Reacting to their votes: Instructor agility

by Stephanie Chasteen on April 10, 2015

You don’t know how your students will vote on a clicker question, but you can anticipate and prepare yourself for the likely outcomes. It’s really important to use a clicker system which lets you have a sneak-preview of student responses – as i>clicker does, shown below. This lets you “hold back” the histogram from students until you’ve decided where to go with the question. Remember, giving the answer stops student thinking!

i>clicker base

I’m assuming you’re using Peer Instruction – where students vote on a question individually first, and then talk to their neighbors and re-vote after that peer discussion. I think it’s important to give students the chance to vote on their own first, because it gives everyone a chance to process the question (even those who are underperforming, or are English language learners).

So, your main decision point comes after you’ve gotten student responses to the first vote.

Scenario #1
What might you do if you get the following vote distribution (where C is the correct response)?

Graph

A lot of you might say that you’d just discuss the question and move on, without having students talk to their neighbors. That’s fine, but with two caveats:

  1. A lot of students didn’t get the right answer – if you add up all those numbers on the unpopular choices, it’s a sizeable fraction of the class. So, I think it’s really important to always talk about why the right answers are right, AND why the wrong answers are wrong.
  2. What is your cutoff at which you no longer have students turn and talk to their neighbors? For us, it’s 80% – if less than 80% got it right, we want students to talk it out, because the question is challenging enough. You should determine your own cutoff in advance, so you’re not hemming and hawing in front of the class.

Scenario #2
What if you got the above distribution, but C was NOT the correct response? What might you do then?

Well, perhaps you expected them all to get it wrong – maybe C is a really tempting distractor. If so, you might have them turn to their neighbors, but throw them a hint (like asking them to consider a certain situation, or think about a certain idea). If you’re surprised that they’re all getting it wrong, you might ask a few students to explain their answers, so you can see if there is a problem with the interpretation that you need to clear up.


Scenario #3

What if you get this distribution after the first vote?

Graph

You might have good luck having students turn and talk to their neighbor on this – but I’d say that a distribution like this suggests that the students are just guessing. Is the question confusing? Are they not awake? Are they missing some key piece? You might fish around to see if you need to give them some more information.

Scenario #4

Lastly, what about this distribution, after the first vote?

Graph

This is a beautiful distribution – students are strongly drawn by two of your choices, and they’ll have good productive arguments. Have them turn and talk to their neighbors, this is a picture-perfect opportunity for peer instruction.

Note that I don’t show students the distribution before I have them talk to their neighbors – I want the mystery to be maintained to spark that discussion. The only exception is if (a) I have a distribution which is pretty near 50/50, so students realize there is really split opinion I the room, or (b) if I’m using a survey/discussion question, where I feel that seeing the diversity of opinion in the room will spark student discussion.

Images courtesy of Peter Newbury and Cynthia Heiner.

This is reposted from my article on the i>clicker blog.

Comments on this entry are closed.

Previous post:

Next post: