Hey guys,


I am fairly new to lime survey but have some experience with other survey tools. I wanted to set up a very simple and quick survey but I did encounter that some answers are not recorded, they simply dont show in the response for some questions.

I am using Version 3.21.3+191219 via my University.


I had 17 questions, one single condition (if this answer "yes", then skip the next question). I run a pretest with 9 participants and figured that for some questions answers are always recorded, for some they are never recorded and for some others sometimes yes, sometimes no. The survey itself runs without issues.

I tried quite a bit to fix it but was not successful yet: I even set up a completly new survey with only five demographic questions (age, gender, race, etc.). It has only one question group, only a single langue, no conditions, all questions mandatory. Still I couldnt track the problem: for the question on ethnicity (List (radio)) the answers are not recorded or not showing up in the response section. The question is basically not different to the race question. I also had an issue with an Array question before.

I very much appreciate your help since I am currently lost. I hope I just missed on something simple.


Thank you!

Hi,
what do you expect us to say, if we don't know your survey and the settings?

for some questions answers are always recorded, for some they are never recorded and for some others sometimes yes, sometimes no

You see, this is no information at all.

So, please attach a lss export of your survey and say exactly which answers were always stored and which were never stored.

Joffm

Dear Joffm,

thank you very much for the quick reply and appologies. I thought it might be something like: "Ohh that sounds like this typical but rare mistake beginners always do, you just have to press this button".

Attached you can find the lss file for both surveys. I will first report the question IDs of Survey A (the longer one). Green worked, meaning all answers recorded, orange means some answers were recorded (to a varying degree and not consistent accross participants, sometimes an answer is missing for participant A, sometimes for participant B, etc.), red means that the answer of noone was recorded, grey indicates that I cannot tell because question was not mandatory.


In order how they appear in Survey A:

• 22760
• 22656
• 22657
• 22658
• 22761
• 22762
• 22763
• 22769
• 22768
• 22764
• 22766
• 22778
• 22781
• 22782
• 22783
• 22786
• 22780

For the shorter survey B, everything seems to work, except question 22814.

Survey A was run by 10 random participants and a couple of times myself, survey B ~20 times by myself from different browsers and computers.

Finally, please also find attached a screenshot of how the answer csv looks like. Like a swiss cheese. The online result interface has the similar look.

Thank you so much! I very much appreciate your help.
MiRe

Are you really sure that you use a 3.x version?
I could not import your surveys into 3.x, only into 4.x.

But I could not reproduce anything in 4.1.4.


Joffm

Hey Joffm,

you are right, it gives me an error as well. I must have done something wrong when renaming after exporting. The two new files attached work. The version number should be correct, at least that is what is displayed to me in the lower right corner at lime survey.

Also, I realized that the question IDs change when importing. So here a new list with the question codes that should stay constant after importing the lss:

For survey B, some answers are missing for question with code Ethn
For survey A, again from start to end:
consent, Age, Gender, Ethn, Race, Income, Education, ParEduc, Zip, Community, Country, ParStatus, KnowCorona, Numbers, Corona, Affiliation, Errors

Thanks for helping me! Appreciate your effort

Someone maybe an idea?

Unfortunately the problem still exists. Although all mandatory, for some questions answers are not always stored Hence, Lime survey is completle unfuctional for me right now.

I am also not able to import the survey.

I would copy it and start removing groups until you get a viable export - that may lead you to where the problem is.

Thanks for coming back. Yeah, the first uploads are unfortunately corrupted - I made some kind of mistake. But the second ones should work (in my 3rd post). I attached those lss files again.

Also, in survey B there are no question groups (if you mean this by groups). There are only 5 simple questions from the drop down of ready made question types, no conditions, no programming, one language, and all in one group.

I really appreciate your help. Thank you!

I have no problem when testing survey B.

Maybe there is an issue with your template(theme). Try with one of the core themes.

Thanks @tpartner !

Unfortunately I am using my university's theme and it seems I have no rights to change anything regarding themes. However, I registered a new and free normal account and tested the survey there and you are right, it is perfectly running. It very much seems as my university's theme is corrupted. I will write the IT guys of my university asking to fix it.

Maybe any hints how I could fix this myself in the meantime?

Thank you very much for hinting me in the right direction!

Maybe any hints how I could fix this myself in the meantime?

No, not if you are not allowed to switch to a core theme. (I would also tell the IT folks that that is a ridiculous policy - classic case of IT forgetting who their customer is)

Hey guys,

after almost two weeks of trying to fix it, I now got it working. Just for the record, I wanted to share my rather disappointing and exhausting experience. It appeared it was not the theme. Even when I was able to switch to other themes it did not work. Also playing around with and testing many other options (mandatory questions, question group by question group, cookies switched on or off, etc. etc.) did not help.

However, in the end it seems it was a simple thing: My university's IT-guy updated the limesurvey version that is running at my university (now Version 3.22.10+200323). After that, everything was recorded properly - at least with my current settings; I had no need to play around and test anything else.

I hope this might, in the future, be of help to someone else.

Cheers!