Feedback that guides learning: a model and example for teachers

A grade reports the result of an assessment but can leave the next revision unclear. This guide uses two research works to propose more specific comments and a sheet usable with large classes. The example is illustrative, not a measured SimulAI outcome.

Feedback can help or hinder depending on its information, timing, and how students use it. Hattie and Timperley offer a model for these conditions; no effect is guaranteed for every comment.

Why a Grade Alone Does Not Work

A grade alone reports a result against a criterion but may not explain the next step. Grades also serve institutional purposes. The example below addresses support for learning, not the abolition of assessment.

In Butler’s 1988 experiment, with specific grades and tasks, task-focused comments supported interest and performance more than grades or grades plus comments. The study does not prove how every student interprets a mark or that any qualitative comment will be better.

This does not mean abolishing grades — they serve important institutional functions. It means recognizing that, to promote learning, qualitative commentary is irreplaceable.

The Hattie and Timperley Model: Three Questions Feedback Must Answer

In The Power of Feedback, Hattie and Timperley organize three questions: what is the goal, how does this work compare with it, and what is the next step? This is a tool for planning feedback, not a guarantee of positive effect.

• Where am I going? (Feed-up): Does the student know what the learning goal is? Without a clear target, any feedback sounds arbitrary. Feed-up establishes the destination before any journey of improvement can begin.
• How am I going? (Feed-back): Where is the student in relation to the goal? This is the dimension most of us recognize as “feedback,” but it only makes sense once the first question has been answered.
• Where to next? (Feed-forward): What are the concrete next steps? This is the most neglected dimension and, paradoxically, the most transformative. Good feed-forward turns a response into an action plan.

To use the model, check whether the student knows the criterion, can locate a success or mistake in the work, and has a specific action for the next attempt. Every individual comment need not repeat all three questions literally.

The Four Levels: Where to Focus

Hattie and Timperley distinguish feedback about task, process, self-regulation, and person. These levels serve different purposes by context; they are not a ladder where one comment is always superior.

• Task level: indicates whether an answer meets a criterion and where an error occurred. It may suffice for some corrections; others require a discussion of strategy.
• Process level: describes a procedure, such as comparing data without checking units. It suggests a revision step, but its effect depends on whether the student can apply it.
• Self-regulation level: invites students to check their own work and choose a next step. It is not automatically superior to feedback about a factual error.
• Self level: praise about a trait, such as “you are smart,” does not indicate which procedure helped. The model favors information linked to task and process; it does not show that every personal compliment is harmful.

Choose the level for the observed obstacle. For a factual error, start with the task; for a recurring pattern, discuss the process; for independent revision, invite students to check their own work. Avoid treating personal praise as guidance on the task.

Effective Feedback: Essential Characteristics

The qualities below are review questions for a teacher, derived from the models discussed. This article does not present a separate experiment proving each in every situation.

• Specificity: “Good work” teaches nothing. “Your central argument is strong, but the first two paragraphs do not yet make clear who you are writing for — try opening with a sentence that orients the reader” gives students exactly what they need to act.
• Timing: offer feedback while there is still an opportunity to use it in the next attempt. The best timing depends on content and activity; there is no universal deadline for every comment.
• Focus on the work: “this paragraph is unclear” points to something editable; “you write badly” labels a person. Prefer comments that locate a problem and next action.
• Actionability: The comment must point to a path forward. “Needs improvement” is not enough. “Revisit the transition between the second and third paragraphs: your main idea is not yet connected to the example you introduce” tells students exactly where and how to act.

The Danger of Vague Praise and Feedback That Is Only Positive

“Great!” alone does not say which part of the work can be repeated or improved. Specific praise may make the criterion visible; this does not mean every generic compliment causes harm.

Positive feedback is powerful when it is specific. Compare: before — “Well done, very good!”; after — “Your analysis of the economic causes is very well supported: you drew on three different sources and connected them clearly. This structure could serve as a model for the next sections.” The second version teaches the student to recognize what worked and to replicate that strategy.

The cited sources do not establish that students always read exclusively positive praise as a lack of commitment. If there is a useful next step, state it respectfully and check whether it was understood.

Feedback as Dialogue, Not Monologue

Feedback can include a student response: identify a question, test a revision, and check the result. This is one way to apply the three-question model; the cited sources do not measure a universal superiority for dialogue here.

One option is to ask students to mark their question before reading a comment and then state what they will change. This can make understanding of the criteria visible; calling it “metacognition” does not prove a lasting academic gain in this example.

Asking students to suggest a change may reveal whether the comment was understood. The two works cited here do not show that such conversation always reduces defensiveness or increases receptiveness.

Worked example of usable feedback

In a large class, teachers need to choose where individual feedback is most useful. A collective review of frequent errors and a short rubric are organizing options; their effectiveness depends on use.

Managing Feedback in Large Classes

One of the most legitimate objections to qualitative feedback is time: how do you give individualized responses to 35 or 40 students? The honest answer is that perfect feedback for everyone, every time, is not possible. But there are strategies that maximize impact within real constraints:

• Collective feedback on patterns: After marking a set of assignments, identify the three or four most common errors and address them with the whole class before returning individual papers. This reaches everyone at once and normalizes mistakes as part of the process.
• Selective and rotating feedback: You do not need to comment on everything in every student’s every assignment. Define a feedback focus by stage (for this submission I will focus on argumentation; for the next, on coherence) and rotate which students receive more detailed responses.
• Structured peer feedback: Teaching students to give each other feedback — with clear criteria and a defined protocol — multiplies the class’s feedback capacity and also develops critical evaluation skills in the students themselves.
• Rubrics before submission: share criteria so students can check their own work. Observe later whether the rubric was understood; its presence does not guarantee less rework.

Question tools can display an answer and explanation, but teachers must check the source and assess conceptual errors. This article does not measure time saved or learning gains from automation.

Giving good feedback is a skill that is learned and refined over time. No teacher is born knowing how to do it, and few have received explicit training on the subject. The simplest starting point is this: the next time you write a comment on a student’s work, ask yourself — after reading this, will they know exactly what to do? If the answer is no, the comment is not ready yet.