Misconceptions report
Marks tell you how your class did. The Misconceptions tab tells you why. It groups the wrong and partial answers across your class into common patterns, so you can see which mistakes are widespread — and target your re-teaching accordingly.
Getting there
Open any assessment with marked submissions and click the Misconceptions tab. (Hidden until you’ve marked at least one submission.)
Misconceptions analysis is part of assessment insights, available on any active plan or free trial. On a top-up-only balance it’s locked — see What’s included on your plan.
You need at least 10 marked submissions
Misconceptions look for patterns across a class, so they need at least 10 marked submissions before they can surface anything meaningful. Below that, the tab tells you how many you currently have and asks you to mark more.
Generating the report
Unlike the other insight tabs, misconceptions are generated on demand rather than calculated instantly. Click Generate misconceptions and Marking.ai runs an AI pass over your class’s answers. This usually takes 20–60 seconds depending on how large the assessment is — you can stay on the page or come back later. You’ll get a toast when it’s ready.
Already got results and want a fresh pass (for example, after marking more submissions)? Use Regenerate. This replaces the current analysis with a new one.
What you’ll see
Results are laid out as cards, grouped two ways using the toggle at the top:
- By question (the default) — misconceptions grouped under each question number
- By theme — misconceptions grouped by category
Each card shows:
- A short description of the misconception
- A category chip (e.g. the type of error)
- The question number it appears on
- A student count badge — how many students showed this pattern
- Show examples — expand to read a few real answers that illustrate it
The cards are a read-only summary. There aren’t clickable student or question links on them — use the count and examples to decide where to dig in, then open the individual submissions from the Submissions tab.
Use the cards to plan a re-teach, a follow-up question, or a class discussion.
How Marking.ai groups misconceptions
Marking.ai looks across every wrong or partial answer in the assessment and clusters similar mistakes together. Two students writing different words but making the same underlying error end up in the same group.
This isn’t deterministic — it’s a pattern Marking.ai surfaces, not a definitive taxonomy. Use the groups as a starting point, then read the example answers (and the underlying submissions) to confirm.
What if I don’t see any misconceptions?
A few reasons this can happen:
- You have fewer than 10 marked submissions — the tab will say so.
- Your class genuinely answered consistently, with no clustering of errors worth flagging.
- You haven’t run Generate misconceptions yet — nothing appears until you do.
Need help?
Open the Account menu and choose Chat to a human, ask Marko, or email connect@marking.ai.
FAQs
How does the AI decide what counts as a misconception?
It looks for repeated patterns in the answers that lost marks — wrong reasoning, partial answers, confused concepts — and clusters similar instances together. It’s a heuristic, not a guarantee.
Do misconceptions update automatically as I mark more?
No — they’re a snapshot from when you last generated them. After marking more submissions, click Regenerate for a fresh pass.
Can I edit the misconception cards?
Not directly. If a card looks wrong, the underlying student submissions are the source of truth — review those and regenerate if needed.