I blame inquiry. Not inquiry in the generalized, overloaded, science teaching approach sense. Just the word. "Inquiry."
Even now, when I hear the word "inquiry" I still think mainly of asking questions and designing experiments. A bad side effect of thinking in this way was that I would spend far too long having students ask questions and design experiments and very little time evaluating evidence and generating claims. On good days, when we miraculously got cleaned up before the bell, I might have spent 10 minutes at the end of class telling students what they were supposed to have figured out and students would answer questions like, "Explain how you know mass is conserved in a chemical reaction." (Answer: Because you just told me by asking that question.)
We were very busy and very engaged and learned very little.
There are a few structures I've been using to help shift the focus on the class to analysis and argument. One of them is the claim-evidence-reasoning framework. I'm not sure who the originator was but most of what I do came from Katherine McNeill who has published a ton on it. She wrote a book too.1
Claim-Evidence-Reasoning (pdf and pdf) is a framework for writing scientific explanations. Occasionally I use it as a probe in the style of Paige Keeley. After most labs, students are asked to write about a paragraph worth in this format and a more extended version when we wrap up a unit.
As part of their lab handout they get a prompt that looks like this:
As the year goes on I remove most of the scaffolds until ultimately the students just get a prompt or question.
I've been happy with it. There's a fourth step, Rebuttal, which I've never gotten off the ground.
I like frameworks a lot. I like having specific language I can refer to over and over. In a typical initiate-respond exchange, I follow-up by ask students for their reasoning. When students want to whine or talk back to me, I get to ask them, "What is your evidence? How does that evidence support your claim that I'm the most annoying teacher in the world?" (Evidence #1: The fact that I ask that question.)
The key to implementation is that the structure of the class really has to be designed around C-E-R. Even if I'm just direct teaching something, I need to model how to think about the evidence that led to the claim.
I also give my students a whiteboard format now. It usually looks like this:
I used to structure it in Claim-Evidence-Reasoning order but I realized students would then write their claim before they got their evidence.
McNeill uses C-E-R for essentially everything. Any major idea can be written up in that format. I think the framework works best when students truly have something to argue about. Of course, I'd say that all learning works best when students have something to argue about.
I've had a lot of success with using it when students are constructing the kinetic molecular theory or deciding if mass is conserved in a reaction or not. Why things float or sink was fun. "Is Pluto a planet?" is a good one. A lot of Dr. McNeill's papers are written around a multi-week unit designed around the question, "Are soap and fat different substances?"
I've been happy with the results. We used to spend all of our time just generating data. Now that data is being put to use.
For more on writing, I show up in an interview in ASCD Education Update titled Improving Student Writing Through Formative Assessments. It's only for ASCD members but most of what I have to say is in Managing Feedback.
Addendum: Kirk shared a free NSTA article called Engaging Students in the Scientific Practices of Explanation and Argumentation. Berland and Reiser in particular have written a number of good articles on the topic.
1: I can't fully recommend the book. It's pretty basic and you can find most of the information from googling around for her various research papers. If you're not interested in dealing with that you might want to give it a go.