Randomization of mixed design | XM Community
Question

Randomization of mixed design

  • 30 April 2024
  • 1 reply
  • 21 views

Badge +1

Hi Everyone,

I am looking for some guidance on how to implement randomisation for the following experimental design: 

There are 6 evidence conditions (within), 6 policies (within), and 36 stimuli (between) (i.e., one stimulus for each evidence condition and policy combination). I am trying to implement randomisation for evidence condition and policy such that each participant is exposed to all 6 evidence conditions and all 6 policies, but only one policy across per evidence condition (i.e., 6 of the 36 stimuli). In other words, if a participant is exposed to policy X in evidence condition Y, they will not be exposed to evidence for policy X in any other evidence condition. Additionally, I want to expose participants to the 6 evidence conditions in random order so that each policy is evenly selected within each evidence condition. 

I set up some test blocks to try and figure this out (see example with 2 evidence conditions and 3 policies below).

However, I cannot figure out how to implement my plan for randomisation (see attempt below) as randomizers do not allow display logic (to my knowledge) and it is not clear how I might use question randomisation/display logic instead. 

 

I have looked at some similar posts about mixed design radomisation; however, I cannot figure out how these apply to the current example. Any help is greatly appreciated! 


1 reply

Userlevel 5
Badge +12

Hi @akrodger You can use survey flow elements like groups, branch logics, embedded data to resolve your query, please refer to a similar post :- 

 

Leave a Reply