Did you give the same randomization group name to all questions? From your description you seem to give the randomization group name only to groups. But you don't want to randomize the groups, you want to randomize the questions. I know it is confusing that there are question groups and that there are randomization groups and they are totally different things.

First of all: Take everything out of the randomization group field for the two QUESTION groups. Give the two QUESTION groups different names (you called them both "Grammar"). Then give all the questions in the two question groups the same name in the "randomization group name".

If that instruction doesn't work: Upload what you have done as an example (please only these two question groups that are relevant for this issue).

Much better for the helpers if you do an upload of your problem, rather than us to create an example and upload. We do this all in our spare time and the easier you make it for us to help you, the better.

Here it comes. So I want that every participants gets 1 random question - that's why I created 1 group with 1 question and another one with 2 questions - so it should be picked out of a pool of 3 questions.

Thanks in advance! I use Version 3.1.0+180124

As previously stated...

1) Give the second group a relevance of 0.

2) DO NOT use the "Randomization group" field for groups. (see attached screenshot)

3) DO use a common value in the "Randomization group name" field for all questions. (see attached screenshot)

Sample survey attached:

You obviously did NOT follow the innitial instructions:

hide the second group (relevance equation = 0)

You gave 0 to the questions, not the group!

then make sure all 50 questions are part of _the same_ randomization group
the questions will get random order _between_ the two groups

Attached a version of your test. I deleted the final question which was irrelevant for the problem and switched of the welcome text, which is taking valuable time when testing this specific issue.

Called the survey 10 times for test purposes, received 4x Text1, 3x Text2 and 3x Text3. Seems to work perfectly.

And once again, Tpartner beat me again. However, why did I not see his answer before. It didn't take me 50min to solve this!!!

holch wrote: However, why did I not see his answer before. It didn't take me 50min to solve this!!!

You posted 2min after Tony. Nobody to blame. Since you can edit your posts, why not post a placeholder?

Nope, my second post was 2 minutes after my first post.

Tony's post was "2 hours 35 minutes ago", mine "1 hour 35 minutes ago".

I guess his answer was just on the next page and I did not notice that he had already answered and started to solve the problem as well.

Why would I post a place holder? To "reserve" the first position for my answer? Hahahaha.

Not that big of a deal. Tony has a lot more practical experience with LS, so he will beat me any time, I guess. My only chance is getting there way before him and get a headstart...

I was following this post since morning and I could not get it right.
Until after lunch. I changed the General Options - Format to "All in one"
and it worked!

I don't know what I have been missing but maybe this can help.

Thanks guys. This post is very helpful to me.
I am using v3.2.1

This seems to work. However, the total count of questions at the start of the test and at the progress bad seems to consider all questions, even the ones on the relevance 0 groups.

Is there a way to fix this? Or maybe a newer and updated method for doing this random questions thing?

The counter at the start considers all questions in your survey, no matter their relevance (which in many cases can change, depending on your relevance equations). My solution would be to either put my own count or leave the count out completely. Because in my opinion it doesn't help much anyway.

You can have a survey with two questions, both with enormous lists of answer options, e.g. 100 each. This will take significantly longer to answer than a survey with 20 YES/NO questions.