At 3:47 on Thursday afternoon, Mara Patel, a fictional seventh-grade science teacher, is alone in Room 214 with 28 exit tickets.
Her students have spent the period learning how plants make food. Mara’s prompt was familiar: What was the most important thing you learned today? Explain.
Most responses say some version of “Plants need sunlight.” A few mention roots taking in nutrients. Several define photosynthesis with the same sentence from the textbook. The answers show that students were present and can repeat vocabulary. They do not show whether students understand where a plant’s mass comes from, why leaves need gas exchange, or which misconception needs attention first.
Tomorrow’s lesson is leaf anatomy. Mara has a department meeting at 4:00, so she places the tickets in a blue folder, writes “look at these” on a sticky note, and leaves. On Friday, she starts the planned lesson. By Monday, a student confidently says that plants get their food from dirt. Mara recognizes the idea from the tickets she never used.
The problem was not a lack of care or even a lack of data. Mara collected evidence without deciding what the evidence would change. That is how exit tickets die in drawers: the paper becomes the endpoint instead of part of a teaching decision.
The stack is a planning problem
An exit ticket becomes formative assessment when a teacher uses the response to adjust what happens next. Black and Wiliam’s influential work on formative assessment made this point decades ago, though classroom practice often reduces it to a collection of techniques: a ticket, a thumbs-up, a quiz, a poll. The tool matters less than the feedback loop around it.
Checking for understanding means eliciting evidence that helps you choose the next instructional move. A thumbs-up can tell you that students are willing to signal agreement. It cannot reliably tell you whether they can distinguish a plant’s energy source from its source of matter. A question that asks students to explain a choice gives you more to work with.
A useful test is to finish this sentence before writing the prompt: Because of what I see, tomorrow I will…
If the ending is “know who paid attention,” the ticket needs work. If the ending is “model the difference between energy and matter for the whole class,” “meet with the four students who still reverse the steps,” or “move on and give students a transfer problem,” the purpose is clearer.
This is where many formative assessment ideas become more manageable. The teacher does not need a new routine for every day. The teacher needs a small amount of evidence tied to a real fork in the road.
Mara’s first question was too broad for the decision she needed to make. She was planning to teach leaf structures, but the more urgent question was whether students had a usable model of photosynthesis. She needed evidence about one idea, not a summary of the entire period.
Write the ticket backward from tomorrow
Start with the learning target, then name the likely error. A strong exit ticket asks students to produce evidence that separates understanding from a familiar phrase.
What makes an exit ticket effective? It reveals whether students can perform one specific piece of thinking and gives the teacher a plausible next action. It does not need to measure everything taught that day.
For a next-day adjustment, one strong question or two brief questions are usually enough. Give students about five to seven minutes. More questions can create a misleading result: you may end up measuring writing stamina, reading load, or end-of-day fatigue rather than the concept you meant to check. If you need ten questions, you may be planning a quiz rather than an exit ticket.
The following Thursday, Mara writes a narrower prompt:
A student says, “Most of a plant’s new dry mass comes from the soil because its roots are in the soil.” Do you agree or disagree? Use one observation or model from today to explain where most of that mass comes from.
The prompt requires retrieval and explanation. Roediger and Karpicke’s research on retrieval practice supports asking learners to produce knowledge rather than merely recognize it, although retrieval alone does not guarantee a correct mental model. The explanation is the important part for Mara. It can reveal whether a student knows that carbon dioxide enters through the leaves, is relying on a memorized phrase, or is still treating sunlight as a material that enters the plant.
Mara also decides what she will do before collecting the papers. Her class has 28 students, so she writes three planning rules in the margin: if eight or more students use soil or sunlight as the main source of new mass, she will pause for a whole-class model; if three to seven do so, she will run a short small-group lesson while others work; if fewer than three do so, she will proceed with leaf anatomy and add a transfer question.
Those numbers are not a research-backed mastery cutoff. They are a practical decision rule for one teacher, one class, and one lesson. Setting the rule in advance keeps Mara from moving the goalposts after she sees the answers or treating a class average as more precise than it is.
The ticket also needs a response plan. Mara tells students it is ungraded, but she still expects an explanation. Students who need a sentence frame can use one: I disagree because most of the new dry mass comes from ____, which enters the plant through ____. The language support reduces writing barriers without removing the scientific reasoning.
Read the wrong answer for the idea underneath
After the bell, Mara reads all 28 responses. She does not assign points. She makes three loose groups: a usable explanation, a partial explanation, and a misconception or missing response.
Nine students agree with the fictional student and say the mass comes from soil. Five choose sunlight as the source. Six mention water or carbon dioxide but offer no evidence. Eight explain that carbon from the air enters through the leaves and becomes part of new plant tissue.
The count is useful, but the wording is more useful. The nine soil answers are not identical. Six say roots “absorb food,” which suggests a model of roots as the place where food enters. Three simply repeat that roots are in soil without explaining how mass moves. The five sunlight answers appear to confuse energy with matter. The six thin answers may reflect partial understanding, weak retrieval, or uncertainty about what counts as evidence.
This is a subtle reason to read exit tickets as evidence rather than grades: the most useful response may be the confidently wrong one. A blank response is important, but it is ambiguous. A student may have missed the lesson, run out of time, misunderstood the language, or declined to write. An articulated misconception shows a model that can be addressed. It should not be treated as a character judgment or a permanent label.
Mara chooses one follow-up hinge question for Friday: Which change would add carbon atoms to a growing plant—more soil, more carbon dioxide entering the leaves, or a brighter lamp with no extra carbon dioxide? Choose one and explain why.
Dylan Wiliam has written about hinge questions as quick checks that help a teacher decide whether to move on. Mara can use mini-whiteboards, fingers, or simultaneous cards to see the pattern, then ask for a brief reason. The point is not who answers first. The point is whether the class can now connect the source of matter to the process they studied.
She keeps student names private and shares only anonymous answer patterns with the class. “Several responses treated sunlight as a substance,” is useful. “Jordan still thinks sunlight is matter,” is neither necessary nor fair.
Make the next lesson visibly different
On Friday, Mara spends the first nine minutes responding to Thursday’s evidence. She draws two simple models on the board: one sends material from soil to leaf, and the other shows carbon dioxide entering the leaf. Students annotate which arrows represent matter and which represent energy. Then they compare their annotations with a partner and revise one claim.
She has not thrown away the leaf-anatomy lesson. She has moved a short model-building task ahead of it and changed the opening question. When she introduces stomata, she asks students to connect the structure to gas exchange instead of memorizing a definition. The planned lesson remains, but the route into it has changed.
Before students leave, Mara returns the original responses without scores and asks them to add one sentence: I changed or kept my answer because ____. That second response gives students a chance to correct their thinking and gives Mara a quick check on whether the morning’s explanation helped. The exit ticket has become a correction loop rather than a collection ritual.
Students also notice the difference. Their answers led to a model, a new question, and a chance to revise. Over time, that makes the request to explain feel less like paperwork. The teacher has shown that their thinking affects the next lesson.
When this routine fails
This routine fails when the next class has no meaningful connection to the ticket. If Friday begins an unrelated unit, forcing Thursday’s responses into the lesson creates busywork. Use a short pre-assessment at the start of the new unit instead, or wait until the end of the current sequence when the evidence can inform a real choice.
It also fails when the response asks for more writing than the instructional question requires. A student with a useful idea may produce a weak paragraph because of dysgraphia, limited English proficiency, fatigue, or a processing difference. Accept a labeled diagram, brief oral explanation, sentence frame, or dictated response when those formats fit the learner’s accommodations. Do not infer conceptual failure from grammar alone.
The routine is a poor fit when the teacher cannot read the answers in time. A daily exit ticket that creates an hour of evening marking will not last. For a class of 28, Mara can read one- or two-sentence responses in roughly ten minutes because she is looking for two or three codes, not correcting every comma. Longer answers require a different schedule. Two carefully chosen tickets a week are better than five abandoned ones.
Finally, one prompt cannot establish mastery. A student may guess correctly, misunderstand the question, or know the answer in one context and fail to transfer it. Research on formative assessment supports responsive use of evidence, but it does not provide a universal frequency, response threshold, or magic prompt. Mara’s ticket informs Friday; it does not settle what every student knows about plant biology.
By the following week, Mara still has a tray on her desk. The difference is that each paper has a planned destination. Before the next lesson, she writes the target, the likely misconception, the response rule, and the specific change she will make. After class, she sorts the evidence and records one sentence about what she will adjust. The next morning, she names the pattern without naming students, makes the adjustment, and checks the idea again.
Try that routine with one lesson this week. Before your next planning period, complete this sentence: If students show _____, tomorrow I will _____. Then write one exit-ticket prompt that can reveal the blank in five minutes. If you cannot name the action, leave the ticket off the lesson.