Due to quite frequent requests in the forum, I would like to see the possibility to randomize questions, question groups and if possible answer groups (which do not exist yet, but see www.limesurvey.org/en/support/forums/20-...option-group-headers ).

It should be possible to define a group of questions, and those question should be possible to randomize.

Example:

Q1
Q2
Q3a
Q3b
Q3c
Q4
Q5

So the different Q3 questions should be randomized, while the order of the rest stays the same:

Q1, Q2 -start randomize- Q3b/Q3a/Q3c -stop randomize- Q4

The same thing should be possible with groups of questions, example:

Q1
Q2a,b,c
Q3a,b,c
Q4a,b,c
Q5

So this could look like this:
Q1 -start randomize- Q4a,b,c - Q3a,b,c - Q2a,b,c -stop randomize- Q5.

Ideally the questions within the group should be possible to randomize as well (see above).

I agree, this is a very useful feature.

When discussing features, please always add the according idea to the idea tracker and link it here so that other users can vote for it.

This makes sense, Mazi. I will do that.

Mazi wrote:

When discussing features, please always add the according idea to the idea tracker and link it here so that other users can vote for it.

Since the idea cannot be changed after submitting (or is my impression wrong about that and it can be changed?) I see the purpose of this forum to develop the idea and submitting it after a few pro and cons have been discussed. If nobody is submitting a few words here the submission won't make any sense.

jelo wrote:

Since the idea cannot be changed after submitting (or is my impression wrong about that and it can be changed?) I see the purpose of this forum to develop the idea and submitting it after a few pro and cons have been discussed.

I just checked and you can at least edit the proposed solution. So one can outline the problem in details, discuss one or several possible solutions and edit these solutions at a later stage.

I am also interested in having the feature Holch describes. I often used the randomize_groups.pdf workaround in 1.87+, but it does not work reliably in 1.90+. (See Bug Tracker 00557 for additional detail.) Thanks!

At least the randomization of groups seems to be implemented for the upcoming 1.91 version: limesurvey.svn.sourceforge.net/viewvc/li...vision&revision=9528

Outstanding!!!

Do we know a release date yet?

This would be really great, if we would have at least the randomization of groups. This would help quite a lot already.

You should be able to already test the feature at the dev demo system: www.limesurvey.org/en/support/forums/20-...elopment-test-system

Currently there are several new minor features added by Google Code in students so it might take a while thought most of the feature are ready for a first RC version.

Hmmm, I just had a quick look at the randomization of question groups.

Either I don't get it, or the way this feature is implemented isn't really helpful for most market research applications.

From what I understood, you have for example 3 groups A, B, C (let's better say pages) and each group is supposed to have say 3 questions in it.

So you have like question Q1.1, Q1.2, Q1.3. Those are the choices for the first position in the groups A, B, C.

Then there would be Q2.1, Q2.2 and Q2.3. Those would be the choices for the second question on each page, and so on.

However, when I speak of group randomization, I want to create groups of questions (usually those are the same questions, just for different products, aspects, etc.). Then the groups are shown in a random order.

The way it is right now, I can't find a use for it for now. But of course there might be some and I just don't see it. But for most market research projects that I can imagine, the current setup isn't really helpful.

I have uploaded the randomization test survey to the devtest, so it would be good if you guys could have a look at it and give some feedback as well.

I might be totally wrong, but I don't see much of a use the way it is implemented at the moment. I think the implementation at the moment is more complicated than actually necessary.

Mazi was talking about a different thing. The feature in the dev branch is question randomization. So basically you can assign any number of questions to a virtual randomization group and within that randomization group the questions are shown in a randomized order (be aware that a randomization group is not connected to a guestion group in any way).
It is different from your desired feature of 'question group randonization', where you have question groups which have their group order randomized.