Please help us help you and fill where relevant:
Your LimeSurvey version: [see right hand bottom of your LimeSurvey admin screen]
Own server or LimeSurvey hosting:
Survey theme/template:
==================
Hello, I have an issue in my survey. When I preview my survey and look at the question UGorder, a multiple numerical input question that I sue to track the order questions in my survey, I sometimes get NAN as an answer. I thought the issue was with the default answers I set. However, when I preview the question on it's own or preview the questions who's value are referenced in UGorder's default answer, I get no NAN message. The question Faceorder is very similar to UG and yet does not have any issues.

Is it possible to solve it, if so how ?

I use  LimeSurvey Cloud
Version 3.28.0

Hi,

I sometimes get NAN as an answer

Did you try to see a pattern where there sometimes is a NAN ?
You see, the timestamp of each question is stored.
And furthermore you the order of the images is stored in QH1-QH6.
So you should be able to calculate the order of the questions and the value that should be saved, but is not. 
Edit:
I saw that you set a randomization group name on the questions of UG.
Why? You already randomized this by the selection of the images in QH1-QH6.
One and a half year ago I showed that with this javascript selection the randomization by randomization group is not necessary.


Personally I do not see a reason to randomize all 5 groups (which causes your "track order" problem)
In my opinion it would have been sufficient to keep UG at a fixed place and randomize the other four around.
Like:
BPNI  : Randgroup0
ACME  : Randgroup0
UG groups
BSL  : Randgroup0
FACES  : Randgroup0

Now you could have splitted UG into several groups each containing the three related questions
Like
UG_AF37 : Randgroup1
UG_HM19 : Randgroup1
...

Everything would have been fine.

Joffm
 

It seemed to occur when we set ''Always hide'' on OffsetUG, it seems to work fine now that we simply hide OffsetUG with javascript.

To answer your other questions, I am in the same lab. We had proposed the solution of making the UG multiple groups but our advisor turned it down. His reasoning was that, to avoid priming effects (we are a social psychology lab) we would have to create multiple surveys placing the UG and Face Recognition in different places and it would be preferable to have just one survey for data analysis and to avoid having the same participants take part in multiple versions of the same survey, unknowingly.

The reason we use QH1 type questions for the faces but not for the Offer pictures was that when we used the method you had originally proposed over a year ago there was no means to guarantee that, for instance, Angry Male faces wouldn't, through randomness, appear disproportionately with a specific type of offer. So within the question of UG, the offers remained the same but the faces varied while remaining in the same category (Neutral Male, for instance). It was important for the sake of the experiment that we would be able to separate the effect of offer type from the effect of face type and that all participants perceived the face/offer combinations for the same amount of time, hence the question codes like AF37, Angry Female who proposes to give 0.3$ and keep 0.7$.

So we opted for the solution you cooked up with the Offsets a few months ago instead. There may have been a better solution.

thank you for your help