Generating Formative Assessments for Reading With AI
A formative reading assessment isn't a graded test — it's a quick, in-the-moment check that tells you whether a student is comprehending a text right now, while there's still time to adjust the lesson before moving on. AI can generate these checks fast enough to keep pace with a lesson's actual rhythm, but only if the prompt asks for the right grain size: small, frequent, ungraded, not a scaled-down quiz.
Quick Answer: AI-generated formative reading assessments work best as short, ungraded checks — a single comprehension question, a quick retell prompt, or a vocabulary-in-context check — embedded at natural pause points in a text, rather than a full comprehension quiz saved for the end. The goal is minute-by-minute information, not a grade.
This guide covers how to design formative reading checks that give you real-time comprehension data, where to place them in a text, and how this format differs from the true/false and summative formats covered elsewhere in this series.
What Makes an Assessment "Formative" for Reading
Formative assessment is defined by its purpose and timing, not its format — it's assessment used to adjust instruction while learning is still happening, checked minute-by-minute or day-by-day rather than at the end of a unit (Wiliam & Thompson, 2007). A formative reading check can be as simple as a single question asked mid-chapter, as long as the answer changes what you do next.
This distinguishes it clearly from a summative check:
| Dimension | Formative | Summative |
|---|---|---|
| Purpose | Adjust instruction in real time | Evaluate learning at the end |
| Timing | During reading, at natural pause points | After a full text or unit |
| Grading | Ungraded or very low stakes | Graded, often high stakes |
| Length | 1-3 quick items per check | A full quiz or exam |
The National Council of Teachers of English has emphasized that comprehension monitoring works best when checked close to the moment of reading, not reconstructed from memory well after the fact (NCTE, 2019). That's the practical case for formative checks over waiting for an end-of-chapter quiz.
Three Formative Check Types That Work for Reading
Three formative check types cover most reading-comprehension needs without requiring a full quiz: a quick comprehension question, a retell prompt, and a vocabulary-in-context check — each targets a different part of what "understanding a text" actually means.
Quick Comprehension Questions
A single, direct question about what was just read, answerable in one sentence or less. These work best asked at a natural stopping point — the end of a chapter, a scene, or a clearly marked section — rather than saved until the whole text is finished.
- "What problem is the main character facing right now?"
- "Why do you think the author included this detail?"
- "What just changed in the story?"
Retell Prompts
Asking a student to briefly summarize what just happened tests whether they're tracking the text's structure, not just isolated facts. A retell prompt can be spoken (a 30-second partner share) or written (2-3 sentences), depending on your time budget.
Vocabulary-in-Context Checks
A quick check on whether a student can infer an unfamiliar word's meaning from surrounding context — a core reading strategy distinct from simple definition recall. "Based on how it's used in this sentence, what do you think reluctant means?"
Placing Checks at Natural Pause Points
A formative check embedded at an arbitrary page number interrupts reading flow; one placed at a natural pause point — a chapter break, a scene shift, a clear plot turn — fits the text's own structure and produces a more honest comprehension signal.
For a class working through a chapter book or a longer article, a reasonable placement pattern:
- After the opening section, checking whether the setup (characters, setting, initial conflict) landed
- At the midpoint or a major turn, checking whether students are tracking the developing complication
- Near the end, before revealing the resolution, checking prediction skills — a strong formative move, since asking what a student expects to happen next reveals different information than asking what already happened
You could use EduGenius to generate a set of formative check questions tied to specific pause points in an assigned text, holding a class profile's grade level constant so question complexity matches what the class can reasonably handle mid-read.
Structuring the Prompt for Formative, Not Summative, Output
A vague "make comprehension questions for this text" prompt tends to generate a batch of quiz-style items meant to be answered all at once at the end — the fix is specifying check-in points and ungraded, low-stakes framing explicitly.
A well-structured formative prompt should include:
- The specific text or passage, or at minimum a clear description of its content and structure
- Natural pause points where checks should be placed, named explicitly if you already know the text's structure
- Check type per point — comprehension question, retell prompt, or vocabulary check, varied rather than repeating the same type throughout
- Explicit ungraded framing, so generated language doesn't default to quiz-style phrasing ("Select the best answer") when a conversational check ("What do you think just happened?") fits the formative purpose better
Using What You Learn From a Formative Check
A formative check that doesn't change anything about the next few minutes of instruction isn't really formative — it's just an ungraded quiz. The value comes from acting on the signal immediately, while the class is still working through the text.
- If most students answer confidently and correctly: move forward at the planned pace
- If a few students struggle: a quick small-group re-read or a clarifying question to the whole class, without stopping the lesson entirely
- If most of the class struggles: that's a signal to slow down, re-read the section aloud, or explicitly model the comprehension strategy being tested
This is the practical difference formative assessment is designed to create — a decision point embedded inside the lesson, not a data point collected for later.
What to Avoid
Grading Formative Checks
Attaching a grade to a formative check changes its purpose and its honesty — students answer differently when stakes are attached, which undermines the real-time comprehension signal you're trying to get.
Saving All Checks for the End of the Text
A check placed only after the full text is finished can't inform instruction while students are still reading it — that's a summative check wearing a formative label.
Using Only One Check Type Throughout
Repeating the same question format (always literal recall, for example) misses other dimensions of comprehension, like inference, prediction, and vocabulary-in-context understanding.
Treating a Weak Formative Result as a Final Judgment
A struggling response to one formative check is information for adjusting the next few minutes of instruction, not a fixed measure of a student's overall reading ability.
Key Takeaways
- Formative reading assessment is defined by purpose and timing — adjusting instruction in real time — not by any specific format (Wiliam & Thompson, 2007).
- Three check types cover most needs: quick comprehension questions, retell prompts, and vocabulary-in-context checks, each targeting a different comprehension skill.
- Placing checks at natural pause points in a text's structure produces a more honest signal than arbitrary interruptions.
- The prompt needs to explicitly request ungraded, conversational framing, or generated questions default to quiz-style phrasing that doesn't fit formative use.
- The real value is in acting on the result immediately — a formative check that doesn't change the next few minutes of instruction isn't functioning formatively.
- AI can generate check questions tied to a text's structure while holding a class profile's grade level constant; deciding how to respond to what the checks reveal stays a teacher's real-time judgment.
Frequently Asked Questions
What's the difference between formative and summative reading assessment?
Formative assessment checks comprehension during reading to adjust instruction in real time, usually ungraded and low-stakes. Summative assessment evaluates learning after a text or unit is complete, usually graded. Both are legitimate — they serve different purposes at different points in the lesson (Wiliam & Thompson, 2007).
How often should formative reading checks happen during a lesson?
At natural pause points in the text — chapter breaks, scene shifts, or major plot turns — rather than at fixed intervals like every five pages. Two to four checks across a single reading session is typical, depending on the text's length and structure.
Should formative reading checks be graded?
No. Grading a formative check changes how honestly students respond and shifts its purpose away from real-time instructional feedback toward evaluation, which undermines the format's core value.
Can AI generate formative reading questions from any text?
AI can generate questions from a described or provided text, but the prompt needs to specify pause points and explicitly ungraded, conversational framing — a generic request tends to produce quiz-style items meant for a single end-of-text check rather than checks distributed throughout the reading.
Sources
- Wiliam, D., & Thompson, M. (2007). Integrating assessment with learning: What will it take to make it work? In C. A. Dwyer (Ed.), The Future of Assessment. Erlbaum.
- NCTE (2019). Formative Assessment and Reading Comprehension: Position Statement. National Council of Teachers of English.
For the broader landscape this format sits inside, see AI Quiz & Assessment Generators: The 2026 Teacher's Guide and How to Create Spanish Quizzes With AI. For related formats, see Generating Fill-in-the-Blank Quizzes for Spanish With AI, Generating True/False Quizzes for Reading With AI, Generating Question Banks for Spanish With AI, and Best AI Study Guide Generators in 2026.