about patterns? What do we know about normals?
Student3 - [mumbles something then retracts]
Teacher - You learned about the bell curve in math? what does the bell curve represent?
Student3 - [inaudible]
Student4 - There's a few at one end and a bunch in the middle.
Teacher - OK, that's a real important thing to keep in mind as a scientist, do you suppose that scientists just say "op, that's it, I've got this one piece of evidence, or this one individual characteristic that doesn't fit so I'm going to throw this theory out, this theory is no good"
This example is from a whole-class discussion that served as a culminating activity for the Finch Scenario. In this discussion students and teacher, together, describe a detailed causal explanation for the episode presented in the problem. The discussion in the following excerpt has two interesting characteristics: the level of detail and extent of the causal story described, and the conversation structure. In the example the teacher responds to students by asking for extensions or continuations, rather than with evaluations. In this way she bridges student remarks and helps create a structure that is closer to student cross talk. Here students are the "talkers of science," unlike traditional question-answer-evaluation structures where the teacher is the "talker of science" ( Lemke, 1990).
[Students from different groups report on their final findings, the teacher synthesizes responses from the different groups and restates a common causal story. The discussion began with a description of environmental conditions that introduces a pressure].
Teacher: And that pressure selected what? This is where I'm not sure I understand. It selected what?
Students - [students call out responses, but it is hard to distinguish statements].
Teacher - The beak? It selected which beak?
Student1 - The longer.
Teacher - The longer beak. Now, I really get fuzzy -- why?
Student2 - Because the only food around was harder.
Teacher - The only food that was left during the dry season was hard?
Student3 - And it was hard to break open.
Teacher - And it was hard to break open. OK. And . . . finish the story . . .
Student2 - So the weaker ones . . .
Teacher - So the weaker ones, the weaker finches could what?
Student4 - The females and the younger ones had a hard time . . .
Students - Cracking
Student6 - Consumption went down
Teacher - How did you know that their food consumption went down?
Student6 - Said so in the field notes
Teacher - Said so in the field notes? OK. I just wanted to make sure you had some evidence for that. Did you all agree? You don't have to -- some of you told me about mating, some of you told me that you thought the beak length let them reach, reach higher, Student7 I think you told me that?
Student7 - Yeah
Teacher - Did you have evidence to support that?
Student7 - Yeah.
Teacher - What did it say?
Student7 - I don't know its not here.
Teacher - What do you remember?
Student7 - Something about reaching for the food.
The level of detail in the explanation that was constructed during the discussion above exemplifies the advantage of whole class, cross-group discussions. Very few groups in the class constructed as complete and detailed an explanation as the one laid out in the quote above, however, this culminating discussion drew on several groups' explanations in order to construct a more complete, and detailed explanation.
Students learning science by collaborating on investigations can be more effective if these collaborations occur in a context that combines three types of supports. First, domain-specific strategic support focuses investigations and discussions on key principles of the domain. Second, teacher-student interactions during small-group investigations augment the support provided in the computer environment by modeling profitable strategies, and answering questions. Third, teacher-student interactions in whole-class forums enable the class to