Randomizing order

Randomization changes the order of answer options or statements. It is configured on the question; there is no general “shuffle the whole survey” control. To send respondents to different versions, use Split respondents between paths.

Question	Setting
Multiple Choice, Checkboxes, Dropdown	Randomize Order is visible, but does not currently shuffle the respondent view
Likert	Randomize Order for statements
Ranking	Randomize Initial Order
Matrix	Randomize Row Order and Randomize Column Order

Select the question and open Content to find its setting. Preview several fresh runs to inspect the presentation. Do not rely on every run producing a visibly different order.

Branches and pipes refer to the underlying options, so changing their displayed order does not require rewriting a condition. Still test questions that combine randomization with conditional logic.

Choose what you are trying to vary

Goal	Use
Present ranking items in different initial orders	Ranking's Randomize Initial Order
Change the order of Likert statements	Likert's Randomize Order
Vary Matrix presentation	Its separate row and column randomization settings
Show one of several survey versions	A Split rule in Conditional Logic

These choices answer different design questions. Randomizing a ranking's initial order changes where items begin; the respondent's final order is still their answer. A Split determines which route they receive and does not shuffle all of that route's questions.

Example: workshop topics

For a Ranking question containing Research methods, Data analysis and Presenting results, enable Randomize Initial Order when no initial ordering is intended. In Preview, place a known topic first, continue, and check a later top-ranked-item pipe. It should reflect the final ranking the respondent submitted, not the position the topic started in.

Verify the settings you depend on

Run several fresh previews and inspect the relevant items, statements, rows or columns. A repeated order can occur by chance; one repeated pair of runs does not establish a failure. Conversely, a setting being visible does not prove it is applied: the choice-question limitation above matters when planning your survey.

Test alongside branching and piping, and record which kind of randomization your survey uses. For weighted assignments and the difference between Balanced and Random, see Split respondents between paths.

1234

Enlarge screenshot