Please help us help you and fill in where relevant:
Your LimeSurvey version: 5.6.18
Own server or LimeSurvey hosting: Free LimeSurvey Cloud
Survey theme/template: Default
==================
Using another Survey provider, I created a Survey that has this structure:

• Page P1 with an introduction;
• Page P2 with 5 questions;
• Group G1 of pages from P3 to P9. Each page has 4 questions related to the same topic;
• Group G2 of pages from P10 to P16. Each page has 4 questions related to the same topic;
• Group G3 of pages from P17 to P23. Each page has 4 questions related to the same topic.

The survey presented has the following logic:

• The survey shows everything one page at a time and questions are all mandatory;
• Each group appears in random order;
• Each question within the group but the last one appears in random order;
  • The last question, once answered, jumps to the end of the survey.

The purpose of all this mess is to have just one survey to submit instead of three different ones represented by each group.

I tried to emulate this survey's structure and logic on LimeSurvey. I mainly faced two problems. LimeSurvey does not have the "page" concept, and there is no intuitive way to jump to the end of the survey after a question is answered. 
To be fair, after a while, I just achieved the randomization of groups and questions.
Is it possible to obtain what was just explained on LimeSurvey? Maybe by self-hosting the community edition of LimeSurvey?

Yes, you are right.
There is no "group in group" concept.

But to be honest: I do not understand what you mean by

• The last question, once answered, jumps to the end of the survey.

Does it mean a respondent only answers one group?

You should show us your survey! Send a lss export.
So we may see something, advise something , or whatever.

Joffm

Joffm wrote:

Does it mean a respondent only answers one group?


 

Yes, you've got the point. I just want a respondent only answers the initial questions and then to just one randomly picked group. Sorry If it was not clear, I used the terminology of the other provider. 

Joffm wrote:

You should show us your survey! Send a lss export.
 

Unfortunately, I am not able to do such an export from that provider.

So, you still don't have anything developed in Limesurvey?
But, what you say, is easy to implement.
Create a random number (1-3) and display only the group according to it.

Nevertheless, create a prototype in Limesurvey, which you show us.

Joffm 

Joffm wrote:

Create a random number (1-3) and display only the group according to it.

 

Ehm... since it does not seem the most intuitive thing in the world, can I find somewhere a guide about this?

Thanks,
Nino

Is there a question that defines which of the 3 groups is shown or is this random?

Because to me it sounds like the group to be shown is determined by an answer to one of the 5 questions at the beginning.

Your description of your survey isn't very clear.

Hi, @holch,
I understood

• Each group appears in random order;

And he wants to terminate after the first group.
So it is:
Ask P1 and P2
select one group of G1, G2 or G3 randomly 

Joffm

There are many descriptions of this approach in the forum. Here a short summary:

1. You need to create an equation type question to generate the random number from 1-3 (loads of examples in the forum on ho to do this). Once everything is tested and it works this equation type question will be hidden.

2. Show / hide the 3 different questions groups via relevance equation in the group depending on the random number created in the equation question.

This would look something like this, assuming that the equation question is called random:

So the group where this is written in the relevance equation will only appear if the random number is 1. Do the same for the other 2 groups with 2 and 3 as the random number.

And should you understand German, then Joffm has 3-4 excellent tutorials in the German forum, one of which is covering exactly this approach.

Once you got the gist of it, it will take you less than 5 minutes to set this up.

So I recommend to create a simple structure of your survey with the 3 groups, etc. so that you can play around with the approach (you don't need to setup the full survey to test this).

If you get stuck somewhere, you can export your test survey as LSS and upload it here, which makes it a lot easier for us to see what your specific issues are.

Without a test survey all our explanations will be very vague.

A survey structure like the one you want to setup, depending on the details, should not take much more than 30min.