Blotnicky et al.

International Journal of STEM Education (2018) 5:22

https://doi.org/10.1186/s40594-018-0118-3
International Journal of
STEM Education

RESEARCH Open Access

A study of the correlation between STEM

career knowledge, mathematics self-
efficacy, career interests, and career
activities on the likelihood of pursuing a
STEM career among middle school
students
Karen A. Blotnicky1*, Tamara Franz-Odendaal2, Frederick French3 and Phillip Joy4

Abstract
Background: A sample of 1448 students in grades 7 and 9 was drawn from public schools in Atlantic Canada to
explore students’ knowledge of science and mathematics requirements for science, technology, engineering, and
mathematics (STEM) careers. Also explored were their mathematics self-efficacy (MSE), their future career interests,
their preferences for particular career activities, and their likelihood to pursue a STEM career.
Results: Analysis revealed that while older students had more knowledge about mathematics/science requirements
for STEM careers, this knowledge was lacking overall. Also, students with higher MSE were more knowledgeable
about STEM career requirements. Furthermore, students with higher MSE and STEM career knowledge were more
likely to choose a STEM career. Students with greater interest in technical and scientific skills were also more likely
to consider a STEM career than those who preferred career activities that involved practical, productive, and concrete
activities.
Conclusions: The results of this study show that students in middle school have a limited STEM career knowledge
with respect to subject requirements and with respect to what sort of activities these careers involve. Furthermore,
students with low MSE have a declining interest in STEM careers. Our data thus support the need to improve
access to knowledge to facilitate students’ understanding of STEM careers and the nature of STEM work.
Exposure of students to STEM careers can enhance their interest in pursuing careers involving science,
technology, engineering, and mathematics.
Keywords: STEM career, Mathematics self-efficacy, Technical skills, Career knowledge, Career awareness, Career
interests, Career activities, Education, Subject requirements

* Correspondence: [email protected]
1
Department of Business Administration and Tourism and Hospitality
Management, Mount Saint Vincent University, Halifax, NS B3M 2J6, Canada
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made.
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 2 of 15

Background (middle) school age that a student’s beliefs about compe-

Globally, youth vary considerably in their level of science, tency and interests begin to solidify (Simpkins et al.
technology, engineering, and mathematics (STEM) career 2006). It is at this time that student engagement activ-
knowledge, their career interests, and their intentions of ities and career knowledge should be at its highest.
pursuing a STEM career. STEM career knowledge, Social cognitive career theory (Lent 2005) acknowledges
defined as a student’s familiarity with a particular STEM and hypothesizes that career interests, choice, and per-
career, varies considerably based on the school’s STEM sonal goals form a complex human agency process that
career guidance. The level of STEM career knowledge an includes performance, self-efficacy, and outcome expec-
individual has will directly affect one’s intentions of pursu- tations. For example, self-efficacy is positively related to
ing a STEM career in the future (Compeau 2016; Nugent student academic performance and science self-efficacy
et al. 2015; Zhang and Barnett 2015). Without adequate has been shown to impact student selection of science-
knowledge, there is a risk that students will dismiss a related activities, which impacts their ultimate success
STEM-based career path as a potential option for their and helps maintain interests (Britner and Pajares 2006;
future. Consequently, student interest in a particular Parker et al. 2014; Richardson et al. 2012).
STEM career will wane, which will negatively influence Early interest in STEM topics is an excellent predictor
their desire to participate in activities that serve to for later learning and eventual career interests and
increase STEM career knowledge and awareness. Indeed, choice (DeBacker and Nelson 1999). Contextual and
interventions have shown that equipping students with individual variables influence these social cognitive vari-
STEM career knowledge early increases their motivation ables including factors such as parental, teacher, and
to take more science and mathematics courses in high peer cultural expectations (Lent et al. 1994). Nugent et
school (Harackiewicz et al. 2012). al. (2015) found support for the social cognitive career
Students’ career interest and their preferred future car- theory (Lent et al. 1994) as a framework for examining
eer activities will also affect their intention of pursuing a STEM learning and career orientation outcomes by pro-
STEM career. A key predictor of STEM career interest viding a way in which to view the socio-contextual, mo-
at the end of high school is interest at the start of high tivational, and instructional factors that can impact
school (Sadler et al. 2012). However, the positive atti- youth STEM interests.
tudes towards science identified in youth age 10 sharply Although 88% of parents believe they can help guide
declines by age 14 (Murphy and Beggs 2005; Tai et al. their children’s learning, less than 28% actually discuss
2006); the junior high school years are typically ages 12– the value of a STEM education with their children
14 years. An extensive study in 2015, surveying 24,000 (“Let’s Talk Science Canada Annual Report,” 2015).
students, showed that occupational intentions change Recent studies have also indicated that junior high stu-
dramatically between the 9th and 11th grade and that dents have an unclear view about engineering (Compeau
the relationship between STEM intention and motiv- 2016; Karatas et al. 2011) and science (Masnick et al.
ation is highly time-sensitive (Mangu et al. 2015). 2010) yet these are critical years in which to build STEM
Both STEM career knowledge and career interests are interest. The present paper builds on our previous study
also influenced by society at large. These society influen- (Franz-Odendaal et al. 2016) and explores students’
cers include role models that students are exposed to knowledge of STEM career mathematics/science re-
either in person or through the media, the individual stu- quirements and their mathematics self-efficacy (MSE)
dents interact with on a daily basis such as teachers, family and how these shape students’ career interests and pre-
members, and peers, as well as students’ extracurricular ferred career activities. Differences among grade 7 and 9
experiences (Dabney et al. 2012; Harackiewicz et al. 2012; students with respect to career interests and activities,
Nugent et al. 2015; Sahin et al. 2014; Sahin et al. 2015; and the likelihood of pursuing a STEM career will be ex-
Schumacher et al. 2009; Sjaastad 2012; Steinke et al. 2009; amined. While gender differences are important because
Zhang and Barnett 2015). Collectively, these influencing STEM stereotypes are heavily biased towards males,
factors predict the self-efficacy (i.e., one’s belief in one’s these differences are beyond the scope of the current
ability) youth hold about their career options as well as study. This study will examine who, what, and how
their outcome expectancies (Mangu et al. 2015). Self- youth are influenced in STEM career choice.
efficacy is considered a major predictor guiding the selec- This study captured five main areas of interest: student
tion of majors during high school and post-secondary knowledge of mathematics and science requirements
education (Heilbronner 2009; Kelly et al. 2013). that lead to STEM careers, MSE, career interests, career
The grades 7 through 9 years (12–15-year-olds) are activity preferences, and their correlation with the likeli-
the key time period for influencing STEM career interest hood to consider pursuing a STEM career among youth.
and for building this self-efficacy with respect to math- Based on the literature, the following research questions
ematics and science. Thus, it is during the junior high were developed to guide this research.
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 3 of 15

RQ1: What is the correlation between grade level and career activity preferences, career interests, and likeli-
students’ knowledge of high school requirements for hood to choose to pursue a STEM career. These mea-
STEM careers? sures were incorporated into the study based on earlier
RQ2: What is the correlation between MSE and reviews that found that studies’ examining factors influ-
students’ knowledge of high school requirements for encing career choice have been criticized for failing to
STEM careers? account for the complexity of career choices and career
RQ3: What is the correlation between MSE and decision-making (Patton and McMahon 2006) and for
students’ career interests and/or their preference for being too static in their view of career development
particular career activities? (Hirshi 2011).
RQ4: What is the association between student
preferences for career interests and preferred career STEM career knowledge score
activities with grade level? A STEM career knowledge (SCK) score was created to
RQ5: What are the relationships between the following capture students’ knowledge about the requirements for
factors and the likelihood that students will choose a high school mathematics and science in STEM careers.
STEM career: grade level, MSE, student knowledge of Students were presented with 12 STEM careers and
mathematics/science requirements for post-secondary asked to indicate whether they believed that the training
study for STEM careers, career interests and preferred for each of the careers required having taken mathemat-
career activities? ics or science in high school. Students could respond
“yes” if they believed the career required high school
These research questions have not been explored in mathematics or science based on their knowledge of the
the context of Atlantic Canada, thus making this study entrance requirements for Canadian colleges and univer-
relevant to the education system within Canada and sities. They could respond “no” if they believed that the
globally. career did not require high school mathematics or
science, or they could choose “uncertain” if they were
Methods not sure that high school mathematics and/or science
The sample were required for that career. The list included careers
Grade 7 and 9 students in the four Canadian Atlantic students are commonly exposed to (such as veterinarian,
provinces (New Brunswick, Nova Scotia, Prince Edward pharmacist, and oral hygienist) as well as careers that
Island, and Newfoundland) completed an online survey are likely less familiar to them (such as mechanical
during their school hours. This research was approved engineer, geologist, and land surveyor). The list included
by the university research ethics board. Permission to mechanical engineer, computer hardware designer,
collect data in the schools was obtained from school pharmacist, medical technologist, geologist, veterinarian,
board superintendents and parents. Schools were pur- oil industry engineer, physiotherapist, oral hygienist,
posefully chosen from school families in geographic nutritionist, land surveyor, and ophthalmologist. The list
areas across Atlantic Canada. English and French lan- was provided to students in no particular order.
guage schools were included in the study. Data were A score was calculated to capture students’ knowledge
weighted to ensure that the sample was representative based on these responses. “Yes” responses were scored
by grade level, from each of the four Atlantic provinces. as “1,” “uncertain” responses as “0,” and “no” scored as
A total sample size of 1448 students was obtained across “− 1.” The responses were then summed to obtain a
all four provinces in Atlantic Canada: New Brunswick basic SCK score per student. The SCK score was calcu-
(33%), Nova Scotia (38.4%), Prince Edward Island (6.5%), lated only for students who had rated at least one third
and Newfoundland-Labrador (22.1%). The sample was of the careers in the list. The SCK was validated using
split almost evenly between grade 7 (48%) and grade 9 confirmatory factor analysis (CFA) and reliability
(52%). The sample was balanced with respect to gender analysis.
(58% female to 42% male). Students ranged in age from
11 to 20 years with an average age of 13.5 years and a Mathematics Self-Efficacy Scale
median age of 14 years (SD = 1.1). Grade 7 students had In an attempt to offer a more complete perspective on the
an average age of 12.6 years (SD = .6) with a median age process of career decision-making, Hackett and Bertz
of 13 years. Grade 9 students had an average age of 14. (Hackett and Betz 1981) drew on the work of Bandura
5 years (SD = .6) with a median age of 14 years. (1977) to introduce the concept of self-efficacy to the car-
eer development literature noting its potential to help
Measures understand the complexity of career decision-making such
Five different measures were used in this study. These as the underrepresentation of women in traditional male-
included measures of STEM career knowledge, MSE, dominated career fields. Self-efficacy referred to the belief
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 4 of 15

that a person had in their own ability to successfully per- efficacy and outcome expectations, and that self-efficacy
form a particular behavior based on their perception of relates to outcome expectations across Holland’s (1997)
their capability and the likelihood of their achieving suc- broad occupational themes as utilized in the current study
cess in that activity. (Sheu et al. 2010). This alignment was felt to provide a
The second measure used in this analysis was a MSE rationale for the use of student ratings of interests, activ-
scale. Students were asked to describe their experiences in ities, and Holland’s broad occupational themes as a com-
mathematics by rating each of the following statements on prehensive way of gaining insight into the complexity of
a scale ranging from (1) Strongly disagree to (5) Strongly career decision-making of junior high students.
agree: I get good grades in mathematics; I learn quickly in Hollands’ Theory of Career Choice and Development
mathematics; I look forward to my mathematics class; I (Holland 1973) focused on six basic personality types:
feel tense doing mathematics problems; I feel helpless realistic (practical); investigative (analytical, curious); art-
doing mathematics problems. Negatively phrased items istic (expressive, original); social (working/helping
were reverse-coded to maintain consistency in the MSE others); enterprising (goal oriented); and conventional
scale. Students’ responses to these five statements about (ordered). Individuals are not limited to one personality
their experiences in mathematics were then coded into di- type and many exhibit characteristics on more than one
chotomies to create five separate measures. Dichotomies type. Holland (1973) argued that everyone has career de-
were created by assigning a value of “1” to those who were cisions to make at various stages of their lives. As well,
most comfortable with Likert scale ratings of 4 or 5 to the he argued that everyone can serve as both a coach and/
statements and assigning a value of 0 to those who were or a player in those decisions depending on their role,
less confident and comfortable with Likert scale ratings of situation, and knowledge. Reflecting on the life stage, the
1 through 3. These five measures were then summed to environment, and the knowledge one has of their own
get a single MSE scale that would reflect higher measures particular type of preferred approach to life plus know-
for those who were the most confident and comfortable ledge of the interaction among a variety of factors such
with mathematics. The MSE scale was validated using as the cultural, social, academic, and family influences
confirmatory factor analysis (CFA) and reliability analysis. on the decisions that each individual makes about their
The resulting MSE score ranged from 0 to 5. The MSE life career. These are not perfect, single, nor static events
scale was then divided into two subgroups to create an and depend on self and other perceptions of a wide
MSE score for further analysis. The MSE score consisted range of factors. However, at a point in time, they repre-
of those with low MSE (scores of 0 through 3) and high sent what each person conceptualizes as a satisfying car-
MSE (scores of 4 or 5). The goal in using this breakout eer for them. Holland (1973) argued that his theory of
was to identify students who were the most comfortable careers was really intended to help practitioners,
and confident in their mathematics experiences. researchers, and students in education and social science
to address a fuller understanding of vocational choice
Career activities and career interests and to be helpful in professional counseling. Miller
Social Cognitive Career Theory (SCCT) (Lent et al. (1998) stated that Holland’s theory can be used to help
2010) has continued to evolve to include person and individuals explore career choices. More recently,
environmental and socio-demographic variables as Olitsky (2014) used Holland’s theory of career and edu-
well as interest and career choice models. The SCCT cational choice when researching the earnings of STEM
argues that people develop interests (actively likes and majors, indicating that the underlying theory is still rele-
dislikes) largely on the basis of their beliefs about vant. Since career interests and career activities are
their self-efficacy and the outcomes their efforts could highly correlated, they were measured separately.
achieve. Ultimately, people become interested in activ- The third measure used in this analysis was a ranking
ities they believe they can perform well. Therefore, of the preferred career activities using Holland’s Theory
people develop goals to pursue academic and career of Career Choice and Development (1973). Students
activities that are consistent with their interests as were asked to rank six different career activities from (1)
well as with their self-efficacy and outcome expecta- Most favorite to (6) Least favorite. Each of the career
tions (Sheu et al. 2010). Thus, career activities and activities was then analyzed based on the percentage of
career interests are highly correlated. For these students who rated it in their top 2 favorites. The career
reasons, student ratings of self-perceptions of their activities studied in this research included the following:
career interests, and also their career activities, were (1) artistic, unusual, and creative activities; (2) working
included. on practical, productive, and concrete activities; (3) tak-
The SCCT has been found to support self-efficacy and ing responsibility, providing leadership, and convincing
outcome expectations as significant predictors of inter- others; (4) things being organized into routines and hav-
est, that interests partially mediate the relation of self- ing an order; (5) learning by reading, study, analysis, or
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 5 of 15

investigation; (6) helping others and being concerned for Bivariate logistic regression was conducted to explore
the welfare of others. the relative contribution of the following factors on the
The fourth measure used in this analysis was a ranking likelihood that students would choose a STEM career:
of career interests also based on Holland’s Theory of Car- SCK score, MSE score, grade level, career interests, and
eer Choice and Development (Holland 1973). Students career activities. Grade level, career interest, and career
were asked to rank six different career interests from (1) activities were coded as dichotomies for the regression
Most favorite to (6) Least favorite. Each of the career analysis as follows: grade level (grade 9 = 1, grade 7 = 0),
interests was then analyzed based on the percentage of career interests (rated in top 2 favorites = 1, not rated in
students who rated it in their top 2 favorites. These inter- top 2 favorites = 0), career activities (rated in top 2
ests were (1) working with people; (2) creative skills and favorites = 1, not rated in top 2 favorites = 0).
expression; (3) technical and scientific skills; (4) manual Three regressions were created to explore the research
and mechanical skills; (5) leading, persuading, and direct- questions. The first analysis regressed grade level, SCK
ing others; and (6) routines and adhering to standards of score, and MSE score against the likelihood to pursue a
performance. STEM career. Two more regressions were conducted:
one to regress career activities and a second to regress
Likelihood to pursue a STEM career career interests against the likelihood to pursue a STEM
The final measure used in this analysis was the likeli- career as a dependent variable. Measures for career ac-
hood that students would consider choosing a STEM tivity and career interests showed a high level of multi-
career in their future. Students were asked how likely collinearity between the two sets of variables. Separating
they would be to choose a career that is science- these predictors into two different regressions eliminated
related (including science, engineering, health, or problems with multicollinearity.
technology). Likelihood was measured using the fol-
lowing Likert scale: (1) Very unlikely, (2) Somewhat Results
unlikely, (3) Somewhat likely, and (4) Very likely. This We first describe the results for each of the measures
scale was recoded into a dichotomous variable for use used in this study and then answer our research ques-
in bivariate logistic regression: Students who were tions (RQ1–5).
somewhat likely or very likely to choose a STEM car-
eer were coded as “1,” and those who were somewhat Student knowledge of mathematics and science
unlikely or very unlikely to choose a STEM career requirements for STEM careers
were coded as “0.” We assessed student’s knowledge of high school require-
ments for STEM careers, by asking students to indicate
Data analysis whether a career required mathematics and/or science
Data were analyzed using the SPSS software (IBM Corp (Table 1). Mechanical engineer was noted by 71.4% of
2013). Descriptive statistics were used to provide an students as having a high school mathematics/science
overall analysis of the data. Various statistical tests were requirement. Two careers (land surveyor and ophthal-
selected based on the level of data measurement and mologist) were noted by less than half of the students as
data distributions (McDaniel et al. 2014; Hair Jr. et al. requiring high school mathematics or science. Five
2010). t tests were used to explore differences in average careers were classified as requiring mathematics and sci-
ratings between groups. Chi-square was used to analyze ence by 65.6 to 68.2% (veterinarian, geologist, medical
associations between nominal and ordinal variables. technologist, pharmacist, computer hardware engineer).
Analysis of variance (ANOVA) was used to evaluate sig- Four of the careers were listed as requiring mathematics
nificant differences between average ratings and mea- and science by 51.8 to 58.6% of the students (nutrition-
sures involving categorical variables with more than two ist, oral hygienist, physiotherapist, oil industry engineer).
response levels (McDaniel et al. 2014). Logistic regres- What is notable in the students’ responses are that most
sion was used to explore research questions involving students seemed confident of their career classification
interval and ratio-scaled variables (Hair Jr. et al. 2010). in that they answered “yes” or “no” and not the option
Brown-Forsyth exact tests were used with ANOVA to of “uncertain,” indicating that they were confident in
compensate for violations of homogeneity of variance their choice. The percentage of students saying that they
(IBM Corp 2013). Bonferroni post hoc tests were used were uncertain if a career required mathematics or sci-
to detect significant differences between groups for sig- ence for post-secondary study was low and ranged from
nificant ANOVA results (IBM Corp 2013; Hair Jr. et al. 12.5 to 32.6% across all of the careers with half of the
2010). Data were weighted to reflect the population of uncertain responses ranging from 12.5 to 13% of stu-
students by grade level and province across Atlantic dents. Table 1 shows the results of high school mathem-
Canada. atics/science requirements for STEM careers.
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 6 of 15

Table 1 Classification of careers as requiring mathematics and/or science by students in Atlantic Canada
Career Does the career require high school mathematics and/or science for entry into university?
Yes (%) No (%) Uncertain (%) Number
Mechanical engineer 71.4 16.1 12.5 1250
Computer hardware designer 68.2 17.3 14.5 1259
Pharmacist 67.9 18.6 13.5 1257
Medical technologist 67.1 18.8 14.0 1243
Geologist 67.0 20.0 13.0 1249
Veterinarian 65.6 19.3 15.1 1243
Oil industry engineer 58.6 21.3 20.1 1262
Physiotherapist 54.8 24.1 21.1 1254
Oral hygienist 53.1 25.7 22.5 1252
Nutritionist 51.8 25.7 22.5 1247
Land surveyor 48.2 27.7 24.1 1232
Ophthalmologist 46.6 20.9 32.6 1223
Sample size = 1448. Data weighted by grade and province

Students’ responses were summed to obtain an over- Overall, these results suggest that STEM career know-
all SCK score. A factor analysis of the career ratings ledge is limited among middle school students. Results
was used to ensure it was unidimensional, and reliabil- also reveal that students seem to be unaware of their
ity of the score was measured using Cronbach’s alpha. limited knowledge regarding STEM career preparation.
The confirmatory factor analysis was statistically sig-
nificant (KMO = .961, p < .01). Cronbach’s alpha was .95 Students’ mathematics self-efficacy
which meets the criterion for reliability. In order to determine whether MSE was correlated
The SCK score ranged from − 12 to + 12, with an with students STEM career knowledge (RQ2) and/or
average score of 4.6 (SD = 7.6; Fig. 1). The average between MSE and career interests and/or preferred
SCK score was low, indicating a lack of familiarity career activities (RQ3), we first determined the MSE
with the mathematics and/or science requirements of scale for the cohort. The MSE scale ranged from 0
STEM careers. Approximately 8% of students did not (No self-efficacy) to 5 (High self-efficacy). The distri-
correctly classify any of the careers as having a high bution of the Math Self-Efficacy Scale is shown in
school mathematics and/or science requirement. Only Fig. 2.
36.4% of students had high SCK scores having cor- A confirmatory factor analysis of the measures in the
rectly classified 10 to 12 careers. The top quartile of MSE scale indicated that it was unidimensional and reli-
students scored 11 or better while the bottom quar- able. The factor analysis was statistically significant
tile scored 0 or less than 0 out of the 12-point score. (KMO = .698, p < .01). Cronbach’s alpha was .72 which is
A summary of the SCK score is in Fig. 1. acceptable for a scale analysis. These results suggest that

Fig. 1 STEM career knowledge (SCK) score

Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 7 of 15

Fig. 2 Mathematics Self-Efficacy (MSE) Scale

over half of the students had a relatively high MSE and favorites ranged from 21.9% for routines and adhering to
about one third of students had low MSE. standards to working with people at 49.8%. The results
are summarized in Table 3.
Career activities and interests
In order to assess students’ preferred career activities Likelihood of choosing a STEM career
and their career interests, students were asked to select Next, we assessed whether students were interested in
their favorites. Students were presented with a list of six pursuing a STEM career. Nearly 70% percent of students
career activities and asked to indicate which activities surveyed revealed that they were either somewhat likely
were in their top 2 favorites. These measures were or very likely to pursue a STEM career. On a scale of (1)
recoded into dichotomies for further analysis. There was Very unlikely to (4) Very likely, an average rating of 2.9/4
a very even spread of students rating career activities in indicated that students were somewhat likely to pursue a
their top 2 favorites, ranging from 32.1 to 45.9%. Most STEM career. The results appear in Table 4.
of the activities were listed in their top 2 favorites by
about one third of the students. The results revealed that The association of grade level and STEM career
artistic, unusual, and creative activities were most com- knowledge
monly listed in the top 2 favorite career activities. The The first research question (RQ1) explored the correl-
career activity with the lowest rating was helping others ation of grade level and STEM career knowledge.
and being concerned for their welfare. The results are There was a statistically significant difference in the
summarized in Table 2. average SCK score by grade, with grade 9 students
Students were also presented with a list of six career scoring higher than grade 7 students (5.7 vs 3.3, t = −
interests and asked to indicate which career interests 5.69, df = 1209.7, p < .01). While it is good to see that
were in their top 2 favorites. These measures were students appear to acquire more knowledge of STEM
recoded into dichotomies for further analysis. The per- career requirements in middle school grades, it is
centage of students rating career interests as their top 2 concerning that students in grade 9 still had a low
average SCK score since this is the year in which stu-
dents begin to choose subject classes in Atlantic
Table 2 Career activities rated by grade 7 and 9 students in
Atlantic Canada as being in the top 2 favorites
Table 3 Career interests rated by grade 7 and 9 students in
Career activity Percent Number
Atlantic Canada as being in the top 2 favorites
Artistic, unusual, creative activities 45.9 548
Career interest Percent Number
Practical, productive, concrete activities 38.5 457 Working with people 49.8 593
Taking responsibility, providing leadership, 36.3 430 Creative skills and expression 41.2 487
convincing
Technical and scientific skills 39.4 463
Organized into routines and having order 34.7 410
Manual and mechanical skills 32.9 387
Reading, study, analysis, investigation 33.7 390
Leading, persuading, and directing others 32.7 384
Helping others and being concerned for welfare 32.1 374
Routines and adhering to standards 21.9 255
Sample size = sizes range from 1182 to 1199 across career activities. Data
weighted by grade and province Sample size = 1448. Data weighted by grade and province
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 8 of 15

Table 4 Likelihood of choosing a STEM career Correlation between mathematics self-efficacy and
Percent Number students’ preferred career activities and their career
(4) Very likely 32.7 444 interests
(3) Somewhat likely 36.8 499
Our third research question (RQ3) explored whether there
was a correlation between MSE and students’ career inter-
(2) Somewhat unlikely 15.9 215
ests and preferred career activity. There were statistically
(1) Very unlikely 14.6 197 significant differences by students’ preferred career activ-
Total 100.0 1356 ities for the MSE scale. MSE scale totals were sorted into
Sample size = 1356. Mean = 2.9, Median = 3.0, SD = 1.0. Data weighted by two groups to create an MSE score for further analysis.
grade and province
Those with low MSE scale totals (0 through 3) were
assigned an MSE score of 0, and those with high MSE
Canada. This indicates that more work is needed to scale totals (4 and 5) were assigned an MSE score of 1. A
ensure students have the correct information about chi-square analysis revealed that only one career activity
STEM career requirements in time for them to make differed significantly based on students’ MSE scores. Read-
informed decisions about high school course ing, study, analysis, and investigation was listed in the top
selection. 2 favorites for career activities by 36.5% of students who
had high MSE scores (between 4 and 5) when compared
Correlation between mathematics self-efficacy and the to 28.4% of students with low MSE (0 through 3) (χ2 = 7.
STEM career knowledge 979, df = 1, p < .01). The remaining career activities did
The second research question (RQ2) focused on not differ significantly based on students’ MSE. The
whether there is a correlation between students with results are summarized in Table 6.
higher MSE and knowledge of STEM career require- These results show that most of the preferred career
ments. An analysis of variance revealed that students activities had no correlation at all with MSE scores.
with high self-efficacy (MSE scale = 4 and 5) had sig- However, reading, study, analysis, and investigation are
nificantly higher SCK scores than students who did the hallmarks of a mathematics-, science-, or
not score as highly in the MSE scale (BF = 8.7, df = 5, technology-based activity. Therefore, it is reasonable that
p < .01). Students with high MSE had a SCK score of students who are confident and comfortable with math-
6.6 out of 12, while students with lower MSE scores ematics would also enjoy reading, study, analysis, and
had average SCK scores ranging from 2.8 to 4.8. The investigation.
results are shown in Table 5. In order to explore whether there is a correlation be-
These results for RQ2 show that students who re- tween MSE and student’s career interests, a chi-square
port more confidence and comfort in mathematics analysis was conducted. The chi-square analysis revealed
tend to be more knowledgeable about mathematics/ that only one career interest differed significantly based
science requirements for STEM careers. This is a cor- on students’ MSE score. This career interest was technical
relation only and cannot be interpreted as a causal and scientific skills. This career interest was listed in the
relationship since survey data cannot be used to top 2 favorites for career activities by 43.8% of students
measure causality. with high MSE score (between 4 and 5) compared to 36.
0% of students with low MSE score (0 through 3) (χ2 = 6.
558, df = 1, p = .01). The remaining career interests did
Table 5 The correlation between mathematics self-efficacy not differ significantly based on students’ MSE scores. As
(MSE) on STEM career knowledge (SCK) scores with the results for career activities, these results show
Mathematics STEM career knowledge score
that most of the career interests were not significantly
Self-Efficacy correlated with MSE and all of the career interests were
Average Std. deviation Number
Scale rated in the top 2 favorites by less than half of the
5 6.6 6.6 365 students. It is reasonable that students who are confident
4 4.8 7.3 304 and comfortable with mathematics would also be
3 3.6 7.9 194 interested in careers involving technical and scientific
skills. The results are summarized in Table 7.
2 3.0 8.3 197
1 2.8 8.5 101
The correlation between grade level and students’ career
0 3.3 7.5 105 interests and preferred career activities
Sample size = 1266. Data weighted by grade and province. The SCK score for The fourth research question (RQ4) addressed whether
students with high MSE is significantly higher than the average ratings for any
of the lower self-efficacy scores (Brown-Forsyth exact test = 8.7, df = 5/824,
grade level was associated with student preferences for
p < .01). Minimum = − 12, maximum = + 12 career interests and activities. There were statistically
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 9 of 15

Table 6 Analysis of students’ favored career activities by mathematics self-efficacy (MSE)

Mathematics self-efficacy score
Career activity listed in top 2 favorites 0 to 3 4 to 5
Low self-efficacy High self-efficacy
Helping others and being concerned for welfare 146 (31.9%) 200 (57.8%)
Artistic, unusual, creative activities 230 (49.6%) 286 (45.0%)
Practical, productive, concrete activities 178 (38.7%) 247 (38.7%)
Taking responsibility, providing leadership, convincing 166 (35.9%) 235 (37.1%)
Organized into routines and having order 145 (31.2%) 233 (36.8%)
Reading, study, analysis, investigation^ 131 (28.4%) 232 (36.5%)
Differences are statistically significant at the .01 level or better (^). Data weighted by grade and province. Sample size varies from 1088 to 1100 for each activity.
Results show the percent within each MSE group (low or high MSE) that rated the career activity shown in their top 2 favorites. Percentages do not sum to 100%

significant differences by grade regarding some of the The correlations between students’ STEM career
career interests, thereby satisfying the first part of the knowledge, mathematics self-efficacy, and grade level on
fourth research question. More grade 7 than grade 9 stu- their likelihood to choose a STEM career
dents listed manual and mechanical skills in their top 2 The fifth research question (RQ5) focused on how sev-
favorites (36.5 vs 29.4%, χ2 = 6.84, df = 1, p < .01), as well eral aspects might relate to students’ likelihood of
as creative skills and expression (45.4 vs 37.0%, χ2 = 8.73, choosing a STEM career. These areas included grade
df = 1, p < .01). More grade 9 than grade 7 students level, MSE, knowledge of STEM careers, and preferences
ranked “working with people” in their top 2 favorites for various career interests and activities.
(52.8 vs 46.8%, χ2 = 4.21, df = 1, p < .05). These results First, a logistic regression was conducted to determine
are summarized in Table 8. whether or not grade level, STEM knowledge, and MSE
More grade 7 than grade 9 students listed practical, score were associated with students’ likelihood to pursue
productive, and concrete activities in their top 2 favorites a STEM career. The hypothesized regression model was
(42.1 vs 34.7%; χ2 = 6.9, df = 1, p < .01). More grade 9 likelihood of choosing a STEM career (ODDS) = f(grade
than grade 7 students rated helping others and being level, STEM knowledge score, MSE score). A test of the
concerned for their welfare in their top 2 favorite career full regression model against an intercept-only model
activities (34.9 vs 28.6%; χ2 = 5.4, df = 1, p < .05) as well was statistically significant (χ2 = 76.85, df = 3, p < .01).
as having things organized into routines and having The regression was strong with a McFadden’s R2 = .85.
order (39.4 vs 29.4%; χ2 = 13.2, df = 1, p < .01). There The regression analysis correctly classified 70.6% of all
were no statistically significant differences by grade level cases and 95.3% of those who were likely to choose a
for the other career activities studied. STEM career. The regression revealed that students with
This trend is similar to that emerging in the analysis stronger SCK scores were marginally more likely to pur-
of career interests. In general, students in the higher sue a STEM career than were students with weaker SCK
grade focused more on activities involving helping others scores (odds ratio = 1.04, probability = .51). However,
and being less attracted to careers that involved practical students with high MSE scores were 1.3 times more
applications or routines. The results are summarized in likely to pursue a STEM career than were those who
Table 9. had lower MSE scores (probability = .56). Grade level

Table 7 Analysis of students’ favored career interests by mathematics self-efficacy (MSE)

Mathematics self-efficacy score
Career interest listed in top 2 favorites 0 to 3 4 to 5
Low self-efficacy High self-efficacy
Working with people 215 (46.8%) 328 (51.1%)
Technical and scientific skills^ 159 (36.0%) 280 (43.8%)
Creative skills and expression 195 (43.5%) 267 (41.5%)
Manual and mechanical skills 150 (33.6%) 205 (32.2%)
Leading, persuading, and directing others 139 (31.3%) 205 (32.1%)
Routines and adhering to standards 104 (23.8%) 130 (20.4%)
Differences statistically significant at the .01 level (^). Data weighted by grade and province. Sample sizes varied from 1073 to 1101 by career interest. Results
show the percent within each MSE group (low or high MSE) that rated the career interest shown in their top 2 favorites. Percentages do not sum to 100%
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 10 of 15

Table 8 Analysis of students’ favored career interests by grade

Career interest listed in top 2 favorites Grade 7 Grade 9
% Number % Number
Working with people^ 46.8 277 52.8 316
Technical and scientific skills 40.3 236 38.6 227
Creative skills and expression^ 45.4 268 37.0 220
Leading, persuading and directing others 31.7 185 33.7 199
Manual and mechanical skills^ 36.5 214 29.4 173
Routines and adhering to standards 19.8 115 24.1 140
Sample size ranges from 255 to 593 across career interests. Differences are statistically significant at the .05 level or better (^). Data weighted by grade and province

was not a statistically significant predictor of the likeli- and expression; working with people; leading, persuad-
hood of pursuing a STEM career. ing, and directing others; routines and adhering to stan-
These results showed that students’ knowledge of dards). A test of the full regression model against an
STEM careers and their self-efficacy in mathematics intercept-only model was statistically significant (χ2 =
were statistically significant factors in the likelihood that 119.94, df = 6, p < .01). The regression was reasonably
they would pursue a STEM career, while STEM career strong with a McFadden’s R2 = .73. The regression
knowledge was a modest contributor. Also, students in analysis correctly classified 72% of all cases and 96.6% of
grade 9 were not more likely to pursue a STEM career those who were likely to choose a STEM career.
than were students in Grade 7. However, research has The regression revealed that students who rated
shown that occupational intentions change dramatically technical and scientific skills in their top 2 favorite
between 9th and 11th grades and the relationship be- career interests were 5.4 times more likely to pursue
tween STEM intention and motivation is very time- a STEM career (probability = .84). Students who rated
sensitive (Mangu et al. 2015, p.55). The results are sum- working with people in their top 2 favorites were 1.5
marized in Table 10. times more likely to pursue a STEM career (probabil-
These regression results reveal that individual student ity = .61). Students who rated creative and expressive
characteristics, MSE, and SCK are better predictors of skills in their top 2 favorite career interests were less
the likelihood to pursue STEM careers than student likely to pursue a STEM career than those who rated
grade level. Individual strengths and weaknesses, as well creative and expressive skills highly. Their odds of
as students’ knowledge and competency, are better indi- pursuing a STEM career were only .70 of those who
cators of future career paths than grade level. did not rate creativity and expressiveness among their
favorite career interests. Their probability of pursuing
The correlation between students’ career interests and a STEM career was .41. The remaining career inter-
their likelihood to pursue a STEM career ests were not statistically significant predictors of the
A second logistic regression was conducted to explore likelihood of pursuing a STEM career (manual or
whether or not students’ preferred career interests was mechanical skills; leading, persuading, or directing
correlated with their likelihood to pursue a STEM car- others; routines and adhering to standards). These re-
eer. Six career interests were explored in the analysis. sults provide evidence for the fifth research question
The hypothesized regression model was likelihood of in that three out of the six career interests measured
choosing a STEM career (ODDS) = f(manual and mech- did have a statistically significant correlation with the
anical skills; technical and scientific skills; creative skills likelihood that a student would consider pursuing a

Table 9 Analysis of students’ favored career activities by grade

Career activity listed in top 2 favorites Grade 7 Grade 9
% Number % Number
Artistic, unusual, creative activities 47.2 283 44.5 265
Organized into routines and having order^ 29.4 173 39.4 236
Taking responsibility, providing leadership, convincing 36.4 217 36.0 213
Helping others and being concerned for welfare^ 28.6 170 34.9 204
Practical, productive, concrete activities^ 42.1 249 34.7 207
Reading, study, analysis, investigation 33.6 199 32.0 191
Sample size ranges from 374 to 548 across career activities. Differences statistically significant at the .05 level or better (^). Data weighted by grade and province
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 11 of 15

Table 10 Bivariate logistic regression of the correlations between grade level, STEM knowledge score and mathematics self-efficacy
by students’ likelihood to pursue a STEM career
B Wald χ2 Sig. Exp(B) Probability
(ODDS)
STEM career knowledge score .042 25.053 .000 1.043 .51
Mathematics self-efficacy score .236 34.146 .000 1.266 .56
Grade 9 .232 3.073 .080 1.261 .56
Constant − .145 .964 .964 .857 –
Sample size = 1215. Data weighted by grade and province

STEM career. The results are summarized in analysis correctly classified 72% of all cases, and 100% of
Table 11. those who were likely to choose a STEM career.
These results indicate that student preference for tech- The regression revealed that students who preferred
nical and scientific skills and careers involving working career activities involving reading, study, analysis, and
with people enhance the likelihood of pursuing a STEM investigation were 1.8 times more likely to pursue a
career, while students who prefer careers involving cre- STEM career (probability = .65) than those who did not
ative skills and expression are less likely to do so. While prefer such activities. Students’ rating career activities
a focus group could better explore the students’ prefer- involving routines and having an order were 1.5 times
ences for creativity and creative careers, this level of de- more likely (probability = .60) to pursue a STEM career
tail is not possible in large sample survey-based research than those who did not prefer such activities, while stu-
and is outside of the scope of this study. Other career dents with preferences for practical, productive, and
interests that focus on mechanical, manual, or routine concrete career activities were 1.5 times more likely to
activities, or those involving leadership, do not predict pursue a STEM career (probability = .60) compared to
the likelihood of students pursuing a STEM career and those who did not prefer such activities. The remaining
are not significantly correlated with STEM career career activities were not statistically significant predic-
choice. tors of the likelihood to pursue a STEM career (artistic,
A third logistic regression analysis was conducted to unusual, and creative activities; taking responsibility,
determine whether students’ career activity preferences providing leadership, and convincing others; helping
were correlated whether or not they were likely to pur- others and being concerned for their welfare). These re-
sue a STEM career. Six career activities were explored in sults revealed three out of the six career activities mea-
the analysis. The hypothesized regression model was sured did have a statistically significant correlation with
likelihood of choosing a STEM career (ODDS) = f(prac- the likelihood that a student would consider pursuing a
tical, productive, concrete activities; reading, study, ana- STEM career. The results are summarized in Table 12.
lysis, and investigation; artistic, unusual, and creative These results stand in contrast to those for students’
activities; taking responsibility, providing leadership, and career interests and the likelihood of pursuing a STEM
convincing; and helping others and being concerned for career. Unlike the career interest analysis, students seek-
their welfare). A test of the full regression model against ing routine career activities are more likely to pursue a
an intercept-only model was statistically significant (χ2 = STEM career. Also, students who ranked career interests
32.883, df = 6, p < .01). The regression was reasonably involving helping others were more likely to pursue
strong with a McFadden’s R2 = .78. The regression STEM careers, but this analysis showed that student

Table 11 Bivariate logistic regression of the correlation between career interests and students being very likely to pursue a STEM
career
Career interests listed in top 2 favorites B Wald χ2 Sig. Exp(B) Probability
(df = 1) (ODDS)
Technical and scientific skills 1.687 83.115 .000 5.402 .84
Working with people .431 7.657 .006 1.538 .61
Creative skills and expression − .356 5.309 .021 .700 .41
Routines and adhering to standards .247 1.905 .167 1.280 .56
Manual and mechanical skills − .16 .825 .364 .853 .46
Leading, persuading, and directing others .073 .200 .655 1.076 .52
Constant .335 2.621 .105 1.397 –
Sample size = 1126. Data weighted by grade and province
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 12 of 15

Table 12 Bivariate logistic regression of the influence of career activity preferences on students being very likely to pursue a STEM
career
Career activity listed in top 2 favorites B Wald χ2 Sig. Exp(B) Probability
(df = 1) (ODDS)
Reading, study, analysis, investigation 0.606 14.49 0.000 1.834 0.65
Organized into routines and having order 0.412 7.301 0.007 1.51 0.60
Practical, productive, concrete activities 0.389 6.518 0.011 1.475 0.60
Helping others and being concerned for welfare 0.179 1.441 0.23 1.197 0.54
Taking responsibility, providing leadership, convincing 0.143 0.915 0.339 1.154 0.54
Artistic, unusual, creative activities − 0.091 0.399 0.528 0.913 0.48
Constant 0.42 5.466 0.019 1.522 –
Sample size = 1129. Data weighted by grade and province

preference for career activities involving helping others (2009), and Mangu et al. (2015) have noted how young
and being concerned for their welfare was not a statisti- women have a lower self-efficacy in STEM during high
cally significant indicator of their likelihood to pursue a school years. Previous research has also shown that
STEM career. Further, these results differ somewhat in interest in STEM and motivation to pursue STEM activ-
terms of students’ preferences for practical activities. ities tends to wane over time for all high school stu-
While career interests involving manual or mechanical dents. The results of the current study agree with earlier
skills were not statistical indicators of the likelihood of findings that lower levels of MSE exist; we found
pursuing a STEM career, career activities involving prac- approximately 34% of participants had low MSE scale
tical, productive, and concrete activities were statistically totals. These findings raise concerns about the combined
significant. The career activity involving reading, study, effects of students’ low MSE and their declining interest
analysis, and investigation was also statistically linked to in STEM from early through to later grades and on the
students’ likelihood to pursue a STEM career, which numbers of graduating high school students who will be
seems reasonable given that such activities are at the inclined to choose a STEM career.
heart of many STEM careers.
Career activities and interests
Discussion Results of the current study demonstrated that stu-
Youth vary widely in their career knowledge, interest, dents in grades 7 and 9 had a broad range of favorite
and intentions. Factors investigated in the present study career activities with the majority (approximately
examined STEM career knowledge, MSE, career activ- 46%) stating that their strongest preference was for
ities, career interests, and the likelihood of students to artistic and creative types of activities. Also, all of the
pursue a STEM career. possible activities were selected by at least one third
of the group. Interestingly, approximately 50% of par-
Knowledge and self-efficacy ticipants selected their career interest as being “work-
Results of the present study align with recent findings by ing with people,” but relative to career activities, only
Compeau et al. (2016), Nugent et al. (2015), and Zhang one third of participants selected “helping people.”
and Barnett (2015) show that self-efficacy along with However, this is not surprising given that one can
knowledge of STEM careers are significant factors in have an interest but may not want to have a career
whether or not adolescents pursue STEM careers. Find- working in that activity. For example, one may be in-
ings also indicated that career knowledge is limited terested in art, but have no interest, or lack sufficient
among middle school students and students seem to be talent, to pursue a career in the field (Holland 1973).
unaware of their limited knowledge regarding STEM Also, middle school students may not be able to dis-
preparation. While approximately 70% of students re- criminate between the nuances between career activ-
ported that mathematics was an important requirement ities and interests in the way that older students and
for a career in mechanical engineering, computer hard- young adults would. Although a focus group study
ware design, and pharmacy, 50% or less were aware that may be able to further elucidate this issue, this is be-
it was also important in careers for ophthalmology, land yond the scope of the current study.
surveyor, nutrition, and oral hygienist. Working with others and participating in creative
The issue of self-efficacy takes on particular signifi- types of activities are important findings that relate to
cance as students progress through high school. Previous current issues in education in Canada. A recent study
research by Murphy and Beggs (2005), Heilbronner (Ayar and Yalvac 2016) found that many STEM careers
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 13 of 15

are team-based, creative, and require technical, scientific, to solidify. Further research may help to uncover reasons
and problem solving skills. However, in Canada, many for not seeing such a correlation in this analysis.
post-secondary programs continue to focus more on
memorizing and replicating science content knowledge. Influence of grade level on STEM career knowledge,
Further study of this possible implication would be interest, and activities
worthwhile. Results indicate that there were significant differences
between grade 7 and grade 9 students in the present
Likelihood of choosing a STEM career study relative to STEM career knowledge. Overall, stu-
While approximately 70% of participants stated they dents in grade 9 were more knowledgeable than grade
were likely to choose a STEM career, 30% were less 7 students about STEM careers. The differences formed
likely to do so. Not all students have the financial means an interesting and consistent pattern that more grade 7
to pursue a career interest. In addition, career interest students expressed interest in manual and mechanical
and motivation are highly time sensitive (Mangu et al. skills than grade 9 students who tended to have more
2015, p.55). Considering findings from studies such as interest in working with people. Further, more grade 7
Ayar and Yalvac (2016), as well as what we know about students expressed interests in practical and concrete
the decline in interest in STEM careers as students types of activities while more grade 9 students
mature, these results suggest that there is room for in- expressed interests in helping people and being con-
creasing awareness, STEM career interest, and providing cerned for their welfare. Reasons for this shift are not
for better knowledge acquisition in the area of STEM clearly understood. As noted by Lent (2005), career
careers. As well, our data suggests that alternative ways interest, choice, and personal goals form a complex
of teaching and evaluating STEM courses should be chain involving performance, self-efficacy, and outcome
considered. Perhaps a greater emphasis on authentic expectations. As well, socio-cultural factors also need
means of teaching and evaluating STEM content that in- to be considered along with opportunity for exposure
volves collaboration, problem solving, and application of (Fouad and Smith 1996; Kuncel et al. 2005; Lent et al.
STEM knowledge might serve to engage learners in 1994). As with the relationship among self-efficacy,
more meaningful ways, thereby enabling continued knowledge, interests, and activities, in-depth research
motivation and interest in STEM careers as students involving student interviews may result in greater un-
progress through secondary and post-secondary educa- derstanding of the reasons for these shifts and their im-
tion. Does a higher MSE lead students to consider pur- pact on later careers.
suing STEM careers and lead them to becoming more
informed about the career requirements or do students Factors influencing positive statements involving the
who have higher knowledge of STEM careers become likelihood of choosing a STEM career
more competent in mathematics? Are these factors sim- Regression analyses revealed that participants with
ply correlational and reflect students who are high in stronger STEM career knowledge were slightly more
both measures or low in both measures? While these likely to pursue a STEM career and that students with
questions cannot be answered in this research, it is inter- higher MSE scores were also slightly more likely to
esting to note that MSE may play a role, or be a leading choose a STEM career. Also, grade level was not a dif-
indicator, for STEM career knowledge. ferentiating factor, which was anticipated given the small
distance between the experiences of grade 7 versus grade
Influence of mathematics self-efficacy on career 9 students. As noted in many previous studies (Lent et
knowledge, interest and activities al. 1994; Kuncel et al. 2005), knowledge of STEM careers
Our results indicate that while there is a relationship and self-efficacy in mathematics are statistically signifi-
between career knowledge and MSE, we did not find a cant factors in the likelihood that participants will pur-
relationship among career interests or activities with sue STEM careers.
MSE. The assumption that having a positive sense of Interest in technical and scientific skills is a strong
mathematics skill would correlate with STEM career predictor of the likelihood of pursuing a STEM career
interests and activities was not supported. Follow-up re- with those who indicated a preference for technical and
search involving interviews with participants about their scientific skills being 5.4 times more likely to indicate
understanding of career interests, activities, and MSE, the likelihood of choosing a STEM career compared to
would provide more an in-depth understanding. Based those who rated working with people as their stronger
on findings by Simpkins et al. (2006) it was expected interest. Indeed, preferences for practical, productive,
that there would be a relationship among interests, and concrete activities also indicated a stronger likeli-
activities and MSE as their findings indicated that in hood of pursuing STEM careers than those who do not
junior high beliefs about competency and interests begin prefer such activities. Implications of these findings
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 14 of 15

point to improving methods for providing information Publisher’s Note

on the skills and nature of the work in STEM careers Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
particularly in fields such as engineering and technology
(which have an important focus on team work, problem Author details
1
solving, and creativity) as well as on technical and scien- Department of Business Administration and Tourism and Hospitality
Management, Mount Saint Vincent University, Halifax, NS B3M 2J6, Canada.
tific skills. 2
Department of Biology, Mount Saint Vincent University, Halifax, NS B3M 2J6,
Canada. 3Department of Education, Mount Saint Vincent University, Halifax,
NS B3M 2J6, Canada. 4School of Health and Human Performance, Dalhousie
Conclusions University, Halifax, NS B3H 4R2, Canada.
Overall, results of the present study show that career
Received: 28 July 2017 Accepted: 24 April 2018
knowledge is limited among middle school students and
that they have a declining interest in STEM and have
low MSE scores. Students are interested in careers that
involve a wide variety of activities but do not appear to References
Ayar, MC, & Yalvac, B. (2016). Lessons learned: authenticity, interdisciplinarity, and
relate these activities to STEM careers. Our results point mentoring for STEM learning environments. International Journal of Education
to the importance of finding and expanding on ways to in Mathematics, Science and Technology, 4(1), 30–43.
increase authentic learning opportunities in secondary Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavior change.
Psychological Review, 84, 191–215.
school in Atlantic Canada such that students are better Britner, SL, & Pajares, F. (2006). Sources of science self-efficacy beliefs of middle
able to participate in collaboration, problem solving, and school students. Journal of Research in Science Teaching, 43(5), 485–499.
the application of scientific knowledge in their classes. Compeau, S. (2016). The calling of an engineer: High school students’
perceptions of engineering. Retrieved from http://qspace.library.queensu.ca/
Such learning opportunities would ensure that students jspui/handle/1974/13924
have access to more information on the actual nature of Dabney, KP, Tai, RH, Almarode, JT, Miller-Friedmann, JL, Sonnert, G, Sadler, PM,
work in the STEM field and what is required to pursue Hazari, Z. (2012). Out-of-school time science activities and their association
with career interest in STEM. International Journal of Science Education, Part B,
these careers. This strategy would also serve as a motiv- 2(1), 63–79.
ator to those who are not aware that STEM careers in- DeBacker, TK, & Nelson, RM. (1999). Variations on an expectancy-value model of
volve people skills, creativity, and problem solving. motivation in science. Contemporary Educational Psychology, 24(2), 71–94.
Fouad, NA, & Smith, PL. (1996). A test of a social cognitive model for middle school
students: math and science. Journal of Counseling Psychology, 43(3), 338.
Abbreviations Franz-Odendaal, TA, Blotnicky, K, French, F, Joy, P. (2016). Experiences and
CFA: Confirmatory factor analysis; MSE: Mathematics self-efficacy; SCK: STEM perceptions of STEM subjects, careers, and engagement in STEM activities
career knowledge score; STEM: Science, technology, engineering, and among middle school students in the maritime provinces. Canadian Journal
mathematics of Science, Mathematics and Technology Education, 16(2), 153–168.
Hackett, G, & Betz, NE. (1981). A self-efficacy approach to the career development
Acknowledgements of women. Journal of Vocational Behavior, 18, 326–336.
Funding was provided to author by a grant from the Natural Sciences and Hair Jr, JF, Blacks, WC, Barry, BJ, Anderson, RE (2010). Multivariate data analysis, 7/e.
Engineering Research Council of Canada. We are grateful to the individuals Upper Saddle River, NJ: Prentice-Hall, Inc.
who helped our team obtain parental consent forms and parents for providing Harackiewicz, JM, Rozek, CS, Hulleman, CS, Hyde, JS. (2012). Helping parents to
consent for the children to participate in this study. motivate adolescents in mathematics and science an experimental test of a
utility-value intervention. Psychological Science, 956797611435530.
Heilbronner, NN (2009). Pathways in STEM: factors affecting the retention and
Availability of data and materials attrition of talented men and women from the STEM pipeline. ERIC Retrieved
University ethics approval does not include release of the raw data. Data was from http://eric.ed.gov/?id=ED513162.
collected from minor-aged children under the strict condition of anonymity
Hirshi, A. (2011). Career-choice readiness in adolescence: developmental
and cannot be shared. Please contact the corresponding author for more
trajectories and individual differences. Journal of Vocational Behavior, 79,
information.
340–348.
Holland, JL (1973). Making vocational choices: a theory of careers. Englewood Cliffs:
Authors’ contributions Prentice-Hall, Inc.
The second author conceptualized the broad goals of the study, and all four IBM Corp. Released 2013. IBM SPSS Statistics for Windows, version 22.0. Armonk,
authors were part of a research team that designed all the study components NY: IBM Corp.
and executed the research. The first author was primarily responsible for Karatas, FO, Micklos, A, Bodner, GM. (2011). Sixth-grade students’ views of the
designing the methodology and quantitative data analysis, while the fourth nature of engineering and images of engineers. Journal of Science Education
author assisted with the data analysis. The second and third authors were and Technology, 20(2), 123–135.
primarily responsible for conducting the literature review, developing the Kelly, K, Dampier, DA, Carr, K. (2013). Willing, able, and unwanted: high school
theoretical underpinnings of the research, and interpreting the results of the girl’s potential selves in computing. Journal of Women and Minorities in
analysis from a theoretical perspective. All authors read and approved the final Science and Engineering, 19(1) Retrieved from http://www.dl.begellhouse.
manuscript. com/journals/00551c876cc2f027,2d4a5d286768c212,26d9b12a206d2bc6.html.
Kuncel, NR, Credé, M, Thomas, LL. (2005). The validity of self-reported grade point
averages, class ranks, and test scores: a meta-analysis and review of the
Ethics approval and consent to participate
literature. Review of Educational Research, 75(1), 63–82.
This approval was granted by the Research Ethics Board of Mount Saint
Lent, R. W. (2005). A Social Cognitive View of Career Development and
Vincent University in Halifax, NS Canada.
Counseling. Retrieved from http://psycnet.apa.org/psycinfo/2004-21312-005
Lent, RW, Brown, SD, Hackett, G. (1994). Toward a unifying social cognitive theory
Competing interests of career and academic interest, choice, and performance. Journal of
The authors declare that they have no competing interests. Vocational Behavior, 45(1), 79–122.
Blotnicky et al. International Journal of STEM Education (2018) 5:22 Page 15 of 15

Lent, RW, Paixao, MA, da Silva, JT, Leitao, LM. (2010). Predicting occupational
interest and choice aspirations in Portuguese high school students: a test of
social cognitive career theory. Journal of Vocational Behavior, 76, 244–251.
Let's Talk Science Annual Report. Spotlight on Science Learning - Exploring
Parental Influence: Shaping Teen Decisions Regarding Science Education.
(2015). Retrieved from http://letstalkscience.ca/Portals/0/Documents/RPS/
Spotlight/LTS-Exploring-parental-influence-EN.pdf
Mangu, DM, Lee, AR, Middleton, JA, Nelson, JK (2015). Motivational factors
predicting STEM and engineering career intentions for high school students.
In Frontiers in education conference (FIE), 2015. 32614 2015. IEEE, (pp. 1–8). IEEE
Retrieved from http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7344065.
Masnick, AM, Valenti, SS, Cox, BD, Osman, CJ. (2010). A multidimensional scaling
analysis of students’ attitudes about science careers. International Journal of
Science Education, 32(5), 653–667.
McDaniel, C, Gates, R, Sivaramakrishnan, S, Main, K (2014). Marketing research
essentials, 2/ce (). Mississauga, ON: Wiley & Sons Canada.
Miller, MJ. (1998). Broading the use of Holland’s hexagon with specific
implications for career counselors. Journal of Employment Counseling, 35, 2–5.
Murphy, C, & Beggs, J. (2005). Coteaching as an approach to enhance science
learning and teaching in primary schools. Counterpoints, 294, 207–231.
Nugent, G, Barker, B, Welch, G, Grandgenett, N, Wu, C, Nelson, C. (2015). A model
of factors contributing to STEM learning and career orientation. International
Journal of Science Education, 37(7), 1067–1088.
Olitsky, NH. (2014). How do academic achievement and gender affect the
earnings of STEM majors? A propensity score matching approach. Research in
Higher Education, 55, 245–271.
Parker, PD, Marsh, HW, Ciarrochi, J, Marshall, S, Abduljabbar, AS. (2014).
Juxtaposing math self efficacy and self-concept as predictors of long-term
achievement outcomes. Educational Psychology, 34(1), 29–48.
Patton, W, & McMahon, M (2006). Career development and systems theory, (2nd
ed., ). Pacific Grove, CA: Brooks/Cole.
Richardson, M, Abraham, C, Bond, R. (2012). Psychological correlates of university
students’ academic performance: a systematic review and meta-analysis.
Psychological Bulletin, 138(2), 353.
Sadler, PM, 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.
Sahin, A, Ayar, MC, Adiguzel, T. (2014). STEM related after-school program
activities and associated outcomes on student learning. Educational Sciences:
Theory and Practice, 14(1), 309–322.
Sahin, A, Gulacar, O, Stuessy, C. (2015). High school students’ perceptions of the
effects of international science Olympiad on their STEM career aspirations
and twenty-first century skill development. Research in Science Education,
45(6), 785–805.
Schumacher, MM, Stansbury, KN, Johnson, MN, Floyd, SR, Reid, CE, Noland, MP,
Leukefeld, CG. (2009). The young women in science program: a five-year
follow-up of an intervention to change science attitudes, academic behavior,
and career aspirations. Journal of Women and Minorities in Science and
Engineering, 15(4) Retrieved from http://www.dl.begellhouse.com/journals/
00551c876cc2f027,293840bd7415e516,4ac4cd3d67f80a79.html.
Sheu, HB, Lent, RW, Brown, SD, Miller, MJ, Hennessy, KD, Duggy, RD. (2010). Testing
the choice model of social cognitive career theory across Holland themes: a
meta-analytic path analysis. Journal of Vocational Behavior, 76, 252–264.
Simpkins, SD, Davis-Kean, PE, Eccles, JS. (2006). Math and science motivation: a
longitudinal examination of the links between choices and beliefs.
Developmental Psychology, 42(1), 70.
Sjaastad, J. (2012). Sources of inspiration: the role of significant persons in young
people’s choice of science in higher education. International Journal of
Science Education, 34(10), 1615–1636.
Steinke, J, Lapinski, M, Long, M, Van Der Maas, C, Ryan, L, Applegate, B. (2009).
Seeing oneself as a Scientist: media influences and adolescent girl’s science
career-possible selves. Journal of Women and Minorities in Science and
Engineering, 15(4) Retrieved from http://www.dl.begellhouse.com/journals/
00551c876cc2f027,293840bd7415e516,0b5caaa338bc00a0.html.
Tai, R. H., Liu, C. Q., Maltese, A. V., & Fan, X. (2006). Planning early for careers in
science. Life Sciences, 1, 0–2.
Zhang, L, & Barnett, M. (2015). How high school students envision their STEM
career pathways. Cultural Studies of Science Education, 10(3), 637–656.