Balanced Design
Last updated
Last updated
(c) 2024 Data Illusion Zumbrunn. All rights reserved. Reproduction strictly forbidden.
The balanced design is an experimental setup where each level of every attribute appears an equal number of times across all choice sets presented to the respondents. Each choice set is referred as a task for the respondent. This balance ensures that the design is not biased towards any particular attribute level, allowing for a fair and accurate estimation of the relative importance and utility values of all attribute levels being studied.
In the example below the respondents will be presented with 4 choice sets resulting in 4 tasks to accomplish by choosing for each task 1 choice out of 3 choices. In order to cover all the possible versions to have an accurate analysis you would need 330 respondents to answer this question.