Home Articles Reader Opinion Editorial Book Reviews Discussion Writers Guide About TCRecord
transparent 13
Topics
Discussion
Announcements
 

Individual and Institutional Factors of Applied STEM Coursetaking in High School


by Cameron Sublett & Michael A. Gottfried - 2017

Background/Context: One approach to address the shortage of STEM-proficient high school graduates has been the development of applied STEM coursework, which seeks to increase STEM interest and retention by illustrating the interconnectedness and accessibility of STEM concepts. Importantly, however, no research has yet examined which student and institutional factors are associated with applied STEM coursetaking.

Purpose/Objective/Research Question/Focus of the Study: The current study poses three questions: (1) What student factors are related to applied STEM coursetaking in high school? (2) What school factors are related to applied STEM coursetaking in high school? (3) How is the influence of these factors different based on the timing of course taken?

Population/Participants/Subjects: This study uses data from the Education Longitudinal Study of 2002 (ELS:2002). ELS:2002 provides complete transcript data for ~ 90% of the students in the sample. Importantly, parent and teacher questionnaires were also gathered in the base year (2002) of the study. ELS:2002, therefore, represents a rich, nationally representative multilevel dataset with complete student coursetaking records along with parent, teacher, and school administrator data.

Research Design: This study relies on secondary data analyses and a series of logistic and multinomial logistic regression models. There were four key outcome variables, including: (1) a binary indicator of enrollment in at least one applied STEM course during high school, (2) a binary measure of whether a student had enrolled in at least one SRE-specific applied STEM course, (3) a binary measure of whether a student had enrolled in at least one IT-specific applied STEM course. To construct the fourth and final outcome variable, we divided applied STEM coursetaking into three categories: early, late, and never.

Findings: The findings for the first research question suggested that women were significantly less likely to enroll in applied STEM courses. The findings for the second research question suggested that institutional factors were only weakly related to applied STEM enrollment. The findings for the third research question did not suggest major trends in when students enrolled in applied STEM. Rather, it appears that students enroll in applied STEM courses throughout high school.

Conclusions/Recommendations: This study illustrated that rather than increasing the number of traditionally underrepresented student groups in STEM, applied STEM courses may be contributing to the much-discussed “gender gap” in STEM education. Policymakers and school leaders must examine the mechanisms behind this stratification; future research should explore, perhaps through qualitative inquiry, why women and students with disabilities choose to not take these courses.



High school coursetaking continues to serve as a key determinant of academic achievement (Adelman, 1999). This is particularly true for science, technology, engineering, and mathematics (STEM) courses where progression through the STEM pipeline—from high school, to college, to career (Tyson, Lee, Borman, & Hanson, 2007)—is dependent, in part, on hewing closely to strict coursetaking sequences (Grossman & Stodolsky, 1995). Students who complete advanced mathematics and science coursework are more likely to enroll in postsecondary institutions (Long, Conger, & Iatarola, 2012; Schneider, 2003), pursue a STEM area of study (Federman, 2007; Schneider, Swanson, & Riegle-Crumb, 1998; Wang, 2013a), and persist through to degree completion (Schneider, Swanson, & Riegle-Crumb, 1998). It is a positive sign, therefore, that high school students today are earning more STEM credits in high school compared to students in 1990 (Laird, Alt, & Wu, 2009).


Still, the increased enrollment in STEM coursework has not necessarily translated into an increased number of high school graduates who are engaged or proficient in STEM. Many argue that there remains a dearth of high school graduates who are qualified and interested in pursuing a career in STEM (Tyson et al., 2007). Moreover, minority students continue to underperform and be underrepresented in STEM compared to their nonminority counterparts (National Science Foundation, 2013). As a result, there have been growing concerns about the lack of diversity in the STEM workforce (Hu, 2014; Tsui, 2007) and the nation’s ability to maintain its global competitiveness in the face of demographic shifts at home and growing technical expertise abroad (National Science Board, 2007; President’s Council of Advisors on Science and Technology, 2012).


A number of solutions have been proposed to address the shortage of STEM-proficient high school graduates (H.R. 3373, 2011; Schultz et al., 2011). One approach has been the development of applied STEM coursework (e.g., the Carl D. Perkins Career and Technical Education Improvement Act), which seeks to increase STEM interest and retention by illustrating the interconnectedness and accessibility of STEM concepts (Plank, DeLuca, & Estacion, 2008). In high school, applied STEM courses differ from traditional STEM courses in several ways. Most notably, applied STEM courses stress the applicability of STEM content and in so doing are less theory-driven and more grounded in “real-world” uses (Gottfried, Bozick, & Srinivasan, 2014). In contrast, traditional STEM courses often emphasize theoretical underpinnings or concepts. Traditional mathematics courses, for example, are often typified by a heavy reliance on text-based instruction and lecture where students are expected to learn abstract concepts (e.g., the fundamental theorem of algebra) through practice problems and other assignments. Traditional coursetaking is also distinct from applied STEM coursetaking in the variety of courses students can take. Whereas a traditional math coursetaking sequence may begin with arithmetic and move on to introductory algebra, trigonometry, precalculus, and calculus, applied STEM may infuse mathematical concepts and practice in the instruction of computer programming.


As noted by Gottfried, Bozick, and Srinivasan (2014), there are two categories of applied STEM courses in high school: scientific research and engineering (SRE) courses and information technology (IT) courses. Examples of applied SRE courses include: Architectural Technologies, Structural Engineering, Civil Technologies, and Surveying. Examples of applied IT courses include: Computer and Information Sciences, Computer Mathematics, Computer Science and Computer Programming (e.g., C++, PASCAL). Approximately 26% of high school students enroll in an at least one applied STEM course (SRE or IT) by graduation.


A growing body of research has established the importance of applied STEM coursetaking on both secondary and postsecondary outcomes. For example, previous research has determined that applied STEM courses may help to propel students to take advanced traditional math and science courses in high school and to major in STEM fields in college (Bozick & Dalton, 2013; Gottfried et. al., 2014). What remains unknown, however, is which students are more or less likely to enroll in applied STEM courses by the time they graduate high school. This critical gap is concerning for several reasons. First, we still do not fully know which students are benefiting from these courses. To assist policymakers with the evaluation of programs like Perkins IV, it is essential that we systematically investigate which students are taking most advantage of applied STEM coursework. It is equivalently important that we determine which students are not enrolling in these courses, particularly given the aforementioned concerns about a lack of diversity in STEM fields.


Second, and related, the degree to which applied STEM courses may be reducing or widening the STEM “coursetaking gap” among student populations is unclear. Previous research and federal policy has long established that females, students with disabilities, and students from minority racial/ethnic backgrounds are underrepresented in STEM areas of study and the workplace (National Science Board, 2014; National Science Foundation, 2013; Tsui, 2007; Tyson et al., 2007). Applied STEM coursework has often been framed as a model with which to better engage students from these backgrounds (Baker & Leary, 1995; Chubin & Ward, 2009; Gottfried, Bozick, & Srinivasan, 2014; Hunter, Laursen, & Seymour, 2007; Seymour & Hewitt, 1997; Thompson & Windschitl, 2005; Weinberger, 2004; F. Wilson, 2003). Aggregately, research supports that both females and racial/ethnic minorities were uninspired by traditional STEM content due to its lack of engaging subject matter. Experiences that were more hands-on and active—two key characteristics of applied STEM courses—are supported as extremely influential for these groups. However, the lack of empirical investigation into whether these groups are more or less likely to take applied STEM courses prevents us from knowing the degree to which applied STEM has succeeded at increasing representation.


A third reason why the lack of research into applied STEM coursetaking is concerning is that we do not yet know how school characteristics are enabling or potentially deterring applied STEM participation. Previous research has indicated that school factors are indeed correlated with student coursetaking patterns particularly in STEM areas (Adelman, 2006; Benbow & Arjmand, 1990; Cullota, 1992; National Center for Education Statistics, 1994; Oakes, Gamoran, & Page, 1992). While this study is the first to evaluate school-level predictors of applied STEM coursetaking, it is entirely possible that students’ decisions to enroll in applied STEM are influenced by school characteristics, as guided by prior research. For example, school-level characteristics, such as increased graduation requirements (Teitelbaum, 2003) or school type (Lee, Chow-Hoy, Burkam, Geverdt, & Smerdon, 1998) may facilitate traditional STEM coursetaking. On the other hand, school-level characteristics, such as a lack of school resources (Gamoran, 1987) may discourage traditional STEM coursetaking. It is important, therefore, for researchers and policymakers to know if applied STEM coursetaking is similarly influenced by school factors.


A final reason why the lack of research into applied STEM coursetaking is concerning is that it remains unclear not only who enrolls in applied STEM, but when. This is a critical gap in the literature because the sequence in which students complete high school coursework is important. For instance, Schneider et al. (1998) found that the strongest predictor of mathematics and science coursetaking sequences in the 12th grade was students’ course sequences in the 10th grade. Perhaps of greater interest for the present study, however, was that Gottfried (2015) showed early applied STEM coursetaking (i.e., 9th and 10th grade) was predictive of advanced academic math and science coursetaking in the 12th grade. As the existing literature indicates, it is crucial that students adhere to rigid coursetaking sequences in high school (Adelman, 2006). This is particularly true for STEM courses, which often require students to meet prerequisites before they can take advanced STEM courses. It is clear that timing matters, hence the focus of the present study.


Therefore, the purpose of the current study is to identify the student and school characteristics predictive of applied STEM course enrollment in high school. The overarching aim of this study is to provide policymakers, school leaders, and education researchers much-needed insight into who is and who is not participating in applied STEM during high school. If it is the case that applied STEM courses attract particular student subgroups and populations and not others, and, in addition, if school characteristics appear to influence applied STEM coursetaking, appropriate policy action can be taken to ensure that policies which seek to broaden the number and variety of students entering the STEM pipeline are successful.


WHO ENROLLS IN STEM COURSES IN HIGH SCHOOL?


No research has yet examined which factors are associated with applied STEM coursetaking. However, previous research (e.g., Adelman, 1999, 2006) has found a number of significant student-level factors to be associated with traditional STEM coursetaking patterns in high school. Therefore, we rely on the vast extant literature on individual factors of academic STEM coursetaking to guide which student factors we explored in our assessment of applied STEM coursetaking.


STUDENT DEMOGRAPHICS


It is well established that gender differences are associated with academic STEM coursetaking patterns. For instance, Laird et al. (2009) found females completed more units in algebra, advanced biology, and chemistry. Males, on the other hand, completed more units in physics, engineering, and computer science. Comparing the advanced coursetaking of males and females, Tyson et al. (2007) found that while females in high school completed more advanced courses across all subjects, they were less likely than males to complete advanced math and science courses in particular. Even though those females who did enroll in advanced mathematics and science courses in high school had similar achievement levels as males in those courses, females enrolled in STEM-related postsecondary majors at lower rates.


In addition to gender, race/ethnicity appears to be linked to STEM coursetaking patterns among high school students (Blanchett, 2006; Nord et al., 2011). For instance, the aforementioned reported increases in STEM units earned among high schools in recent years does not necessarily extend to Black and Hispanic students (Laird et al., 2009); students in these groups continue to complete fewer STEM units than White and Asian students (Hoffer, Rasinski, & Moore, 1995). More so, Black and Hispanic students are more likely to enroll in remedial mathematics courses relative to White and Asian students (Tyson et al., 2007), making them less likely than White and Asian students to advance to high-level coursework by high school graduation (National Science Foundation, 1999). Black and Hispanic students are also less likely than White and Asian students to complete advanced placement (AP) STEM courses (R. Wilson, 2000).


Students’ English language proficiency appears to be linked to STEM coursetaking patterns as well. For example, Kanno and Kangas (2014) illustrated that even though they are the fastest growing segment of K–12 education, English language learners (ELL) are often underprepared for postsecondary success because this segment of the student population commonly attends schools without advanced-level, honors, or AP courses (Callahan, Wilkinson, & Muller, 2010). Therefore, it is important to examine to what degree ELL students participate in applied STEM courses during high school to see if, perhaps, teachers and counselors are steering ELL students away from applied STEM courses (Kanno & Kangas, 2014).


FAMILY CHARACTERISTICS


There is a long-established relationship between socioeconomic status (SES) and school achievement (e.g., Davis-Kean, 2005; Reardon, 2011; White, 1982). Evidence suggests this relationship carries over to STEM coursetaking as well (Adelman, 1999, 2006). Low-SES students complete fewer STEM units (Adelman, 2006; Hoffer et al., 1995) and are less likely to pursue STEM fields of study in high school (National Science Board, 2014; Tyson et al., 2007) and college (Clark, 1986; Wang, 2013b). However, King (1996) found that low-SES students who enrolled in advanced math and science courses were more likely to enroll in a 4-year college after graduation compared to low-SES students who did not. This provides evidence of the ways in which STEM coursetaking may be particularly influential for underrepresented groups. Similarly, measures of both parental education and occupation are also associated with STEM coursetaking patterns (Pearson, 1986). High school students whose parents were college educated and or worked in technical or managerial occupations were found to be more likely to pursue and persist in STEM (Astin & Astin, 1992; Grandy, 1998). While the current study examines coursetaking and not academic outcomes, it is plausible that, in addition to influencing students’ postsecondary majors, parents who are in STEM-oriented professions might encourage STEM coursetaking among their children as a way of engaging them into STEM-related field (Astin & Astin, 1992).


ACADEMIC ATTITUDES


Evidence suggests that high school STEM coursetaking patterns are related to a number of measures pertaining to students’ academic attitudes. For example, previous research has found that students’ reported levels of math and science self-efficacy can influence the degree to which they pursue more and advanced STEM courses (Colbeck, Cabrera, & Terenzini, 2001; Stevens, Olivarez, Lan, & Tallent-Runnels, 2004). Because self-efficacy in math and science is related to the degree to which a student believes STEM courses are important or useful (Pajares & Miller, 1994), it also seems plausible to assert that students who feel math and science are important will be more likely to take such courses. Indeed, Maltese and Tai (2011) argued that a student’s interest in STEM is more influential on STEM coursetaking and, ultimately, a student’s chances of earning a STEM-related degree than whether or not a student complete advanced-level STEM courses.


WHAT SCHOOL FACTORS CONTRIBUTE TO STEM COURSETAKING PATTERNS?


Previous research indicates that student behaviors in high school are a function of both student- and institutional-level characteristics (e.g., Anderman, 2002; Sebring, 1987). But just as with individual student characteristics related to applied STEM coursetaking, there is no existing research into the school-level factors that are correlated with students’ applied STEM coursetaking patterns during high school. A large body of literature exists that examines how school characteristics link to traditional STEM coursetaking. This body of literature, therefore, guided the present investigation.


As a key starting point, it is reasonable to assert that school curriculum offerings are highly associated with whether students can enroll in STEM courses (e.g., Kanno & Kangas, 2014). This curriculum is often influenced by school-level characteristics, such as size (Monk & Haller, 1993). The racial and socioeconomic composition of a school can also influence student behaviors, including coursetaking (Fowler & Walberg, 1991; Kelly, 2009).


A great number of studies have examined the association between school sector and student academic performance. For instance, Lee et al. (1998) found that students enrolled in private schools completed more advanced mathematics courses than public school students. Part of the justification for this finding was that Catholic schools exhibited greater control over student coursetaking than public and independent secondary schools; Catholic school administrators appeared to ensure that STEM coursetaking was distributed equitably among its student bodies. Similarly, Carbonaro and Covay (2010) found students enrolled in private schools completed more math courses than public school students, even after controlling for family background and prior achievement.


While researchers have examined the link between school urbanicity and STEM coursetaking, there has been little indication that a correlation exists between the two. Schiller and Muller (2003) found urbanicity was not significantly associated with the number of mathematics courses students completed, while Chaney, Burgdorf, and Atash (1997) found only marginal support for an association between the two. Still, because of the mixed evidence presented here, our study will consider whether school urbanicity is linked to students’ applied STEM coursetaking. Also, because the present study seeks insight into applied STEM coursework, which falls under the purview of CTE, we will test for an association between a student’s participation in applied STEM and whether a school offers vocational-technical education programs.


THIS STUDY


Very little research has systematically examined applied STEM coursetaking. None has examined what factors might be linked to the students taking these courses. Hence, the current study represents the first empirical investigation into the individual and institutional factors predictive of applied STEM coursetaking in high school. To that end, we posed the following three research questions:


1. What student factors are related to applied STEM coursetaking in high school?

2. What school factors are related to applied STEM coursetaking in high school?

3. How is the influence of these factors different based on the timing of course taken?


To answer these questions, we relied on nationally representative data that contained student, parent, and school characteristics as well as complete student academic transcripts. For advocates of STEM education, applied STEM coursework is often framed as a way to increase STEM interest and retention, especially among students who are traditionally underrepresented in STEM fields of study. While applied STEM is associated with improved STEM outcomes in high school and college (Bozick & Dalton, 2013; Gottfried et. al., 2014), we still do not know who is enrolling in these courses to begin with and when these students take them. Furthermore, we do not yet know to what degree applied STEM courses have worked to reduce the “STEM gap” in high school. The current study seeks to address these critical lapses in the literature and to provide important information for policymakers, school leaders, and educational researchers as they attempt to increase the number of high school graduates prepared for STEM careers and/or postsecondary study. If it is the case that disparities in applied STEM coursetaking emerge from an analysis of the data, then appropriate next steps can be taken to encourage more equal participation. Also, if school-level factors are related to increased or decreased applied STEM coursetaking, then we can direct future research and suggest appropriate interventions.


METHOD


DATASET


This study used nationally representative longitudinal data collected by the National Center for Education Statistics (NCES). The dataset, the Education Longitudinal Study of 2002 (ELS:2002), was gathered from a cohort of 10th-grade students from 750 schools beginning in 2002, the 1st year of data collection (i.e., base year). Follow-up data were then gathered in 2004, 2 years later, when the majority of students were in 12th grade. The students in the ELS:2002 study were followed after high school graduation as they entered the workforce or began postsecondary enrollment. For the purposes of the present study, which focused on high school coursetaking, we focused on the data collected in the base year and first follow up. NCES collected official high school transcripts for the students, which were merged with the 10th- and 12th-grade data. This merging provided complete high school transcript data for approximately 91% of the students included in the base year sample. Importantly, parent and teacher questionnaires were also gathered in the base year (2002) of the study. ELS:2002, therefore, represents a rich, nationally representative multilevel dataset with complete student coursetaking records along with parent, teacher, and school administrator data.


Though the first wave of data collection occurred in 2002, the educational experiences of students in the sample remain relevant to today’s landscape. While there have been some changes in the types of technology injected into applied STEM curricula, these have mostly been isolated instances and not implemented nationally. In other words, the purpose and curriculum have generally remained the same since the time of the collection of this dataset. Courses that use more advanced technology still rely on applied STEM curricula developed in previous years, and thus, this nationally representative dataset here in this study remains highly relevant.


Included in our analysis were students within the ELS:2002 sample who had participated in the first and second waves of data collection and for whom detailed coursetaking information (i.e., applied STEM participation) was available. Importantly, there were a number of missing observations on the key variables of interest for the current study, and in order to maintain sufficient statistical power while also controlling for the potential influence of missing observations, we chose to substitute missing observations with mean values as done in prior work using large-scale data from the U.S. Department of Education (e.g., Gottfried et. al., 2014). After mean value substitution, we created dummy variables that indicated whether a participant had a missing observation; while we do not report the estimated effects of these indicators in our tables, they were included in all regression analyses. After mean value imputation, our final analytic sample was composed of 16,190 student observations nested within 750 high schools.


MEASURES


Outcomes


There were four key outcome variables in the present study. For the first and second research questions, the main outcome variable of interest was a binary indicator of enrollment in at least one applied STEM course during high school. In fact, most students only take one applied STEM course in high school—in the dataset, less than 5% of all students take more than one course. To construct this variable, we relied on the Secondary School Taxonomy published by NCES. This taxonomy divided all high school courses into four mutually exclusive categories: academic, career and technical education (CTE), enrichment/other, and special education. All applied STEM courses fall within the CTE curriculum and in that way are distinct from academic math and science courses, which fall within the “academic” curriculum. Note that courses classified as applied STEM cannot be classified as part of any other category according to this taxonomy.


Within the CTE curriculum, there are 16 categories of applied courses. Two of these categories contain applied STEM coursework: scientific research and engineering (SRE) and information technology (IT). Course titles are identified within these two clusters, which we then aligned with the Secondary School Taxonomy. Students who had enrolled at least one applied STEM course (SRE or IT) by high school graduation were coded “1”; students who never enrolled in applied STEM were coded “0.” Approximately 26% of the students in our final analytic sample enrolled in at least one applied STEM course by the time they graduated from high school. Note that students take applied STEM in addition to academic STEM courses. While the classification of these courses is mutually exclusive, coursetaking in both areas is not. Therefore, students’ being coded as 1 for applied STEM coursetaking does not imply that they did not take academic STEM courses.


Within research questions 1 and 2, to test for differential effects pertaining to the type of applied STEM courses (SRE or IT), we created two additional outcome variables. The first was a binary measure of whether a student had enrolled in at least one SRE-specific applied STEM course. Students who had enrolled at least one SRE course by high school graduation were coded “1”; students who never enrolled in an SRE course were coded “0.” Approximately 9% of the students in our final analytic sample enrolled in at least one SRE course by the time they graduated from high school. The second was a binary measure of whether a student had enrolled in at least one IT-specific applied STEM course. Students who had enrolled at least one IT course by high school graduation were coded “1”; students who never enrolled in an SRE course were coded “0.” Approximately 20% of the students in our final analytic sample enrolled in at least one IT course by the time they graduated from high school.


We constructed a fourth and final outcome variable to answer the third research question of how the relationship between student and school factors of applied STEM coursetaking was different based on when a student enrolled in applied STEM. In more detail, we divided applied STEM coursetaking into three categories: early, late, and never. We used the Secondary School Taxonomy again to create this variable. Early applied STEM enrollment was defined as having enrolled in applied STEM (SRE or IT) in either 9th or 10th grade; late applied STEM enrollment was defined as having enrolled in applied STEM (SRE or IT) in either 11th or 12th grade. Approximately 16% of students in our analytic sample enrolled in applied STEM early in high school. Approximately 15% of students enrolled in applied STEM late (i.e., 11th or 12th). Just 745 students, or 4% of the students in our analytic sample, took applied STEM in both periods. Roughly 74% of students never enrolled in an applied STEM course.


Student-Level Factors


We included a number of student-level measures based on previous research into high school STEM coursetaking. Independent student measures in the analyses were derived from the first wave of data collection. Table 1 presents descriptive statistics broken out by students who took applied STEM courses during high school versus those who did not. Students who enrolled in applied STEM during high school were significantly different from their counterparts who did not in many of the measures; this illustrated that applied STEM coursework appears to draw a distinct cohort of high school students and that further analyses were necessary.


Table 1. Descriptive Statistics and Mean Comparisons of Student-Level Controls (N = 16, 190)

 

 

 

 

 

 

 

 

 

 

 

Never Taken Applied STEM

 

Taken Applied STEM

 

Mean

 

Standard deviation

 

Mean

 

Standard deviation

Student Demographics

 

 

 

 

 

 

 

 

 

Female

0.55

***

 

0.50

 

0.36

***

 

0.48

Race

 

 

 

 

 

 

 

 

 

         Hispanic

0.15

***

 

0.36

 

0.12

***

 

0.33

         White

0.71

 

 

0.45

 

0.72

 

 

0.45

         Black

0.17

*

 

0.38

 

0.16

*

 

0.37

         Asian

0.12

 

 

0.33

 

0.13

 

 

0.33

Ever in ESL program

0.09

*

 

0.26

 

0.07

*

 

0.27

English is primary language

0.83

 

 

0.38

 

0.84

 

 

0.37

Family Characteristics

 

 

 

 

 

 

 

 

 

Socioeconomic status

0.04

 

 

0.75

 

0.06

 

 

0.72

Parents graduated high school

0.19

 

 

0.39

 

0.19

 

 

0.39

Parents graduated from college

0.23

 

 

0.41

 

0.23

 

 

0.41

Academic Background

 

 

 

 

 

 

 

 

 

Number of academic risk factors

1.02

***

 

1.12

 

0.91

***

 

1.03

Student is in vocational program

0.09

***

 

0.29

 

0.11

***

 

0.32

Cumulative GPA

2.70

***

 

0.80

 

2.76

***

 

0.72

Student has an IEP

0.14

***

 

0.34

 

0.09

***

 

0.29

Ever in remedial math

0.10

 

 

0.30

 

0.10

 

 

0.30

Thinks math is important

2.54

***

 

2.51

 

2.44

***

 

2.42

Student academic aspirations

5.18

 

 

1.46

 

5.23

 

 

1.36

***p<.001,**p<.01, *p<.05

 

 

 

 

 

 

 

 

 


Student demographic measures used in the current study included indicators for: gender, race/ethnicity, whether a student was enrolled in an English as a second language (ESL) program, and whether English was a student’s native language. Family characteristics included student SES, which was a standardized composite measure provided in ELS:2002, and parental education level, which was measured using two dichotomous indicators of whether a student’s parents had completed high school or earned a bachelor’s degree. The gender, race/ethnicity, and language variables were sourced from the base year student questionnaire; variables related to SES and parent’s education level were sourced from the base year parent questionnaire.


A final category of individual student measures pertained to students’ academic background. These variables included the number of academic risk factors a student had in 10th grade, the base year of data collection. This particular measure was created by NCES. Academic risk factors included: (1) coming from a single parent household, (2) having two parents without a high school diploma, (3) having a sibling who has dropped out of school, (4) changing schools two or more times, (5) having repeated a grade, and (6) having a family income below the poverty line. Additional academic background controls included whether or not a student was enrolled in a vocational program by the 10th grade and a student’s cumulative grade point average (measured on a 4-point scale) throughout high school. We also included dummy indicators for whether a student had been flagged as having an individualized education plan (IEP) in Grade 10, and whether a student needed remedial math instruction. ELS:2002 transcript files were used to source information pertaining to GPA, remedial course enrollment, and vocational program participation. A students’ IEP indicator was derived from the school administrator survey, and academic risk was derived from the parent questionnaire.


Finally, we included academic attitudes. A measure for how important a student viewed math was based on a self-report scale that ranged from 1 = “strongly agree” to 4 = “strongly disagree.” To aid interpretation, we chose to reverse code this variable such that a student who “strongly agreed” with the statement that “mathematics is important” was coded 4, and a student who “strongly disagreed” was coded 1. A measure of academic aspiration was included based on a survey measure ranging from 1 to 7, where 1 = a student who planned to earn less than a high school degree and 7 = a student who planned to obtain an advanced graduate degree (e.g., Ph.D. or MD). The base year student questionnaire provided the variables related to math importance and academic aspiration.


School-level Measures


School variables were also included in our models of applied STEM coursetaking. These variables were all measured in the base year, wave 1, of data collection and included school control (i.e., public vs private), the percentage of racial/ethnic minority students, and the percentage of students receiving free and reduced lunch. Additional school-level controls included a binary indicator of charter school status, school size measured by total student enrollment, whether a school was located in an urban setting, and whether the school offered vocational-technical programs to students. We coded all dichotomous variables such that 0 = no and 1 = yes. Lastly, we included two variables that measured the number of full-time math and science teachers in a school during the base year of collection.


These variables and their descriptive statistics are presented in Table 2. This table also provides the results of mean comparison tests between schools that offered applied STEM courses and those that did not. Results of the tests indicate that schools that offered applied STEM courses significantly differed from those that did not with regard to racial/ethnic composition and school size. Significant differences were also observed in the number of full-time math and science teachers. Schools that offered applied STEM to students were more likely to also offer vocational-technical programs. No significant differences were observed with regard to school control, charter status, urbanicity, or the percentage of students receiving free or reduced lunch.


Table 2. Descriptive Statistics and Mean Comparisons for School-Level Controls (N = 750)

 

 

 

 

 

 

 

 

 

Does Not Offer Applied STEM

 

Offers Applied STEM

 

Mean

Standard deviation

 

Mean

Standard deviation

School Characteristics

 

 

 

 

 

 

 

School control

0.76

 

0.43

 

0.84

 

0.36

Percent minority

25.26

***

29.98

 

35.26

***

30.86

Percent free lunch

25.58

 

18.17

 

25.97

 

18.51

Charter school

0.00

 

0.00

 

0.10

 

0.10

School size

1,142.20

***

757.53

 

1,437.74

***

859.51

Urban

0.29

 

0.46

 

0.26

 

0.44

Offers vocational programs

0.64

***

1.83

 

0.90

***

0.96

Number of full-time math teachers

8.01

***

5.10

 

10.42

***

6.11

Number of full-time science teachers

7.60

***

4.69

 

9.19

***

5.53

***p<.001, **p<.01, *p<.05

 

 

 

 

 

 

 


To provide additional clarity around our analytic sample and to yield some additional insight into applied STEM coursetaking behaviors in high school, Table 3 presents the mean and standard deviations of the previously described student controls divided into four categories: no applied STEM coursetaking in high school, early applied STEM coursetaking in high school (i.e., 9th or 10th grade), late applied STEM coursetaking (11th and 12th grade), and “distributed” applied STEM coursetaking (i.e., both early and late). Of the students who took applied STEM, there was a relatively even split between early and late enrollment. In other words, we see no evidence that a substantial proportion of applied STEM coursetakers did so early over late or vice versa. Table 3 also shows that only 745 students in our sample took applied STEM courses both in early and later years.


The similarity in means across the four columns reveals few observable differences between early and late applied STEM coursetakers. One noticeable exception, however, is that the proportion of females from our sample who did not enroll in an applied STEM course is markedly higher than the proportion of females who enrolled in applied STEM courses in high school. Overall, Table 3 shows that students in our sample took applied STEM courses either early or late and few students took courses in both periods. The decomposition of student characteristics does not provide support for the existence of systematic variation between students belonging to the four applied STEM coursetaking sequences.


Table 3. Patterns of Applied STEM Coursetaking

 

 

 

 

No Applied STEM

 

Applied STEM Early (9 & 10)

 

Applied STEM Late (11 & 12)

 

Applied STEM Early & Late

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mean

sd

 

Mean

sd

 

Mean

sd

 

Mean

sd

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Student Demographics

 

 

 

 

 

 

 

 

 

 

 

 

 

Female

 

 

0.55

 

 

0.39

 

 

0.39

 

 

0.25

 

 

Race

 

 

 

 

 

 

 

 

 

 

 

 

 

 

         Hispanic

 

 

0.14

 

 

0.11

 

 

0.13

 

 

0.11

 

 

         White

 

 

0.71

 

 

0.73

 

 

0.69

 

 

0.75

 

 

         Black

 

 

0.13

 

 

0.11

 

 

0.14

 

 

0.10

 

 

         Asian

 

 

0.08

 

 

0.09

 

 

0.09

 

 

0.10

 

 

Ever in ESL program

 

 

0.09

 

 

0.08

 

 

0.09

 

 

0.07

 

 

English is primary language

 

 

0.83

 

 

0.85

 

 

0.82

 

 

0.84

 

 

Family Characteristics

 

 

 

 

 

 

 

 

 

 

 

 

 

Socioeconomic status

 

 

0.04

0.72

 

0.06

0.69

 

0.06

0.72

 

0.06

0.67

 

Parents graduated high school

 

 

0.20

 

 

0.20

 

 

0.19

 

 

0.22

 

 

Parents graduated from college

 

 

0.23

 

 

0.23

 

 

0.23

 

 

0.21

 

 

Academic Background

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of academic risk factors

 

 

1.01

0.96

 

0.93

0.91

 

0.93

0.88

 

0.93

0.91

 

Student is in vocational program

 

 

0.09

 

 

0.11

 

 

0.10

 

 

0.15

 

 

Cumulative GPA

 

 

2.70

0.75

 

2.72

0.76

 

2.77

0.68

 

2.83

0.67

 

Student has an IEP

 

 

0.13

 

 

0.11

 

 

0.11

 

 

0.10

 

 

Ever in remedial math

 

 

0.10

 

 

0.10

 

 

0.09

 

 

0.10

 

 

Thinks math is important

 

 

2.47

0.75

 

2.51

0.77

 

2.54

0.76

 

2.58

0.76

 

Student academic aspirations

 

 

5.19

1.34

 

5.20

1.32

 

5.26

1.22

 

5.21

1.22

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n

 

 

11,985

 

1,810

 

1,760

 

745

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


ANALYSIS PLAN


Research Question 1


The first research question sought to discover the individual student characteristics predictive of applied STEM coursetaking. To answer this question, we employed two models. Our first model included the student and family sociodemographic characteristics described in the previous section. Our second model then introduced students’ academic history and attitudes measures. This final model can be written as follows:


[39_21859.htm_g/00002.jpg]


where [39_21859.htm_g/00004.jpg] is a dichotomous outcome as to whether student i in school s enrolled in at least one applied STEM course during high school. The independent predictors previously described are also included: [39_21859.htm_g/00006.jpg] refers to a set of student sociodemographic characteristics (e.g., gender, ethnicity, English proficiency),  [39_21859.htm_g/00008.jpg] refers to the set of student family background characteristics (e.g., SES and parent’s education level), and [39_21859.htm_g/00010.jpg] refers to students’ academic history and attitudes (e.g., GPA, academic risk factors, academic aspirations). To avoid any potential bias from unobserved school-level influences, the error term [39_21859.htm_g/00012.jpg] in the model above is estimated with robust standard errors adjusted for school clustering.1


Research Question 2


The second research question inquired into the effects of the institutional characteristics of applied STEM coursetaking. We sought to test this relationship after controlling for student characteristics predictive of applied STEM enrollment and therefore constructed the following model:


[39_21859.htm_g/00014.jpg]


where again [39_21859.htm_g/00016.jpg] is a dichotomous outcome as to whether student i in school s enrolled in at least one applied STEM course during high school. The same set of student control measures were also included; however, this final model introduced a new term, [39_21859.htm_g/00018.jpg], which represents a set of school-level characteristics hypothesized to be associated with applied STEM coursetaking. Therefore, in sum, we employed a total of three analytic models to address our first two research questions. Models 1 and 2 tested for associations between applied STEM coursetaking and student, or within school, characteristics; Model 3 built upon these by testing for associations between applied STEM coursetaking and institutional, or between-school, characteristics net of any student-level influences.


Because the outcomes for both research questions 1 and 2 represented binary measures (e.g., enrollment vs no-enrollment) all models described were estimated using binary logistic regression. All results attributable to binary predictors are reported as odds ratios, which can be interpreted as an increase or decrease in the odds of applied STEM coursetaking at any point during high school. Effect sizes are also reported in text in order to aid the interpretation of odds ratios associated with continuous covariates.


Research Question 3


Our third research question inquired into whether the associations between student- and school-level characteristics and applied STEM coursetaking were different based on when a student enrolled in these courses. To answer this question, we followed a similar analytic procedure to the one we used to address research questions 1 and 2. The difference between the models we specified for research question 3 and those previously described concerns the outcome variable, which in the case of research question 3 was a multinomial outcome with m = 3 values: “early applied STEM enrollment” (i.e., 9th and 10th grades), “late applied STEM enrollment” (i.e., 11th and 12th grade), and “never enrolled in applied STEM in high school”:


[39_21859.htm_g/00020.jpg]


For each model, “never enrolled in applied STEM in high school” was used as the reference category so that results could be interpreted as relative to never taking applied STEM in high school. Multinomial binary logistic regression was used to answer research question 3. All estimates related to binary predictors are presented as relative risk ratios (RRR), which are interpreted as the odds of enrolling in applied STEM such that a risk ratio of < 1 represents decreased odds of applied STEM enrollment and a risk ratio of > 1 represents increased odds.


RESULTS


RESEARCH QUESTION 1


Table 4 presents the odds ratios from the logistic regression models predicting applied STEM coursetaking in high school based on the wide span of individual and institutional measures. Odds greater than a value of 1 suggest greater chances of applied STEM coursetaking. For binary measures, the comparison is made between the indicator and the reference group (e.g., males versus females). For continuous measures or scales, the comparison is made between having higher or lower values of the independent measure. Robust standard errors adjusted for school clustering associated with each control variable are reported in parentheses. Significance levels are indicated as well.


Table 4. Odds Ratios From Logistic Regression Models Predicting Applied STEM Coursetaking in High School

 

 

 

 

 

 

 

Model 1

Model 2

Model 3

 

 

 

 

 

Student Demographics

 

 

 

 

Female

 

0.45***

0.42***

0.42***

 

 

(0.03)

(0.02)

(0.02)

Hispanic

 

0.78**

0.82*

0.99

 

 

(0.08)

(0.09)

(0.11)

Black

 

0.92

0.99

1.16

 

 

(0.10)

(0.10)

(0.10)

Asian

 

1.08

1.01

1.20*

 

 

(0.12)

(0.11)

(0.13)

Student is in ESL program

 

0.87*

0.91

0.90

 

 

(0.07)

(0.07)

(0.07)

English is primary language

 

1.01

1.00

0.92

 

 

(0.09)

(0.09)

(0.09)

Family Characteristics

 

 

 

 

Socioeconomic status

 

1.00

0.90**

0.94

 

 

(0.05)

(0.04)

(0.04)

Parents graduated from high school

 

1.00

0.99

0.97

 

 

(0.05)

(0.05)

(0.05)

Parents graduated with BA degree

 

0.97

0.96

0.97

 

 

(0.05)

(0.05)

(0.05)

Academic Background

 

 

 

 

Academic risk factors

 

 

0.95**

0.95**

 

 

 

(0.03)

(0.02)

Vocational or technical program

 

 

1.24***

1.23**

 

 

 

(0.10)

(0.10)

Cumulative GPA

 

 

1.19***

1.15***

 

 

 

(0.04)

(0.04)

Student has an IEP

 

 

0.64***

0.64***

 

 

 

(0.07)

(0.07)

Student enrolled in remedial math

 

 

0.95

0.94

 

 

 

(0.07)

(0.07)

Math is important

 

 

1.06**

1.06**

 

 

 

(0.03)

(0.03)

Academic aspirations

 

 

1.03*

1.04**

 

 

 

(0.02)

(0.02)

School Characteristics

 

 

 

 

Comprehensive public school

 

 

 

0.86

 

 

 

 

(0.12)

Percent ethnic minority students

 

 

 

0.99*

 

 

 

 

(0.00)

Percent free and reduced lunch

 

 

 

1.00

 

 

 

 

0.03

Charter school

 

 

 

0.50

 

 

 

 

(0.19)

School size

 

 

 

1.00**

 

 

 

 

(0.00)

Urbanicity

 

 

 

1.06

 

 

 

 

(0.12)

School offers vocational programs

 

 

 

1.25

 

 

 

 

(0.21)

Number of full-time math teachers

 

 

 

1.01

 

 

 

 

(0.02)

Number of full-time science teachers

 

 

 

0.98

 

 

 

 

(0.02)

 

 

 

 

 

Students

 

16,190

16,190

16,190

Schools

 

750

750

750

 

 

 

 

 

Note: Estimated coefficients are reported in odds ratios. Robust standard errors adjusted for school clustering are in parentheses.

***p<.001,**p<.01, *p<.05

 

 

 

 


Model 1 begins with student and family sociodemographic characteristics. The results suggested that females were much less likely to enroll in applied STEM courses during high school compared to males. In addition, Hispanic students were significantly less likely to take applied STEM courses compared to White students. However, Black students and Asian students were not more or less likely to enroll compared to White students. There was no association between applied STEM coursetaking and SES, ESL participation, English proficiency, or parent education level.


Model 2 introduced students’ academic variables. Note that once we accounted for academic controls such as academic risk factors, enrollment in a vocational-technical program, GPA, disability, remedial math enrollment, math importance, and academic aspirations, females remained less likely to enroll in applied STEM courses relative to Model 1 (in fact, the odds are lower in the table). However, the association between applied STEM coursetaking and race was tempered for Hispanic students. Again, ESL participation, English proficiency, and parent education were not associated with applied STEM coursetaking.   


Measures pertaining to students’ academic characteristics appeared to be related to applied STEM coursetaking. In more detail, academic risk factors were negatively associated with taking an applied STEM course in high school (d = -0.03). On the other hand, students enrolled in a vocational program and those with higher cumulative GPAs (d = 0.11) were more likely to enroll in applied STEM during high school. Students with disabilities were much less likely to take these courses. However, students who strongly agreed that math was important were statistically more likely to enroll in applied STEM at some point in high school (d = 0.04). Whether a student needed remedial math instruction was unrelated to applied STEM participation. Students’ academic aspirations were unrelated as well.


RESEARCH QUESTION 2


We answered the second research question by building on Model 2. More specifically, we added school-level variables to the model while still clustering the errors at the school level in order to estimate the association between institutional characteristics and applied STEM coursetaking. The results of this final logistic regression model are listed in far right column of Table 4.


As for student-level characteristics, looking at the estimates of Model 3, we see that females continued to be much less likely than males to enroll in applied STEM courses in high school. After including school characteristics in the model, the associations between being Hispanic and applied STEM enrollment were completely mediated. Note that any statistical relationship between student SES and applied STEM coursetaking also disappeared with the inclusion of school-level variables. Importantly, the associations found between the previously discussed student academic characteristics in Model 2 remained statistically significant even after including school factors.


The introduction of institutional factors in Model 3 allowed us to answer the second research question. Once we added these variables, we noticed that while the percentage of minority students in a school and school size were both significantly associated with a student’s decision to enroll in applied STEM, the estimated effect sizes associated with percentage of minority students in a school (d = -0.01) and school size (d = 0.00) were both close to 0. In other words, there lacks practical significance given the magnitude of the effects. It appears, therefore, that school-level characteristics do not necessarily predict applied STEM coursetaking, though including them in the model did mediate many of the individual-level factors that were previously statistically significant in Models 1 or 2.


DIFFERENCES BY COURSE TYPE


To determine if the individual and institutional factors of applied STEM coursetaking would change based on the type of applied STEM course took during high school, we disaggregated applied STEM into scientific research and engineering (SRE) and information technology (IT) specific courses. We then re-ran the third model listed in Table 4 for each of these outcomes: (1) a binary indicator of SRE enrollment and (2) a binary indicator of IT enrollment. Table 5 presents the odds ratios from these two new models along with the fully expressed model from Table 4. Again, robust standard errors adjusted for school clustering are reported in parentheses.


In terms of applied STEM coursetaking in general, females were much less likely than males to enroll in all types of applied STEM courses. For both SRE- and IT-specific courses, the odds of enrollment in either types of courses among females remained low. In fact, females were even less likely to take IT courses than previous models would have suggested when we examined applied STEM coursetaking more aggregately. No other sociodemographic characteristic besides gender shone through so consistently across all models.


There was a degree of consistency in interpretation when examining academic characteristics across the applied STEM model and specific the SRE and IT models. However, differences arose when specifically examining SRE versus IT. As in the general applied STEM model, having more academic risk factors was linked to reduced IT coursetaking (d = -0.06), though there was no link necessarily to SRE coursetaking. Participation in a vocational-technical program was associated with increased odds of IT coursetaking but not SRE coursetaking. A student’s cumulative GPA was not positively related to IT enrollment as it was with applied STEM enrollment in general, though SRE enrollment was predicted with higher GPAs (d = 0.13). Having an IEP in the 10th grade was negatively associated with SRE coursetaking but not IT coursetaking. Also, strongly believing math was important was positively related to IT enrollment (d = 0.06) but not SRE; having higher academic aspirations was positively related to SRE but not IT.


Table 5. Odds Ratios from Logistic Regression Models Predicting Applied STEM Coursetaking in High School

 

 

 

 

 

 

 

General Applied STEM

SRE Applied STEM

IT Applied STEM

 

 

 

 

 

Student Demographics

 

 

 

Female

 

0.42***

0.56***

0.24***

 

 

(0.02)

(0.03)

(-0.02)

Hispanic

 

0.99

1.05

0.77**

 

 

(0.11)

(0.14)

(-0.09)

Black

 

1.16

1.31***

0.89

 

 

(0.10)

(0.13)

(-0.12)

Asian

 

1.20*

1.25

1.05

 

 

(0.13)

(0.15)

(-0.15)

Student is in ESL program

0.90

0.91

0.95

 

 

(0.07)

(0.08)

(-0.11)

English is primary language

0.92

0.88

1.00

 

 

(0.09)

(0.10)

(-0.11)

Family Characteristics

 

 

 

Socioeconomic status

0.94

1.00

0.81***

 

 

(0.04)

(0.05)

(-0.05)

Parents graduated from high school

0.97

1.01

0.93

 

 

(0.05)

(0.06)

(-0.07)

Parents graduated with BA degree

0.97

0.96

0.97

 

 

(0.05)

(0.05)

(-0.07)

Academic Background

 

 

 

Academic risk factors

0.95**

0.99

0.91**

 

 

(0.02)

(0.03)

(-0.04)

Vocational or technical program

1.23**

1.07

1.59***

 

 

(0.10)

(0.10)

(-0.19)

Cumulative GPA

1.15***

1.23***

0.98

 

 

(0.04)

(0.05)

(-0.05)

Student has an IEP

0.64***

0.54***

0.82

 

 

(0.07)

(0.07)

(-0.11)

Student enrolled in remedial math

0.94

0.93

1.00

 

 

(0.07)

(0.08)

(-0.10)

Math is important

1.06**

1.03

1.10**

 

 

(0.03)

(0.03)

(-0.04)

Academic aspirations

1.04**

1.06***

1.01

 

 

(0.02)

(0.02)

(-0.02)

School Characteristics

 

 

 

Comprehensive public school

0.86

0.76

1.21

 

 

(0.12)

(0.12)

(-0.25)

Percent ethnic minority students

1.00*

1.00

0.99***

 

 

(0.00)

(0.00)

(0.00)

Percent free and reduced lunch

1.00

1.00

1.00

 

 

(0.00)

(0.00)

(-0.01)

Charter school

 

0.50*

0.51

0.57

 

 

(0.19)

(0.23)

(-0.21)

School size

 

1.00**

1.00***

1.00

 

 

(0.00)

(0.00)

(0.00)

Urbanicity

 

1.06

1.22

0.87

 

 

(0.12)

(0.15)

(-0.13)

School offers vocational programs

1.25

1.07

1.81**

 

 

(0.21)

(0.19)

(-0.51)

Number of full-time math teachers

1.01

1.03

0.95

 

 

(0.02)

(0.03)

(-0.03)

Number of full-time science teachers

0.98

0.98

1.02

 

 

(0.02)

(0.02)

(-0.04)

 

 

 

 

 

Students

 

16,190

16,190

16,190

Schools

 

750

750

750

 

 

 

 

 

Note: Estimated coefficients are reported in odds ratios. Robust standard errors adjusted for school clustering are in parentheses.

***p<.001,**p<.01, *p<.05



As before, few school-level characteristics were linked to any type of applied STEM coursetaking. In cases where there was statistical significance, the coefficients did not indicate any practical significance with applied STEM coursetaking (i.e., the odds were close to 1-to-1 and the effect sizes were practically 0). In one notable exception, however, students who attended a school with vocational-technical programs were nearly twice as likely to enroll in IT courses compared to students in schools that did not have these programs. This relationship was not observed when examining aggregated applied STEM coursetaking.


In summary, the results of the analyses of applied STEM enrollment disaggregated by course type (SRE/IT) revealed how the individual and institutional factors of applied STEM coursetaking differently predicted SRE versus IT enrollment. For the most part, differences arose in the factors that predicted SRE versus IT enrollment, though the overall interpretation remained consistent when looking across SRE and IT models. However, there was consistency across the models when it came to gender: regardless of model, females were less likely to enroll in any type of applied STEM course.


RESEARCH QUESTION 3


We followed the same modeling strategy we used to address the first two questions to answer question 3. Our first model included covariates related to students’ sociodemographic and family background characteristics. Our second model then introduced students’ academic history and attitudes measures. Our third and fully expressed model added school characteristics. The outcome variable for the third research question was a multinomial measure of applied STEM coursetaking: early (i.e., 9th and 10th), late (i.e., 11th and 12th) and never. We used “never” as the reference category so that results can be interpreted as the likelihood of enrolling in applied STEM coursework early in high school versus never enrolling and the likelihood of enrolling in applied STEM later in high school versus never enrolling. The results from these models are listed in Table 6 as relative risk ratios, where again coefficients > 1 represent increased likelihood and coefficients < 1 represent decreased likelihood. Cluster adjusted robust standard errors are presented in parentheses.


The results from the first model comparing early applied STEM coursetaking versus never taking an applied STEM course revealed that females were significantly less likely to enroll in an applied STEM course early in high school. Hispanic students also appeared to be less likely to enroll in applied STEM early in high school. On the other hand, students whose primary language was English were more likely to enroll in applied STEM early in high school.


Model 2 introduced students’ academic history and attitudes variables. After accounting for academic history, females remained less likely to enroll in applied STEM courses early in high school. Similarly, Hispanic students continued to experience reduced odds of enrollment. Students whose primary language was English were still more likely to enroll in applied STEM early in school. With regard to academic history and attitudes measures, only having an IEP in 10th grade was significantly associated with applied STEM coursetaking. In more detail, the table shows that students with an IEP were less likely to enroll in applied STEM courses early in high school.



Table 6. Risk Ratios From Multinomial Regression Models Predicting Applied STEM Coursetaking in High School

 

 

 

 

 

 

 

 

 

 

 

 

Model 1

Model 2

Model 3

Model 1

Model 2

Model 3

 

 

Early vs Never

Early vs Never

Early vs Never

Late vs Never

Late vs Never

Late vs Never

 

 

 

 

 

 

 

 

Student Demographics

 

 

 

 

 

 

Female

 

0.50***

0.48***

0.48***

0.42***

0.38***

0.38***

 

 

(0.04)

(0.04)

(0.04)

(0.03)

(0.02)

(0.02)

Hispanic

 

0.77**

0.78**

0.97

0.79*

0.85

1.00

 

 

(0.09)

(0.09)

(0.11)

(0.10)

(0.11)

(0.14)

Black

 

0.77

0.80

0.95

1.03

1.15

1.32***

 

 

(0.12)

(0.12)

(0.13)

(0.11)

(0.13)

(0.13)

Asian

 

1.18

1.13

1.39**

1.02

0.94

1.08

 

 

(0.16)

(0.16)

(0.19)

(0.13)

(0.12)

(0.13)

Student is in ESL program

0.85

0.86

0.85

0.89

0.94

0.93

 

 

(0.10)

(0.10)

(0.10)

(0.09)

(0.09)

(0.09)

English is primary language

1.20*

1.20*

1.08

0.89

0.89

0.82*

 

 

(0.13)

(0.13)

(0.12)

(0.10)

(0.10)

(0.10)

Family Characteristics

 

 

 

 

 

 

Socioeconomic status

0.96

0.90

0.92

1.02

0.90*

0.95

 

 

(0.06)

(0.06)

(0.05)

(0.05)

(0.05)

(0.05)

Parents graduated from high school

0.96

0.95

0.93

1.02

1.02

1.00

 

 

(0.07)

(0.07)

(0.07)

(0.07)

(0.07)

(0.07)

Parents graduated with BA degree

0.98

0.98

0.98

0.96

0.95

0.95

 

 

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

Academic Background

 

 

 

 

 

 

Academic risk factors

 

0.95

0.95

 

0.94**

0.94**

 

 

 

(0.04)

(0.04)

 

(0.03)

(0.03)

Vocational or technical program

 

1.16

1.16

 

1.30***

1.28***

 

 

 

(0.11)

(0.11)

 

(0.12)

(0.12)

Cumulative GPA

 

1.06

1.02

 

1.29***

1.26***

 

 

 

(0.06)

(0.05)

 

(0.05)

(0.05)

Student has an IEP

 

0.64***

0.65***

 

0.65***

0.64***

 

 

 

(0.10)

(0.10)

 

(0.08)

(0.08)

Student enrolled in remedial math

 

1.00

0.98

 

0.92

0.91

 

 

 

(0.10)

(0.10)

 

(0.08)

(0.08)

Math is important

 

1.04

1.04

 

1.07**

1.08**

 

 

 

(0.04)

(0.04)

 

(0.03)

(0.03)

Academic aspirations

 

1.02

1.03

 

1.03*

1.05**

 

 

 

(0.02)

(0.02)

 

(0.02)

(0.02)

School Characteristics

 

 

 

 

 

 

Comprehensive public school

 

 

0.77

 

 

0.94

 

 

 

 

(0.14)

 

 

(0.13)

Percent ethnic minority students

 

 

0.99*

 

 

1.00

 

 

 

 

(0.00)

 

 

(0.00)

Percent free and reduced lunch

 

 

1.00

 

 

1.00

 

 

 

 

(0.01)

 

 

(0.00)

Charter school

 

 

 

0.39

 

 

0.56*

 

 

 

 

(0.25)

 

 

(0.17)

School size

 

 

 

1.00

 

 

1.00**

 

 

 

 

(0.00)

 

 

(0.00)

Urbanicity

 

 

 

1.25

 

 

0.94

 

 

 

 

(0.20)

 

 

(0.10)

School offers vocational programs

 

 

1.43

 

 

1.13

 

 

 

 

(0.31)

 

 

(0.21)

Number of full-time math teachers

 

 

0.99

 

 

1.03

 

 

 

 

(0.03)

 

 

(0.02)

Number of full-time science teachers

 

 

1.00

 

 

0.97

 

 

 

 

(0.03)

 

 

(0.02)

 

 

 

 

 

 

 

 

Students

 

16,190

16,190

16,190

16,190

16,190

16,190

Schools

 

750

750

750

750

750

750

 

 

 

 

 

 

 

 

Note: Estimated coefficients are reported in relative risk ratios. Robust standard errors adjusted for school clustering are in parentheses.

***p<.001, **p<.01, *p<.05



After introducing school-level factors, the results from Model 3 showed that females were still far less likely to enroll in applied STEM courses, though Hispanic students were no longer less likely to enroll applied STEM early in high school. The inclusion of school characteristics in the model also revealed that Asian students were significantly more likely to enroll in applied STEM early in high school. Having English as a primary language was no longer positively associated early applied STEM coursetaking after accounting for school factors. When examining the second set of results pertaining to late coursetaking, females were less likely to enroll in applied STEM courses later in high school as well. This prediction persisted even after including students’ academic history and attitude measures. The inclusion of these variables also revealed that student SES was significantly associated with reduced applied STEM enrollment later in high school (d = -0.03), indicating that higher-income students were less likely to enroll in applied STEM courses in later grades.


Table 6 also shows that the number of academic risk factors was negatively associated with applied STEM enrollment in courses later in high school (d = -0.04). There was also a significant association between enrollment in a vocational-technical program in high school and applied STEM coursetaking in a student’s later years of high school. Students with higher GPAs were significantly less likely to enroll in applied STEM courses later in high school (d = 0.06). Students who strongly believed that math was important (d = 0.05) and students with high academic aspirations (d = 0.03) were also more likely to take applied STEM later in high school.


These results pertaining to student-level characteristics were fairly consistent after incorporating school characteristics. Females remained far less likely to enroll in applied STEM courses toward at end of high school. Controlling for school, Black students became more likely to enroll in applied STEM later in high school. Consistent with the previous models in this study, students in vocational or technical programs and with higher GPAs were more likely to enroll later in applied STEM, and students who believed math is important and those with high academic aspirations were more likely to enroll later in high school. Consistent with the early versus never model, students with an IEP were less likely to enroll toward the end of high school.


At the school level, students who attended charter schools were less likely to take applied STEM courses later in high school. School size was statistically significant. However, the size of this effect (d = 0.00) indicates little practical significance.


DISCUSSION


The purpose of this study was to identify the individual and institutional factors that predict applied STEM coursetaking in high school. Whereas previous research has investigated the relationship between applied STEM coursetaking and a range of subsequent student outcomes, this was the first study to examine which students enroll in applied STEM to begin with and as such represents an important contribution to the work on high school coursetaking in general and applied STEM coursetaking in particular. There are number of important conclusions to be drawn from the results, all of which have direct and relevant policy implications.


First, based on our first research question (i.e., student-level predictors), the most striking result that emerged was that females were significantly less likely to enroll in applied STEM courses in high school relative to males. These results were consistent in every model and after disaggregating applied STEM enrollment by SRE and IT-specific courses. Furthermore, results from the analysis of applied STEM coursetaking by timing of enrollment suggested that the odds of enrollment among females were still lower than males regardless of timing. An important conclusion of the current study, therefore, is that contrary to increasing the number of females within the STEM pipeline, applied STEM courses appear to document another “coursetaking gap” between females and males within STEM fields. Notably, this finding suggests that applied STEM courses alone are not sufficient to reduce the gap in STEM coursetaking in high school and that policymakers and school leaders need to find additional ways of engaging females into these courses, particularly given the observed benefits of these courses on both high school and college-level outcomes (Gottfried et. al., 2014). Second, particular concern should focus around increasing female representation in IT courses, which evidence suggested were even more sparsely attended by females compared to males.


The second research question inquired into the institutional factors related to applied STEM coursetaking. While mean differences between schools that did and did not offer applied STEM courses were observed, the results from our analyses of applied STEM enrollment actually provided little support for the existence of substantial between-school variation in applied STEM coursetaking among the sample of students. In the majority of cases where the associations were significant, the coefficients associated with institutional factors were modest to negligible. The one exception was the association between a school’s decision to offer vocational programs and increased odds of IT applied STEM coursetaking.


The general lack of significant associations between institutional factors and applied STEM coursetaking was surprising to the researchers in light of prior research, which found characteristics such as school size (Monk & Haller, 1993), racial and socioeconomic composition of a school (Fowler & Walberg, 1991; Kelly, 2009), and urbanicity (Schiller & Muller, 2003) to be significantly related to STEM coursetaking among students. While more research is needed in this area, our results suggest that policies geared toward influencing coursetaking in high school, and applied STEM coursetaking in particular, would be more effective at targeting student-level factors rather than school-level characteristics.


Also important to applied STEM enrollment was timing (i.e., our third research question). Research has consistently demonstrated that the odds of success in STEM increase with early exposure and participation (Anderson & Dongbin, 1990; Fries-Britt, Younger, & Hall, 2010; Fullilove & Treisman, 1990). Alternatively, students become less likely to enroll in advanced traditional STEM coursework, succeed in their traditional STEM courses, and choose a STEM-related major if they delay STEM coursetaking in high school. Therefore, early participation in applied STEM coursework is ideal in order to maximize these documented benefits of applied STEM participation (Gottfried et. al., 2014). The results of our models suggested differences in timing. For instance, females and students with disabilities were much less likely to take applied STEM courses early in high school. This is concerning because it represents a potential lost opportunity to funnel these student groups into the applied STEM pipeline early.


Overall, our results have direct policy relevance. First, with regard to who takes applied STEM in high school, we found that female students were significantly less likely to enroll in applied STEM courses relative to their male counterparts. This finding was robust to model specification and after disaggregating the applied STEM coursetaking outcome into SRE and IT applied STEM courses. This finding suggests that rather than closing the long-documented STEM gender gap, applied STEM courses appear to be reinforcing it. If one promise of applied STEM coursework was to generate STEM interest among a broad spectrum of students, our results indicate a high degree of stratification insofar as gender is concerned. Interested legislators, policymakers, and school leaders may consider interventions that would seek to drum up enrollment in these courses among females, especially since prior research (Gottfried et. al., 2014) has documented a positive effect associated with applied STEM coursetaking on later STEM outcomes.


Second, policymakers should be concerned about stratification in students with disabilities. Students who had an IEP in the 10th grade were also consistently less likely to enroll in applied STEM courses. Again, this finding is unfortunate considering the documented benefit of these courses. In both cases, we are curious if the decreased odds of applied STEM enrollment among female and disabled students is evidence of tracking, such that females and disabled students are re-directed away from these courses. We must consider this finding and uncover, perhaps through qualitative investigation, if and why females and disabled students are being directed out of these courses in order to make adjustments to ensure student success.


The second main question that motivated the current study revolved around the timing of applied STEM coursetaking. We uncovered policy-relevant findings here as well. First, and broadly speaking, we found that just a few student characteristics were related to early/late applied STEM coursetaking, which we feel is an important finding itself. It appears that, with some exception, applied STEM coursetaking is dispersed throughout high school and is not sensitive to student demographics, family background, academic profile, or school factors. The few exceptions were Asian students, who were more likely to enroll in applied STEM early relative to White students. Black students, on the other hand, were statistically more likely to take the courses late in high school. The same was true for students in vocational education programs and students with higher cumulative GPAs. Early participation in applied STEM courses is ideal considering the linkage between these courses and subsequent advanced STEM coursetaking in 11th and 12th grade. However, we feel the “grander” narrative that policymakers should take away from this study is that the results related to who enrolls in applied STEM are more noteworthy and perhaps deserving of intervention than the question of when students take them.


Because applied STEM courses are a component of the broader CTE framework in high school, the results of the current study might also have implications for education policies related to CTE. Our finding that female students were significantly less likely to enroll in applied STEM courses suggests that efforts to expand the participation of women in STEM may not necessarily be achieved through simply offering of applied STEM courses from within the CTE curriculum. That is, expanding CTE course offerings may not necessarily reduce gender gaps in STEM, and in fact, these offerings may sustain current STEM gender gaps. Therefore, as CTE becomes a larger part of the high school curriculum, additional efforts in CTE are needed to counteract any stratification by gender.


LIMITATIONS


This study was limited in several ways. As the first study of its kind, this study was limited to general questions regarding who was and who was not enrolling in applied STEM coursetaking. Future studies should consult the findings presented here for more specific inquiries in applied STEM, such as mechanisms driving the gender gap in applied STEM coursetaking. This study helped to fill an important gap in the literature on applied STEM in high school, but it was unable to address many of the questions raised by the gaps discovered within the analysis.


As a quantitative analysis, this study was also limited to discerning broad and underlying statistical relationships. Additionally, the data utilized in this study were from the prior decade, and thus, the statistical relationships determined here might have different implications if the data were collected more recently. Therefore, an important next step is systematic qualitative inquiry within today’s high schools in order to fill the questions left by the current analysis. These lingering questions include:


1.

Why do females choose to not enroll in applied STEM coursework high school?

2.

What barriers, individual and institutional, exist that prevent increased participation among females and students with disabilities?

3.

Why do students chose to take applied STEM later rather than earlier in high school?

4.

Why do students choose SRE over IT-specific applied STEM courses and vice versa?


Future quantitative studies should examine the effect of SRE and IT participation on postsecondary outcomes including enrollment, major of study, and even workforce entry to see if, in fact, stratification in applied STEM coursework high school leads to stratification in the long term.


To maintain its competitiveness abroad, to inspire innovation, and to fill growing demand in the labor market, the nation must increase the number of STEM interested and qualified high school graduates. The challenge has been finding novel and effective ways of increasing the number and diversity of students moving through STEM pipeline. As a promising means of attracting and retaining students into STEM fields of study, applied STEM coursework needs increased empirical investigation and inquiry in order to maximize potential returns to the Perkins legislation. Previous research has established the efficacy of applied STEM coursework on a range of positive student outcomes. Yet missing from the literature was an examination into which and when students take applied STEM courses in high school. The conclusions of the current study suggest that these courses were serving some student groups and not others, and that student and school factors related differentially to the type of applied STEM courses. The time at which students are more or less likely to enroll in applied STEM courses during high school was dependent on these factors as well.


Notes


1. An alternative method of accounting for nested data structures is multilevel mixed effects modeling. However, as previous authors have illustrated (Arceneaux & Nickerson, 2009), mixed effects models and cluster adjusted robust standard errors yield the same estimates with the latter being more parsimonious and interpretable than the former. The authors of the current study found similar results after utilizing both mixed effects models and cluster adjusted models and therefore chose to only report coefficients produced with cluster adjusted models.


References


Adelman, C. (1999). Answers in the tool box: Academic intensity, attendance patterns, and bachelor’s degree attainment. Washington, DC: U.S. Department of Education.


Adelman, C. (2006). The toolbox revisited: Paths to degree completion from high school through college. Washington, DC: U.S. Department of Education.


Anderman, E. M. (2002). School effects on psychological outcomes during adolescence. Journal of Educational Psychology, 94(4), 795–809.


Anderson, E., & Dongbin, K. (1990). Increasing the success of minority students in science and technology. The unfinished agenda: Ensuring success for students of color. Washington, DC: American Council on Education.


Arceneaux, K., & Nickerson, D. (2009). Modeling certainty with clustered data: A comparison of methods. Political Analysis, 17(2), 177–90.


Astin, A. W., & Astin, H. S. (1992). Undergraduate science education: The impact of different college environments on the educational pipeline in the sciences [Final Report]. Los Angeles, CA: Higher Education Research Institute.


Baker, D., & Leary, R. (1995). Letting girls speak out about science. Journal of Research in Science Teaching, 32(1), 3–27. Retrieved from http://www.weizmann.ac.il/weizsites/blonder/files/2011/02/girls-speak-about-sci-Baker-95_Bat-Shahar.pdf


Benbow, C. P., & Arjmand, O. (1990). Predictors of high academic achievement in mathematics and science by mathematically talented students: A longitudinal study. Journal of Education Psychology, 82, 430–441.


Blanchett, W. J. (2006). Disproportionate representation of African American students in special education: Acknowledging the role of white privilege and racism. Educational Researcher, 35(6), 24–28.


Bozick, R., & Dalton, B. (2013). Balancing career and technical education with academic coursework: The consequences for mathematics achievement over the last two years of high school. Educational Evaluation and Policy Analysis, 35, 123–138


Callahan, R., Wilkinson, L., & Muller, C. (2010). Academic achievement and course taking among language minority youth in U.S. schools: Effects of ESL placement. Educational Evaluation and Policy Analysis, 32(1), 84–117.


Carbonaro, W., & Covay, E. (2010). School sector and student achievement in the era of standards based reforms. Sociology of Education, 83(2), 160–182.


Carl D. Perkins Career and Technical Education Act of 2006 [Web page]. (n.d.). Retrieved from http://www2.ed.gov/policy/sectech/leg/perkins/index.html


Chaney, B., Burgdorf, K., & Atash, N. (1997). Influencing achievement through high school graduation requirements. Educational Evaluation and Policy Analysis, 19(3), 229–244.


Chubin, D. E., & Ward, W. E. (2009). Building on the BEST principles and evidence: A framework on broadening participation. In M. K. Boyd & J. L. Wesermann (Eds.), Broadening participation in undergraduate research: Fostering excellence and enhancing the impact (pp. 13–30). Washington, DC: Council of Undergraduate Research.


Clark, M. L. (1986). Predictors of scientific majors for Black and White college students. Adolescence, 21, 205–213.


Colbeck, C. L., Cabrera, A. F., & Terenzini, P. T. (2001). Learning professional confidence: Linking teaching practices, students’ self-perceptions, and gender. Review of Higher Education, 24(2), 173–191.


Cullota, E. (1992). Scientists of the future: Jumping high hurdles. Science, 258, 1209–1213.


Davis-Kean, P. E. (2005). The influence of parent education and family income on child achievement: The indirect role of parental expectations and the home environment. Journal of Family Psychology, 19(2), 294–304.


Federman, M. (2007). State graduation requirements, high school course taking, and choosing a technical college major. B.E. Journal of Economic Analysis & Policy, 7, 1–32.


Fowler, W. J., Jr, & Walberg, H. J. (1991). School size, characteristics, and outcomes. Educational Evaluation and Policy Analysis, 13(2), 189–202.


Fries-Britt, S. L., Younger, T. K., & Hall, W. D. (2010). Lessons from high-achieving students of color in physics. New Directions for Institutional Research, 148, 75–83.


Fullilove, R. E., & Treisman, P. U. (1990). Mathematics achievement among African American undergraduates at the University of California, Berkeley: An evaluation of the mathematics workshop program. Journal of Negro Education, 59(3), 463–478.


Gamoran, A. (1987). The stratification of high school learning opportunities. Sociology of Education, 60, 135–155.


Gottfried, M. A. (2015). The Influence of Applied STEM Coursetaking on Advanced Math and Science Coursetaking. Journal of Educational Research, 108(5), 382–399.


Gottfried, M. A., Bozick, R., & Srinivasan, S. V. (2014). Beyond academic math: The role of applied STEM coursetaking in high school. Teachers College Record, 116(7), 1–35.


Grandy, J. (1998). Persistence in science of high-ability minority students: Results of a longitudinal study. Journal of Higher Education, 69(6), 589–620.


Grossman, P., & Stodolsky, S. S. (1995). Content as context: The role of school subjects in secondary school teaching. Educational Researcher, 24(8), 5–11, 23.


Hoffer, T. B., Rasinski, K. A., & Moore, W. (1995). Social background differences in high school mathematics and science coursetaking and achievement (NCES 95-206). Washington, DC: National Center for Education Statistics.


Hu, E. (2014, June 17). How Yahoo’s diversity numbers compare with Google’s. NPR: All Tech Considered. Retrieved from http://www.npr.org/blogs/alltechconsidered/2014/06/17/323040120/how-yahoos-diversity-numbers-compare-to-googles


Hunter, A. B., Laursen, S. L., & Seymour, E. (2007). Becoming a scientist: The role of undergraduate research in students’ cognitive, personal, and professional development. Science Education, 91, 36–74.


Kanno, Y., & Kangas, S. E. N. (2014). “I’m not going to be, like, for the AP”: English language learners’ limited access to advanced college-preparatory courses in high school. American Educational Research Journal, 51(5), 848–878.


Kelly, S. (2009). The Black-White gap in mathematics course taking. Sociology of Education, 82(1), 47–69.


King, J. E. (1996). The decision to go to college: Attitudes and experiences associated with college attendance among low-income students. New York, NY: College Board.


Laird, J., Alt, M., & Wu, J. (2009). STEM coursetaking among high school graduates, 1990–2005. Berkeley, CA: MPR Associates, Inc.


Lee, V. E., Chow-Hoy, T. K., Burkam, D. T., Geverdt, D., & Smerdon, B. A. (1998). Sector differences in high school coursetaking: A private school or catholic school effect? Sociology of Education, 71(4), 314–335.


Long, M. C., Conger, D., & Iatarola, P. (2012). Effects of high school course-taking on secondary and post-secondary success. American Educational Research Journal, 49, 285–322.


Maltese, A. V., & Tai, R. H. (2011). Pipeline persistence: Examining the association of educational experiences with earned degrees in STEM among U.S. students. Science Education, 95(5), 877–907.


Monk, D. H., & Haller, E. J. (1993). Predictors of high school academic course offerings: The role of school size. American Educational Research Journal, 30(1), 3–21.


National Center for Education Statistics. (1994). Curricular differentiation in public high schools (NCES 95-360). Washington, DC: Author.


National Science Board. (2007). A national action plan for addressing the critical needs of the U.S. science, technology, engineering, and mathematics education system. Arlington, VA: National Science Foundation.


National Science Board. (2014). Science and engineering indicators 2014. Arlington, VA: National Science Foundation.


National Science Foundation. (1999). Women, minorities, and persons with disabilities in science and engineering: 1998 (NSF 99-338). Arlington, VA: Author.


National Science Foundation. (2013). Women, minorities, and persons with disabilities in science and engineering: 2013 (NSF 13-304). Arlington, VA: Author.


Nord, C., Roey, S., Perkins, R., Lyons, M., Lemanski, N., Brown, J., & Schuknecht, J. (2011). The nation’s report card: America’s high school graduates (NCES 2011-462). Washington, DC: National Center for Education Statistics.


Oakes, J., Gamoran, A., & Page, R. N. (1992). Curriculum differentiation: Opportunities, outcomes, and meanings. In P. W. Jackson (Ed.), Handbook of research on curriculum (pp. 570–608). New York, NY: Macmillan.


Pajares, F., & Miller, D. M. (1994). Role of self-efficacy and self-concept beliefs in mathematical problem solving: A path analysis. Journal of Educational Psychology, 86(2), 193–203.


Pearson, W. J., Jr. (1986). Black American participation in American science: Winning some battles but losing the war. Journal of Educational Equity and Leadership, 6, 45–59.


Plank, S., DeLuca, S., & Estacion, A. (2008). High school dropout and the role of career and technical education. A survival analysis of surviving high school. Sociology of Education, 81, 345–370.


President’s Council of Advisors on Science and Technology (PCAST). (2012). Engage to excel: Producing one million additional college graduates with degrees in science, technology, engineering, and mathematics. Washington, DC: Author.


Reardon, S. F. (2011). The widening academic achievement gap between the rich and the poor: New evidence and possible explanations. In R. Murnane & G. Duncan (Eds.), Whither opportunity: Rising inequality and the uncertain life chances of low-income children (pp. 91–115). New York, NY: Russell Sage Foundation Press.


Schiller, K. S., & Muller, C. (2003). Raising the bar and equity? Effects of state high school graduation requirements and accountability policies on students; mathematics course taking. Educational Evaluation and Policy Analysis, 25(3), 299–318.


Schneider, B. (2003). Strategies for success: High school and beyond. In D. Ravitch (Ed.), Brookings papers on education policy (pp. 55–79). Washington, DC: Brookings Institution Press.


Schneider, B., Swanson, C. B., & Riegle-Crumb, C. (1998). Opportunities for learning: Course sequences and positional advantages. Social Psychology of Education, 2, 2553.


Schultz, P. W., Hernandez, P. R., Woodcock, A., Estrada, M., Chance, R. C., Aguilar, M., & Serpe, R. T. (2011). Patching the pipeline: Reducing educational disparities in the sciences through minority training programs. Educational Evaluation and Policy Analysis, 33, 95–114.


Sebring, P. A. (1987). Consequences of differential amounts of high school course work: Will the new graduation requirements help? Educational Evaluation and Policy Analysis, 9(3), 258–273.


Seymour, E., & Hewitt, N. M. (1997). Talking about leaving: Why undergraduates leave the sciences. Boulder, CO: Westview.


STEM Education Innovation Act of 2011, H.R. 3373, 112th Cong. (2011). Retrieved from https://www.congress.gov/bill/112th-congress/house-bill/3373/text


Stevens, T., Olivarez, A., Lan, W. Y., & Tallent-Runnels, M. K. (2004). Role of mathematics self-efficacy and motivation in mathematics performance across ethnicity. Journal of Educational Research, 97(4), 208–221.


Teitelbaum, P. (2003). The influence of high school graduation requirement policies in mathematics and science on student course-taking patterns and achievement. Educational Evaluation and Policy Analysis, 25(1), 31–57.


Thompson, J. J., & Windschitl, M. (2005). "Failing girls": Understanding connections among identity negotiation, personal relevance, and engagement in science learning from underachieving girls. Journal of Women and Minorities in Science and Engineering, 11, 1–26.


Tsui, L. (2007). Effective strategies to increase diversity in STEM fields: A review of the research literature. Journal of Negro Education, 76(4), 555–581.


Tyson, W., Lee, R., Borman, K. M., & Hanson, M. A., (2007). Science, technology, engineering, and mathematics (STEM) pathways: High school science and math coursework and postsecondary degree attainment. Journal of Education for Students Placed at Risk, 12, 243–270.


Wang, X. (2013a). Modeling entrance into STEM fields of study among students beginning at community colleges and four-year institutions. Research in Higher Education, 54, 664–692.


Wang, X. (2013b). Why students choose STEM majors: Motivation, high school learning, and postsecondary context of support. American Educational Research Journal, 50(5), 1081–1121.


Weinberger, C. J. (2004). Just ask! Why surveyed women did not pursue IT courses or careers. IEEE Technology and Society Magazine, 23(2), 28–35. doi:10.1109/MTAS.2004.1304399


White, K. R. (1982). The relation between socioeconomic status and academic achievement. Psychological Bulletin, 91(3), 461–481.


Wilson, F. (2003). Can compute, won’t compute: Women's participation in the culture of computing. New Technology, Work and Employment, 18(2), 127–142. doi:10.1111/1468-005X.00115


Wilson, R. (2000). Barriers to minority success in college science, mathematics, and engineering programs. In G. Campbell, R. Denes, & C. Morrison (Eds.), Access denied: Race, ethnicity, and the scientific enterprise (pp. 193–206). New York, NY: Oxford University Press.




Cite This Article as: Teachers College Record Volume 119 Number 10, 2017, p. 1-38
https://www.tcrecord.org ID Number: 21859, Date Accessed: 12/8/2021 11:49:59 AM

Purchase Reprint Rights for this article or review
 
Article Tools
Related Articles

Related Discussion
 
Post a Comment | Read All

About the Author
  • Cameron Sublett
    University of California Santa Barbara
    E-mail Author
    CAMERON SUBLETT is a doctoral candidate in the Gevirtz Graduate School of Education at the University of California Santa Barbara. His research relates to education policy, leadership, and research methods.
  • Michael Gottfried
    University of California Santa Barbara
    E-mail Author
    MICHAEL A. GOTTFRIED is an assistant professor in the Gevirtz Graduate School of Education at the University of California Santa Barbara. His research interests pertain to issues including school quality and effectiveness, classroom peer effects, and STEM. Recent articles include: “School Entry Age and Children’s Socio-Behavioral Skills: Evidence from a National Longitudinal Study of U.S. Kindergarteners” (Educational Evaluation and Policy Analysis) and “Classmates with Disabilities and Students’ Non-Cognitive Outcomes” (Educational Evaluation and Policy Analysis).
 
Member Center
In Print
This Month's Issue

Submit
EMAIL

Twitter

RSS