LimeSurvey version: Version 5.4.3+220926
Own server or LimeSurvey hosting:  umfragen.uni-oldenburg.de/
Survey theme/template: vanilla
==================

Hello, 

I have a semantic differential array type question with 33 subquestions. I have a survey structure as follows: 

Question Group1 (rand)
a related question to group 1
array question with randomized display of subquestions

Question Group 2 (rand)
a related question to group 2
array question with randomized display of subquestions

Question Group 3 (rand)
..
Question Group 6 (rand)

When I run the survey in preview mode it works as I want it: The question groups are viewed in randomized order and array subquestions are also in randomized order in each question group.  When I activate the survey the array question with randomized display of subquestions gets only shuffled once and the same shuffled order repeats through all question groups. How can I solve this problem? 

Kind regards

Hi,
please attach a lss export of a prototype of your survey, so we can reproduce (or not)

Joffm 

Dear Joffm, 

Thank you for your answer. Here is a mock of my survey structure. Same problem occurs in this one too. Preview is fine, activation kills the randomization of array subquestions after the first shuffle. not sure if it helps, but I copy the questions to re-use it in each group.

Kind regards, 

Dear Joffm,

Could you get a chance to take a look at the prototype?

Kind regards,
lile

Hi,
yes, I can confirm this.
Now, the question is: Is it a bug or is it a feature?

I remember some requests to keep an initial randomized order throughout the whole survey.
So you may file a bug report to see what the developers say.

But in my opinion it is a useful feature.
Example:
You ask several questions about car manufacturers, like
Alfa Romeo
FIAT
Citroen
Opel
Ford
...

1. In your opinion which company was founded first?
2. In your opinion which company had the highest revenue in 2020?
3. In your opinion which company sold the most cars in 2010?

In this case it is really annoying and confusing for a participant to see always a different order of the answer options.
Once he sees
FIAT, Ford, Opel,...
Next question
VW, Peugeot, Ford,...

And if there are different answer options, well, the participant doesn't notice it at all.

Main reason for randomization is to avoid a bias by order. And this you do by displaying a new randomized order for each new participant.
This is my opinion about this.

But as said before: File a bug report.

Joffm