Citekey: @Woodruff1997

Woodruff, E., & Meyer, K. (1997). Explanations from intra- and inter-group discourse: Students building knowledge in the science classroom. Research in Science Education, 27(1), 25–39. doi:10.1007/BF02463030


An earlier paper on science education following the “progressive discourse” approach.


We have conducted three separate studies in science classrooms that implemented this approach within a science unit. Across the studies we have varied the grade level of the students (grade five, grade seven), varied the topic of inquiry (images and shadows, density), and documented “on-line” inquiry discourse within a computer software environment called CSILE, standing for Computer Supported Intentional Learning Environment (Scardamalia, Bereiter, McLean, Swallow, & Woodruff, 1989) that networked small groups of students in the grade five class. Our findings show inquiry discourse, a methodologyof explaining, to be a promising approach to learning within both the topics and for students in both grade levels. (p. 3)

The purpose of this article is to integrate the findings of these three studies from the prospect that explanation and intraand inter-group discourse can be integral parts of school science. Our discussion will focus on three perspectives: inquiry and the generation of explanations in small groups of students, and the validation of explanations in large groups of students; the evolution of student explanations within inquiry discourse; and finally, pedago~cal interpretations and considerations. Thus this article discusses the process of explaining through inquiry discourse by small and large groups of students, how those explanations may change during the process and the implications of this approach as interpreted by ourselves and the collaborating teacher. (p. 3)

Explanations in Science (p. 3)

In this case, what triggers the explanation is the inquiry, How do objects float?, which arises from the domain of the observer’s experience. First we observe what we want to explain. The object doesn’t fall through the liquid like it does in air. Next, its floating is described in particular lan~,maagewithin a context of relations. The latter is a critical point because the objectis not isolated but within a system. We further propose a mechanism, (p. 3)

such as an upward force, that would generate floating. As Maturana and Varela argue, we conserve coherence with experience as we construct a mechanism that will generate the phenomenon we are explaining. (p. 4)

While the garden hose explanation aptly satisfies many children, it would not be adequate for the scientific community. We can say then the adequacy of an explanation is determined by the listener/s. But of course there are multiple explanations deemed adequate by multiple perspectives. The child “knows” somethir~g. Is it adequate knowledge? The scientist knows something else. In science teaching what is the common ground between these two perspectives? What are criteria for evaluating students’ explanations in school science as scientifically valid? (p. 4)

School science is barraged with scientific explanations. They are, however, pre-packaged and validated by communities of scientists rather than communities of students. Questions these explanations answer do not arise directly from the students’ domain of experience or fantasy. And if the students do not ask the questions to which these explanations belong, they are not actively involvedin inquirynor in explaining. It is a vicarious participation at best, an attempt at making sense of an abstract concept that did not arise from students’ manner of experience. Comprehending textbookexplanations becomes problematic because understanding requires the moment of relevance that experiencebrin~ forth. (p. 4)

authentic problem
tracing source of principles (p. 4)

Two TypesofScientific Communitiesand Two TypesofDiscourse (p. 4)

Researchers have observed that scientists construct knowledge in two very different kinds of communities. One community thrives within the scientist’s laboratory, while the other exists in the communityat large. Each has distinctly different types of social interactions and purposes concerning the nature of knowledge growth. Dunbar (1995), for example, provides a detailed account of the constructive and cooperative, interpersonal processes that exist among members of small teams working together successfully within the same laboratory. Latour (1987) and other sociological researchers, on the other hand, have described a very different inter-laboratory community that is competing for grants, status, and prestige as rewards for the establishment and acceptance of new knowledge. (p. 4)

VirmaUy all scientists are members of both types of communities. (p. 5)

In general, we are describing the main difference between intraand inter-laboratory communities along a private and public dimension. The inter-laboratory community provides the public forum for scientists. This forum sets and applies a discipline’s standards and benchmarks and supports the arbitration that lets the discipline advance. These environments are high risk forums for scientists’ egos and careers. Intralaboratory communities, by and large, are private and low risk environments. Scientists use this private forum to discuss ideas that are not fully worked out without high risk to their ego or career. As such, the discourse can be inquiry oriented, jointly constructed, fragrnented, and extended. The value of such a community is that ideas can be nurtured toward maturity since the inquiry focus reduces the risk that it will be discarded prematurely. To maintain a private, low risk environment, intra-laboratory groups need to be reasonably small and cohesive. (p. 5)

It is our contention that the classroom can support both intraand inter-laboratory type communities. (p. 5)

Iterations of lntraand InterGroup Discourse in the Classroom (p. 6)

some great examples here. (p. 6)

Sinking and Floating (p. 6)

improvable ideas (p. 8)

While small groups manipulated ideas and materials, and generated explanations, the large group and on-line discussions pushed the students to increase their clarity and coherence of their explanations. The adequacy of explanations was therefore determined by the class at large. In the above set of class responses, for example, students asked for clarity on the idea of one liquid being “stronger” than another. In general, we believe these intermittent class discussions provided the means to disseminate students’ ideas that were formulated in small groups, raise the level of explanations as a whole, and raise’the benchmark of student learning within the topic (p. 8)

examples of improvable ideas (p. 8)

The Evolution of Students’ Explanations (p. 8)

We were interested in these explanations as starting points for the documentation of how group explanations of light propagation or floating evolved over time. (p. 8)

Our analyses examined the nature and sequence of explanations citing causal ideas and variables that were abandoned, or that survived in various forms to the final explanation. Across the studies are three salient levels of explanations that have particular foei and qualities. (p. 8)

First order explanations: Expected outcomes (p. 8)

In Studies 2 and 3, beginning explanations captured in pre-testsand early group conversations usuallydescribedpropertiesofobjectsratherthanmechanisms, orrelations. (p. 8)

Theseexplanationsarelikelybasedoncommon observationsand inferencesaboutmaterialobjectswithinthephenomenon. (p. 8)

Similarly, in Study 1, after observing the fwst shadow effect, students first talked to each other about shadows according to how they usually appear: they are dark, size is variable, the shape of the object and “its” shadow are congruent; the shape of the hole and the image are also congruent. (p. 9)

Generally in all three studies, students’ initial explanations did not provide any causal justifications as to why heavy objects sink, or why the size of a shadow varies. Second, these explanations did not involve a dynamic system of relations, such as the relation between the floating object and liquid; or relative sizes, distances or shapes of the light source, aperture (or object) and screen. (p. 9)

These first order explanations correspond to expected outcomes. The contextual relations of components (e.g., between object and medium) are missing. (p. 9)

Second order explanations: Convergence on variables (p. 9)

Across the studies, student intra-group explanations of floating or shadows/images began to change over time as students manipulated the materials themselves, and observed the related phenomena that we sequenced across the unit. Students’ convergence on variables involved a deconstruction of what matters to a particular phenomenon in determining, for example, whether an object will float, or what determines a shadow’s size. (p. 9)

Within this exploratory space, explanations began to take explicitly the form of, “It depends…” The relations of a system (e.g., light, aperture, screen) became articulate indicating a shift of some sort in students’ causal thinking. (p. 9)

The following is part of a conversation between three grade seven girls about floating and sinking “things. (p. 9)

Gid I: Girl 2: Girl 3: Girl 2: Girl 1: Girl 2: Girl 3: Girl 2: Girl 3: Girl 1: We think that lighter things will float… Lighter things float on heavier things. On heavier liquid On heavier liquid. That’s what I meant. Because if you take like a brick or fill up this container with a brickor whatever, like cement, and then you take the same container and fillit with water and you weigh them together, the brick will be heavier. But with a boat… Because the boat’s hollow. Yea and so take the boat and weigh it on the ocean water, how much it takes up, what the space takes up, the water would be heavier…. Yea but if you mould the clay it takes up more room. So if you take that much room. Okay,lighterthingsfloatonheavierthingsiftheybothtakeupthesamespace,right? Is that better? Yeah, that’s good. (p. 10)

This dialogue began with a simple statement of a First Order type, much like explanations we heard from students in the beginning of the unit, but quickly pro~essed toward an explanation of relative densities (“lighter things” and “heavier things” of the same volume). (p. 10)

Due to their experimentation, the girls’ prior idea that shadow shape is invariable (a shadow is always the same shape as the object blocking the light) becomes interrupted. (p. 10)

Explanations typically describe a set of relations among variables they believe to be relevant to the context. The adequacy of their explanations-in-progress are negotiated within their mutual discourse. (p. 10)

Third order explanations: Coherenceof related phenomena (p. 10)

Through class debate and sharing of ideas, students went “back to the drawing board” several times prior to final explanations. (p. 10)

This last class discussion was a shared effort of fitting group explanations into one explanation, the products of thinking about all the activities and discussions and arriving at one way to interpret them. It is clear that the students had progressed from individual statements such as, “Light things floaL” In the end, students had expressed ideas, without specific scientific vocabulary, about density (in terms of “spreading weight”), buoyancy (as a force that pushes up), and displacement (the spreading of water), and how each related to what they observed. In this sense, the final class explanation proposed not only a mechanism, buoyancy, that explained floating but addressed relevant relations (weight to volume; and densities). (p. 11)

In general, the students’ final explanations across the three studies focused on relative conditions while the beginning explanations did not. These latter explanations involved a complex system of priorities applied to the conditions (p. 11)

Groups’ concluding explanations were more than descriptive and dealt with issues of “fit” regarding coherence of various conditions and effects within the activities (p. 11)

In diSessa’s early work on scientific reasoning (1983), he explored differences between novices and expert reasoning and concluded, “Experts have a vastly deeper and more (p. 11)

complex priority system” (p. 32). We argue that such priority systems can be developed by: actively manipulating materials to produce a phenomenon and trying to explain that mechanism that generates what is happening; and testing the coherence of that mechanism with related phenomena. The latter conclusions get debated in the iterations of intraand inter-group discourse. (p. 12)

Implications for the Classroom (p. 12)

BeginningReflections (p. 12)

She noticed that some students in small groups were dominating the discussions and making unilateral decisions for the whole group while other students remained passive. She wrote that small groups might work better ff more of the students were held accountable for their understanding and group tasks. “I wonder if duties can be rotated. (p. 12)

A few times when students couldn’t convince each other about their ideas and arrive at consensus, they wrote down all the ideas and presented them in the large group discussion. (p. 12)

At this midpoint Susan noted that the students’ debating “skills in the large class discussions could be improved.” (p. 12)

By the last day Susan wrote that the verbal exchange between the groups had changed and was more like a discussion rather than the groups defending their ideas. (p. 12)

She noted that the students’ confidence levels in discussing their explanations had increased because their knowledge base had increased. What was swiking to us from her notes was a description of a quiet girl, Randa, who became actively involved in the later class discussions. (p. 12)

Pedagogical Considerations (p. 13)

Both teaching and learning science content are formidable for several reasons. (p. 13)

Second, students are expected to induce such complex coherent ideas from textbooks, laboratory activities and lectures. However, these sources typically present powerful ideas as distilled facts and explanations without prerequisite grounding, historical context, or explanatory justification (i.e., why do scientists think this way). (p. 13)

Third, the formal language of science is problematic. Solomon (1994) describes learning science as “arrival on a foreign shore or as struggling with conversation in ~n unknown language” (p. 16). In school science, students’ response of memorising scientific explanations does not in itself push students toward coherent thinking or an appreciation for scientific explanations. (p. 13)

We propose that students become engaged with inquiry and explaining. (p. 13)

We suggest that the iterations of intraand inter-group inquiry discourse motivates students to increase explanatorycoherence by abandoning some ideas and advancing others. We have found that students’ explanations show a progressive shift in this process from explanations that focus on properties of objects to dynamic ones that incorporate complex priority systems and relative conditions. (p. 13)

Our concerns about this approach relate to its viability in the classroom. We are asking teachers to combine small group learning techniques with a methodology that also requires multiple large group sessions. While the approach is wholly consistent with the literature on small group learning (see for example, Cohen, 1994; Johnson & Johnson, 1987; Nastasi & Clements, 1991; Slavin, 1989) the requirement of producing coherent and powerful explanations is an ill-structured task. (p. 13)

challenges of KB (p. 13)

Like learning models addressing ill-slructured tasks, our approach requires that students become accustomed to generating, evaluating and sharing ideas, and possibly debating them. In addition, like generative teaching models, what we have proposed requires collaboration and blocks of time. (p. 13)

we asked students to provide a critique of the related textbook explanation, by asking, “What is left out of this explanation?” or “What do you understand about this explanationT’ (p. 14)

From a learning perspective, the approach we propose falls in the growing camp that argues that cognition is a result of social interchange and is constructed, shared and spread among groups of individuals (Gruenfeld & Hollinshead, 1993). (p. 14)

Specifically, our approach takes advantage of the intm-‘and inter-group dynamics to promote progressive discourse by cycling the students between the differing discourse environments. (p. 14)

Blog Logo

Bodong Chen



Crisscross Landscapes

Bodong Chen, University of Minnesota

Back to Home