“We do STEM”: Unsettled conceptions of STEM education in middle school S.T.E.M. classrooms
This article features a Research into Practice Companion Article. Please click on the supporting information link below to access.
Abstract
STEM education has received attention both as a reform pedagogy and as a public means for economic growth and national security. Yet, despite this attention, clarity about what “counts” as STEM and how the individual disciplines relate to an integrated approach remains ambiguous. This article elicits the conceptions about what constitutes STEM education from middle grade teachers across the United States who teach a STEM discipline. Based on coded interviews and drawn conceptual models, participating teachers were more likely to describe essential aspects of integrated STEM found in the literature, but when asked to draw their conception of STEM education in their context, they often represented less ambitious conceptions. Furthermore, nontraditional subjects like engineering and technology were sometimes represented in subordinate ways to traditional subjects like science and mathematics. These findings have implications for preparing teachers to engage students in integrated STEM learning experiences as well as policy implications for what requirements must be in place to fund new STEM learning opportunities.
“We do STEM.” The first author—the director of a research and programmatic center focused on issues of teaching and learning in science, technology, engineering, and mathematics (STEM) education—first heard these words from a teacher of the STEM disciplines on a school visit in 2013. The anecdotal statement suggested that its speaker conceived of clearly defined boundaries and a common understanding of what comprises STEM education. Yet research on one area of STEM education—STEM-focused schools—suggests that existing criteria and rubrics are taken up and applied locally, resulting in a wide variety of perspectives on what actually defines STEM education (Eisenhart et al., 2015).
- How do middle school (MS) educators who teach science, technology, engineering, or mathematics classes, conceptualize the elements that define STEM education in their contexts?
- Within teachers’ conceptions of STEM education, what is the relationship among the science, technology, engineering, or mathematics disciplines?
This article contributes to the empirical work of STEM conceptualizations in at least three important ways. First, data are derived from classroom teachers responsible for elements of STEM teaching and learning within their schools. As these teachers have responsibility for implementing STEM instruction for children, identifying areas of consensus and confusion has important implications for future communication, policy, and professional development. Second, participants in this article represent diverse geographic backgrounds with representation from 10 states including both coasts of the United States, the South, and the Midwest. As geographical diversity may shape conceptualizations of STEM, representative data from multiple contexts may further illuminate the language and conceptions of STEM that exist among middle school teachers. Third, this article uses multiple modalities for probing teachers’ conceptions. Participating teachers constructed both drawn conceptual models as well as oral interviews about their conception of STEM education.
1 BACKGROUND LITERATURE AND FRAMING
1.1 STEM education conceptual frameworks
Definitions and conceptions of STEM range from the simple elaboration of the S.T.E.M. acronym (science, technology, engineering, and mathematics) to more integrated pedagogical visions (National Research Council, 2011a). Although research on, practice in, and funding for science, technology, engineering, and mathematics has occurred for decades, the use of the STEM acronym formally relating these disciplines is a more recent phenomenon. The National Science Foundation (NSF) began using the acronym “SMET” for these subjects in the 1990s, and for literary reasons adopted the now-commonplace term, STEM, in 2001 (Sanders, 2009). In the nearly 30 years since the first use of the SMET acronym, multiple reports have defined STEM in ways ranging from a simple listing of the disciplines (in this article we refer to the discrete disciplines as the acronym, “S.T.E.M.”) to more substantive interactions among the disciplines that result in particular pedagogical approaches (in this article we refer to this integrated notion in the non-acronym form, “STEM”) (Honey, Pearson, Schweingruber, & others, 2014; National Research Council, 2011a; President's Council of Advisors on Science & Technology, 2010).
Lack of clarity around the use of STEM has been explored at the postsecondary level. Breiner et al.’s (2012) study indicates that faculty within their institution of higher education conceptualize STEM most often as a siloed entity in which the acronym simply represents each discipline, but not necessarily in a connected way. Approximately 72% of the respondents described STEM as related to the disciplines of science, technology, engineering, and mathematics while nearly 28% did not have a definition for STEM. More importantly, among those surveyed who provided a relevant definition, nearly 60% simply defined the letters of the acronym. Surveyed participants were also asked how STEM affected their lives and some responded in ways that reflected an integrated perspective, but most maintained a siloed perspective, referencing individual disciplines. The authors conclude that even at a university where faculty are often involved in multiple, interdisciplinary STEM projects, there is no common operational conception of STEM and that definitions generally align with naming the parts of the acronym.
A siloed conceptualization referencing the individual disciplines may not be surprising at the postsecondary level as the distinction between departments, in terms of tenure, funding, and often identity, can be sharp. However, in K–12 contexts much more interaction exists among the disciplines. Perspectives exist in the literature and in practice that move beyond a siloed definition. The most prevalent perspective is an orientation toward “integrated STEM” (Honey et al., 2014). Most often, this conceptualization has been explored in the context of studies focused on STEM-focused schools (e.g., LaForce, Noble, King, Holt, & Century, 2014; Lynch et al., 2015; Scott, 2012). We analyze the goals, pedagogical approaches, and roles of the various stakeholders to frame conceptualizations of STEM that go beyond the acronym.
1.1.1 Goals
The goals of a STEM-focused school vary in terms of how well they address students’ learning experience. Organizationally, some schools target students who have shown elite academic achievement within the STEM disciplines (National Research Council, 2011a) so as to better prepare future scientists, mathematicians, and engineers. Other schools take a more inclusive approach and target populations that are currently underrepresented in the STEM fields (LaForce et al., 2014; Lynch et al., 2015). At the classroom level, research has shown that some schools focus on developing students’ competencies that can be either content-specific or general across the STEM disciplines. Content-specific goals focus on core ideas and practices within each of the STEM disciplines, while more general goals include the improvement of critical thinking and problem solving or “soft skills” such as collaboration and communication (Honey et al., 2014; National Research Council, 2011b).
1.1.2 Pedagogical approaches
Research from the STEM-schools literature highlights the centrality of the pedagogical approach. Siloed science or mathematics instruction is conventionally found in schools. That is, students have specific classes that address these subjects, and while the interaction of disciplines may occur, it is not a systematic routine and is at the discretion of the individual teacher. In contrast, an integrated approach places students in contexts where multiple disciplines are intended to be addressed consistently and authentically (Honey et al., 2014). It is not expected, however, that all four of the STEM disciplines within the acronym are present in the same learning experience for integration to occur. Rather, some definitions cite the needed presence of two or more disciplines to “count” as integration (Honey et al., 2014).
Within this multidisciplinary context, students are placed in constructive and interactive learning spaces generally focused on ill-defined problems. Research across all classrooms, not just STEM classrooms, has shown that creating an environment in which students construct knowledge and work socially to address problems has significant cognitive and achievement outcomes for students (Chi, 2009). In integrated STEM classrooms, this occurs most often through a problem-based learning format in which students focus not on the discrete solving of a problem that likely has one “right answer,” but rather on ill-defined problem spaces in which students must make decisions about what knowledge and skills are needed, how that knowledge and those skills will be learned, and in what ways they will be used to solve the problem (Bragg, 2005; Chin & Chia, 2006; Gallagher, 1995; Hmelo, Gotterer, & Bransford, 1997).
Further, studies on STEM schools have identified the importance of relevance and context that frame the learning experience. Whereas, traditional instruction may focus more on abstract problems (such as solving traditional physics problems from a textbook), more integrated approaches to STEM learning emphasize the importance of problems to which students can relate. The presence of a contextualized, relevant problem is also a necessary, but not sufficient element of effective problem-based learning (Goodnough & Cashion, 2006).
1.1.3 Stakeholders
Teaching and learning STEM in formal settings obviously involves relationships and interactions between teachers and students. However, in STEM settings, it is likely that even more burden falls on the ability of the teacher to invest time and resources into the planning and direction of instruction. A study of 25 inclusive STEM high schools noted that teachers in these contexts were responsible for creating all or part of the curriculum (LaForce et al., 2014). Teachers created special projects that placed students in authentic or real-world contexts, thus requiring a strong understanding of the content and practices of the STEM disciplines. Given the increased time commitment, this study also noted the important role that administrators play in helping support STEM instruction. Recent research suggests that STEM schools are highly collaborative environments that benefit from distributed leadership models and a re-envisioning of the relationship between teachers, students, and knowledge (Spillane, Lynch, & Ford, 2016).
STEM teaching and learning is not bound to traditional class time and thus, occurs at times outside of the classroom. Learning, especially at the high school level, can occur through internships, apprenticeships, and informal engagements with industry (Peters-Burton, Lynch, Behrend, & Means, 2014). At younger grades, engaging the community includes interactions with STEM professionals in ways that can shape a students’ STEM identity and learning. A central element of STEM-schools places students in contexts of service learning as well as career and technical training both within and outside the walls of their school (LaForce et al., 2014). Even at the elementary level where internships and college credit are developmentally inappropriate, the community can play a significant role in the definition of the school's STEM focus. For example, in South Carolina (2016) the state provides a special STEM designation for STEM-focused schools that requires a variety of elements including “community stakeholders [that] assembl[e] quarterly to discuss challenges and solutions to…improving a STEM program” (p. 16). Thus, schools desiring a more intentional STEM focus must seek support from those in their community. This work can take many forms, including using members of industry to help articulate design challenges for students, to provide images of the possible for STEM careers, or to help provide supplies to which many schools would not normally have access.
1.2 Mental models and conceptions of phenomena
Understanding how individuals conceptualize particular phenomena has consequences for behavior. As Johnson-Laird (1983) notes, “Human beings understand the world by constructing models of it in their minds” (p. 10). These models are central to how individuals understand phenomena or environments and influence ways in which they then behave. Mental models require linguistic and symbolic representations that show how concepts are related to other concepts within the mental model.
Furthermore, research on mental models suggests that the cognitive structure can be modeled explicitly using symbols; this model is a sample of the full representation of an idea or system of ideas that rests within someone's cognitive structure (Carley & Palmquist, 1992). Viewing this explicit representation of a mental model is best accomplished when “verbal statements are represented as visual structures in which concepts and the relationships between those concepts are specified” (Carley & Palmquist, 1992, p. 603). While drawn models may be limited in their ability to convey abstract ideas and interviews may over-represent “buzzwords” or language used without clarity of meaning, triangulating the data with both data forms likely provides a more robust representation of an individual's conception of the topic at hand.
Drawing on the STEM education literature as well as the literature for identifying individuals’ mental conceptions of phenomena, we elicited the thinking of a variety of teachers’ ideas about STEM education. The following sections detail the sampling and methods used to capture and code the irregular data produced by drawn mental models and the procedures put in place to capture the range of perspectives of middle school STEM teachers and how they view interactions among the disciplines within their classrooms.
2 METHODS
2.1 Sample
Data for this article was collected from 64 middle school teachers from nineteen different schools. All teachers had been selected as a participant in a 3-year STEM teaching fellowship for middle grade teachers. Fellows had to participate as a school-based team of at least three and no more than five teachers whom all had responsibility for teaching at least one class that is part of the STEM acronym or a specific integrated STEM course. Thirty-five of the teachers identified primarily as a science teacher, 20 teachers identified primarily as math teachers, 5 teachers identified as technology or computer science teachers, and 4 participants identified as non-traditional teachers such as teachers of an elective STEM or engineering course.
The fellowship from which participants were sampled focused on three main areas of professional development: (a) STEM integration, (b) core instructional practice such as the facilitation of discussion and assessing student thinking, and (c) the development of a school-based STEM impact plan. Two themes of equity and leadership spanned each of the three main areas. It is important to note that although the nature of the fellowship may have influenced the type of participants in this sample, all data were collected prior to any intervention or any detailed distribution of the fellowship's goals or outcomes.
The sample consisted of 44 female teachers and 21 male teachers spread across 2 rural-based, 6 urban-based, and 11 suburban-based schools. Approximately 40% of teacher participants taught in schools with at least 50% of the student body comprised of underrepresented minorities. In terms of experience, only one teacher was coded by the program's criteria as a novice (1–2 years of experience), 44 were considered early career (3–10 years), 14 were considered mid-career (11–20 years), and 5 teachers were considered late career (21+ years). Nearly half of the sample possessed master’s degrees and about 40% had an undergraduate or graduate degree in one of the S.T.E.M. disciplines or S.T.E.M.-focused education, such as a degree in education with a science certification.
2.2 Data sources
2.2.1 Drawn models
Representations of mental models have been studied in the context of cognitive psychology, symbolic interactions, and schema theory (Carley & Palmquist, 1992; Johnson-Laird, 1983). Based on previous studies using this methodology, participants provided drawn representations of their conceptions of STEM education on an 8.5 × 11 sheet of paper. Cognitive or mental maps attempt to represent the cognitive structures in memory, that is, the organization and relationship of concepts in memory (Shavelson, 1972). We chose drawn mental models as one data source to identify participants’ conceptions of STEM education because it provided an opportunity for participants to show not only concepts, but also the relationship among those concepts. While difficult to code because of their idiosyncratic nature, the drawn models provided an artifact for comparing what can be implicit among STEM teachers. As Carley and Palmquist (1992) state, “Cognitive mapping is perhaps the most useful means of exploring the nature of shared knowledge in social groups” (p. 605). Drawn models were collected and scanned for coding.
2.2.2 Participant interviews
- The term “STEM” has been used in schools in many different ways. In your own words:
- How would you explain to a parent the term “STEM Education” in your school setting?
- Let's go into some more detail based on your description.
- What kinds of teaching practices fit into your STEM education definition?
- How about attitudes—both teachers’ and students’?
- What kinds of (classroom, school, community) activities do you expect to see in STEM education?
- What kinds of resources for STEM education might a school be expected to provide?
- Based on your school's or your understanding of STEM education, please identify one or two activities, teaching practices, or attitudes that some teachers or schools may call “STEM,” but you think fall outside of the intended definition.
Due to audio difficulties, one interview was incomplete and not used for the interview analysis described below.
2.3 Procedures
2.3.1 Drawn models
Participants constructed representations of the mental models of their conceptualization of STEM on three occasions: (a) As part of the first activity within their fellowship, prior to any instruction or professional development about STEM; (b) As part of the final activity of their first Summer Institute, 2 weeks and approximately 65 hours of professional development after the initial model; and (c) As part of the final activity of their second Summer Institute, exactly 1 year after their second model and after approximately 100 more hours of STEM professional development. As this article focuses solely on initial conceptions of STEM prior to any intervention, only the initial models were used for analysis.
In each replication, participants were given a blank sheet of paper and the prompt: Develop a visual representation to show how you think about STEM Education in your school context. Please include: (a) the important elements; (b) how these elements are related; and (c) who is involved in STEM education. Participants had approximately 10 minutes to create their drawn conceptual models. At the end of the 10 minutes they were asked to discuss their models with an assigned partner before submitting their models for analysis.
2.3.2 Interviews
An external research firm, using the questions noted above, conducted semistructured interviews. Similar to the drawn models, interviews were conducted annually with the first interview taking place after participants were selected to the fellowship, but prior to any professional development or specific dissemination of information about the program. Interviews lasted approximately 25 minutes, were conducted over the phone, audiotaped, and transcribed for analysis. As this article is focused on initial conceptions prior to any intervention, data from only the first interview was used for analysis.
2.4 Data analysis
Confirmatory and exploratory data analyses were conducted for both the drawn models and the interview transcripts (Carley & Palmquist, 1992). For research question 1, confirmatory analysis included deductive coding from existing frameworks of STEM education described above (e.g., Honey et al., 2014; LaForce, Noble, King, Holt, & Century, 2014; Scott, 2012). Drawn models and interviews were coded according to three Level 1 codes: (a) the goals of STEM education; (b) the elements of enactment/pedagogical approach; and (c) the stakeholders involved in STEM education (excluded from analysis in this article due to confounding data) (Table 1). Level 2 codes were developed from the literature and several emergent codes were added inductively after coding a small sample of the data. Analysis occurred at the level of the Level 2 codes.
| Level 1 Code | Level 2/Level 3 codes | Code description |
|---|---|---|
| Goals | ||
| STEM Workforce Readiness | The goals include preparing students for college and careers in STEM | |
| STEM Learning | The goals focus on STEM learning of fundamental concepts and practices | |
| Cognitive Competencies | The goals focus on improving cognitive skills | |
| Inter- and Intrapersonal Skills | The goals focus on affective measures not related to cognition. | |
| Interest and Engagement | The goals include increased interest and engagement in STEM disciplines | |
| Service/Stewardship | The goals are toward helping others immediately and in the future | |
| Elements of Enactment | ||
| Approach | The pedagogical approach | |
| No Reference | ||
| Disc rete | Focused on traditional methods in which teachers teach discrete pieces of information or practices. | |
| Hands-on | Focused on activity in which students interact with the material world, but it is inconclusive whether the activity is focused on an ill-defined or rich problem. | |
| Problem-based | Focused on problem-based methods in which a scenario or problem drives instruction. This can include elements of traditional instruction, but even these elements contribute to a broader problem or design challenge. | |
| Disciplinary Integration | The level of integration among STEM and other disciplines | |
| No Reference | ||
| Siloed | The disciplines are mentioned, but they are not connected or they are seen as siloes. | |
| Integrated | Reference is made to the interaction of the disciplines in authentic ways with at least two disciplines interacting. | |
| Contextualization | The relationship between the STEM learning and the local and personal contexts of students | |
| No Reference | ||
| Abstract | STEM education is focused on doing problems or exercises from books that are disconnected from actual contexts in the students' lives | |
| Relevant/Contextualized | STEM education draws on elements of the students' lives and context |
For the goals and stakeholder domains, each model was scored in a binary way—whether the model included the sub-category or not. A single instance of a subcategory received a code of present for that category. For the second domain focused on enactment and pedagogical approach, models were coded along a continuum. The three subcategories—instructional approach, level of integration, and level of contextualization—were coded at either end of the continuum for each of these categories or marked as “no reference” for the given category. During the analysis, it became clear that coding representations of hands-on material activity, or phrases related to “hands-on learning” required too much inference from the coders to determine whether the instances were embedded in problem-based contexts or not. Therefore, a Level 2 code was added to the instructional approach continuum with “hands-on” falling between the neutral-oriented “no reference” code and the “problem-based” code.
For research question 2, three main themes were coded based on the relationship of the disciplines present in the models and interviews: (a) Subordinate/superordinate relationship of disciplines; (b) Co-relational disciplines; and (c) Missing disciplines. For each discipline within the STEM acronym, the models and interviews were coded to indicate whether across the body of the data for a particular data mode, a discipline was subordinate to other disciplines. For instance, a one-way arrow from a mathematics representation in the drawn model to engineering suggests that mathematics is used in the service of doing engineering. In contrast, co-relational disciplinary representation existed when double arrows connected two disciplines or other symbols indicating equivalence. Data were coded for the possibility of each pairwise relationship among the STEM disciplines. Finally, if across the entirety of either the drawn model or the interview a reference to one of the discrete STEM disciplines was missing, it was coded as a missing discipline. Statistical comparisons were made among frequencies for each distribution, disaggregated by interview and drawn model, using the omnibus Kruskal–Wallis non-parametric test. When significant, Dunn's post hoc test was used to determine which elements were responsible for the significant difference.
Coding of the drawn models occurred blind to the demographics of the Fellow or the cohort. All authors coded the same, initial 25% of the drawn models and interviews, reconciling all disagreements before coding the remaining data in pairs. Two authors coded the remaining data individually and then reconciled their codes to 100% agreement.
3 RESULTS
3.1 Research question 1: Teachers’ conceptualizations of STEM education in their contexts
- Goals of STEM education; and
- Pedagogical strategies to teaching STEM.
3.1.1 Goals
Due to the abstract nature of the “goals” construct, teachers’ conceptions of the purpose of STEM education in their contexts were captured more frequently in their interview responses than in their drawn mental models (Figure 1). Sixty-four of the participants articulated a goal for STEM education while 25 of those same participants visually represented a goal. The most frequently held views about the goals of STEM education from the participating teachers focused on affective measures. Specifically, teachers focused most on promoting interest and engagement in STEM, as well as developing character traits, such as “grit” (Duckworth, Peterson, Matthews, & Kelly, 2007) that lead to on-going effort in light of difficult problems.
This perspective was clearest in the interview responses of teachers, approximately two-thirds of whom discussed the value of making STEM interesting and engaging for students. For example, a middle school science teacher from the Southwest named Kelly, described the importance of connecting with students through STEM saying, “They really enjoy [STEM lessons], because they're more interesting, there's less direct instruction, there's more of them doing things and trying to figure things out on their own and coming up with their own solutions. So it really helps with interest in the learning process.” Similarly, Sinead, teaching at a school on the East Coast serving a student body of 100% underrepresented students in STEM, also focused on the overlapping goals of fostering interest and interpersonal skills within a problem-based context, saying, “I think it's all about excitement. It's about excitement. It's about growing. It's about learning. It's about struggling. I think all of those—I think all of those attitudes combined it's—are great. I think it's about the will to want to solve a problem.” Interestingly, the pattern of increased mentions of a goal in an interview aligned with much lower, but relatively increased representations of the construct in the drawn mental models (Figure 1) does not hold for the interest/engagement construct. Given the data used in this article, it is unclear why this code follows a different pattern than the other codes.
3.1.2 Pedagogy
I would I guess I kind of look at it there's the silo approach and then there's that integration. You know you can have just that silo of I'm just teaching this subject. We're done with this subject. I move on. Teach the next subject, which you know I'm 42, so I go back to my childhood and I think that's a lot of how my education was is that everything was compartmentalized … But this is the big world problem solving skills. How do we present a problem and not just pull the science, but pull in your math, pull in your writing skills, pull in all these things to integrate if together to move students from that silo to here's your big picture. Here's the real world. Here's life.
Fewer teachers, 30, mentioned the contextualization axis. Of these teachers, all mentioned the importance of a relevant or contextualized learning environment and none talked about an abstract environment or tasks. And finally, on the integration axis, 47 of the teachers addressed the importance of integration of STEM disciplines in their school contexts with none making references to siloed disciplines.
Evidence from the drawn conceptual models suggests that slightly more variance might exist in teachers’ conceptions than represented by the interviews. While there was only a single instance in the interviews of a teacher mentioning an element of STEM that does not reflect the high-quality enactments found at the positive end of each axis, drawn representations in the models illustrated multiple occurrences of discrete, siloed, or abstract STEM elements in these teachers’ classrooms. Specifically, four teachers’ representations of the instructional axis only had symbols of discrete problems, nine teachers drew models in which the S.T.E.M. disciplines were clearly siloed, and one teacher drew an image that reflected an abstract context.
While it is assumed that fewer elements—both positive and negative—would be represented in the drawn models compared to the interviews because of the higher degree of difficulty in representing certain abstract ideas, the distribution of the different categories raises interesting questions about what teachers really believe is encompassed in STEM education. For example, 6 teachers specifically illustrated problem-based learning instruction and 12 teachers drew more general hands-on learning activities, but 4 teachers also drew very discrete types of problems, such as “3x+2 = 8.” Similarly, while 18 teachers represented integrated learning environments in their models, nine teachers also drew images in which the disciplines were siloed. The two panels in Figure 2 represent the conceptions of two different participants. The first model, drawn by Sheila, represents a more traditional classroom environment in which disciplines are seen as separate elements of learning (Panel A). In contrast, Rachel's model reflects more contemporary conceptions of STEM (Panel B). As you can see in Panel A, Sheila explicitly drew dividers among the disciplines and represented learning of those disciplines in discrete ways. In contrast, Rachel's conception made explicit connections among the STEM disciplines as part of a problem-based environment.
Plotting the coded models for each teacher on a three-dimensional coordinate system shows interesting patterns about teachers’ holistic conceptions of STEM (Figure 3). Each point on Figure 3 represents the intersection of a teacher’s three axes: instructional approach, disciplinary integration, and contextualization for both the interviews (Panel A) and drawn models (Panel B). The size of the asterisk is proportional to the number of teachers whose plot fell at the same intersection of the three axes.
In Panel A, the plot shows that teachers’ interviews reflect conceptions of STEM that are fairly aligned with current writing about STEM pedagogical approaches. The top corner—the intersection of the three axes at the positive end of each axis—represents nine teachers who spoke about problem-based, integrated, and contextualized STEM learning environments within the same interview. The most frequent plots on the system were the 22 teachers who spoke about integrated and contextualized pedagogical approaches that included hands-on, but not necessarily problem-based instruction.
Panel B shows a different distribution of plots based on teachers’ drawn models. The most frequent plot exists at the point of no representation for each of the three axes. That is, 20 teachers did not represent any element, either positive or negative, for any of the three categories. The difficulty of representing some elements visually may have led to the high incidence of this point, but the remaining plots indicate a variety of perspectives, including those that are more discrete, abstract, and siloed. While only one teacher illustrated all three of these “negative” components, nine teachers did have conceptions in which abstraction was represented. Highly prominent, though, was the array of drawn models that showed problem-based and hands-on instructional approaches that coincided with different levels of contextualization and integration.
3.2 Research question 2: Relationship among the S.T.E.M. disciplines
Teachers’ conceptions of the relationships among science, technology, engineering, and mathematics showed, on average, three main themes: (a) explicit connections between S.T.E.M. disciplines; (b) individual S.T.E.M. disciplines in service of other S.T.E.M. disciplines; and (c) missing S.T.E.M disciplines.
3.2.1 Connections between disciplines
Science and mathematics were the most represented S.T.E.M. discipline in both drawn models and interviews (Figure 4). Connections between these two disciplines were also the most common with 21 connections in the interviews—nearly a third of the sample—and 10 in the drawn models. The remaining pairwise connections ranged between 5 and 10 instances for both interviews and drawn models with the exception of technology and engineering that had only two connections described in the interviews. Despite the descriptive differences, the Kruskal–Wallis test indicated no statistical difference between the connections between the disciplines for either media, p>.05.
3.2.2 In-service of S.T.E.M. disciplines
Individual disciplines were rarely discussed or drawn in ways that put them in service to other disciplines. Engineering had only one instance in the drawn models and no instances in the interviews of existing at the service of another discipline. Similarly, science and mathematics were at the service of other disciplines no more than three times in either the interviews or drawn models. Although not extensive, technology was seen as in-service to other S.T.E.M. disciplines more frequently than the other disciplines. Nine teachers represented technology in the service of other disciplines in the interview. One teacher talked about technology not as a discipline in its own right, but as an educational tool saying, “Obviously we need some sort of technology component, although I don't think technology should drive the curriculum. I think it's just an aid in teaching.” Another teacher talked about technology as a tool in mathematics classrooms that could be used to “crunch numbers” and “give students a break from having to do calculations.” Four teachers also represented technology in service of other disciplines on their mental models. Although qualitative differences were found among the disciplines, statistical tests indicated no statistical differences, p > 0.05.
3.2.3 Missing disciplines
Science and mathematics were least likely to be missing from teachers’ drawn models or interviews with both disciplines not mentioned in one interview, respectively, and 5 and 4 drawn models, respectively. An interesting pattern emerged with engineering. While technology was also missing from only one interview, seven interviews were missing references to engineering. Drawn models were more likely to be missing both technology and engineering with 10 and 11 models missing these disciplines, respectively. Similar to the previous two constructs, no statistical differences were found among the disciplines in either the interviews or drawn mental models, p > 0.05.
4 DISCUSSION
STEM education has received enormous investments from both public and private entities over the past two decades. For instance, just one spending source—the federal budget—reported $380 million spent on STEM education programs not including any funding directed to the Department of Education, the NSF, NASA, or NOAA in 2016 (American Institute for Physics, 2015). Despite significant funding and attention, ambiguity remains about what constitutes the goals and component elements of STEM education. This article drew on a geographically diverse sample of teachers from STEM disciplines who often define what “counts” as STEM in their own contexts each day. As S.T.E.M. and STEM education appear to have implications for the development of a literate citizenry, students’ “21st century skills,” and problem-solving competencies in ill-defined contexts (Honey et al., 2014), better understanding the range of conceptions of STEM may lead to more coherence among local and regional organizations like the school or district.
Different stakeholders have made various claims about the goals of STEM education. Perhaps it is unsurprising, then, that conceptions of STEM also vary among practitioners. At the national level, two common narratives have cited our nation's security and preparing a future workforce to spur economic growth and stability as central STEM education goals (National Research Council, 2011a; President's Council of Advisors on Science & Technology, 2012). However, our sample of middle grades teachers did not echo these narratives. Rather then target long-term STEM persistence and career outcomes for adults, conceptions elicited in interviews and drawn mental models focused more proximally on student-centered goals such as generating long-term interest in the STEM disciplines and creating non-discipline-specific skills such as critical thinking, problem solving, and positive social interaction. While these affective and interpersonal goals do not contradict economic or national security goals—in fact, they may be central to long-term STEM workforce initiatives (President's Council of Advisors on Science & Technology, 2010)—more student-centered narratives may influence how professional development, curricular materials, and communication among stakeholders are framed. Previous work in science classrooms has shown that mismatched goals for reform-based science teaching among teachers, administrators, and districts can result in ineffective professional development, stunted professional growth, and at times confusion and animosity among levels of school-wide organization (Allen & Penuel, 2015). Ensuring coherence among the stakeholders about the goals and nature of STEM education are essential to its long-term success.
To promote coherence, professional organizations, professional development providers, and policymakers must distinguish between programs or goals for siloed S.T.E.M education and integrated STEM education. It is not that the professional development of either siloed or integrated STEM should be the only focus of professional development, but as professional development on science or math instruction have been ubiquitous for decades, more attention may need paid to providing a coherent vision of the elements of integrated STEM. As no one professional organization can claim authority over shaping the narrative for integrated STEM, members from all communities need to create a framework that is coherent and useful for teachers. Comparing our data from interviews and drawn conceptual models to the research literature on defining integrated STEM approaches, our sample reflects many aspects of a problem-based, integrated, and contextualized approach rooted within the core disciplinary ideas and practices of the STEM disciplines, but whether this conception is held by S.T.E.M. teachers more broadly and influential stakeholders from the administration or higher education is not clear.
Our article focused on middle grades S.T.E.M. teachers as their daily work depends on their conception of the STEM construct, but previously reported data by Breiner et al. (2012) suggests a possible tension between these conceptions and those surveyed among university faculty. Postsecondary faculty interviews showed much higher frequencies of conceptualizing STEM as defined by the individual parts of the S.T.E.M. acronym. Reflecting on these results they concluded that “an operational definition would be at best difficult to achieve” and that “while [stakeholders] working on a certain STEM initiative [may need] to have a common conceptualization, caution should be paid as many initiatives across the nation are probably too varied to be placed into too narrow a framework” (p. 10). In contrast, in working with a cross-country sample of in-service teachers of STEM disciplines, we believe that an operationalized definition is necessary for both siloed S.T.E.M. and integrated STEM if we seek to improve the quality of STEM education in the United States. This does not suggest that a shared definition need be so limiting that it does not attend to local needs. Highlighting and relating the problem-based, integrated, and contextualized nature of integrated STEM may allow teachers to better share a collective understanding and focus on-going initiatives and professional development.
In comparing our samples’ responses to existing literature on STEM education, our sample consistently conceived of STEM education as instruction that goes beyond traditional didactic interactions with students. As reported above, in both the interviews and the drawn models, action and collaboration were central to teachers’ STEM definitions. Nearly 100% of the interviewed teachers mentioned either problem-based learning or hands-on engagement. The data suggest that students’ active participation with the natural and built world was central to their conceptions. The drawn models reported a larger proportion of teachers illustrating a hands-on approach that was not necessarily problem-based, but viewed together, the problem-based and hands-on approach greatly outnumbered conceptions of discrete approaches that often occur in math and science classrooms (Weiss, Pasley, Smith, Banilower, & Heck, 2003).
This research on teachers’ conceptions of STEM is only a first step toward improving the educational experiences of young people. For example, future studies focused on teachers’ conceptions will be important to elicit teachers’ thinking about the differences between hands-on and problem-based approaches. The former approach, while likely more favorable than a traditional focus on discrete, and often decontextualized problems, can result in manipulation of the material world without the intended cognitively challenging “minds-on” work. Furthermore, empirical work must be done that investigates the connections, or lack thereof, between conceptions and practice. The theory of planned behavior (Ajzen, 1985), building on the theory of reasoned action (Fishbein & Ajzen, 1975), posits that behaviors are influenced not only by conceptions and attitudes, but also by subjective norms of the environment and by one's perceived behavioral control. For example, teachers may talk in-depth about constructivist approaches and believe them to be the most effective form of instruction, but given an environment that emphasizes a didactic approach or given inexperience and the lack of tools to support a constructivist approach, the behavior that occurs in instructional practice may contradict the teacher's conceptions and beliefs. Future studies need to explore and compare individual teachers’ conceptions and observations of what is carried out in the STEM classroom.
Finally, more research must be done about teachers’ conceptions of the relationship among the S.T.E.M. disciplines. From an integrated perspective, the disciplines of science, technology, engineering, and mathematics should all contribute unique core ideas and disciplinary practices that can be used to solve complex problems. Thus, when asked to think generally about STEM education, the disciplines should be on equal footing—pairwise connections among the disciplines in drawn models should be evenly distributed, one-way relationships should be absent, and each of the four main disciplines should be present in a teacher's mental model. While most teachers in our sample represented all four disciplines and models showed few instances of subordinate disciplines, the data raised questions about teachers’ conceptions of engineering and technology in relation to traditional subjects like science and mathematics.
Engineering was the discipline most commonly missing from teachers’ responses. This finding is perhaps unsurprising as engineering is not a traditional school subject and even though its status has been elevated with inclusion in the Next Generation Science Standards (NGSS Lead States, 2013), even its inclusion places engineering within the context of science classes and as such, may often be perceived by teachers and students as simply an application of science and not an autonomous discipline. However, several authors have proposed engineering as the central hub for defining STEM integration (Roehrig, Moore, Wang, & Park, 2012). Roehrig and colleagues write that engineering and the engineering design cycle naturally provide a problem-based, relevant, and contextualized opportunity for students to apply science, math, and technology principles. Whether engineering becomes the hub for STEM integration learning environments or not, work must be done with teachers to understand the role that engineering plays, what constitutes engineering practice, and how it can be addressed in K–12 classrooms.
Similarly, technology did not hold the same conceptual place as science and mathematics. While technology was present in most models and interviews, more often than the other subjects, participating teachers viewed technology as a tool to help the other disciplines. Their drawn models showed technology as necessary to carry out other disciplinary work, but the reverse relationships were not illustrated. While computer science standards exist (e.g., CSTA Standards Task Force, 2016), more work must be done to help teachers integrate conceptual frameworks for technology that includes, but extends beyond computer science and coding. A greater understanding of the core ideas and disciplinary practices of technology may help teachers envision its role in the STEM classroom.
4.1 Limitations
The use of drawn models requires a level of inference by coders that can limit the claims made about the data. When in doubt, we abstained from coding a particular valence to a representation, thus providing a more conservative analysis of the conceptions. Although interviews were also conducted, the interviews probed slightly different questions in order to understand the participants’ own conception of STEM and how their school conceptualized STEM at the time of data collection. Future studies using this methodology would benefit from using think-aloud protocols that directly follow the creation of drawn models.
Furthermore, the sample includes 64 participants and although they represent teachers from diverse school environments and geography, they are not fully representative of all teachers from the STEM disciplines. Rather, they represent those teachers most engaged in their profession as the application and selection process for the fellowship to which they applied was quite extensive. Yet, the results of this article are still important as no other studies have been found that include samples of teachers from such a broad geographic and school reach on this topic.
4.2 Conclusions
STEM education is a popular term that drives not only instruction, but also school-wide reform. Yet despite its role in educational dialogue, conceptions about the goals and component parts of STEM education vary greatly. Findings from this article suggest that many teachers of the S.T.E.M. disciplines recognize important instructional factors central to a student-focused approach, but coherence and common understanding are still absent from the national discussion across K-20, with policy makers, and the general public. More work must be done to clarify what constitutes STEM education without limiting the various approaches to achieving its goals. And once a more collective understanding has been developed, the most difficult work of preparing a highly qualified field of administrators and teachers to enact an ambitious vision of problem-based, integrated, and contextualized instruction must be taken up collaboratively by researchers, professional development providers, and practitioners.
