Please don’t leave - but I am actually going to start this post with a quick personal anecdote. I started wrestling in early elementary school and it quickly took over my identity. I’ll skip all the boring details in the middle - but it led me to get VERY interested in the idea of failure in an athletic competition. During my graduate degree in special education at the University of Maryland (Go Terps! 🐢) I completed a research methods course on phenomenology1. Phenomenology is an approach to qualitative research focused on building a deep understanding of the lived experiences of people related to a specific phenomenon. As part of the course I wrote a capstone paper focused on how wrestlers interpreted and perceived the concept of failure and what their lived experience were related to this in their athletic careers. I’ll not bore you with the details – but a significant finding was that failure did live deep within the lived experiences of the people I interviewed and this was the lasting taste with which they left the sport (same for me).
Awkward Transition
Now, of course “failure” can be used in a different sense when discussing academic skill acquisition. Learning new “stuff” will eventually lead to an experience where the learner encounters something they are asked to do that is not known or that they are currently doing that is incorrect. These experiences could perhaps be described as current “failure” in performing a task. So the questions I have are
Is failure inevitable?
Is failure (in isolation) beneficial to student learning?
Is Failure Inevitable?
Yes.
Is failure (in isolation) beneficial to student learning?
This is the heart of the conversation. “Failure” - in the way I defined above is inherently going to happen in the learning process. So how should failure be used in the instructional process.
I had a conversation with a friend (Thanks Boston) recently and they asked the question, “How do you respond differently if you did something correctly versus incorrectly?” I provided a longwinded answer. He then said, “No, you feel the exact same way until you get some semblance of feedback on whether that response was correct or incorrect.” That stopped me in my tracks.
Failure, in isolation, does not actually promote learning if there is not a feedback loop letting you know what was provided was incorrect. Clearly, in one context a problem could be presented that is WAY advanced for a student so they may not even attempt the problem - this could be counted as “failure.” And honestly, no feedback was even provided that suggests this was a failed attempt at the problem. However, in many other contexts, students solve math problems with higher or lower self-confidence in whether it was solved correctly - yet the feedback loop is not provided until their is a chance to self-evaluate their solving of the problem with different solution plans. If no feedback is provided, students can continually solve the same types of problems until the cows come home - probably would not result in additional learning. I’m going to take you to a nerdy, weird place - statistical programming :). I am attempting to learn how to use the program R more proficiently. I can code until the cows come home different models for meta-analysis. If I NEVER click “run program” then I can continually be writing code that is inaccurate yet I think is correct. It is not until I click “run” and error messages pop up everywhere that I get a feedback loop that something I am doing is wrong.
So in sum, just failing doesn’t promote learning. Something else needs to be included.
How does failure become useful?
If the failure is used as a learning experience. A simple model could be (1) failure, (2) teacher explanation (or perhaps worked example on problem), (3) presentation of similar problem, (4) student solves correctly (please read caveats here2). The initial error is being used as a way to serve as a signal or stimulus for an incorrect response. This can help children learn to discriminate between incorrect versus correct responses.
This is an aside for the researchers reading this. The second class the professor asked what we hoped to get out of the course. I said something to the effect of understanding how children respond to math interventions and using this to generalize out children’s experience to better design interventions. In a kind way the professor said, “that is not the goal of phenomenology.” I quickly realized qualitative research was an entirely different paradigm than my prior learning history and how I thought.
This simple model works excellently for declarative facts. If there is an error in a procedural task - a task analysis identifying where in the chain the error occurs can be helpful (i.e., step 5 of 9 was incorrect). You can then help students work through a scaffolded chain of problems to clarify this error/misconception. For errors at more of an abstract or conceptual level a deeper level of explanation might be needed.
thanks Corey, gr8 analysis. Another aspect is often we have the kids for a year or more, in my case I will have Year11's then take them for Year 12 the following year. Most of the time we use worked examples. Also, most text books are based on worked examples. So over the course of a year just using that approach can be boring and disengaging. I try to use "productive failure" sparingly to break up the monotony. An example in our Y11 Probability unit is a simple activity of rolling 2 die and subtracting them then creating a game where kids lay bets. Most don't realise what is going on and don't know the theory behind the game, the subtle change of subtracting rather than adding die confuses them. After trialing and playing the game I then go through the lattice diagram and how it helps to analyse the game. Then show the simple probability calculations afterward - as per your steps above. So the change of strategy does (anecdotally) engage the kids. Also the kids seem to remember, as a year or so later, they do state exams where these type of problems are given. Kids often report back to me they remembered that game we played and "aced" those questions. So in the context of trying to engage the kids over a year or more, this strategy is useful.