Images v1 and v2 show the same question after a reload of the page.

The problem is still present in version 4.1.17+. In some very rare cases, it might be useful to randomize answer options, but certainly not for Likert or semantic differential scales whre, precisely, randomization of subquestions if often required.
There should be 2 (two) randomization switches: one for subquestions and a different one for answer options.

I agree with YvesRoy.
a) it's still present in 4.1.17
b) two options would also be easier to understand

I think this is a serious problem. This "feature" forces a lot of people to use randomization for subquestions, because for many scales it makes no sense what so ever to be randomized. Imagine randomizing a 1-5 or 1-10 scale. Will be fun to respond, but I wouldn't trust the outcome of such a survey.

It's a feature not a bug ....

Hi,

either way. As some (psychological) scales are designed to have randomized questions, this should work.
The idea from YvesRoy to have the option to also randomize the answers is imho a (new) feature.

I tested with a 5 point Array and they answer options are NOT randomized, only the subquestions, when you choose randomization.
Now for the question type Array (Type F), both, the subquestions AND the answer options are randomized. But there is NO way to prevent the answer Options to be randomized.

A lot of people will use individual answer scales from 1-5, 1-7 or 1-10, or like shown ONG, you might have text, that needs to stay in the same order. In my area, for about 99% of all cases where I use an Array question I can't use randomization of subquestions with LS4, because it would mess up the answer options, that in most cases need to stay fix.

I like the idea of the feature to be able to randomize also the answer options, but only if you can decide if you want to apply randomization to subquestions or to answer options or to both. But as is, LS4 for probably 90% of market research purposes it is not usable with this "feature". For me this feature is a step back. If adding the possibility to randomize answer options means that it is always conected with the randomization of subquestions, then I'd rather prefer not to have this feature.

The correct (and only sensible) approach in my opinion would be two settings:
Randomization of Subquestions: Yes/No
Randomization of Answer options: Yes/No.

the subquestions AND the answer options are randomized. But there is NO way to prevent the answer Options to be randomized.

Oups ... sorry : clearly a bug here (and in 3.X : only subquestions are randomized)

The correct (and only sensible) approach in my opinion would be two settings:

Yes and here : it's a feature , but more need (for array text / arry number)

• Randomization of Subquestions X axis: Yes/No
• Randomization of Subquestions Y axis: Yes/No
• Randomization of Answer options: Yes/No.

Oups ...
That's what I said, when I discovered this "feature" in LS4. The person who did this never programmed a survey in their life. ;-)

Will fix, but in the future, please always add a survey file (lss) which reproduces the issue.

Hi,

ok, here you go: All questions (except the last one) have the decribed problem.
The last one has another problem, reported here: https://bugs.limesurvey.org/view.php?id=16227

ong

Thank you, @ong. I tested with you survey and it seems to work now. If you use the github version of LimeSurvey, you can checkout branch bug/16032-randomize-also-randomizes-answers and test yourself.

Hi ollehar,
thanks for your reply.
No, I do not use the github version (yet, and I doubt I can do it in a short amount of time).
I guess I will wait for the next release.

Sure, next bugfix release is next Monday.

Hi,
sorry, - in my case the error still exists in 4.2.3+200511

Same here. Probably didn't get into the release.

No, if status is still "testing", then it's waiting for someone in the team to test it.

Would be great if we could get this tested and included in the next release, because for most uses of array questions with randomization the current behaviour is a show stopper. Many array questions are used with scales (answer options) like 1-5, 1-7, 1-10, etc. which just don't make sense randomized.

@holch If you want to join the testing team we'd be more than happy to welcome you. :D

@ollhar: What do I have to do and what skills and setup do I have to have to be able to test?

@holch You need a local installation and git checkout <branch name.

@holch Are you available on any chat channel? We sometimes use Telegram.

I am a little bit on a war path with git. Haven't used it for a long time because I find it very cryptic to use (not a programmer and I don't use it on a regular basis, so once I get it working, I don't use it for a couple of months and then need to learn how to use it all over again. But I will see if I can get it running.

Chat channel: not really. I am showing up at the LS Team chat from time to time. I think I have Telegram installed on my phone, need to check. But here in Brazil the communication channel is Whatsapp and thus this is what I use most.

OK, well, maybe the forum can help you set it up? Are you using Windows? In any case, there are a number of graphical interfaces to git available for free.

Whatsapp is not optimal, since you can't easily use it on your computer (independent of your phone).

Fix committed to master branch: http://bugs.limesurvey.org/plugin.php?page=Source/view&id=29968

Fixed in Release 4.2.4+200520