Survey Question Order Randomiser

What this generator does

Shuffles the middle and leaves the ends alone. A list shuffler moves everything, which for a questionnaire means consent lands in the middle and the survey is unusable. This treats anchors as positions rather than items: they are placed first and the shuffled body fills the gap between them, so an anchor cannot end up in the wrong place however the shuffle lands. Block mode goes further, shuffling the blocks as units and the items within each one, so a question never leaves the block it was written for.

How to use this tool

1. Paste the questions you want shuffled, one per line.
2. Put consent and screening in the pinned-to-the-start box, demographics in the pinned-to-the-end box.
3. Add block labels line for line if some questions must stay together.
4. Shuffle, and copy the order into your survey tool.

Understanding the controls

Questions to shuffle

One per line. Blank lines are dropped, and a question listed both here and as an anchor is used once, in its anchored position, with a note saying so.

Pinned to the start

Consent, screening and anything that must be answered before the respondent sees the rest. They appear in the order you list them.

Pinned to the end

Demographics conventionally go last, partly because they are dull and partly because asking them first can prime the answers that follow.

Block per question

One label per line, matching the questions line for line. Only used when the block option below is ticked.

Keep each block together

Shuffles the blocks as units and the items inside each one, so a validated scale stays intact while its position in the questionnaire varies.

Worked examples

6 questions, 1 pinned first, 2 pinned last

Nine items: the pinned one, six shuffled, then the two pinned in the order given.

Three blocks of items

The blocks come out in a random order, each one unbroken, with the items inside each shuffled.

A question listed twice

Kept once, in its anchored position, with a note explaining what happened.

Pressing shuffle again

A different middle order every time; the pinned items never move.

Common use cases

• Building several question orders for a paper questionnaire
• Working out a randomised order for a platform that has no randomisation feature
• Keeping a validated scale intact while shuffling everything around it
• Producing a fixed alternative order for a second wave of a survey
• Checking what a randomised version of your draft actually looks like

How this generator works

The anchors are removed from the body and placed at the two ends first. The remaining questions are shuffled and dropped into the middle, so the anchored positions are never candidates for a swap. In block mode the body is grouped by label, the group order is shuffled, and each group is shuffled internally — which is why a block occupies one contiguous run in the result. The output is then checked against the input: it must be a permutation of the same items, the anchors must be in their stated positions, and each block must appear as a single unbroken run.

Randomness and fairness

Randomness decides the order of the body and, in block mode, the order of the blocks. The anchors are never drawn — they are placed. Pressing shuffle again gives a different middle every time, and there is no seeded mode because a reproducible questionnaire order is not usually what you want.

For how randomness is produced across the whole site, see how Generate Random works.

Assumptions this tool makes

• Every question in the body can sensibly appear anywhere between the anchors, unless it has been given a block.
• The block labels line up with the questions line for line.

Limitations and good to know

• It produces one order at a time. Genuine per-respondent randomisation has to happen in your survey platform — a single random order for everybody does not remove an order effect, it just picks a different one.
• It does not randomise the response options within a question, only the questions themselves.
• It has no view on whether your questions are well written, leading or double-barrelled.
• It does not know which of your questions are logically dependent on each other; a follow-up that only makes sense after another must be blocked with it.
• Each order exists only until you shuffle again. Copy any order you intend to use before generating the next one.

Common mistakes

Shuffling a follow-up away from the question it depends on

Give both the same block label and tick keep blocks together. A shuffle has no idea that one question only makes sense after another.

Using one random order for the whole sample

That replaces a known order effect with an unknown one. Where your platform supports it, randomise per respondent; where it does not, produce several orders and rotate them.

Forgetting to pin consent

It will be shuffled into the middle like anything else. The anchors exist precisely because a plain shuffle does the wrong thing here.

Practical tips

• Generate three or four orders and rotate them across your sample if per-respondent randomisation is not available.
• Keep a validated scale in one block — separating its items can invalidate the published norms.
• Read the shuffled version through once; randomisation often exposes questions that only made sense in the order you wrote them.

Privacy and your data

Your questions, anchors and block labels stay in your browser. Nothing is uploaded, nothing persists between visits, and nothing is written into the page address — which matters here, because draft survey questions are often confidential. Analytics records that the tool ran and how many questions it ordered, never any text.

Frequently asked questions

Why not shuffle everything?

Because a questionnaire has items that must not move. Consent before anything else is usually a requirement rather than a preference, and demographics are conventionally last because asking them first can prime later answers.

Does one random order remove order bias?

No. It replaces the order effect you designed with one you did not. Randomising per respondent spreads the effect across the sample; a single order simply moves it.

How is this different from the counterbalance generator?

That produces a small set of orders with a formal balance property for an experiment. This shuffles a long list around fixed points, which is a questionnaire problem rather than a design-of-experiments one.

Can I randomise the answer options too?

Not here. Option order matters as well — particularly for lists of items — but it is a per-question setting that belongs in your survey platform.