Here’s a snapshot of our model as it existed around March — I have removed all student names for privacy, but I would normally track who proposed or asked what, and keep notes about the context in which questions arose.

We gathered evidence from our own measurements and from published sources.  The students proposed most of the additions to the model, but occasionally I proposed one. I’ll write in detail about each method next — this post covers some basic ideas common to all approaches. Additions to our shared mental model of electricity tended to be short, simple conceptual statements, like “An electron cannot gain or lose charge” or “Between the nucleus and the electrons, there is no ‘stuff.'”

Ground Rules

Here are the ground rules I settled on (though introduced gradually, and not phrased like this to the students).  I was flying by the seat of my pants here, so feel free to weigh in.

1. Every student must contribute ideas to the model.
2. For an idea to get added to the model, it must get consensus from the class.
3. Deciding to accept, reject, or change a proposal must be justified according to its clarity, precision, causality, and coherence (not “whether you like the idea/presenter”).
4. Each student is responsible for maintaining their own record of the model.
5. Students may bring their copy of the model to any quiz.
6. Quizzes will assess whether students can support their answers using the model.

1.  What if the class rejects someone’s proposal?

Adding an idea to the model is a skill on the skill sheet.  As with any other skill, students get lots of feedback, then they improve.  They have until the end of the semester.  I don’t allow them to move to the “class discussion” stage until I’m satisfied that they have a well-reasoned attempt that addresses previous feedback with a solid chance of getting something accepted. Unlike other skills, though, they’re getting individual feedback sheets from each of their classmates, not just me. Most people wrote two drafts.  No one wrote more than three.  More on this soon.

2. Consensus from the class — are you kidding?

I had a small class this year (12 students at the high-water mark) and I’m not sure how it will work with 20-24 next year — I’m thinking hard about how to scale this.  But yes, it really worked.  We used red-green-yellow flashcards to vote quickly.  I did not vote.  My role was to

• Teach them how to have a consensus-focussed conversation
• Point out supporting or conflicting ideas in the model, if the students didn’t — in other words, to “keep the debate alive [if] I’m not convinced they have compelling arguments on either side”
• Do some group-dynamics facilitation, such as recording new questions that come up (for later exploration) as a result of the discussion, making sure everyone gets a chance to speak and is not interrupted, and sometimes doing some graphical recording on the whiteboard if the presenting students are too busy answering questions to do their own drawing.
• Determine when the group is ready to move to a vote or needs to table the conversation

That made me a cross between the secretary and the speaker of the house.  Besides being productive, it was fun.  A note about group dynamics facilitation: I’m often frustrated by the edu-fad of re-labelling teachers “facilitators.”  I’ve done a lot of group dynamics facilitation for community groups.  It is a different role than teaching, and it’s disingenuous to pretend that students need only facilitation, not teaching.  However, in this situation, facilitation was called for.  The group had a goal (to accept or reject an idea for the model) and they had a process (evaluate clarity, precision, etc. etc.); my job was to attend to the process so they could attend to the goal.

3. What about people blocking consensus out of personal dislike, and other sabotage maneuvers?

There is a significant motivation for students to take part in good faith.  First off, no one wants these conversations to go on forever: they’re mentally challenging and no one has infinite stamina.  Second, if a well-supported, generative idea is left out of the model, no one will be able to use it on the next quiz.  Third, if the class accepts a poorly supported idea, it will cause a huge headache down the road when ideas start to contradict; we’ll have to backtrack, search out the flaw in our reasoning, and then uproot all the later ideas that the class accepted based on the poorly reasoned idea.  They were darned careful what they allowed into their model.

Other than that, I used my power as the facilitator.  Conflict happens in any group, and the usual techniques apply.  It was almost always possible to resolve the conflict fairly quickly: modifying the wording, adding a caveat.  We’d vote again, the idea would get accepted, and we’d move on. Sometimes a student felt strongly enough about their objection to propose an experiment that we should try; unless the group could come up with a reason why we shouldn’t do that experiment, we’d just put the idea on hold, and I’d add that experiment to the list for that week.

Once, a student voted against a proposal but was unable to explain what they needed made clearer, more precise, more causally linked, etc.  It just “didn’t sound right.”  None of my usual techniques worked to draw out of that student what they needed to be confident that the proposal was well-reasoned, or just to feel heard.  So I reminded them that the model is always modifiable, that we can remove ideas as well as add them, and that we have committed to base our decisions on the best-quality judgement we can make, not “truth” or “feeling.”  I told them that I would consider the idea accepted for the purposes of using it on quizzes, but record the disagreement, and that if/when new information came up, we would revisit it.

An important point about facilitation: these conversations were sometimes fun and lighthearted, but sometimes gruelling for the students, especially as we moved into more abstract and unintuitive material.  The most important mistake I made was letting the conversation drag.  To remedy this, I used what I consider fairly aggressive facilitation — quickly redirecting rambling speakers, proposing time limits, summarizing and restating sooner than I otherwise might, etc.  If the conversation was so unclear that students weren’t able to even give feedback about how to improve, I had to diagnose that as soon as possible.  I would say something like “It looks like we’re not ready to move forward here.  Joe, let’s meet after class and see if we can figure out how to strengthen your proposal.”

4.  Students maintain their own record of the model

Benefits:

• Students had a reason to take decent notes, at least about certain key ideas.
• The list got really, really, long , and included a lot of topics, many of linked to each other.  It was a powerful visual indicator of just how huge of an endeavour this really is.
• Most ideas fit clearly under more than one category.  It was up to the students to choose.  It was a good reminder that the new ideas aren’t divorced from the old ideas, that all the chapters in the textbook really do connect.

5. Every test is “open notes”?

In a way, yes.  I felt a bit weird about this — there are some things they really need memorized.  But it worked out really well.  I allow students to bring their copy of the model to any quiz (no other notes).  It must contain only the ideas voted on and accepted by the class — no worked problems, etc.  I circulate during the quiz and randomly choose some paper copies to take a close look at.   It was a non-issue.

About halfway through the semester, students gradually stopped bringing it.  We used that thing so much, for so many purposes, that they mostly didn’t need it.  Besides, like any quiz, if you have to look everything up, you won’t have time to finish.  Then you’ll have to apply for reassessment, which means doing some practice problems, which means building speed and automaticity, which means needing the notes less.

6.  “You’re going to grade us based on what we say?!”

I modified many quiz questions so that they said things like

Explain, using a chain of cause and effect, one possible reason for the difference in energy levels on either side of S1.  It doesn’t have to be “right” – but it must be backed up by the model and not have any gaps.

This worked extremely well.  It helped students enter the tentative space between “I get it” and “I don’t get it,” saying things like “Based on our model, it is possible that…”.  It gave me the opportunity to show a variety of well-reasoned answers (I sometimes used excerpts from student answers to open conversation in a later class).  It helped me banish my arch nemesis: the dreaded “But You Said.”  (Because I didn’t say.  They did.)