based explanations in science may have a very complex structure, this instructional approach identifies the most central elements and makes them explicit to students. These elements consist of the claim, the empirical evidence in support of the claim, and the reasoning that articulates why the evidence supports the claim. This instructional framework is developed with students’ input, as they learn first about evidence supporting claims and then consider how to organize a written or oral presentation to defend a claim. Teachers discuss the three elements with students, who then begin to use this framework to represent their written explanations in response to research questions. As they construct and refine these explanations, students use worksheets with scaffolding prompts that remind them of the elements and the criteria for them. This framework becomes a repeated structure that they use to guide their investigation, and it guides the synthesis of results into an explanation. Empirical studies of students’ explanations reveal that students using this instructional framework improve in their ability to cite relevant data and connect it with claims within their written explanations (McNeill and Krajcik, in press). A similar approach embedded in software tools rather than paper and pencil worksheets was used in the Knowledge Integration Environment (Linn, 2000). In these tools, a checklist of tasks specifying important steps in inquiry was provided to help students coordinate the different steps in the activity. Similar approaches have been explored for scaffolding prompts to help learners articulate their experimental designs (Kolodner et al., 2003).
Another approach is to structure the tools that students use to represent their ideas in order to make the important aspects of the task more explicit. This is apparent in several software tools for argumentation. For example, SenseMaker (Bell and Linn, 2000) provides a representation that helps students develop and record their arguments. Students explicitly identify relevant evidence and code it as supporting or refuting sets of competing claims. Belvedere (Toth, Suthers, and Lesgold, 2002) supports students in constructing argument graphs, in which claims and evidence are visually distinguished, and students construct a chain of reasoning that includes claims, subclaims, and their supporting evidence. These tools can help learners develop more accurate and elaborate arguments, focusing them on the distinctions relevant for the domain (such as claim or theory versus evidence).
We have discussed the promise and the complexity of social interactions in doing science. Scaffolding science learning through students’ social interactions can harness the complexity of a scientific task and students’ varied experiences and observations with it to build understanding in a student group. This type of approach has its roots in the reciprocal teaching ap-