Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Why Do Students Choose Engineering?

A
Qualitative, Longitudinal Investigation of
Students’ Motivational Values
HOLLY M. MATUSOVICH, RUTH A. STREVELERa, AND RONALD L. MILLERb
Colorado School of Minesb, Purdue Universitya, Virginia Polytechnic Institute and State University

BACKGROUND method approach, participants included eleven students (five men and six
Recently published reports call for an increase in the number of engineering women) at a U.S. technical school.
graduates and suggest appropriate characteristics that these graduates
should embody. Accomplishing such change first requires understanding RESULTS
why students choose to pursue engineering degrees. Results demonstrate that different patterns exist in the types of value or per-
sonal importance that participants assign to earning an engineering degree.
PURPOSE (HYPOTHESIS) Moreover, a primary differentiating feature of these patterns is whether or not
Framed in motivation theory, our purpose was to better understand how stu- participants choose engineering because it is consistent with their personal
dents choose engineering by answering the question: How do engineering stu- identity or sense of self.
dents’ engineering-related value beliefs contribute to their choices to engage
and persist in earning engineering degrees? CONCLUSIONS
We conclude that values are very important in students’ choices to become
DESIGN/METHOD engineers. To increase persistence rates we must focus on values, especially by
This research uses Eccles’ expectancy-value theory in a qualitative, longitudi- helping students connect their personal identities to engineering identities.
nal examination of undergraduate students’ choices to enroll and persist in
engineering majors. In particular, the focus of this work is Eccles’ subjective
task value (STV) construct, which incorporates the personal importance an KEYWORDS
individual assigns to engaging in an activity. Using a multiple case study expectancy-value theory, identity, motivation

I. INTRODUCTION To address this knowledge gap, we examined persistence using


motivation theory and specifically expectancy-value theory (Eccles,
A. Problem Statement and Research Questions 2005, 2007; Eccles et al., 1983). In its simplest form, Eccles’ ex-
Fueled by reports calling for increases in science, technology, pectancy-value theory suggests that choices to engage in activities,
engineering, and math (STEM) graduates (Committee on Sci- such as earning an engineering degree, are shaped by competence
ence, Engineering, and Public Policy (COSEPUP) and Policy and and value beliefs (Figure 1). Competence beliefs address questions
Global Affairs (PGA), 2006; Jackson, 2002; NAE, 2004, 2005), of ability, “CAN I do this task?” and value beliefs consider the per-
considerable research efforts have focused on attracting and retain- sonal importance of a task, i.e., “Do I WANT to do this task?” This
ing engineering students. Students who graduate in engineering study focuses on value beliefs to offer explanatory insights into per-
are called persisters and students who leave engineering majors are sistence choices by answering the question: How do engineering
called non-persisters. Significant findings include descriptions and students’ engineering-related value beliefs contribute to their choic-
characteristics of persisters and non-persisters (e.g., French, Im- es to engage and persist in earning engineering degrees? To answer
mekus, and Oakes, 2005; Mendez et al., 2008; Nicholls et al., this question, we used case study methods to qualitatively examine a
2007). However, what is lacking in current research findings is a series of extensive, matched longitudinal interviews with five male
general understanding about how and why students choose to and six female engineering students. This study was part of a larger
enter and persist in earning engineering degrees. This greatly in- body of work, the Academic Pathways Study (APS), conducted by
creases the difficulty associated with attracting and retaining the Center for Advancement of Engineering Education (CAEE)
students in engineering programs. It is hard to recruit and train fu- (Clark et al., 2008; Sheppard et al., 2004).
ture engineers when prospective candidates’ goals, objectives, and
decision-making criteria are not well understood. Although
Ohland et al. (2008) recently found that persistence rates of engi- II. SITUATION IN CURRENT LITERATURE
neering students are actually not that different than those for
students in other majors, their findings continue to highlight the Many factors contribute to the choice to become an engineer,
gender gap in engineering majors. This makes understanding per- which makes attracting and retaining engineering students a com-
sistence choices vital in creating a sufficiently large and diversified plex task. Researchers suggest that pedagogical and programmatic
engineering workforce. changes are necessary to influence current and perspective students’

October 2010 Journal of Engineering Education 289


choices to be engineers (e.g., NAE, 2008; Sheppard et al., 2009). (Lent, Brown, and Hackett, 1994). However, the defined value
However, before we can effectively design these pedagogical and categories in expectancy-value theory provide a more explicit
programmatic changes, we need a deep-rooted understanding of framework to examine students’ beliefs about the importance of
the students’ perspectives and why they choose engineering. In choosing engineering careers and how such beliefs shape the per-
order to connect with current and prospective students, we need to sistence process. SCCT frames values, primarily interest, within
know their starting point. outcome expectations, meaning that interest in an activity is fu-
Surprisingly little existing research on persistence in engineering eled by the anticipated outcomes from an activity and the impor-
has been conducted from the student perspective, and even less re- tance of such outcomes to the individual (Lent, Brown, and
search explains how persistence happens. One notable exception is Hackett, 1994). In contrast, expectancy-value theory posits four
Seymour and Hewitt’s (1997) large-scale study on persistence in sci- value categories, including interest, importance, cost, and
ence, math, and engineering (SME) which explicitly addresses rela- attainment, to describe how individuals assign importance to
tionships between reasons for choosing SME fields and tendencies engaging in an activity (Eccles, 2005). The four value categories
to complete degrees in those fields. Their findings show that leaving and relationships to activity engagement, such as career choice,
SME fields was usually not a result of poor preparation or conceptual are described in greater detail in the following section.
difficulty. Instead, choices to leave SME fields tend to reflect the
reasons they originally chose SME fields. Persisters were more likely A. Expectancy-Value Theory and Persistence Choices
to have chosen engineering based on interest. Non-persisters tended As previously described, expectancy-value theory posits two pri-
to choose SME fields for reasons not related to the nature of the mary sets of beliefs, competence and value, which factor into choic-
work associated with the major such as the influence of family mem- es to engage in activities such as earning an engineering degree
bers, high school teachers, and others, for materialistic reasons, (Figure 1). Competence and value beliefs are shaped by many con-
and/or through uninformed choices, such as choosing engineering tributing factors including: past experiences, the influences of so-
because they did well in high school math and science courses. Sey- cializers (e.g., parents, teachers and peers), personal identity beliefs
mour and Hewitt (1997) suggest that intrinsic interest must be a (the aspects of ourselves that make us unique), and collective identi-
strong element by itself or in combination with other reasons for ty beliefs (the aspects of ourselves that tie us to others) (Eccles,
choosing SME fields to favor persistence. 2005, 2009; Eccles et al., 1983). Consistent with the tenets of social
This current study builds on the work of Seymour and Hewitt cognitive theories, Eccles’ model is based on an individual’s percep-
(1997) by framing persistence choice in motivation theory. Since tion of his or her abilities and task values that shape engagement
motivation theories attempt to explain the process whereby individ- and persistence behavior; importantly it is not his or her actual abili-
uals choose and continue to engage in activities, they have the po- ty or task completion that is important.
tential to facilitate a move from describing characteristics of persis- Within expectancy-value theory, competence beliefs are identi-
ters and non-persisters to explaining how and why persistence fied as expectancies of success and include an individual’s belief as to
happens. Specifically, this research uses expectancy-value theory how well he or she will perform in an upcoming task or activity (Ec-
(Eccles et al., 1983), which has rich history of application in under- cles et al., 1983). Expectancies of success are similar to both Ban-
standing career choices (Eccles, 1986; Eccles et al., 1983; Eccles, dura’s (1997) self-efficacy construct and also to self-concept of abili-
Barber, and Jozefowicz, 1999; Frome et al., 2008). We intentionally ty (Shavelson, Hubner, and Stanton, 1976). Although these three
chose a motivational perspective, and specifically expectancy-value concepts are distinct by definition, they have proven difficult to sep-
theory, over several other models of career choice including social arate empirically (Eccles and Wigfield, 1995; Eccles et al., 1993;
cognitive career theory (SCCT) (Lent, Brown, and Hackett, 1994). Watt and Eccles, 2008; Wigfield, 1994; Wigfield and Eccles, 2000)
SCCT has found application in understanding the career choices of and are often measured in the same ways (Bong, 2001).
engineering students (Lent et al., 2003, 2008a, 2008b; Trenor et al., Competence beliefs have been studied more widely than value
2008) and has some overlapping constructs with expectancy-value beliefs across all age groups. For example, competence beliefs: 1)
theory such as ability beliefs, outcome expectations, and interests link to better task performance among fifth to twelfth graders

Figure 1. Simplified view of Eccles' expectancy-value theory showing the relationship between competence and value beliefs and the
motivated action of earning an engineering degree.

290 Journal of Engineering Education October 2010


(Eccles et al., 1983) and among college students (Bong, 2001), 2) to achievement, little is known about the role of personal values and
relate to beliefs about task value or importance (Jacobs et al., 2002; interests related to learning. He recommends, “It is time that we re-
Wigfield et al., 1997), 3) tend to decrease with age for primary and dress this imbalance and give more attention to the valuing aspects
secondary school children (Jacobs et al., 2002), and 4) predict career of motivation” (p. 192). Echoing this call, we need more research on
aspirations (Correll, 2001; Eccles, Barber, and Jozefowicz, 1999). engineering students’ values so that we can fully understand and
Competence beliefs are also more widely studied than value be- then influence persistence in engineering.
liefs among engineering students and are most often grounded in
self-efficacy theory. Research on competence beliefs shows: 1) self-
efficacy mediates the relationship between positive feedback and III. METHODS
better performance (Alias and Hafir, 2009), 2) mastery experiences
and social comparisons are extremely important sources of self- In conjunction with expectancy-value theory, case study meth-
efficacy beliefs among first year engineering students (Hutchison ods guided this research. Case study research differs from other
et al., 2006; Hutchison-Green, Follman, and Bodner, 2008); and 3) forms of qualitative research, such as ethnography and grounded
among female students, self-efficacy is related to intention to persist theory, because case study research tests existing theory (Yin, 2003).
(Hutchison et al., 2006; Hutchison-Green, Follman, and Bodner, Together, expectancy-value theory and case study methods shaped
2008; Marra et al., 2009). decisions regarding participant selection, use of data sources, the
Although studied less often, value beliefs are no less important. analysis process and interpretation of results.
Whereas competency beliefs look at a person’s ability to do a task or Case study research methods are used across many disciplines,
engage in an activity, value beliefs look at the desire to engage, or e.g., psychology, social science, business, and economics, when re-
the importance of engaging, in a task or activity. Eccles defines sub- searchers want to understand events or social systems while main-
jective task values (STV) as the individual’s incentives for engaging taining the realistic characteristics and details of such systems or
in different tasks or activities based both on the nature of the task events (Yin, 2003). In this application, we examined students’
and how well it aligns with personal values, goals, and needs choices to pursue engineering degrees, which we argue constitutes
(Eccles, 2005; Eccles et al., 1983). There are four categories of both an event and a social system. As an event, choosing to pursue
STV: 1) attainment value, the individual’s perception of how per- an engineering degree is a formative life event that extends over the
formance on the task reflects on the individual and how this reflec- undergraduate years, i.e., each year until degree completion stu-
tion matches with self-concept; 2) intrinsic or interest value, the dents must choose to continue an engineering major or to leave en-
enjoyment experienced in doing the task; 3) utility value, the per- gineering. As a social system, students are making persistence
ceived future importance of engaging in the task that may be direct- choices based on their experiences, perceptions, and with input
ly or indirectly related to the task itself; and 4) relative cost (cost), the from faculty, peers, and family members. To maintain realistic
price of success or failure in terms of effort, time, and/or psychologi- characteristics of the choice process, we considered the participants
cal impact (Eccles, 2005; Wigfield and Eccles, 1992). holistically in the context of being engineering students. This
Research directed towards understanding STV shows that they: means we did not just examine their choice process as a separate and
1) relate to future task engagement plans such as enrollment in ad- distinct activity; rather we considered the choice process as an expe-
vanced mathematics courses among fifth to twelfth graders (Eccles rience situated in the context of other simultaneous and related ex-
et al., 1983; Eccles, Barber, and Jozefowicz, 1999) or occupational periences such as taking classes, participating in campus activities,
choices among fifth to twelfth graders (Eccles, Barber, and Joze- interacting with family members, etc.
fowicz, 1999) and among early career professionals (Frome et al., Using multiple cases allows deeper exploration of, and additional
2008), 2) predict undergraduates’ intentions to attend graduate context for, the individual cases (Miles and Huberman, 1994;
school (Battle and Wigfield, 2003), and 3) are better predictors of Stake, 2006). For example, in this research students have different
activity choice than competence beliefs (Wigfield, 1994; Wigfield campus and classroom experiences; therefore, a broader perspective
and Eccles, 2000) among children and adolescents. Limited work on the choice to be an engineer is obtained than if a single case, or
on STV has been conducted specifically regarding engineering al- student, had been the sole focus. This deeper, contextual under-
though there are several notable exceptions. First, a survey found standing can promote greater transferability of results (Miles and
that both engineering and non-engineering college students gener- Huberman, 1994) and enable generalization to theory (Yin, 2003).
ally believe that engineering benefits society but that it takes too In all case study research, it is important to set spatial and tem-
much effort to earn a degree and the resulting careers are very de- poral bounds around the cases under analysis (Creswell, 1998;
manding (Li et al., 2008). More recently, Jones et al. (2010) found Stake, 2006). For this study, the bounds are established as five male
that among freshman, value-related constructs predicted career and six female students enrolled in engineering majors, at a techni-
choice better than expectancy constructs. Additionally, a study cal public institution we refer to as TPub, from fall 2003 through
comparing the possibility of several models for predicting persis- spring 2007. The participants and the context of TPub are de-
tence in engineering found no predictive power for Eccles’ ex- scribed in more detail in later sections. Each participant represents
pectancy-value model (although the authors conceded that the pri- an individual case and each case was analyzed separately before
mary scales used had low internal consistency) (Schaefers, being analyzed together.
Epperson, and Nauta, 1997). With the exception of the last set of
questionable findings, overall these results suggest that values play a A. Introduction of CAEE and APS
very critical role in understanding task persistence. All data used in this analysis were collected as part of a larger
In a review of current findings related to goal theory, motivation, body of work, the Academic Pathways Study (APS) conducted by
and school achievement, Covington (2000) noted that, in contrast the Center for Advancement of Engineering Education (CAEE)

October 2010 Journal of Engineering Education 291


and funded by the National Science Foundation (NSF) grant ESI- complete sets of interviews (i.e., one for each of the four years).
0227558. CAEE represents a multi-institutional collaboration of The eleventh participant, Julie, left engineering in her third year.
more than 40 scholars. Both CAEE and APS have been described Her interview data set included three interviews with the third in-
previously in the literature (Clark et al., 2008; Sheppard et al., terview specifically addressing her choice to leave. Having a total
2004). of eleven participants is consistent with Stake’s (2006) suggestion
of five to fifteen cases to insure sufficient but not overwhelming
B. Context quantity of data.
TPub, a public research university located in the western moun- Participants’ demographic information is shown in Table 1
tain region of the United States, is a highly competitive school at- using pseudonyms. Gender and ethnicity are self-reported in the
tracting top performing high school students as measured by grades interviews and GPA were obtained from participants’ transcripts.
and SAT scores. TPub specializes in undergraduate engineering All students entered college directly from high school and are typi-
education with very few majors outside of engineering. Non-engi- cal college-aged students (by Sax’s (2008) suggestion of typical
neering majors are still math and science-related and include meaning entering college in the age range of 17-19). Participants
Chemistry, Economics and Business, Math, and Computer Sci- self-reported the following engineering majors: Chemical Engi-
ence. The overall undergraduate population is heavily weighted to- neering, Materials Science/Metallurgy, Mechanical Engineering,
wards engineering majors. Students choosing to leave engineering Physics, Civil Engineering, Environmental Engineering, and Pe-
have few options at TPub for alternative majors. troleum Engineering. Participants also represent a range of back-
It is possible that TPub being primarily an engineering school grounds and campus experiences including various financial aid
contributes in two ways to increased persistence in earning engineer- arrangements, a member of an athletic team, a fraternity member,
ing degrees. First, students choosing to enter engineering programs club sport participants, engineering society members, and some
at TPub may be self-selected towards engineering persistence since spent one or two semesters studying outside of the United States.
they know they will have few options of major outside of engineer- While not every possible case is represented, we believe that this
ing. Second, students considering leaving engineering might have an case variation covers a sufficiently broad space to explore the phe-
increased tendency to stay since choosing another major usually nomenon of choosing to pursue an engineering degree while not
means they have to leave TPub entirely. This makes the TPub con- being overwhelming in diversity.
text useful for studying persistence choices by highlighting the ways
that students negotiate choices to stay in engineering even when D. Data Sources
staying is difficult. Understanding how students navigate difficult 1) Primary Data Source: The primary data source for this study
persistence paths and ways that they succeed in engineering can help was individual semi-structured interviews collected during the
us uncover possible ways we need to support and encourage students spring semester each year for four years. A loosely structured frame-
to stay in engineering, not only in contexts like TPub but also in con- work of guiding questions prompted the students to think about
texts where there are more choices in major. Moreover, this study topics central to APS research questions. The design of the inter-
has only one non-persister and therefore is limited in what it can view protocol, described in detail elsewhere (Stevens, O’Connor,
offer with regard to understanding non-persistence. Therefore, we and Garrison, 2005), intentionally included questions that drove
consistently frame our study and its findings as contributing to un- students to reflect on past, present, and future experiences in engi-
derstanding persistence rather than non-persistence. neering to capture changes over time. Two interviewers at TPub
Women are underrepresented at TPub (roughly 18 percent) rel- followed a written APS protocol guide to conduct the audio-
ative to the general population but not necessarily with regard to recorded interviews, which were later transcribed verbatim. Al-
other engineering programs. However, since TPub is predominant- though all interview responses were included in the present analysis,
ly an engineering school, the representation of women in the engi- questions with particular salience for this study include:
neering program is reflective of the entire school. This is unlike 1. Can you tell me how you became interested in engineering?
other universities that may offer programs attracting predominantly 2. Think of other engineering students that you have come
female students thereby raising the overall gender equity. across here. Would you say that in general they are more dif-
Other than Native American students, ethnicity distributions at ferent from you or more similar?
TPub are similar to other engineering programs. While the nation- 3. Think about your professors here at TPub. What would you
al average for Native American students in undergraduate engineer- say they think it means to be a good engineer?
ing majors is 0.6 percent (NSF, 2006), the percentage at TPub is a. How does that fit with your own image of an engineer?
nearly doubled. This may result from TPub being located proximal 4. Okay, let’s imagine it’s a few years from now, and you’ve
to areas with large Native American populations. One APS partici- graduated with a degree in (student’s planned major).
pant, Mark, reported being of Native American descent. a. What’s next for you? (if not engineering explore why they
have made this choice)
C. Participants b. What do you imagine yourself doing on a day-to-day
The subset of APS data used for the analysis described herein basis?
included longitudinal interview and survey data for eleven stu- c. What would you say it takes to be a good (insert student’s
dents (five males and six females) at TPub. These 11 participants career choice)?
represent a purposeful sample of the 40 TPub APS participants. Interview questions were generally the same from year to year.
In accordance with APS design, 16 of the 40 TPub APS partici- However, minor modifications recognized that questions had been
pants engaged with researchers in semi-structured interviews each asked in previous years but that the interviewers wanted to explore
year for all four years. Of these sixteen participants, only ten had them in more detail.

292 Journal of Engineering Education October 2010


Table 1. Participants' demographic information.

2) Additional Data Sources: Case study analysis often uses multi- (Matusovich and Streveler, 2009). Miles and Huberman (1994) de-
ple data sources (Creswell, 1998; Stake, 2006; Yin, 2003), and scribe three categories of data analysis strategies including: a) case-
therefore this analysis also incorporated information obtained oriented, where the individual cases are of central importance; b)
through academic transcripts and informal conversations. Academ- variable-oriented, where the variables identified across the cases are
ic transcripts provided GPA information which can be considered of central importance and details of the cases may be lost; and c)
an indication of academic performance. Seven of the eleven partici- mixed strategies, where a combination of the two is used. This
pants also participated in informal conversations. Unlike the semi- study incorporated a mixed strategy where the eleven individual
structured interviews, these informal conversations did not follow cases and the four STV categories (value constructs) were equally
even a loosely scripted list of questions. This allowed greater free- important. Results are presented as variable-oriented assertions
dom for participants to discuss topics of importance to them and for such that they are organized around the four STV categories in-
the interviewer to continue probing. These conversations were also cluding attainment, cost, interest, and utility.
recorded and transcribed verbatim. There are a total of 22 informal This analysis began as part of a broader research project with 10
conversations for seven participants. Not all participants engaged in cases (participants) that examined competence, value, and identity
recorded informal conversations because APS was designed to have beliefs (Matusovich, 2008). Building on that project and focusing
differing levels of interaction with different groups of participants. on values, this analysis includes an eleventh case. Consistent with
The students having informal conversations had the greatest case-study methods, findings from the original analysis were tested
amount of interaction with the research team. The informal conver- by adding a new case with distinguished characteristics (Kirk and
sations provided supplementary information and were used to tri- Miller, 1986; Stake, 2006; Yin, 2003). The added case represented
angulate findings from the primary data source (semi-structured in- a non-persister, Julie who left engineering and TPub after her third
terviews with all participants); in no cases were the informal year in an engineering major. The description of methods reflects
conversations found to contradict information provided in the the process of adding an additional case.
semi-structured interviews. Therefore, it is believed that there is no Starting with a case-based approach, interviews (a total of 40; 4
negative impact on the overall results of this study as a consequence for each of the 10 participants excluding Julie), were read repeat-
of having informal conversations available for some but not all of edly and then coded using Atlas Ti software. Initially we used
the participants. open-coding strategies, developing codes inductively from the
data (Patton, 2002) with expectancy-value theory in mind. The
E. Data Analysis resulting initial list of codes was refined by examining the codes
Our data analysis started with the semi-structured interviews, for uniqueness and combining codes where sufficient overlap ex-
which are considered the primary data source. We followed an isted. We then applied this refined list to all 40 interviews to in-
analysis approach similar to the methods described by Miles and sure consistency across cases.
Huberman (1994). First, the interviews were grouped by case, i.e., This initial coding pass maintained a focus on the participants’
each of the eleven participants represents a case and each case in- words. This allowed us to develop operationalized definitions of
cludes four semi-structured interviews (three for Julie). Then, each Eccles’ value categories that are reflective of the participants’ per-
case was analyzed separately before looking across cases for themes spectives. By “operationalized definitions” we mean that we created
and patterns. Our analysis included a combination of coding output useful, measureable definitions relevant to this study but that corre-
from Atlas Ti and pictorial and graphical data representations spond to definitions given in the literature. For example, in one

October 2010 Journal of Engineering Education 293


source Eccles (2005) defines attainment value as “the value an activi- increasing undergraduate years. A meta-matrix graphic display
ty has because engaging in it is consistent with one’s self-image” (Miles and Huberman, 1994) was created for cross-case analysis.
(p. 139). Our operationalized definition translates “the activity” to We grouped, sorted and organized the information presented in the
“choosing an engineering career” such that in our study, attainment individual tabular displays into a single display for each variable, e.g.,
value is “a reason for pursuing (or not pursuing) engineering that is data related to attainment values for all participants across all years
related to being the type of person who is an engineer.” This broad- was organized into a single table. This allowed cross-case patterns to
er code encompasses multiple inductively developed codes which emerge. By organizing the tabular displays based on the variables, we
encompass the characteristics the participants attribute to engi- shifted our analysis to a variable-based approach in which the four
neers. For example, codes such as “being good at math and science” value categories had priority over the individual cases.
or “being a problem-solver” are attainment values because they de- As a final step in the analysis, we rated each participant with
scribe choosing engineering because one is good at math and sci- regard to each value category for each of the four years. Ratings of
ence, or problem-solving, and believes that being an engineer high (H), moderate (M), and low (L) were used. High attainment
means being good at these things. The other three value categories means engineering is consistent with sense of self. High cost repre-
were operationalized in a similar manner. All 40 interviews were re- sents making sacrifices to become an engineer. High interest in-
coded with the broad codes in place to insure that all relevant sec- cludes enjoying activities believed to be associated with engineer-
tions of responses were appropriately captured. Additionally, as part ing. High utility is the perceived future usefulness of an
of the data triangulation, the set of informal conversations (a total of engineering degree. The ratings represent a combination of fre-
22, for 7 participants) were coded using the operationalized a priori quency and quality of coded segments. This combination balances
codes and the inductively developed codes. participants who talk more with those who say less. For example,
In addition to data triangulation, this analysis incorporated re- a participant who repeatedly mentions a value would be rated as
searcher triangulation. The first author was responsible for most of “High” for the appropriate category. Ratings of “Low” mean that
the coding. The first author trained an additional researcher, only a particular value was mentioned infrequently or not at all. Corre-
mildly familiar with the project, in using the codebook by coding spondingly, “Moderate” ratings are in the middle. Alternatively, a
two interviews with the primary author and discussing code appli- participant who uses minimal words and has less of a tendency to
cations. The additional researcher then independently coded one repeat himself or herself but gives an emphatic response related to
interview from each participant (third year interviews were used) a particular value category would also be rated “High” for that cat-
using the established code book. The author and secondary re- egory. As shown in the results section, this rating system made
searcher then discussed any coding discrepancies until an agreement patterns across variables and cases highly visible.
was reached. This resulted in a final clarification of codes. The final Analysis proceeded similarly for Julie’s case, which includes three
code list, including literature definitions, operationalized defini- interviews. The stabilized list of codes was applied to each interview,
tions and examples is shown in Table 2. outcomes were added to the meta-matrix, and ratings were assigned
Following methods outlined in Miles and Huberman (1994), the for each value category. We recognized there could be potential bias
coding output for each case was examined by creating tabular dis- by having researchers familiar with the outcomes from analysis of the
plays. Using one table for each case, quotes were summarized and previous ten interviews analyzing the eleventh interview. Therefore, a
displayed in columns by value category. Rows were organized by third researcher coded Julie’s interviews. This researcher was trained

Table 2. Final code list applied across all interviews showing code definitions and examples of code applications.

294 Journal of Engineering Education October 2010


using the final codebook and interviews from two of the ten previous- were not fictitious and were based on interview data. Rather than tar-
ly analyzed cases. The author and third researcher coded segments to- gets for marketing purposes, our personas were holistic representations
gether and independently until the third researcher was comfortable of the participants intended to mitigate researcher bias with virtual
with the codebook. They then coded all three of Julie’s interviews member-checking and hypothesis testing during data analysis.
separately and assigned value-category ratings. Discrepancies were In creating these pictorial personas, the first objective was to dis-
discussed until agreement was reached. play salient information about the participants in an “easy-to-see”
format. The second objective was to represent the participants
F. Research Quality holistically, i.e., as complete people and personalities having experi-
Validity and reliability are measures of research quality (Patton, ences beyond those specifically related to the analysis. Although
2002) and are as important to qualitative research as they are to seemingly contradictory, we accomplished both objectives through
quantitative research (Yin, 2003). Strategies used to enhance validi- the systematic creation and organization of the posters to include
ty and reliability were mentioned throughout the methods section participants’ described identities, future career goals, internships,
and are summarized in Table 3. One additional strategy, virtual and other important on and off-campus experiences. Each poster
member-checking, is discussed following the table. The definitions entry was color-coded by year to visually highlight developmental
in Table 3 are based on Yin’s (2003) discussion of measures of qual- changes. All posters were organized to have the same types of infor-
ity as applicable to case study research. mation in the same places to create uniformity.
1) Virtual Member-Checking: Member-checking, a procedure As the primary analyst, the first author took great efforts to identi-
that can mitigate researcher bias, is a process where the participant fy and continually acknowledge her biases by making a concerted ef-
reviews the researcher’s interpretation of the interview (Creswell, fort to see aspects of herself and her story in her participants. Promi-
1998; Patton, 2002). However, member-checking was not possible nent pictorial personas displayed in her primary research workspace,
in this study. The longitudinal nature of the study with analysis oc- provided a visual reminder to consciously and continuously ask herself
curring a year or more after all data were collected meant that verify- if she was telling her story or theirs and if her assertions were consis-
ing interviews after such a long delay would provide limited mean- tent with her participants’ stories. As themes developed during analy-
ingful information. Moreover, the annual interviews were designed sis, they could easily be checked against the representations of the par-
to allow participants to reflect on their experiences over time ticipants to assess goodness of fit with each case.
(Stevens, O’Connor, and Garrison, 2005) perhaps making it more
difficult for participants to reconstruct a previous mindset. In lieu of
actual member-checking, the researchers developed a means for vir- IV. RESULTS
tual member-checking.
We created poster-sized representations of the participants. We Framed in motivational theory, this research proposed to an-
called these representations “pictorial personas.” Described in great de- swer the question: How do students’ engineering-related values
tail previously (Matusovich and Streveler, 2009) the creation and pur- contribute to their choices to engage and persist in earning engi-
pose is summarized here. Pictorial personas were modeled after a user- neering degrees? In answering that question, this research shows:
centered design approach where fictitious persona representations of • all four of Eccles’ value categories are represented in partici-
clients/customers become the ideal person for whom the product is pants’ choices to earn engineering degrees, although attain-
being designed (Pruitt and Adlin, 2006). In our study the personas ment value plays a prominent role

Table 3. Definitions for measures of quality and descriptions of implementation in this qualitative research study.

October 2010 Journal of Engineering Education 295


• the four categories are not mutually exclusive A. Attainment Values
• distinct patterns within values emerged Not surprisingly, the data show that students choose to en-
The results are summarized in Figure 2, followed by details of gage and persist in earning engineering degrees for a variety of
each category. When participant quotes are included in the results, reasons. Participants’ reasons always connect to an attainment
they are followed by an indication of the participant’s name, the in- value, i.e., a sense of self. However, this sense of self is not always
terview type with semi-structured interviews denoted by “SS” and engineering-related. Consider the following two examples. Joe
informal conversations denoted by “INF”, and an indication of the considers himself to be an “engineering-type person.” During an
semester and undergraduate interview year. informal conversation (INF) in his second year he says,
Recalling the methods section, participants were rated
(Figure 2) each year with regard to expressions of value used in de- I mean it is what I like doing. It’s what I do. And, it’s just
scribing choices to be engineers. Note that Julie has no ratings for her rather convenient for me to think that I can get a job doing
fourth year since she left engineering and TPub after her third year. it. (Joe, INF, Spring 2nd year)
High (H) attainment means engineering is consistent with sense of
self, i.e., consistent with the type of person each participant thinks he While also showing an interest value, this quote demonstrates
or she is. Low (L) attainment represents engineering as inconsistent Joe’s belief that an engineering career matches who he is as a person.
with sense of self. High cost represents sacrifice in becoming an engi- In contrast, consider Anna. Anna views an engineering degree as a
neer while low cost means the participant perceives he or she is mak- base for other more interesting future career options. In the spring
ing no sacrifices in becoming an engineer. High interest equates to of her first year she says in a semi-structured interview (SS),
high enjoyment associated with engineering activities while low
interest is a lack of enjoyment associated with engineering activities. …engineering, uh like I see it as a tool for what I wanna
High utility is the perceived future usefulness of an engineering de- do… I will learn engineering and physics because it is going
gree whereas low utility is a lack of usefulness. Moderate (M) ratings to lead me into what I’m doing. (Anna, SS, Spring 1st year)
fall in the middle in each category.
The figure highlights several important patterns in the data This statement is part of a much longer response (several min-
which are summarized here and expanded upon in following sec- utes) where Anna lists many potential future career options that she
tions. All participants are categorized as having either high or low sees as more consistent with the person she considers herself to be.
attainment values; these ratings remained consistent and unchang- For example, she describes using her engineering degree to qualify
ing across all four years. There were no ratings of moderate for this her for advanced degrees culminating in work in psychology or biol-
value category. For participants with high attainment values, interest ogy. It is this pathway to other careers that she refers to when saying
values stay the same or increase with time and cost values are consis- “what I’m doing.” Anna sees an engineering degree as useful for fu-
tently low. Conversely, for these same participants, utility values stay ture, non-engineering career plans but does not see engineering as
the same or decrease with time. For participants with low attain- consistent with her sense of self.
ment values, cost and interest values differ for each participant and In theory, attainment value is how important one perceives a task
utility values stay the same or increase. to be and how consistent that task is with his or her sense of self

Figure 2. Changes in participants' engineering persistence-related values (by category) with time across four undergraduate years.

296 Journal of Engineering Education October 2010


(Eccles, 2005; Eccles et al., 1983; Wigfield and Eccles, 1992). Op- Even as a fourth year student, Tim doubts if engineering is the
erationalized for this study, attainment values are reasons for pursu- right career choice for him. Similarly, Anna initially pursues engi-
ing (or not pursuing) engineering degrees that are related to percep- neering as a foundation for a career in another field and, like Tim,
tions of the self as an engineer. These perceptions of the self as struggles with fit. Leslie believes she wants a career in engineering
engineers are similar to professional identities (e.g., Ibarra, 1999) or then finds it is not a good fit for her because she doubts her abilities.
engineering roles (Davis, Beyerlein, and Davis, 2005). By this defi- She does not believe she has the “mechanical intuition” which is in-
nition, Joe has a high attainment value and Anna has a low attain- herent in engineers. Conversely, Marie does not talk about how en-
ment value. Joe and Anna experience the continual choice process to gineering matches or does not match her sense of self; she plans to
pursue engineering very differently, as do their peers who have simi- pursue a graduate degree and has postponed career choices until she
larly rated attainment values. As mentioned, all participants are reaches graduate school. In talking about choosing engineering,
rated as either high or low in engineering-related attainment value; Julie also talks about aspects of teaching that are more and less con-
no participants had moderate ratings. sistent with her sense of self:
Hillary, Mark, Will, Max, and Beth also have high attain-
ment values for being engineers. Hillary thinks of herself as a Julie: Um, well when I was little I wanted to be a teacher
problem-solver and looks forward to working in the field solv- but I mean…
ing the real-time problems that arise during operation. Mark Int: Talk about that?
thinks of himself as a designer and initially hopes to design au- Julie: The being a teacher? [Int: Mm-hmm.] I don’t
tomobile parts in the future. Will is a designer and builder and know, it was probably cause like um, there’s a gap
he thinks of engineering as the career name for his hobbies. between me and like my next cousin, so like I was
Max describes himself as a materialistic person and he believes the oldest among the other cousins, so I really liked
engineers are well-paid. Beth enjoys math, science, and hands- teaching them how to do things, and I kinda like
on activities and believes those are characteristics of engineers. being in charge for some odd reason but, as I got
For all six of these participants, their beliefs about being engi- older like I just couldn’t see myself being
neers match beliefs they hold about themselves; they have at- responsible for a lot of kids and their education so,
tainment values related to being engineers. and then I really got into like the math and the
As shown in Figure 2, five participants, Anna, Marie, Leslie, science and, and ‘Julie, what do you like?’ ‘Math
Tim, and Julie, have low ratings for attainment values; how they see and science.’ ‘Oh you should be an engineer.’ ‘Oh
themselves may not match how they see engineers. For example, okay.’ So you know like never really thought about
Tim waivers in choosing a career in engineering. Tim thinks of anything else really. (Julie, SS, Spring 1st year)
himself as a businessman and pursues engineering primarily because
he believes it will help him be a better businessman in the future. In Julie describes enjoying teaching activities. The responsibility
his third year, Tim describes how he initially became interested in teachers have is inconsistent with her sense of self so she was direct-
engineering, ed towards engineering out of interest and skill with math and sci-
ence courses. Julie’s passage suggests engineering is a default career
I didn’t think I could go straight into business because rather than something she is drawn to because of the types of work
that is not my personality. And, it was kind of my engineers do.
strength but like I thought I wanted more. (Tim, SS,
Spring 3rd year) B. Cost Value
According to Eccles (Eccles, 2005; Eccles et al., 1983; Wigfield
In this excerpt, Tim describes needing something “more” and Eccles, 1992), cost is the time, effort, and psychological price of
than a business major, and he believes engineering provides that succeeding or failing in a given task. In this study, costs associated
added something. While also encompassing one of Tim’s utility with being an engineer or not being an engineer in the future are
beliefs, i.e., that a background in engineering as something use- considered. For example, an individual might view working long
ful for working in business, Tim describes choosing engineering hours on an isolated oil rig as a cost of being a petroleum engineer or
because he is the type of person that needs something more than they might view failing to become a petroleum engineer as a missed
a straight career in business could provide. Although he intends opportunity to work in a foreign country.
to complete his degree and practice engineering, in his fourth The costs of being a practicing engineer are different from the
year Tim expresses doubts that engineering is consistent with costs of being an engineering student. For example, the costs of being
his sense of self, an engineering student might include heavy course loads or the
emotional and psychological toll associated with the financial bur-
Like the theory, what makes things work, taking things den of paying for engineering courses. This study acknowledges
apart, and figuring all this stuff out, and all the intricacies, that there are costs of being an engineering student but focuses on
sometimes I just don’t care about that. I’m like, ‘OK so the costs of being a practicing engineer. This is necessitated by insuf-
these guys are engineers, that’s what their thing is. Why am ficient information to distinguish the costs of being an engineering
I an engineer?’ And I look at like where I wanna’ do kind of student from being a college student. For example all participants
with my career, and you see this shift away from mention at least once having a difficult semester, not having enough
engineering. And I ask myself, ‘Well so why did I get an free time, having a hard class, etc. However, the challenges they
en… [sic], involved in engineering in the first place?’ (Tim, mention could be associated specifically with being engineering stu-
SS, Spring 4th year) dents or they could be associated with attending a competitive

October 2010 Journal of Engineering Education 297


school like TPub. Therefore, for this study, costs are limited to those C. Interest Value
associated with a career in engineering. Interest value refers to doing a task for the enjoyment experi-
Two different patterns emerge across the eleven participants enced while doing the task (Eccles, 2005; Eccles et al., 1983;
with regard to the cost values associated with being an engineer Wigfield and Eccles, 1992). Interest is operationalized in this study
and with attainment values. The six participants shown in Figure 2 as the enjoyment experienced in performing activities perceived to
rated with high attainment values are also rated as low with regard be engineering activities and/or enjoyment related to becoming an
to the importance of relative costs in their choices to become engi- engineer in the future.
neers. In fact, these participants generally do not talk about costs As shown in Figure 2, the six students with high attain-
related to a career in engineering. ment values also have moderate to high interest levels. Hillary,
The five participants shown in Figure 2 with low attainment val- Will, Beth, and Joe have a high interest level in engineering
ues have different cost rating patterns from each other and some have that is unchanging over their four years. Max and Mark have
cost ratings which change with time. Marie does not talk about the moderate interest values that increase over their four years be-
costs of being or not being an engineer extensively. Anna, Tim, coming high by the end. Following is an example quote from
Leslie, and Julie do consider costs as they consider if becoming an Hillary:
engineer is the right career choice. Some examples of the costs ex-
pressed by Anna and Leslie are given here. I used to go down there after class and help [name] set up
Anna has a love of art and believes that choosing a career in engi- for labs and stuff, and that’s kind of where I really fell in love
neering potentially means not having time for the art which she en- with it, like how things worked, because I’d go and there
joys in her spare time, would be bowling balls hanging from the ceiling and we had
ramps and tracks and everything and I’d just get to go down
Well there’s a constant struggle with me and art. Because I there and play… (Hillary, SS, Spring 1st year)
love it. And it’s a part of who I am. And I wish I could do it
more, you know? (Anna, SS, Spring 4th year) In this quote, Hillary describes pre-college enjoyment for what
she perceived as an engineering-related activity, i.e., investigating
Leslie fears that choosing engineering means choosing a career how things work. She describes her time in the lab as “play,” an ex-
she will not enjoy, pression of enjoyment.
Anna, Tim, Leslie, Marie, and Julie, the five participants with
And I kind of feel like I’m not really going to enjoy it if I low attainment value ratings, have low to moderate ratings for in-
actually go into engineering. (Leslie, INF, Fall 3rd year) terest values. There is not a discernable pattern within these five
participants related to changes in interest levels. One increases,
While this is also an interest value, Leslie expresses this in a cost for- one remains constant, two decrease, and one increases then re-
mat; a potential cost of being an engineer is not enjoying her work. turns to starting level. These participants express interest in engi-
Julie does not mention the costs associated with being an engi- neering activities less frequently than the other six and often with
neer until her third year when she has decided to leave engineering less passion. For example, unlike Hillary’s quote, these partici-
and TPub to pursue a career in teaching. For Julie, part of the cost of pants do not easily use words like “love” or “play.” While Leslie
becoming an engineer is being someone that she is not. In her third has typically expressed a generic or matter-of-fact interest in engi-
year, she describes her choice to leave: neering, in this quote she realizes she does not have the deeper
level of interest or “passion” she believes is needed to be happy in
…well another thing that scared me was I knew I was an engineering career:
graduating in like a year and a half and it just scared me that
I couldn’t be an engineer. Like I don’t see myself fitting in And, I just really realized you have to be really passionate
there anymore (Julie, SS, Spring 3rd year) about what you do. ‘Cause it’s not about your 40-hour week
and then it’s over. Like, usually you’re gonna’ be working a
Also related to attainment value, being an engineer has a psycho- lot more [I: Um-hum.] and you should be excited about
logical cost for Julie associated with her identity. Additionally and what you’re doing. [I: Um-hum.] And, I kind of feel like
much like Anna, being an engineer also has a cost associated with a that I’m not really going to enjoy it if I actually go into
perceived inability for Julie to simultaneously pursue dance-related engineering. (Leslie, INF, Fall 3rd year)
activities. In her third year, Julie describes how teaching will allow
her to also dance while being an engineer might not: This realization comes during her first internship when Leslie
finds she does not have passion for the type of engineering work she
But, I think that’s another reason that I wanna’ go into and those around her are doing.
teaching is because there’s you know, like dance teams, and
drama. And that maybe I could choreograph musicals. D. Utility Value
Whatever. I mean, I feel like I have more opportunity being According to Eccles, utility value incorporates the perception of
a teacher than I do as an engineer. the future use of engaging in the task (Eccles, 2005; Eccles et al.
1983; Wigfield and Eccles, 1992). For this study, utility value is the
Dance is one of Julie’s hobbies and she perceives that being a perceived future use or advantage of becoming an engineer and the
teacher will afford her a better opportunity to continue her involve- usefulness of an engineering degree. The following are examples of
ment in dance than engineering. how students expressed utility values:

298 Journal of Engineering Education October 2010


‘Cause I wanna’ make money. That’s a huge goal of mine, is tend to have moderate or high cost values, have low or moderate in-
money. That’s such a large motivator for me. (Max, SS, terest values and have consistently moderate or high utility values.
Spring 4th year) The one student in this study who left the engineering program had
a low attainment value towards engineering. These relationships,
I decided engineering because it’s more applicable. I think it also shown in Figure 3, are intended as a way to summarize the gen-
has more benefits to humanity than just straight science or eral patterns that emerged from the data not as direct correlations.
more immediate benefits. (Hillary, SS, Spring 1st year) Notably, Max is an exception having a high attainment value and
also a consistently high utility value across all four years. Important-
For Max, an engineering degree is useful because it leads to a ly though, his interest value still grows from moderate to high by his
profitable career. For Hillary, an engineering career is useful to hu- fourth year showing that it is not high utility alone contributing to
manity. For students with low attainment values, an engineering de- his persistence.
gree may be seen as providing a useful back-up career:

…if you have an engineering degree and you are an artist, V. DISCUSSION
then, you still have something to fall back on. (Anna, SS,
Spring 2nd year) A. Values and Persistence Patterns
Career decision-making models, such as those by Holland
Anna believes that if she finished her engineering degree then (1997), Lent, Brown, and Hackett (1994), and Super (1980,
pursued her art passion, she could fall back on a career in engineer- 1990) highlight the importance of interest-based choice. In their
ing if her career in art was not successful. longitudinal study on persistence, Seymour and Hewitt (1997)
As shown in Figure 2, for participants with high attainment also highlight interest as salient to career decisions. In contrast, the
value ratings, utility values stay the same or decrease with time, results of this current study show a career choice process hinging
while the opposite is generally true of the utility values for partici- on engineering as being consistent with students’ sense of self (at-
pants with low attainment values. Julie is an exception and across all tainment value) rather than interest. Moreover, this study confirms
three years has a low utility value for engineering. In the first two that one value category alone is not enough to explain persistence;
years, Julie does not express a utility value and in the third mentions while Seymour and Hewitt (1997) found that low interest when
that an engineering degree would not be useful to her because she coupled with other factors contributed to persistence, this study
does not intend to practice engineering. In her third year Julie sug- finds that low attainment values coupled with high utility value
gests that continuing in engineering would not be useful. and/or moderate interest can lead to persistence. Situations in
which attainment, interest, and utility values are low and cost values
And, I’m gonna’ get a degree [in engineering] that I am are high challenge students’ intentions to persist the most. For ex-
probably not going to use. And, I’d rather backtrack and get ample, students such as Anna, Marie, and Tim seem likely possi-
a degree [in teaching] that I want. It’s not just the fact that bilities to leave engineering based on reporting little connection
I’m just going to get a degree. I’m going to get a degree that between engineering and their sense of self (low attainment value).
I’m gonna’ use. And, I’ll be proud of… However, they persist because they have other reasons for wanting
to be engineers. For example, Anna feels pressure to earn a high
Julie believes that a degree in teaching will be much more useful paying salary and help support her parents. The perceived useful-
to her than an engineering degree. ness of an engineering degree helps propel her along towards
E. Summary graduation. On the contrary, Julie, who has low attainment value,
In summary, this research finds that participants can be catego- generally low interest and low utility value for engineering, opts to
rized as having either high or low engineering-related attainment leave engineering (and TPub) during her third year. Julie does not
values. These attainment values remain consistent and unchanging have a strong value sustaining the pursuit of an engineering de-
over their four-year academic careers. Participants with high attain- gree.
ment values all have low cost values, have consistently high or in- This research does not suggest that reasons for pursuing engi-
creasing from moderate to high interest values and have consistent neering degrees that are associated with low attainment values,
or decreasing utility values. Participants with low attainment values such as those of Tim, Anna, Marie, Leslie, or Julie, are invalid or

Figure 3. Relationships among value constructs derived through this study, showing the central importance of attainment value.

October 2010 Journal of Engineering Education 299


are somehow less meaningful than high attainment-value-related C. Implications for Research and Practice
reasons. Instead, this research shows that students with low attain- Smart, capable students may leave engineering fields (Seymour
ment values experience the persistence process differently and with and Hewitt, 1997) and this study contributes to understanding pos-
greater difficulty. Tim, Anna, Marie, and Leslie frequently renego- sible reasons. While Anna, Leslie, Marie, Tim, and Julie may not
tiate or rejustify the choice to be (or not be) an engineer. Only Julie be at risk for leaving college based on academic achievement, they
ultimately decides not to pursue an engineering degree. Perhaps may be considered the highest risk for leaving engineering due to
students with low attainment values are the students on whom engi- their limited connection between engineering and personal sense of
neering educators and policy-makers can have the greatest impact; self. Generally, the students with low attainment values also have
these students show resiliency and tenacity and actively look for rea- low competence beliefs and/or uncertainty about becoming engi-
sons to stay in engineering. neers (Matusovich et al., 2008; Matusovich et al., 2009b, 2009a). It
Moreover, this research highlights individual differences in val- is entirely possible that had TPub offered degrees in other fields,
ues. While there are differences in ratings across categories, there Anna, Leslie, Marie, and Tim might have changed majors rather
are differences within categories. For example, utility values can take than persisting in engineering. Therefore, the findings of this study
on many different forms, such as Max’s desire for a high paying job could be even more significant in different academic situations.
and comfortable life and Anna’s desire for a high paying job to help While answering questions about the development of engineer-
support her parents and siblings. While both related to earning ing-related task values over time, and the relationships among dif-
money, these are very different yet equally powerful personal moti- ferent categories of such values, this research also raises a number of
vators for earning engineering degrees. To persist in earning an en- important questions. First, this research focuses on students who
gineering degree, not every student must be motivated in the same persist in earning an engineering degree as opposed to students who
way by the same things. By showing the great variety in the values leave engineering. With only one non-persister included in this
that students assign to earning an engineering degree, this research study, insights into choices to leave engineering are limited. To gain
suggests there are many ways to reach students and help them con- a more complete understanding of value patterns as related to per-
nect to the personal possibilities an engineering degree could bring. sistence in engineering, future research should examine value pat-
terns associated with choices to leave engineering majors.
B. Gendered Patterns Second, this research scope was intentionally limited to relation-
In this study, more women than men participants experienced a ships between task values and the choice to become an engineer.
lack of connection between their engineering-related values and This simplified and isolated a portion of the Eccles model for deep
sense of self (low attainment values). Four of the six women partici- investigation. It is now important to resituate what has been learned
pants (Anna, Marie, Leslie, and Julie) experienced low attainment about value beliefs back into the complete model to fully understand
values. Of the five men in this study, only Tim (notably an ethnic the implications for the findings. For example, understanding
minority) experienced low attainment values. Seemingly contradic- sources of, and influences on, value beliefs, such as socializers (e.g.,
torily, these women (other than Julie) generally had higher GPAs parents, faculty, and peers), gender identity, ethnic identity, and
than the men, including two women with low attainment value for cultural influences, could help us better understand ways educators,
engineering having cumulative GPAs over 3.9. While it would be advisors, and policy-makers can influence value beliefs. It is also im-
inappropriate to generalize gendered findings based on the limited portant to reconnect value and competence beliefs by directly prob-
data presented herein, it would be equally inappropriate to ignore ing the relationship between the two. For example, prior research
the qualitative pattern that emerged through this study. with these same participants shows changes in competence beliefs
These findings related to gender differences in values and aca- with time and in relation to campus and internship experiences
demic achievement are consistent with current literature. For ex- (Matusovich et al., 2008; Matusovich et al., 2009a). We should
ample, research shows that women, both persisters and non-per- specifically target understanding the causes of changes in compe-
sisters, are more likely than men to lose interest in the various tence and value beliefs. Moreover, it is important to understand
STEM disciplines with time (Brainard and Carlin, 1998; Sey- how specific classroom and program practices contribute to chang-
mour and Hewitt, 1997). Also, female engineering students have ing and developing value beliefs.
a greater perceived lack of fit than male engineering students The findings from this study suggest a need to increase stu-
(Heyman, Martyna, and Bhatia, 2002) where perceived lack of fit dents’ attainment values related to engineering in order to increase
is similar to attainment values as defined herein. Finally, Sax persistence. In other words, we can encourage students to stay in
(2008) found that women college students tend to attain higher engineering by helping them associate a perceived engineering
grades than men from their final year of high school through col- identity with their personal identity and demonstrating the value
lege. Consistent with current literature, our research continues to of this association. Doing this requires an understanding of what
highlight the existing apparent contradiction that women have students value and then connecting this to one or more of the
higher grades but lower values for engineering. This also rein- many different ways of practicing engineering. We must help stu-
forces Pascarella and Terenzini’s (2005) findings that “college dents understand what it means to be an engineer not only by
grades may well be the single best predictor of student persistence, teaching a variety of engineering skills, but also by exemplifying
degree completion and graduate school enrollment” (p. 396). Al- the breadth of activities engineers perform in their daily work.
though the women doubt their choice, they have high grades and Perhaps this is another way of saying that engineers work in many
they persist. This research offers insight into understanding this varied jobs and situations as they participate in a wide variety of
apparent contradiction. By examining multiple categories of val- activities. We need to make sure students do not develop a narrow
ues and the interplay between such values, we can see that these view of engineering that omits activities and goals that might par-
high-achieving women find pathways to persistence. ticularly interest them as individuals. We must continually (over

300 Journal of Engineering Education October 2010


all four years) and explicitly reference various possible engineering Denzin, N.K., and Y.S. Lincoln. 2003. Strategies of qualitative inquiry.
identities. Authentically exposing students to a variety of engi- Thousand Oaks, CA: Sage Publications.
neering career possibilities and activities would create opportuni- Eccles, J.S. 1986. Gender-roles and women’s achievement. Educational
ties for students to find such connections. While important for Researcher 15 (6): 5–19.
both genders, the lower attainment values among women and Eccles, J.S. 2005. Subjective task value and the Eccles et al. Model of
greater uncertainty about engineering and engineering abilities Achievement-Related Choices. In Handbook of competence and motivation,
found in women suggest it is even more important to develop in- eds. A.J. Elliot and C.S. Dweck. New York: The Guilford Press.
terest and competence beliefs in women to retain them in engi- Eccles, J.S. 2007. Families, schools, and developing achievement-related
neering fields. motivations and engagement. In Handbook of socialization: Theory and
research, eds. J.E. Grusec and P.D. Hastings. New York, NY: Guilford Press.
Eccles, J.S. 2009. Who am I and what am I going to do with my life?
ACKNOWLEDGMENTS Personal and collective identities as motivators of action. Educational Psy-
chologist 44 (2): 78–89.
This material is based on work supported by the National Sci- Eccles, J.S., T.F. Adler, R. Futterman, S.B. Goff, C.M. Kaczala, J.L.
ence Foundation under Grant No. ESI-0227558, which funds the Meece, and C. Midgley. 1983. Expectancies, values, and academic behav-
Center for the Advancement of Engineering Education (CAEE). iors. In Achievement and achievement motivation, ed. J.T. Spence. San
Any opinions, findings and conclusions or recommendations ex- Francisco, CA: W. H. Freeman.
pressed in this material are those of the author(s) and do not neces- Eccles, J.S., B.L. Barber, and D. Jozefowicz. 1999. Linking gender to
sarily reflect the views of the National Science Foundation. CAEE educational, occupational, and recreational choices: Applying the Eccles
is a collaboration of five partner universities. et al. Model of Achievement-Related Choices. In Sexism and stereotypes in
The authors would also like to acknowledge Aidsa Santiago and modern society: The gender science of Janet Taylor Spence. Washington, DC:
Katherine Winters for assistance with inter-rater coding and the re- American Psychological Association.
viewers from this journal for their helpful comments. Eccles, J.S., and A. Wigfield. 1995. In the mind of the actor—the
structure of adolescents achievement task values and expectancy-related be-
liefs. Personality and Social Psychology Bulletin 21 (3): 215–25.
REFERENCES Eccles, J.S., A. Wigfield, R.D. Harold, and P. Blumenfeld. 1993. Age
and gender differences in childrens’ self and task perceptions during ele-
Alias, M., and N. Hafir. 2009. The relationship between academic self- mentary-school. Child Development 64 (3): 830–47.
confidence and cognitive performance among engineering students. In French, B.F., J.C. Immekus, and W.C. Oakes. 2005. An examination
Proceedings of the Research in Engineering Education Symposium (REES). of indicators of engineering students’ success and persistence. Journal of En-
Palm Cove, Australia. gineering Education 94 (4): 419–25.
Bandura, A. 1997. Self-efficacy: The exercise of control. New York: W.H. Frome, P.M., C.J. Alfeld, J.S. Eccles, and B.L. Barber. 2008. Is the de-
Freeman. sire for a family-flexible job keeping young women out of male-dominated
Battle, A., and A. Wigfield. 2003. College women’s value orientations occupations? In Gender and occupational outcomes: Longitudinal assessments of
toward family, career, and graduate school. Journal of Vocational Behavior 62 individual, social, and cultural influences, eds. H.M.G. Watt and J.S. Eccles.
(1): 56–75. Washington, DC: American Psychological Association.
Bong, M. 2001. Role of self-efficacy and task-value in predicting col- Heyman, G.D., B. Martyna, and S. Bhatia. 2002. Gender and achieve-
lege students’ course performance and future enrollment intentions. Con- ment-related beliefs among engineering students. Journal of Women and
temporary Educational Psychology 26 (4): 553–70. Minorities in Science and Engineering 8 (1): 41–52.
Brainard, S.G., and L. Carlin. 1998. A six-year longitudinal study of Holland, J.L. 1997. Making vocational choices: A theory of vocational per-
undergraduate women in engineering and science. Journal of Engineering sonalities and work environments. Odessa, FL: Psychological Assessment
Education 87 (4): 369–75. Resources.
Clark, M., S.D. Sheppard, C. Atman, L. Fleming, R.L. Miller, R. Hutchison-Green, M., D.K. Follman, and G.M. Bodner. 2008. Pro-
Stevens, R. Streveler, and K. Smith. 2008. Academic Pathways Study: viding a voice: Qualitative investigation of the impact of a first-year engi-
Processes and realities. In Proceedings of the 2008 American Society for Engi- neering experience on students’ efficacy beliefs. Journal of Engineering Edu-
neering Education Annual Conference and Exposition. Pittsburgh, PA. cation 97 (2): 177–90.
Committee on Science, Engineering, and Public Policy (COSEPUP) Hutchison, M.A., D.K. Follman, M. Sumpter, and G.M. Bodner.
and Policy and Global Affairs (PGA). 2006. Rising above the gathering 2006. Factors influencing the self-efficacy beliefs of first-year engineering
storm: Energizing and employing America for a brighter economic future. students. Journal of Engineering Education 95 (1): 39–47.
Washington, DC: National Academies Press. Ibarra, H. 1999. Provisional selves: Experimenting with image and
Correll, S.J. 2001. Gender and the career choice process: The role of bi- identity in professional adaptation. Administrative Science Quarterly 44 (4):
ased self-assessments. American Journal of Sociology 106 (6): 1691–1730. 764–91.
Covington, M.V. 2000. Goal theory, motivation, and school achieve- Jackson, S.A. 2002. The quiet crisis. San Diego, CA: Building Engi-
ment: An integrative review. Annual Review of Psychology 51:171–200. neering and Science Talent.
Creswell, J.W. 1998. Qualitative inquiry and research design: Choosing Jacobs, J.E., S. Lanza, D.W. Osgood, J.S. Eccles, and A. Wigfield. 2002.
among five traditions. Thousand Oaks, CA: Sage Publications. Changes in children’s self-competence and values: Gender and domain dif-
Davis, D.C., S.W. Beyerlein, and I.T. Davis. 2005. Development and ferences across grades one through twelve. Child Development 73 (2): 509–27.
use of an engineer profile. In Proceedings of the 2005 American Society for En- Jones, B.D., M.C. Paretti, S.F. Hein, and T.W. Knott. 2010. An
gineering Education Annual Conference and Exposition. Portland, OR. analysis of motivation constructs with first-year engineering students:

October 2010 Journal of Engineering Education 301


Relationships among expectancies, values, achievement, and career plans. Nicholls, G.M., H. Wolfe, M. Sacre Besterfield, L.J. Shuman, and S.
Journal of Engineering Education 99 (4). Larpkiattaworn. 2007. A method for identifying variables for predicting
Kirk, J., and M.L. Miller. 1986. Reliability and validity in qualitative STEM enrollment. Journal of Engineering Education 97 (1): 33–44.
research, Qualitative research methods. Beverly Hills, CA: Sage Publications. National Science Foundation. 2006. Women, minorities and persons with
Lent, R.W., S.D. Brown, and G. Hackett. 1994. Toward a unifying so- disabilities in science and engineering. http://www.nsf.gov/statistics/
cial cognitive theory of career and academic interest, choice, and perfor- wmpd/pdf/tabc-5.pdf (last accessed, July 2010).
mance. Journal of Vocational Behavior 45 (1): 79–122. Ohland, M, S. Sheppard, G. Lichtenstein, O. Eris, D. Chachra, and
Lent, R.W., S.D. Brown, J. Schmidt, B. Brenner, H. Lyons, and D. R.A. Layton. 2008. Persistence, engagement, and migration in engineering
Treistman. 2003. Relation of contextual supports and barriers to choice be- programs. Journal of Engineering Education 97 (3): 259–78.
havior in engineering majors: Test of alternative social cognitive models. Pascarella, E.T., and P.T. Terenzini. 2005. How college affects stu-
Journal of Counseling Psychology 50 (4): 458–65. dents: A third decade of research. 1st ed, The Jossey-Bass higher and adult ed-
Lent, R.W., A.M. Lopez, F.G. Lopez, and H.B. Sheu. 2008. Social ucation series. San Francisco, CA: Jossey-Bass.
cognitive career theory and the prediction of interests and choice goals in Patton, M.Q. 2002. Qualitative research and evaluation methods. Thou-
the computing disciplines. Journal of Vocational Behavior 73 (1): 52–62. sand Oaks, CA: Sage Publications.
Lent, R.W., H.B. Sheu, D. Singley, J.A. Schmidt, L.C. Schmidt, and Pruitt, J., and T. Adlin. 2006. The personal lifecycle: Keeping people in
C.S. Gloster. 2008. Longitudinal relations of self-efficacy to outcome ex- mind throughout product design, The Morgan Kaufmann series in interactive
pectations, interests, and major choice goals in engineering students. Jour- technologies. Boston, MA: Elsevier.
nal of Vocational Behavior 73 (2): 328–35. Sax, L.J. 2008. The gender gap in college: Maximizing the developmental
Li, Q., D. McCoach, H. Swaminathan, and J. Tang. 2008. Develop- potential of women and men. San Francisco, CA: Jossey–Bass.
ment of an instrument to measure perspectives of engineering education Schaefers, K.G., D.L. Epperson, and M.M. Nauta. 1997. Women’s
among college students. Journal of Engineering Education 97 (1): 47–56. career development: Can theoretically derived variables predict persistence
Marra, R.M., K.A. Rodgers, D. Shen, and B. Bogue. 2009. Women in engineering majors? Journal of Counseling Psychology 44 (2):173–83.
engineering students and self-efficacy: A multi-year, multi-institution Seymour, E., and N.M. Hewitt. 1997. Talking about leaving: Why un-
study of women engineering student self-efficacy. Journal of Engineering dergraduates leave the sciences. Boulder, CO: Westview Press.
Education 98 (1): 27–38. Shavelson, R.J., J.J. Hubner, and G.C. Stanton. 1976. Self-concept:
Matusovich, H.M. 2008. Choosing engineering: Can I succeed and do Validation of construct interpretations. Review of Educational Research 46
I want to? Doctoral Dissertation, Engineering Education, Purdue Univer- (3): 407–41.
sity, West Lafayette, IN. Sheppard, S., C. Atman, R. Stevens, L. Fleming, R. Streveler, R. Adams,
Matusovich, H.M., and R.A. Streveler. 2009. Pictorial personas: A way and T. Barker. 2004. Studying the engineering student experience: Design of
to look at longitudinal qualitative data. Paper presented at Ethnographic and a longitudinal study. In Proceedings of the 2004 American Society for Engineer-
Qualitative Research Conference. Cedarville, OH. ing Education Annual Conference and Exposition. Salt Lake City, UT.
Matusovich, H.M., R.A. Streveler, H.G. Loshbaugh, R. Miller, and B. Sheppard, S., K. Macatangay, A. Colby, and W. Sullivan. 2009. Edu-
Olds. 2008. Will I succeed in engineering? Using expectancy-value theory cating engineers: Designing for the future of the field. San Francisco, CA:
in a longitudinal investigation of students’ beliefs. In Proceedings of the 2008 Jossey-Bass.
American Society of Engineering Education Annual Conference and Exposition. Stake, R.E. 2006. Multiple case study analysis. New York: The Guilford
Pittsburgh, PA. Press.
Matusovich, H.M., R.A. Streveler, R.L. Miller, and B.A. Olds. 2009a. Stevens, R., K. O’Connor, and L. Garrison. 2005. Engineering student
Competence in engineering: A tale of two women. In Proceedings of the identities in the navigation of the undergraduate curriculum. In Proceedings
2009American Society of Engineering Education Annual Conference and Expo- of the 2005 American Society for Engineering Education Annual Conference
sition. Austin, TX. and Exposition. Portland, OR.
Matusovich, H.M., R.A. Streveler, R.L. Miller, and B.A. Olds. 2009b. Super, D.E. 1980. A life-span, life-space approach to career develop-
I’m graduating this year! So what IS an engineer anyway? In Proceedings of ment. Journal of Vocational Behavior 16:282–92.
the 2009 American Society of Engineering Education Annual Conference and Super, D.E. 1990. A life-span, life-space approach to career develop-
Exposition. Austin, TX. ment. In Career choice and development: Applying contemporary theories to
Mendez, G., T.D. Buskirk, S. Lohr, and S. Haag. 2008. Factors associ- practice, eds. D. Brown and L. Brooks. San Francisco, CA: Jossey-Bass.
ated with persistence in science and engineering majors: An exploratory Trenor, J.M., S.L. Yu, C.L. Waight, K.S. Zerda, and T.L. Sha. 2008.
study using classification trees and random forests. Journal of Engineering The relations of ethnicity to female engineering students’ educational expe-
Education 97 (1): 57–70. riences and college and career plans in an ethnically diverse learning envi-
Miles, M.B., and A.M. Huberman. 1994. Qualitative data analysis. ronment. Journal of Engineering Education 97 (4): 449–65.
Thousand Oaks, CA: Sage Publications. Watt, H.M.G., and J.S. Eccles, eds. 2008. Gender and occupational out-
National Academy of Engineering (NAE). 2004. The engineer of 2020: comes: Longitudinal assessments of individual, social, and cultural influences.
Visions of engineering in the new century. Washington, DC: National Acade- Washington, DC: American Psychological Association.
mies Press. Wigfield, A. 1994. Expectancy-value theory of achievement motivation:
National Academy of Engineering (NAE). 2005. Educating the engineer A developmental perspective. Educational Psychology Review 6 (1): 49–78.
of 2020: Adapting engineering education to the new century. Washington, Wigfield, A., and J.S. Eccles. 1992. The development of achievement
DC: National Academies Press. task values—A theoretical analysis. Developmental Review 12 (3): 265–310.
National Academy of Engineering (NAE). 2008. Changing the conver- Wigfield, A., and J.S. Eccles. 2000. Expectancy-value theory of
sation: Messages for improving public understanding of engineering. Washing- achievement motivation. Contemporary Educational Psychology 25 (1):
ton, DC: National Academies Press. 68–81.

302 Journal of Engineering Education October 2010


Wigfield, A., R.D. Harold, C. Freedman-Doan, J.S. Eccles, K.S. Yoon, University of Hawaii at Manoa. Prior to coming to Purdue, she was
A.J.A. Arbreton, and P.C. Blumenfeld. 1997. Change in children’s compe- the founding director of the Center for Engineering Education at
tence beliefs and subjective task values across the elementary school years: A the Colorado School of Mines. She has been a co-PI on several pro-
3-year study. Journal of Educational Psychology 89 (3): 451–69. jects funded by the National Science Foundation and served a stint
Yin, R.K. 2003. Case study research: Design and methods. 3rd ed, Applied as acting director of the Center for the Advancement of Engineer-
social research methods series. Thousand Oaks, CA: Sage Publications. ing Education. She earned the 2006 and 2007 Helen Plants
Awards for best non-traditional session at the Frontiers in Educa-
tion Conference.
AUTHORS’ BIOGRAPHIES Address: School of Engineering Education, Purdue University,
Neil Armstrong Hall of Engineering, West Lafayette, IN, 47907;
Holly M. Matusovich is an assistant professor in the Department telephone: (1) 765.496.9031; fax: (1) 765.494.5819; e-mail:
of Engineering Education at Virginia Polytechnic Institute and State streveler@purdue.edu.
University. She holds a B.S. in Chemical Engineering from Cornell
University, and a M.S. in Materials Science from the University of Ronald L. Miller is a professor of Chemical Engineering and
Connecticut and a Ph.D. in Engineering Education from Purdue director of the Center for Engineering Education at Colorado School
University. Dr. Matusovich has more than 12 years of experience in of Mines. He earned degrees in chemical engineering from the
engineering practice including work as an engineering consultant and University of Wyoming and Colorado School of Mines. Dr. Miller
later in a variety of roles in a manufacturing environment. has received three university-wide teaching awards and has held a
Address: Department of Engineering Education (0218), 626 Jenni teaching fellowship at CSM. He has received grant awards for
McBryde Hall, Blacksburg, VA 24162; telephone: (1) education research from the National Science Foundation, the U.S.
540.231.4205; fax: (1) 540.231.6903; e-mail: matushm@ Department of Education FIPSE program, the National Endowment
vt.edu. for the Humanities, and the Colorado Commission on Higher Edu-
cation and has published widely in the areas of engineering education
Ruth A. Streveler is an assistant professor in the School of Engi- assessment, pedagogy, and curricular design.
neering Education at Purdue University. She holds a B.A. in Biolo- Address: Chemical Engineering Department, Colorado School
gy from Indiana University, a M.S. in Zoology from The Ohio of Mines, Golden, CO, 80401; telephone: (1) 303.273.3892; fax:
State University and a Ph.D. in Educational Psychology from the (1) 303.273.3730; e-mail: rlmiller@mines.edu.

October 2010 Journal of Engineering Education 303

You might also like