The care and feeding of helpful feedback

This article originally appeared earlier this week at Grading for Growth, a blog about alternative grading practices that I co-author with my colleague David Clark. I post there every other Monday (David does the other Mondays). Check the end of this post for some extra thoughts that don't appear in the original.

Click here to subscribe and get Grading for Growth in your email inbox, free, once a week.

---

In one of the earliest posts on this blog, we took a look at what all of the different alternative grading systems out there seem to have in common. This was not only one of our earliest posts but, as it turned out, one of the most influential. The centerpiece of that post was something we called the Four Pillars of Alternative Grading:

In discussions that David and I have had with others since that post, and in our own thinking and writing, this framework has been really helpful in understanding what alternative grading is all about — indeed what any sort of grading is about. It’s been especially useful for instructors new to the concept of alternative grading, by helping them keep focus on what really matters and not get lost in the weeds.

The Four Pillars concept has been so helpful, in fact, that I’ll be spending my next three posts, over the next several weeks, going into depth on each of the last three — explaining what each pillar entails and how grading systems may or may not implement the pillar, and giving practical advice on how to enact each one.

My post from two weeks ago, on Clearly Defined Standards, got us started. Let’s take a look at the next one: Helpful Feedback.

What is feedback? When is it helpful?

Let's define terms first. By "feedback", we mean evaluative information about the outcome of an event or action that is given back to the source of that event or action.  In diagram form, the feedback process looks like this:

The feedback loop

This diagram should feel very familiar in the context of learning. When you’re learning something, you do something to learn. That “something”, represented here by the box with the capital “P” in it, takes inputs — time, energy, existing knowledge — and results in something new, an output. Then someone or something evaluates it, and that’s then "fed back" to the you, the source. Then, it’s coupled with new inputs to the same process, where the new outcome is evaluated again and the information fed back. And so on, until the output passes some critical condition.

In this model and definition, the purpose of feedback is iteration. The feedback is not just data, but information that the learner is then intended to use, along with new input, to loop through the process again. Therefore feedback is helpful when, and to the extent that, it helps learners iterate productively and make steps toward meeting that stopping condition. In other words — feedback is helpful insofar as it helps learners to grow.

Helpful feedback and grading

What makes helpful feedback tricky isn't the "feedback" part but the "helpful" part. Unhelpful feedback is easy to give — it happens all the time. But being helpful requires work.

Traditional grading systems typically give feedback in two forms: Points, and annotations or side notes. For example, suppose Alice takes a 20-point quiz in one of her classes, consisting of five 2-point multiple choice questions and two 5-point short-answer questions. Two of her multiple choice answers are incorrect, and one of her short-answer responses wasn’t clear. The instructor puts an “X” through the wrong answers on the multiple choice questions and circles the right ones; and strikes through the unclear portions of the short-answer response and writes “Not clear” in the margin. The feedback given to Alice is the score, 13 out of 20, along with the markings and the margin comment.

How helpful is this feedback?

• The score itself (13) is not very helpful. It contains no information about Alice’s inputs or outputs. It only says she has earned 65% of the possible points. But as to what Alice is supposed to do about it, the score is silent. Is 65% bad? It sounds bad, but maybe that’s average, or acceptable given the setup of the course. Alice has no way to know — and no way to know how to make it better.
• The X’s and circles on the multiple choice items are a little more helpful, because at least they pinpoint the issue. But the helpfulness ends there, because since the right answer was revealed, Alice is now no longer in a feedback loop. She’s been boxed out of any questions she could explore to address the shortcomings in her knowledge. It would have been more helpful just to mark the wrong answers and stop there.
• The annotations and margin notes on the verbal response are about as helpful as the X’s and circles on the multiple choice items. The phrase “Not clear” is basically just an “X” through the work. There’s no details about what was not clear, why it wasn’t clear, and suggestions or questions that would help Alice clarify it.

The second two objections are academic anyway, since research suggests that when given both points and verbal feedback on a task, students will ignore the verbal feedback and focus solely on the points, which we’ve said are definitely unhelpful. Even worse, traditional grading rarely in my experience offers anything like retakes of quizzes and so on, so there is not even any feedback loop from which students are excluded. There is one “retake”: The final exam.

So, not very helpful. What Alice really needs, if she is to grow, is a feedback loop in which she can participate, and feedback that sets her up to engage successfully with that loop.

Some instructors who practice traditional grading do give helpful feedback; but it’s not the default, definitely not the norm, and in my experience the helpfulness tends to disappear as the semester grinds onward. But all the alternative grading systems you’ve learned about here and elsewhere, have helpful feedback at their core. And that feedback tends to look the same no matter which system or combination of systems you choose.

How to give helpful feedback

So what does helpful feedback look like, and more importantly, how do you give it? As is often the case, our colleagues in the business world have a wealth of experience from which we can learn, since feedback is considered one of the cornerstones of agile business practices.

This Harvard Business Review article by Cynthia Phoel gives a useful guide on giving good feedback. Here’s a summary, with my translation from business-speak to academic language.

• Focus on learning outcomes and growth. Feedback isn't the same thing as "criticism". When framed in terms of specific clearly defined standards or learning goals, feedback presents "an opportunity to solve a problem rather than to criticize" as Phoel puts it. So keep feedback specific and focused. Feedback that is focused on specific goals puts the emphasis on problem-solving and depersonalizes the issues that are present. For example, when grading an essay or a mathematical proof, compare a vague and pejorative comment like “Your argument here doesn’t make any sense” with the more specific and focused comment “The argument here has incorrect assumptions, and there is a logical error in the third sentence”. Remember you’re grading the work, not the learner!
• Ask questions. Phrasing feedback in the form of questions to consider is an excellent way to invite learners into a feedback loop and help them take ownership of their growth. For example, when grading an essay or proof where the learner sets out an argument with incorrect assumptions, instead of just telling them what was wrong such as “Your assumptions here are incorrect for this kind of argument”, turn it into a question: “These assumptions don’t seem correct; What assumptions should you be making for this kind of argument?”
• Give feedback often. You should definitely give feedback on assignments that figure into a student’s grade, and it’s better to give too much than too little. But also, don’t keep feedback restricted just to the graded items. You can give feedback any time there’s an opportunity: During a class activity, while waiting for class to begin, while class is ending, during office hours… As Phoel says, “Praise good performance right away. When negative feedback is required, talk with the [student] within 24 hours".
• Keep your feedback data-driven. This is an aspect of making feedback specific as I mentioned above. Phoel suggests that the goal of a feedback session — whether it’s face-to-face or written — is to gather evidence that will help you describe the specific issues or behaviors that the student has or hasn’t done, the impact of those issues or behaviors, and what you want the student to do differently (perhaps phrased as a question, like above). What’s unhelpful is feedback about an issue that isn’t supported by any particular form of evidence.  Therefore it’s important to tie feedback to specific items in a student’s work, and avoid generalizations like “You always…” or “You never…” or referring to something that isn’t present in the item the student has provided.
• Don't assume that you're right. Sometimes when grading, you know you’re right. But as your assignments climb higher in Bloom’s Taxonomy, it’s more likely that your evaluation of student work could be wrong, and that what looks like an “issue” in student work is more of a disagreement. So stay humble and helpful, realizing in such cases that you may not have the full picture.

In addition to those basic principles, I’d add the following life-hacks that have been helpful to me in giving helpful feedback:

• Be mindful of your own state of mind, and avoid grading when angry, frustrated, tired. The hope is that when we grade, we will be like Gandalf — a perfect mixture of wisdom, kindness, and knowledge that dispenses exactly the right feedback in exactly the right amounts. But even Gandalf had bad days, and when you’re depleted, we tend to lash out, even at students when we grade. I think we’ve all done it, and it’s never good. So if you need to break from grading for an hour, an evening, a day, or a weekend — do it, if it will make your feedback helpful, and don’t put yourself in a position you might regret.
• Watch your language. Along the same lines, be mindful of what you say or write. Don’t use language that’s condescending, mansplaining, or angry. Some specific items to watch are second person pronouns, exclamation points, and capitalizations. For example compare “You don’t have the correct set of assumptions for a contradiction proof” with “The contradiction proof doesn’t have the correct set of assumptions”. Yes, the student needs to own the error and fix it. But we don’t have to “dunk” on students in the process.
• Text snippet tools are your friends. I use TextExpander for frequently-used comments. This saves a lot of time — for example to enter in the phrase “Your work on Problem 1 shows progress in a useful direction, but there are also some gaps or errors that you will need to address in a revision of that problem” I just enter in the snippet ;;needsrev and the software does the rest. But not only is it a great time saver, it also lets me compose the feedback when I am feeling calm and friendly, and then later if I can enter that feedback even when I am tired or grouchy and therefore likely to say something I’ll regret.
• Try audio or video feedback instead of/in addition to written feedback. Some students respond more favorably to short video or audio snippets that give feedback than they do to text feedback, which can lose much of the emotional content of face-to-face communication (as we all know too well from bad emails). You might explore using tools like Loom for giving quick, portable video feedback or your computer’s audio recording software for lower-bandwidth audio-only clips — and seeing if it helps.

Feedback loops are the “roof” that the Four Pillars supports, so it’s crucial that we give feedback that’s as helpful as possible if we want students to grow. That helpfulness extends to the marks that we give students on their work and in the gradebook — and that concept (Marks Indicate Progress) is the subject of the next post in this series.

---

Bonus extra thoughts

• You might disagree that learning something always involves doing something, i.e. that there is a "process" in the feedback loop corresponding to every learning event we experience. My view corresponds to the constructivist point of view and I think there is more than enough experimental evidence and way more than enough anecdotal evidence in my experience to justify that view. But for example, if you are listening to a podcast and some interesting fact comes across, you learned something – but was there a process? I would say yes, and the process consists of whatever it is you do to "file it away" – making a note, tweeting about it, reinterpreting it in your own mind so it makes sense to you. The learning doesn't just happen on its own. And, the higher up Bloom's Taxonomy we go, the less persuasive the countering points of view become. In my opinion.
• This pillar seems to me like one that, if you are thinking about alternative grading but not yet ready to go all-in with it, you could just focus on this one thing in your existing courses and it would improve learning outcomes greatly. That is, if you could do only one thing with your next course that has a low barrier to entry and would produce the biggest gains, simply giving more and higher-quality feedback would be a good thing to focus on. I think that if an instructor did this – just focus on better feedback – the benefits of having clear and measurable standards, moving from points to more informative marks, and allowing reattempts without penalty will become clearer.
• If you're a Mac user, there is now a text snippet tool built right into the OS. Haven't used it myself so I can't compare it to TextExpander, which is cross-platform.