Exploring the Intersection of Science Education and 21st Century Skills: A Workshop Summary (2010)

This chapter and the next focus primarily on two questions addressed at the workshop:

1. What are the promising models or approaches for teaching these abilities in science education settings? What, if any, evidence is available about the effectiveness of those models?

2. What are the unique, domain-specific aspects of science that appear to support development of 21st century skills?

Four papers prepared for the workshop describe promising curriculum models. In order to ensure that the papers would address both of these questions and to increase uniformity across papers, the workshop planning committee provided a set of guiding questions to each author:

1. Curriculum goals: To what extent does the curriculum model target the five 21st century skills emerging from the May 2007 workshop or similar skills defined in the context of science education for instruction?

2. Alignment with learning research: To what extent does the curriculum treat 21st century skills and content knowledge as separate or intertwined? To what extent does the model reflect research on children and adolescents’ learning and development in science?

3. Assessment and evidence: Where has the model been implemented? Does the model incorporate assessment of 21st century skills? What evidence is available about development of one or more 21st

century skills among students and/or teachers engaged with the model?

1. Effectiveness and implications: What does the available evidence indicate about the impact of the model on development of 21st century skills among diverse groups of science learners? What does the evidence indicate about unique, domain-specific aspects of science that may support development of 21st century skills? Does the available evidence point to principles of instructional design for development of 21st century skills that may be applicable to other science curricula and/or teaching strategies?

This chapter summarizes the two papers presented on the first day of the workshop, and Chapter 5 summarizes the two papers presented on the second day. Chapter 8 synthesizes the evidence of intersections between science education and 21st century skills from all four papers.

Douglas Clark (Arizona State University) presented an overview of his team’s paper, which considers how engaging students in argumentation in online environments can help promote the development of 21st century skills (Clark et al., 2009). First, he explained that the team focused on argumentation, because inquiry and argumentation are at the heart of current efforts to help all students develop scientific literacy (American Association for the Advancement of Science, 1993; National Research Council, 1996). Scientific literacy, he said, involves understanding how knowledge is generated, justified, and evaluated by scientists and how to use such knowledge to engage in inquiry in ways that reflect the practices of the scientific community. Engaging students in argumentation can build this understanding and application of science processes.

Clark used the term “scientific argumentation” to describe a process in which students learn, whether in the domain of science or in another domain, to:

• develop, warrant, and communicate a persuasive argument in terms of the processes and criteria valued in science and

• construct, critique, and communicate sound and valid arguments in terms of the connections between and among the evidence and theoretical ideas.

He proposed that both of these facets of scientific argumentation are central 21st century skills. However, he cautioned, developing scientific argumentation can be quite challenging for students (e.g., Abell, Anderson,

and Chezem, 2000; Bell and Linn, 2000; Kuhn and Reiser, 2005; McNeill and Krajcik, 2008a; Ohlsson, 1992; Sandoval, 2003).

To address these challenges, education researchers have focused over the past 15 years on developing computer-enhanced environments to support students in constructing arguments and engaging with one another in argumentation. Clark briefly described four examples of such learning environments. Although only one of the environments—the Web-based Inquiry Science Environment (WISE) Seeded Discussions—was developed specifically to support argumentation in the domain of K-12 science, the other three develop argumentation that is similar to scientific argumentation in terms of argumentation structure, what counts as evidence, and goals.

The Computer-supported Argumentation Supported by Scripts-experimental Implementation System (CASSIS) environment has been used by undergraduate education psychology students in Germany to collaboratively solve problem cases related to attribution theory. CASSIS uses a similar approach to that developed by Guzdial and Turns (2000) in their CaMILE online learning environment, which has been used to support collaborative learning in science and in other domains. The Virtual Collaborative Research Institute (VCRI) learning environment has been used to support argumentation and collaborative learning among secondary students in the Netherlands in the domains of history, Dutch, and social studies. The Dialogical Reasoning Educational Web Tool (DREW) learning environment has been used to develop argumentation in the domain of science policy among secondary and undergraduate students in Finland.

WISE Seeded Discussions environment focuses on grouping students together with others who have expressed differing perspectives or stances. WISE Seeded Discussions first engages students in exploring the phenomenon to be discussed through probe-based labs and virtual simulations and then supports them in constructing an explanation for the phenomenon. In order to help students focus on the salient issues and articulate clear stances, they are given drop-down menus to construct their explanation from sentence fragments identified through research on students’ alternative conceptions. Once the students have submitted their explanations, they are organized into discussion groups with other students who have created explanations conceptually different from one another. Students participate in asynchronous online discussion of their explanations, in which they are encouraged to propose, support, critique, evaluate, and revise ideas. Finally, they reflect on how their ideas have changed through the discussion.

WISE Seeded Discussions have been implemented in a broad range of public middle and high school science classrooms, with data generally col-

lected on three to six classes of students for each study. Assessments initially focused on analyses of the structural quality of argumentation among students and later expanded to investigate the conceptual and grounds quality, in addition to the structural quality, of students’ argumentation (Clark and Sampson, 2005, 2007, 2008). The rubrics used in these assessments incorporate elements of complex communication/social skills and nonroutine problem-solving skills. More recent studies have used pretests and posttests to analyze gains in content knowledge.

CASSIS is designed to facilitate argumentation in asynchronous online discussions through collaboration scripts, which specify and sequence collaborative learning activities. Developed at the University of Munich, CASSIS engages groups of three students in analyzing problem cases using a specific theory. Usually, the group’s task is to first analyze three problem cases and then develop a joint solution for each case, collaboratively constructing an argument. An asynchronous, text-based discussion board is built into the environment so group members can communicate with each other as they work, and different collaboration scripts are implemented to promote and support productive collaboration among the students. For example, a script for the construction of single arguments consists of three text boxes that require students to input a claim, grounds, and qualifications as they construct the argument.

As an experimental learning environment, CASSIS has not yet been fully integrated into the core curriculum of an entire course. However, several hundred students enrolled in an educational psychology class at the University of Munich have participated in experimental sessions using CASSIS that take the place of a three-hour lecture on attribution theory—a theory of how people explain their successes and failures. Assessments of student learning through the use of CASSIS have focused on the quality of collaborative argumentation, based on analysis of transcripts of individual contributions to the online discussion. Some of this assessment has been automated.

VCRI is a groupware program developed in the Netherlands, designed to support collaborative learning on inquiry tasks and research projects, allowing students to communicate with each other, access information sources, and coauthor texts and essays. While working with VCRI, students share tools designed to support the collaborative inquiry process over the

course of approximately eight lessons. They start by investigating a topic, using a sources tool. Students are able to discuss the information found in these sources with other group members, using the synchronous chat tool. Students use the debate tool to help them examine and explore the arguments contained in these information sources. The debate tool enables the students to collaboratively create an argumentative map, a visual representation of the arguments in a single source or across sources. Once the argumentative maps are complete, students can transfer the lines of reasoning to the co-writer tool, a text processor that allows simultaneous editing by multiple users, to write a final report using the lines of reasoning identified and highlighted with the debate tool as a guide.

The DREW environment, developed at the University of Jyväkylä, Finland, consists of several different tools designed to support collaborative activities, including a chat environment, a collaborative writing tool, and an argument diagram tool. The argument diagram tool enables users, either individually or collaboratively through a shared screen from different workstations, to construct argument boxes that include claims, arguments, and counterarguments. The boxes are connected with each other by arrows indicating whether the content of the box supports or criticizes the content of the box to which the arrow points. A completed diagram depicts the argumentative structure of a text or discussion by indicating the main thesis of the materials and showing how the thesis is supported and criticized by illustrating other arguments and counterarguments and their interconnections.

Researchers studied use of the DREW environment among secondary students working in dyads engaged in chat discussions about current science policy issues (e.g., genetically modified organisms, nuclear power, and animal experimentation). They also studied secondary school students’ use of the DREW diagram tool to organize and structure arguments and counterarguments gathered from different Internet sources to be used in a joint essay-writing task. In a current study, university students are using DREW to analyze the content of scientific articles by creating argument diagrams using the DREW diagram tool.

All of the online learning environments, Clark said, are designed to implement and test instructional design principles developed through research on argumentation and the learning sciences (e.g., National Research Council, 2000). Although the environments did not specifically target the

five 21st century skills as learning goals, these skills are deeply intertwined with the development of scientific argumentation. The environments thus focus on skills, habits of mind, and communication processes that are central to both science and the development of 21st century skills.

The research team’s review of the research on student learning in these four environments indicates that they can support development of 21st century skills (Clark et al., 2009), Clark observed. He discussed examples related to each of the five skills, noting that they are not equally supported by the environments. Overall, the learning environments support the development of complex communication skills most strongly, followed by problem solving, self-monitoring, adaptability, and systems thinking (see Chapter 8 for a summary of the research).

Finally, Clark noted that the research has implications for the design of other science curricula and teaching strategies. He said that the four online learning environments are organized as scripts and activity structures that orchestrate and structure students’ interactions with each other and the environments. Current research on these learning environments focuses on the efficacy of various configurations and structures of the scripts. These scripts and activity structures could easily be incorporated into other online and offline curricula. He concluded that the research provides evidence that a broad range of collaborative learning skills can be supported by online environments for the development of scientific argumentation.

Rodger Bybee, former director of the Biological Sciences Curriculum Study (BSCS), opened his presentation by noting that, although education policy makers and practitioners have agreed for over a decade on the goal of developing scientific literacy among all students (American Association for the Advancement of Science, 1993; National Research Council, 1996), U.S. students’ scores on international comparative assessments of science show little evidence of progress toward this goal (Lemke et al., 2004).

The current concerns in the business community about the education and skills required for work are not new, Bybee went on, citing reports from the 1980s and 1990s (National Research Council, 1984; U.S. Department of Labor, 1991). More recently, he said, researchers have identified skills required for the workplace (Levy and Murnane, 2004; Murnane and Levy, 1996) that are somewhat similar to widely accepted goals for reform of science education (American Association for the Advancement of Science, 1993; National Research Council, 1996). Bybee noted that reports about workforce skill demands propose somewhat different definitions of the exact types or levels of skills required. For example, although most call for development of broad, transferable skills similar to those identified

in the May 2007 workshop (see Box 1-1), others focus on more specific knowledge and skills in the fields of science, technology, engineering, and mathematics.

Now, Bybee said, educators face the challenge of clarifying the skills that are needed for work and moving from broad statements of purpose to more specific discussions of educational practice. This challenge, in turn, led to the analysis in his paper of the intersection between 21st century skills and the 5E model (Bybee, 2009).

Bybee explained that the current 5E model has its origins in one of several science curriculum study groups established by the National Science Foundation in the 1960s after the Soviet Union succeeded in launching the Sputnik satellite. The Science Curriculum Improvement Study (Atkin and Karplus, 1962) developed the learning cycle model, including the three phases of explore, invent, and discover (Karplus and Thier, 1967).

During the late 1980s, BSCS convened a group of experts to review and revise the learning cycle model. The group added a new first stage—engage—with the goal of increasing students’ interest and motivation. And in response to teachers’ requests for an approach to assess student learning using the model, the group added a final stage—evaluate. The group also changed the name of the second stage in the learning cycle from invent to explain, and the third stage from discover to elaborate. These changes led to the current 5E model: (1) engage, (2) explore, (3) explain, (4) elaborate, and (5) evaluate.

In the 5E model, the curriculum is designed to allow students to explore scientific phenomena and their own ideas. Students are invited to explain their ideas, and explanations can also be provided by the teacher or a textbook or through the use of technology. The curriculum then helps students to clarify the key concepts targeted for instruction by engaging them in new situations in which they can elaborate and extend their learning. Finally, the curriculum invites both students and teachers to evaluate the learning that took place.

Bybee explained that the model reflects research on how students learn. Rather than simply requiring students to progress through a series of exercises that are sequenced to cover certain science topics within a certain number of days, the model aims to expose them to major concepts as they arise naturally in problem situations. The model calls for structuring activities in these problem situations so that students are able to explore, explain, extend, and evaluate their own progress. The model is based on findings from cognitive research that ideas are best introduced when students see a need or a reason for their use. Seeing relevant uses of the knowledge helps

students to derive meaning from the activities (National Research Council, 1999, p. 127).

Bybee said that, although it was developed in the 1980s, the 5E model also reflects more recent research, such as the research reviewed in the National Research Council study (2005) of high school science laboratories. That study concluded that laboratory experiences are more likely to support student science learning when they are integrated with other forms of instruction (National Research Council, 2005, p. 82). The 5E model represents one approach to integrating different forms of science instruction.

In 2006, BSCS was funded by the National Institutes of Health Office of Science Education to assemble and analyze all of the available research on the 5E instructional model. Bybee acknowledged that, although BSCS had been using the model in curriculum development and implementation for 20 years, the organization had conducted little research on its effectiveness. On the few occasions he did submit proposals to conduct research studies, he said, they were rejected because funders viewed research by the developers of the model as self-serving.

The review of the available research found that the model was implemented widely across the United States and in other countries. For example, a simple Google search of the term “BSCS 5E instructional model” returns about a quarter of a million citations. Most frequently, the term appears in:

• documents that frame larger pieces of work, such as curriculum frameworks, assessment guidelines, and course outlines;

• curriculum materials of various lengths and sizes; and

• adaptations for teacher professional development, informal education settings, and disciplines other than science.

At the same time, the team found that there was not very much research available on learning outcomes among students exposed to the 5E model. Nevertheless, some studies suggest that the instructional model is more effective than alternative approaches at helping students master science subject matter (e.g., Akar, 2005; Coulson, 2002). One of the key findings in the research relates to how faithfully teachers follow the model. Students whose teachers taught with medium or high levels of fidelity to it exhibited learning gains that were nearly double those of students whose teachers did not use the model or used it with low levels of fidelity (Coulson, 2002; Taylor, Van Scotter, and Coulson, 2007).

Bybee said that his review of the available research did not find any cases in which 21st century skills were specifically targeted as desired learning outcomes in curricula based on the 5E model. The model is aimed at developing students’ mastery of science subject matter, not at development of skills. With these caveats in mind, he then discussed the available evidence on development of each of the five skills (see Table 4-1). He found no evidence of development of adaptability, some evidence of development of complex communication/social skills (i.e., argumentation), and stronger evidence of development of three other skills: (1) nonroutine problem solving (i.e., scientific reasoning); (2) self-management/self-development (i.e., interest in science and science learning); and (3) systems thinking (i.e., mastery of content knowledge about complex scientific systems) (see Chapter 8).

Bybee concluded with four observations. First, he said that further clarification is needed about what each 21st century skill looks like and how to teach it, in order to provide concrete, explicit guidance to teachers. Second, it is important to identify specific learning outcomes associated with 21st century skills, because the preliminary definitions of these skills used as a framework for the workshop include a mixture of cognitive abilities, social skills, personal attitudes, motivational interests, and conceptual understanding. He suggested a need to tease out these different aspects in order to establish more specific learning outcomes. Third, he called for increased clarity in curriculum goals related to science education and 21st century skills. Finally, he suggested that, because the 5E model is a known quantity in the world of science education, it could serve as an excellent vehicle for developing 21st century skills, if adapted to focus on these skills.

Following both presentations, moderator Arthur Eisenkraft asked whether the environments for scientific argumentation described by Clark and the 5E model described by Bybee are intended for use in creating effective science lessons and if they reflect research on student learning. Bybee responded that the 5E model is based on Piaget’s theory of how children learn at different stages of cognitive development. He tells curriculum developers to focus first on how to evaluate student progress toward learning goals and then work backward from the goals to develop the curriculum.

Eisenkraft asked whether students and teachers are aware of the underlying learning models in curricula based on the 5E model or in the online learning environments. Clark responded that increased student awareness of their own internal learning process is an important part of the scaffolding built into the online environments. If the students are still able to engage in scientific argumentation when the scaffolding is faded, he said, this indicates that they are becoming aware of their own developing skills in argumentation. Bybee said that it is important to sort out some specific learning outcomes from the list of 21st century skills to use as a starting point and then design the curriculum to help students attain these outcomes. One important outcome might be development of durable skills that are transferable to new situations, such as skills in the control of variables developed by science students in a study by Klahr and Nigam (2004).

Marcia Linn asked how to take advantage of the embedded assessments, the tasks that students are actually doing in these more complicated curriculum materials, and Clark responded that a potential advantage of the online environments is that computers continually monitor student behavior and provide some real-time assessment. Eisenkraft asked about the possibilities for wide dissemination and implementation of the two instructional models. Clark responded that the online models for scientific argumentation highlighted in his presentation are not implemented on a wide scale, but other curricula using technology, like Investigating and Questioning our World through Science and Technology (IQWST, see Chapter 5), are being used in a fairly large number of school districts. In addition, he said, the Technology Enhanced Learning in Science Center, funded by the National Science Foundation, which engages students in Internet environments for science education, has grown to include 7 school districts and over 100 teachers.¹ He predicted that other states might soon join the center.

Clark also said that WISE provides an activity authoring system that can be used by others. In addition, Yael Kali has developed a database of design principles that are useful in constructing science learning environ-

ments (Kali, Linn, and Roseman, 2008). Teachers, he said, could take those design principles and create a new online science curriculum unit to meet their specific needs. However, Clark reported, more often a teacher will use these design principles or the WISE authoring system to customize existing curriculum materials to meet her or his local needs.

Eisenkraft asked what specific characteristics of the 5E model would promote 21st century skills, and Bybee responded that the primary goal of the model has been to increase student mastery of subject matter, including science concepts. He said that he has learned that, in order to reach any instructional goal, it is important to focus on that specific goal. Therefore, in order to develop 21st century skills through the 5E model, these skills would have to be explicitly targeted for instruction.