GUEST POST: Feedback -- We Still Don't Know What Works
In order to spark discussion, we feel it is important to publish posts related to learning while offering a different perspective. So, we have begun inviting guest bloggers to write about education and learning from their unique perspectives, so that we can come together to engage in lively discussion.
By: Michael Pershan
Today, Michael Pershan, @mpershan, writes about feedback in the classroom and his experiences as a teacher navigating the literature for classroom-ready applications. Michael is a Math teacher in New York City. He writes his own blog, Teaching with Problems, and edits Math Mistakes, a website dedicated to compiling, analyzing, and discussing the mathematical errors students make.
Each summer since I began teaching has had its own idiosyncratic rhythms. A few summers ago, my routine went like this: after lunch I’d take the subway into the Bronx, get off at the library, download a bunch of articles about giving feedback, and try to make sense of it all on the subway ride home.
I was unhappy with how I’d been giving feedback. Most of my feedback to kids was via grades-and-rubric or a quick conversation. The quick conversations were unsystematic; the grades-and-rubric combo, I was now sure, wasn’t doing much good either. I turned to research for help.
That summer was intellectually thrilling. This was the first time I dove deeply into any research. I especially remember reading a review paper by Kluger and Denisi: according to their review, in over 1/3 of quality experiments, feedback interventions led to worse performance than no feedback at all. Feedback could be detrimental! I had sometimes consoled myself that, no matter the flaws in my feedback, it was better than nothing. Now I wasn’t so sure.
I also read Ruth Butler’s work. She designed a series of experiments that gave reason to believe that grades do not make for good feedback. Just as important as her results was her explanation: grades draw students’ attention away from the task and towards themselves. This, she posited, could be a dangerous thing for a teacher to do.
All of this information was enlightening and thought-changing for me. But there was a problem. For all the interesting research I was finding, I couldn’t figure out how to apply it to my teaching. Two years later, I’m still struggling. Why is this so hard?
Maybe the nature of feedback research itself contributes to the trickiness of translating it to my teaching.
Valerie Shute wrote a comprehensive review of the feedback literature. At its end, she includes a series of helpful tables that summarize the major recommendations to educators. Some of the advice clearly doesn’t apply to a k-12 classroom: “Try to avoid delivering feedback orally.” My 3rd Graders aren’t college students; reading comments is often slow and challenging work for them. Other pieces sound suspiciously generic: “Be specific and clear”; “Present feedback in manageable chunks.”
Maybe I’m being unfair. Some of the advice is quite applicable: “Focus feedback on the task, not the learner,” Shute writes, citing Ruth Butler. “Do not give normative comparisons”; “Be cautious about providing overall grades”; “Minimize use of extensive error analyses and diagnosis.”
In a lot of ways, it seems that the advice on what not to do is more concrete than what to do. When it came to figuring out what sort of feedback might be helpful to give, I’ve had to rely nearly entirely on my own ideas. More and more, I’m thinking this might be a limitation of how researchers have, so far, approached feedback.
All experiments are a sort of approximation of reality. The reality that these carefully controlled studies approximate is (what I call) a feedback-poor environment. A feedback-poor environment is one where an instance of feedback is easily identifiable -- it clearly stands out against a foreground of non-interaction.
I studied philosophy in college, and nearly all my class time was spent listening to lectures. Some of my professors even read directly from their notes -- no digressions! Any feedback would come at clearly designated moments: a comment on a paper, a conversation with a teacher, a question after class. These classrooms were feedback-poor.
In contrast, the math classes that I teach are suffused with feedback. This isn’t because I’m some sort of amazing pedagogue. In a k-12 classroom, students are constantly receiving various sorts of feedback. I go over a homework question on the board. A kid checks his classwork against a friend’s. Students ask questions, we give answers. Whether you’re specifically intending to or not, you give a ton of feedback while teaching math to kids. It’s an environment that’s feedback-rich.
What is the problem here? It seems to me that we’ve built a body of knowledge on feedback that makes sense only in the context of feedback-poor environments. It doesn’t always carry over well to feedback-rich environments.
A k-12 teacher is going to give feedback of one sort or another to their students. The crucial question, I think, is what form should that feedback take? What activities open up opportunities for revising your thinking? Are there instructional routines that lead to great opportunities for comments?
Recently, I’ve stumbled on one of these productive routines. The issue was that, sometimes, my students would simply ignore my comments. I’d ask them to revise their work, but they’d do a poor job. I decided that part of the problem was the comments themselves -- they were too cryptic, or otherwise too lengthy. I was trying to develop conceptual understanding in these comments, but I started to feel like this was the wrong way to do it.
I tinkered with my approach, and now I do something different. I start by studying my students’ work and try to locate the most significant gap in understanding. Then, I design a short activity that will introduce the language or visual I want to use in my feedback comments. I’ll run that activity with the entire class, and then return the students’ work, with comments that sync up with the activity. Finally, I’ll ask students to revise their work in class while I circulate the room and answer questions.
Would this work for other teachers, in other contexts? Are there other such routines out there, like this one, that I could use? Does different sort of learning call for different sorts of feedback routines? These are the questions I have about feedback, and I wish researchers would look into them. After a century of research on feedback, it seems to me that we still don’t know what feedback routines are promising enough to be worth testing in the classroom. If researchers wish to do work that is relevant to the life of the classroom, they’re going to have to make some big changes in how they study feedback.