Conference Discussion Format

Revision history: Original version 2018-02-03. Revised with input from Kate Sanders, 2018-02-04.

This writes up the format for discussion that I have used several times, such as at SNAPL. However, the original is not due to me: I learned about it from ICER, the International Computing Education Research Conference, via Kathi Fisler. I understand that credit for the physical layout is due to Mary Sullivan, who suggested it to the ICER 2011 chairs, and for discussion at tables, Sally Fincher.

There are several problems worth addressing in the conventional q&a format at a conference:

• Relatively few questions get to be asked and answered.
• There's no space to “debug” a question before posing it to the speaker.
• People who are shy may be reluctant to ask a question at all, even if they have excellent ones to ask.
• People who need to vent end up taking up more time than they should, proportionally speaking.
• Simple clarifications either take as much time as complex questions, or don't get asked at all. Correspondingly, there's less opportunity to build on a clarification to get to a question, or to build on simple questions to get to complex ones.
• There's a performative aspect to questions, which isn't always edifying.
• It can be stultifying to sit around all day only listening, saying nothing.
• Finally, and most subtly, it puts too much emphasis on the individual over the wisdom that can emerge from a group.

The ICER format brilliantly addresses all these problems.

Instead of arranging the audience in the conventional rows, people sit around (circular) tables, facing one another. You want about four to six people per table: enough to enable discussion, not so many that there's not enough time per person. You may need to load-balance the room a little to avoid ending up with tables with too few people.

Suppose, conventionally, you have a 30 minute talk slot, with the speaker expected to finish in 25, leaving 5 for q&a. Instead, give the speaker about 22 minutes. Reserve 4 minutes for “table discussion” and 4 for q&a. Similarly, 25 can be broken down into 17 + 4 + 4. And so on.

While the speaker is speaking, audience members rotate their chairs to face the speaker. (Make sure the chairs are light enough to enable this. If you have someone with mobility issues, place them facing the speaker the whole time.)

When the speaker is done, during table discussion, everyone turns back to face the table. During this time, they have a private, table-local discussion. The speaker has no role, though they are welcome to walk around and listen in on the conversations, and maybe also answer any quick clarifications that are needed.

When this time is done, a conventional q&a begins.

So, what's different?

All the magic happens during table discussion.

• Everyone has an opportunity to speak. The session chair can walk around making sure that conversations are not being dominated at any table.
• Parallelism. Lots more questions get asked and discussed (not all can be answered, but you may be surprised how often someone at a table is able to make at least a decent stab at answering).
• Simple clarifications can usually be addressed at the table itself. Having the speaker walk around and answer short, quick questions can aid this process.
• People collaboratively build on one another's work. People consolidate questions. One person suggests a question; another person—who may not have thought of the question themselves—may contribute a superior formulation of the same question. Questions get merged with one another. One person may decide someone else's question supercedes theirs. Many questions are resolved locally; the table will usually vocally identify some as the ones to ask of the speaker.
• Shy people are much more likely to participate since they are only speaking at a local table, not before everyone, and everyone is being expected to speak. Also, while tables do not have formal moderators or leaders, usually some responsible person will notice someone who's been quiet and try to bring them into the conversation.
• Even if someone does not have anything to say, listening to the different perspectives can be awfully educational.
• It's inevitable that people will want to vent (sometimes with justification). The table provides a good medium for that, letting them “get it out of their system”, without having to take up the whole room's time. Someone else at the table may even be able to contradict their rant or shape it into a constructive question.
• People tend to be much more awake and engaged: a non-trivial percentage of the time of “a talk” is actually time for them to be engaged in active learning!
• Most subtly, it slightly shifts the power dynamic in the room. Whereas conventionally the speaker is mostly in charge, here the focus of power explicitly shifts away from the speaker for a while. That demonstrates that the purpose of the event is to educate everyone, not just to feed the speaker's ego.

Of course, at breaks, suggest or require people to switch around tables. That way they get a different set of perspectives for the next session. Also, people who may not be rubbing along well have a natural way to separate without social awkwardness.

This process works poorly if several people at a table are more engaged with their email than with the discussion. Why people travel all the way to a conference, spending money, dealing with jetlag, etc., just to read the same email they could have read from the comfort of their room, I have no idea. But they do. Consider nudges or rules to get people to engage.

Finally, here's the litmus test of when the process has really worked: it's when a person stands up and asks, “We were wondering...”. Notice the we. That's a sign that the table came up with a question. Usually such questions are rich, interesting, and reflect the combined inputs of several people.