Monday, April 28, 2014

Immediate Feedback Is Too Soon For Feedback

The Quickest Feedback You Can Get, via Frank
(Blogging between classes here, so I've set myself a 7 minute time limit on writing this post. Apologies if it shows!)

(The obvious questions is, can't the post wait? That's a good question.)

What makes for good feedback on a quiz?

The most common answer around the internet is that it all has to do with timing. Immediate, personal feedback is the best, and we should take steps to shrink the amount of time between an assessment and our input. I totally subscribed to this approach for a few years. Frank's brilliant management system made it possible for me to pull this off in my classes.

Lately, I've been questioning the assumption that the best feedback is immediate. For one, there's a sort of emotional investment that a kid has with a quiz right after they've finished it. A lot of the kids who checked their quizzes right after they finished had this sort of "I got this wrong? WHY?????" reaction that I didn't love.

I also suspect that there are learning benefits to delaying the feedback on a task a day or two. Delayed exposure to ideas, and all that.

And the other thing is that, realistically, the only way to pull this off is for kids to be all on their own with an answer key. Maybe you guys are better at making answer keys than I am, but I wasn't sure how much my kids were getting out of them. What they really wanted was a conversation about the problems, but I couldn't give the kids that conversation because their classmates were still taking the quiz.

(Actually, I'm sure you guys are better at making answer keys than I am because I always end up hastily scribbling them on my way down to the photocopier machine.)

(I've just blown past my 7 minute deadline. We're up to 11 minutes now. Oh dear.)

In short, I wanted to give my kids feedback with a little bit of a delay, but not too long of a delay. I wanted kids to be able to reflect on their own work, but I wanted it to be social.

Here's what I do now:

  1. Kids take quizzes on Friday.
  2. I collect them, photocopy them, take notes and make grades. Digitally scan them for reference.
  3. Kids get back their originals on Monday with no notes on them from me. They get back what they handed in.
  4. They get into groups, go over their quizzes together, iron out any differences or disagreements. I always make my quizzes a little bit on the long side, so there's usually a question or two they didn't get to on Friday that they can work on if they finish.
  5. I have notes about areas of concern. I check in on kids who I had flagged for a conversation. I ask groups questions if there's a disagreement that they're not seeing yet. I'll ask the whole class questions if there are problems with widespread errors.
And the only part of this that I lied about is the part about giving grades, since I'm currently teaching at a school that doesn't give grades. But that's how I would've graded them, if I had to. 

It's entirely possible that this system wouldn't work in a place where grades were given, like where I'd been teaching before this year. Kids would probably go nuts not knowing what their grades were. I'm optimistic that, with enough time, I could have convinced them that this way was better, but dunno. 

If you try this, let me know how it goes!

(Clocking in at 19 minutes. Gah!)


  1. Whoa! I have been experimenting with something similar this year! I give the quiz one day on all the "standards" (as defined by me) that we have covered thus far. Then I photocopy them (without any grades or marks from me). The next day I give back the photocopied version and give them time to work in groups to figure out if they were correct or not, and to correct what they need to (by crossing out and writing in). Then I collect and grade the quiz, slightly taking into account if they made small errors and were able to fix them without my help. They also do a written reflection that second day on how they think they did (before I grade it but after talking to classmates).

    After a while, I give back the quiz (graded) and they chart their progress on each "standard" on a bar graph. They choose review homework for the next quiz accordingly.

    The issues I've been having are:
    - the quizzes I've been giving are more basic skillsy type things, and I aim for a low common denominator, things that everyone really should know. This means that I have several kids that get most questions mostly correct, they are not really very creative questions. These kids have little investment then in talking with the other kids during the correction time.

    - The actual correction still comes from me, and they don't really go and look at what I marked. This seems like a flaw in the loop. I'd rather not spend the time doing the correcting, but it's not really happening in the second day for everyone. But then they are not really looking at what I write.

    1. Woot! I've got Katie in my corner, guys! We'll show you how it's done. :)

      Re your issues: One of the main reasons why I like what I'm doing is because kids never paid attention to my written corrections anyway. (On twitter, Kelly O'Shea has suggested that this is because I haven't coached my kids properly. Maybe?)

      I basically have dropped corrections. The written feedback I give now is almost entirely in the form of questions that I think would be helpful for kids to think about while they're reflecting on their work.

  2. What kind of questions do you put on a quiz (in terms of cognitive demand)? My thinking was that I would assess "doing mathematics" type things in the form of projects, or group quizzes (with everyone held accountable) or other more creative ways, and assess more procedural "can you do it" type things with individual quizzes, which is why the answers are so straightforward and correct or not correct. It sounds like maybe you have more creative questions on your tests?

    1. I can send you some of my quizzes sometime, but I definitely want my quizzes to assess "doing mathematics." I worry that "can you do it" questions aren't rigorous enough, and that they won't be able to apply their procedural skills in unfamiliar contexts. Of course, I also want these to be questions that the kids can do on their own.

      One thing that helps me strike a balance is that anything from the entire year can be on the quiz, and I don't tell them in advance. Sometimes, but not always, this can infuse a procedural question with a new challenge.