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:
==================
Hi,

I programmed the following alternative case in LimeSurvey:

I draw a random number between 1 and 100 :

if(is_empty( randvalue ), rand(1, 100), randvalue.NAOK )    

[Question code: randvalue, question type: equation]

Then, if this number is strictly lower than 4, I want the variable DIS to be equal to 'YES' and if this number is equal or higher than 4, I want the variable DIS to be equal to 'NO':

{if(randvalue.NAOK<4,'YES','NO')}

[Question code: DIS, question type: equation]

The two questions belong to the same group. Then, I have additional equations in the group that depend on whether DIS='YES' or DIS='NO'.

The problem that I get is the following:

when I test my survey inside LimeSurvey, I sometimes get a randvalue <4 and in that case, I get on my screen DIS='YES'. Then, on my screen, all the following equations in the group that depend on whether DIS='YES' or DIS='NO' are correctly calculated.

However, when I send the results and go look at them in the answers part of my LimeSurvey, the randvalue that appears is indeed lower than 4, but DIS='NO' (instead of 'YES' in my screen when I tested the survey) and all the following equations are calculated based on the alternative DIS='NO'. Therefore, all the calculations are different from those that appeared on my screen.
See the last line in the attached image, randvalue is equal to 2, but DIS is equal to 'NO'.  

How come that the corresponding results on my screen when running the survey and those recorded in LimeSurvey are different and the latter are wrong?

This seems very odd to me and I don't know what to do, because when I run the previous code in LimeSurvey, everything works well, this is just not the correct results that are recorded in LimeSurvey.

Many thanks for your help!

Not sure if this is the issue, but I would take out the ".NAOK" part. This stands for "No Answer is OK". But in your case, there will always be an answer, at least in this part of your equation. The NAOK is not necessary. Not sure if it can cause trouble though.

Ideally it would be great to send a LSS file with just the relevant questions, so that we can have a look at the actual survey and do some tests.

holch wrote:

Not sure if this is the issue, but I would take out the ".NAOK" part. This stands for "No Answer is OK". But in your case, there will always be an answer, at least in this part of your equation. The NAOK is not necessary. Not sure if it can cause trouble though.
 

I really think it can not break with NAOK. Always add NAOK personaly.

NO idea on issue here.

Thank you for your answers.

I have tried to suppress NA.OK and I still get the same problem.

I have also tried to have " " instead of ' ', in {if(randvalue<4,'YES','NO')}, what's the best?

I have also tried to suppress is_empty in {if(is_empty(randvalue),rand(1,100),randvalue.NAOK} because I don't have previous reference to randvalue.

When I run a minimal example, it works (see the two questions lsq attached). But when I run my whole survey, I always get a NO in the recorded answers in LimeSurvey even though it displayed a YES in the screen when I executed the survey.

I really don't understand what is going on here. What could explain that when recording the answers, the results are not the same than on the screen? Is it some kind of bug?

Many thanks for your help.

 

Please, neither lsq nor lsg exports.
Only lss.
Copy the survey, remove everything not related to the issue, send this lss.

Joffm 

Here you are.

The thing is that the minimal example is working whereas the whole survey is not, so the minimal example is not really instructive.

Thanks

PG0124 wrote:

When I run a minimal example, it works (see the two questions lsq attached). But when I run my whole survey, I always get a NO in the recorded answers in LimeSurvey even though it displayed a YES in the screen when I executed the survey.

Many thanks for your help.

 

Check if you have a condition with only one =

Check survey logic file :
 

You must have one warning only (on your equation question type)
 


 

Thanks.

I have checked and I don't have such warning messages or any missing = in the equations.

Should the hidden equations be made mandatory or can they be optional?

For randvalue and DIS, I have :hidden1clear_defaultNsave_as_defaultN
Could it be related to my problem?

Thanks a lot!!

You check whole survey logic file ?

And about warning : you're right : no warning the way you do it …

Yes, I did check the whole survey logic file.

I really don't know what else I can do.

It is as if randvalue is well recorded in the LimeSurvey database but not DIS.

So the problem should be with the DIS code.

Try intval(randValue)

The thing is that the minimal example is working whereas the whole survey is not, so the minimal example is not really instructive.

Well, than you should send the whole survey

Joffm