students to understand
identity development. It is important to think about the expectations and previously-held understandings of
students as they come into the engineering programs, for these may help explain how and why some students
develop a strong engineering identity and others do not. More work is needed in understanding the role that
alienation and motivation play in developing an engineering identity and who becomes marginalized as a result.
Further research is needed to understand more fully the roles that URM status and gender play in the identity
development process as well as the cultural norms that remain in place that deter these students from being
successful. Attention to the culture and climate present at both the departmental and institutional levels is
needed to understand how professional identity development is systematically weaved into the college
experience and ways in which institutions can do more to support students.
Future work on engineering identity development may choose to focus on the roles that attitude and major
selection decisions have on the development of an engineering identity. Studies may include exploring how
intersectional identities play into the way that students view their disciplinary confidence, competence, and self-
efficacy, and how those issues relate to their identity development. Further scholarship may hone in on specific
teaching strategies that seek to support engineering identity development, which may increase the participation,
retention, and eventual persistence of engineering students at multiple levels of the engineering pipeline.
Table1. Summary of findings & implications for future research

Findings from Current Engineering Identity Research Implications for Future Research
 Engineering identity research concentrated  Shifting the Design of Engineering Identity
in engineering education, general education, Studies
and sociological/psychological journals o Longitudinal studies
o Defining measures
 Research currently focused on identity o Replication and updating
experiences of women and racial ethnic o Differing student characteristics and
minority students in engineering institutional contexts
 Shifting the Focus of Engineering Identity
 Role theory most often used to frame Studies
understandings of engineering identity o Student development centered
development o Exploring subdisciplines
o Organizational and engagement
experiences
o Expectations and marginalization
o Attitudes, culture, climate,
intersectionality
References
Atadero, R. A., & Paguyo, C., & Rambo-Hernandez, K. E., & Henderson, H. L. (2016, June), Promoting
inclusive engineering identities in first-year engineering courses. Paper presented at the American
Society for Engineering Education Annual Conference, New Orleans, Louisiana. 10.18260/p.25994
ASEE Conference Proceedings Search. (2017). https://www.asee.org/search/proceedings
Baumeister, R. F. (2013). Writing a literature review. In M. J. Prinstein & M. D. Patterson (Eds.), The portable
mentor: Expert guide to a successful career in psychology (pp. 119-132; 2nd ed.). New York: Springer
Science+ Business Media.
Baumeister, R. F., & Leary, M. R. (1997). Writing narrative literature reviews. Review of general
psychology, 1(3), 311.
Bergerson, A. A., Hotchkins, B. K., & Furse, C. (2014). Outreach and identity development: New perspectives
on college student persistence. Journal of College Student Retention: Research, Theory & Practice,
16(2), 165-185.
Bix, A. S. (2004). From" engineeresses" to" girl engineers" to" good engineers": a history of women's US
engineering education. NWSA journal, 16(1), 27-49.
Black, L., & Williams, J. (2013). Contradiction and conflict between ‘leading identities’: Becoming an engineer
versus becoming a ‘good Muslim’ woman. Educational Studies in Mathematics, 84(1), 1-14.
Black, L., Williams, J., Hernandez-Martinez, P., Davis, P., Pampaka, M., & Wake, G. (2010). Developing a
‘leading identity’: The relationship between students’ mathematical identities and their career and higher
education aspirations. Educational Studies in Mathematics, 73(1), 55.
Brookstein, D. & Sadeghipour, K. (2016). Shaping new student identity as ‘creatives’ in the 21 st century global
economy. Paper presented at the American Society for Engineering Education Annual Conference, New
Orleans, LA.
Burack, C., & Franks, S. (2004). Telling stories about engineering: Group dynamics and resistance to diversity.
NWSA Journal, 16(1), 79-95.
Carlone, H. B., & Johnson, A. (2007). Understanding the science experiences of successful women of color:
Science identity as an analytic lens. Journal of Research in Science Teaching, 44(8), 1187-1218.
Cech, E., Rubineau, B., Silbey, S., & Seron, C. (2011). Professional role confidence and gendered persistence in
engineering. American Sociological Review, 76(5), 641-666.
Chachra, D., Kilgore, D., Loshbaugh, H., McCain, J., & Chen, H. (2008). Being and becoming: Gender and
identity formation of engineering students. Research Brief. Center for the Advancement of Engineering
Education (NJ1).
Creswell, J. W. (2013). Research design: Qualitative, quantitative, and mixed methods approaches. Sage
publications.
Cross, K. J., & Paretti, M. C. (2012, June), Identification with academics and multiple identities: Combining
theoretical frameworks to better understand the experiences of minority engineering students. Paper
presented at the American Society for Engineering Education Annual Conference, San Antonio, Texas.
https://peer.asee.org/21467
Day, R. A. & Gastel, B. (2012). How to write and publish a scientific paper. 7th ed. Cambridge, England:
University Printing House.
DePoy, E. & Gitlin, L. N. (1993). Introduction to research: multiple strategies for health and human services.
St. Louis: Mosby-Year.
Du, X. Y. (2006). Gendered practices of constructing an engineering identity in a problem-based learning
environment. European Journal of Engineering Education, 31(1), 35-42.
Dutta, D. (2015), Sustaining the pipeline: Experiences of international female engineers in U.S. graduate
programs. Journal of Engineering Education, 104(3): 326–344.
Fleming, L. N., Smith, K. C., Williams, D. G., & Bliss, L. B. (2013). Engineering identity of Black and
Hispanic undergraduates: The impact of minority serving institutions. Paper presented at the American
Society for Engineering Education Annual Conference, Atlanta, GA.
Foor, C. E., & Walden, S. E. (2009). " Imaginary Engineering" or" Re-imagined Engineering": Negotiating
Gendered Identities in the Borderland of a College of Engineering. NWSA journal, 21(2), 41-64.
Foor, C. E., Walden, S. E., & Trytten, D. A. (2007). “I wish that I belonged more in this whole engineering
group:” Achieving individual diversity. Journal of Engineering Education, 96(2), 103-115.
Forin, T. R., Adams, R., & Hatten, K. (2012). Crystallized identity: A look at identity development through
cross-disciplinary experiences in engineering. Paper presented at the American Society for Engineering
Education Annual Conference, San Antonio, TX.
Gee, J.P. (1999). An introduction to discourse analysis: Theory and method. New York: Routledge.
Godwin, A.G. (2016). The development of a measure of engineering identity. Paper presented at the American
Society for Engineering Education Annual Conference, New Orleans, LA.
Godwin, A., & Potvin, G. (2017). Pushing and pulling Sara: A case study of the contrasting influences of high
school and university experiences on engineering agency, identity, and participation. Journal of Research
in Science Teaching, 54(4), 439-462.
Godwin, A.G., Potvin, G., Hazari, Z., & Lock, R. (2013, October). Understanding engineering identity through
structural equation modeling. Paper presented at the Frontiers in Education Conference, IEEE (pp. 50-
56). IEEE.
Godwin, A.G., Potvin, G., Hazari, Z., & Lock, R. (2016). Identity, critical agency, and engineering: An affective
model for predicting engineering as a career choice. Journal of Engineering Education, 105(2), 312-340.
Grant, M. J. & Booth, A. (2009). A typology of reviews: an analysis of 14 review types and associated
methodologies. Health Information and Libraries Journal, 26(2), 91–108. doi: 10.1111/j.1471-
1842.2009.00848.x.
Green, M. R. A., & Mohammadi-Aragh, M. J., & Warnock, J. (2015, June). Promoting Engineering Identity
through a Pre-Semester Freshman Design Competition. Paper presented at the American Society for
Engineering Education Annual Conference, Seattle, WA.
Groen, C. J., Simmons, D. R., & McNair, L. D. (2016, June). Disciplinary influences on the professional
identity of civil engineering students: Starting the conversation. Paper presented at American Society for
Engineering Education Annual Conference, New Orleans, LA.
Hernandez-Martinez, P. (2016). “Lost in transition”: Alienation and drop out during the transition to
mathematically-demanding subjects at university. International Journal of Educational Research, 79(1),
231-239.
Hug, S., Jurow, A. S., & Chi, W. C. (2011). Evolving identities: Undergraduate women pursuing the
engineering professoriate. Paper presented at the American Society for Engineering Education Annual
Conference, Vancouver, Canada.
Hughes, B. E. (2017). " Managing by Not Managing": How Gay Engineering Students Manage Sexual
Orientation Identity. Journal of College Student Development, 58(3), 385-401.
Hutchinson, B. G. (1993). Critical Appraisal of Review Articles. Canadian Family Physician, 39, 1097–1102
IEEE Xplore Digital Library. (2017). http://ieeexplore.ieee.org/Xplore/home.jsp
Johnson, D. R. (2011). Women of color in science, technology, engineering, and mathematics (STEM). New
Directions for Institutional Research, 2011(152), 75-85.
Jones, B. D., Tendhar, C., & Paretti, M. C. (2016). The effects of students’ course perceptions on their domain
identification, motivational beliefs, and goals. Journal of Career Development, 43(5), 383-397.
Kirn, A.K., Godwin, A., Benson, L., Potvin, G., Doyle, J., Boone, H., & Verdin, D. (2016). Intersectionality of
non-normative identities in the cultures of engineering. Paper presented at American Society for
Engineering Education Annual Conference, New Orleans, LA.
Lang, G. & Heiss, G. D. (1998). A practical guide to research methods. Lanham, MD:
University Press of America.
Lent, R. W., Sheu, H. B., Singley, D., Schmidt, J. A., Schmidt, L. C., & Gloster, C. S. (2008). Longitudinal
relations of self-efficacy to outcome expectations, interests, and major choice goals in engineering
students. Journal of Vocational Behavior, 73(2), 328-335.
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Beverley Hills, CA: Sage.
Litchfield, K., & Javernick‐Will, A. (2015). “I am an Engineer AND”: a mixed methods study of socially
engaged engineers. Journal of Engineering Education, 104(4), 393-416.
Llewellyn, D. C., Pyke, P., Paterson, S., Landrum, R. E., Scarritt, A., Cullers, J., & Warner, D. L. (2016). Better
Together: Connecting with Other Disciplines Builds Students' Own Skills and Professional Identity.
Louis Kajfez, R. L. & McNair, L. D. (2014). Graduate student identity: A balancing act between roles. Paper
presented at the American Society for Engineering Education Annual Conference, Indianapolis, IN.
Lu, C. (2015). Finding "Los Científicos" within: Latino Male Science Identity Development in the First College
Semester. Journal of College Student Development, 56(7), 740-745.
Martin, J. P., Simmons, D. R., & Yu, S. L. (2013). The role of social capital in the experiences of Hispanic
women engineering majors. Journal of Engineering Education, 102(2), 227-243.
Matusovich, H. M., Barry, B. E., Meyers, K., & Louis, R. (2011). A multi-institution comparison of students’
development of an identity as an engineer. Paper presented at the American Society for Engineering
Education Annual Conference, Vancouver, Canada.
Matusovich, H. M., Streveler, R. A., & Miller, R. L. (2010). Why do students choose engineering? A
qualitative, longitudinal investigation of students' motivational values. Journal of Engineering
Education, 99(4), 289-303.
McCave, E.J., Gilmore, J., Burg, K (2014). Engineering and science student preparedness for research:
Exploring the connections between student identity and readiness for research. Paper presented at the
American Society for Engineering Education Annual Conference, Indianapolis, IN.
Miles, M. & Huberman, A. (1994) Qualitative Data Analysis: An Expanded Sourcebook. London: SAGE
Publications.
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). Math= male, me= female, therefore math≠
me. Journal of personality and social psychology, 83(1), 44.
Ong, M., Wright, C., Espinosa, L., & Orfield, G. (2011). Inside the double bind: A synthesis of empirical
research on undergraduate and graduate women of color in science, technology, engineering, and
mathematics. Harvard Educational Review, 81(2), 172-209.
Owen, C. & Rolfes, D. (2015). Communication class size and professional identity. Paper presented at the
American Society for Engineering Education Annual Conference, Seattle, WA.
Oxman, A. D. (1994). Systematic reviews: checklists for review articles. British Medical Journal, 309 (6955),
648–51.
Oxman, A. D & Guyatt, G. H. (1988). Guidelines for reading literature reviews. CMAJ, 138, 697–703.
Patrick, A.D. & Borrego, M. (2016). A review of the literature relevant to engineering identity. Paper presented
at the American Society for Engineering Education Annual Conference, New Orleans, LA.
Perez, T., Cromley, J. G., & Kaplan, A. (2014). The role of identity development, values, and costs in college
STEM retention. Journal of Educational Psychology, 106(1), 315.
Pierrakos, O., Beam, T. K., Constantz, J., Johri, A., & Anderson, R. (2009, October). On the development of a
professional identity: Engineering persisters vs engineering switchers. Paper presented at the Frontiers in
Education Conference, San Antonio, TX.
Polgar, S. & Thomas, S. A. (1995). Introduction to research in the health sciences. 3rd ed. Melbourne:
Churchill Livingstone.
Portney, L. G. & Watkins, M.P. (2000). Foundations of clinical research: applications to practice. 2nd ed.
Upper Saddle River, NJ: Prentice Hall.
Prybutok, A., Patrick, A., Borrego, M., Seepersad, C. C., & Kirisits, M. (2016, January). Cross-sectional Survey
Study of Undergraduate Engineering Identity. Paper presented at the American Society for Engineering
Education Annual Conference, New Orleans, LA
Ross, M. M. S., & Godwin, A. (2016). Engineering identity implications on the retention of Black women in
engineering industry. Paper presented at the American Society for Engineering Education Annual
Conference, New Orleans, LA.
Settles, I. H., O’Connor, R. C., & Yap, S. C. (2016). Climate Perceptions and Identity Interference Among
Undergraduate Women in STEM: The Protective Role of Gender Identity. Psychology of Women
Quarterly, 40(4), 488-503.
Slavin, R. E. (1995). Best evidence synthesis: An intelligent alternative to metaanalysis. Journal of Clinical
Epidemiology, 48(1), 9–18.
Stoup, K.D. & Pierrakos, O. (2016). Engineering students’ self-concept differentiation: Investigation of identity,
personality, and authenticity implications for program retention. Paper presented at the American Society
for Engineering Education Annual Conference, New Orleans, LA.
Strauss, A. L. & Corbin, J. M. (1990) Basics of Qualitative Research: Grounded Theory Procedures and
Techniques. Newbury Park, CA: Sage Publications.
Tatar, N., Van Beek, L., & Lilienkamp, L.A. (2016). Conceptualizing student identity development through self-
directed learning opportunities in the first year of an engineering program. Paper presented at the
American Society for Engineering Education Annual Conference, New Orleans, LA.
Thomas, L.D. (2014). Identity-trajectory as a theoretical framework in engineering education research. Paper
presented at the American Society for Engineering Education Annual Conference, Indianapolis, IN.
Tonso, K. L. (2006). Teams that work: Campus culture, engineer identity, and social interactions. Journal of
Engineering Education, 95(1), 25-37.
Tonso, K. (2014). Engineering identity. In A. Johri & B. Olds (Eds.), Cambridge handbook of engineering
education research (pp. 267-282). New York, NY: Cambridge University Press.
Trytten, D.A., Pan, R., Shehab, R.L., Walden, S.E. (2015). Inclusion or exclusion? The impact of the
intersection of team culture and student identity and pathway on team diversity. Paper presented at the
American Society for Engineering Education Annual Conference, Seattle, WA.
Trytten, D. A., Wong Lowe, A., & Walden, S. E. (2012). “Asians are Good at Math. What an Awful
Stereotype” The Model Minority Stereotype's Impact on Asian American Engineering Students. Journal
of Engineering Education, 101(3), 439-468.
United Nations Educational, Scientific and Cultural Organization [UNESCO]. (2010). Engineering: issues,
challenges, and opportunities for development. Paris, France: United Nations Educational, Scientific and
Cultural Organization. Retrieved from: http://unesdoc.unesco.org/images/0018/001897/189753e.pdf
Ventres, W., Kooienga, S. & Marlin, R. (2006) EHRs in the exam room: tips on patient-centered care. Family
Practice Management, 13 (3), 45–47.
Walther, J., Kellam, N., Sochacka, N., & Radcliffe, D. (2011). Engineering competence? An interpretive
investigation of engineering students' professional formation. Journal of Engineering Education, 100(4),
703-740.
World Economic Forum. (2015). The Human Capital Report 2015. Retrieved from:
http://www3.weforum.org/docs/WEF_Human_Capital_Report_2015.pdf
Author Information
Sarah L. Rodriguez Charles Lu
Iowa State University University of California – San Diego
2666B Lagomarcino Hall 9500 Gilman Drive
901 Stange Road San Diego, CA 92093
Ames, Iowa 50011-1041 U.S.A.
U.S.A.
Contact e-mail: srod@iastate.edu
Morgan Bartlett
College Possible Minnesota
540 Fairview Ave N #304
St Paul, MN 55104
U.S.A.
From Deficit Thinking to Counter Storying: A Narrative Inquiry of

Nontraditional Student Experience within Undergraduate Engineering
Education
Angela Minichiello

Article History This paper reports on narrative inquiry research conducted with nontraditional
undergraduates as they pursued a common goal of becoming engineers.
Received: Participants were enrolled in a two-year engineering transfer program offered in
30 June 2017 the evenings, via synchronous broadcast distance instruction, by a public, land-
grant university situated in the western United States. The purpose of the
Accepted:
research was to understand the experiences of nontraditional students during the
20 December 2017
two-year program, which they accessed at regional campuses within or near their
local communities, and later as they transitioned to the four-year university
Keywords
campus located in another city away from these communities. In this paper,
Adult learners personal, social, and institutional tensions, arising from instances of deficit
Broadening participation thinking and emerging from within the participants‟ lived and told stories, are
Dewey identified and examined. Despite deep, personal reactions to the bias they
Engineering transfer experienced, participants overcame perceptions of personal deficiency to pursue
student and achieve their goal of becoming engineers. Participant counter stories further
revealed ways in which their unique life experiences alternately served to
enhance and deepen their engineering education.
I know I have put myself in this situation, as far as being a nontraditional student, but it sure is tough
trying to hang in there with the traditional students. I keep feeling like the instructors/advisors must
think I am a terrible student and that engineering may not necessarily be what I should be pursuing. If
I had a chance to explain my situation, that school is really my third (major) priority, maybe they
would better understand my seemingly lackadaisical effort, when sometimes that effort is truly all I
can give.
(Joe, reflecting on his interim research text, January 2016)
Introduction
Within the United States today, policymakers search for effective ways to expand the nation‟s engineering
workforce while addressing underrepresentation in science, technology, engineering, and mathematics (STEM).
Recent recommendations call for establishment of innovative programs capable of broadening STEM
undergraduate participation in ways that include the talent residing within diverse student groups. Specifically,
current federal policy guidance explicitly states that nontraditional students should be top priorities for STEM
recruitment initiatives, degree programs, and interventions (President's Council of Advisors on Science and
Technology (PCAST), 2012, p. vii). As a group, nontraditional students include older, working adults and others
whose life circumstances differ markedly from those possessed by undergraduates who enter college directly
after high school. Moreover, the United States is not alone in its concerns regarding the likelihood for engineer
shortages; many countries of Europe, as well as other industrialized nations throughout the world, share unease
regarding the resistance of today‟s youth to pursue engineering careers, high attrition rates among current
engineering students, and sustained low enrollment of women in engineering (Becker, 2010, p. 349; Morice,
1990). Thus, efforts made toward building in-depth understanding of nontraditional student participation,
achievement, and retention in STEM may be considered widely relevant across a global landscape.
The purpose of this research is to examine the lived experience of a group of nontraditional students who
participated in a distance-delivered, undergraduate engineering transfer program. Offered from 2009-2016 by a
mid-size, public, land grant institution situated in the western United States, the program aimed to improve
participation of geographically dispersed, rural, and/or working residents located throughout the state by
providing access to the first two-years of engineering education within (or close to) residents‟ local
communities. The current study, part of a larger project focused on developing a deeper understanding of the
experience of success among nontraditional engineering undergraduates through narrative inquiry, was guided
by the following research question: How do nontraditional undergraduates within engineering education
experience bias, related to themselves as financially self-supporting, adult learners, and with what outcome(s)?
Literature Review
The literature review consists of five sections. The first section focuses on how the nontraditional student
construct has been defined and operationalized within the research literature. The second section reviews
research related to nontraditional student outcomes in higher education; the third section examines factors found
to inhibit nontraditional student participation and success. The fourth section evaluates research related to
perceptions of nontraditional and adult undergraduates. The review concludes with a discussion of this study‟s
contribution to the existing literature.
Who are Nontraditional Undergraduates?
For nearly five decades, the term nontraditional (see, e.g., Cross, 1980, 1981) has been used to describe post-
secondary students whose set of experiences differ from those historically considered as typical or normative
(i.e., “traditional”) among undergraduates enrolling at four-year colleges and universities. Choy (2002, p. 1)
defined the traditional undergraduate as “… one who earns a high school diploma, enrolls full time
immediately after finishing high school, depends on parents for financial support, and either does not work
during the school year or works part time.” In contrast, it has proven difficult to define who nontraditional
undergraduates are as directly and succinctly. Scholars (e.g., Bean & Metzner, 1985; Horn, 1996, p. 3; Jones &
Watson, 1990) have identified several characteristics as critical facets of nontraditional undergraduate
experience, including: personal factors such as age, race, ethnicity, and gender; situational factors such as
residence (i.e., commuter students) and enrollment in non-degree (i.e., certificate) programs; socio-economic
factors such as socio-economic status (SES), employment level, and financial independence; and academic
factors such as first generation college status and enrollment patterns (e.g., full-time or part-time enrollment),.
Amid critique of an ongoing monolithic treatment of nontraditional students within the research literature
(Pascarella & Terenzini, 1998; Richardson & King, 1998), Horn (1996) proposed a continuum approach to
identify nontraditional students in a way that captures as well as categorizes the complexities inherent to their
experiences. The model locates undergraduates along a continuum of experience; demarcations along the
continuum are defined as the number of statistical risk factors that each student possesses. These risk factors, or
nontraditional student “characteristics” (Horn, 1996, p. i), reflect the “choices and behavior that may increase
students‟ risk of attrition and as such, are amenable to change or intervention at various stages in a students‟
school life” (Horn, 1996, p. 3). They include: (a) delaying college enrollment by one year or more; (b) attending
college part-time; (c) supporting themselves financially while enrolled; (d) working full-time while enrolled; (e)
having dependents other than a spouse; (f) being single parents; and (g) having earned a GED or other
equivalency certificate in place of a high school diploma (Horn, 1996, p. i). While Horn (1996) classifies
undergraduates who display at least one characteristic as generally nontraditional, she categorizes
undergraduates as minimally (possessing one characteristic), moderately (possessing two or three
characteristics), or highly nontraditional (possessing four or more characteristics). Doing so accounts for an
increasing risk of attrition among students who possess more than one to several characteristics (e.g., see
Berkner, Cuccaro-Alamin, & McCormick, 1996; Berkner, He, & Catadli, 2002).
This study adopts Horn‟s (1996) model to guide purposeful sampling. Because sufficient but not extreme
intensity (Patton, 2002) of experience is important to an examination of nontraditional student experience in
engineering, volunteers who self-reported having three to six nontraditional student characteristics were
purposefully selected for this study. In doing so, participants could be categorized as solidly within, but not at
the extremes of, the moderately and highly nontraditional undergraduate categories proposed by Horn (1996)
(see also the participant selection section).
268 Minichiello
Nontraditional Undergraduate Outcomes
Researchers in the United States report that nontraditional undergraduates are a growing population who
currently outnumber those who enter post-secondary institutions possessing traditional backgrounds. Several
studies indicate that more than 70% of U.S. undergraduates (i.e., aggregated across community and 2-year and
4-year colleges and 4-year universities) possess at least one nontraditional characteristic and can be categorized
as nontraditional (Choy, 2002; Horn, 1996; Radford, Cominole, & Skomsvold, 2015). Choy (2002) also
reported that the number of U.S. undergraduates who are highly nontraditional, when aggregated across all post-
secondary institutions (i.e., 2-year and 4-year institutions) is equivalent to the number of traditional U.S.
undergraduates. However, traditional undergraduates continue to predominate at 4-year institutions where most
professional STEM degrees (e.g., engineering degrees) are awarded (Choy, 2002; Horn, 1996).
Despite increasing numbers of nontraditional students enrolling in higher education, traditional undergraduates
continue to achieve higher outcomes on average. As a group, nontraditional undergraduates are twice as likely
as traditional students to leave college during their first year and remain less likely than traditional students to
earn undergraduate (i.e., associate‟s or bachelor‟s) degrees (Choy, 2002; Horn, 1996; NCES, 2002). Choy
(2002, pp. 16-17), for example, found that the first-year attrition rate for U.S. nontraditional students pursuing
bachelor‟s degrees in 1989-1990 was 27%, compared to 14% for traditional students. During the same period,
the first-year attrition rate for nontraditional students pursuing associate‟s degrees was 46%, compared to 23%
for traditional students (Choy, 2002, pp. 16-17). Others reported that the five-year degree completion rate for
nontraditional undergraduates enrolling with the goal of earning a bachelor‟s degree (37%) lags well behind that
of traditional students (54%) (NCES, 2002, p. 34). Moreover, Horn (1996) found that attrition rates increase
with nontraditional student status; 42% of minimally nontraditional students earned a bachelor‟s degree in five
years, compared to 17% of moderately nontraditional students and 11% of highly nontraditional students (p. ii).
Nontraditional Undergraduate Outcomes in STEM
STEM undergraduates are often considered separately from those who matriculate in non-STEM majors due to
the marked contextual variations between their experiences. Drawing from several nationally representative,
longitudinal datasets of postsecondary undergraduates in the United States (i.e., Beginning Postsecondary
Students Longitudinal Study (BPS: 96/01), National Postsecondary Student Aid Study (NPSAS:04), and the
Education Longitudinal Study of 2002/06 (ELS:02/06)), Chen and Weko (2009) reported that STEM
undergraduates who were over age 20 or financially self-supporting were statistically less likely than younger
and financially dependent undergraduates to enroll, persist, or complete bachelor‟s degrees. While younger and
financially dependent students were likely to leave STEM programs by changing to a non-STEM major, older
and financially self-supporting students tended to leave STEM programs by exiting postsecondary education
altogether (Chen & Weko, 2009). Although research that examines nontraditional student achievement and
retention within engineering is nascent, it is on the rise. A substantial portion of current work focuses on
understanding engineering transfer student pathways (e.g., Mobley & Brawner, 2013; Mobley, Shealy, &
Brawner, 2012, 2013; Ogilvie, 2014; Ogilvie et al., 2015, 2016; Shealy, Brawner, Mobley, & Layton, 2013;
Sullivan et al., 2012). Ogilvie (2014), for example, points to a gap in the literature related to transfer student
pathways in engineering and calls for more research to better understand engineering transfer student success.
Others (McNeill, Long, & Ohland, 2014; McNeill & Ohland, 2015, 2016) examined data from the Multiple-
Institution Database for Investigating Engineering Longitudinal Development (MIDFIELD). MIDFIELD
contains longitudinal, engineering undergraduate record level data gathered within 11 large and public research
universities in the United States. McNeill and Ohland (2015) reported that nontraditional engineering
undergraduates (i.e., those age 24 or older at first matriculation) were more likely than traditional (i.e., younger)
engineering students to graduate with engineering degrees in six years; they also found nontraditional (i.e.,
older) engineering students were to be more likely to leave the institution without a degree. These seemingly
contradictory findings suggest that nontraditional (i.e., older) engineering undergraduates may possess a high
level of commitment to earning an engineering degree and/or entering the engineering profession. Yet, if they
cannot complete the engineering degree, they may be unaware of opportunities within majors outside of
engineering, uninterested in other majors, or lack adequate resources to change majors and continue (McNeill &
Ohland, 2015, p. 5). Others examining MIDFIELD data (McNeill et al., 2014; McNeill & Ohland, 2016)
reported that the academic outcomes of nontraditional students (i.e., those age 24 or older at first matriculation),
measured as mean engineering grade point average (GPA), mean final cumulative GPA, and six-year
engineering graduation rate, were on par with or better than those of their traditional counterparts.
Because MIDFIELD contains few categories that reflect nontraditional student characteristics (see McNeill &
Ohland, 2015; McNeill & Ohland, 2016, for discussions of current MIDFIELD limitations and plans for its
upcoming expansion), researchers who explore nontraditional student outcomes in engineering using
MIDFIELD data often define nontraditional students through age and/or enrollment status only. McNeill and
Ohland (2016), who disaggregated MIDFIELD data related to student transfer status as well as age and
enrollment status, reported that transfer status had a larger effect on student outcomes than age and/or
enrollment status. Since transfer status is often considered to be a proxy for financial self-support, full-time
employment, and/or having dependents, this finding suggests that consideration of additional nontraditional
student factors is important for understanding the nontraditional student experience in engineering.
Barriers Faced by Nontraditional Undergraduates
Researchers point to several barriers that inhibit full participation of nontraditional students in postsecondary
education; time spent fulfilling other responsibilities (e.g., family obligations and work requirements) is perhaps
the factor that is most often highlighted for impacting nontraditional student achievement and retention
(Gonclaves & Trunk, 2014; Kenner & Weinerman, 2011; National Survey of Student Engagement Annual
Report, 2006; Wyatt, 2011). Other factors include administrative inflexibility toward nontraditional student
needs and scheduling constraints (Benshoff & Lewis, 1992; Gonclaves & Trunk, 2014; Zacharakis, Steichen,
Diaz de Sabates, & Glass, 2011), inattention and uncaring attitudes of academic advisors and staff members
towards nontraditional students (Gonclaves & Trunk, 2014; Wyatt, 2011), a lack of nontraditional student
organizations and peer networking opportunities (Benshoff & Lewis, 1992; Gonclaves & Trunk, 2014), and
pervasive feelings of isolation and not fitting in among student peers (Gonclaves & Trunk, 2014). Scott and
Lewis (2012) observed that older (age 48+) students held differing “perceptions of acceptance” among faculty
and younger student peers (p. 1). While the nontraditional students felt that interactions with their instructors
were positive both in and out of the classroom, they considered their interactions with younger student peers to
be “not as communicative,” “distant,” “minimal,” and even “hostile” (Scott & Lewis, 2012, p. 6). Hostile
interactions were attributed to several factors including: (a) the substantial interaction among nontraditional
students and the instructor; (b) a majority traditional student population in the class; (c) ethnic diversity within
the class; and (d) minimal support for student peer interaction provided within the classroom curricular
framework (Scott & Lewis, 2012, p. 6).
Other findings report that nontraditional students tend to ask more questions in class, prepare more drafts of
writing assignments, and come to class unprepared less frequently than traditional students (National Survey of
Student Engagement, 2006). Others scholars (Bye, Pushkar, & Conway, 2007; Zacharakis et al., 2011) report
that interactions with caring faculty, who treat each student as “an active partner in a shared learning
experience” (Bye et al., 2007, p. 155), have positive effects on nontraditional students‟ motivations and interest.
These results point to the profound impact that interactive classroom experiences and supportive student-
instructor relationships may have on nontraditional student success.
Perceptions of Nontraditional Undergraduates
In an era characterized by substantial growth in the number of nontraditional students pursuing postsecondary
education, several scholars (Donaldson & Townsend, 2007; Pascarella & Terenzini, 1998) critique a lack of
research attention paid to “adults students‟ presence and their impact upon nonprofit higher education”
(Donaldson & Townsend, 2007, p. 28). Quinnan (1997) suggested that there has been little advancement, in
terms of research questions, methodologies, and protocols, in adult undergraduate research since its inception.
Donaldson and Townsend (2007) contend that the discourse related to adults students within the literature is
inconsistent and “… frequently treats them as different, but not positively different, from traditional students, or
else it accepts them but does not embrace them” (p. 45), pointing to the need for further research concerning
nontraditional student experience in higher education.
Others (Haselgrove, 1994; Kasworm, 1990; Marshall & Nicolson, 1991; Richardson & King, 1998) agree that
“discussions about the role of adult students in higher education tend to stress their supposed needs rather than
the potential benefits they can bring” (Richardson & King, 1998, p. 66) and suggest that it is common for adults
to be considered, a priori, deficient in several areas: study skills, time management skills, intellectual ability,
and potential for academic performance (Richardson & King, 1998, p. 70). Kasworm (1990) labels these
disparaging views of the capabilities of adult learners as “… an insidious negative mythology regarding adults
students‟ impact on the campus culture” that is akin to perceiving adult students “… as retreads in a form of
270 Minichiello
salvage operation” (p. 156). Kasworm (1990) suggests that faculty and administrators often act out of the “fear
that the inclusion of adult students will lower standards and create a new set of problems demanding special
resources from the institution” (p. 156). Finally, Kasworm (1990) warns that those without prior experience
working with adult students are most likely to operate from these negative biases (p.157).
Perceptions of Nontraditional Undergraduates in STEM
There is evidence to suggest that nontraditional students in STEM may encounter pronounced bias due to
cultural norms existing within STEM disciplines. Conefrey (2001), for example, described widely held cultural
beliefs that work to maintain an insider/outsider binary within science-related fields. Of the twelve “myths” she
exposed, the following portend negative outcomes for nontraditional students in STEM: (a) science is gender-
neutral and value-free, (b) science is a meritocracy, (c) curricular and pedagogical change in science is
unnecessary, (d) challenge and competition is necessary for science, (e) failure is the individual‟s fault, and (f)
each scientist‟s classroom is his castle (Conefrey, 2001, pp. 174-178, 182). Describing so-called “second tier
students” in science, Tobias (as cited in Foor, Walden, & Trytten, 2007) writes,
Unless they are unusually self-motivated, extraordinarily self-confident, virtually teacher- and
curriculum-proof, indifferent to material outcomes, single-minded and single-track, in short unless
they are younger versions of the science community itself, many otherwise intelligent, curious, and
ambitious young people have every reason to conclude that there is no place for them in science [47].
(p. 112)
Foor et al. (2007) highlight the lack of opportunities within engineering education for students whose life
experiences differ from the “mythical norm of engineering … that of the high-achieving, elite, white male” (p.
103). They recommend that engineering faculty (a) be wary of “„us‟ and „them‟ classroom dynamics” that form
due to differences in capital possessed by students of varying backgrounds, (b) recognize that students who earn
average grades often become skillful engineers, and (c) be more accessible for working students who are unable
to attend office hours (p. 113).
Significance of this Study to the Literature
Within the United States, nontraditional students are a promising source of future engineers. While data indicate
that this population is increasing across postsecondary institutions generally, nontraditional student
achievement, particularly in STEM, continues to lag behind that of traditional students. Because current work
focuses on the examination of nontraditional enrollment and retention trends using quantitative research
approaches, nontraditional student experience in engineering remains largely unexplored. This narrative inquiry
project adds to the literature by employing counter story to provide a deeper understanding of how
nontraditional engineering students strive for and achieve success, despite facing social and institutional bias
related to their ways their socio-economic backgrounds and life situations differ from engineering cultural
norms.
Theoretical Framework
This study is framed theoretically by foundational work related to deficit thinking (Irizarry, 2009; Valencia,
1997) and Feminist Theory counter story (Nelson, 1995).
Deficit Thinking
Deficit thinking occurs when negative perceptions of the “ability, aspirations, and work ethic of systematically
marginalized peoples,” particularly those who are socio-economically disadvantaged and/or belong to
ethnic/racial minorities, are used to account for their disproportionately low academic achievement (Irizarry,
2009). Within K-12 education, deficit thinking is harshly criticized for its effect of relieving schools of their
“responsibilities to educate all students” by assigning root cause of failure to individual students and/or to their
communities and families (Irizarry, 2009). Deficit thinking, therefore, is not only psychologically harmful to
diverse students, but also works to limit their academic achievement indirectly through institutional policy and
school curricula.
Nelson (1995) tells us that persistent expression/embodiment of negative attitudes, stereotypes, and bias by
those in authority can easily become “dominant stor[ies]” of a faceless “Other” (p. 34). If dominant stories are
(re)-told sufficiently and well, their (re)-telling can ingrain conscious and subconscious thought patterns among
education stakeholders: administrators, faculty, and even students themselves. In this way deficit thinking—and
the dominant stories it engenders—is able to invalidate diverse students‟ individual academic abilities and
commitments to learning and to negate the relevance and value of the nonacademic knowledge and skills they
bring to their education (Haselgrove, 1994).
Counter Storying
Nelson (1995) reminds us “anyone can tell a story” (p. 34). Polkinghorne (1988) explains how the temporal,
chronological organization of narrative and storytelling is necessary for understanding human activity and
recognizing how our actions affect our goals (p. 18). In suggesting that “the stories we tell ourselves and the
stories others tell about us, [are] … in essence our reality … who we are and who we are becoming” (Schaefer,
Lessard, & Lewis, 2017, Theoretical Frame section, para. 1), Schaefer et al. (2017) propose a unique and
powerful link between narrative story-telling, experience, and being.
Nelson (1995) continues by saying that “just as anyone can tell a story,” “anyone can tell a counter story” (p.
34). While counter stories are written and told just as other stories are, they have particular characteristics that
allow them to “counter” or dispute specific elements of dominant narratives: Counter stories are authored from
the standpoint of the marginalized “Other,” are “capable of attending to difference,” and are “ told for the
specific purpose of undermining a dominant story” (Nelson, 1995, p. 34). The true purpose of counter stories,
then, is not wholly to undo the dominant story, but rather to “set into equilibrium certain details or moral
interpretations the dominant story ignores or underplays, thus allowing … for dissent from the interpretation and
conclusions the dominant story invites” (Nelson, 1995, p. 34). Like all stories, counter stories offer rallying
points for group action and may lead to lasting impact. Within educational settings, counter stories allow
marginalized students to “… enter the [dominant] story sideways, … correcting for the biases and distortions of
the dominant interpreters” and opening up “access” to the valuable “goods” of school communities to all
students (Nelson, 1995, p. 36).
Methodology
This study used the qualitative methodology known as narrative inquiry (Clandinin, 2013; Clandinin &
Connelly, 2000) to examine and map the lived stories of undergraduates who pursued engineering degrees as
nontraditional students. Dewey (1938, p. 23) asserts that the “connection … between the achievements of the
past and the issues of the present” exists “within experience.” Narrative inquiry, which “derives from the
Deweyan view of experience (particularly situation, continuity, and interaction),” is a relational, transactional
research methodology framed by 3-dimensional narrative-inquiry space consisting of the personal/social
(interaction), time (continuity), and place (situation) (Clandinin & Connelly, 2000, pp. 49-50). While narrative
inquiry involves storytelling, it is more; “Narrative inquiry is a way of understanding experience. It is a
collaboration between researcher and participants, over time, in a place or series of places, and in social
interaction with milieus” (Clandinin & Connelly, 2000, p. 20). Thus, narrative inquiry is often described as “…
people in relation studying people in relation” (Clandinin, 2013, p. 23). The strength of narrative inquiry lies in
its ability to develop detailed understanding of human action through in-depth and personal examination of
experience.
Method
Participants
Volunteers were recruited via email from a list of students who had graduated from the transfer program or
completed (at a minimum) the Engineering Statics course taught within the program. Because Engineering
Statics is (commonly) the first engineering course that engineering students encounter, Statics completion was
considered a reasonable requirement for participation because it ensured that participants had experienced actual
engineering instruction. Volunteers who responded to the initial email were provided with a link to an online
survey used to screen for nontraditional undergraduate characteristics (Horn, 1996). To insure sufficient but not
272 Minichiello
extreme intensity (Patton, 2002) of nontraditional student experience, only those volunteers who responded as
having three to six characteristics were accepted as participants.
All participants were white and one was female. Programmatic data show that the group from which participants
were recruited was characterized as less than 4% female and fewer than 22% nonwhite. Among those students
who had graduated from the program by the time recruitment began, there were no females and fewer than 13%
were nonwhite. These data suggest that participants represented the larger student body on gender and racial
characteristics. The nontraditional characteristics most frequently shared among the participants included having
breaks of one year or more between high school and college, working full-time and/or enrolling part-time at
some point during the program, being financially responsible for themselves and any dependents, and having
dependents other than a spouse. Only one participant had earned a GED in place of a high school diploma and
one other was a single parent.
Data Collection
As the researcher, I met individually with each participant, either in person or virtually, twice over the initial 7-
month period. During this period, I conducted formal data collection and analysis by jointly constructing
longitudinal narratives with each participant. For most participants, these research conversations strengthened
existing relationships that had naturally developed during the transfer program as I taught and mentored students
there. Our meeting arrangements were mutually agreed upon and I led our conversations using an in-depth,
semi-structured approach (Hesse-Biber, 2014). Conversations were digitally recorded (audio only), transcribed,
and reviewed for accuracy before being used in analysis.
During each initial meeting, I posed discussion prompts modeled after the interview protocol used by Pawley
(2013) to conduct narrative research with underrepresented minorities in engineering education. The prompts
worked to fulfill the project goal of eliciting participant reflection and storytelling and included questions such
as: (a) Where are you in your life now? (b) How would you describe your background? (c) What were your
goals? (d) What obstacles did you face? and (e) What successes did you achieve? At the end of each initial
meeting, I asked participants to draw/write personal journey maps (Nyquist et al., 1999) to prompt deeper
reflection and elicit further detail about their personal experiences during our second meeting. Shortly after each
meeting (within the next day), I completed a post-interview reflective memo, based on Miles and Huberman‟s
(1994) contact summary form, as proposed by (Pawley, 2013) to capture my initial impressions of the stories
shared, summarize the information gathered, and reflect on the information that would be useful to get during
follow on conversations.
After the conclusion of each initial meeting, I developed unique questions to pose to each participant during
subsequent meetings. These questions were based upon analysis and chronological ordering of participants‟
personal stories shared during the initial meeting. During subsequent meetings, each participant and I jointly
reviewed the written transcript from the first conversation, the participant‟s journey map, and the initial analysis
of their stories. After the completion of all formal meetings, participants were provided with a $30 gift card to
either the university bookstore or big box store of their choosing.
Once these formal research conversations were complete, I began the process of co-constructing a longitudinal,
experiential narrative (interim research text) with each participant. During this period, I shared electronic
versions of narratives with each participant. I solicited written feedback via email periodically until each
narrative converged to joint agreement. As I worked to produce final research texts (i.e. journal articles) from
interim research texts, I was in contact participants to request feedback on final research text drafts and updates,
if any, on their experiences since our earlier conversations.
The field texts (data) for this study include screening survey results, transcribed conversations, participant
journey maps, researcher written reflections, written participant feedback on drafts of the longitudinal narratives
(interim research texts), and written participant feedback on final research texts and updates about their
experiences since our initial meetings. This study was conducted with approval from our university‟s
institutional review board; all participants are required to give informed written consent prior to data collection.
Data Analysis
Narrative analysis (Polkinghorne, 1995) was conducted by co-developing longitudinal narrative accounts,
similar to life stories (Atkinson, 2007), with each participant to communicate their experiences of engineering
education. Polkinghorne (1995) explains that, in performing narrative analysis, “the researcher‟s task is to
configure the data elements into a story that unites and gives meaning to the data as contributors to a goal or
purpose” (p. 15). These narrative accounts served as interim research texts; elements from the collected
narrative accounts are also being developed into final research texts (Clandinin, 2013). The work of constructing
and negotiating narrative accounts with participants helped to ensure that analysis and interpretation was
accomplished collaboratively while honoring the ethical considerations and relational nature of narrative inquiry
methodology. Trustworthiness of the accounts was supported by considering each participant (i.e., “storyteller”)
as “… the expert and the authority on his or her own life, thus having final say in what gets told” (Atkinson,
2007, p. 239). Participants chose how to represent themselves within their stories, using either first name
pseudonyms or their actual first names.
Findings
Findings from this study are presented as nontraditional undergraduate narratives of encountering, embodying,
and countering deficit thinking in engineering education. Due to the length of the detailed individual narratives,
this work was limited in sharing from only two of the 14 participant stories developed during the project: Clair,
a 22-year old, moderately nontraditional (three characteristics) white male; and Joe, a 29-year, old highly
nontraditional (five characteristics) white male. Because deficit thinking emerged across the range of participant
stories, the choice of whose stories to present involved several considerations. First, as the data generally broke
down as younger, moderately nontraditional students and older, highly traditionally students, one story from
each category (i.e., one younger/moderately nontraditional story and one older/highly nontraditional story) was
selected for presentation. Second, to present findings centered within potential extremes of experience and more
apt to resonate with most readers, participants whose time spent in the transfer program (i.e., time to associate‟s
degree completion) was centered within, and not at the limits of, the range of associate‟s degree completion
times represented within each category were selected for presentation. One unintentional benefit of presenting
the stories of Clair and Joe, both white males and participating members of the dominant local culture, was to
highlight how being nontraditional, in and of itself, may be enough to catalyze deficit thinking within
engineering education.
Clair
Now a mechanical engineer/structural analyst for a large aerospace engineering firm, Clair thought back to the
time when he and his wife, as newlyweds, enrolled as undergraduates at the regional campus, he in the
engineering transfer program and she in elementary education. Shaking his head with a half-grin, half-grimace,
he remarked how “Those were the longest days of my life. We both look back and think that was a miserable
point in our lives.”
The irony of “those” days is not lost on Clair. Finally getting about the business of becoming an engineer—the
thing he had wanted to do since he was a young boy building his inventions in his parents‟ backyard—Clair had
been “scared to death” for the entire first year he was in the transfer program. He explained how he enrolled in
the program “not knowing how I was going to accomplish my goal but knowing I had to take a step forward.”
Clair‟s goal, ever since he could remember, was one of becoming an engineer. He talked fondly about his
childhood—especially the hours and hours he spent building and creating in the wide-open space behind his
childhood home, nestled in a corner of the Texas ranch where his family lived and his dad worked as an animal
physiologist. During these hours in the backyard, Clair—in effect—was figuring out who he wanted to become.
Clair said:
Even when I was really, really young, I‟d be out back building my ideas and coming up with how to
do things differently. And I remember that I asked my mom, “Who does this type of stuff in the real
world? … I didn't have the slightest clue. … I asked my mom and she said, “You know, engineers do
that type of thing.” And so after that I said, “ I want to be an engineer.” And I told everybody, for the
rest of my life, that I wanted to be an engineer.
274 Minichiello
Today, Clair laughingly admits that he isn‟t sure if he stuck with his engineering story all those years “… out of
stubbornness or because that‟s what I had told everybody.”
Embodying Deficit Thinking
Clair describes his high school self as a “terrible student.” In fact, he was the kind of student who would “… sit
down one night and figure it out—right before the test—and then ace the test” without really trying. Clair
“didn‟t try” in his most of his classes because none of his high school subjects interested him, except for the
automotive class. The highlight of his high school days, in fact, came at a national automotive competition as he
was awarded a scholarship to a 2-year Texas college to earn an associate‟s degree in automotive technologies.
Clair figured that, once he earned his associate‟s degree and gained marketable skills for a “decent paying job,”
he would go to a 4-year university—“somewhere”—to study engineering. Eventually, he hoped to become an
automotive engineer and maybe even start his own company to design and fabricate custom parts. Since Clair‟s
parents had instilled in him an admirable and inexorable resistance to accepting loans of any kind—without
having means to support him financially—that initial 2-year scholarship was a critical part of Clair‟s plan.
But Clair never used that scholarship. Much happened in the four-plus years that passed between Clair‟s high
school graduation and entry into college. After graduating, Clair temporarily deferred his scholarship, worked
any part-time job he could get, and waited for a mission call from his church. The mission call came 14 months
later and Clair spent the next two years abroad in St. Petersburg, Russia. Clair remembered how, “The whole
time I was on my mission I thought, „I‟m going to use that scholarship down in Texas.‟ The whole time, that is,
until my mom got cancer again.” Upon returning stateside, Clair made a difficult decision to give up his
scholarship and spend time, whatever it took, helping her get well.
Thankfully, his mother fully recovered from her second breast cancer diagnosis. During the time he spent at
home, Clair had the good fortune of finding a job he “loved, “working with his hands” while restoring old
houses, as well as meeting, becoming engaged to, and marrying his wife. Soon after the wedding, the newlywed
pair agreed that it was finally time for both of them go to college. So Clair, at 22 years old, and his wife packed
up and moved to a different city in another state so that Clair could enroll in the engineering transfer program.
He chose the program because of the regional reputation of the university in engineering, as well as the fact that
the evening programming would allow both he and his wife to work while attending college.
Clair described how they walked into the regional campus advising office together to enroll, just after arriving in
town, having “no idea what they were doing.” Clair explained,
At the time I started, it was four and a half years I hadn‟t been in school—I hadn‟t done math in four
and a half years! I‟m embarrassed to say this—I couldn‟t test out of trigonometry because I couldn‟t
remember it. All these kids going into the engineering program skipped out of calculus and I was
starting in Algebra. I was scared because I wasn‟t a very good student in high school. I just didn‟t
think I was smart enough to do it.
I met Clair at about this time. He was, in fact, the first student I met in my new position as an engineering
instructor in the transfer program. I remember Clair knocking on my office door to ask for help with some math
problems he was studying in advance of taking the placement exam. He was bright and energetic, but seemed a
bit nervous and apologetic. I remember being inwardly surprised he was studying algebra, and not calculus,
problems.
While Clair considered both his lack of effort in high school and his break between high school and college as
major obstacles to becoming an engineer, it was also clear that he doubted his own personal abilities and
intelligence, particularly with respect to his skills in math. He explained how engineering was somehow
different than other subjects and, “even though I studied every day on my mission because I had to learn
Russian, it didn‟t prepare me for studying engineering.” As Clair wondered whether he was “smart enough” to
do engineering, I imagined the courage he summoned walking into that advising office the day he came to town.
Encountering Deficit Thinking
Despite the personal stories of deficit that he told about himself early on, Clair began to craft a new story for
himself as he worked his way through the transfer program. He and his wife each found part-time work situated
at the regional campus. Flexible work schedules enabled them to take day or evening classes as needed. By
reducing their schoolwork commute time down to zero, their jobs enabled them to spend maximum time and
effort on their studies.
As he began taking classes, Clair was somewhat surprised that “all of my classes interested me except for,
maybe, psychology. I just needed some interest and I started doing really well.” Clair received all A‟s that first
semester. “Oh, I can do this,” Clair began to tell himself. “I can—if I just sit down and work hard, I can get
good grades.” Clair talked about how:
One of my earliest goals was to get good enough grades. Once I realized I could get good
grades, then my goal became to get really good grades so that I could pay for school
[through scholarships], because I had no other way [besides working] to pay for school.
Clair became self-motivated—just the way he remembered being in his high school automotive class. He started
asking lots of questions in class—even, one day, to the point of annoying his calculus instructor. Clair told me:
“I understood ninety percent of everything. But I wanted the last ten percent and I was working my butt off to
get it.” He started to experience a synergy by being—studying and working—on campus. Faculty and staff
members “saw me working hard to do this and they appreciated it. Their appreciation motivated me to work
even more.”
Clair experienced considerable success in the transfer program. Just shy of his 25th birthday, Clair graduated as
Valedictorian of his transfer program graduating class. Now, he had a new story to tell:
When I left the … program, I felt like it was a success because when I was done I felt like if
somebody asked me how to do something, I could do it. I remember loving thermodynamics
and thinking, “I could dominate in the thermodynamics world. This is awesome.” And
granted, that was a pretty basic class. As you get farther into it, you learn that there is way
more to this. But that was one thing that I remember always feeling proud of—that I actually
knew what I was doing.
Clair‟s academic success earned him two scholarships which together completely funded his tuition for his final
two years of engineering study. Riding the high of his successes, Clair took newfound confidence with him to
the main campus mechanical engineering program.
Somehow, though, things were different for Clair after he started classes on main campus. He felt that
“atmosphere changed”—it became “… as if nobody cares about you [and] you‟re just a number.” One
recollection of a classroom experience really bothered him. One day, an engineering instructor told the class that
this was their “weeding out period, and if you have anything else you want to do right now, put it on the back
burner.” Clair remembered thinking how “rude” the statement and how uncaring that instructor were.
Clair‟s feelings of isolation went deeper still. Clair talked about how he:
… told some of the [engineering] instructors that I came from the [transfer] program and they kind of
got this look on their face—like depressed looking. And they said something like, “Oh, okay, well,
yeah, good luck over here.”
These unsettling experiences with engineering faculty began to work—negatively—on Clair‟s motivation. After
those experiences, he stopped trying so hard. Clair said:
My motivation decreased in a major way.… And by that, I mean the motivation to excel. The
motivation to become an engineer was still there. The motivation to be at the top of my class left. I
didn‟t try extra hard. I just did what I needed to and I didn‟t go after that last ten percent like I did in
[the transfer program]. I settled for understanding enough to get a good grade.... And then I moved
on.
His classroom behavior changed, too. Clair found himself too embarrassed to ask questions in the large
engineering classes that were common on main campus. Sitting in class with a question on his mind, Clair often
thought to himself: “I‟m sure so many other kids understand it. I‟m just going to go home and try to figure it
out.” In this new environment, Clair‟s story of becoming changed, as he found that he “had to figure out how to
learn without asking questions in class.”
276 Minichiello
Countering Deficit Thinking
For Clair, one bright spot during the professional program came during his senior capstone design course. Clair
became the leader of a team of mechanical engineering seniors who, as part of the Crossing the Gap competition
for university engineering students run by the Air Force Research Laboratory (AFRL), was tasked with
designing and building a collapsible structure weighing approximately twenty pounds, measuring under four and
one-half cubic feet when collapsed, that was quickly deployable and could hold 350 pounds. The structure was
intended for use by “U.S. Special Forces agents [who] need to scale a wall, traverse a canal, or cross between
rooftops” (Hwang, December, 2013).
For Clair, this senior design project was the “ideal” engineering learning experience because he was given
“freedom to create something that came from an idea that I had.” He said:
When they came to us … and said, “Hey, here‟s a list of requirements figure out how to accomplish
it,” that was really fun to me—I enjoyed that. I like the fact that somebody just said, “Here are the
requirements. Figure out how to accomplish it kind of creatively and inventively so.”
The design and prototype developed by Clair‟s team turned out to be wildly successful, winning the AFRL
competition and being featured in the December 2013 issue of Popular Science magazine (Hwang, December,
2013). Now, Clair told a new story of passion and confidence as he discussed how his team worked together on
the project:
Even though senior design was super difficult and I felt like beating my head against the wall a lot of
times, all my initial ideas were eventually polished into something that was better than I had come up
with because of the group discussions. No one part of that design was any one person‟s idea
exclusively. The concept might have been one person‟s idea, but it all got improved because of
everybody‟s work.
Amid disappointing and demotivating interactions with engineering faculty during the professional program, the
senior design AFRL competition provided Clair a rewarding, memorable experience in which he took an
opportunity to lead his peers in an area in which he naturally excelled (engineering design). While Clair‟s senior
design experience did little to improve or reshape his relationships with the engineering faculty, it did provide
him an alternative environment within which to find renewed motivation to excel.
Clair completed his bachelor‟s degree in the spring of 2013, four and one-half years after he began. He quickly
landed an aerospace engineering job that was “better than most people dream of” and put him right where he
wanted to be—except, perhaps, for the industry. Still, Clair keeps his entrepreneurial dream alive by working on
his inventions nearly every day.
Thinking back to our last conversation, I remember that Clair told me how he was “a little bit bitter” about the
way he was treated in the professional engineering program. I admit that I was startled when he confessed he
hadn't attended his bachelor‟s degree graduation ceremony. In response to my overt reaction, Clair described
how he holds no lasting sense of enthusiasm or loyalty for his alma mater, saying,
I don‟t feel like they did me any favors ... I mean, you transfer [to the four-year engineering program]
and they‟re anticipating you to live, breathe, and do school. I mean, that is what they expect of you.
And if you‟re nontraditional, you have to figure it out, you know? And maybe if I would have just
been like I‟ll go into debt, whatever, and I‟m just kind of out of [high] school … that might have been
different.
Frustrated with faculty‟s seeming distrust of his abilities and lack of accommodation for his needs as a
nontraditional student, Clair could only imagine how his story of becoming an engineer would have been
different had he been different, that is, had he been more like his traditional undergraduate peers. Alternatively,
lingering in the aftermath of Clair‟s story, I am left wondering how Clair‟s experiences—as well as those of
students like him—could be different if institutions themselves were asked to change. What could nontraditional
student outcomes in engineering look like if faculty became better able recognize professional engineer potential
among nontraditional undergraduates, and if engineering programs became better able to meet nontraditional
undergraduate needs and constraints?
Joe
Joe was 29 and married when he enrolled in the engineering transfer program in the fall of 2009. He and his
wife welcomed their first child, a son, in 2010 when Joe was in his second year of the program. During these
years, Joe supported himself and his family working as a full-time heating, ventilation, and air conditioning
(HVAC) technician at a national aerospace engineering firm located roughly 30 miles—about a 45-minute
drive—from his home. He had started there as a part-time worker in 2007, and within four short months had
been hired full-time. Joe‟s decision to enroll in the transfer program and earn his engineering bachelor‟s degree
while supporting his growing family came after spending over 10 years pursuing other serious endeavors after
high school, including serving a two-year mission for his church in the eastern United States, working as a
HVAC technician for a small local company, working as an independent mechanical contractor, and seriously
considering training for a career as a commercial airline pilot, either by joining the U.S. Air Force or enrolling
in private flight school.
Encountering Deficit Thinking
When Joe enrolled in the transfer program, it had been ten years since he had graduated high school. In the fall
following his high school graduation, Joe had enrolled for a single semester in the university‟s engineering
program to fill the time as he waited for his mission call. At that time, Joe didn‟t make much progress toward a
degree but blamed his lackluster performance on just “not being ready for college” yet. While he considered
himself to be a good student who knew how to study, Joe was certain that he had forgotten most of what he had
ever learned about college-level mathematics by the time he enrolled in the transfer program.
To make up for what he felt was lacking in his preparation as he entered the transfer program, Joe and his wife
decided that it would be best for Joe to restart his mathematics education at the very beginning with Basic Math.
While this decision added a few years and several courses to his program of study, Joe never regretted this
decision in terms of his academic preparation; Joe found his math courses to offer some of the biggest yet most
rewarding challenges he faced in the program. Joe did, however, regret some of the personal interactions he
experienced as he entered into engineering math courses. One experience in calculus particularly bothered him:
The first day of class the instructor told us, “All right, this is how many points you need to get a C
minus. This is how many points you need to pass the class”—and that was it and then he moved on. I
was like, “You‟re not going talk about what it takes to get a B or an A?” But that was the
expectation—we were all just trying to pass and could not really do well in the course. It made the
biggest difference when the teacher just expected you to fail. That‟s what I felt like in some of those
classes.
After making a substantial commitment of additional time and money to help insure his mathematics
preparedness and future success in engineering, Joe found it not only surprising and confusing, but also
frustrating and demeaning, being met by engineering calculus instructors who assumed that he and his
nontraditional classmates had little interest in or aptitude for learning calculus.
Embodying Deficit Thinking
By all accounts, Joe had been a solid performer throughout the transfer program. As his thermodynamics
instructor, I had noticed his keen interest in the course material. Joe told me how thought it was “cool” that,
after taking that course, he could “run calculations for HVAC systems as well as fix them.” He was proud how
his strong efforts in the transfer program enabled him to work closely with and for the facilities engineers at his
place of employment, noting how the other technicians employed there could not do so. Joe talked about how
“Taking thermodynamics has really helped me understand refrigeration at a level where I just get it and can
readily explain things about it to other technicians and engineers. I couldn‟t do this before going back to
school.”
Sometime after Joe had completed the thermodynamics course, his manager surprised him with an offer to
participate in an engineering internship at the company. After some serious deliberation, Joe accepted the offer,
“even though I was very busy with school, because they said „You'll be doing HVAC, you know, BTUs in and
BTUs out, that kind of thing.‟”
278 Minichiello
[But] it was all sorts of different things. I was like “Holy crap, I don't know how to do any of this
stuff!” I was really nervous and felt very inadequate at first. I had to learn all of the other computer
programs that … the engineers deal with on a regular basis that I‟d never used before…. That was the
first month. I was like “Why am I doing this? I hate this.” But later on, I kind of enjoyed it.
In the end, Joe was grateful for the experience; he decided that it was a clear sign that his management was on
board with his educational pursuits and perhaps even was interested in the prospect of him moving into an
engineering position at the company in the future.
Joe graduated from the transfer program in the spring of 2015 after completing the entire mathematics
sequence—from basic mathematics through differential equations and linear algebra—as well as the first two-
years of engineering curriculum within six years. To Joe, graduation was an important personal milestone—not
because of any financial benefit he hoped to receive from earning his associate‟s degree, but rather because
graduation from the transfer program marked a midpoint in his studies:
[Graduation from the transfer program was] just over halfway for me because I had to retake all of
those math courses. At the beginning, I looked at all those math classes and I was like, man, I'm never
going to get through all this. I hated the beginning of every semester but I felt so good at the end of
every semester, being that much closer. I had papers that list all the classes I had to take. When I …
cross[ed] off one … it felt so good.
Joe transferred to the main university campus the next fall and began the rest of his engineering studies
alongside more traditional student peers. At that point, he estimated that it would take him an additional 3-4
years to complete his bachelor‟s degree in mechanical engineering.
Early in the spring of 2016, I emailed Joe and asked him for his feedback on an interim research text I was
preparing. Soon, we began talking about his current experiences on main campus. I knew from our previous
conversations that he had been very concerned about the transfer—how it would all play out for him as he tried
to juggle day-time classes and a full-time work schedule yet still give time and effort to his roles as husband and
father. He had talked about the most difficult days in the transfer program “… where I didn‟t see my son at all,”
acknowledging how he and his wife were jointly committed to enduring these sorts of hardships “for the well-
being of our family in the future.” By the time he was ready to transfer to the main campus, Joe had grown very
anxious about what this next phase would require of him and his family. When I asked him about his first
semester on main campus, Joe told me this story:
This semester was really tough attempting to adjust to daytime class and nighttime work. My average
workday was to leave for work at 1:00 pm and then return from work around 11:30 pm. On school
days (MWF last semester) … [my round-trip commute] was just over 2 hours and about 100 miles a
day. Several times I planned on leaving work a little early to do homework right after work and then
things would get really busy at work so I would "lose" time I had budgeted to complete some
assignments.
Joe went on to compare his experiences within the transfer and main campus programs:
I think the biggest change for me … [was] the additional commute [time] and [the] time required to
get around campus. Another thing I felt was that, as class sizes grow [on the main campus], the
instructors seem to have less time to answer questions and seem more concerned about getting
through material. Getting help takes a lot more effort/time because it has to be face to face, so
because of my schedule that was impractical.
Joe had taken a lighter than average load for engineering student, which consisted of “only three courses,” that
spring. He told me how he received a “B” in one course and barely passed the other two. I could tell, from his
voice, how disappointed he was. He told me how he was also facing administrative repercussions—being barred
from enrolling in some 3rd year engineering courses—because his semester grade point was below a
predetermined limit.
These repercussions would undoubtedly increase his time to degree. Obviously frustrated, Joe remarked, “I wish
I would have been aware of some of these „rules‟ at the beginning of the semester, because now it almost seems
as though they are being made-up as I go.” Joe ended our conversation saying,
I know I have put myself in this situation, as far as being a nontraditional student, but it sure is tough
trying to hang in there with the traditional students. I keep feeling like the instructors/advisors must
think I am a terrible student and that engineering may not necessarily be what I should be pursuing. If
I had a chance to explain my situation, that school is really my third (major) priority, maybe they
would better understand my seemingly lackadaisical effort, when sometimes that effort is truly all I
can give.
Joe‟s very candid and personal response troubled me. In his words, I heard him questioning his own abilities and
internalizing negative perceptions (e.g., “terrible student,” “lackadaisical”) about his student self. At this
moment, his words stood in stark contrast to the expressions of pride and confidence he displayed as he
graduated from the transfer program.
I thought about how, after just one semester taking classes in the engineering education environment of main
campus, Joe had begun to question himself—whether he was good enough— and to perceive personal
deficiencies within himself that came as a result his multiple identities as husband, father, employee, and
student. Within his words, I heard an apology for being a nontraditional student, as if it was a mistake he had
made or a misfortune he chose to endure. As the strength of his concerns led him to question his emerging
identity as an engineer, I was left wondering how Joe would continue to fare as a minority within the dominant
engineering culture.
Countering Deficit Thinking
I continued to converse with Joe, intermittently, via email. A little more than a year later, I emailed and asked
him to review a draft research text I was preparing. Admittedly, I was nervous each time we communicated—I
carried around with me lingering doubt that he would, one day, complete his degree. As I read Joe‟s message
that day, I was struck hard by my own lack of faith and negative perceptions. “I have transferred to another
university as a distance [engineering] student,” Joe wrote. I sat there, staring at my computer, trying to imagine
the circumstances that had driven Joe away from the physical university located approximately 30 miles from
his home to a distance education program administered from approximately 1,800 miles away. Joe went on to
write,
Two semesters at the ... [main] campus trying to attend the classes I needed whenever they were
offered and having to adjust my work schedule accordingly was a little too painful…. I had looked
into online [engineering bachelor‟s degree] programs in the past but could not find any that were
accredited. However, after that first semester [on main campus] … I started looking again and found
two.
Joe found two accredited mechanical engineering programs online and was accepted as a transfer student at both
of them. The transfer process—to another university system in another state—however, had been hard. “It
literally took dozens of emails and phone calls over several weeks to sort through how all the credits …
would/could be applied.” In the end, Joe chose the program that required the fewest on-campus visits. Joe told
me:
When the time comes, I will have two classes that will require travel … two or three times a semester.
While that may sound ridiculous to most people, it is far more valuable to me to have a consistent
schedule at work, than the cost of several trips … to attend labs on the weekend. The main reason is I
have a pretty good job now, and I just don‟t want to deal with the risks of having to change my work
schedule each semester.
As I read and responded to Joe‟s message, I was overcome with relief at having heard Joe‟s new story of
becoming. I am excited by Joe‟s possibilities within the profession as well as humbled by the drive, dedication,
and sacrifice he has shown in discovering his own pathway to becoming an engineer. While Joe‟s narrative is
not yet complete, he is once again optimistic and looking ahead to a future as an engineer.
Discussion
Clair and Joe‟s experiential narratives of encountering, embodying, and countering deficit thinking serve to
demonstrate how students who are older, enrolled part-time, and/or working full-time working—even those with
280 Minichiello
familial responsibilities—can be successful in undergraduate engineering education as well as in the profession

of engineering. Their narratives serve to counter common deficit storylines that are (re)-told about nontraditional
students in higher education. As depicted in the narratives, Clair and Joe faced substantial challenges as they
negotiated a complex mix of personal, social, and institutional tensions while pursuing their engineering
degrees. Yet, instead of merely identifying the barriers they faced, their stories describe and examine how each
persevered and succeeded in engineering despite frequently bumping against deficit perspectives.
Early Undergraduate Experiences in Mathematics
Clair and Joe each experienced perceptions of personal deficiencies as incoming engineering students related to
what they thought about their innate mathematical abilities and/or the current state of their mathematical skills.
Clair admitted to being embarrassed about testing into algebra while his student peers, “these kids,” were all
starting out ahead of him in calculus. Being well behind his peers in math made Clair question his ability to
become an engineer. These doubts were intensified because he had not gained sufficient confidence in math in
high school. Joe, who had felt reasonably comfortable with math in high school, more readily accepted
placement within Basic Math. Yet, despite rational acceptance of the need to hone his math skills, Joe was
insulted by the words of his mathematics instructor, whom he perceived as jumping to the foregone conclusion
that he and his peers did not have potential or desire to excel in math.
Because being calculus-ready is considered as the point of entry for engineering programs (National Academies
of Sciences Engineering and Medicine, 2016), it is likely that many nontraditional engineering students face
similar challenges related to mathematics instruction. Clair and Joe‟s stories reveal to us that nontraditional
students may be inclined to engage in personal deficit thinking about their intellectual abilities to do
engineering, based solely upon the current state of their knowledge and skills in mathematics. While choosing to
(or being required to) take remedial mathematics courses is often a good course of action for nontraditional
students academically speaking, it can exacerbate personal (e.g., financial, familial, employment) challenges
faced by nontraditional students by necessarily increasing the time and expense associated with earning the
degree. Joe, for example, took six years to complete the two-year transfer program. Joe‟s completion time was
due in large part to his placement within Basic Math. One year of calculus is a standard prerequisite for even the
most fundamental engineering courses. As Clair and Joe‟s stories show, remedial math placement may increase
the social challenges that nontraditional students face if peers and/or instructors view the need for remedial
mathematics training from deficit perspectives.
Thus, it appears that mathematics instruction plays a key role in the early experiences of nontraditional students
in engineering. This seems true not only because mathematics courses are among the first that engineering
undergraduates encounter, but also because nontraditional students may be particularly prone to personal deficit
thinking about their abilities in mathematics. Moreover, early course failures in mathematics may be especially
difficult for nontraditional students to rebound from, not only due to administrative rules concerning allowable
course repeats and the inherent linearity of mathematics/engineering curricula, but also due to the need for
students to overcome strong personal perceptions of math inadequacy reinforced through course failures. It is
important for institutions desiring to promote nontraditional student participation and retention in engineering to
make active, engaged, and supportive instruction, especially within mathematics, a top priority.
Engineering Faculty Interaction and Instruction
As Clair and Joe transferred to the 4-year university campus and entered the professional program, interaction
with engineering faculty and engineering course instruction became obstacles to their success. Clair and Joe
each cited the large class sizes and distant, seemingly uncaring faculty as factors that negatively affected their
motivation for learning within that transitional year. After rising to a position of relative confidence and
excitement about his ability to learn engineering within the transfer program, Clair very quickly began to feel
isolated and ignored as he transferred to the professional program. These changes rapidly manifested in Clair‟s
classroom actions; Clair went from being an avid classroom participant who worked to understand “that last ten
percent” to being too embarrassed to ask questions, sure that “so many other kids” knew the answers to the
questions he wanted to ask. Ultimately, Clair disengaged from classroom instruction to the point of just doing
what he needed to and moving on.
Joe expressed how large class sizes within the professional program reduced the quality of the instruction,
making it seem rushed and as if faculty were more interested in getting through the material than instructing. Joe
found it nearly impossible, given his carefully choreographed work-school-commute schedule, to get out-of-
class help. The standard face-to-face office hours support was completely “impractical” for his needs and
schedule; Joe could find no other available options. All in all, these stories help to highlight the importance of
continuing faculty interaction and student-centered instructional approaches for nontraditional students
throughout the engineering curriculum. It may prove particularly difficult for nontraditional students—
especially those who transfer from a two-year program to a four-year program—to preserve their confidence and
motivation. Moreover, it may be that the transfer (third) year is a particularly difficult juncture for nontraditional
transfer students and that engineering curriculum stakeholder should focus resources to bolster instruction and
student support during that year.
Resilience and Goal Orientation
While the narratives serve to reinforce the importance of instructor relationships and classroom instruction for
nontraditional student success in engineering, they also demonstrate the personal resilience, perseverance, and
ability of the nontraditional students to manage their actions based upon higher goals. Both students came into
the transfer program with well-defined and interdependent goals. For Clair, these goals included the concurrent
education of he and his spouse, his future career as an engineer—preferably within the discipline of automotive
engineering— and his dream of, one day, becoming a self-employed entrepreneur. For Joe, goals for his
education, for his current career as HVAC technician, and for his future career as an engineer—perhaps with his
current employer—combined to create an overarching purpose of providing a stable future for his family.
Unlike many traditional students “… who select engineering as a major without actually knowing what
engineering is” (National Academies of Sciences Engineering and Medicine, 2016, p. 26), Clair and Joe enrolled
with firm understandings of their educational/career/familial aspirations as well as personal convictions,
grounded in experience, that they wanted to be engineers. Yet, because their educational goals were not singular
and premier but interdependent, Clair and Joe were apt to shift their priorities in relationship to school, work, or
family longitudinally over time. These shifts did not reflect a lack of time management or study skills as perhaps
engineering faculty might perceive, but rather pointed to sophisticated long-range planning and a necessary
reordering of responsibilities as they balanced their everyday lives in the hopes of reaching their long-term
goals. In many ways, these stories reveal how nontraditional students often manage the same or similar
personal, familial, and career-related activities as do educators, administrators, and staff and, perhaps, point to
how nontraditional students could benefit from being treated by faculty and staff more as adult peers, from the
sense of life experience, than as students.
Limitations
The findings presented in this study represent the experiences of only two nontraditional undergraduates. While
their stories provide rich details of nontraditional student experience in engineering, they do not necessarily
represent the experiences of all nontraditional students enrolled in the engineering transfer program or the
broader group of nontraditional engineering undergraduates generally. Insights originating from the in-depth
examination of only a few stories of becoming in engineering—as told by students who embody characteristics
that serve as markers for deficiency and risk—are nonetheless valuable and worth (re)-telling. While the lack of
cultural, ethnic, and gender diversity among the stories presented reflects the cultural influences of the
geographical region, it also suggests that the stories presented represent but a glimpse of the broader landscape
of nontraditional student experience. In addition, my own background—first as a nontraditional engineering
graduate student 20 years earlier, as a practicing professional engineer, and as engineering instructor within the
transfer program—undoubtedly influenced my desire to examine nontraditional student experience as well as
my choice of theoretical frameworks to underpin this study. While the existing student-instructor relationships I
shared with many participants made it easier for me to negotiate entry into their personal stories, they also
required me to continuously manage my own preconceived interpretations of the stories they told. In retrospect,
I believe it was perhaps my unique background that allowed me to first perceive and later acknowledge, after
years of enculturation in engineering professional and academic culture, my own complicity in nontraditional
student bias and to choose to examine this phenomenon from a collaborative, student-centered perspective.
Conclusions and Scholarly Significance

While Clair‟s narrative of living and learning as a nontraditional undergraduate in engineering may be
considered complete, Joe‟s is still unfolding. As a former instructor in the transfer program, I was moved by the
282 Minichiello
stories that we, as “people in relation studying people in relation” (Clandinin, 2013, p. 23), experienced and
shared. The deficit perceptions described by Clair and Joe often echoed the isolation I experienced as I entered
engineering graduate school upon my military discharge. At other times, their stories spotlighted the cultural
perceptions and habits of mind to which, having spent many years immersed in engineering culture, I had
become accustomed and even, at times, helped to perpetuate.
By accessing these unique views into the personal, social, and institutional tensions experienced by
nontraditional transfer students, this narrative inquiry awakens new attention and sensitivity to the perceptions
and bias that faculty, staff, and administrators may bring to their roles in engineering education. These in-depth
insights, provided by nontraditional engineering students in the midst, present important implications for
improving nontraditional student experience in undergraduate engineering education. Tentative conclusions
from this study suggest that nontraditional students represent an important source of engineering human capital,
with strong potential for completing undergraduate engineering degrees and becoming productive practicing
engineers. Academic institutions should strive to reduce institutional tensions arising from rigid educational
structures and policies, as well as social tensions stemming from poor instructor/advisor-student relationships
and inadequate classroom engagement, since these may be among the most disruptive and difficult for
nontraditional students to negotiate. To promote nontraditional student participation and success in engineering
education, faculty, and administrators should consider the long-range goals of nontraditional students, as well as
the unique knowledge and skills that they are apt to possess. Moreover, engineering curriculum stakeholders are
encouraged to look for ways to make use of the range of experiences that nontraditional students bring to the
classroom so that they might improve engineering education for all students.
References
Atkinson, R. (2007). The life story interview as bridge in narrative inquiry. In D. J. Clandinin (Ed.), Handbook
of narrative inquiry: Mapping a methodology (pp. 224-245). Thousand Oaks, CA: Sage Publications.
Bean, J. P., & Metzner, B. S. (1985). A conceptual model of nontraditional undergraduate student attrition.
Review of Educational Research, 55(4), 485-540.
Becker, F. S. (2010). Why don't young people want to become engineers? Rational reasons for disappointing
decisions. European Journal of Engineering Education, 35(4), 349-366.
Benshoff, J. M., & Lewis, H. A. (1992). Nontraditional college students. ERIC Digest ED347483.
Berkner, L., Cuccaro-Alamin, S., & McCormick, A. (1996). Descriptive summary of 1989–90 beginning
postsecondary students: Five years later (NCES 96-155). Washington, D.C.
Berkner, L., He, S., & Catadli, E. (2002). Descriptive summary of 1995-1996 Postsecondary students: Six years
Later (NCES 2003–151). Washington, D.C.: U.S. Government Printing Office.
Bye, D., Pushkar, D., & Conway, M. (2007). Motivation, interest, and positive affct in traditional and
nontraditional undergraduaate students. Adult Education Quarterly, 57(2), 141-158.
Chen, X., & Weko, T. (2009). Stats in Brief: Students who study science, technology, engineering, and
mathematics (STEM) in postsecondary education (NCES 2009-161). National Center for Education
Statistics, U.S. Department of Education. Washington, D.C.
Choy, S. (2002). Findings from the Condition of Education, 2002. Nontraditonal undergraduates (NCES 2002-
012). National Center for Education Statistics, Washington, D.C.: U.S. Government Printing Office.
Clandinin, D. J. (2013). Engaging in Narrative Inquiry. Walnut Creek, CA: Left Coast Press.
Clandinin, D. J., & Connelly, F. M. (2000). Narrative inquiry: Experience and story in qualitative research. San
Francisco, CA: Jossey-Bass.
Conefrey, T. (2001). Sexual discrimination and women's retention rates in science and engineering programs.
Feminist Teacher, 13(3), 170-192.
Cross, K. P. (1980). Our changing students and their impact on colleges: Prospects for a true learning society.
Phi Delta Kappan, 61(9), 630-632.
Cross, K. P. (1981). Adults as learners: Increasing participation and facilitating learning. San Francisco, CA:
Jossey-Bass.
Dewey, J. (1938). Experience & Education. New York: Simon & Schuster.
Donaldson, J. F., & Townsend, B. K. (2007). Higher education journals' discourse about adult undergraduate
students. The Journal of Higher Education, 78(1).
Foor, C. E., Walden, S. E., & Trytten, D. A. (2007). "I wish that I belonged more in this whole engineering
group:" Achieving individual diversity. Journal of Engineering Education, 96(2), 103-113.
Gonclaves, S., & Trunk, D. (2014). Obstacles to success for the nontraditonal student in higher education. Psi
Chi Journal of Psychological Research, Winter, 164-172.
Haselgrove, S. (1994). The student experience. Bristol, PA: Open University Press.
Hesse-Biber, S. N. (2014). Feminist approaches to in-depth interviewing. In S. N. Hesse-Biber (Ed.), Feminist

research practice: A primer (pp. 182-232). Los Angeles, CA: Sage Publications.
Horn, L. (1996). Nontraditional undergraduates: Trends in enrollment for 1986 to 1992 and persistence and
attainment among 1989-1990 beginning postsecondary students (NCES 97-578) National Center for
Educational Statistics, Washington, DC: U.S. Government Printing Office.
Hwang, L. S. (December, 2013). Bridge in a Backpack. Popular Science.
Irizarry, J. (2009). Cultural deficit model. [White paper]. Retrieved from
http://www.education.com/reference/article/cultural-deficit-model
Jones, D. J., & Collier Watson, B. (1990). High risk students in higher education: Future trends. ASHE-ERIC
Higher Education Report No. 3. Washington D.C.: The George Washington University, School of
Education and Human Development.
Kasworm, C. E. (1990). Adult students in higher education: Myths and realities. Community/Junior College
Quarterly, 14, 155-175.
Kenner, C., & Weinerman, J. (2011). Adult learning theory: Applications to nontraditional college students.
Journal of College Reading and Learning, 41(2), 87-96.
Marshall, H., & Nicolson, P. (1991). Why choose psychology? Mature and other students' accounts at
graduation. In J. Radford (Ed.), The choice of psychology (Group of Teachers of Psychology, Occasional
Paper No. 12) (pp. 22-29). Leicester, U.K.: British Psychological Society.
McNeill, J. C., Long, R., & Ohland, M. W. (2014). Getting better with age: Older students achieve higher
grades and graduation rates. Paper presented at the IEEE Frontiers in Education, Madrid, Spain.
McNeill, J. C., & Ohland, M. W. (2015). Pathway, choice of major, and peer economic status of nontraditional
students in engineering. Paper presented at the IEEE Frontiers in Education Conference, El Paso, TX.
McNeill, J. C., & Ohland, M. W. (2016). Entry pathways, academic performance, and persistence of
nontraditional students in engineering by transfer status. Paper presented at the IEEE Frontiers in
Education Conference, Erie, PA.
Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook. Thousand Oaks,
CA: Sage.
Mobley, C., & Brawner, C. E. (2013). Engineering transfer students' views on orientation and advising. Paper
presented at the ASEE Annual Conference & Exposition, Atlanta, GA.
Mobley, C., Shealy, E. G., & Brawner, C. E. (2012). Work in progress: Transfer students in engineering: A
qualitative study of pathways and persistence. Paper presented at the Frontiers in Education Conference,
Seatlle, WA.
Mobley, C., Shealy, E. G., & Brawner, C. E. (2013). First-generation engineering transfer students: A
qualitative study of social and cultural capital. Paper presented at the IEE Frontiers in Educaton
Conference, Oklahoma City, Oklahoma.
Morice, P. B. (1990). Recruitment of engineering students in the UK. European Journal of Engineering
Education, 15(4), 357-360.
National Academies of Sciences Engineering and Medicine. (2016). In S. Malcolm & M. Feder (Eds.), Barriers
and opportunities for 2-year and 4-year STEM degrees: Systemic change to support students' diverse
pathways. Committee on Barriers and Opportunities in Completing 2-Year and 4-Year STEM Degrees.
Washington, D.C.: The National Academies Press. doi:10.17226/21739
National Center for Education Statistics (NCES). (2002). The Condition of Education 2002. NCES 2002-025.
Washington, DC: U.S. Department of Education, NCES.
National Survey of Student Engagement. (2006). Engaged learning: Fostering success for all students.
Retrieved from http://www.nsse.jub.edu
National Survey of Student Engagement Annual Report. (2006). Engaged learning: Fostering success for all
students. Retrieved from
http://nsse.indiana.edu/NSSE_2006_Annual_Report/docs/NSSE_2006_Annual_Report.pdf
Nelson, H. L. (1995). Resistance and insubordination. Hypatia, 10(2), 23-40.
Nyquist, J. D., Manning, L., Wulff, D. H., Austin, A. E., Sprague, J., Fraser, P. K., . . . Woodford, B. (1999). On
the road to becoming a professor. Change, 31(3), 18-27.
Ogilvie, A. M. (2014). A review of the literature on transfer student pathways to engineering degreees. Paper
presented at the ASEE Annual Conference & Exposition, Indianapolis, IN.
Ogilvie, A. M., Knight, D. B., Borrego, M. J., Fuentes, A. A., Nava, P. A., & Taylor, V. E. (2015). Transfer
Student Pathways to Engineering Degrees: A Multi-Institutional Study Based in Texas. Paper presented
at the IEEE Frontiers in Education Conference, El Paso, TX.
Ogilvie, A. M., Knight, D. B., Borrego, M. J., Fuentes, A. A., Nava, P. A., & Taylor, V. E. (2016). Transfer
student pathways to engineering degrees: Progress and preliminary findings from a multi-institutional
study. Paper presented at the ASEE Annual Conference & Exposition, New Orleans, LA.
284 Minichiello
Pascarella, E. T., & Terenzini, P. T. (1998). Studying college students in the 21st century: Meeting new
challenges. The Review of Higher Education, 21(2).
Patton, M. (2002). Qualitative research and evaluation methods (3rd ed.). Thousand Oaks, CA.: Sage
Publications.
Pawley, A. L. (2013). "Learning from small numbers" of underrepresented students' stories: Discussing a
method to learn about institutional structure through narrative. Paper presented at the 120th ASEE
Annual Conference and Exposition, Atlanta, GA.
Polkinghorne, D. E. (1988). Narrative knowing and the human sciences. Albany, NY: State University of New
York Press.
Polkinghorne, D. E. (1995). Narrative configuration in qualitative analysis. International Journal of Qualitative
Studies in Education, 8(1), 5-23.
President's Council of Advisors on Science and Technology (PCAST). (2012). Report to the President. Engage
to excel: Producing one million additional college graduates with degrees in science, technology,
engneering, and mathematics. Washington, D.C.
Quinnan, T. W. (1997). Adult students "at risk": Culture bias in higher education. Westport, CT: Bergin
&Garvey.
Radford, A., Cominole, M., & Skomsvold, P. (2015). Web tables—Demographic and enrollment characteristics
of nontraditional undergraduates 2011-12 (NCES 2015-025). Washington, D.C.
Richardson, J. T. E., & King, E. (1998). Adults in higher education: Burden or boon? Journal of Higher
Education, 69(1).
Schaefer, L. M., Lessard, S. M., & Lewis, B. (2017). Engaging in a co-created narrative inquiry research space
alongside urban aboriginal youth: Pedagogical implications for preservice teachers. Paper presented at
the American Educational Research Association Annual Meeting, San Antonio, TX.
Scott, L. M., & Lewis, C. W. (2012). Nontraditional college students: Assumptions, perceptions, and directions
for a meaningful academic experience. The International Journal of Interdisciplinary Social Sciences,
6(4), 1-10.
Shealy, E. G., Brawner, C. E., Mobley, C., & Layton, R. A. (2013). A descriptive study of engineering transfer
students at four institutions: Comparing lateral and vertical transfer pathways. Paper presented at the
ASEE Annual Conference & Exposition, Atlanta, GA.
Sullivan, M. D., de Cohen, C. C., Barna, M. J., Orr, M. K., Long, R. A., & Ohland, M. W. (2012).
Understanding engineering transfer students: Demographic characteristics and educational outcomes.
Paper presented at the IEEE Frontiers in Education Conference, Piscataway, NJ.
Valencia, R. R. (1997). Conceptualizing the notion of deficit thinking. In R. R. Valencia (Ed.), The evolution of
deficit thinking, Educational thought and practice (pp. 18-29). London, UK: RoutledgeFalmer.
Wyatt, L. G. (2011). Nontraditional student engagement: Increasing adult student success and retention. The
Journal of Continuing Higher Edcuation, 59(1), 10-20.
Zacharakis, J., Steichen, M., Diaz de Sabates, G., & Glass, D. (2011). Understanding the experiences of adult
learners: Content analysis of focus group data. Adult Basic Education and Literacy Journal, 5(2), 84-95.
Author Information
Angela Minichiello
Department of Engineering Education
Utah State University
4160 Old Main Hill Logan, Utah 84322
U.S.A.
Contact e-mail: angie.minichiello@usu.edu
“Just put it together to make no commotion:” Re-imagining Urban

Elementary Students’ Participation in Engineering Design Practices
Christopher G. Wright, Kristen B. Wendell, Patricia P. Paugh

Article History In the growing field of K-12 engineering education, there is limited research that
highlights the experiences of youth from historically marginalized communities
Received: within engineering learning environments. This study offers insights into the
07 July 2017 ways in which two groups of elementary school students constructed approaches
for participating in the engineering design practice of collaborative reflective
Accepted:
decision-making. Findings suggest that students conceptualized urban,
21 December 2017
engineering learning environments as spaces for risk management. This notion of
managing risks informed their participation in collaborative decision-making,
Keywords
and the ways in which they viewed themselves as doers of engineering.
Engineering practices Implications for this study include the continued need for the development of
Urban schools methodologies and frameworks that provide opportunities to uncover these
Team potential risks, and design supports for student participation in engineering
Decision-Making design practices.
“I’m a ten because I’m a good student. I’m a good listener. I follow directions, and I do what I’m
supposed to do.” – Cynthia, 4th grade engineering student
Introduction
When five African American students from two different elementary schools discussed their views of themselves as
“doers of engineering,” their self-appraisals varied across a scale from 1-10. Despite this variation in self-
evaluations, students’ responses were similar to the response that Cynthia provided above. Self-appraisals were
heavily informed by students’ perceptions of their abilities to negotiate the perceived risks associated with
participating in an engineering design practice referred to as collaborative reflective decision-making (Wendell,
Wright, & Paugh, 2017). Specifically, students’ views of their adeptness to enact the qualities of a doer of
engineering were situated in their abilities, or inabilities, to successfully negotiate the intellectual and social risks
associated with engaging in collaborative reflective decision-making within the context of their classroom.
Collaborative reflective decision-making was envisioned as a combination of two disciplinary practices in the Next
Generation Science Standards’ (NGSS Lead States, 2013): (a) designing solutions and (b) engaging in argument
from evidence. The practice required students to reflect on previously gathered information and/or ideas while
making team design decisions. Unexpectedly, when presented with opportunities to evaluate their participation in
collaborative reflective decision-making, the five elementary school students characterized their “engineering
abilities” through a lens that emphasized displaying good behavior and following strict instructions for developing
“right answers.”
Highlighting the experiences of youth within urban classrooms, this research argues for the use of theoretical
frameworks that provide opportunities to critically examine and frame the complexities of student engagement in
engineering design practices, such as collaborative reflective decision-making. Specifically, this study looks to offer
insights into the ways that a group of students within an urban context developed situated identities (Gee, 1999;
Lave & Wenger, 1991; Leander, 2002) as doers of engineering. Empirical work examining students’ engineering
experiences within urban learning environments is limited (Denson, Avery, & Schell, 2010; Mehalik, Doppelt, &
Schunn, 2008; Silk, Schunn, & Cary, 2009), and even more so at the elementary school level (Capobianco, Diefes-
286 Wright, Wendell, & Paugh
Dux, Mena, & Weller, 2011; Capobianco, Ji, & French, 2015; Marulcu & Barnett, 2013). We contend that as
engineering’s visibility in elementary school contexts increases (Aguirre-Muñoz & Pantoya, 2016; Cunningham,
2009; Hegedus, Carlone, & Carter, 2014; Lottero-Perdue, Bodwitch, Kagan, Robinson-Cheek, Webb, Meller, &
Nosek, 2016), it is imperative to empirically call attention to the experiences of youth from historically marginalized
communities to acknowledge engineering’s history of both marginalization and underrepresentation (Frehill, 2004;
McGee & Martin, 2011; Moore, Madison-Colmore, & Smith, 2003; Tonso, 2006) and to proactively explore
opportunities for addressing these patterns at the elementary school level (Cunningham & Lachapelle, 2014).
Situated within Spencer’s (1995) phenomenological variant of ecological systems theory (PVEST) framework, this
research provides insights into a group of urban students’ emergent engineering identities. In the context of this
study, emergent engineering identities were conceptualized as the ways in which students viewed themselves, as
well as their peers, as “doers of engineering” as it was realized within the specific contexts of their elementary
classrooms. Succinctly, students described what kind of engineering students they were expected to be within social
contexts and if they were able to meet these expectations. In the following sections, we present the theoretical
framework and construct that guided this work and conceptualized the practice of collaborative reflective decision-
making within this framework. Next, we present and analyze data from student interviews and classroom
observations to unpack students’ perceptions of their participation in collaborative reflective decision-making and
their experiences with negotiating the challenges associated with this engineering design practice. Finally, we
conclude with a discussion of the importance of utilizing alternate theoretical frameworks for providing insights into
students’ experiences.
PVEST and Emergent Engineering Identities

For this study, our sociocultural perspective deployed Spencer’s (1995) phenomenological variant of ecological
systems theory (PVEST framework (see Table 1) by specifically foregrounding the tenet of students’ emergent
engineering identities. Generally, PVEST is a framework that “builds a bridge between identity and context”
(Swanson, Spencer, Harpalani, & Spencer, 2002, p. 75), whereby an individual’s self-appraisal and meaning making
are analyzed by also accounting for the contextual contributors. “Thus, an individual’s perceptions about settings
and their experiences in them matter” (Spencer, 2006, p. 697). As a tenet within the larger PVEST framework,
emergent identities are defined as the ways in which individuals view themselves within various contexts of growth
and development (Lee, Spencer, & Harpalani, 2003). This situated notion of identity was utilized to inform the
construct of emergent engineering identities, which explored how elementary school students viewed themselves, as
well as their peers, as doers of engineering within the contexts of their specific classrooms. Calling upon PVEST, in
general, and the tenet of emergent identities, we foregrounded the belief that all humans are burdened by varying
levels of vulnerabilities, or risks, that impacts their growth and development. We argue that the physical and
imagined realities of being a student within urban contexts (Emdin, 2010; Martin & Larnell, 2014) impacted
students’ developing views of themselves as doers of engineering. Although all students experience varying levels
of vulnerability within school contexts, we insist that deficit perceptions of “urbanness” (Watson, 2011) position
students from urban contexts in spaces of increased intellectual and social vulnerabilities and risks. For instance,
teachers, both pre-service and in-service, have been found to share concerns around personal safety, student abilities,
cultural conflicts, and language barriers when faced with the task of working within urban contexts (Hampton, Peng,
& Ann, 2008; Knoblauch & Hoy, 2008; Lynn, Bacon, Totten, Bridges, & Jennings, 2010; Siwatu, 2011). In this
research, we contend that to understand students’ development and engagement in disciplinary practices in
engineering design that we must examine their experiences through lenses that recognize the imposed expectations
of students within urban communities.
In an effort to begin to interrupt the deficit perspectives that are often associated with students within urban schools,
we utilized the construct of emergent engineering identities to uncover the vulnerabilities and risks students
associated with their participation in collaborative reflective decision-making. Furthermore, focusing on emergent
engineering identities provided opportunities to hear directly from elementary school students about the risks they
perceived and their approaches for managing these risks. Framing students’ participation as strategies for managing
risks provides an alternate view of student engagement in engineering design practices. The next section looks to
further conceptualize collaborative reflective decision-making through a lens of risk management.
Table 1. Tenets of the phenomenological variant of ecological systems theory (PVEST) (Swanson et al., 2002, p. 76)
PVEST Tenet Description of PVEST Tenet Translation to the current study
Net Vulnerability Consists of the context and characteristics that Conceptualized as the intellectual and
Level can potentially pose a challenge during one’s behavioral stigmas that are often
development at any life stage. A risk contributor associated with students within urban
is a factor that may predispose an individual for contexts.
adverse outcomes during a particular
developmental stage.
Net Stress Refers to the actual experiences of a situation or Conceptualized as the risks that students
(Engagement) context that challenges the individual’s well associated with their participation in
being. These are the challenges that one actually collaborative reflective decision-making.
encounters and are juxtaposed against any
available support(s).
Reactive Coping Problem-solving strategies that an individual Conceptualized as the practices or

Strategies employs to deal with stress and dissonance. approaches that students co-constructed in
Reactive coping responses can lead to either an response to the risks associated with
adaptive or maladaptive solution. participating in collaborative reflective
decision-making.
Emergent As an individual employs various coping Translated through the idea of emergent
Identities: Stable strategies, self-appraisal develops. The strategies engineering identities, or the ways
Coping Responses that produce desirable results are replicated. students conceptualized what it meant to
Accordingly, these become stable coping be a doer of engineering within their
responses, and, coupled together, they yield classroom.
emergent identities.
Life-Stage Coping Identity lays the foundation for future Translated as students’ future perceptions,
Outcome perceptions, self-appraisals, and behavior, self-appraisals, and behaviors as
yielding adverse or productive life-stage, specific engineering students. Examining
coping outcomes. elementary students’ life-stage coping
outcomes were not within the scope of
this study.
Conceptualizing Collaborative Reflective Decision-Making through a Lens of Risk Management
Engineers are often described as professional decision-makers (Chen &Wassenaar, 2003; Hazelrigg, 1998), thus
drawing attention to the importance of effectively and efficiently making sound decisions throughout the
engineering design process. Design decisions within professional engineering contexts typically carry high
consequences, potentially impacting the success of a design solution, the engineer herself or himself, the business
affiliated with the engineer(s), and the society at large (Gandy, Jager, Bertsche, & Jensen, 2007; Jankovic, Stal-Le
Cardinal, & Bocquet, 2010). The complexities associated with generating optimal engineering design solutions
involve prioritizing a variety of information that is available to the design team, including, but not limited to, social
goals, scientific understandings and applications, ecological impacts, and financial considerations and implications.
Thus, the research presented here recognized and positioned decision-making as a vital disciplinary practice within
the engineering design process.
Collaborative reflective decision-making was envisioned as the practice of collectively reflecting on previously
gathered information and/or generated ideas in order to make in-the-moment, intentional, and informed design
decisions within the contexts of engineering design activities (Wendell, Wright, & Paugh, 2017). The presumption is
that this in-the-moment reflective practice of making design decisions based on “evidence” that a given design or
idea would work would minimize a design team’s dependence on trial-and-error methodologies, and foster an
efficient and effective design experience. In this research, disciplinary norms for effectively participating in this kind
of learning environment required students to carefully navigate through the process of critiquing the ideas of
teammates, arguing points through the use of tangible evidence, defending potential design choices and ideas, and
negotiating multiple individual ideas within a design team (Jordan & McDaniel, 2014). For instance, elementary
school doers of engineering were expected to verbally articulate ideas and/or solutions within a design team,
evaluate the pros and cons of each potential solution, and intentionally select a solution based on the previous
analyses of their various options (Wendell, Wright, & Paugh, 2017).
Although students were expected to participate in these disciplinary practices associated with collaborative reflective
decision-making, this study acknowledges that these practices could potentially position students in vulnerable
spaces, especially those students attending schools within urban communities. Emdin (2016), a noted urban
education scholar, points out that “students who populate urban schools are generally beholden to a pedagogy of
poverty that rewards them for being docile and punishes them for being overly vocal and expressive” (p. 66).
Emdin’s observations call attention to the didactic and teacher-controlled instruction prevalent within urban school
contexts (Ferguson, 2000; Rivera-Maulucci, 2010; Thadani, Cook, Griffis, Wise, & Blakey, 2010) that also
emphasizes students’ maintenance of “appropriate behavior” to succeed. Thus, the research presented here
recognized the potential for students in urban communities to encounter various risks when participating in
collaborative reflective decision-making, where instances of critique, argumentation, debate, and negotiation are
encouraged within classrooms that typically reward students for contrasting practices of compliance and docility.
Addressing the Need for Continued Research on Argumentation in Elementary School Classrooms
The practices of argumentation and critique, foundational practices for engaging in collaborative reflective decision-
making, have received increased attention within the science and engineering education communities. For instance,
the National Research Council Framework (2012) contends, “students should argue for the explanations they
construct, defend their interpretations of the associated data, and advocate for the designs they propose” (p. 73).
Empirical research around these practices often highlights the ways in which these practices enhance students’
conceptual understanding and scientific reasoning (Duschl & Osborne, 2002; Osborne, Erduran, & Simon, 2004),
and role of social interaction in learning and thinking processes (Jiménez-Aleixandre & Erduran, 2008).
Previous science education research has often cited elementary school students’ capabilities for developing quality
scientific arguments. For instance, McNeill (2011) focused on how 5 th grade students’ abilities to engage in
argumentation changed over the course of a school year. She found that teachers were able to positively impact
students’ developing stronger arguments through the use of instructional supports that built upon students’ everyday
resources. Ryu and Sandoval (2012) also found that an instructional focus on argumentation supported elementary
students’ improved abilities in constructing and evaluating arguments. This finding was most attributed to
improvements in the development of several classroom norms that supported students’ defining “good arguments.”
Finally, while examining learning progressions for scientific argumentation, Berland and McNeill (2010) suggest
that the instructional context supports students’ engagement in argumentation in complex ways. While the
discussion of these three studies is not intended to serve as an exhaustive review of argumentation research in
elementary science education (Cavagnetto, Hand, & Norton-Meier, 2010; Martin & Hand, 2009; Richmond &
Striley, 1996; van Zee, Hammer, Bell, Roy, & Peter, 2005), collectively, they suggests potential supports for
addressing some of the complexities for engaging elementary children in argumentation. In our research we applied
these findings that instructional contexts and the development of classroom norms have implications for students’
abilities in developing quality arguments. Further, by situating our work within a phenomenological variant of
ecological systems theory, we feature the roles that the classroom contexts and norms play in students’ development
in collaborative reflective decision-making.
Building upon the increased focus on argumentation and critique in science education, our current work looks to
explore the demands of argumentation and critique during collaborative reflective decision-making in elementary
engineering contexts. Research on argumentation in engineering education is not as extensive as that in science
education; however, previous studies in this area do provide important insights for our current research. Similar to
the aforementioned reasons for highlighting argumentation in science education, Garcia and Mazzotti (2016) argue
that approaching engineering education from an argumentation perspective has the potential for providing students
with opportunities to develop disciplinary literacies. They argue that by enabling critical thinking and empowering
decision-making, students are allowed to refine their abilities for presenting plausible justifications for selected
solutions. However, earlier work called attention to difficulties in leveraging argumentation for engineering learning
in undergraduate contexts. Kittleson and Southerland (2004) found that undergraduate engineering students rarely
engaged in concept negotiation, “a form of collaborative interaction in which more than one participant actively
contributes to the evolving conceptual content of the conversation” (p. 271). They identified several themes that
played a role in how and when groups engaged in concept negotiation, including (a) assumptions about the purpose
of group work, (b) views about effective groups, and (c) their epistemologies and ontologies. They conclude that it is
important for researchers and educators to uncover these types of orientations when examining practices such as
argumentation and critique within group interactions. In another study, Purzer (2011) hypothesized a significant
relationship between undergraduate engineering students’ achievement and how many “challenge-oriented
discourse” (p. 670) actions they received from their peers. Challenge-oriented discourse actions included instances
of disagreement, argumentation, and defending one’s points. She found that students were more likely to show
agreement and ask questions rather than challenge each other’s ideas. Purzer argues that opportunities exist for
creating additional learning opportunities if students are taught to engage in effective argumentation. These findings
highlight a need for continuing research that looks to address the social challenges of engaging in argumentation
within engineering design contexts. To address the issues highlighted in this review, the research questions that
guide this work are: (a) How did students conceptualize what it meant to be doers of engineering within their
specific classrooms? (b) How did students respond to perceived risks associated with engaging in collaborative
reflective decision-making? (c) What are the risks that students associated with this practice?
Methods
Study Contexts and Sample
This study took place in two engineering classrooms within two different schools located in the southeastern part of
the United States. For the privacy of the participating schools, teachers, and students, pseudonyms were utilized. The
two schools, Medgar Evers Elementary School and Fannie Lou Hamer Elementary School, were considered schools
with “urban characteristics” (Milner, 2012, p. 559) and located within the same school district. Here, “urban
characteristics” are used to describe schools that are not located in large cities but nonetheless are experiencing
increased challenges often associated with larger urban contexts. These challenges may include, but are not limited
to, inadequate resources for authentically engaging students in engineering design or science inquiry or the lack of
culturally and linguistically responsive approaches and assessments for effectively engaging diverse populations of
students. The larger project included two classrooms from each school, for a total of four classrooms; however, only
one classroom from each school was included in this smaller research study. Adhering to a case study methodology
(Stake, 1995; Yin, 2011), one engineering design team from each participating classroom became the focus of this
research. All work produced from two teams were collected, including engineered artifacts and design drawings
from engineering classroom interactions. Specific details of each school, engineering classroom, and design team are
provided below.
Description of Evers Elementary School
The student demographics at Evers Elementary were 68% African American, 21% white, 6% biracial, and 5%
Hispanic and 76% of students received free or reduced lunch. The featured design team at Evers consisted of three
5th grade African American male students named Corey, Keith, and Toren. This team was selected as the case study
for this classroom based on a recommendation from Ms. Simpson, the students’ general classroom teacher and
facilitator of the students’ engineering experiences. Ms. Simpson recommended the selected team because we
wanted a heterogeneous ability group to see how different members interacted with one another. Design teams at
Evers Elementary participated in a total of 18 hours of engineering design activities that included three units from
the Engineering is Elementary curricula (Museum of Science, Boston): (a) A slick solution: Cleaning an oil spill, (b)
The best of bugs: Designing hand pollinators, and (c) Now you’re cooking: Designing solar ovens. A vital
component of collaborative reflective decision-making is the ability to engage in argument through the use of
tangible evidence (Wendell, Wright, & Paugh, 2017). Lesson Three of each of these units provided students
opportunities to collect the necessary evidence for engaging in collaborative reflective decision-making. Lesson
Three is described as the lesson that details how scientific data inform engineering design (Museum of Science,
Boston). For instance, the best of bugs unit provided students the opportunity to perform controlled experiments to
identify effective materials for picking up and dropping off pollen. This process included testing and describing the
properties of materials of objects, such as, erasers, marbles, pipe cleaners, tape, and pompoms. Neither the students,
nor Ms. Simpson, had previous experience with engaging in engineering design activities. Due to the school’s below
average test scores in mathematics and reading, the majority of the children’s day consisted of test preparation and
enrichment activities within the areas of mathematics and reading. School personnel considered students’
engagement in engineering design activities as a fun, hands-on opportunity for students. Students were not given
grades for their participation in these activities.
Description of Fannie Lou Hamer Elementary STEM School
The student demographic at Hamer Elementary were 74% African American, 15% white, 6% biracial, and 5%
Hispanic, and 81% of students received free or reduced lunch. The featured design team at Hamer consisted of three
4th grade female students named Abigail, Cynthia, and Jata. Abigail was a white female student, and Cynthia and
Jata were African American female students. Although Abigail served as an integral member of this design team,
she was not included in the collection of interview data. At the time of the interview data collection, Abigail had
transferred schools and was unavailable for participation. This team was selected as the Hamer Elementary case due
to the recommendation of Ms. Humphrey, Hamer’s technology specialist and facilitator of the students’ engineering
experiences. Similar to the student design team at Evers Elementary, this design team was recommended because we
wanted a heterogeneous ability group to see how different members interacted with one another. The design team at
Hamer participated in a total of 12 hours of engineering design activities specific to this project that included two
units from the Engineering is Elementary curricula (Museum of Science, Boston): (a) Water, water, everywhere:
Designing water filters and (b) The best of bugs: Designing hand pollinators. Hamer students were provided the
same opportunities to engage in data collection to inform their engineering designs, as described in the Evers
Elementary description above. At the time of the study, Hamer was a school that was in transition and attempting to
change its culture. Mathematics and reading scores were well below the state and national average, and the school
was recently reconstituted as a mathematics and science magnet school. With this reconstitution came modest
increases in student participation in engineering design activities within the school’s engineering design laboratory.
Despite the once a week engineering experiences, the school was heavily focused on improving its test scores
through increased enrichment activities and test-taking preparations.
Data Collection and Analysis
Data sources for the study included semi-structured interviews and classroom observations of team interactions
during engineering design activities. Study participants took part in face-to-face interviews lasting no longer than 30
minutes. Each interview was video recorded and transcribed for later analysis. During the interviews, students
shared insights into their approaches for working within the design team and coming to team consensus on design
decisions, as well as their assessments of being doers of engineering. Data included video recordings of 12 hours of
engineering team interaction at Hamer Elementary and 18 hours of engineering team interaction at Evers
Elementary, transcripts of these interactions, and field notes generated during each interaction.
Interview data for this study were analyzed using interpretive phenomenological analysis (Larkin, Watts, & Clifton,
2006; Quinn & Clare, 2008; Smith, 2004) to make sense of how students characterized what it meant to be a doer of
engineering within their classrooms, their abilities or inabilities to enact these characteristics, and their approaches
for engaging in collaborative reflective decision-making. Interpretive phenomenological analysis as a qualitative
framework for data collection provided opportunities to understand the students’ lived experiences – that of a
student within an urban engineering classroom – and how they made meaning of these experiences. Using a
thematic analysis approach (Marshall & Rossman, 2015), the research team utilized three categories derived from
the PVEST framework (see Table 1) in analyzing data: (a) reactive coping strategies, (b) stress engagements, and (c)
emergent identities. Reactive coping strategies, or the strategies that students employed to resolve dissonance-
producing situations in response to perceived challenges for engaging in collaborative reflective decision-making,
were identified through examining students’ interview responses. Each student provided details into how their
design team approached collaborative reflective decision-making and the research team analyzed for any emerging
patterns across these responses. In addition, video recording and transcripts of design team interactions were
analyzed to identify instances in which teams engaged in collaborative reflective decision-making, or the critique,
argumentation, and negotiation of ideas to produce a single design solution for the team. These data were
secondarily utilized to examine instances of team decision-making and coordinate with students’ descriptions of
their approaches for engaging in collaborative reflective decision-making from their interviews.
Student interview data were utilized to analyze the final two categories of stress engagement and emergent
identities. Stress engagements, or the specific challenges students associated with engaging in collaborative
reflective decision-making, were identified as students discussed the approaches they utilized for engaging in the
disciplinary practice. Adhering to a PVEST framework, we expected to identify and connect students’ perceived
challenges for engaging in collaborative reflective decision-making with their reactive coping strategies. As students
talked about the approaches they utilized for engaging in the disciplinary practice, they also described the
experiences that led them to utilize these approaches. These described experiences were coded as “challenges for
engaging in collaborative reflective decision-making.” Finally, student interview data were also utilized to analyze
students’ emergent identities, or the ways that students viewed themselves as doers of engineering, and to identify
the similarities and variations in the experiences of different students. The next section opens by highlighting these
data, and addresses the question of how students conceptualized what it meant to doers of engineering within their
specific classrooms.
Findings
Examining Students’ Emergent Engineering Identities
This section looks to build upon PVEST’s fourth tenet, emergent identities (see Table 1), to unpack and characterize
students’ emergent engineering identities. Emergent engineering identities were conceptualized as the ways in which
students viewed themselves, and their classroom peers, as doers of engineering. To explore students’ emergent
engineering identities, students were presented with the question, “if you had to place yourself on a scale of 1-10, 1
being the lowest and 10 being the highest, where you would place yourself in regards to engineering and why?” As
communicated in the opening paragraph of this manuscript, students’ self-appraisals as doers of engineering varied
and included scores ranging from 5-10 (see Table 2). We contend that the variations in students’ self-appraisals were
informed by their evaluation of being able to maintain “appropriate behavior” during engineering class.
Table 2. Emergent engineering responses from the five participating elementary school students
Student School Response to emergent engineering question
Name
Corey Evers Six or seven. I was being a little crazy with someone that I always get in trouble with and I
think it’s a bad thing to always to be around that person that get you in trouble. So, I learned
my lesson and I stopped being around the people that makes me crazy and that’s why I’m a
good student now.
Cynthia Hamer Ten, because I’m a good student. I’m a good listener, I follow instructions, and I do what I’m
supposed to do. And that’s it.
Jata Hamer Five because I’m in the middle of it. I’m not really that bad, but I’m not really that good.
Because I know – how to work together with people.
Keith Evers Seven to eight because sometimes I can be off task and sometimes I can – sometimes I have
my moments, like paying attention for the whole class and doing good.
Toren Evers Ten. Well, because I’ve never really done anything to get kicked out of it [engineering class]
or anything else or a teacher had to actually say something to me; tell me to get out the room.
Despite the variations in students’ self-appraisals, the above data suggest consistency in students’ conceptions of
what constitutes a doer of engineering. Within and across the two engineering design teams, student responses
focused on their ability, or inability, to “behave correctly” or following specific instructions. Absent from any of the
above responses are specific engineering or engineering practice references. The similarities in student responses are
consistent with Emdin’s (2016) aforementioned observations where students within urban academic settings are
often rewarded for their docility and compliance. Also visible from students’ self-assessments (see Table 2) were the
ways in which their emergent engineering identities developed during their engagement in engineering design
activities, while also being re-evaluated by the ways in which they were able to adjust their behavior throughout the
year. For instance, Corey and Keith (see Table 2) provided responses that articulated a fluid understanding of their
self-appraisal that was connected with their abilities to alter their classroom behavior over the course of the
academic year. Corey provided a narrative that described getting into trouble with a peer, learning his lesson at some
juncture during the year, and believing that this was the reason that we was a “good student now.” In a similar
response, Keith also specified instances in which he considered himself to “be off task” and acknowledged these
moments as contributing to his self-assessment of 7-8. Other students, specifically Toren and Cynthia, articulated
more stable associations with their ability to abide by the emergent theme of “good behavior” and confidently
viewed themselves as doers of engineering. Although they evaluated themselves as “10’s,” their assessments were
connected to their abilities to avoid “getting kicked out” and being able to “follow instructions” instead of any
specific engineering practice.
To further exemplify the relationship between students’ emergent engineering identities and the idea of behaving
appropriately in class, we present additional interview data from Evers Elementary. Students were asked to identify
three classroom peers who they thought were smart engineering students. Student responses (see Table 3) provide
additional support for the highlighted theme of doers of engineering being coupled with the assessment of one’s
behavior during engineering team interactions.
Table 3. Evers’ student responses to naming smart engineering peers

Students’ Student Responses
Name
Corey Olivia, Toren, and Aaliyah, because they [Olivia and Aaliyah] both don’t do anything, they just know
not to be with people that get them in trouble. They just know straight right there, if you know
somebody gone get you in trouble don’t go with them. So, she [Olivia or Aaliyah] just go with girls
that won’t get her in trouble. Like, Olivia usually goes with Aaliyah and Aaliyah with Olivia. They
just think the right way and they don’t talk to each other like that…
Keith I might say Aaliyah or Olivia because they don’t – they don’t really get in trouble like that.
Toren Olivia, Corey, and Aaliyah.
Corey, Keith, and Toren all identified Aaliyah and Olivia as class exemplars of “smart engineering students” and
described the girls’ abilities to “think the right way” and avoid “getting in trouble.” Coupled with student responses
in Table 2, these responses suggests that students conceive doers of engineering as students who demonstrate the
ability to avoid “getting into trouble” during team interactions. This ability to avoid getting into trouble included
students’ ability to follow teachers’ strict instructions of rules and knowing how to talk to or work with specific
teammates. PVEST describes emergent identities (see Table 1) as developing from students’ deployment of reactive
coping strategies (McGee & Spencer, 2013) that produce desirable results. From this perspective, the next section
examines a reactive coping strategy that we believe contributed to students’ emergent engineering identities within
these two classrooms.
Situating the Approach of Combining Multiple Ideas as a Coping Reactive Coping Strategy
The goal of this section is to make sense of and characterize an unexpected approach for students’ participation in
collaborative reflective decision making that is referred to as “combining ideas.” Specifically, we look to situate
students’ combining ideas approach within the PVEST tenet of reactive coping strategies (McGee & Spencer, 2013;
Spencer, Dupree, & Hartman, 1997; Spencer, Noll, Stoltzfus, & Harpalani, 2001), or as a problem-solving strategy
that students utilized to avoid “getting into trouble” while being asked to engage in collaborative reflective decision-
making.
As the research team reviewed and analyzed data from two different elementary school engineering design teams,
this approach of combining ideas emerged as a common approach across various design teams. The primary practice
of combining ideas consisted of students incorporating several individual ideas within the team’s final design
solution without the interrogation or discussion of the pros and cons of these ideas. This approach contradicted the
research team’s expectations for student participation in collaborative reflective decision-making. To provide further
context for what this approach looked like within a design team, an illustrative case is provided below. In this
example, Abigail, Cynthia, and Jata are engaged in the water, water everywhere: designing water filter curriculum
developed by the Engineering is Elementary team (Museum of Science, Boston). In this environmental engineering
unit, students explored properties of potential water filter materials (e.g., gravel, cotton balls, cheesecloth), applied
their knowledge of water, and planned and constructed team water filters.
In the excerpt below, the three 4th graders were making decisions regarding the team’s water filter design. Prior to
the brief discussion documented below, each girl individually designed a solution, shared their individual design
during a team discussion, and came together to decide on a single team design solution or to participate in
collaborative reflective decision-making. This specific excerpt was selected because of the students’ explicit
reference to adhering to an approach that included “adding ideas together,” or combining multiple ideas.
Excerpt 1. Illustrative case of a design team combining multiple ideas
1 Abigail Well, can we draw something?

2 Jata Okay, are we going to use my idea or are we going to add all ideas together?
3 Cynthia Add all our ideas together//
4 Abigail // Coffee filter. ((Begins drawing coffee filter on the team’s whiteboard))
5 Cynthia What is this? ((Referring to Abigail’s drawing of a coffee filter))
6 Abigail Coffee filter. [See Figure 1]
7 Cynthia That’s water.
8 Abigail That ain’t no water!
In the above excerpt, Jata (line 2) initiated a move for the team to make a decision on how to proceed with the
team’s design. She explicitly offered her teammates two potential methods for participating in collaborative
reflective decision-making: (a) use the idea that she previously put forward to the team or (b) add all of their ideas
together. Analyzing this exchange, we argue that Jata expressed confidence that her design would be an effective
solution (e.g., are we going to use my idea?); however, she also felt the need to offer the team an additional option
of “adding all our ideas together” (line 2). Following Jata’s offer, Cynthia (line 3) verbally voted for the use of a
combining ideas approach, while Abigail (line 4) physically validated the approach by initiating the drawing of her
contribution towards the team’s final design solution. Thus, the team’s incorporation of this approach was solidified.
This interaction continued with Cynthia’s critique (lines 5 and 7) of the “appearance” of the coffee filter drawn by
Abigail, as opposed to a debate regarding the potential benefits or disadvantages of including this material in the
design of the water filter. Following the conclusion of this team’s exchange, the team of girls continued the practice
of combining ideas that included Cynthia’s addition of cheesecloth and Jata’s addition of cotton balls and a piece of
screen to the team’s final design (see Figure 1).
Figure 1. Drawing of the design team’s water filter design

A combining ideas approach could have been recognized as collaborative reflective decision-making if teammates
had debated and unpacked the pros and cons of the individual ideas with the purpose of developing an optimal
design solution for the design challenge. However, this specific design team, as well as others observed throughout
the study, utilized an approach where un-interrogated ideas were added within the team’s final design. For instance,
Cynthia questioned Abigail’s drawing of a coffee filter by critiquing the aesthetics of the coffee filter and describing
it as “water” (line 7). Due to previous teacher modeling of engaging in collaborative reflective decision-making and
insights into developing an optimal solution, the expected approach for engaging in collaborative reflective decision-
making would have included a discussion about the potential effectiveness for incorporating the coffee filter into the
final design. From this perspective, a combining ideas approach would have been conducive for participating in
collaborative reflective decision-making if teammates had debated the pros and cons for incorporating various
design elements.
To situate a combining ideas approach as a reactive coping strategy (McGee & Spencer, 2013; Spencer et. al., 1997;
Spencer et al., 2001), associated risks must also be identified and unpacked. During individual student interviews,
Jata was asked why she offered her team the option of “adding all their ideas together” (line 2) instead of pushing
for the adoption of her individual idea. Jata responded:
Because, sometimes, I can get in arguments. Because you don’t want Ms. Humphrey thinking
you’re in an argument with nobody. You get in an argument then it turns out to be a fight then
you get suspended, and you get a whoopin’ at home, you know?
Jata’s response provided initial insights into the potential risks that students associated with participating in
collaborative reflective decision-making. Here, Jata introduced the risk of being identified as participating in an
argument when debating the pros and cons of individual ideas. With the risk of Ms. Humphrey thinking you’re in an
argument, Jata processed the potential for being reprimanded and the potential consequences that could follow, e.g.,
suspension from school and punishment at home. From Jata’s insight, we argue that the approach of combining
ideas served as a reactive coping strategy (McGee & Spencer, 2013; Spencer et al., 1997; Spencer et al., 2001) for
many teams as they participated in collaborative reflective decision-making. Jata employed this approach as a
problem-solving strategy for alleviating the potential for being reprimanded for the perception of engaging in
negative behavior. The following section utilizes student interview data to highlight additional risks that students
associated with collaborative reflective decision-making, while corroborating students’ use of a combining ideas
approach.
Identifying Perceived Risks Students Associated with Participating in Collaborative Reflective Decision-
Making
The previous section argued that the students’ approach to combining ideas should be positioned as a reactive
coping strategy (McGee & Spencer, 2013; Spencer et al., 1997; Spencer et al., 2001). This characterization as a
reactive coping strategy required the research team to also uncover the various risks that students were “reacting” to
while participating in collaborative reflective decision-making. Lee and colleagues (2003) describe PVEST’s second
tenet, stress engagements (see Table 1), as the “actual experiences or situations that challenge an individual’s well
being” (p. 9). For this study, stress engagements were conceptualized as the specific risks that students associated
with participating in a learning environment that promoted engagement in critique, argumentation, debate, and
negotiation. The previous section introduced the risk of being reprimanded for the perception of engaging in
detrimental behavior, while this section looks to identify additional risks students highlighted during student
interviews.
To explore students’ perceptions for participating in collaborative reflective decision-making, students were
presented the question, “what kinds of things did you need to do in order to come up with one solution as opposed to
having different ones?” Student responses provided insights into the perceived risks for participation (PVEST tenet
of stress engagement) and their responses to these risks (PVEST tenet of reactive coping strategies). For instance,
Corey, a student at Evers Elementary, stated:
Try to put it [ideas within the team] together. Because if you think one idea is good, that means
you just have to put it together because it ain’t gonna be right to just use one person’s. So – so, if
you like, we all have good ideas, and you could try to connect parts with each other and just try
to like work with it because sometimes, sometimes, people get real mad when you don’t use their
ideas, and they just go off and just get real mad and just say they don’t care. So, you should just
put it together just to make no commotion, or not commotion, but – arguing.
Analyzing Corey’s response, we initially highlight Corey’s acknowledgement of the team’s use of a combining
ideas approach, e.g., “you should just put it [ideas within the team] together.” Consistent with positioning this
approach as a reactive coping strategy with Jata’s team, we argue that Corey’s team also utilized this approach in
response to potential risks for participating in collaborative reflective decision-making. Two risks that emerged from
Corey’s response were the potential for (a) unfairness within the team when making collective design decisions and
(b) marginalizing teammates from the decision-making process. Corey calls attention to his desire to avoid potential
unfairness within the team as he stated, “it ain’t gonna be right to just use one person’s …” Although the objective
of participating in collaborative reflective decision-making is to generate an optimal design solution to the
engineering design problem, Corey incorporated the additional objective of maintaining a positive team atmosphere
by avoiding “commotion or arguing....” In addition, Corey recognized that this potential for “unfairness” while
making team decisions could possibly marginalize his teammates’ participation in the design process, as he stated,
“… sometimes people get real mad when you don’t use their ideas. And they just go off.…” Corey identified the risk
of maintaining a positive team environment within the design team and utilized the approach of combining ideas in
response to these risks.
To further highlight the approach of combining ideas as a reactive coping strategy and the perceived risks for
participating in collaborative reflective decision-making, we call attention to Toren’s response when posed with the
same question presented to Corey. Toren stated:
Pretty much what we did, we can come up with our own ideas and then we talked together as a
group and see if we can find one good; like the best idea we can out of all ideas combined. Well,
some people think differently and you combine three different things and made something very
good or something not good.
Analyzing Toren’s response, we initially call attention to his acknowledgment of incorporating a combining ideas
approach as he stated that they “combined three different things.” Toren and his teammates were not confident in
the development of an optimal solution by combining ideas, as evident by his concluding statement, “made
something very good or something not good.” The research team interpreted Toren’s acknowledgment of combining
ideas as reacting to the risk that was previously articulated by Corey; the risk of maintaining positive team
relationships.
The previous two sections have highlighted an approach of combining ideas that student teams utilized for
participating in collaborative reflective decision-making. This approach was positioned as a reactive coping strategy
(McGee & Spencer, 2013; Spencer et al., 1997; Spencer et al., 2001) that students utilized to participate in critique,
argumentation, and debate while also minimizing the risks that were associated with these practices. By combining
ideas, students reduced the chances of being improperly reprimanded for arguing or fighting or marginalizing
teammates.
Discussion
What it meant to be Doers of Engineering at Evers and Hamer Elementary
Focused on supporting students’ engagement in collaborative reflective decision-making, the research team
conceived the notion of “doers of engineering” to include students who effectively participated in the practices of
critique, argumentation, and negotiation. Specifically, in these classrooms, we envisioned doers of engineering as
students who utilized these discursive practices during the process of developing team design solutions. From this
perspective, we did not define student “success” through a lens of the development of “right answers,” but through a
lens that valued the various ways in which teams developed productive processes for engaging in collective
problem-solving. Previously described findings illustrate that students’ developed contrasting conceptions of “doers
of engineering” that revolved around their ability to avoid trouble. In these classrooms, from the student perspective,
being a doer of engineering meant being able to collectively develop a team solution while also avoiding the
potential risks of being reprimanded by the teacher or alienating teammates. Similar to findings from Kittleson and
Southerland (2004), we argue that “it is important to recognize that it is not just contextual features such as objects
and other group participants that shape knowledge, but also seemingly invisible factors such as assumptions and
ideologies …” (p. 269). The issue for the current study is that these factors affected student engagement in
collaborative reflective decision-making, and to consider the development of these invisible factors.
This study conceptualized emergent engineering identities as the ways in which students viewed themselves, as well
as their peers, as doers of engineering as it was realized within the specific contexts of their engineering classrooms.
While participating teachers, Ms. Humphrey and Ms. Simpson, continually modeled varied ways for engaging in
collaborative reflective decision-making and encouraged team discussions during design phases, these engineering
classrooms also lived within a larger school contexts that could have informed students’ perceptions of “smart
students.” Evers and Hamer Elementary were both considered schools with “urban characteristics” (Milner, 2012),
where students were often subjected to teaching acts that included giving directions, asking questions for right
answers, settling disputes, and punishing noncompliance (Haberman, 2010). This is important to note because while
students were provided opportunities to co-construct approaches for engaging in critique, argumentation, and
negotiation, these opportunities were limited to a total of 18 hours at Evers and 12 at Hamer. We contend that as
students went through a process of making sense what it meant to be a doer of engineering, their conceptions were
also informed by previous experiences at Evers and Hamer that privileged compliance and docility.
How Students Responded to Perceived Risks Associated with Participating in Collaborative Reflective
Decision-Making
Collaborative reflective decision-making was envisioned as a disciplinary practice where students would rely on
instances of argumentation and critique in the efforts of making informed engineering design decisions. Although
the research team identified positive beginnings of student engagement in this disciplinary practice (Wendell,
Wright, & Paugh, 2017), students also exhibited apprehension about collectively engaging in these discursive
practices. Specifically, students frequently utilized an approach that we refer to as “combining ideas” that included
students combining several ideas within a team while limiting the discussion around the pros and cons of these
ideas. We contend that the approach of combining ideas served as both a “facilitating and inhibitory role” (Kittleson
& Southerland, 2004, p. 267) related to the teams’ decision-making interactions. Conceptualized as a reactive coping
strategy (McGee & Spencer, 2013; Spencer et al., 1997; Spencer et al., 2001), a combining ideas approach was co-
constructed to provide students opportunities to engage in the decision-making process while minimizing the
potential risks of being reprimanded by a teacher or marginalizing a teammate. Additionally, within classrooms
where students perceived the importance of developing and producing “right answers,” a combining ideas approach
also served the purpose of distributing the potential for developing “wrong answers” throughout the design team.
From this perspective, the approach of combining ideas facilitated team decision-making interactions by reducing
the social and intellectual risks that students associated with collaborative reflective decision-making. In contrast, a
combining ideas approach also served an inhibitory role by minimizing students’ use of argumentation and critique
during the decision-making process. Believing that “reasoning and argumentation are needed to identify the best
solution to a design problem” (NGSS Lead States, 2013, p. 13), we contend that a combining ideas approach was not
very productive for students developing optimal design decisions. While Abigail, Cynthia, and Jata were able to
provide a design solution for the water filter challenge (see Figure 1), they could not articulate some of the
intricacies of the design, such as the physical properties of the cotton balls that could potentially contribute to a
successful design. We contend that a discussion around the pros and cons of these materials, or engaging in
collaborative reflective decision-making, could have positively contributed to students’ final designs and contributed
to their developing important disciplinary literacies (Garcia & Mazzotti, 2016). In conclusion, we would like to
clarify that we are not arguing that a combining ideas approach would be the expectation within urban engineering
classrooms. On the contrary, we contend that this approach was co-constructed by students in response to the
specific risks that students perceived within these two classrooms. The larger contention is for the need to recognize
that there may be perceived risks for engaging in engineering practices, and for researchers and educators to
acknowledge as reactive coping strategies.
The Risks that Students Associated with Engaging in Collaborative Reflective Decision-Making
Analyzing students’ participation in collaborative reflective decision-making through a phenomenological variant of

the ecological systems theory lens presented the opportunity to identify potential vulnerabilities for engaging in
group interactions. While students, in general, could face vulnerabilities with engaging in argumentation and
critique, we contend that “the nature of the vulnerabilities that human individuals and human communities face are
clearly differentiated by an array of societal positionings, particularly with regard to race, ethnicity, class, gender,
sexual orientation, and constructions of ability” (Lee, 2017, p. 262). Specifically, we found that the intellectual and
behavioral stigmas often associated with students attending schools within urban communities impacted the ways in
which students conceptualized their participation in this disciplinary practice. Findings from student interviews
suggested that students aimed to minimize the following risks: (a) being reprimanded for behavior deemed
inappropriate by the classroom teacher, (b) marginalizing teammates from the decision-making process, and (c)
having proposed design solutions evaluated as a “wrong answer.” Despite these findings, we do not claim that these
risks will appear within every engineering classroom in an urban school. However, we are calling attention to the
need for engineering education researchers and educators, alike, to consider the role of context when examining
student participation. This study by acknowledged the potential vulnerabilities and conceptualized students’
interactions as reactive coping strategies.
Recommendations
A goal of this study was to explore elementary students’ conceptions of “doers of engineering” in relation to their
participation in a disciplinary practice referred to as collaborative reflective decision-making. These findings call
attention to the importance of context when considering participation in engineering design practices, and the
essential role that teachers play in supporting the development of productive learning environments. Although
participating students maintained positive views of themselves as doers of engineering, their justifications for those
views were absent of any references to engineering or engineering practices. Here, we recommend that engineering
educators devote the necessary time for co-constructing a shared understanding of “success” and “competence”
within their engineering classrooms.
The educator’s role in developing these shared understandings, or disciplinary norms within their classrooms, is
important in students’ development. McClain and Cobb (2001) examined the teacher’s role in renegotiating the
sociomathematical norm of mathematical difference in a first grade classroom. Through the explicit solicitation of
different solution processes during whole-class discussions, the teacher contributed to the class’ development of
criteria used to judge a variety of mathematical solutions. In another study, Andrews (2017) uncovered a teacher’s
role in developing norms around testing and failure within an elementary engineering classroom. The teacher’s
moves included continued references to a poster, or anchor chart, about the importance of testing early and often,
and numerous questions asking students to share why they made the changes they made. Success and competence in
varied elementary engineering contexts can look different than other disciplines (e.g., mathematics and language
arts), and explicit discussions about these specific competencies could potentially contribute to different conceptions
of “doers of engineering.”
Additionally, we recommend that educators examine the potential for engaging students in engineering design
practices, specifically argumentation and critique, in culturally sustaining ways (Alim & Paris, 2017). Culturally
sustaining pedagogy “seeks to perpetuate and foster – to sustain – linguistic, literate, and cultural pluralism as part of
schooling for positive social transformation” (p. 1). Such an approach would require educators to attune their
attention to their students’ heterogeneous linguistic and literacy practices, and recognize their potential for engaging
students in engineering design practices. For instance, Wilson-Lopez, Mejia, Hasbun, and Kasun (2016) found that
Latina/o funds of knowledge mapped onto the application of engineering design processes, and could be
reconceptualized as cultural practices to be fostered and sustained within K-12 engineering classrooms. Additional
research highlighting students’ cultural engagement in argumentation (Hudicourt-Barnes, 2003) and critique
(Wright, 2016) in science learning environments could inform the implementation of these practices within
engineering classrooms.
Acknowledgements
The National Science Foundation, under grants 1411660 and 1316762, supported this work. We are grateful to our
collaborating teachers and their students and to the Engineering is Elementary team.
References
Aguirre-Muñoz, Z., & Pantoya, M. L. (2016). Engineering literacy and engagement in kindergarten classrooms.
Journal of Engineering Education, 105(4), 630-654.
Alim, H. S., & Paris, D. (2017). What is culturally sustaining pedagogy and why does it matter? In D. Paris & H. S.
Alim (Eds.), Culturally sustaining pedagogies: Teaching and learning for justice in a changing world (pp. 1-
21). New York, NY: Teachers College Press.
Andrews, C. (2017). Elementary students’ engagement in failure-prone engineering design tasks. Unpublished
doctoral dissertation, Tufts University, Medford.
Berland, L. K., & McNeill. K. L. (2010). A learning progression for scientific argumentation: Understanding student
work and designing supportive instructional contexts. Science Education, 94(5), 765-793.
Capobianco, B. M., Diefes-Dux, H. A., Mena, I., & Weller, J. (2011). What is an engineer? Implications of
elementary school student conceptions for engineering education, Journal of Engineering Education, 100(2),
304-328.
Capobianco, B. M., Ji, H. Y., & French, B. F. (2015). Effects of engineering design-based science on elementary
school science students’ engineering identity development across gender and grade. Research in Science
Education, 45(2), 275-292.
Cavagnetto, A., Hand, B. M., & Norton-Meier, L. (2010). The nature of elementary student science discourse in the
context of the science writing heuristic approach. International Journal of Science Education, 32(4), 427-
449.
Chen, W., & Wassenaar, H. J. (2003). An approach to decision-based design with discrete choice analysis for
demand modeling. Journal of Mechanical Design, 125(3), 490-497.
Cunningham, C. M. (2009). Engineering is elementary. The Bridge, 30(3), 11-17
Cunningham, C. M., & Lachapelle, C. P. (2014). Designing engineering experiences to engage all students. In S.
Purzer, J. Strobel, & M. Cardella (Eds.), Engineering in pre-college settings: Synthesizing research, policy,
and practices (pp. 117-142). Lafayette, IN: Purdue University Press.
Denson, C. D., Avery, Z. K., & Schell, J. W. (2010). Critical inquiry into urban African American students’
perceptions of engineering. Journal of African American Studies, 14(1), 61-74.
Duschl, R. A., & Osborne, J. (2002). Supporting and promoting argumentation discourse in science education.
Studies in Science Education, 38(1), 39-72.
Emdin, C. (2010). Chapter 7: What is urban science education? Counterpoints, 215, 101-111.
Emdin, C. (2016). For white folks who teach in the hood… and the rest of y’all too: Reality pedagogy and urban
education. Boston, MA: Beacon Press.
Ferguson, A. A. (2000). Bad boys: Public schools in the making of Black masculinity. Ann Arbor, MI: University of
Michigan Press.
Frehill, L. M. (2004). The gendered construction of the engineering profession in the United States, 1893-1920. Men
and Masculinities, 6(4), 383-403.
Gandy, A., Jager, P., Bertsche, B., & Jensen, U. (2007). Decision support in early development phases – A case
study from machine engineering. Reliability Engineering & System Safety, 92(7), 921-929.
Garcia, A. J., & Mazzotti, T. B. (2016). Argumentation in engineering education. Paper presented at the Canadian
Engineering Education Association. https://queens.scholarsportal.info/ojs-
archive/index.php/PCEEA/article/view/6483.
Gee, J. (1999). Critical issues: Reading and the new literacy studies: Reframing the National Academy of Sciences
Report on reading. Journal of Literacy Research, 31(3), 355-374.
Haberman, M. (2010). The pedagogy of poverty versus good teaching. Phi Delta Kappan, 92(2), 81-87.
Hampton, B., Peng, L., & Ann, J. (2008). Pre-service teachers’ perceptions of urban schools. The Urban Review,
40(3), 268-295.
Hazelrigg, G. A. (1998). A framework for decision-based engineering design. Journal of Mechanical Design,
120(4), 653-658.
Hegedus, T. A., Carlone, H. B., & Carter, A. D. (2014). Shifts in the cultural production of “smartness” through
engineering classrooms. In Proceedings of the Annual Meeting of the American Society of Engineering
Education, Indianapolis, IN.
Hudicourt-Barnes, J. (2003). The use of argumentation in Haitian Creole science classrooms. Harvard Educational
Review, 73(1), 73-93.
Jankovic, M., Stal-Le Cardinal, J., & Bocquet, J. C. (2010). Collaborative decision-making in design project
management: A particular focus on automotive industry. Journal of Decision Systems, 19(1), 93-116.
Jiménez-Aleixandre, M. P., & Erduran, S. (2008). Argumentation in science education: An overview. In S. Erduran
& M. P. Jiménez-Aleixandre (Eds.), Argumentation in science education: Perspectives from classroom-based
research (pp. 3-28). Dordre-cht: Springer.
Jordan, M. E., & McDaniel, R. R. (2014). Managing uncertainty during collaborative problem solving in elementary
school teams: The role of peer influence in robotics engineering activity. Journal of the Learning Sciences,
23(4), 490-536.
Kittleson, J. M., & Southerland, S. A. (2004). The role of discourse in group knowledge construction: A case study
of engineering students. Journal of Research in Science Teaching, 41(3), 267-293.
Knoblauch, D., & Hoy, A. W. (2008). “Maybe I can teach those kids:” The influence of contextual factors on
student teachers’ efficacy beliefs. Teaching and Teacher Education, 24(1), 166-179.
Larkin, M., Watts, S, & Clifton, E. (2006). Giving voice and making sense in interpretive phenomenological
analysis. Qualitative Research in Psychology, 3(2), 102-120.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge: Cambridge
University Press.
Leander, K. M. (2002). Locating Latanya: “The situated production of identity artifacts in classroom interaction.”
Research in the Teaching of English, 37(2), 198-250.
Lee, C. D. (2017). An ecological framework for enacting culturally sustaining pedagogy. In D. Paris & H. S. Alim
(Eds.), Culturally sustaining pedagogies: Teaching and learning for justice in a changing world (pp. 261-
273). New York, NY: Teachers College Press.
Lee, C. D., Spencer, M. B., & Harpalani, V. (2003). “Every shut eye ain’t sleep:” Studying how people live
culturally. Educational Researcher, 32(5), 6-13.
Lottero-Perdue, P., Bowditch, M., Kagan, M., Robinson-Cheek, L., Webb, T., Meller, M., & Nosek, T. (2016). An
engineering design process for early childhood: Trying (again) to engineer an egg package. Science and
Children, 54(3), 70-77.
Lynn, M., Bacon, J. N., Totten, T. L., Bridges, T., & Jennings, M. (2010). Examining teachers’ beliefs about African
American male students in a low-performing high school in an African American school district. Teachers
College Record, 112(1), 289-330.
Marshall, C., & Rossman, G. B. (2015). Designing qualitative research – 6th edition. Thousand Oaks, CA: Sage
Publishing.
Martin, A. M., & Hand, B. (2009). Factors affecting the implementation of argument in the elementary science
classroom: A longitudinal case study. Research in Science Education, 39(1), 17-38.
Martin, D. B., & Larnell, G. V. (2014). Urban mathematics education. In H. R. Milner & K. Lomotey (Eds.),
Handbook of Urban Education (pp. 373-393). New York, NY: Routledge.
Marulcu, I., & Barnett, M. (2013). Fifth graders’ learning about simple machines through engineering design-based
instruction using LEGO materials. Research in Science Education, 43(5), 1825-1850.
McClain, K., & Cobb, P. (2001). An analysis of development of sociomathematical norms in one first-grade
classroom. Journal of Research in Mathematics Education, 32(3), 236-266.
McGee, E. O., & Martin, D. B. (2011). “You would not believe what I have to go through to prove my intellectual
value!” Stereotype management among academically successful Black mathematics and engineering
students. American Education Research Journal, 48(6), 1347-1389.
McGee, E. O., & Spencer, M. B. (2013). The development of coping skills for science, technology, engineering, and
mathematics students: Transitioning from minority to minority environments. Urban ills: Post recession
complexities of urban living in the twenty first century, 351-378. Retrieved from
http://repository.upenn.edu/ges_pubs/265.
McNeill, K. L. (2011). Elementary students’ views of explanation, argumentation, and evidence, and their abilities
to construct arguments over the school year. Journal of Research in Science Teaching, 48(7), 793-823.
Mehalik, M. M., Doppelt, Y., & Schunn, C. D. (2008). Middle-school science through design-based learning versus
scripted inquiry: Better overall science concept learning and equity gap reduction. Journal of Engineering
Education, 97(1), 71-85.
Milner, H. R. (2012). But what is urban education? Urban Education, 47(3), 556-561.
Moore III, J. L., Madison-Colmore, O., & Smith, D. M. (2003). The prove-them-wrong syndrome: Voices from
unheard African American males in engineering disciplines. The Journal of Men’s Studies, 12(1), 61-73.
National Research Council. (2012). A framework for K-12 science education: Practices, crosscutting concepts, and
core ideas. Washington, DC: The National Academies Press. https://doi.org/10.17226/13165
NGSS Lead States. (2013). Next Generation Science Standards: For States, By States. Washington, DC: The
National Academies Press.
Osborne, J., Erduran, S., & Simon, S. (2004). Enhancing the quality of argumentation in school science. Journal of
Research in Science Teaching, 41(10), 994-1020.
Purzer, S. (2011). The relationship between team discourse, self-efficacy, and individual achievement: A sequential
mixed-methods study. Journal of Engineering Education, 100(4), 655-679.
Quinn, C., & Clare, L. (2008). Interpretive phenomenological analysis. Nursing research: Design and methods, 375-
384.
Richmond, G., & Striley, J. (1996). Making meaning in classrooms: Social processes in small-group discourse and
scientific knowledge building. Journal of Research in Science Teaching, 33(8), 839-858.
Rivera-Maulucci, M. S. (2010). Resisting the marginalization of science in an urban school: Coactivating social,
cultural, material, and strategic resources. Journal of Research in Science Teaching, 47(7), 840-860.
Ryu, S., & Sandoval, W. A. (2012). Improvements to elementary children’s epistemic understanding from sustained
argumentation. Science Education, 96(3), 488-526.
Silk, E. M., Schunn, C. D., & Cary, M. S. (2009). The impact of an engineering design curriculum on science
reasoning in an urban setting: Journal of Science Education and Technology, 18(3), 209-223.
Siwatu, K. O. (2011). Preservice teachers’ sense of preparedness and self-efficacy to teach in America’s urban and
suburban schools: Does context matter? Teacher and Teacher Education, 27(2), 357-365.
Smith, J. A. (2004). Reflecting on the development of interpretative phenomenological analysis and its contribution
to qualitative research in psychology. Qualitative Research in Psychology, 1(1), 39-54.
Spencer, M. B. (2006). Phenomenology and ecological systems theory: Development of diverse groups. In W.
Damon & R. M. Lerner (Eds.), Child’s and Adolescent Development: An Advanced Course (pp. 696-735).
Hoboken, NJ: John Wiley & Sons, Inc.
Spencer, M. B. (1995). Old issues and new theorizing about African American youth: A phenomenological variant
of ecological systems theory. In R. L. Taylor (Eds.), Black youth: Perspectives on their status in the United
States (pp. 37-70). Westport, CT: Praeger.
Spencer, M. B., Dupree, D., & Hartmann, T. (1997). A phenomenological variant of ecological systems theory
(PVEST): A self-organization perspective in context. Development and Psycopathology, 9(4), 817-833.
Spencer, M. B., Noll, E., Stoltzfus, J., & Harpalani, V. (2001). Identity and school adjustment: Revisiting the “acting
white” assumption. Educational Psychologist, 36(1), 21-30.
Stake, R. E. (1995). The art of case study research. Thousand Oaks, CA: Sage.
Swanson, D. P., Spencer, M. B., Harpalani, V., & Spencer, T. R. (2002). Identity processes and the positive youth
development of African Americans: An explanatory framework. New Directions for Student Leadership,
2002(95), 73-100.
Thadani, V., Cook, M. S., Griffis, K., Wise, J. A., & Blakey, A. (2010). The possibilities and limitations of
curriculum-based science inquiry interventions for challenging the “pedagogy of poverty.” Equity &
Excellence in Education, 43(1), 21-37.
Tonso, K. L. (2006). Teams that work: Campus culture, engineer identity, and social interactions. Journal of
van Zee, E. H., Hammer, D., Bell, M., Roy, P., & Peter, J. (2005). Learning and teaching science as inquiry: A case
study of elementary school teachers’ investigations of light. Science Education, 89(6), 1007-1042.
Watson, D. (2011). “Urban, but not too urban:” Unpacking teachers’ desires to teach urban students. Journal of
Teacher Education, 62(1), 23-34.
Wendell, K. B., Wright, C. G., & Paugh, P. P. (2017). Reflective decision-making in elementary students’
engineering design. Journal of Engineering Education, 106(3), 356-397.
Wilson-Lopez, A., Mejia, J. A., Hasbún, I. M., & Kasun, G. S. (2016). Latina/o adolescents’ funds of knowledge
related to engineering. Journal of Engineering Education, 105(2), 278-311.
Wright, C. G. (2016). Constructing a collaborative critique-learning environment for exploring science through
improvisational performance. Urban Education, 1-30. https://doi.org/10.1177/0042085916646626
Yin, R. K. (2011). Applications of case study. Thousand Oaks, CA: Sage.
Author Information
Christopher G. Wright Kristen B. Wendell
Drexel University Tufts University
School of Education School of Engineering
3401 Market Street Science & Engineering Complex, Robinson Hall
Philadelphia, PA 19104 Medford, MA 02155
U.S.A. U.S.A.
Contact e-mail: cgw57@drexel.edu
Patricia P. Paugh
University of Massachusetts - Boston
College of Education & Human Development
Wheatley Hall
Boston, MA 02125
U.S.A.
Exploring Moments of Agency for Girls during an Engineering Activity

Gina Navoa Svarovsky, Catherine Wagner, Monica Cardella

Article History The persistent underrepresentation of women in engineering continues to be a
complex and difficult challenge. The interactions of young women and their
Received: parents during early, family-oriented engineering design experiences can provide
21 July 2017 girls with opportunities to express agency during an engineering activity, which
can ultimately contribute to the development of sustained interest and self-
Accepted:
efficacy in engineering. However, few studies have examined these parent-child
22 January 2018
interactions to date, and none have specifically focused on moments when girls
express agency during an engineering design process. In this paper, we examine
Keywords
one such setting: a museum exhibit that engages visitors in engineering design
Informal engineering activity. A qualitative content analysis was performed on transcripts from a total
education of 39 family groups videotaped at the exhibit, each involving a daughter between
Adult-child interactions the ages of 5-12 and at least one parent. Qualitative codes describing the ways
Young women and children expressed agency and led interactions with their parents included
engineering directing, proposing design ideas, and asking questions. Interestingly, the
analysis also suggests that the young women in this study tended to direct their
mothers more than their fathers. Although focused specifically on parent-child
interactions, this study can inform both formal and informal engineering
educators who engage young students in engineering activities.
Introduction
Ongoing efforts to recruit and retain women in STEM fields within the United States have had mixed success
over the past three decades (Hill, Corbett, & St. Rose, 2010). While some areas of STEM, such as those
connected to the life sciences and experimental psychology, have seen substantial increases in the percentages
of women entering and completing undergraduate programs, other fields such as engineering, computer science,
and physics have actually seen decreases over the past ten years in the overall percentage of women entering
those programs (National Science Foundation, 2017).
Specifically within the field of engineering, there are several theories that attempt to shed light on the low
numbers of women. Traditionally, engineering education researchers relied on the “chilly climate” and “leaky
pipeline” metaphors to describe potential explanations for the dearth of women in engineering (Hoegh &
Pawley, 2010). More recent studies, however, have begun to move away from these models, recognizing the
need for a more complex, nuanced, and interconnected understanding of how and why women choose to enter or
exit engineering pathways at different points in their lives (Matusovich, Streveler, & Miller, 2010; Sheppard,
Atman, Stevens, et al., 2004; Trenor, Yu, Waight, et al., 2008).
To date, much of the research on recruiting and retaining women in engineering has focused on the retention of
college-aged females in engineering majors and programs (Hill, Corbett, & St. Rose, 2010; Marra et al., 2009;
Seymour, 1995; Seymour & Hewitt, 1997). While these studies at the undergraduate level are essential, an
increased focus on the engineering perceptions and experiences of young women prior to attending college is
necessary to more fully address the underrepresentation of women in engineering fields. Certainly over the past
two decades exposure to engineering at the pre-college level has become increasingly common, particularly with
the inclusion of engineering-focused practices within the Next Generation Science Standards and the rise in
popularity of curriculum packages targeting pre-college learners such as Engineering is Elementary (eie.org)
and Project Lead the Way (pltw.org). Moreover, early engineering and design experiences have long been a
possibility for young learners outside the classroom, within informal learning environments such as museums,
afterschool clubs, and libraries—and these opportunities only continue to increase, particularly with the rise of
the Maker Movement (Svarovsky, 2014). Informal learning opportunities are often very social in nature,
commonly engaging not only the young learner in an experience, but their peers, siblings, and parents as well
(Feder, Shouse, & Lewenstein, 2009). Given the ways that parents in particular increasingly have been shown to
impact career choices in STEM for women (Matusovich et al., 2010), examining these early engineering
experiences within informal learning environments may lead to a greater understanding of how young women
can begin to develop interest and confidence in fields like Engineering.
In this paper, we address this challenge by exploring one such engineering learning experience: an engineering
exhibit within a large Midwestern science center, where parents and children engage collaboratively on an
engineering design challenge. Specifically, this study investigates whether and how young women demonstrate
agency during these design experiences, with the goal of advancing the understanding of what conditions can
promote the engagement, confidence, and persistence of young women in engineering specifically and in STEM
overall.
Background and Theoretical Framework

Persistence in STEM, Agency, and Design Practice
The construct of persistence in undergraduate STEM programs has been explored in multiple disciplines. For
example, a comparative study following seven women showed the complexity of STEM persistence in the male-
dominated STEM fields, finding that ways of persisting vary from situation to situation and that there is not a
single, universal solution that women pursue to persist in these fields (Hughes, 2011). Another study found that
factors related to successfully increasing persistence for students in STEM fields were to involve them in the
“doing” of STEM, help them to create a “growth mindset,” build their STEM identity, and foster a sense of
belonging in the STEM community through a climate that favors peer interaction (Metevier et al., 2015). This
work on persistence draws on several bodies of literature, as evidenced by the persistence framework advanced
by Graham and colleagues that links persistence to learning theory, confidence, and motivational theory, and
theories of identity development (Graham, Frederick, Byars-Winston, et al., 2013). However, while many
studies have examined persistence at the undergraduate level, fewer studies have been conducted on persistence
in STEM for pre-college students. Unlike at the undergraduate level, where persistence can be measured by the
extent to which college students enter, stay in, and complete specific degree programs, studies of persistence at
the pre-college level often focus on exploring the factors and conditions that lead to STEM-related motivation,
achievement, and identity development (Banks, McQuater, & Hubbard, 1978; Calabrese Barton & Tan, 2010;
Singh, Granville, & Dika, 2002).
One construct that has been helpful in discussions of STEM persistence at the pre-college level is the idea of
agency. Drawing on the empowerment and agency indicators identified by Ibrahim and Alkire (2007), the
definition of agency in this study is articulated as an individual’s ability to assert control or choice within a
given interaction. Developing a sense of agency around a particular topic has been shown to be connected to
positive identity development and increased self-efficacy in a range of domains, including STEM (see, for
example, Calabrese Barton & Tan, 2010, and Carlone, 2004). In some ways, the nature of engineering design
may be a particularly fruitful context to explore opportunities for the expression of agency, given the structure
of engineering as a discipline and the types of practices involved with the engineering design process.
Engineering design is generally oriented toward creating a solution to meet or exceed a set of design goals, often
under specific constraints (Crismond & Adams, 2012). During the engineering design process, an engineer
demonstrates agency throughout, such as when advancing a particular design idea, making a decision about a
type of material to use or the placement of a particular structure, or choosing to prioritize certain design
constraints in response to client need (Dym et al., 2009). The nature of engineering design activities may afford
young learners several opportunities to express agency by asserting their own ideas and decisions around the
specific design being developed in the activity.
Expectancy-Value Theory and Persistence in STEM
Of course, engaging in engineering and design activities rarely happens in isolation, and therefore the broader
contexts in which pre-college learners encounter and potentially develop persistence in engineering should be
considered. A useful tool for examining these ideas is the expectancy-value theory (Wigfield & Eccles, 2000),
which suggests that young people make choices about their careers by considering how successful they believe
they will be at a given profession as well as how much they care about that profession. The creation of an
expectancy-value model for a given context further highlights the connection between expectancies, values, and
304 Svarovsky, Wagner, & Cardella
motivation, particularly around constructs of performance, effort, and persistence (Wigfield & Eccles, 2000). In
other words, the alignment between what people believe they can do, what they value and are interested in
doing, and their choices about what to do all come together to impact decision making and persistence in a given
course of action—and in particular decisions about careers and college choices (Eccles, 2005; Eccles, Barber, &
Josefowicz, 1999). A recent study applied the expectancy-value theory to their discussion of the impact of
adolescent girls’ experiences and beliefs on their mathematics and science motivation (Leaper, Farkas, &
Brown, 2012). The study examined social and personal factors and how they relate to girls’ motivation in the
subjects of mathematics and science in comparison to the subject of English/language arts. The social factors
they examined were the perceived support in mathematics and science from parents and peers, while the
personal factors were gender identity and attitudes as well as exposure to feminism. Overall, they found that for
adolescent girls their mathematics and science motivation was influenced by mother and peer mathematics and
science support, as well as exposure to feminist and gender-egalitarian beliefs (Leaper et al., 2012).
The Role of Parents in Persistence in STEM
As illustrated by the study by Leaper and colleagues described above, parents have been shown to play a crucial
role in children’s achievement. One study investigating the effects of parents on their child’s achievement found
that parents’ motivational practices positively influence their children’s achievement in mathematics regarding
where they start in seventh grade and how much they learn through the twelfth grade (Ing, 2014). These
motivational practices were also shown to have positive effects on the children’s later STEM careers. However,
not every type of parental motivational practice was shown to influence children’s mathematics achievement or
persistence; instead, only the “mathematics-specific, intrinsically focused parental motivational practices,” as
opposed to extrinsically-focused practices, resulted in significant influences on persistence and achievement in
STEM careers (Ing, 2014). Overall, this study shows the positive impact parents can have on their child’s
achievement, specifically by their motivational practices.
Certainly, the role that parents play in STEM persistence is essential, but it seems that this is particularly the
case within the field of engineering. Several studies have suggested that both girls and their parents often
underestimate their abilities in STEM (Frome et al., 2006; Herbert & Stipek, 2005; Lloyd, Walsh, & Yailagh,
2005; Tenenbaum, 2009; Voyles & Williams, 2004), which can lead to inaccurate beliefs about the likelihood of
success within engineering fields. On the other hand, recent research has highlighted the crucial part that parents
play in women’s decisions to enter an engineering field, with one study suggesting that female engineers are
significantly more likely than male engineers to have an engineer as a parent (Mannon & Schreuders, 2007).
Another study identifies parents as influential figures who contributed to the students’ decisions to major in
engineering (Matusovich et al., 2010). However, few studies have been done specifically on girls’ interactions
with their parents while engaging in engineering design activities (Cardella, Svarovsky, & Dorie, 2013; Dorie,
Cardella, & Svarovsky, 2014, 2015; Svarovsky et al., 2017). Understanding how these types of interactions
occur and how girls can begin to demonstrate signs of agency during these experiences can inform how adults in
general engage young women in engineering activities in productive ways.
The Gender Research on Adult-child Discussions within Informal ENgineering environmenTs (GRADIENT)
study investigates interactions between parents and young women during a range of engineering activities
within the informal learning context of a science museum. In particular, the present analysis explores whether
and how young girls express agency (Calabrese Barton & Tan, 2010) during collaborative engineering activities
with a parent by asserting some level of control (Ibrahim & Alkire, 2007) during the interaction, such as leading
portions of the activity with their parent. This study addresses the following research questions:
1) Do young women demonstrate agency (as demonstrated by leading interactions) during engineering
activities with a parent?
2) If so, what are the most common ways that young women lead interactions, and what are the most
common ways that parents respond?
3) What patterns or relationships, if any, appear between the different ways that young women lead
interactions and the ways that parents respond?
4) What differences, if any, exist between the ways young women interact with their mothers and their
fathers?
Method
Research Context
To address the research questions posed above, a subset of video data from the larger GRADIENT Study
(Cardella et al., 2013; Dorie et al., 2014, 2015; Svarovsky et al., 2017) was acquired and analyzed. The analysis
for this study focused on data collected at an exhibit called the Pneumatic Ball Run, as seen in Figures 1a and 1b
below.
(a) (b)
Figure 1. The pneumatic ball run component, as part of the Engineering Studio exhibit.
This exhibit is a component within a larger area of the museum called the Engineering Studio, which includes
several interactive exhibits that invite visitors to engage in design activities. The Pneumatic Ball Run presents
visitors with the design challenge of getting a small ball from the “Start” position on the left side of the exhibit
to the “Finish” position on the right side of the exhibit. Visitors can use any of the materials within the exhibit,
including simple, hand-operated pneumatic pistons, ramps, and cardboard tubes, to build a system that moves
the ball across the face of the exhibit. Interestingly, the “Finish” position is higher than the “Start” position, thus
making the design challenge more difficult and requiring visitors to consider both the materials and sequencing
of pistons to propel the ball to a greater height. The Pneumatic Ball Run provided an engaging and multi-faceted
engineering design experience for visitors on the museum floor.
Description of Sample
Overall, the GRADIENT study focused specifically on exploring interactions between young women and their
parents during engineering activities. A purposeful sampling technique recruiting families with a young female
child on the museum floor was used throughout the study (Patton, 2002). The data collected for the broader
GRADIENT study included an emphasis on family interactions for girls at two age levels: pre-school (ages 3-5)
(Cardella et al., 2013; Dorie et al., 2014, 2015; Svarovsky et al., 2017), and elementary/middle. For this study,
the analysis focused on analyzing the video data from the elementary/middle school girls and their families,
with a total of 39 daughter-parent groups included. Families were recruited to participate as parent-child dyads;
however, at times other family members—such as a second parent or additional siblings—were also present.
The videos analyzed for this study focused on female children engaging in the Pneumatic Ball Run with a
parent. The female children ranged in age from 5 to 12, with median age of 10. In 20 cases, the primary parent
participant in the study was male. Females were the primary parent participant in 13 cases, and 6 cases involved
both parents. Group size varied from 2 to 6 people, with 19 cases involving parent-child dyads. The remaining
20 cases involved an additional parent or siblings during the design activity. Within this sample of 39 families,
10% of primary parent participants identified as African-American, 6% as Native American, 2% as
Hispanic/Latino, 2% as Asian, and 78% as White/Caucasian. The remaining 2% identified as “Other
race/ethnicity.” This demographic distribution is typical of museum visitors, as described in field-wide studies
(Farrell & Medvedeva, 2010).
Segmentation of Data
Interactions between the parent and child participating in the engineering design activity were videotaped and
transcribed, with 39 cases provided for the current analysis. Although 49 cases were collected for the
GRADIENT Study, 6 of these were identified as pilot cases used for developing data collection methods and
training data collection staff and 4 videos were removed from the dataset because they were deemed unusable
by the original research team. The remaining 39 videos and transcripts were examined initially to identify
patterns in the interactions between parents and daughters. Focused on exploring moments of agency, our units
of analysis for the study were comprised of “child-led interactions,” which we defined as any verbal interaction
that the daughter initiates with her parent(s). Operationally, these child-led interactions began with a question or
statement that the child initiated, followed by a parent’s response (or lack thereof). These units of analysis were
then compiled and coded using Dedoose, an online qualitative analysis program. In addition, it should be noted
that a parent’s response to child-led interaction did not have to occur immediately following the child’s catalyst
remark. In some cases following a short period of quiet building the parent responded in regard to the
aforementioned topic or question.
Description of Coding Schemes
Child-led Interactions Coding Protocol
Each turn of talk within the child-led interactions was coded inductively for different types of interaction. Initial
ideas for coding these interactions built on earlier coding schemes developed in the GRADIENT study, which
also focused on turn-of-talk level interactions between adults and children (Cardella et al., 2013; Dorie et al.,
2014, 2015; Svarovsky et al., 2017). This process led to two sets of refined codes: one set that focused on what
the young women were doing during the interactions they were leading, and the other on how the parent(s) were
responding to the child. The Child Initiation codes describe the different ways children led interactions with
their parents. These codes included design idea promotion, direct, question, statement of problem, or other, as
described in Table 1. For clarity, it should be noted that throughout the transcription and analysis conducted for
this work participants were labeled as “CF” for Child, Female; “AM” for Adult, Male; and “AF” as Adult,
Female.
Table 1. Child initiation codes for pneumatic ball run child-led interactions
Child Code Abbreviation Definition Example
Design Idea DIP Proposing an idea or suggestion CF: "Oh yeah, we could
Promotion for the ball run make it; we could just make
it lower" (Case 37, 7:44)
Direct D Telling parents what to do or CF: "Dad, you have to try to

explaining how something works get it in there. And Dad, this
can do this" (Case 16, 14:49)
Question – Child QC Asking a parent a question, such CF: "So, wait a minute, the
as for clarification or if the adult goal is to get it here, right?"
is ready to start a new trial (Case 32, 1:51)
Statement of SP Commenting on a problem with CF: "Dad, it can't get to the

Problem the design or a piece not working finish line." (Case 33, 21:01)
Other – Child OC Any other comment such as CF: "We can do this!"
encouragement (Case 33, 17:20)
Parent Response Coding Protocol
The Parent Response codes describe the different ways that parents responded to their children during the child-
led interactions. These codes included agreement, question, explanation, suggestion, non-response, or other, as
described in Table 2.
Table 2. Parent response codes for pneumatic ball run child-led interactions
Parent Code Abbreviation Definition Example
Agreement AG Agreeing with child’s proposed CF: "Oh yeah, we could

idea/instructions or replying in the make it; we could just make
affirmative it lower"
AM: "Right" (Case 37, 7:47)
Question – QA Asking child a question; however, CF: “No, I need your help.”
Adult if it is more explaining than a real AF: “Okay, what do you
question, it is coded as S need?”
(Case 15, 31:25)
Explanation E Explaining a design piece or CF: "We should put that

concept to child or giving an over there, maybe, so once it
answer to what child proposed can make it over"
AF: "We could. But if we
put it here, it will lift it up a
lot higher.”
(Case 41, 8:00)
Suggestion S Suggesting an idea or command to CF: "Okay. Are you ready?"

child AM: "Wait, you gotta start
there though" (Case 37,
6:13)
Non-Response NR No verbal response or saying CF: "Dad, you have to try to

something unrelated get it in there. And Dad, this
can do this"
AM: [no verbal response]
(Case 16, 14:49)
Other – Adult OA Any other statement such as praise, AM: "Yeah, I see what
laughter, or disagreement you're saying. Yeah so it…
So there's not resistance? It's
a great idea"
(Case 38, 24:50)
Inter-rater Reliability
After the qualitative codes were identified and initially applied by a primary coder to 20% of the sample, a
second researcher conducted an inter-rater reliability (IRR) analysis and achieved a percent agreement level of
90% for the established child codes and 92% for the established parent codes. Differences were resolved by
discussing the discrepancies and refining code definitions. After the IRR process, the remaining units of analysis
were coded by the primary coder with these child and parent codes in the online coding software program,
Dedoose (available at Dedoose.com).
Data Analysis
Although this study is qualitative in nature, a small amount of quantitative correlational analysis was conducted
to further explore potential patterns of relationships among variables. Once the coding process was completed,
the data were exported into Excel and subsequently into SPSS, where a bivariate correlation matrix was
generated to identify any statistically significant relationships between each of the child and parent codes.
Theoretically meaningful correlations that were statistically significant—even if the strength of relationship was
quite small—were explored further in the qualitative data, which provided additional insight into the nature of
the interactions between young women and their parents within these contexts. To be clear, these correlations
are only meant to further understand the qualitative patterns in the coded data. They are not intended to be
predictive or broadly generalizable beyond the participants of this study (Shaffer & Serlin, 2004).
Results
Findings from the study are presented in four parts, with each part aligning with a particular Research Question.
Part One: Girls Lead Interactions with Parents during Engineering Activities
Research Question 1 asks whether girls demonstrate agency during family-based engineering activities. The data
suggest that girls in fact do express agency when interacting with their parents during these experiences. Across
the 39 cases examined, 5,798 of the 17,178 total turns of talk were identified as part of a child-led interaction,
which translates to an overall percentage of 33.8% of all possible turn of talk segments. These identified child-
led interactions then became the units of analysis for the subsequent coding of Child Initiation codes and the
Parent Response codes discussed below.
The percentage of how many segments within each case that were identified as part of a child-led interaction
varied, as shown in Figure 2. The majority of the cases had anywhere from 21% to 40% of the turns of talk
coded as child-led interactions. Two had 10% or fewer interactions coded as child-led, whereas only one had
more than 60% of the turns of talk coded as child-led interactions.
Frequency of Child-Led Interactions

12
10
Number of Cases
0
0-10% 11-20% 21-30% 31-40% 41-50% 51-60% 61-70%
Percent of Total Segments in Cases that are Included in Child-Led
Interactions
Figure 2. Frequency of child-led interactions
Part Two: Common Types of Child Initiation Moves and Parent Responses During Child-led Design
Activities
Research Question 2 asks about the most common ways that young women led interactions with their parents as
well as the most common ways that parents responded to their children. The Child Initiation codes described in
the Methods section above reflect the most common ways that children tended to demonstrate agency during
these exchanges. The distribution of the Child Initiation code frequencies can be seen in Figure 3. The “Direct”
child code was the most frequent (36%) type of Child Initiation observed during the interactions. The second-
most frequent is the “Design Idea Promotion” child code (17.9%), followed closely by the “Question” child
code (17.7%). The “Statement of Problem” child code constitutes 14.2% of child-led interactions, and the
“Other” child code had the smallest percentage (13.7%).
Child Code Frequencies

40%
Percent of Child-Led Interactions
35%
30%
25%
20%
15%
10%
5%
0%
Design Idea Direct Question Statement of Other
Promotion Problem
Child Code
Figure 3. Distribution of the common ways young women led interactions with their parents
Similarly, the Parent Response codes listed in the Methods section above reflect the most ways that parents
tended to respond to their daughters during these exchanges. The distribution of the Parent Responses code can
be seen in Figure 4. The “Agreement” parent code was the most frequent (27.6%) type of Parent Response
observed during the interactions, followed by the “Suggestion” code (20.4%), the “Question” parent code
(15.9%), ”Other” (14.9%), “Explanation” (14.0%), and “Non-Response” (7.2%).
Parent Code Frequencies

30%
Perent of Child-Led Interactions
25%
20%
15%
10%
5%
0%
Agreement Explanation Non-Response Question Suggestion Other
Parent Code
Figure 4. Distribution of the common ways parents responded during child-led interactions
These results from Parts One and Two suggest that girls in this study do in fact express agency by leading their
parents in engineering design activities. Girls do this by initiating conversation primarily in the form of directing
their parents and promoting design ideas. Parents in this study, in turn, generally respond to their daughters in
agreement, but they also make suggestions, ask questions, and provide explanations. In the next section, we
examine the relationships between the Child Initiation codes and the Parent Response codes.
Part Three: Patterns of Interaction between Children and Parents
Research Question 3 asks about the patterns or relationships between the ways that young women in this study
led interactions with their parents and the ways in which their parents responded. To more clearly identify the
patterns derived from the qualitative results, correlations between the Child Initiation codes and the Parent
Response codes were estimated. From these analyses, four main interaction patterns are discussed.
Interaction Pattern #1: Parents tend to agree with a child’s direction during design.
For the participants in this study, there is a link between children directing the decisions and activities during the
design experience and the parent agreeing with the child, as seen in a small positive correlation between the
“Direct” code for children and the “Agreement” code for parents (r=0.116, p<0.01). For example, Excerpt 1
demonstrates an instance in which the child directs her father who responds in agreement. Speakers are
abbreviated as “CF” for Child, Female; and “AM” for Adult, Male.
Excerpt 1: Case 47, 22:11-28
Code Transcript Segment
Statement of Problem CF: It just can’t, it can’t make that.

Suggestion AM: You gotta go faster to get over there.
Direct CF: Cause this can’t go any faster. See, just go like that.
Agreement AM: Okay. Try it again.
Other, Child CF: Oops. A little too fast.
Suggestion AM: Oop, wrong way. Put it down again. Ha. Yeah, lower down a little bit and let it
go again.
In Excerpt 1, the child describes a problem with the design, which is followed by a suggestion from her father
that provides a solution to this problem. The daughter responds by directing her father, showing him what she
wants him to do to fix the problem. Her father responds in agreement. The child expresses a difficulty to which
the father responds by suggesting another way to improve the ball run design. This interaction shows a few
exchanges, but the emphasis at this stage is the positive correlation evident by the father agreeing in response to
his daughter’s direction.
Interaction Pattern #2: Parents tend to respond to a child’s statement of problem not by agreeing but by
suggesting.
For the participants in this study, there is a small relationship between children verbalizing a problem with the
design or describing a piece not working and the parent responding in a way other than simply agreeing with the
child, as seen in a small negative correlation between “Statement of Problem” code for children and the
“Agreement” code for parents (r=-0.051, p<0.01). When stating a problem, the child is trying to determine a
solution, gain feedback, or receive another form of support; thus, agreement would not achieve this goal. An
example of this is seen in Excerpt 2. Speakers are abbreviated as “CF” for Child, Female; and “AF” as Adult,
Female.
Excerpt 2: Case 15, 23:01-23:21
Statement of Problem CF: I just need it, something to like escalate it a little more.
Suggestion AF: Use this one.
Design Idea Promotion CF: You know something, cause this it could like push it up, and then, to this, and
then this could be a little more further up, and all these, they are like pushable ones,
like go up.
Agreement AF: I like it. Okay, try it.
In Excerpt 2, the child begins by stating a problem with one of the components of the ball run. She recognizes
that she needs a piece to help the ball “escalate” more. The mother suggests a design piece to her daughter,
which shows another trend observed that parents tend to give suggestions to their daughters when they state a
problem, as seen in the small positive correlation between the “Statement of Problem” child code and the
“Suggestion” parent code (r=0.047, p<0.05). The mother’s suggestion encourages the daughter to continue
working, and it even may have sparked the next design idea the daughter mentions, to which her mother agrees.
With the help of suggestions from her parent, the child in Excerpt 2 is eventually able to reason through this
current obstacle and eventually goes on to complete this segment of the ball run.
Interaction Pattern #3: Children’s questions tend to be followed by parent explanations.
For the participants in this study, there is a relationship between children’s questions of clarification to be
followed by parents’ explanations, as evident by the small positive correlation between “Question” child code
and “Explanation” parent code (r=0.110, p<0.01). This is not entirely surprising, considering that when a child
asks a question requesting either advice or instruction the expected response is an explanation from a parent,
perhaps to give information about a concept. Excerpt 3 shows an instance where parent explanations followed a
child’s questions. Speakers are abbreviated as “CF” for Child, Female; and “AF” as Adult, Female.
Excerpt 3: Case 16, 1:58-3:14
Question, Child CF: Mom, where’s the other one?

Explanation AF: You don’t want your ball to get caught here. So you gotta have it at the very end.
Question, Child CF: Almost there. What’s this?
Explanation AF: This one, it’s a...See how that goes like that? So you could have it flip over and
go into the...You might want to move it over. Or the…
Question, Child CF: Do we just like test it out quick, or something?
Non-Response AF: [no response]
Question, Child CF: How does it go up? Oh.
Explanation AF: Oh, it doesn’t go anywhere. So you gotta go what?
Direct CF: Does it go any...down…
Question, Adult AF: Okay. Then what’s gonna happen to that one?
Direct CF: This one has to go up. This one has to go…
Suggestion AF: So you’re gonna bring it down.
Question, Child CF: Move it down. But then how’s it gonna get up to there?
Suggestion AF: Well, then that’s probably where you’re gonna wanna use one of these.
In Excerpt 3, the child initiates the interaction by asking a question, which is followed by an explanation of the
design process by the mother. The daughter then asks another clarifying question, to which her mother explains
again, this time by showing her daughter how a specific building piece works. These exchanges foster a
continually questioning dialogue on the daughter’s part. The last part of this excerpt shows the parent offering a
suggestion to the child’s question. Sometimes, a parent’s response contained elements of both explanation and
suggestion, which indicates that these types of responses are both common for parents in response to their
child’s queries. There is a small relationship between a parent responding to a child’s question in a way other
than with a question, as evident by the significant negative correlation between “Question” child code and
“Question” parent code (r=-0.049, p<0.01). This correlation suggests that the parents in this study responded in
a variety of ways that helped continue the design process. By responding to her daughter’s question with a
suggestion, this mother was able to help her daughter continue to persist in this engineering activity.
Interaction Pattern #4: Parents tend to engage in helpful conversations about children’s design ideas.
For the participants in this study, there is a relationship between children proposing an idea or suggestion
relating to the design of the ball run and parents asking questions about these ideas, as seen in a significant
positive correlation between “Design Idea Promotion” child code and “Question” parent code (r=0.073,
p<0.01). Parents generally responded by asking for further clarification, thus allowing the child to develop her
idea. As reported in the section above, another small relationship observed was parents responding in ways other
than a question, as evident by the negative correlation between “Question” child code and “Question” parent
code (r=-0.049, p<0.01). Excerpt 4 shows this pattern of parents engaging in helpful conversation by responding
to their child’s question in a way other than posing a question, thus sparking further design ideas. Speakers are
abbreviated as “CF” for Child, Female; and “AM” for Adult, Male.
Excerpt 4: Case 41, 24:26-39
Question, Child CF: This more up?

Suggestion AM: I think what we can do is with just what we have here, and just how you pushed
the labyrinth. How you push the, you need to maybe push it faster or slower?
Design Idea Promotion CF: Faster.
Agreement AM: Faster, why don’t we try that?
In Excerpt 4, the daughter asks her father a question, which he responds to by proposing a suggestion to help
clarify her understanding. In her next statement, she demonstrates her comprehension by suggesting to go faster,
which was initiated as a result of her father clarifying some concepts during his suggestion. In this case, the
father not only responds with a helpful suggestion to his daughter’s question, but in his response sets her up for
success by guiding her to the correct answer: to make this specific segment of the ball run work, his daughter
has to push the piston faster. This effective exchange of ideas was evident throughout this case as the daughter
was continually engaged even when she was confused, which enabled her to persist with the help of her father.
The results above suggest that there are several patterns in the interactions between daughters and parents during
family-based engineering activities for the participants in this study. Parents tend to agree with a child’s
direction during design, and, instead of responding to a child’s statement of problem in agreement, they tend to
reply with a suggestion. When children ask a question, parents tend to give an explanation, and parents also tend
to engage in helpful conversations about children’s design ideas. These interaction patterns suggest that in a
range of different interactions during this exhibit experience, parents are able to support their daughters’ ideas
and respond to their questions, which can help the daughter to further persist in this engineering design activity.
Part Four: Differences based on Parent Gender
Research Question 4 asks about what differences, if any, exist between the ways young women interact with
their mothers and their fathers while engaging in the engineering exhibit. To more clearly identify the patterns
derived from the qualitative results, correlations were estimated between the Child Initiation codes, the Parent
Response codes, and parent gender. From this analysis, four main interaction patterns are discussed.
Gender Difference #1: Children tend to direct their mothers more than their fathers.
One interesting pattern in the data was the relationship between the Child Initiation code of Directing and the
gender of the parent. In this dataset, daughters tended to direct their mothers than their fathers, as seen by the
small positive correlation between “Direct” child code and “Female” parent (r=0.070, p<0.01) and the small
negative correlation between “Direct” child code and “Male” parent (r=-0.055, p<0.01). Excerpt 5 shows an
example of a child directing her mother. Speakers are abbreviated as “CF” for Child, Female; and “AF” as
Adult, Female.
Excerpt 5: Case 15, 25:45-26:15
Statement of Problem CF: Yeah, here too. It’s too much…

Suggestion AF: It stops there, you’re gonna need to push it up there.
Other, Child CF: Okay. I thought there’s a pushy there.
Other, Adult AF: No.
Question, Child CF: There isn’t?
Suggestion AF: We will need a push right here.
Direct CF: Put lower this.
Other, Adult AF: That’s a good idea.
Direct CF: Hold that, pull that up.
Question, Adult AF: Ready?

Direct CF: Push it. Yeah it stops.
Question, Adult AF: Down more?
In Excerpt 5, the child begins by stating a problem with one of the components of the ball run, to which her
mother responds not by agreeing but by offering a suggestion to continue the dialogue. The daughter thinks
aloud, which then leads to asking a question that results in her mother’s suggestion to add a design piece. The
daughter then directs her mother to put a certain design piece lower, an idea her mother praises. Next, the child
directs her mother to pull a piece up. The final direction by the daughter is to push a certain design piece, to
which the mother asks a clarifying question. In this example, the mother is receptive to her daughter’s
instructions, which leads to a dynamic interaction during this part of the assembly process.
Gender Difference #2: Children tend not to promote design ideas with their mothers.
Children in this study were less likely to promote their design ideas with their mothers, as seen by the small
negative correlation between “Design Idea Promotion” child code and “Female” parent (r=-0.067, p<0.01).
Following the first child-parent gender relationship above, it seems puzzling that, although children direct their
mothers more than their fathers, children are less likely to promote or advance their own design ideas with their
mothers. This suggests that in this study there may be something different in the environment that affected how
a daughter felt she could direct her parent. Excerpt 6 shows several instances of a child promoting her ideas
when paired with her father. Speakers are abbreviated as “CF” for Child, Female; and “AM” for Adult, Male.
Excerpt 6: Case 24, 2:21-3:06
Direct CF: That’s magnetic.

Agreement AM: Yeah. How does that work?
Direct CF: That’s a strong magnet.
Suggestion AM: Maybe we could try to get it. Whoops!
Design Idea Promotion CF: Like that.
Agreement AM: Like that.
Design Idea Promotion CF: Then we could use this some…
Suggestion AM: And then use something like that to bump the ball up?
Design Idea Promotion CF: No. Then the ball couldn’t get through if we used that. I just found something out
that might help us.
Question, Adult AM: What happens? Yeah.
Design Idea Promotion CF: Wait a minute, I’ve got an idea.
Question, Adult AM: What do you think? So, maybe we can put. See how that works?
Excerpt 6 is a dialogue between a child and her father that includes several moments of the child promoting her
design idea. It begins with the child directing her father by her explanation of the magnetism of a design piece.
The father responds first in agreement and then offers a suggestion. The daughter shows her father how to
assemble a part of the ball run, to which her father agrees. This spurs the next few exchanges, which consist of
another suggestion of an idea from the daughter and then a suggestion from the father. In response to his
daughter’s final two suggestions her father asks clarifying questions, remaining engaged with his daughter over
the course of this entire segment. This excerpt also provides additional evidence to support the trend that parents
in this study tended to agree with their children’s directions.
Gender Difference #3: Mothers tend to agree.
Female parents in this study were more likely to agree to the child’s statement or question, as evident by the
significant positive correlation between “Agreement” parent code and “Female” parent (r=0.080, p<0.01).
Excerpt 7 shows an instance where a mother agrees with her daughter. Speakers are abbreviated as “CF” for
Child, Female; and “AF” for Adult, Female.
Excerpt 7: Case 26, 8:30-9:03
Question, Child CF: Okay, so how would that go through?

Suggestion AF: Well try it.
Direct CF: It’s gonna go down. Okay.
Agreement AF: Yeah, go.
Statement of Problem CF: What? Yep, but I won’t go over.
Suggestion AF: If so, get ready, you can wait the ball. Stop. There you go. Now let’s just finish
on this side.
Question, Child CF: Oh, will it go this way?
Explanation AF: No, cause it doesn’t have a main on that side. So you just use these.
This excerpt begins with a query from the child about how the ball run would work, to which her mother
responds with a suggestion to try and see. In the second child-led interaction, the daughter directs her mother by
explaining what will happen next—specifically, that the ball will go down—to which the mother agrees. After
the mother’s agreement, the dialogue continues with problem-solving and suggestions back and forth. The end
of this excerpt also further supports the previously-stated trend that parents offer explanations to their daughter’s
questions.
Gender Difference #4: Fathers tend to respond more to their daughters.
Fathers in this study were less likely to provide a non-response to a statement or question initiated by their
daughters, as seen in the small negative correlation between “Non-Response” parent code and “Male” parents
(r=-0.046, p<0.05). By being less likely to not respond, fathers created more opportunities for engagement with
their daughters by some form of response. Excerpt 8 shows an example of a father playing a critical role in
responding to his daughter’s ideas. Speakers are abbreviated as “CF” for Child, Female; and “AM” for Adult,
Male.
Excerpt 8: Case 41, 7:45-8:00
Design Idea Promotion CF: Or maybe we should put this more right here, so can lift it up like nothing and
can go up.
Question, Adult AM: Say what?
Design Idea Promotion CF: We should put that over there, maybe, so once it can make it over.
Suggestion AM: We could. But if we put it here, it will lift it up a lot higher. So what we wanted
to do, maybe what we’re gonna need to do is, uh, lower this side down so it rolls more
into it. So the force of gravity gets the ball going, cause it kind of has to land like
right into here.
In Excerpt 8, the daughter proposes a design idea, to which her father responds by asking her to repeat herself.
After she clarifies her design idea, her father provides a suggestion that both explains to his daughter why her
idea structurally will not work and also gives another option to solve this problem. The father in this case had
the opportunity not to respond to his daughter, but instead he asked a clarifying question to continue to assist his
daughter with the activity. This excerpt was one of the most prominent examples of the constructive role parents
could play in their daughter’s learning and interest in the activity.
These findings suggest that for the participants in this study there is variation in the ways that female children
engage with parents of different genders when engaged in engineering activities. Female children are more
likely to direct their mothers than their fathers but less likely to promote their design idea with their mothers.
Mothers are more likely to respond by agreeing with their daughters but less likely to respond by explaining.
Fathers tend to respond more to their daughters. Further examination of these patterns can help shed light on the
ways mothers and fathers can play a role in the development of engineering interest, identity, and agency in
young women.
Discussion
This study found that interactions during family-based engineering activities can be productive learning
experiences for girls to begin demonstrating early elements of persistence. The proportion of overall turns of
talk coded as child-led interactions in Figure 2 shows that girls are leading interactions during design activities.
When girls lead activities they are more likely to be interested in the activities, which can result in greater levels
of persistence. The majority of the child-led interactions consisted of children directing their parents, but other
patterns included children promoting their design ideas and asking questions. The majority of parents’ responses
were in agreement with their daughter, but parents also were observed to make a suggestion and ask a question.
Several patterns also were discovered in the interactions between daughters and parents. For example, parents
tend to agree with their child’s directions and respond to their statements of a problem not by agreeing but by
suggesting. Mothers were more likely to agree, but fathers were more likely to respond. Parents tend to provide
explanations to children’s questions and engage in helpful conversations about their child’s design ideas. A
striking difference was discovered between the ways girls interact with their mothers and fathers; girls were
significantly more likely to direct their mothers than their fathers. They were also less likely to promote their
design ideas with their mothers.
These findings suggest that there might be connections between the ways that parents engage with their
daughters during engineering design activities and the ways that girls exhibit elements of persistence through
demonstrating agency in STEM activities. This study shows a strong link between children directing and parents
agreeing. In Excerpt 1, the daughter directs her father by showing him what to do to fix a problem. He responds
in agreement but in a way that sets up further dialogue, thus fostering effective communication. By agreeing and
then later also providing a helpful suggestion, the father plays a role in his daughter’s continuation of the
activity and ultimately her achievement in successfully completing the activity. Parents’ motivational practices
can positively influence their children’s achievement and parents’ intrinsic motivational practices can even
result in higher persistence in STEM careers for their children (Ing, 2014). As expectancy-value theory
emphasizes, individuals’ motivation, such as how much they value an activity or how they think they will do,
has a large impact on their persistence and performance (Wigfield & Eccles, 2000). Because the father was
agreeable in Excerpt 1, he provided a positive response to his daughter’s leadership during this engineering
activity, which may contribute to her overall confidence and self-efficacy in engaging with these types of
engineering design activities.
These findings support previous research showing that parents can be positive influences on student motivation
(Farmer, 1985). The positive effect of parental support is seen by the trend that when a child mentions a
problem, parents usually respond in a productive way instead of simply agreeing with their child. As seen in
Excerpt 5, in response to the child’s statement of a problem, the mother provides a suggestion to her daughter
that initiates further dialogue. The result was collaborative problem-solving by the mother and daughter, which
led to eventual completion of that immediate segment of the ball run. For adolescent girls in particular, their
math and science motivation is influenced by mothers and by peer math and science support (Leaper et al.,
2012). In this example, the daughter’s motivation to continue to persist in the activity is sparked by her mother’s
suggestion, which acts as a springboard for further discussion. In this case the mother’s support has lasting
effects, as seen by the daughter later directing her mother and showing leadership in the activity. As a result of
parental support positive consequences are seen later in the activity. This suggests that parents can encourage
their children’s later success and ownership of an activity, which has been shown through expectancy-value
theory to impact their persistence and performance (Wigfield & Eccles, 2000). Research showing that teachers
can be positive influences on student motivation suggests that the findings of this study may also apply to how
teachers and students interact in a classroom (Farmer, 1985).
These results also emphasize the impact of incorporating gender-specific strategies in these engineering design
activities. Strong verbal tendency is commonly associated with females. Studies have shown that teachers are
aware of the strong verbal tendencies of females and thus try to incorporate this verbal component into science
by asking girls higher-level questions or incorporating verbal components in hands-on lab experiences
(Subrahmanyan & Bozonie, 1996). Excerpt 3 is an example of an educationally-enriching dialogue between
mother and daughter that demonstrates the efficacy of conversation in engineering design activities. The child
initiates the interaction by asking a question to which her mother responds by explaining part of the design
process. This leads to several more exchanges of dialogue, one of which includes the mother asking a thought-
provoking question that challenges her daughter to think critically about how a certain piece will influence the
design. By providing constructive responses, the mother creates an environment that fosters constructive
conversation that leads to a continuation of the design process, despite the daughter’s initial questions. Not only
does the strong verbal communication between mother and daughter help provide a productive learning
environment, but the hands-on aspect of this activity further strengthens the daughter’s persistence.
Additional observations from the dataset suggest that these parent and child interactions might not always be in
direct support of persistence. For example, in Case 12, despite no obvious external signs of frustration on the
part of either participant, when the father asks his daughter if they want to try another activity she instantly
agrees. Up until this point the daughter had been leading the activity, and they were working together without
any major problems. This example shows a downside of what can happen if a parent does not feel the need to
persist in the STEM activity and the child follows his lead. While this museum environment lends itself to the
distraction of other activities, this case nonetheless shows the impact that parents’ decisions not to persist can
have on their impressionable children. In Case 12, the father and daughter did not seem as attached to the
activity given that they were able to leave quickly and easily. There were other cases, however, where the
parents had to ask several times for their daughter to be done with the activity. While these examples were
positive in that the daughters were determined to persist, they also show the negative effect that parents can have
when they are attempting to persuade their daughters to stop working. Case 31 even ended with the child in
tears, as she was frustrated with both the activity and her father’s insistence that she need to leave immediately
to go home. Overall, while the majority of the correlations described above can have positive implications,
larger themes seen in the videos depicted underlying challenges when parents are involved with their daughters’
persistence.
Limitations, Implications, and Future Work
This study has multiple limitations. First of all, the scope of this study is limited by the sample of families who
participated in the study. Parents who visit a science museum with their children may be of a similar
demographic in terms of educational attainment or socioeconomic status (Farrell & Medvedeva, 2010), or they
already may be more inclined to engage and collaborate with their children. Therefore, it is important to
consider this bias when interpreting these findings. Second, there was not a closed system during the
experimental data collection. The exhibit was in the middle of a busy museum floor, which allowed for some
interaction with people outside the scope of the study. Siblings were also a distraction, as well as parents who
entered midway through the design project. These added family members were not prevented from joining the
original parent-child dyad since they could create more opportunities for family interactions. Finally, the lack of
additional demographic data for the participants limits the types of analyses that can be done on this subset of
data from the GRADIENT study. Additional information about the family, such as whether it is a single-parent
household, the typical temperament and disposition of the parent and child, and general parenting styles and
techniques used in the daily routines of family life, all could better inform the interpretation of these data
analysis results.
Despite these limitations, our findings still have implications for potentially increasing the persistence of
females in STEM. These results give us an initial picture of the specific ways daughters engage with their
parents during a challenging engineering design activity, whether by directing, stating a problem, promoting a
design idea, or questioning. The specific ways that parents respond to their daughters by agreeing, explaining,
suggesting, questioning, or not responding also provided insight. These findings can be useful for both informal
and formal educators, such as parents and teachers, as they continue to develop their understanding of how to
foster environments where girls can lead interactions during STEM activities by being actively engaged and
expressing agency. Since much of the research on persistence in STEM focuses on college-aged students, any
teacher introducing engineering concepts to pre-college students would be able to draw upon this research to
learn about productive ways to structure learning environments and practices to engage young women in
developing confidence, interest, and self-efficacy during engineering activities. Parents might also be able to
leverage these findings and engage their female children in authentic collaboration during family-based STEM
activities, which might impact their daughter’s ongoing development of motivation and persistence in STEM.
Future research on interactions during family-based engineering activities could focus on a variety of topics. A
follow-up study could focus on exploring these patterns in a sample of parent-child dyads where all the children
are boys, thus providing the comparison group that would allow researchers to identify patterns between the
ways that fathers or mothers interact with their sons and then compare these patterns to those identified in this
study. This shift in focus would allow for further knowledge of the dynamics between parents and their sons and
daughters in an attempt to identify any similarities or differences between these child-parent interactions. In
addition, further analysis of the variation in patterns based on time-on-task, the presence of siblings (both older
and younger), and additional demographic information such as the occupations of the primary parent participant
could also continue to refine the interpretations of these data and the findings of this study.
Using this same dataset, different codes during analysis could be applied to answer other research questions
such as how children respond when faced with a specific challenge in a design activity and what parents can do
to prompt a productive response. This study focused on the expression of agency through primarily verbal
communications between children and parents, but another topic of interest could be to examine nonverbal cues.
Identifying what actions girls take when they come across a challenging part of the activity could provide
further insight into ways parents can help foster persistence through their actions as well as words. The context
of this study could be expanded to include a different engineering activity in this same museum but could also
be translated to other environments such as a classroom. All of these future directions could lead to positive
insights into the effects of environments and adults on young children’s persistence in STEM activities, which
could have implications for the development of more productive learning environments and instructive practices
for young STEM learners in the future.
Acknowledgements
The authors of this study wish to thank the members of the GRADIENT research team, especially Brianna
Dorie, Zdanna King, and Scott Van Cleave. In addition, the authors thank the reviewers and special issue editors
for their thoughtful and productive feedback on this manuscript. This material is based upon work supported by
the National Science Foundation under Grant No. (1136253) and the Glynn Family Honors Program at the
University of Notre Dame. Any opinions, findings, and conclusions or recommendations expressed in this
material are those of the author(s) and do not necessarily reflect the views of the listed funders.
References
Banks, W. C., McQuater, G.V., & Hubbard, J.L. (1978). Toward a reconceptualization of the social-cognitive
bases of achievement orientations in blacks. Review of Educational Research, 48(3), 381-97.
Calabrese Barton, A., & Tan, E. (2010). We be burnin'! Agency, identity, and science learning. The Journal of
the Learning Sciences, 19(2), 187-229.
Cardella, M., Svarovsky, G.N., Dorie, B. (2013). Gender research on adult-child interactions in informal
engineering environments (GRADIENT): Early findings. Proceedings of the 120th ASEE Annual
Conference & Exposition, J Atlanta, GA, June 2013.
Carlone, H. B. (2004). The cultural production of science in reform‐ based physics: Girls' access, participation,
and resistance. Journal of Research in Science Teaching, 41(4), 392-414.
Crismond, D. P., & Adams, R. S. (2012). The informed design teaching and learning matrix. Journal of
Dorie, B.L., Cardella, M.E., and Svarovsky, G. (2014). Capturing the design behaviors of a young children
working with a parent. Proceedings of the 121st American Society of Engineering Education Annual
Conference & Exposition, Indianapolis, IN, June 2014.
Dorie, B.L., Cardella, M.E., and Svarovsky, G. (2015). Engineering Together: Context in Dyadic Talk During
an Engineering Task. Proceedings of the 122nd American Society of Engineering Education Annual
Conference & Exposition, Seattle, WA, June 2015.
Dym, C. L., Little, P., Orwin, E. J., & Spjut, E. (2009). Engineering design: A project-based introduction. John
Wiley and Sons.
Eccles, J. S. (2005). Studying gender and ethnic differences in participation in math, physical science, and
information technology New Directions for Child and Adolescent Development, 110(Winter), 7-14.
Eccles, J., Barber, B., & Josefowicz, D. (1999). Linking gender to educational, occupational, and recreational
choices: Applying the Eccles et al. model of achievement-related choices. In W. B. Swann, J. H.
Langlois & L. A. Gilbert (Eds.), Sexism and stereotypes in modern society: The gender science of Janet
Spence (pp. 153-192).
Farmer, H. S. (1985). The role of typical female characteristics in career and achievement motivation. Youth &
Society, 16(3), 315-334.
Farrell, B., & Medvedeva, M. (2010). Demographic transformation and the future of museums. AAM Press.
Feder, M. A., Shouse, A. W., Lewenstein, B., & Bell, P. (Eds.). (2009). Learning science in informal
environments: People, places, and pursuits. National Academies Press.
Frome, P. M., Alfeld, C. J., Eccles, J., & Barber, B. (2006). Why don't they want a male-dominated job? An
investigation of young women who changed their occupational aspirations. Educational Research and
Evaluation, 12(4), 359-372.
Graham, M. J., Frederick, J., Byars-Winston, A., Hunter, A. B., & Handelsman, J. (2013). Increasing persistence
of college students in STEM. Science, 341(6153), 1455-1456.
Herbert, J., & Stipek, D. (2005). The emergence of gender differences in children's perceptions of their
academic competence. Journal of Applied Developmental Psychology, 26(3), 276-295.
Hill, C., Corbett, C., & St. Rose, A. (2010). Why So Few? Women in Science, Technology, Engineering, and
Mathematics. American Association of University Women. Retrieved from
https://eric.ed.gov/?id=ED509653
Hoegh, J., & Pawley, A. L. (2010). “Modeling the career pathways of women STEM faculty through oral
histories and participatory research methods.” Proceedings of the 117th ASEE Annual Conference &
Exposition, Louisville KY, June 2010.
Hughes, R. (2011). Are the predictors of women’s persistence in STEM painting the full picture? A series of
comparative case studies. International Journal of Gender, Science and Technology 3(3): 547-570.
Ibrahim, S., & Alkire, S. (2007). Agency and empowerment: A proposal for internationally comparable
indicators. Oxford development studies, 35(4), 379-403.
Ing, M. (2014). Can parents influence children’s mathematics achievement and persistence in STEM
careers?. Journal of Career Development, 41(2), 87-103.
Leaper, C., Farkas, T., & Brown, C. S. (2012). Adolescent girls’ experiences and gender-related beliefs in
relation to their motivation in math/science and English. Journal of youth and adolescence, 41(3), 268-
282.
Lloyd, J. E. V., Walsh, J., & Yailagh, M. S. (2005). Sex differences in performance attributions, self-efficacy,
and achievement in mathematics: If I'm so smart, why don't I know it? Canadian Journal of Education,
28(3), 170-182.
Marra, R. M., Rodgers, K. A., Shen, D., & Bogue, B. (2009). Women engineering students and self-efficacy: A
multi-year, multi-institution study of women engineering student self-efficacy. Journal of Engineering
Education, 98(1), 27–38. https://doi.org/10.1002/j.2168-9830.2009.tb01003.x
Mannon, S. E., & Schreuders, P. D. (2007). All in the (engineering) family? The family occupational
background of men and women engineering students. Journal of Women and Minorities in Science in
Engineering, 13(4), 333-351.
Matusovich, H. M., Streveler, R. A., & Miller, R. L. (2010). Why do students choose engineering? A qualitative,
longitudinal investigation of students' motivational values. Journal of Engineering Education, 99(4),
289-303.
Metevier, A., Seagroves, S., Shaw, J., & Hunter, L. (2015). ISEE’s Contributions to STEM Persistence and
Effective Mentoring Practices: A Report Summarizing Fourteen Years of Design, Practice and Outcomes
Studies. Institute for Scientist & Engineer Educators, 1.
National Science Foundation, National Center for Science and Engineering Statistics. (2017). Women,
minorities, and persons with disabilities in science and engineering: 2017. Special Report NSF 17-310.
Arlington, VA. Available at www.nsf.gov/statistics/wmpd/.
Patton, M. (2002). Qualitative research & evaluation methods (3rd ed.). Thousand Oaks, CA: Sage
Publications.
Seymour, E. (1995). The loss of women from science, mathematics, and engineering undergraduate majors: An
explanatory account. Science Education, 79(4), 437–473. https://doi.org/10.1002/sce.3730790406
Seymour, E. & Hewitt, N. (1997). Talking about leaving: why undergraduates leave the sciences. Boulder,
Colo.: Westview Press.
Shaffer, D. W., & Serlin, R. C. (2004). What good are statistics that don’t generalize?. Educational
Researcher, 33(9), 14-25.
Sheppard, S., Atman, C., Stevens, R., Fleming, L., Streveler, R., Adams, R., & Barker, T. (2004). Studying the
Engineering Student Experience: Design of a Longitudinal Study. Proceedings of the 2004 American
Society for Engineering Education Annual Conference & Exposition, Salt Lake City, UT.
Singh, K., Granville, M., & Dika, S. (2002). Mathematics and science achievement: Effects of motivation,
interest, and academic engagement. The Journal of Educational Research 95(6): 323-332.
Svarovsky, G. N. (2014). Engineering learning in museums and other designed settings: Towards a theoretical
framework. In Strobel, J., Purzer, S. & Cardella, M. (Eds.) Engineering in Pre-College Settings:
Research into Practice. Purdue University Press, West Lafayette, Indiana.
Svarovsky, G., Cardella, M., Dorie, B., and King, Z. (2017). Productive forms of facilitation for young girls
during engineering activities within informal learning settings. Paper presented at the American
Educational Research Association Annual Meeting, April 2017, San Antonio, TX.
Subrahmanyan L., & Bozonie, H. (1996). Gender equity in middle school science teaching: Being “equitable
“should be the goal. Middle School Journal 27(5):3–10.
Tenenbaum, H. R. (2009). 'You'd be good at that': Gender patterns in parent-child talk about courses. Social
Development, 18(2), 447-463.
Trenor, J. M., Yu, S. L., Waight, C. L., Zerda, K. S., & Sha, T.-L. (2008). The relations of ethnicity to female
engineering students' educational experiences and college and career plans in an ethnically diverse
learning enviornment. Journal of Engineering Education, 97(4), 449-465.
Voyles, M., & Williams, A. (2004). Gender differences in attributions and behavior in a technology classroom.
Journal of Computers in Mathematics and Science Teaching, 23(3), 233-256.
Wigfield, A., & Eccles, J.S. (2000). Expectancy–value theory of achievement motivation. Contemporary
educational psychology 25(1), 68-81.
Author Information
Gina Navoa Svarovsky Catherine Wagner
University of Notre Dame University of Notre Dame
107 Carole Sandner Hall, Notre Dame, IN 46556 107 Carole Sandner Hall, Notre Dame, IN 46556
U.S.A. U.S.A.
Contact e-mail: gsvarovsky@nd.edu
Monica Cardella
Purdue University
516 Northwestern Avenue, Purdue University, West
Lafayette, IN 47906
U.S.A.

International Journal of Education in Mathematics, Science and Technology

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

International Journal of Education in Mathematics, Science and Technology

Uploaded by

Copyright:

Available Formats

International Journal of Education in Mathematics, Science and Technology

Volume 6, Number 3, July 2018 ISSN: 2147-611X

Volume 6, Number 3, July 2018 ISSN: 2147-611X

Mack SHELLEY - Iowa State University, U.S.A.

Guest Editor (Special Issue on Engineering Education)

Şenay PURZER- Purdue University, U.S.A.

International Journal of Education in Mathematics, Science and Technology (IJEMST)

International Journal of Education in Mathematics, Science and Technology (IJEMST)

Volume 6, Number 3, July 2018 ISSN: 2147-611X

Culture, Identity, and Motivation in Engineering Education [Editorial Essay] i

Engineering Identity Development: A Review of the Higher Education Literature 254

Exploring Moments of Agency for Girls during an Engineering Activity 302

Volume 6, Number 3, 2018 DOI:10.18404/ijemst.428160

Culture, Identity, and Motivation in Engineering Education

Article Info Abstract

Themes of This Special Issue

Self-efficacy, Academic Performance, and Student Study Strategies

Unique Groups and Contexts

Volume 6, Number 3, 2018 DOI:10.18404/ijemst.428165

Relationship of Mathematics Self-efficacy and Competence with Behaviors

Article Info Abstract

Motivation for this Study

This research was guided by the following research questions:

Descriptive statistics Quantitative analysis Next participant

Interviews n = 11 Qualitative data collection

Coding Memos, codes, and categories

Deep understanding of students’

Quantitative Phase Methods

Qualitative Phase Methods

Mixed Phase Methods

Table 1. Cronbach’s α values of the items of each construct on the MSES

Mathematics Activities (18 items) 0.91

Fall 2015 Mathematics Courses (14 items) 0.83

Mathematics Problems (18 items) 0.90

Mathematics Activities (18 items) 0.69

Spring 2016 Mathematics Courses (14 items) 0.80

Mathematics Problems (18 items) 0.90

Participants’ Mathematics Competence and Mathematics Self-Efficacy

P Spring 2014 6.6 NA 8.1 Medium Higher

L Summer 2014 7 NA 9.1 Medium Higher

P Fall 2014 7.6 NA 1.5 High Lower

P Fall 2014 8 NA 2.9 High Lower

L Spring 2015 6.2 NA 7.3 Medium Higher

P Spring 2015 6.8 NA 6.7 Medium Matched

L Fall 2015 7 5.8 8.5 Medium Higher

L Spring 2016 6.5 7.9 6.6 Medium Matched

P Spring 2016 9 9 8.3 High Matched

L Spring 2016 6.9 8.8 6.5 Medium Matched

L Spring 2016 7.8 8.6 8.2 High Matched

1  Medium Low 6.6 24 1

1  Medium Low 6.3 6 0

Mixed Methods Results

Mismatch between High Mathematics self-efficacy and Lower Mathematics Competence

Mismatch between Medium Mathematics self-efficacy and Higher Mathematics Competence

Matched High Mathematics self-efficacy and High Mathematics Competence

Conclusion and Recommendations

(3rd ed.). Thousand Oaks, CA: Sage.

Differences, 16, 1–12.

Volume 6, Number 3, 2018 DOI:10.18404/ijemst.428171

An Instrument for Examining Elementary Students’ Engineering Interests

Article Info Abstract

Children’s Interests in and Attitudes toward Science and Engineering

Table 1. Subscales identified and named at each stage of analysis

Self-Efficacy in Science Self-Efficacy in Engineering (not retained)