Translations of this page:

Randomization: within-subject (Measurement Repetition)

Measurement repetitions are used in quite different designs. Here you will find an overview and the appropriate technical solution in each case:

Classic Measurement Repetition

In a within-subject design, in the simplest case, all respondents receive the same questionnaire. With such a design, all you need to do is duplicate the questions or sections so that they can display a different question (with the same content, if applicable) for each stimulus.

Measurement Repetition with Rotating Sequence

It gets a little more challenging when all respondents are given the same stimuli – but their order is rotated between-subject. How this works you can read here: Rotation

Measurement Repetition with Part Sets

If you want to present only a part of the stimuli to each respondent, you can also use the solutions for Rotation of pages. In this case, set the random number generator to draw only a certain number of slips in each interview. So for example 3 of 10 stimuli.

Work with a Large Number of Stimuli

Normally you create a question or a set of questions for each stimuli. For stimulus 1, for example, questions AB01, AB02, and AB03 would be asked. For stimulus 2 the questions BB01, BB02 and BB03. By copying rubrics (Manage Questions and use as Templates) such question sets are quickly duplicated. For a handful of stimuli, this is still reasonably implementable – but if you have 10, 20, or even more stimuli, it won't give you any more meaningful data sets. In this case you use the Multi-Level Structure.


A special case of within-subject design is a conjoint analysis. To implement such a design, please follow the instructions Multifactorial Experimental Design.

en/create/randomization-within.txt · Last modified: 03.03.2021 20:04 by sophia.schauer
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International
Driven by DokuWiki