What is Formative Feedback and How is it Different From a Grade?

A grade tells a student where they landed. Formative feedback tells them what to do next. These are not two versions of the same information; they answer entirely different questions, and confusing them is one of the most persistent errors in higher education assessment design.

Formative feedback is information delivered during the learning process, intended to close the gap between a student's current understanding and a specific learning goal. A grade, by contrast, is a summative judgment: a symbol attached after the fact that certifies a level of performance. One is a tool for revision. The other is a record.

Why the Distinction Matters More Than It Seems

Research on this question goes back decades, and the findings are less tidy than most education blog posts suggest. A frequently cited 1988 study by Ruth Butler found that students who received only task-related comments improved more than students who received grades alone, or even grades paired with comments. When a grade was present, students tended to read it first and skip the reasoning underneath it, regardless of how detailed that reasoning was.

Thomas Guskey, a research scholar who has spent years reviewing this literature, cautions against oversimplifying Butler's findings into a blanket "no grades, only comments" rule. His analysis for Kappan Online found that the impact of feedback varied by subject area, generally proving more effective in language arts than in math or science, and by how it was implemented, with structured, supported approaches outperforming ad hoc ones. The real variable driving outcomes was the substance of the feedback, not just its format.

That nuance matters for how universities should read the research: the case against grade-only feedback is not that grades are inherently harmful, but that a number attached to work carries no information about what to change. A student who receives a 68 knows they underperformed. They do not know which part of their reasoning was wrong, why it was wrong, or what a corrected version would look like.

What Formative Feedback Actually Requires

Effective formative feedback is not simply detailed feedback. A systematic review of formative assessment and feedback practices in higher education, published in Review of Education, found that the research base is still thin relative to how widely these practices are recommended, and that the specific implementation, not just the presence of feedback, determines whether learning improves. One of the reviewed studies found that regular low-stakes quizzing had a substantial positive impact on final grades compared to a no-quizzing condition, with gains visible both on the quizzes themselves in the short term and on the end-of-term exam later on.

That finding is consistent with a broader principle in the feedback literature associated with John Hattie and Helen Timperley: useful feedback answers three questions for the learner. Where am I going. How am I doing relative to that goal. What should I do next. A grade answers none of these. It communicates a final position without a direction.

This is where the distinction between assessment for learning and assessment of learning becomes practical rather than theoretical. Assessment for learning treats the act of assessing as part of the instructional process itself, feeding directly back into what the student does next. Assessment of learning, which is what a grade represents, happens after instruction is over and simply reports the outcome.

Why Most Formative Tools Still Behave Like Grades

Many platforms marketed as formative assessment tools still produce a score at the end of a quiz or a poll. That score functions exactly like a grade: it tells the student a number without explaining the misconception that produced it. Calling a tool formative because it happens mid-semester, rather than because of what it actually reveals to the student, is a category error that has become common in ed-tech marketing.

A genuine formative assessment platform should surface why a student got something wrong, not just that they got it wrong. Axiom Flow goes beyond the formative assessment platform model by running teaching-based assessment: the student's explanation itself becomes the evidence, and the resulting misconception-based evaluation shows exactly where understanding broke down. The distinction is not cosmetic. It is the difference between a score and a diagnosis.

In Axiom Flow's process, this separation is structural rather than stylistic. Atlas, the platform's assessment designer, generates a configurable set of misconceptions from the course material and maps one exam question to each. Sam, an AI student, starts every session holding those misconceptions and has no independent way to check what is true. The student's job is to teach Sam and correct what he believes. This teaching phase is unscored: it is pure formative feedback, delivered in real time, because Sam reflects back exactly what he understood from the explanation, and where his understanding is still off.

Only after teaching ends does Atlas step in to evaluate Sam's answers to the exam questions, producing a score and a gap report showing which misconceptions were resolved and which remain. The formative phase and the summative phase are kept distinct on purpose. A student gets to work through confusion without a grade attached to the process of getting there, and the eventual score reflects understanding that was actually built, not guessed at under time pressure.

What This Means for Assessment Design

The practical takeaway for instructional designers is not "eliminate grades." Grades serve a certification function that institutions genuinely need. The takeaway is that formative feedback and grades should not be asked to do the same job. If a course's only feedback mechanism is a graded quiz, students are not getting formative information; they are getting a preview of their transcript.

Rebuilding a course around actual formative practice does not require eliminating grades. It requires separating the moment a student is allowed to be wrong, and learn from being wrong, from the moment their performance gets recorded. That separation is what the research keeps pointing back to, and it is the design principle worth building around.