Randomization

Randomization in computers is actually "pseudo-random" – it is based on an algorithm that when given a seed will return a seemingly random order. Normally, time is used to provide the seed-aspect. However in Authoring the responseid value is used (this being the unique system id that is assigned to a respondent the first time he/she opens the survey). This means that the order for a given respondent will be the same every time such a randomization takes place. For example, if the order of some questions on a page is randomized and the user re-opens the page, or clicks Next and then Back, the order of the questions is preserved for that user.

The same randomization pattern is also used when an answer-list is reused in several questions. For a given respondent, the answers listed will therefore be presented in the same randomized/rotated/flipped order for all questions when you use the same answer-list. It is the number of items in the answer-list that determines the randomization pattern.

Authoring has several areas in the functionality where "answer-list randomization" is used. These areas include answer list display orders for all question types, column display orders for 3D-Grid questions, loops and blocks, rotation and flip.

Rotation displays answers in a rotated order – for example 3,4,5,1,2, and the start number will be random. Flip displays answers either incrementally or decrementally (i.e. either 1,2,3,4,5 or 5,4,3,2,1) – which of the two orders is selected for a particular instance is random.

If the user wishes to create his/her own randomization, or for example randomly pick x number of respondents, this can be done through scripting with the Math.random() method.