LimeSurvey Cloud
Versão 5.3.24
Fruity Theme ( fruity)
==================
Hello eveyone,
Can anyone help me on how to setup cross quota, based on 2 discrete variables:

Var 1
Value A
Value B
...
Value H
Value I

Var 2
Value A
Value B
...
Value R
Value S

My goal is to ser a quota for:
VAR1.A X VAR2.A
VAR1.A X VAR2.B
...
VAR1.A X VAR2.S

///

VAR1.B X VAR2.A
VAR1.B X VAR2.B
...
VAR1.B X VAR2.S

///

...153 quotas in total.

Thanks in advance!

Add both answer options to your quota.
 

Joffm

Thank you for your assistance, Joffm.

Hey @Joffm,
I've 153 quotas to input manually.
There is any way or any strategy to automate this process by uploading a .csv or txr .file with all quotas designed?
Something similar to the Quick CSV Report about quotas' progress but this time as system input.
Thank you.

There is no bulk import for quotas - it is very unusual to have more than 2 or 3.

I've 153 quotas to input manually.

You CAN do something.
I never did this, but you may give it a try.
Export the survey as *.txt with one of your quotas.
Open in Excel or OpenOffice (TAB-separated) and see how the quota is defined (class "QTAx"). Then you may copy and adapt it as many times as you need and import the file again.

BUT: 
Do you really need this amount of quotas?
What for?
Do you have to pay for completed surveys and want to avoid unnecessary responses?

Otherwise I do not see a real reason for that in your scenario.
Store all responses and after you exported your data, remove the overquota.

OR: Enter manually.

Joffm

Thanks Joffm.

There is definetly a way of doing this by the method you're suggesting.

It seems that info from quotas are not written in the .txt file IF they have no answers added (at least I can't find them). Like this one, 155, I can't find it / them:

 

But those that have added answers can be found in the .txt file. Like this 102:

 

A.
There is a reference to the quota at the end of the file (n set to 9). Those reference numbers 123 / 131 seem to have no direct relationship with any other detail in the .txt file.

 

B.
In the .txt file, quotas appear allocated to a QTAM class (all my quotas are nested quotas) at the beginning of the file.

In my case of nested quotas, it seems the system saves the two references at different locations: [example for quota 124] the code @ var1

 

and the code @ var2

 

However, I can't bridge with these 2 reference points (at B.) with the information about the quota at the end of the file (A.)

Okay! This was not conclusive, but it's a good starting point for the next time I will deal with this issue. I'm in a rush and my design is a bit messy at the moment because I'm redesigning quotas instead of doing this from scratch. Next time I will start with one quota -> download the txt -> add another quota -> download txt again -> explore variations -> implement quotas through this method.

Thanks and feel free to drop any additional inputs.

@tpartner
"it is very unusual to have more than 2 or 3. "

@Joffm
"BUT:
Do you really need this amount of quotas?
What for?"

Let's say that you've this sample plan:


How do you guarantee that you'll reach the right amount of sample for each point, without specifying one by one as a quota?

Thanks for your inputs.

A quota does not guarantee that you get 10 or 40 responses of certain combinations.
They only limit. You won't get more.
So what do you win if you screen out respondents with a combination that is full?

In any case you have to wait until all combinations are filled with your desired amount of responses.

Now it is easy to remove the overquota.
You may set a quota on the "forbidden" combinations.

But as you use version 5.x. there might ne a way by using the plugin "statFunctions".
count the sum of each combination, compare with the actual combination of the respondent in an equation and screen out if necessary.

This way you only need one quota.
And the equation you can generate easily in Excel or a text editor.

Tomorrow I'll show something.

Joffm

 

That limit is also relevant to me. It doesn't increase cost when I use it on tel. interviewing. I do control the screening stage through the contact list I've available. Not 100% accurate, for sure.

Always learning, mate! This project in specific can have this kind of mechanic, but I've another one, 100% online, where I'm spending my response balance on screenings that gives no value back. I look forward to seeing your suggestions for tomorrow. It sounds promising.

I've read shortly about the plugin "statFunctions". But still had no time to incorporate it in my practice. Thx!

It doesn't increase cost when I use it on tel. interviewing

So, lack of information. 
 

Also. But it's more a problem of systems integration.

Like all things is life, we've to take it one step at a time. Rome wasn't built in a day