Many claims are made about the benefits to students’ learning and thinking of integrating education across science, technology, engineering, and mathematics (STEM). In this chapter we explore the evidence relevant to whether and how integrated approaches to STEM education support a range of outcomes within and across the disciplines. The full range of outcomes was described in Chapter 2. Here, we consider two main types of outcomes: those related to learning and achievement and those related to interest and identity.
As noted in Chapter 2, integrated STEM instruction is typically accomplished through the use of problem-, project-, or design-based tasks to engage students in addressing complex contexts that reflect real-world situations. For example, students might be invited to build an oven that is environmentally friendly or functional in settings where people do not have access to electricity. The students would use the engineering process to create a solar oven and in doing so investigate a wide range of STEM concepts such as the thermal properties of materials and how density affects a material’s performance as a thermal insulator. They might use mathematics for measuring, and for graphing and interpreting data, and even develop a mathematical model of device behavior to inform the process of design.
1 This chapter is based on the literature review overseen by David Heil and Associates and on commissioned papers by Angela Calabrese Barton, Michigan State University, Mary Gauvain, University of California, Riverside, and K. Ann Renninger, Swarthmore College.
Through iterative design cycles the students would engage in planning, creating, testing, and improving their inventions.
As illustrated in this example, we define integration to mean working in the context of complex phenomena or situations on tasks that require students to use knowledge and skills from multiple disciplines.
Research on the impact of integrated experiences on students’ achievement, disciplinary knowledge, problem-solving ability, and ability to make connections between domains is not extensive, and concerns related to both the design of studies and the reporting of results hamper the ability to make strong claims about the effectiveness of integrated approaches. Nonetheless, preliminary conclusions can be drawn from the well-designed studies. The findings suggest that integration can lead to improved conceptual learning in the disciplines but that the effects differ, depending on the nature of the integration, the outcomes measured, and the students’ prior knowledge and experience.
Most studies of STEM learning consider each discipline singly and do not measure students’ ability to make connections across disciplines or their proficiency with skills such as collaboration or general problem solving. In addition, learning is often assessed using standardized achievement tests, which may not effectively measure the full range of learning and reasoning outcomes supported by integrated experiences. Assessment instruments on integration are rare because theories and tests have generally focused on content area–specific concepts and procedures and because, as explained in Chapter 2, there is no widely accepted definition of integrative thinking.
Beyond these assessment challenges, there are fundamental conceptual difficulties as well. A major difficulty follows from the simple fact that disciplinary knowledge is structured—understanding disciplinary ideas depends on understanding how they fit with other, related ideas. Concepts make sense not as isolated facts but as elements of integrated bodies (or structures) of knowledge, and learning means developing or “building” those structures, often over extended spans of time.
Although education research has made some progress in understanding how to help students construct coherent bases of disciplinary knowledge, domain-general learning principles provide limited guidance. Instead, how
to support the development of disciplinary knowledge remains largely an empirical enterprise, in which cycles of research and trials with students and teachers gradually yield information about the most fruitful starting points, what conceptual resources students bring, and the kinds of instruction that are needed. Because integrated knowledge structures are developed gradually, it takes time—weeks, months, or years—for researchers to track their growth of student knowledge. Consequently, information about how to best help students learn with understanding is still limited to relatively few topics and has not yet resulted in widespread changes in educational practices.
Given these difficulties, it is not surprising that very little is known about how to organize curriculum and instruction so that emerging knowledge in different disciplines will mesh smoothly and at the right time to yield the kind of integration that supports coherent learning. Without very careful attention to developing coherent knowledge structures, the danger is that one or more of the “integrated” disciplines will receive short shrift in its development.
Integrating Mathematics and Science
The most well-studied integrated STEM education pairing is that of mathematics and science (e.g., Berlin and Lee 2003, 2005; Czerniak et al. 1999; Hurley 2001; Pang and Good 2000), but the number of studies that report the effects of integration on student learning in these two subjects separately is small. Moreover, the studies often are not explicit about the theory guiding how learning in the two subjects is coordinated and developed. Czerniak and colleagues (1999) noted in a review of the literature that there were few empirical studies of the integration of mathematics and science; many of the published articles promoted assumed benefits of integration or were theoretical in nature. Yet among the few empirical articles, Czerniak and colleagues saw a general trend toward a positive influence of integration on science and mathematics learning, although they pointed out that the descriptions of integration were so impoverished that it is difficult to make generalizations about the different approaches described.
Hurley (2001) conducted a meta-analysis of 31 studies that compared integrated mathematics and science instruction to a nonintegrated control group and reported mathematics and/or science achievement measures. She
found positive effects of integration on scores in both math (ES = .27)2 and science (ES = .37), which is consistent with other meta-analyses that report small to medium positive effects of integration (Hartzler 2000), although the effects varied both by subject and by the year the study was conducted. The lowest overall effect size for math achievement (ES = .07) was observed in the 10 most recent studies reviewed (1980s–1990s) and was lower than the effect for science achievement in all time periods.
Hurley also separated the achievement results by the level of integration (as described in the study reports) using the following categories:
• Sequenced : science and mathematics are planned and taught sequentially, with one preceding the other.
• Parallel : science and mathematics are planned and taught simultaneously through parallel concepts.
• Partial : science and mathematics are taught partially together and partially as separate disciplines in the same classes.
• Enhanced : either science or mathematics is the major discipline of instruction, with the other discipline apparent throughout the instruction.
• Total : science and mathematics are taught together in intended equality.
The effect size for mathematics achievement was positive and large when using a sequenced integration model (for mathematics ES = .85, for science ES = .34) but much lower for all other models of integration, ranging from -.11 for parallel instruction to .20 for total integration; parallel instruction also produced a negative effect size in science (ES = -.09). Both enhanced instruction (.66) and total integration (.96) produced large positive effect sizes for science.
Hurley also examined the 31 studies by grade level. At the elementary level there was only one study that examined mathematics. At the middle school level, two studies had outcomes for both science and mathematics. At the high school level, six studies had science outcomes and four mathe-
2 Effect size (ES) was calculated by subtracting the control group mean from the treatment group mean and dividing by the combined standard deviation of the treatment and control groups, following the recommendation of Hedges et al. (1989). Small effect sizes are around .3 or less, medium effect sizes around .5, and large effect sizes .8 or above. A negative effect size indicates that the traditional group outperformed the experimental group.
matics outcomes. At the college level, two studies had outcomes for science and three for mathematics.
At both the middle and high school levels the effect sizes for science were higher than those for mathematics, indicating that it may be difficult to enhance mathematics achievement by integrating the math into another disciplinary context. Similar results in an unpublished meta-analysis of math and science integration also suggest that there are fewer positive benefits of integration for mathematics outcomes compared to science outcomes (Hartzler 2000). One possible explanation is that attempts to integrate science ideas with ideas from mathematics may interrupt a sequential approach thought to help students investigate and elaborate the rich relations among mathematical concepts and procedures (Lehrer and Schauble 2000).
In contrast, Lehrer and Schauble (2006) found enhanced development of scientific concepts known to be challenging to students in the elementary grades when the students use mathematics as a resource for representing and modeling natural systems. These more carefully articulated studies of the use of mathematical systems as tools for learning about natural systems suggest that effect sizes may depend on details of the instructional approach that are obscured by simple characterizations of the temporal sequence of integration.
According to other studies, the nature of the mathematical tools and systems of representation available to students determine the depth and breadth of learning about core ideas in science because mathematical forms correspond to forms of understanding natural systems. For example, Sherin (2001) noted that university students’ models of force and motion were bound with symbolic equations. When students worked with the relations among quantities expressed by equations, they occasionally generated novel equations that prompted elaboration and reconsideration of core concepts. DiSessa (2000) posits that new forms of mathematical expression supported by computational media can make new ways of understanding science and mathematics accessible to larger numbers of students. For example, studies of student learning about complex systems indicate that agent-based descriptions—descriptions that represent phenomena as a large collection of interacting individuals—support learning about phenomena that are traditionally difficult to learn, such as electricity (Sengupta and Wilensky 2011), statistical mechanics (Wilensky 2003), and natural selection and population dynamics (Dickes and Sengupta 2012; Wilensky and Reisman 2006).
Collectively, these studies suggest that the integration of mathematics and science can be supported by engaging students in the invention and
revision of mathematical models of natural systems. A strong implication is that learning science entails learning to express the behavior of natural systems as mathematical models, making this form of integration not merely supportive of but indispensible to learning science.
Learning Science and Mathematics in the Context of Engineering Design3
Design-based approaches, a hallmark of engineering education, have received particular attention for their potential as a rich context for integrated STEM. The effect of engineering on learning in science and mathematics was examined in the NAE/NRC report Engineering in K–12 Education (2009). The authoring committee found preliminary but promising evidence of a positive impact of engineering on learning in science and mathematics. However, two published empirical studies of Project Lead the Way (PLTW), a major program in engineering education for middle and high schools, showed mixed results when state achievement test scores were the basis of comparison. In schools serving a high proportion of low-income families, all students showed significant overall gains in mathematics and science achievement scores between 8th and 10th grade regardless of their course enrollment. However, students enrolled in one or more PLTW engineering classes showed statistically less improvement in mathematics scores and a nonstatistical difference in science achievement scores over that period, compared with a control group (Tran and Nathan 2010a). In schools serving predominantly affluent families, PLTW students exhibited small gains in mathematics achievement but no improvement in science achievement compared with students in a control sample (Tran and Nathan 2010b).
The results of these two studies provide additional evidence that enhancing math achievement through integration with other disciplines is difficult to do, and it is likely that students need additional support in place to see how specific mathematics concepts and skills are integrated with the engineering activities in order to exhibit substantial gains in mathematics achievement. These studies also fail to show substantially larger gains for students participating in project-based engineering courses, underscoring the inconsistency in current research on integrated STEM instruction.
3 This section is based in part on a commissioned paper by Petrosino et al. (2008) for the NAE/NRC Committee on Engineering Education K–12.
Other research has demonstrated the effectiveness of learning science concepts through design in some but not all situations (Baumgartner and Reiser 1997; Fortus et al. 2004; Mehalik et al. 2005, 2008; Penner et al. 1997, 1998; Sadler et al. 2000). This approach can be effective if concepts are introduced when students engage with the design activity (Baumgartner and Reiser 1997; Fortus et al. 2004; Mehalik et al. 2007) or when design failure provokes conceptual change as students redesign an artifact to meet a goal (Lehrer et al. 2008). In addition, participant structures such as research groups (Lehrer et al. 2008) and design sharing sessions (pinup sessions) (Kolodner 2002) can provide conversational forums for clarifying and elaborating relations between designed artifacts and scientific concepts. These collective forms of activity are described more fully in Chapter 4.
Studies reveal that students may not spontaneously make connections between the devices being designed and the related scientific concepts (Crismond 2001; Kozma 2003; Nathan et al. 2013) and that they tend to focus on aesthetic or ergonomic aspects of design (Crismond 2001; Penner et al. 1998). Connections between the representations and notation systems used for design and for science need to be made explicit to students (Fortus et al. 2004; Nathan et al. 2013), or the material must be presented in such a way that students grasp that they can invent and revise systems of representation to understand how a natural or designed system works. Furthermore, the scientific knowledge gained through design may be highly contextualized, unless the activities are developed to support transfer of knowledge from one context to another, for example by using designs that highlight similar concepts across contexts (Fortus et al. 2004, 2005).
Design can elicit naïve conceptions from students. Explaining how a device functions presents an opportunity for the exploration of appropriate scientific concepts, especially in the case of redesign. However, without instructional support nothing inherent in the design process will necessarily challenge students’ ideas (Crismond 2001; Penner et al. 1997). Sadler and colleagues (2000) demonstrated the potential of redesign as an avenue to challenge naïve conceptions through rapid cycles of design activity that allow for many iterations to refine the student’s understanding (see also Penner et al. 1997, 1998). Redesign may be particularly useful for instruction because many elements of the designed object are already working, and only a few need to be focused upon and changed (Crismond and Adams, 2012).
When students engage in an engineering design task, they are likely to develop contextually dependent ideas about designing (e.g., “rules of thumb” and “how-to” knowledge). At least initially, without instructional support,
their design ideas are unlikely to connect to or be coherent with normative science ideas that might inform their designs.
Crismond (2001) showed that whereas experts recognize opportunities to connect with science ideas, nonexpert designers miss them. Even after lots of experience in given design contexts, individuals can reach an expert level but connect very different ideas to the context depending on their own conceptual frame. For example, aquarium hobbyists are likely to consider the practical challenges of designing an aquarium to support a specific range of aquatic organisms, whereas academic biologists may be more likely to focus on very general notions about how energy exchanges drive the system (Hmelo-Silver et al. 2007).
These findings highlight the need to carefully frame the instructional goals and settings to support students in making links to concepts in science. Box 3-1 provides an example of design as a context for integration.
A study of two elective digital electronics classes in two urban high schools examined instructional strategies that can support students in building connections across different representations of a phenomenon or situation when they are engaged in the complexities of design (Nathan et al. 2013). One classroom in each school was videotaped over 3 or 4 contiguous days; the participating students were in grades 10–12. In one school students participated in a unit on a voting booth security system; in the other they designed and built a digital circuit that tallied votes and passed resolutions only when a majority affirmed the resolution (with a tie favoring the vote of the president).
Analyses of the instructional moves made by the teachers and interactions between the teachers and students suggest that a key mechanism of integrated STEM education is cohesion of central concepts across the mathematics and science representations, engineering objects, design and construction activities, and social structures in the classroom. When cohesion was supported, students made useful connections across STEM disciplines, as was evident by their ability to move more fluidly among discipline-specific representations (e.g., Boolean algebraic expressions, schematized logic gates, and wiring of the digital circuits) and perform effective troubleshooting. Cohesion was effected through four pedagogical mechanisms:
1. identification of invariant relations and disciplinary concepts regardless of the surface features (Nathan et al. 2013);
2. coordination that “supports students’ reasoning and meaning making by constructing clear links across representations and activities” (Nathan et al. 2013, p. 110);
Example of Using Design as a Context for Integration
In a study with 6th graders, the activity of designing vessels that float was used to make learning from experimentation more relevant to the students (Schauble et al. 1995). After being given a design brief, students individually constructed vessels and added weight until the vessel sank. They then graphed their vessel with others that had similar carrying capacities. This was followed by further individual work in which students drew designs from various views and reflected on their previous design in a journal. Working in teams, students negotiated their designs by experimenting with various aspects of them. These efforts were supplemented by teacher and whole-class discussions of concepts such as buoyancy and relative density. By synthesizing the data from the experimentation, students could go on to plan their final design.
During this activity across several classrooms, a number of instructional challenges emerged. Although reflection is critical to learning, it was difficult to balance reflection activities with time spent on the more dynamic portions of the design process. It was also difficult to keep students focused on the design rather than on diversions while still valuing their background knowledge. And it was challenging to ensure that students not only remained focused on their goal of making the best vessel but also understood how various aspects of design could lead to improvements.
Analysis of interviews with the students before and after the activity revealed that they learned science through design and showed an improved understanding of experimentation. It also revealed that from an instructional perspective it was important to change only one variable at a time. This was true even when variables that would not affect the outcome of an experiment were altered. Instances in which teachers substituted or altered one irrelevant variable (such as using different types of weights that look different but are the same weight) led to confusion for the students, who were still developing an understanding of experimental procedure. Furthermore, teachers rarely discussed patterns in data, assuming that they were obvious to the students; this was demonstrated not to be the case. Finally, students were not spontaneously aware of the value of examining the unsuccessful vessels for attributes to be excluded; this useful skill can be nurtured by explicitly drawing attention to it (Schauble et al. 1995).
This example highlights the importance of framing and instructional support in design activity for integrated STEM learning.
3. forward projection to orient students to connections between current events or representations and future ideas and activities, which “facilitates planning, highlights pending importance, and prepares students for future learning opportunities” (Nathan et al. 2013, p. 110); and
4. backward projection to previously encountered ideas and events, which “prompts students to engage in reflection and emphasizes making connections between new and prior knowledge” (Nathan et al. 2013, p. 110).
Learning Mathematics in the Context of Technology
Although evidence reviewed thus far indicates that it may be difficult to support mathematics learning in integrated contexts, at least two studies suggest it can be done when explicit attention is given to mathematics learning.
Stone and colleagues (2008) studied mathematics-enhanced career and technical education (CTE) courses in high school that covered multiple occupational contexts—business and marketing, auto technology, health and information technology, and agriculture (but not engineering). CTE teachers were randomly assigned to teach courses either with enhanced mathematics or using traditional approaches. The teachers in the enhanced courses received guidance on how to structure their classes and additional professional development and were partnered with a mathematics teacher. They provided explicit opportunities for students to focus on the mathematics concepts, rather than just using math in the occupational context. Students in the two courses performed at similar levels in terms of technical skills, but those in the math-enhanced courses did better on measures of general math ability compared to students in the regular technical education courses.
A study of efforts to “infuse” mathematics in a 20-day middle school engineering/technology (ETE) course (referenced in Chapter 2) also showed promising results (Burghardt et al. 2010).4 Mathematics concepts and skills were introduced in the ETE curriculum at critical points through focused lessons to facilitate students’ ability to make connections between the disciplines. The mechanism used was a bedroom design activity, engaging students in the planning, design, and physical modeling of a “bedroom” that must meet specific cost and building requirements (e.g., the window
4 Infusion (the term used by the study authors) is similar to the enhanced approach to integration described by Hurley (2001).
area must be at least 20 percent of the floor area, the minimum room size is 120 square feet, the minimum closet size is 8 square feet). Eighth-grade students from 13 middle schools participated in the curriculum. Each teacher involved in the infusion curriculum was compared with a teacher in a “business as usual” technology class.
Students in both the infusion and comparison classrooms completed an assessment of mathematics concepts that were relevant to the bedroom design unit before and after instruction in the unit. Students in the infusion classes showed greater gains in scores from pre- to post-test than those in the control classes. It is important to point out, though, that the concepts on the mathematics test were closely aligned to the bedroom design unit and it is not clear from the study whether the students in the comparison classrooms were exposed to these concepts.
In a recent analysis of nationally representative data from the Education Longitudinal Study of 2002, Bozick and Dalton (2013) explored the effects of enrollment in CTE courses on mathematics achievement. Controlling for the characteristics of the students’ background and those of the school or district, the authors found that enrollment in occupational courses did not compromise mathematics achievement when such courses were taken instead of academic courses. When examined alone, engineering and technology courses—a subset of occupational courses that the authors say incorporate quantitative skills, problem solving, and logic—were unrelated to mathematics achievement.
Learning about Engineering and Technology
Very few studies have examined outcomes related to understanding engineering and technology, but pilot studies conducted as part of a large-scale curriculum intervention in New Jersey show some promising results.
Engineering Our Future New Jersey (EOFNJ) is a collaborative effort of Stevens Institute of Technology, the New Jersey Department of Education, the National Center for Technological Literacy (NCTL) at the Museum of Science, Boston, and others to bring exemplary technology and engineering curricula, such as Engineering is Elementary (EiE) and A World in Motion, to mainstream New Jersey K–12 education. The goal of EOFNJ is to ensure that within the next five years all K–12 students in New Jersey experience engineering curricula with a focus on innovation, as a required component of their elementary, middle, and high school education. Pilot studies were conducted at each school level.
At the elementary level, two modules from the EiE curriculum were implemented in 13 schools. One module focused on water quality, and students designed a water filter. The second focused on wind energy, and students designed a windmill that could lift a small weight. Results of tests administered before and after indicate that students improved in their ability to identify examples of technology and in their knowledge of water filters, filter materials, the science involved with the water filter module, and windmills and blade materials.
At the next level, 11 middle schools implemented a 4-week module from A World in Motion that involved designing a simple, mechanically propelled toy. Results on before-and-after tests of students’ conceptions of engineering and technology indicate that they improved their understanding of engineering.
At the high school level, 11 teachers from 10 high schools implemented 2 modules from the NCTL curriculum Engineering the Future: Designing the World of the 21st Century (NCTL 2005). One module was on fluid and thermal systems and included redesign of a boat to improve an aspect of the design. The second involved electrical and communication systems in which students worked with snap circuits. Results on pre-/posttests showed improvement in students’ understanding of fluid and thermal systems and of electrical circuits.
The studies reviewed indicate that the integration of STEM concepts in applied settings can yield increased conceptual learning in the disciplines but that there remain too many inconsistencies and gaps to effectively implement or assess integrated STEM programs.
For example, the positive impact on learning appears to differ for science and mathematics—it is less evident for mathematics outcomes. For both science and mathematics, the impact on learning and achievement varies depending on the approach to integration and the kinds of supports both embedded in the task and provided through instruction. Integration shows improved results on assessments of specific concepts related to the intervention, but not on general mathematics or science achievement tests like those administered by states.
Furthermore, the evidence presented above has several limitations that need to be considered when identifying directions for future research and
development. One of the most significant is the lack of a commonly agreed-upon definition for integration. Without one, it is difficult to consistently describe pedagogy or compare results across studies to develop a nuanced picture of whether and how different approaches to integration support learning. Likewise, without a common set of measures or criteria for documenting integrated learning, there is no clear basis on which to compare results. Moreover, there are few direct measures of integration as a construct or of outcomes that show how well students are able to make connections across disciplines. In the absence of standardized measures of integrated learning, researchers may use assessment instruments that are biased in favor of the particular intervention being studied, thus calling into question the validity of measures of STEM integration.
Finally, the research base includes a relatively small number of studies, with limited samples and often with potential problems with selection bias (e.g., only students who already do well in STEM or are interested in STEM participate). Studies span multiple age groups, include a variety of measures of learning or achievement, and effect sizes are generally small. In order to advance research on integrated STEM education, researchers need to consider a range of designs and methodological approaches. These are discussed in more detail in Chapter 6.
Fostering the development of students’ interest and identity in STEM is an important potential outcome of integrated STEM experiences. Interest and identity are thought to lead to continued engagement in STEM-related activities as reflected in course selection and choice of out-of-school activities, college major, and career path. In this section we review the evidence to determine whether and how integrated approaches support the development of interest and identity and lead to continued engagement in STEM fields. The committee found that out-of-school programs or experiences emphasized these outcomes, whereas school-based programs were more likely to focus on achievement outcomes. In both types of settings, however, direct measures of interest and identity were infrequent, although there was somewhat more attention to continued engagement (e.g., course taking or career aspirations).
In the following sections we explain how interest and identity have been defined by researchers and describe the ways they have been explored in
research. Next we examine evidence indicating whether integrated STEM experiences support the development of students’ interest and identity in STEM.
What Are Interest and Identity?
Interest develops over time, beginning with the triggering of attention and extending to voluntary reengagement, often characterized in terms of curiosity, persistence, and resourcefulness (Hidi and Renninger 2006; Renninger and Hidi 2011). Research findings clearly show that the presence of interest positively affects learner attention, goals, and levels of learning (see Hidi and Renninger 2006; Renninger and Hidi 2011) and that learners of all ages can be supported to develop interest (see Renninger 2010).
Interest is also related to other outcomes that can influence learning such as self-efficacy, an individual’s sense that s/he can be successful in a given domain. With more developed interest, the learner often has stronger feelings of self-efficacy and can better self-regulate behaviors to persevere on challenging tasks (Hidi and Ainley 2008; Sansone 2009).
Once an interest begins to develop, it can be sustained through instruction and/or out-of-school experiences, during which the learner often comes to identify with those who represent and pursue the interest professionally (Krapp 2007; Renninger 2009).
Identity generally refers to who one is or wants to be, as well as to how one is recognized by others—as a particular kind of person, with particular interests, expertise, and ways of being in particular social contexts, such as the classroom. Identity with respect to STEM has implications for how or why one might engage in classes, enroll in STEM courses, or use ideas and practices from STEM disciplines outside the classroom.
People have multiple shifting identities based on the diverse contexts and communities they encounter. As they move through time and space, they create, through their talk, actions, and interactions, different stories or narratives about who they are and want to be. These identities are always under negotiation, are contingent on the resources one has access to, and are shaped by a person’s social, cultural, and historical context, both in the moment and over time (Holland et al. 2001; Wortham 2006). These complexities are illustrated in the case study of a middle school student, Chantelle, presented in Box 3-2.
Case Study of Identity Development in STEM
Because identities are always in the making and socially negotiated, they are difficult to isolate or to name, raising questions about how to study them and what role they might play in helping an individual make sense of best practices for integrating learning in STEM.
Take, for example, the case of Chantelle (see Calabrese Barton et al. 2012). In the 6th grade, Chantelle, a soft-spoken African American girl growing up as the only daughter of a single mother, disappeared from view in science class. She infrequently volunteered in class and her average grades made her neither a concern nor an interest of her teacher. She had an avid desire to be a dancer when she grew up and pursued all the dance-related opportunities available at her arts-based public magnet school. She had little interest in science and mathematics; she’d never met an engineer and did not know what they did. And yet, at the very time when interest and motivation to pursue STEM drops precipitously, especially among girls, Chantelle’s interest—and scholastic achievement—in science increased. By 8th grade, she declared her interest in science and mathematics and stated her career goal to be a green architect, bringing together her love of the arts with science and engineering.
Why is it that Chantelle’s interest in science increased and her identity in science developed into one of a confident and competent student of STEM? There is clear evidence that one reason for the change was her participation in a technology-rich science and engineering club grounded in project- and place-based approaches. She had joined the club because her friends were involved, and initially her participation mirrored that in the science classroom: She arrived on time and finished her work, but she talked only with her small peer group and appeared more interested in watching YouTube videos of singers and dancers than in the science at hand.
However, through a series of events, Chantelle’s participation began to change. The turning point was her involvement in an after-school lightbulb audit at her school. Near the end of a unit on energy efficiency, Chantelle and two of her friends developed a project to determine how much energy and money their school was wasting by using incandescent lightbulbs. Saving money was important to them as budget cuts at their school loomed. They counted the number of incandescent bulbs in the school, documented their kilowatt-hour expenditure, and calculated how much money and CO2 emissions would be saved if they replaced those
bulbs with CFLs. They used a video recorder to document the process and to interview teachers and students on the topic. Chantelle’s two friends led the effort, organized the spreadsheet, and made the suggestions for where to go in the building; Chantelle pointed to the lightbulbs in each video shot.
Chantelle’s role changed, however, when the girls began to edit the video into a short documentary. She directed the editing, choreographed each new scene, and added text and graphics to pull out the message. As the group began to run out of time to finish the movie, Chantelle edited the film in her spare time. The project took about 6 weeks.
The lightbulb audit received such rave reviews by peers in the club that the girls were persuaded to seek permission to present their findings to their school’s student congress and school leaders. When the local electric company got word of the video from the school principal, it donated 1000 CFLs for the youth to distribute to their peers at school.
Furthermore, Chantelle asked to present the project to her science class, a level of active participation that stood in stark contrast to her previous everyday participation. Not only did she present the material, she engaged the class by asking her peers questions about why they should care about lightbulbs. She positioned herself both as the expert and as someone who cares about her fellow students and about the connections between science and their world. The following school year, when her 7th-grade class studied energy transformations, Chantelle eagerly volunteered in class discussion. She became deeply engaged in her science class across a variety of lessons and was described by her teacher as someone he wishes he could “clone.”
Chantelle’s story is illustrative of one of the more positive identity pathways Calabrese Barton and her collaborators have observed among middle school youth. Her experience shows that identity work is ongoing and cumulative and can be either facilitated or constrained by opportunities in the spaces where a student encounters science.
This case study also vividly illustrates the role of integrated STEM experiences and place- and project-based learning in fostering a productive science identity, which in turn enabled greater participation in the classroom, greater opportunities to learn, and the sense that a future in science is possible. Had the researchers only studied Chantelle’s achievement, or only studied her at a moment in time, they would have missed her developmental pathway.
Many studies of identity in STEM disciplines have been tied, in some form, to concerns about equity, in the context of underrepresentation and as a factor in pipeline losses. Studies have documented K–12 classroom and school practices that may contribute to certain students’ choices to disengage from STEM, such as African American girls (Calabrese Barton et al. 2012) who felt they had to choose friendships over extracurricular science in order to make academic success acceptable. Brown (2004, 2006) similarly observed that students “disidentified” with science to avoid cultural conflict.
Identity research may also help to explain why some instructional reforms succeed or fail even when they take into account gender, race, and language concerns (e.g., Carlone et al. 2011).
Evidence that Integrated STEM Supports Development of Interest and Identity
In addition to the case study illustrated in Box 3-2, evaluations of and research on integrated STEM programs provide preliminary evidence that such programs support the development of interest, identity, and continuation in STEM. As noted, however, measures of interest and of continuation in STEM are more common in studies of out-of-school programs, and in most cases the outcomes are measured without careful attention to the specific mechanisms that support the development of interest. Documentation of the development of identity is less common, and the few studies that have examined it in the context of integrated STEM are qualitative.
Studies and evaluations reviewed by the committee provide some evidence that integrated STEM programs can support the development and maintenance of interest in STEM. The programs or interventions considered were school-based projects and curriculum units, afterschool programs, and summer camps.
The study by Burghardt and colleagues (2010) of the infusion of mathematics into an ETE curriculum for middle school students (described in the previous section) documented outcomes related to interest. Students in the infusion curriculum and those in a comparison curriculum completed surveys of their attitudes toward mathematics and technology both before and after the intervention. Survey questions assessed the students’ interest
in mathematics and their perceptions of the importance of mathematics for technology and the relevance of mathematics. Comparison of the postsurvey responses of the two groups showed that students in the mathematics-infused curriculum reported that the subject was more important and interesting than did the students in the comparison group (controlling for responses on the presurvey). There were no significant differences between the groups on relevance. However, changes between the presurvey and postsurvey data revealed a decrease in reports from students in the infusion curriculum about the relevance of mathematics to their lives (Burghardt et al. 2010).
An unpublished study of a school-based engineering project for 6th and 7th graders similarly showed positive effects on students’ attitudes. The study included a comparison group of students who did not participate in the project, and students were surveyed both before and after the project. Students who participated in the project (designing a prosthetic arm) reported increased interest in engineering as a potential career as well as increased confidence in mathematics and science, although girls scored lower than boys in terms of their interest in engineering as a career and in their beliefs that they could become engineers (High et al. 2010).
Turning to out-of-school programs, in an unpublished evaluation of the Techbridge program, 367 girls (44 percent of the total number of girls) who had participated in the program from 2000 to 2007 completed surveys. Nearly 90 percent of the respondents reported that Techbridge had increased their interest in STEM; asked to identify what got them most interested in STEM, 72 percent cited hands-on projects and 16 percent said it was field trips (Ancheta 2008).
Evaluation of another enrichment program for high school youth, integrating engineering with biology concepts in a health care context using lecture and hands-on activities, also revealed positive effects on interest. On post-program surveys 50 percent of participants reported increased interest and more positive attitudes toward science and engineering (Monterastelli et al. 2011).
In a study of an all-girl summer camp with a STEM focus, the girls’ self-report of the likelihood of their pursuing a career in mathematics, science, or engineering rose from an average of 6.3 to 7.4 on a 10-point scale (Plotowski et al. 2008).
The results of other studies have been less clear. An unpublished evaluation of Project Exploration in Chicago, an out-of-school program for middle school–aged girls and minority students, summarized findings from surveys and interviews of participants during and after their participation. The
responses showed greater interest and confidence in science, but these were not assessed at the beginning of the program and no control group was used (Chi and Snow 2010).
Four studies of robotics programs showed somewhat mixed results. A published study of a 4-H robotics program revealed no significant differences in attitude between program participants and a control group of nonparticipants (Baker et al. 2008). But in an unpublished evaluation of FIRST robotics, an out-of-school program where students work in teams to design and build robots, students’ self-report on retrospective surveys (57 percent response rate) indicated higher interest in science and technology (89 percent of respondents) and in science and technology careers (69 percent of respondents) (Melchoir et al. 2005). In an evaluation of an out-of-school program that engages students in computer programming and engineering using robotic kits, 76 percent of students showed an improvement in their attitudes toward science and technology on pre and post surveys (Martin et al. 2011). Finally, in an evaluation of a robotics and geospatial program, about half of students reported more positive attitudes at the end of the program (Nugent et al. 2010).
Few of the studies considered by the committee examined identity. A commissioned paper on the topic reported that only three were conducted in the context of integrated STEM programs, and they were qualitative case studies.
The first study examined identity development in the context of science clubs for low-income middle school youth to pursue projects of their own choosing (Rahm 2008). The study showed that youth who were successful in the science clubs took on positions and roles that integrated their own histories and cultural backgrounds with science and that these roles were recognized by individuals who were more knowledgeable, such as the teachers running the clubs. The researcher posited that the formally acknowledged hybrid roles allowed the youth to try out ideas and ways of being that may have previously seemed out of reach or culturally incongruent (i.e., inconsistent with the culture of the students’ families or communities). She further suggested that the flexibility of the program, the value of doing a project both in and for the community, or the openness that allowed the students to define their own projects may all have been important elements in supporting development of a STEM-related identity (Rahm 2008).
A similar argument is made by Calabrese Barton and Tan (2010a) in the context of a technology-rich integrated science and engineering program focused on green energy. The researchers argue that as the youth in the program appropriated tools and resources through the program in ways that were culturally congruent, they developed roles as “community science experts”—they were seen as experts on matters in the community and in science, able to bring the two together. The study report describes the process by which the youth chose to investigate the urban heat island effect in their city and how they designed their study through scientific, engineering, and place-based concerns. They then wove these concerns together in a series of digital narratives to educate their community about their findings. Their role as experts was recognized and legitimized by teachers, scientists, and community members, and this acknowledgment was essential in supporting both their identity development and their learning (Calabrese Barton and Tan 2010a).
In a follow-up study, Calabrese Barton and Tan (2010b) analyzed the participants’ narratives describing their involvement in the green energy project over multiple years. These narratives revealed how the youths’ identities as community science experts and activists were carried from project to project and into new communities through public service announcements, scientific documentaries, and a new green roof for the building where the club was held, which the youth described as visible reminders of their hard work, what they know, and whom they influenced.
The findings from these three studies suggest that identity development may be supported by integrated experiences because such experiences support a range of ways of knowing, employ project- or problem-based approaches that allow youth to follow their interests, and can focus on problems relevant in local communities.
The findings about whether integrated STEM supports interest and continuation in STEM are mixed; there are promising indications, but the studies vary in quality. The measures of interest are typically not very sophisticated and do not take into account different phases of interest development. Also, many studies use before/after designs without any comparison groups. This is not a very powerful design for determining causal effects, so results are difficult to interpret.
Research on identity is at a very preliminary stage. The studies reviewed were qualitative and involved a very limited number of participants, but seem to indicate that open-endedness and links to students’ culture and community are important, as is the opportunity for students to be recognized as experts.
For both types of research, larger-scale studies and studies that incorporate a wider range of methods are needed.
Research on integrated STEM experiences suggests that they may be promising for supporting both learning in and across the STEM disciplines and the development of STEM-related interest and identity. The research base is limited, however, in terms of the design of the studies, the populations of students involved in them, the outcome measures used, and the extent to which research examines the mechanisms underlying learning in integrated STEM contexts.
In terms of learning and achievement, for integrated STEM education to be successful students need to be able to move back and forth between the acquisition of disciplinary knowledge and skill and their application to problems that call on competencies from multiple disciplines. Students need to be competent with discipline-specific representations and be able to translate between discipline-specific representations thereby exhibiting what some scholars refer to as “representational fluency.” Participation in shared practices, such as modeling in engineering, science, and mathematics, may support such fluency.
Integrated STEM experiences do appear to provide opportunities for students to productively engage in ways that can transform their identity with respect to STEM, and this effect may be particularly strong for populations that have historically struggled in STEM classes and are underrepresented in STEM higher education programs and professions.
The committee’s review of the research illuminated specific areas where further research is needed. For example, there is a need for more studies that measure or document students’ ability to make connections across disciplines or to demonstrate representational fluency. Few studies focus on the development of interest and identity in formal educational settings, and even fewer address their development in the context of integrated STEM, in either formal or informal settings. Finally, although there is a body of
research showing how integrated STEM experiences can be designed to foster connections between science and mathematics, there is a clear need to extend this research to more grade levels and to show more connections with engineering and technology.
More generally, the evidence base needs to be both deepened and broadened to support strong conclusions about the effectiveness of integrated STEM and an understanding of underlying mechanisms. Weaknesses in the research that need to be addressed include impoverished descriptions of interventions, lack of common terms and theories, and the need to use a wider range of methods with a better match of the questions to the designs. Current measures and descriptions of integration, as both a pedagogical method and a student outcome, lack reliability and validity.
All of these research-related issues are explored in greater depth in Chapter 6.
Ancheta, R. 2008. 2008 qualitative and quantitative longitudinal evaluation of Techbridge.
Barker, B.S., G. Nugent, N. Grandgenett, and A. Hampton. 2008 (Spring). Examining 4-H robotics in the learning of science, engineering, and technology topics and the related student attitudes. Journal of Youth Development: Bridging Research and Practice 2(3). Article 0803FA001.
Baumgartner, E., and B.J. Reiser. 1997. Inquiry through design: Situating and supporting inquiry through design projects in high school science classrooms. Annual Meeting of the National Association for Research in Science Teaching, Oak Brook, Illinois.
Berlin, D.F., and H. Lee. 2005. Integrating science and mathematics education: Historical analysis. School Science and Mathematics 105(1):15–24.
Berlin, D.F., and H. Lee. 2003. A bibliography of integrated science and mathematics teaching and learning literature, Vol. 2: 1991–2001. School Science and Mathematics Association Topics for Teachers Series Number 7. Columbus, OH: ERIC Clearinghouse for Science, Mathematics, and Environmental Education.
Bozick, R., and B. Dalton. 2013. Balancing career and technical education with academic coursework: The consequences for mathematics achievement in high school. Educational Evaluation and Policy Analysis 35(2):123–138.
Brown, B. 2004. Discursive identity: Assimilation into the culture of science and its implications for minority students. Journal of Research in Scence Teaching 41(8):810–834.
Brown, B. 2006. “It isn’t no slang that can be said about this stuff ”: Language, identity, and apropriating science discourse. Journal of Research in Scence Teaching 43(1):96–126.
Burghardt, M.D., D. Hecht, M. Russo, J. Lauckhardt, and M. Hacker. 2010. A study of mathematics infusion in middle school technology education classes. Journal of Technology Education 22(1):58–74.
Calabrese Barton, A., and E. Tan. 2010a. We be burnin’: Agency, identity and learning in a green energy program. Journal of the Learning Sciences 19(2):187–229.
Calabrese Barton, A., and E. Tan. 2010b. The new green roof: Activism, science and greening the community. Journal of Canadian Journal of Science, Mathematics and Technology Education 10(3): 207–222.
Calabrese Barton, A., H. Kang, E. Tan, T. O’Neill, and C. Brecklin. 2012. Crafting a future in science: Tracing middle school girls’ identity work over time and space. American Education Research Journal 50(1):37–75.
Carlone, H.B., J. Haun-Frank, and A. Webb. 2011. Assessing equity beyond knowledge- and skills based outcomes: A comparative ethnography of two fourth-grade reform-based science classrooms. Journal of Research in Science Teaching 48(5):459–485.
Chi, B., and J.Z. Snow. 2010. Project Exploration (PE): 10-year retrospective program evaluation summative report, submitted to the Noyce Foundation. Available at www.projectexploration.org/about-us/publications/10 years (retrieved January 21, 2014).
Crismond, D. 2001. Learning and using science ideas when doing investigate-and-redesign tasks: A study of naïve, novice and expert designers doing constrained and scaffolded design work. Journal of Research in Science Teaching 38(7):791–820. doi:10.1002/tea.1032.
Crismond, D., and R. Adams. 2012. The informed design teaching and learning matrix. Journal of Engineering Education 101(4):738–797.
Czerniak, C.M., W.B. Weber, A. Sandmann, and J. Ahern. 1999. A literature review of science and mathematics integration. School Science and Mathematics 99(8):421–430.
Dickes, A.C., and P. Sengupta. 2012. Learning natural selection in 4th grade with multi-agent-based computational models. Research in Science Education 43(3):921–953.
DiSessa, A. 2000. Changing minds. Cambridge, MA: MIT Press.
Fortus, D., R.C. Dershimer, J. Krajcik, R.W. Marx, and R. Mamlok-Naaman. 2005. Design-based science and real-world problem-solving. International Journal of Science Education 27(7):855–879.
Fortus, D., R.C. Dershimer, J. Krajcik, R.W. Marx, and R. Mamlok-Naaman. 2004. Design based science and student learning. Journal of Research in Science Teaching 41(10):1081.
Hartzler, D.S. 2000. A meta-analysis of studies conducted on integrated curriculum programs and their effects on student achievement. Ed.D. Dissertation, Indiana University, Bloomington.
Hedges, L.V., J.A. Shymansky, and G. Woodworth. 1989. A practical guide to modern methods of meta-analysis. Washington: National Science Teachers Association.
Hidi, S., and M. Ainley. 2008. Interest and self-regulation: Relationships between two variables that influence learning. In D. H. Schunk and B. J. Zimmerman (Eds.), Motivation and self-regulated learning: Theory, research, and application (pp. 77–109). Mahwah, NJ: Erlbaum.
Hidi, S., and K.A. Renninger. 2006. The four-phase model of interest development. Educational Psychologist 41(2):111–127.
High, K., J. Thomas, and A. Redmond. August, 2010. Expanding middle school science and math learning: Measuring the effect of multiple engineering projects. Paper presented at the P-12 Engineering and Design Education Research Summit, Seaside, OR.
Hmelo-Silver, C.E., S. Marathe, and L. Liu. 2007. Fish swim, rocks sit, lungs breathe: Expert-novice understanding of complex systems. Journal of the Learning Sciences 16(3):307–331. doi: 10.1080/10508400701413401.
Holland, D., W. Lachicotte Jr., D. Skinner, and C. Cain. 2001. Identity and agency in cultural worlds. Cambridge, MA: Harvard University Press.
Hurley, M.M. 2001. Reviewing integrated science and mathematics: The search for evidence and definitions from new perspectives. School Science and Mathematics 101(5):259–268.
Kolodner, J.L. 2002. Facilitating the learning of design practices: Lessons learned from an inquiry into science education. Available at http://scholar.lib.vt.edu/ejournals/JITE/v39n3/kolodner.html. (retrieved August 1, 2013).
Kozma, R. 2003. The material features of multiple representations and their cognitive and social affordances for science understanding. Learning and Instruction 13:205–226.
Krapp, A. 2007. An educational–psychological conceptualization of interest. International Journal of Educational and Vocational Guidance 7:5–21.
Lehrer, R., and L. Schauble. 2000. Modeling in mathematics and science. In R. Glaser (Ed.), Advances in instructional psychology (Vol. 5, pp. 101–159). Mahwah, NJ: Lawrence Erlbaum Associates.
Lehrer, R., and L. Schauble. 2006. Scientific thinking and science literacy: Supporting development in learning in contexts. In W. Damon, R. M. Lerner, K. A. Renninger and I. E. Sigel (Eds.), Handbook of child psychology, 6th ed. (Vol. 4). Hoboken, NJ: John Wiley and Sons.
Lehrer, R., L. Schauble, and D. Lucas. 2008. Supporting development of the epistemology of inquiry. Cognitive Development 24:512–529.
Martin, F.G., M. Scribner-MacLean, S. Christy, I. Rudnicki, R. Londhe, C. Manning, and I.F. Goodman. 2011. Reflections on iCODE: Using web technology and hands-on projects to engage urban youth in computer science and engineering. Auton Robot 30:265–280.
Mehalik, M., Y. Doppelt, and C.D. Schunn. 2005. Addressing performance and equity of a design-based, systems approach for teaching science in eighth grade. Annual Meeting of the American Educational Research Association, Montreal.
Mehalik, M.M., Y. Doppelt, and C.D. Schunn. 2008. Middle-school science through design based learning versus scripted inquiry: Better overall science concept learning and equity gap reduction. Journal of Engineering Education 97(1):71–85.
Melchoir, A., F. Cohen, T. Cutter, and T. Leavitt. 2005. More than robots: An evaluation of the FIRST Robotics competition: Participant and institutional impacts. Available at http://dev1.raiderrobotix.org/wp-content/uploads/2012/08/FRC_eval_finalrpt.pdf (retrieved January 14, 2014).
Monterastelli, T., T. Bayles, and J. Ross. 2011. High school outreach program: Attracting young ladies with “engineering in health care.” Proceedings of the American Society for Engineering Education, 2011 Annual Conference and Exposition, Vancouver.
NAE (National Academy of Engineering) and NRC (National Research Council). 2009. Engineering in K-12 education: Understanding the status and improving the prospects. L. Katehi, G. Pearson, and M. Feder (Eds.). Committee on K–12 Engineering Education. Washington: National Academies Press.
NCTL (National Center for Technological Literacy). 2005. Engineering the Future: Designing the World of the 21st Century. Boston, MA: Museum of Science.
Nathan, M.J., R. Srisurichan, C. Walkington, M. Wolfgram, C. Williams, and M.W. Alibali. 2013. Cohesion as a mechanism of STEM integration. Journal of Engineering Education. (Special issue on representation in engineering education.)
Nugent, G., B. Barker, N. Grandgennett, and V. Adamchuk. 2010. Impact of robotics and geospatial technology interventions on youth STEM learning and attitudes. Journal of Research in Technology Education 42(4):391–408.
Pang, J., and R. Good. 2000. A review of the integration of science and mathematics: Implications for further research. School Science and Mathematics 100(2):73–82.
Penner, D.E., N.D. Giles, R. Lehrer, and L. Schauble. 1997. Building functional models: Designing an elbow. Journal of Research in Science Teaching 34(2):1–20.
Penner, D.E., R. Lehrer, and L. Schauble. 1998. From physical models to biomechanics: A design-based modeling approach. Journal of the Learning Sciences 7:429–449.
Petrosino, A.J., V. Svihla, and S. Brophy. 2008. Engineering skills for understanding and improving K–12 engineering education in the United States. Presented at the National Academy of Engineering/National Research Council workshop on K–12 Engineering Education. Washington, DC.
Plotowski, P., M.A. Sheline, M. Dill, and J. Noble. 2008. Empowering girls: Measuring the impact of science technology and engineering preview summer camps (STEPS). Proceedings, American Society for Engineering Education. 2008 Annual Conference and Exposition, Pittsburgh.
Rahm, J. 2008. Urban youths’ hybrid identity projects in science practices at the margin: A look inside a school-museum-scientist partnership project and an afterschool science program. Cultural Studies of Science Education 3(1):97–121.
Renninger, K.A. 2009. Interest and identity development in instruction: An inductive model. Educational Psychologist 44(2):1–14.
Renninger, K.A. 2010. Working with and cultivating interest, self-efficacy, and self-regulation. In D. Preiss and R. Sternberg (Eds.), Innovations in educational psychology: Perspectives on learning, teaching and human development (pp. 107–138). New York: Springer.
Renninger, K.A., and S. Hidi. 2011. Revisiting the conceptualization, measurement, and generation of interest. Educational Psychologist 46(3):168–184.
Sadler, P.M., H.P. Coyle, and M. Schwartz. 2000. Engineering competitions in the middle school classroom: Key elements in developing effective design challenges. Journal of the Learning Sciences 9(3):299–327. doi:10.11207/S15327809JLS0903_3.
Sansone, C. 2009. What’s interest got to do with it?: Potential trade-offs in the self-regulation of motivation. In J. P. Forgas, R. Baumeister, and D. Tice (Eds.), Psychology of self-regulation: Cognitive, affective, and motivational processes (pp. 35–51). New York: Psychology Press.
Schauble, L., R. Glaser, R.A. Duschl, S. Schulze, and J. John. 1995. Students’ understanding of the objectives and procedures of experimentation in the science classroom. Journal of the Learning Sciences 4(2):131–166.
Sengupta, P., and U. Wilensky. 2011. Lowering the learning threshold: Multi-agent-based models and learning electricity. In M.S. Khine and M. Saleh (Eds.). Models and modeling in science education (pp. 141–171). Netherlands: Springer.
Sherin, B. 2001. How students understand physics equations. Cognition and Instruction 1:479–541.
Stone, J.R., C. Alfeld, and D. Pearson. 2008. Rigor and relevance: Enhancing high school students’ math skills through career and technical education. American Educational Research Journal 45(3):767–795.
Tran, N.A., and M.J. Nathan. 2010a. An investigation of the relationship between precollege engineering studies and student achievement in science and mathematics. Journal of Engineering Education 99(2):143–157.
Tran, N., and M.J. Nathan. 2010b. Effects of pre-college engineering studies on mathematics and science achievements for high school students. International Journal of Engineering Education 26(5):1049–1060. (Special issue on applications of engineering education research.)
Wilensky, U. 2003. Statistical mechanics for secondary school: The GasLab modeling toolkit. International Journal of Computers for Mathematical Learning 8(1):1–41.
Wilensky, U., and K. Reisman. 2006. Thinking like a wolf, a sheep, or a firefly: Learning biology through constructing and testing computational theories—an embodied modeling approach. Cognition and Instruction 24(2):171–209.
Wortham, S. 2006. Learning identity: The joint emergence of social identification and academic learning. New York: Cambridge University Press.