Class assessment: how to know the lesson worked before the exam tells you

You teach a session. It goes well. Students nod, a few ask decent questions, nobody looks lost. Six weeks later you mark the exam and discover half the cohort never understood the core idea.

The teaching wasn't the problem. The feedback loop was. You got your evidence at the end of the term, when the only thing left to do with it is write next year's syllabus.

Class assessment closes that gap. It means checking what students actually understood inside the class you're already teaching, while you can still change the next forty minutes.

Class assessment is not a smaller exam

The word "assessment" sends most people straight to grading. That's the trap.

A class assessment has one job: tell you what to do next. It doesn't have to be graded, recorded, or even collected. If it produces a number for the gradebook and nothing else, it isn't doing this job — it's just an exam in a smaller box.

Three things separate one from the other:

• It runs during the lesson, not after it.
• It reaches every student, not the three who always raise their hand.
• It changes something — your pacing, your next example, the group you sit together.

Drop the third one and you're collecting data for its own sake.

If you want the formal version of this distinction, it's the line between formative and summative assessment: one shapes the teaching while it's still happening, the other certifies what happened after the fact. Class assessment is the formative side, done in the room, on the clock, with thirty people waiting.

Five checks that fit inside a class

None of these need a platform. All of them survive a room of ninety students.

1. The exit ticket

Last three minutes: one question on the board, one answer per student, on paper or in a form. Make the question diagnostic, not decorative. "What was today's session about?" tells you nothing. "A client argues X. Give one reason they're wrong." tells you everything.

Read them in the corridor. You'll know by the time you reach your office.

2. The misconception poll

Before you explain the hard concept, put up four options — one right, three that each encode a specific wrong mental model. Students vote. You now know which wrong model is in the room, and you can teach against that one instead of the one you assumed.

The value isn't the score. It's knowing which mistake you're dealing with.

3. Show the reasoning, not the answer

Ask for the steps, not the result. "How would you decide?" beats "What's the answer?" because a correct answer hides a broken method, and a broken method is exactly what you need to catch.

This is the check that scales worst and teaches most. Use it on the concept that everything else depends on.

4. Peer explanation

Students explain the idea to the person next to them. You walk the room and listen. Two minutes of eavesdropping on six pairs tells you more about the state of the room than an hour of nodding.

You're not assessing the explanation. You're counting how many of them can produce one at all.

5. The deliberate redo

Hand back a short piece of work with no grade and one instruction: fix the part that doesn't hold up. What they choose to fix shows you what they think "good" means — which is usually the real gap.

The part everyone skips

Running the check is the easy half. Acting on it is where class assessment either works or quietly becomes a ritual.

Decide in advance what you'll do with each outcome. Before the poll, not after:

• Most of the room has it. Move on. Don't re-teach for the three who don't — see them after.
• It's split. Pair them up deliberately, one from each side, and let them argue it out.
• Almost nobody has it. Stop. Re-teach it now, differently. Your plan for the rest of the class

is already wrong; the only question is whether you find out today or in November.

That last one is the hard call, because it means abandoning a session you prepared. Make the rule before the class and you won't have to make it under pressure.

What it costs you, and how to keep it cheap

The honest objection to all of this is time. You have content to cover, and a diagnostic check that eats ten minutes is ten minutes of teaching gone.

Three ways to keep the cost down:

• One check per session, not five. Pick the concept that everything downstream depends on, and

check only that.

• Don't grade what you don't need to grade. Most class assessment works better ungraded.

Students tell you what they don't understand only when saying so is free.

• Reuse the instrument. A good misconception poll works every year, because the wrong mental

models are remarkably stable. Write it once.

The reuse point is where most of the saving lives. The first version of a diagnostic question takes real thought — you have to know the wrong answers well enough to write them convincingly. After that it's a permanent asset, and the second cohort costs you nothing.

Start with one class

Don't redesign your whole course. Take the session that students historically struggle with most — you already know which one — and add a single exit ticket to the end of it.

Read the answers the same day. If they surprise you, you've just found something the final exam would have told you in December.

Then do it again next week, with the same question. The second run is where class assessment earns its keep: you're no longer measuring the cohort, you're measuring your own teaching against the same instrument twice. That comparison is the closest thing a teacher gets to evidence that a change worked.

---

Quindaria builds class activities and rubrics from your own course outcomes — including the short diagnostic tasks that fit inside a session. [Try it free](/sign-up).