How it works
By default, a questionnaire runs question after question, in order. Branching breaks that line: depending on the answer given, the learner is sent to another question — or straight to the end.
Two common uses:
- Remediation — a learner who gets it wrong goes through an explanation; a learner who gets it right skips it.
- Test-out — a learner who shows they have mastered a topic skips the corresponding part.
Two ways to branch
There are two levels available. The first is more precise, the second more universal.
Based on whether the answer is correct — on all questions. Under each question, next to the two feedback fields, there is a "→ go to" menu:
- to the right of the correct feedback: where to send the learner if they got it right;
- to the right of the incorrect feedback: where to send them if they got it wrong.
This is the method to prefer: it works on all question types, including free-text entries, matching, ordering and fill-in-the-blanks.
Based on the option chosen — single answer and true/false. On a single-answer or true/false question, each option has its own "→ go to" menu. So you can direct each answer differently: A to one question, B to another, C to a third.
This is useful when each mistake calls for a different explanation. On other question types, these menus do not appear: there is no "chosen option" in a free-text entry.
Steps
Open the questionnaire
Add an activity → Questionnaire, or edit an existing questionnaire.Write all the questions first. The branching menus offer the existing questions: create them before linking them.
Choose the destination
On the starting question, open the menu you want and select the target question, or End of quiz.
Check by putting yourself in the learner's shoes
Launch the questionnaire and follow each branch.
A menu that sends the learner elsewhere changes colour: you can see at a glance which answers branch.
When the two contradict each other
If an option has a branch and the question has one based on correctness, the option wins. The more precise setting wins — the one you took the trouble to set answer by answer.
The score only counts the questions asked
This is the most important point to understand, and it changes the way you design.
The mark is calculated on the questions actually asked, not on the whole questionnaire. A learner sent down a short branch is therefore not penalised for the questions they never saw: three correct answers out of three questions asked give 100%, even if the questionnaire contains ten.
The progress bar follows the same logic: it counts the answers given, and shows "3 questions answered out of 10" rather than a position in a list the learner will not go through in full.
- A "correct answer → End of quiz" branch set on the first question makes the learner leave with 100% after a single answer. That is the expected behaviour for a test-out — but a short path created by mistake gives the top mark far too easily.
- Go through every branch before publishing. It is the only way to see what your learners will really see.
- A deleted target is ignored — if the target question no longer exists, or if the random drawing from a pool did not select it, the learner simply moves on to the next one. They are never blocked.
- Watch out for loops — nothing prevents you from sending the learner back to an earlier question. This is sometimes intended (redoing a missed point), but check that the learner can always reach the end.
- Without branching, nothing changes — existing questionnaires run and are marked exactly as before.
- Multiple-answer questions have no per-option branching: since several boxes can be ticked, which one would decide? Use branching based on correctness.