Formative Assessment in Math: Quick Checks That Actually Change Tomorrow

There is a stack of exit tickets in a drawer in every school building. Collected Thursday, skimmed Sunday, acted on never. Nobody did anything wrong: the ticket got written, the day kept moving, and by the time anyone looked closely the class had moved on too.

A check that doesn't change what you do tomorrow isn't an assessment; it's paperwork with a nicer name. So the useful question isn't what your exit ticket should look like. It's what you'll do differently because of it, and whether you decided that before you collected it.

Why the principles matter more than the format

Most formative assessment advice is about instruments: ticket templates, thumbs up or down, whiteboards, four corners. Instruments are the easy part, and you probably already have five. What breaks is the loop, the distance between the signal arriving and the teaching changing.

It's worth knowing the evidence here is thinner than the enthusiasm. The What Works Clearinghouse guide on using student achievement data recommends making "data part of an ongoing cycle of instructional improvement," and rates that recommendation, along with its other four, at Minimal Evidence (IES/WWC). That isn't a reason to skip it. It's a reason to keep the loop small, cheap and in your own hands rather than buying a system that promises to run it for you. Three principles carry most of the value.

Principle 1: a check is only formative if it changes tomorrow

Decide your response before you collect the data. 

Before handing out a quick check, name the three moves in advance: if a third of the class misses it, you reteach it tomorrow to everyone; if five or six miss it, you pull a group; if two miss it, that's a two-minute conversation each while the rest start their work. Now the stack can't become paperwork, because each pile already has a decision attached.

This is also why the timing matters more than the polish. Exit tickets earn their keep when the results "inform your instruction for the next day" (Edutopia), which means a scrappy check you read tonight beats an elegant one you read next week.

Principle 2: check one thing, and pick the thing that shows thinking

Narrow beats thorough. The most useful quick checks ask "only one question" and are "simple and had an answer that could be thought of quickly" (Edutopia), because a check covering five objectives tells you a student struggled without telling you where.

Then make that one question do real work. Assigning two or three problems that reveal reasoning beats ten to twenty that only produce a score (Edutopia), and the difference is usually in the prompt. Instead of asking a 2nd grader to solve 8 + 5, ask them what they'd say to a classmate who answered 12. Instead of six problems adding unlike denominators, give a 5th grader one, plus "how did you know what to change the denominators to?" The second version takes the same minute and tells you something a column of checkmarks can't.

Principle 3: read the wrong answers, not the score

Sort by what the error was, not by how many were right. "Correct answers can mask confusion, just like incorrect answers can hide understanding" (Edutopia), so a number correct is the least informative thing on the page.

Two students who both got three out of five may need opposite lessons. One lined the digits up wrong and understood the math fine; the other added the denominators. Grouped by score they land together and one of them gets a lesson they don't need. Grouped by misconception, each gets eight useful minutes. The misconception, not the percentage, is the unit you plan from.

How to apply this in five minutes this week

Read the checks once, sorting into three piles instead of grades:

• Got it. Give an extension that goes deeper on the same math rather than forward into next week's content.
• Almost. Usually a procedural slip, not a concept gap. One targeted hint, a partner, or a second attempt tomorrow.
• Not yet. Tomorrow's small group. Eight minutes on the specific thing the errors point to, then straight back into the grade-level task.

Sorting takes about as long as reading the stack once, because it is reading the stack once. The only change is that you're sorting into decisions instead of scores. Progress monitoring and using the data to decide something are two of the four essential components of a multi-tiered system (MTSS Center), and at the classroom level this is what they look like on a Tuesday.

Where the data can come from without the grading pile

The honest obstacle to all of this is labor. The loop works when reading it is cheap, and it collapses when every check costs you an evening. That's the part software genuinely helps with, and it's a slot Boddle fills. Assign a quick check on a specific standard, and the assignment reports and learning gaps reports tell you who missed what before you've marked anything, so the sorting step starts already done. Practice doesn't cost the student anything either, since there's no score to fall backwards down, and students are markedly more honest on a check that can't hurt them. Read-aloud covers the question and the answer choices, so what you're measuring stays math rather than reading. It's free for teachers and students, with every character earnable without paying, so the data doesn't arrive only for the students whose families upgraded. Teachers in a 2021 LEANLAB study described the useful half plainly: Boddle "helps me catch up with my students...and which skills need to be taught better."

The limit is specific and worth respecting. A report tells you who missed what. It does not tell you why. An auto-scored item can show you that eleven students missed the fraction question; it can't show you that four of them added the denominators while the others never found a common one. That distinction is Principle 3, and it still comes from looking at the work or asking the student a question. Use the reports to find the eleven in seconds, then spend your minutes on the part only you can do.

Frequently asked questions

What's the difference between formative and summative assessment in math? Formative assessment happens during learning and exists to change instruction: a quick check, an exit ticket, a question you ask mid-lesson. Summative assessment happens at the end and exists to report learning: a unit test, a benchmark, a state test. The practical test is what the result changes. If it adjusts what you teach tomorrow, it's formative. If it records what a student learned, it's summative. The same question can be either, depending on whether anyone acts on it.

How often should I use exit tickets in math? Most teachers land on two or three a week rather than daily, which is enough to catch a misconception before it sets without creating a stack you can't read. Frequency matters less than the follow-through: one check a week you act on beats five you file. Keep them to a single question, respond quickly rather than grading later, and hold onto the notable ones (Edutopia). If a routine starts feeling like paperwork, cut the number and keep the loop.

Related content

• Differentiating Math Without Writing Five Lesson Plans (what to do with the three piles)
• Math Centers and Rotations: How to Run Them (where the follow-up small group runs)‍
• 4th Grade Math / 5th Grade Math practice hubs

Close

Formative assessment gets treated as a genre of paperwork and it's really a habit: ask the smallest honest question, read it for thinking instead of scoring, and know your next move before you collect. Everything else is format. Create work students will actually do, at the level each of them is really on, free for everyone in the room, with you deciding what happens next.

Related articles:

• Differentiating Math Without Writing Five Lesson Plans
• Math Centers and Rotations: How to Run Them (and Ideas That Work)
• Low-Pressure Math Practice: Math Games That Don't Punish Wrong Answers
• 4th and 5th Grade Math: Why It Gets Tougher, and What Real Practice Looks Like‍
• Creating Assignments in Boddle: Targeted Practice Made Easy